[Emerging Infectious Diseases | Volume 4 No. 2 / April - June 1998] International Editors—Update * Emerging Infectious Diseases—Southeast Asia Sai Kit Lam Perspectives * Wild Primate Populations in Emerging Infectious Disease Research: The Missing Link? N.A. Wolfe, A.A. Escalante, W.B. Karesh, A. Kilbourn, A. Spielman, and A.A. Lal * A Method Accommodating Analysis of Error Facilitates Comparison and Clustering of Molecular Fingerprints. H. Salamon, M.R. Segal, A. Ponce de Leon, and P.M. Small * Legal Issues Associated with Antimicrobial Drug Resistance. D.P. Fidler * Interactions of Rickettsial Pathogens with Arthropod Vectors. A.F. Azad and C.B. Beard * Could Myocarditis, Insulin-Dependent Diabetes Mellitus and Guillain Barré Syndrome be caused by One or More Infectious Agents Carried by Rodents? B. Niklasson, B. Hörnfeldt, and B. Lundman * Multidrug-Resistant Mycobacterium tuberculosis: Molecular Perspectives. A. Rattan, A. Kalia, and N. Ahmad Synopses * Hospitalization for Unexplained Illnesses Among U.S. Persian Gulf War Veterans. J.D. Knoke and C. Gray * Could the Agricultural Use of Burkholderia (Pseudomonas) cepacia Pose a Threat to Human Health? A. Holmes, J. Govan, and R. Goldstein * The Epidemiology of Haemophilus influenzae Invasive Diseases in the United States, 1994-1995: Near Disappearance of a Child Vaccine Preventable Disease. K.M. Bisgard, A. Kao, J. Leake, P.M. Strebel, B.A. Perkins, M. Wharton * Multiply Resistant Enterococci: The Nature of the Problem and an Agenda for the Future. M.M. Huycke, D.F. Sahm, and M.S. Gilmore * Enteroaggregative Escherichia coli (EAEC) in Diarrhea and Malnutrition. J.P. Nataro, T. Steiner, and R.L. Guerrant * C3 Binding Among Campylobacter jejuni Strains from Patients with Guillain-Barré Syndrome or with Uncomplicated Enteritis. B.M. Allos, F.T. Lippy, A. Carlsen, R.G. Washburn, and M.J. Blaser Dispatches * An Apparently New Virus (Family Paramyxoviridae) Infectious for Pigs, Humans, and Fruit Bats. A.W. Philbey, P.D. Kirkland, A.D. Ross, R.J. Davis, A.B. Gleeson, R.J. Love, A.R. Gould, P.W. Daniels, and A.D. Hyatt * Probable Human Infection with a Newly Described Virus in the Family Paramyxoviridae. K. Chant, R.Chan, M.Smith, D.E Dwyer, P. Kirkland, and the NSW Expert Group * Emergence of the M Phenotype of Erythromycin-Resistant Pneumococci in South Africa. C.D. Widdowson and K.P. Klugman * Reemergence of Plasmodium vivax Malaria in the Republic of Korea. B.H. Feighner, P. Son Il, W.L. Novakoski, L.L. Kelsey, and D. Strickman * Molecular and Epidemiologic Analysis of Dengue Virus Isolates from Somalia, N. Kanesa-thasan, G.J. Chang, B.L. Smoak, A. Magill, M.J. Burrous, and C.H. Hoke, Jr. * Molecular Characterization of a Novel Spotted Fever Group Rickettsia species from Ixodes scapularis in Texas. A.N. Billings, G.J. Teltow, and D.H. Walker * Dual Infection with Erlichia chaffeensis and a Spotted Fever Group Rickettsia: A Case Report. D.J. Sexton, G.R. Corey, C. Carpenter, L.Q. Kong, T. Gandhi, E. Breitschwerdt, B. Hegarty, S.-M. Cheng, H.-M. Feng, X.-J. Yu, J. Olano, D.H. Walker, and S.J. Dumler * New Probable Vectors of Rift Valley Fever in West and Central Africa. D. Fontenille, M. Traore-Lamizana, M. Diallo, J. Thonnon, J.P. Digoutte, and H.G. Zeller * Mycobacterium tuberculosisInfection as a Zoonotic Disease: Evidence of Possible Transmission between Humans and Elephants. K. Michalak, C. Austin, S. Diesel, J.M. Bacon, P. Zimmerman, and J.N. Maslow * Molecular Fingerprinting of Multiresistant Salmonella typhi. M.D. Hampton, L.R. Ward, B. Rowe, and E.J. Threlfall * Using Nurse Hot Line Calls for Disease Surveillance. J.S. Rodman, F. Frost, and W. Jakubowski * Clostridium septicum Infection and Hemolytic Uremic Syndrome. M. Barnham and N. Weightman * Persistent Infection of Pets within a Household with Three Bartonella Species. D.L. Kordick and E.B. Breitschwerdt Commentary * Vector-borne Disease Surveillance and Natural Disasters Roger S. Nasci and Chester G. Moore Letters * Similarity of Chemokines Charge and the V3 Domain of HIV-1 env Protein Ben Berkhout and Atze T. Das * Detection of Glycoprotein of Burkholderia pseudomallei Natalja P. Khrapova, Nikolay G. Tikhonov, Yelena V. Prokhvatilova * Forging New Perspectives on Disease Surveillance Harley Feldbaum * Provide a Context for Disease Emergence Toby D. St. George * Resistance to Dryness of Escherichia coli O157:H7 Strains from Outbreak in Sakai City, Japan, 1996 Yoshio Iijima, Mayumi Matsumoto, Kumiko Higuchi, Taro Furuta, and Takeshi Honda * Irradiation Pasteurization of Solid Foods Stephen Moses and Robert C. Brunham * Emerging Infectious Diseases in Brazil Luiz Jacintho da Silva * Reply to L.J. da Silva Hooman Momen * A Brief Update on Rabbit Hemorrhagic Disease Virus Lorenzo Capucci and Antonio Lavazza * Rabbit Hemorrhagic Disease C. Mead * Reply to Drs. Capucci, Lavazza, and Mead Alvin W. Smith, Neil J. Cherry, and David O. Matson News and Notes * The National Food Safety Initiative * Conference on Global Disease Elimination and Eradication as Public Health Strategies, February 26, 1998 * New and Reemerging Infectious Diseases: A Clinical Course * Erratum Emerging Infectious Diseases National Center for Infectious Diseases Centers for Disease Control and Prevention Atlanta, GA -------------------------------------------------------------------------------------- Update International Editors Update -------------------------------------------------- In this new section, we welcome commentary and updates from our newly formed Board of International Editors. Emerging Infectious Diseases—Southeast Asia Sai Kit Lam University of Malaya, Kuala Lumpur, Malaysia Dr. Lam is professor and head, Department of Medical Microbiology, Faculty of Medicine, University of Malaya. He serves on the editorial board of several medical journals and is the director of the WHO National Influenza Center, the WHO National Center for Rapid Viral Diagnosis, and the WHO Collaborating Center for Arbovirus Reference and Research for Dengue and Dengue Hemorrhagic Fever. The recent emergence of a new strain (H5N1) of influenza A virus, the so-called avian or bird flu, in Hong Kong underlines the importance of the Southeast Asia region as an epicenter not only for influenza A viruses but also for other microbial agents. In late 1992, Vibrio cholerae O139 appeared on the Indian subcontinent; and within a few months, it had spread to China, Nepal, Pakistan, Malaysia, and as far as Russia. The World Health Organization (WHO) has reported that infectious diseases account for more than 17 million deaths per year worldwide and that at least 30 new infectious diseases have emerged within the last 2 decades. Up to half of the 5.8 billion people on earth are at risk for many endemic diseases, with the most overpopulated and economically depressed countries in Southeast Asia at highest risk. Although vaccines and antibiotics are available for many diseases, in 1995 alone, respiratory infections such as pneumonia killed 4.4 million people, 4 million of them children. Diarrheal diseases, including cholera, typhoid, and dysentery killed 3.1 million, most of them children. Tuberculosis (TB) killed almost 3.1 million, malaria 2.1 million, hepatitis B more than 1.1 million, and measles more than 1 million. Faced with the reality of infectious diseases on a daily basis, Southeast Asian nations have not been able to give emerging disease surveillance the priority status it deserves. Much of the surveillance in the region is centered in a few well-established laboratories where there is adequate expertise, sufficient funding, and personal interest. Because of influenza A, it is widely believed that Asia is the epicenter of new influenza strains; the emergence of the avian influenza strain in Hong Kong supports this contention. As a result of the activities of the Pacific Basin Respiratory Virus Research Group, an excellent surveillance program for respiratory viral infections is in place. If this small cluster of avian flu had occurred elsewhere, it would most likely have been missed. All the 110 WHO Influenza Centers, particularly those in Southeast Asia, have been alerted and are frequently updated about the Hong Kong outbreak. Diagnostic reagents for the identification of the H5N1 strain will be made available to these centers as part of the surveillance program. Enteroviruses Enteroviruses are frequently associated with the hot and humid climate of tropical countries in Southeast Asia. The viruses are responsible mainly for inapparent and mild infection, but occasionally they can give rise to infection of the central nervous system, resulting in aseptic meningitis and (rarely) paralysis; polioviruses are a good example. Hand, foot, and mouth disease, caused by enteroviruses, such as coxsackie A16 and enterovirus 71, is common in many countries in the region. In a recent outbreak of hand, foot, and mouth disease in Malaysia, a new clinical entity of encephalomyelitis emerged, which resulted in several deaths among children under 5 years of age. Four such deaths were investigated thoroughly in the University Hospital in Kuala Lumpur. All four children were seen in the Emergency Department in a state of respiratory and cardiovascular instability with a history of fever (3 to 5 days) associated with reduced oral intake. One child had flaccid paralysis with hyporeflexia of the lower limbs. A constant feature in these patients was the very rapid vascular changes, followed rapidly by cardiac decompensation. Pulmonary edema followed, suggesting a concomitant increase in pulmonary vascular permeability. Consent for postmortem was obtained with some difficulty because of religious and cultural reasons. The midbrain, pons, medulla, and spinal cord were extensively damaged, with the worst damage at the level of the medulla. The cerebrum and myocardium were relatively normal. Enterovirus 71 was isolated from the medulla and spinal cord of three children and from the cerebrum of the other child, where no brain stem material was available. Other tissue sites including the cerebrospinal fluid did not yield virus. The need for postmortem examinations in investigating fatal emerging diseases must be addressed in areas where such examinations are not possible for cultural reasons. Arboviruses Arthropod-borne viruses or arboviruses are yet another group of viruses synonymous with countries in the region and cause considerable sickness and death. Some of these viruses, e.g., Murray Valley encephalitis virus in Australia, have restricted geographic distribution but may have the ability to transcend geographical barriers, as in the emergence of Japanese encephalitis virus in the Torres Strait of northern Australia and in Papua New Guinea. Dengue Dengue is by far the most important arbovirus infection in Southeast Asia. According to WHO, dengue has been reported in over 100 countries worldwide and poses a threat to approximately 2 billion people. At the 46th World Health Assembly held in Geneva in May 1993, a resolution was passed to make the prevention and control of dengue a priority. Dengue is an acute viral infection characterized by abrupt onset of fever, severe headache, pain behind the eyes, muscle and joint pains, and rash. The hemorrhagic form of dengue fever, dengue hemorrhagic fever (DHF), was recognized as a new disease in the Philippines in 1953 and has been seen in India, Malaysia, Singapore, Indonesia, Vietnam, Cambodia, and Sri Lanka. During 1956 to 1992, 1,335,049 cases of DHF (13,723 deaths) were recorded in Vietnam alone. The rise of dengue in tropical and subtropical areas of the world is explained by factors such as rapid population growth, expanding urbanization, inadequate municipal water supplies, and difficulties in refuse disposal. These lead to an abundance of new breeding sites for the mosquito vectors, while human migration patterns disperse vectors and viruses into new areas. The resurgence of DHF in several countries where only dengue fever had been reported has become worrisome. Sri Lanka reported its first outbreak in 1989, and India and China also faced the same dilemma. DHF/dengue shock syndrome has also been reported for the first time in New Caledonia and Tahiti. The large outbreak in Cuba in 1981 showed how the disease has spread to the Americas and other countries such as Venezuela and Brazil. In Malaysia, where the disease is endemic, encephalitis cases associated with dengue viruses have been documented, and vertical transmission has also been reported in the last two years. Other unusual manifestations observed rarely include acute renal failure and hemolytic uremic syndrome. The emergence of such unusual clinical manifestations should also be monitored and documented in other dengue-endemic countries. Despite intensive efforts by the countries in the region to control the vectors, the disease is still on the rise. A tetravalent live attenuated vaccine developed at Mahidol University, Thailand, is being field-tested, and the preliminary results are promising. Disease Spread Rapid transportation has been blamed for the spread of diseases, and this is especially so in countries in the region. Tourism is an important industry in Southeast Asia, and a series of conferences on travel medicine has highlighted the problems, particularly in infectious diseases. Another vast movement of people between countries in the region is that of migrant workers. In Malaysia, an estimated 1.7 million workers are migrant workers, of which only 1.2 million came in legally. Most of the workers are from Indonesia, Myanmar, Philippines, and Pakistan, and they work in the agricultural sector as well as in construction and domestic services. Those who come in legally undergo a medical examination that includes screening for some infectious diseases. Diseases associated with migrant workers are chloramphenicol-resistant typhoid, multidrug-resistant TB, leprosy, malaria, HIV/AIDS, and filariasis. In 1993, the first case of kala-azar caused by Leishmania donovani was reported in a Bangladeshi migrant worker in Malaysia, and since then, four other cases have been identified, all of them in Bangladeshi workers. This disease had not been present in Malaysia. In 1997, the theme "Emerging Infectious Diseases: Global Alert—Global Response" was chosen by WHO to celebrate World Health Day. This year, as WHO celebrates its 50th anniversary, the region is still plagued with major public health problems threatening the lives of millions of people in the region. These problems underline the need to strengthen epidemiologic surveillance and concentrate public health activities in Southeast Asia, the epicenter of many emerging diseases. Address for correspondence: S. K. Lam, Department of Medical Microbiology, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia; fax: 603-758-2801; e-mail: lamsk@medicine.med.um.edu.my. Emerging Infectious Diseases National Center for Infectious Diseases Centers for Disease Control and Prevention Atlanta, GA URL: http://www.cdc.gov/ncidod/EID/vol4no2/update.htm Please note that figures and equations are not available in ASCII format; their placement within the text is noted by [fig] and [eq], respectively. Greek symbols are spelled out. The following codes are used: (ft) for footnote; (sup) for superscript; (sub) for subscript; /= for greater than or equal to. ------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------- [Emerging Infectious Diseases] [Volume 4 No. 2 / April - June 1998] Perspectives Wild Primate Populations in Emerging Infectious Disease Research: The Missing Link? Nathan D. Wolfe,* Ananias A. Escalante,† William B. Karesh,‡ Annelisa Kilbourn,‡ Andrew Spielman,* and Altaf A. Lal† *Harvard School of Public Health, Boston, Massachusetts, USA; †Centers for Disease Control and Prevention, U.S. Public Health Service, Chamblee, Georgia, USA; ‡Wildlife Health Sciences, Wildlife Conservation Society, Bronx, New York, USA. ----------------------------------------------------------------------- Wild primate populations, an unexplored source of information regarding emerging infectious disease, may hold valuable clues to the origins and evolution of some important pathogens. Primates can act as reservoirs for human pathogens. As members of biologically diverse habitats, they serve as sentinels for surveillance of emerging pathogens and provide models for basic research on natural transmission dynamics. Since emerging infectious diseases also pose serious threats to endangered and threatened primate species, studies of these diseases in primate populations can benefit conservation efforts and may provide the missing link between laboratory studies and the well-recognized needs of early disease detection, identification, and surveillance. Infectious diseases respect no species or geographic boundaries. For a parasite, closely related hosts offer new environments in which infection, maintenance, replication, and transmission remain possible. The anthropoid primates (which include humans) and to a lesser degree simian primates share broadly similar physiologic and genetic characteristics and thus susceptibility to many viruses, bacteria, fungi, protozoa, helminths, and ectoparasites (1) that have the potential to cross primate-species boundaries (2). Similarities in pathogen susceptibility have made nonhuman primates ideal laboratory models. During the 20th century, laboratory research on captive primates has elucidated the life cycle and pathogenesis of many infectious agents and facilitated drug and vaccine development. Nevertheless, the ecology of infectious agents found in wild populations of primates has only recently been addressed. Just as captive primates have proved invaluable for research at the level of the organism, wild populations can provide the opportunity to study infectious disease phenomena at the population and ecosystem levels. Research at these levels addresses such pressing questions as the origin(s) of pathogens, determinants of pathogen emergence, and factors influencing maintenance of pathogens in animal reservoirs. During the past two decades unknown human diseases, including AIDS, Ebola fever, hantavirus infection, and dengue hemorrhagic fever, have emerged from enzootic foci. The emergence of these and other diseases has been linked to the interface of tropical forest communities with high levels of biodiversity and agricultural communities with relative genetic homogeneity and high population densities of humans, domestic animals, and crops. This interface poses a high risk for the emergence of novel disease (3-5). Since most nonhuman primates live in tropical forest habitats, most interactions between humans and wild nonhuman primates occur in this high-risk interface, which has recently increased because of expanded ecotourism and forest encroachment. These interactions can lead to pathogen exchange through various routes of transmission (Table). Arthropod vectors, shared water, and hunting of wild animals have facilitated pathogen exchange and may have played an important role in pathogen transfers since ancient times. In the recent past, laboratory research has led to accidental human exposure to such agents as primate malaria parasites (12) and a simian immunodeficiency virus (SIV) (16). The potential for exchange through xenotransplantation has been discussed (17), and infection from vaccine contaminated with SV40, a primate papovavirus, led to the exposure of millions of persons in the 1950s (15). Conversely, pathogen transmission from humans to nonhuman primates places both captive and wild animals at serious risk for diseases such as measles and tuberculosis (TB), which are deadly in many nonhuman primate species. In general, as levels of interaction increase, so does pathogen exchange, resulting in further risks to both humans and nonhuman primates. The need for improved surveillance and basic research on emerging infectious disease is well documented (3,18-20). Wild primates can serve as sentinels by signaling which pathogens pose a risk for humans in the immediate area (21) as well as in distant countries (5). Here, we describe potential benefits of incorporating wild populations into emerging infectious disease research. Pathogen Origins Nonhuman primates are infected with the closest relatives of important human pathogens. The construction of molecular phylogenies has played an important role in studying the evolution and classification of many pathogens. While trees that result from these analyses must be interpreted with care, they can provide valuable information on the history of pathogens. Furthermore, the selection of genes that are evolving at the appropriate rate should allow phylogenetic analyses to assess both ancient and more recent epidemic origins (22). Human herpes simplex virus infection is common in nonhuman primates and may reflect contact with humans. A study of viral infection in nonhuman primates found that chimpanzees and gorillas were seropositive to human herpes simplex virus-1 and -2 strains, but orangutans and gibbons were not (23). Research on SIV phylogeny (24,25) has shown that HIV-1 and HIV-2 are each more closely related to primate pathogens than they are to one another. HIV-1 is in a group with SIV(sub CPZ) (26), a chimpanzee (Pan troglodytes) virus, while HIV-2 falls within a clade consisting of West African primate viruses (Figure 1; 24). In fact, Mindell (27) has argued that HIVs and SIVs should be referred to as `primate immunodeficiency viruses' to more accurately reflect their heritage. Preliminary evidence suggests that the origins of global T-cell lymphotropic virus-1 subtype diversity may be analogous, leading Liu et al. (28) to propose that HTLV-I subtypes emerged from three separate nonhuman primate reservoirs. Table. Routes of pathogen exchange between human and nonhuman primates -------------------------------------------------------------------------- Direction of Route of exchange Pathogen exchange Evidence(sup a) Reference -------------------------------------------------------------------------- Nonhuman primate to Animal bite Herpes B human E 6(sup b) Nonhuman primate to Monkeypox human E 7 Human to nonhuman Fecal-oral Poliovirus primate L 2(sup b) Chimpanzee to Poliovirus chimpanzee E 8 Nonhuman Hunting, food prep primate to & eating Ebola human E 9 Mycobacterium Among Nasal secretions leprae primates P, L 10(sup b) Human to nonhuman Respiratory droplet Tuberculosis primate L 11(sup b) Both Vector-borne Malaria directions L,E 12(sup b) Both Filaria directions L,E 8(sup b) Human to nonhuman Water-mediated Dracunculiasis primate L 13 Nonhuman primate to Schistosomiasis human E 14 Nonhuman primate to Xenotransplantation SV40 human E(sup c) 15(sup b) --------------------------------------------------------------------------- (sup a)L = laboratory; E = epidemiologic; P = evidence that parasites live naturally in multiple primate hosts. (sup b)Evidence reviewed. (sup c)The only current evidence for xenotransplantation includes SV40 spread through vaccine production. This pattern is not unique to viruses, as demonstrated through decades of research on primate malarias. More than 26 species of Plasmodia infect primates (12,29). Both morphologic and molecular analyses show human and nonhuman primate malarias interdigitating on phylogenetic trees (Figure 2; 30,33). Preliminary evidence suggests that this group of parasites has a range of coevolutionary scenarios, including the speciation of P. vivax and related parasites in Asian primates, the recent exchange of parasites between humans and New World monkeys (30), and perhaps an ancient exchange of a falciparum-like parasite from a bird or lizard to an African hominoid (31,34,35). Further analyses are likely to lead to surprises. For example, recent research has shown that the diversity of human P. vivax (previously considered to be a single species) also includes a P. vivax-like parasite, a widespread pathogen, which is most closely related to P. simiovale, a primate malaria of Asian macaques (Macaca sp.) (30,36). [fig 1] Figure 1. Relationships among primate and human lentiviruses: Phylogeny of primate lentiviruses based on the gag gene obtained by (25). The names of the strains are indicated in parentheses. Hosts are indicated on trees. The description of the lentivirus strains is provided in (25). A major limitation of these studies is that they rely on very small numbers of nonhuman primate isolates only two SIV(sub cpz) isolates for the HIV-1 clade, a single P. simiovale isolate for malaria phylogeny. In addition, very few data are available on the distribution of these pathogens in wild populations; more data on the distribution and extent of shared pathogens should provide clues to origins (37), thereby pointing to general conditions that may contribute to disease emergence. Molecular phylogenies can also play a direct role in the control of pathogens, for example, in charting their antigenic diversity, a task necessary for vaccine design (38). Because nonhuman primate pathogens are often evolutionary outgroups, vaccines that are targeted at antigens shared by human and nonhuman pathogens should provide more universal coverage. Molecular phylogenies can also assist in determining the rate of evolution of vaccine candidate antigens. Because low rates of mutation often indicate selective constraint, this technique may point to candidate antigens that cannot easily mutate to evade vaccine-induced antibodies. Similar analyses identifying highly constrained gene products or metabolic pathways as targets for drug development may contribute to slowing the emergence of drug resistance. Reservoirs The study of pathogen transmission has been encumbered by the use of inappropriate terminology. Such terms as "primary hosts," "reservoir hosts," "carriers," "arthroponoses," "xenonoses," and "zoonoses" pose premature assumptions and belie aspects of the origins and natural transmission cycles of pathogens. The factors that determine pathogen viability within a vertebrate host may correspond only roughly, or not at all, to species boundaries. Primate pathogens do not adhere to the faithfully maintained sanctity of the distinction between humans and nonhumans. [fig 2] Figure 2. Relationships between primate and human parasites: Malaria phylogeny based on the circumsporozoite protein gene. The alignment does not include the central repeat region. P. falciparum (Pfa), P. vivax (Pvi), and P. malariae (Pma) are from humans; P. cynomolgi (Pcy), P. simiovale (Pso), and P. knowlesi (Pkn) are from macaques; P. simium (Psi) and P. brasilianum (Pbr) are from New World monkeys; P. reichenowi (Pre) is from chimpanzees; P. gallinacium (Pga) is from birds; and P. berghei (Pbe) and P. yoelii (Pyo) are from rodents. The numbers in the names indicate different isolates as described in (30). The sequence of P. gallinacium was reported by (31). The numbers on the branches are bootstrap % based on 500 pseudoreplications. The tree was estimated by the neighbor-joining method with the Tajima and Nei distance (32). The misconception of an evolutionary trend toward increasing host specificity (39) has contributed to the belief that pathogen exchange should be rare. An apparently restricted host range, however, may be the product of specific ecologic conditions and not of an intrinsic characteristic of the pathogen. In many cases, ecologic changes can broaden the host range of a pathogen. For example, the filarid worm Loa loa remains outside human populations, primarily because of vector behavior (40); changes in ecology, such as the availability of a novel host, can change vector behavior and expand the pathogen's host range. Similar phenomena may have played a role in the emergence of a range of flaviviruses (e.g., yellow fever, dengue, and Japanese encephalitis) from primarily forest-dwelling nonhuman primate cycles. For these flaviviruses, primate populations may also continue to play a role by introducing novel genetic variants, which, at least in the case of dengue, may be involved in pathogenesis. Another misconception is that primate populations are too sparse to maintain human pathogens. A number of variables influence whether or not a pathogen is maintained in a given population. The capacity for latency, for example, may decrease the host population size necessary to maintain pathogens. This characteristic has evolved independently in pathogens as diverse as herpesviruses and Plasmodium and may help explain sustained transmission in hosts with low population density. Orangutans (Pongo pygmaeus), for example, are thought to act as hosts to two distinct Plasmodium species (41), despite an estimated population size of two per km2. While molecular phylogenies have not yet been used to verify that these parasites are unique, the presence of a dormant hypnozoite stage might allow for sustained transmission even at this extreme. Where human and nonhuman hosts overlap, both must be factored into epidemiologic models. Small populations of nonimmune humans alone may not be capable of maintaining a pathogen, but when nearby nonhuman hosts are considered, a critical population size may be reached. Surveys to assess pathogen prevalence among nonhumans can play an important role in control strategies. Eradication programs must consider animal reservoirs; even if complete human coverage is achieved, long-term reemergence from animal reservoirs can undo the best eradication efforts. For example, it is not known if poliovirus infections can be maintained in nonhuman primate populations. Seroprevalence surveys conducted before and after eradication may prove invaluable. Accidental exposure to infected laboratory workers has led to poliovirus infections of chimpanzees and gorillas since the 1940s (1). Poliovirus can infect not only our closest living relatives, chimpanzees, gorillas (Gorilla gorilla), and orangutans (Pongo pygmaeus) (2), but also more distantly related anthropoids like the colobus monkeys (e.g., Colobus abyssinicus kikuyuensis [=guereza]) (42). Antibodies and shed virus have also been found in recently imported animals (8), and some chimpanzees may act as symptomless carriers (2). Long-term research by Jane Goodall on wild Tanzanian chimpanzees documented the potential for transmission of poliovirus (or a similar virus) in free-ranging chimpanzee populations (43). Since no samples were collected, it is impossible to determine if the epidemic described by Goodall was part of a natural chimpanzee cycle or the result of introduction from local human populations or researchers. As poliovirus eradication efforts intensify, it may be useful to monitor virus prevalence in humans living near primate habitats. Control efforts that rely on antimicrobial drugs must also take into account the potential for nonhuman primate reservoirs. Despite demonstrations that mass administration of diethylcarbamazine citrate successfully controlled Brugia malayi (a filarid worm), Mak et al. found a high prevalence of B. malayi after a large-scale administration of chemotherapy (44). Research showed that even though periodic prevalence of B. malayi decreased, subperiodic prevalence remained high. The maintenance of subperiodic B. malayi was eventually attributed to mosquitoes infected by leaf monkeys (Presbytis obscura). In this particular free-ranging primate host, approximately 83% of the monkeys were infected (44). Nonhuman "reservoirs" may also confer potential benefits. It is at least theoretically possible that nonhuman primate populations may provide a barrier to the spread of drug-resistant pathogens; while these pathogens benefit from resistance in environments where drug pressure exists, drug resistance can be costly, and resistant pathogens may not compete effectively against susceptible `wild-type' pathogens in the absence of drug pressure. As drug-free populations, reservoirs may provide havens for susceptible pathogens, thereby decreasing the rate at which drug-resistant genes spread and increasing the rate at which susceptibility may return after drug pressure ends. This hypothesis may help explain why the rates of drug-resistant gram-negative enteric bacteria of wild baboons (Papio cynocephalus) living with limited human contact are significantly lower than those of baboons living with human contact (45). Sentinel Surveillance In tropical lowland forests, which contain the greatest biodiversity of terrestrial habitats (46), exist rarely seen or unknown pathogens with the potential to enter human populations. These pathogens may affect residents of and visitors to forested regions (21) and act as the source of introduction of infectious agents to distant susceptible populations (47). Increasing human contact with forested systems almost certainly leads to a corresponding increase in the emergence of infections in the human population. Nevertheless, predicting which pathogens humans may encounter and be susceptible to remains a methodologic challenge. Surveillance methods for predicting emerging pathogens include surveillance of vectors or forest-dwelling human populations and wildlife epidemiology (epidemiologic study of infections in wild populations) (48). These approaches have limitations. While vector sampling may prove the easiest method for widespread surveillance, the pathogens identified from vectors may be difficult or impossible to culture. Even when successful, vector sampling is likely to identify a range of pathogens, only some of which may infect humans. Studies of human populations, while providing valuable information, are limited to regions in which forest-dwelling human populations exist. Epidemiologic research among free-ranging primate populations has the potential to predict which pathogens might enter human populations as contact with forested regions increases. In addition to their physiologic similarities to humans, primates have other characteristics that contribute to their accumulation of infectious agents. Primates live primarily in forested environments (49); in general, they have large bodies and live in large groups-characteristics that may attract vectors (50). Furthermore, dependence on fruit, a characteristic of most primates, requires mobility (both terrestrially and arboreally), a trait that may increase exposure to pathogens (51,52). Despite the lack of organized attempts to document the distribution of pathogens in wild populations, recognized "die-offs" in wild primate populations have played an important role in identifying novel pathogens. In 1956, for example, a novel flavivirus was identified through the investigation of large-scale deaths of bonnet macaques (Macaca radiata) and hanuman langurs (Presbytis entellus) in the Kyanasur Forest of India (Seymour, 1981, cited in [2]) caused by Kyanasur Forest virus. More recently, in 1995, deaths in a chimpanzee population studied by Christoph Boesch in the Täi Forest, Côte d'Ivoire, and a single human case following a necropsy led to the identification of a novel strain of Ebola virus (9,53). The single human case in the Swiss researcher foreshadowed the localized mini-outbreak of Ebola hemorrhagic fever in Mayibout, a village in the northeast of Gabon in January 1996. The Gabon epidemic was linked to the handling, preparation, and consumption of a chimpanzee that had been found dead; 29 of 37 identified cases involved exposure to the dead chimpanzee (54). Close monitoring of such populations, as is being conducted in the Täi Forest, has the potential to identify emergence-linked behavior, such as the consumption of specific plants or insects, which may lead to the still elusive reservoir of Ebola virus. Considering the exceptionally small percentage of wild primate populations under long-term study, these examples represent only the tip of the iceberg. More systematic monitoring of wild primate populations will likely provide a substantial payoff in our understanding, identification, and possible control of novel pathogens, both for humans and endangered primates. Surveillance for certain types of human-nonhuman primate contact may be particularly useful. Hunting, which involves tracking, capturing, handling, transporting, preparing, and consuming meat, may play a particularly important role in pathogen exchange. In addition to the recent evidence of hunting-mediated Ebola transmission, the hunting of a red colobus (Colobus pennanti oustaleti [=badius]) has been implicated in a localized epidemic of monkeypox, an orthopoxvirus similar to smallpox, which continued for four generations of human-to-human contact (55). Another example is the increased risk for feline plague among cats that hunt rodents (7). Necropsies share many characteristics with hunting and are appropriately considered a high-risk activity. Bites from wild primates may also play a role in the transmission of certain pathogens. For example, chimpanzee-to-human transmission of monkeypox occurred when a wild chimpanzee bit a 2-year-old girl (6). Further sampling demonstrated high prevalence in forest squirrel populations (up to 49% among Funisciurus lemniscatus) (56), underlining the need for comprehensive studies before determining the ultimate source and reservoir of pathogens. Bites and scratches have transmitted pathogens to laboratory workers. The transmission of B-virus, a herpesvirus infecting rhesus macaques (Macaca mulatta), has caused death in 18 of 24 known human cases (57). Surveillance and education in human populations that hunt nonhuman primates, as well as follow-up of reported primate bites in nonlaboratory settings, may be indicated. Ecologic and Social Factors Involved in Natural Transmission The potential for disease emergence and reemergence depends on the interaction of complex social, ecologic, and genetic factors at the host, vector, and pathogen levels (58). Because wild populations of primates display diverse social behavior and live in a range of ecologic environments, they exemplify natural transmission. Furthermore, monitoring pathogens in wild primate populations does not involve the treatment or behavior change interventions that monitoring pathogens in humans requires. Exceptions include the increasingly common need for wildlife medicine to maintain the health of endangered species and decrease the impact of pathogens from humans and domestic animals on free-ranging animals (59). Nevertheless, for many pathogens, basic epidemiologic research, and not treatment, remains the primary goal of wildlife medicine (59). Studies of wild populations can highlight factors associated with the pathogen exchange across species boundaries. Some parasites, for example, the chewing lice–pocket gopher system, complete their entire life cycle on a single host. This high level of host specificity may contribute to the close cospeciation between pocket gophers and their respective lice; the phylogenetic trees of host and parasite are nearly mirror images (60). Nevertheless, such host-specific parasites may be the exception rather than the rule. Many intestinal parasites, for example, seem to be generalists. One epidemiologic study of wild primates in the Kibale Forest National Park, Uganda, evaluated the role of primate distribution on the distribution of intestinal amoebas (61); the park contains a number of spatially separate primate groups, each of which consists of multiple primate species. Most variation in amoeba prevalence was explained by group membership, and little was explained by species, which suggests that these parasites treated the primate groups as biologic islands (62), spreading easily among diverse members of the same island but rarely spreading to new islands. Long-term behavioral research combined with occasional pathogen sampling may provide valuable data. Long-term research on the SIVs of East African primates, for example, has provided evidence for a recent cross-species transmission of SIV between baboons and African grivet monkeys (Cercopithecus aethiops aethiops) (62). This research has documented the incidence of SIV(sub agm) among Ethiopian grivet monkeys for more than 20 years (63). By examining prevalence in age groups over time, this research demonstrated the minimal impact of SIV(sub agm) on the survival of grivets, the predominance of sexual transmission, and the lack of maternal transmission. An ecologic approach to pathogen transmission can benefit our understanding of emerging diseases. By determining how humans become a part of the life cycle of pathogens rather than how pathogens enter human populations, we can better understand the factors associated with emergence and improve the quality of public health responses. How humans become part of the life cycle of pathogens depends on human migrations, environmental changes, and cultural and social factors, in the context of evolutionary history of the pathogens, vectors, and hosts, which make up an infectious system. Ecologic and evolutionary studies of wild animals in general, and primate populations in particular, can address questions arising from these complex interactions. The Future Traditionally, the study of wild populations of primates has been the domain of primatologists and wildlife veterinarians, who have worked to overcome logistic difficulties in the field, develop methods, and address ethical issues. Opportunities exist for collaborative work on infectious disease; the benefits of such efforts are considerable. The translocation of animals from vulnerable forest fragments to forest reserves, an increasingly common conservation effort, is an opportunity for pathogen sampling. Other possibilities are provided by collaboration with long-term behavioral research on free-ranging animals. While behavioral research sites may only provide fecal and urine samples, the study of intestinal parasites is often possible; recent advances in urinalysis demonstrate that urine may be a source of both antigen and antibody from systemic infections, such as malaria (64). The distribution of appropriate necropsy protocols and sample collection kits would improve data collection and decrease risks associated with necropsies. Urgently needed are strategies for noninvasive remote diagnosis. For example, combined with sensitive molecular diagnostics, a remote tissue-biopsy dart system (65) may have potential for obtaining epidemiologic samples. Efforts to preserve endangered primates and monitor disease emergence have some common objectives. Pathogen exchange is a two-way street, and exposure to human pathogens poses a serious threat to endangered animals (11). While the main focus of this article has been risk to humans, nonhuman primates are frequently more threatened by microorganisms indigenous to humans than vice versa (2). TB, which is often fatal to nonhuman primates, represents a serious threat to laboratory primate communities, which commonly are infected by humans and can occasionally reinfect laboratory workers (66). TB is prevalent in wild populations, as demonstrated by the presence in wild olive baboons (Papio cynocephalus) of Mycobacterium bovis infection, an infection that most likely originated from cattle (67). Both hunting and forest encroachment threaten endangered primates and increase the possibility of human and nonhuman infectious disease emergence. Research on the pathogens of primates and humans on forest boundaries helps assess risks to wild primates and to humans. In addition to their value in the study of infectious disease and human evolution, many primates are valuable natural resources in their home countries. The veterinary expertise and wildlife management skills of conservation organizations can both supplement the basic pathogen research and control work of the public health community and benefit from it. The examination of pathogen exchange in regions of host overlap may identify social factors that influence pathogen emergence. Data on forest use by human communities surrounding forest reserves and levels of crop-raiding by nonhuman primates have been collected as part of ongoing conservation projects. A comparison of these data with human and nonhuman primate pathogen prevalence may provide a measure of forest-oriented behavioral risks. Further research comparing infection rates among hunters and nonhunters could confirm the findings and determine the role of behavioral control measures to decrease risk. Much remains to be done. Recent evidence has suggested that some wild primates self-medicate with plants in their environment (68). An understanding of the underlying wildlife epidemiology, therefore, combined with long-term dietary data and analysis of plant chemistry, may lead to new chemotherapeutic drugs (69). In addition, research on the dynamics of primate pathogens in their natural hosts may elucidate novel host resistance mechanisms. In particular, evidence of impaired host survival and reproduction, as well as long-term host-parasite association, may point to hosts that are likely to have evolved genetically mediated resistance. Such resistance mechanisms in humans (70) have begun to play an increasingly important role in vaccine and drug development (71,72). While the findings from captive primate studies have played an important role in medicine in the 20th century, this period has also been marked by a notable absence of research on the basic ecology of disease systems. Perhaps by learning from primates in their natural environments we may better prepare ourselves for the disease threats to humans and wildlife populations in the coming century. Acknowledgments We thank Adria Tassy Prosser and Daniel G. Colley for valuable suggestions and Mary Bartlett for editorial assistance. Nathan Wolfe is supported by a Taplin Graduate Fellowship and a National Science Foundation-Predoctoral Fellowship. Ananias A. Escalante is supported by the ASM/NCID Postdoctoral Research Associates Program. Nathan Wolfe is a doctoral candidate in the Department of Immunology and Infectious Disease at the Harvard School of Public Health. Mr. Wolfe has conducted research on self-medicating behavior among chimpanzees and mountain gorillas in Uganda and is collaborating with Drs. Altaf Lal and Ananias Escalante on malaria evolution research. Address for correspondence: Altaf A. Lal, Centers for Disease Control and Prevention, Mail Stop F12, 4770 Buford Hwy, Chamblee, Georgia 30341, USA; fax: 770-488-4454; e-mail: aal1@cdc.gov. References 1. Ruch TC. Diseases of laboratory primates. Philadelphia: W.B. Saunders Company; 1959. 2. Brack M. Agents transmissible from simians to man. Berlin: Springer-Verlag; 1987. 3. Morse SS. Emerging viruses. In: Morse SS, editor. Emerging viruses. New York: Oxford University Press, Inc.; 1993. 4. Levins R, Awerbuch T, Brinkmann U, Eckardt I, Epstein P, Makhoul N, et al. The emergence of new diseases. American Scientist 1994;82:52-60. 5. Wilson ME, Levins R, Spielman A. Disease in evolution: global changes and emergence of infectious diseases. Ann N Y Acad Sci 1994;740. 6. Mutombo M, Arita I, Jezek Z. Human monkeypox transmitted by a chimpanzee in a tropical rain-forest area of Zaire. Lancet 1983;34:735-7. 7. Eidson M, Tierney LA, Rollag OJ, Becker T, Brown T, Hull HF. Feline plague in New Mexico: risk factors and transmission to humans. Am J Public Health 1988;78:1333-5. 8. Douglas JD, Soike KF, Raynor J. The incidence of poliovirus in chimpanzees (Pan troglodytes). Laboratory Animal Care 1970;20:265-8. 9. Morell V. Chimpanzee outbreak heats up search for Ebola origin. Science 1995;268:974-5. 10. Meyers WM, Gormus BJ, Walsh GP, Baskin B, Hubbard GB. Naturally acquired and experimental leprosy in nonhuman primates. Am J Trop Med Hyg 1991;44:24-7. 11. McCallum H, Dobson A. Detecting disease and parasite threats to endangered species and ecosystems. Trends in Ecology and Evolution 1995;10:190-4. 12. Coatney GR, Collins WE, Warren M, Contacos PG. The primate malarias. Bethesda: U.S. Department of Health, Education, and Welfare; 1971. 13. Beverley-Burton M, Crichton VF. Attempted experimental cross infections with mammalian guinea-worms, Drancunculus spp. (Nematoda: Dracunculoidea). Am J Trop Med Hyg 1976;25:704-8. 14. Fuller GK, Lemma A, Haile T. Schistosomiasis in Omo National Park of Southwest Ethiopia. Am J Trop Med Hyg 1979;28:526-30 15. Shah KV. A review of the circumstances and consequences of simian virus SV40 contamination of human vaccines. Symposium on Continuous Cell Lines as Substrates for Biologicals. Developments in biological standardization, Vol. 70; 1989. 16. Khabbaz RF, Heneine W, George JR, Parekh B, Rowe T, Woods T, et al. Brief report: infection of a laboratory worker with simian immunodeficiency virus. N Engl J Med 1994;330:172-7. 17. Michler RE. Xenotransplantation: risks, clinical potential, and future prospects. Emerg Infect Dis 1996;2:64-70. 18. Berkelman RL, Pinner RW, Hughes JM. Addressing emerging microbial threats in the United States. JAMA 1996;275:315-7. 19. Lederberg J, Shope RE, Oaks SC. Emerging infections: microbial threats to health in the United States. Washington: National Academy Press; 1992. 20. LeDuc JW. World Health Organization strategy for emerging infectious diseases. JAMA 1996;275:318-20. 21. Meslin FX. Surveillance and control of emerging zoonoses. World Health Stat Q 1992;45:200-7. 22. Holmes ED, Nee S, Rambaut A, Garnett GP, Harvey PH. Revealing the history of infectious disease epidemics through phylogenetic trees. Philo Trans R Soc Lond B Biol Sci 1995;349:33-40. 23. Kalter SS, Heberling RL, Cooke AW, Barry JD, Tian PY, Northam WJ. Viral infections of non-human primates. Lab Anim Sci 1997;47:461-7. 24. Myers G, MacInnes K, Korber B. The emergence of simian/human immunodeficiency viruses. AIDS Res Hum Retroviruses 1992;8:373-86. 25. Peeters M, Fransen K, Delaporte E, Van Den Haesevelde M, Gershy-Damet GM, Kestens L, et al. Isolation and characterization of a new chimpanzee lentivirus (simian immunodeficiency virus isolate cpz-ant) from a wild-captured chimpanzee. AIDS 1992;6:447-51. 26. Huet TR, Cheynier A, Meyerhans A, Roelants G, Wain-Hobson S. Genetic organization of a chimpanzee lentivirus related to HIV-1. Nature 1990;345:356-459. 27. Mindell DP. Positive selection and rates of evolution in immunodeficiency viruses from humans and chimpanzees. Proc Natl Acad Sci U S A 1996;93:3284-8. 28. Liu HF, Goubau P, Van Brussel M, Van Laethem K, Chen YC, Desmyter J, et al. The three human T-lymphotropic virus type I subtypes arose from three geographically distinct simian reservoirs. J Gen Virol 1996;77:359-68. 29. Collins WE. Major animal models in malaria research: simian. In: Wernsdorfer WH, editor. Malaria: principles and practice of malariology. Edinburgh: Churchill Livingstone; 1988. 30. Escalante AA, Barrio E, Ayala FJ. Evolutionary origin of human and primate malarias: evidence from the circumsporozoite protein gene. Mol Biol Evol 1995;12:616-26. 31. McCutchan TF, Kissinger JC, Touray MG, Rogers MJ, Li J, Sullivan M, et al. Comparison of circumsporozoite proteins from avian and mammalian malaria: biological and phylogenetic implications. Proc Natl Acad Sci U S A 1996;93:11889-94. 32. Tajima F, Nei M. Estimation of evolutionary distance between nucleotide sequences. Mol Biol Evol 1984;1:269-85. 33. Qari SH, Shi YP, Pieniazek NJ, Collins WE, Lal AA. Phylogenetic relationship among the malaria parasites based on small subunit rRNA gene sequences: monophyletic nature of the human malaria parasite, Plasmodium falciparum. Mol Phylogenet Evol 1996;6:157-65. 34. Waters AP, Higgins DG, McCutchan TF. Evolutionary relatedness of some primate models of Plasmodium. Mol Biol Evol 1993;10:914-23. 35. Escalante AA, Ayala FJ. Phylogeny of the malarial genus Plasmodium, derived from rRNA gene sequences. Proc Natl Acad Sci U S A 1994;91:11373-7. 36. Qari SH, Shi YP, Povoa MM, Alpers MP, Deloron P, Murphy GS, et al. Global occurrence of Plasmodium vivax-like human malaria parasite. J Infect Dis 1993;168:1485-9. 37. Desrosiers RC. HIV-1 origins: a finger on the missing link. Nature 1990;345:288-9. 38. Peeters M, Janssens W, Fransen K, Brandful J, Heyndrickx L, Koffi K, et al. Isolation of simian immunodeficiency viruses from two sooty mangabeys in Côte d'Ivoire: virological and genetic characterization and relationship to other HIV type 2 and SIVsm/mac strains. AIDS Res Hum Retroviruses 1994;10:1289-94. 39. Brooks DR, McLennan DA. Parascript: parasites and the language of evolution. Washington: Smithsonian Institution Press; 1993. 40. Gouteux JP, Noireau F. The host preferences of Chrysops silacea and C. dimidiata. Diptera: Tabanidae in an endemic area of Loa loa in the Congo. Ann Trop Med Parasitol 1989;83:167-72. 41. Peters W, Garnham PCC, Rajapaksa N, Cheong WH, Cadigan FC. Malaria of the Orangutan in Borneo. Philos Trans R Soc Lond B Biol Sci 1976;275:439-82. 42. Suleman MA, Johnson BJ, Tarara R, Sayer PD, Ochieng DM, Muli JM, et al. An outbreak of poliomyelitis caused by poliovirus type I in captive black and white colobus monkeys (Colobus abyssinicus kikuyuensis) in Kenya. Trans R Soc Trop Med Hyg 1984;78:665-9. 43. Goodall J. Population dynamics during a 15 year period in one community of free-living chimpanzees in the Gombe National Park, Tanzania. Z Tierpsychol 1983;61:1-60. 44. Mak JW, Cheong WH, Yen PK, Lim PK, Chan WC. Studies on the epidemiology of subperiodic Brugia malayi in Malaysia: problems in its control. Acta Trop 1982;39:237-45. 45. Rolland RM, Hausfater G, Marshall B, Levy SB. Antibiotic-resistant bacteria in wild primates: increased prevalence in baboons feeding on human refuse. Appl Environ Microbiol 1985;49:791-4. 46. Turner IM. Species loss in fragments of tropical rain forest: a review of the evidence. Journal of Applied Ecology 1996;33:200-9. 47. Wilson ME. Travel and the emergence of infectious diseases. Emerg Infect Dis 1995;1:39-46. 48. Grenfell BT, Dobson AP, editors. Ecology of infectious diseases in natural populations. Cambridge: Cambridge University Press; 1995. 49. Fleagle JG, Reed KE. Comparing primate communities: a multivariate approach. J Hum Evol 1996;30:489-510. 50. Davies CR, Ayres JM, Dye C, Deane LM. Malaria infection rate of Amazonian primates increases with body weight and group size. Functional Ecology 1991;5:655-62. 51. Freeland WJ. Pathogens and the evolution of primate sociality. Biotropica 1976;8:12-24. 52. Loehle C. Social barriers to pathogen transmission in wild animal populations. Ecology 1995;76:326-35. 53. Le Guenno B, Formentry P, Wyers M, Gounon P, Walker F, Boesch C. Isolation and partial characterization of a new strain of Ebola virus. Lancet 1995;345:1271-444. 54. World Health Organization. Outbreak of Ebola haemorrhagic fever in Gabon officially declared over. Wkly Epidemiol Rec 1996;71:125-6. 55. Jezek Z, Arita I, Mutombo M, Dunn C, Nakano JH, Szczeniowski M. Four generations of probable person-to-person transmission of human monkeypox. Am J Epidemiol 1986;123:1004-12. 56. Fenner SS. Human monkeypox, a newly discovered human virus disease. In: Morse SS, editor. Emerging viruses. New York: Oxford University Press, Inc.; 1993. p. 176-183. 57. Palmer AE. Herpesvirus simiae: historical perspective. J Med Primatol 1987;16:99-130. 58. Schrag SJ, Wiener P. Emerging infectious disease: what are the relative roles of ecology and evolution? Trends in Ecology and Evolution 1995;10:319-24. 59. Karesh WB, Cook RA. Applications of veterinary medicine to in situ conservation efforts. Oryx 1995;29:244-52. 60. Hafner MS, Page RDM. Molecular phylogenies and host-parasite cospeciation: gophers and lice as a model system. Philos Trans R Soc Lond B Biol Sci 1995;349:77-83. 61. Freeland WJ. Primate social groups as biological islands. Ecology 1979;60:719-28. 62. Jin MJ, Rogers J, Phillips-Conroy JE, Allan JS, Desrosiers RC, Shaw GM, et al. Infection of a yellow baboon with simian immunodeficiency virus from African green monkeys: evidence for cross-species transmission in the wild. J Virol 1994;68:8454-60. 63. Jolly CJ, Phillips-Conroy JE, Turner TR, Broussard S, Allan JS. SIV-agm incidence over two decades in a natural population of Ethiopian grivet monkeys (Cercopithecus aethiops aethiops). J Med Primatol 1996;25:78-83. 64. Rodriguez del Valle M, Quakyi IA, Amuesi J, Quaye JT, Nkrumah FK, Taylor DW. Detection of antigens and antibodies in the urine of humans with Plasmodium falciparum malaria. J Clin Microbiol 1991;29:1236-42. 65. Karesh WB, Smith F, Frazier-Taylor H. A remote method for obtaining skin biopsy samples. Conservation Biology 1987;1:261-2. 66. Centers for Disease Control and Prevention. Tuberculosis in imported nonhuman primates-United States, June 1990-May 1993. MMWR Morb Mortal Wkly Rep 1993;42:572-6. 67. Tarara R, Suleman MA, Sapolsky R, Wabomba MJ, Else JG. Tuberculosis in wild olive baboons, Papio cynocephalus anubis (Lesson), in Kenya. J Wildl Dis 1985;21:137-40. 68. Clayton DH, Wolfe ND. The adaptive significance of self-medication. Trends in Ecology and Evolution 1995;8:60-3. 69. Robles M, Aregullin M, West J, Rodriguez E. Recent studies on the zoopharmacognosy, pharmacology and neurotoxicology of sesquiterpene lactones. Planta Med 1995;61:199-203. 70. Hill AV, Yates SN, Allsopp CE, Gupta S, Gilbert SC, Lalvani A, et al. Human leukocyte antigens and natural selection by malaria. Philos Trans R Soc Lond B Biol Sci 1994;346:379-85. 71. Skolnick AA. Newfound genetic defect hints at clues for developing novel antimalarial agents. JAMA 1993;269:1765. 72. Stephenson J. Findings on host resistance genes for infectious diseases are pointing the way to drugs, vaccines. JAMA 1996;275:1464-5. Emerging Infectious Diseases National Center for Infectious Diseases Centers for Disease Control and Prevention Atlanta, GA URL: http://www.cdc.gov/ncidod/EID/vol4no2/wolfe.htm Please note that figures and equations are not available in ASCII format; their placement within the text is noted by [fig] and [eq], respectively. Greek symbols are spelled out. The following codes are used: (ft) for footnote; (sup) for superscript; (sub) for subscript; /= for greater than or equal to. ------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------- [Emerging Infectious Diseases] [Volume 4 No. 2 / April - June 1998] Perspectives Accommodating Error Analysis in Comparison and Clustering of Molecular Fingerprints Hugh Salamon,* Mark R. Segal,* Alfredo Ponce de Leon,† and Peter M. Small‡ *University of California, San Francisco, California, USA; †Instituto Nacional Nutrición, Zubriran, Mexico City, Mexico; and ‡Stanford University Medical Center, Stanford, California, USA ------------------------------------------------------------------------- Molecular epidemiologic studies of infectious diseases rely on pathogen genotype comparisons, which usually yield patterns comprising sets of DNA fragments (DNA fingerprints). We use a highly developed genotyping system, IS6110-based restriction fragment length polymorphism analysis of Mycobacterium tuberculosis, to develop a computational method that automates comparison of large numbers of fingerprints. Because error in fragment length measurements is proportional to fragment length and is positively correlated for fragments within a lane, an align-and-count method that compensates for relative scaling of lanes reliably counts matching fragments between lanes. Results of a two-step method we developed to cluster identical fingerprints agree closely with 5 years of computer-assisted visual matching among 1,335 M. tuberculosis fingerprints. Fully documented and validated methods of automated comparison and clustering will greatly expand the scope of molecular epidemiology. The combination of conventional epidemiologic investigations with molecular techniques for genotyping pathogens has elucidated the epidemiology of many infectious diseases. The most frequently used genotyping techniques (e.g., pulsed-field gel electrophoresis, restriction fragment length polymorphism [RFLP], and randomly amplified polymorphic DNA) yield fragment-based data. Fewer than 100 patterns can be compared visually. For larger numbers, commercially available computer programs can be used to identify a manageable subset of potentially matching patterns, which are then compared visually. This approach is accurate but cumbersome and excessively labor-intensive as the number of isolates exceeds several hundred. Furthermore, the results of computer-assisted matching are not as reproducible as systematic computational methods. These limitations significantly constrain the size, scope, and standardization of molecular epidemiologic investigations. We present an approach by which identical patterns can be identified from large collections of DNA fingerprints. The number of IS6110 fingerprints continues to increase, with many studies across the globe producing IS6110 data to characterize Mycobacterium tuberculosis isolates. Such molecular epidemiologic studies provide information about tuberculosis (TB) transmission patterns. Studies in Ethiopia, Tunisia, and The Netherlands (1), South Africa (2), India (3), Denmark and Greenland (4,5), the United States (6), and Tanzania (7), among many others, exploit IS6110-based RFLP genetic fingerprints. We developed an automated computational system, in which a statistical analysis of the error in measuring fragment sizes provides a conceptual framework for comparing sets of fragment lengths. The computational approach to lane comparison—align-and-count method (ACM)—permits calculation of the number of fragments that match between two IS6110-based RFLP fingerprints. The parameters of the computational ACM are adjusted to provide the same high sensitivity as the labor-intensive visual inspection used over the last 5 years. We developed an approach to identifying a set of identical fingerprints when the identity of the fingerprints is nontransitive. We also explored improving the specificity among matched fingerprints to reveal additional information in RFLP lanes. Data Acquisition Genotyping M. tuberculosis isolates with IS6100-based RFLP fingerprinting was performed as described in van Embden et al., 1993 (8). Computer-assisted comparison and clustering were performed on the RFLP lanes as described in Woelffer et al., 1996 (9). Internal standards were used to quantitate fragment sizes visualized with a DNA probe to IS6110. Two films' exposures were scanned into Whole Band Analyzer software (BioImage, Ann Arbor, MI, USA); one was obtained when probing for the internal standard, the other when probing for IS6110. The resulting images were aligned with three registration marks that gave reference to the original nylon membrane. The Whole Band Analyzer software quantitated fragment lengths for the IS6110-visualized bands, which were inspected and edited manually in the software package. The resulting collections of fragment lengths for each lane (bacterial isolate or laboratory strain H37Rv) were exported to our ACM software and compared with other lanes. Mathematical Methods The following is a descriptive summary of the principles underlying the analysis (Appendix). Analysis of Error in Data The magnitude and characteristics of experimental error were empirically assessed by analyzing variation in the results obtained from a reference strain included in each experiment (gel). The absolute and proportional differences in the measured fragment lengths of biologically identical samples of this strain were calculated; results showed that the error in measurement was proportional to fragment length and greater between than within gels. Align-and-Count Matching Algorithm A method for matching was developed to accommodate the empirically defined error. The fragment length data from two lanes were scaled through a range of values, and the maximum number of mutually closest bands falling within a threshold tolerance was reported (Figure 1). The acceptable tolerance was smaller when lanes from the same gel were compared than when lanes from different gels were compared. An animated demonstration of this method is accessible on the Internet (URL for use with a graphics-capable Web browser: http://molepi.stanford.edu/hugh/acm/counting). A Graph-Theoretic Approach to Identical Fingerprints We considered pairs of patterns identical if all fragments matched. However, because the results might be nontransitive (A might be identical to B, B identical to C, but A not identical to C), the identification of groups of matched patterns is more complicated. A graph-theoretic approach was used to assemble clusters of matching fingerprints. Alignment and Analysis of Residual Error We further aligned collections of lanes determined to match according to the above algorithms. The optimal alignment was defined as that which minimized the proportional error between putatively identical fragments. This analytic step, which is comparable to the experimental approach of rerunning clustered strains in the same gel, improved the ability to distinguish similar, but nonidentical fingerprints. [fig 1] Figure 1. The align-and-count method finds the maximum number of mutually closest bands within a threshold deviation value[Delta], for a search across a ranges of scaling values. The two lanes are scaled incrementally, thus searching for the best alignment. Error in Analysis of H37Rv Data Investigating the absolute error for pairs of 116 H37Rv lanes (Figure 2), we found that the error was consistently higher in gel-to-gel comparisons than in comparisons of lanes from the same gel. Error was proportional to fragment length for the range of fragment lengths from 0.9 kbp up to at least 5 kbp (Figure 2), which included 90% of the bands in the fingerprint data from San Francisco. The latter empirical result was consistent with the fact that the distance migrated by a DNA fragment on an electrophoresis gel is typically proportional to the logarithm of the fragment length. Furthermore, the error found when we compared one band of a lane to that band in another lane was positively correlated to the error found for the other bands; in other words, if the first band in lane A was larger than the average measurement for that band, it was likely that the other bands in lane A would be larger than the average measurements. This positive correlation was intuitively evident in comparing lane maps (i.e., when comparing graphic representations of the fragment lengths). In Figure 3, the set of lane maps on the left are measurements of a genotype found in San Francisco. With the exception of the fourth lane from the right, they represent a set of identical patterns. Note that the error is mostly a scaling error and that if one fragment is larger than average for that lane, the others are very likely to be larger also. These two observations—that error is proportional to fragment length and positively correlated for bands within a given lane—suggests two classes of error: one is a property of each band; the second is a property of each lane. This analysis motivated us to develop a computational algorithm that scales fingerprints and measures the number of mutually closest bands within threshold sizes of each other for the best alignment, i.e., the scaling that maximizes the number of matching fragments. [fig 2] [fig 2] Figure 2. Means and two Figure 3. Additional alignment standard errors of the mean of very similar patterns can error bars for pairwise identify clearly distinct comparisons among 116 12-banded patterns. Measurement noise H37Rv lanes show that error is obscures the detailed consistently larger when relationships between 26 comparing lanes between gels patterns that were identified than when comparing lanes from from 1,335 as being very the same gel. The x-axis similar. However, after corresponds to w(b), and the alignment to a consensus y-axis to d(b), as presented in pattern, a clearly distinct the text. It is evident that pattern (an outlier from the error is proportional to other members of this fragment length in the range of autocluster) can be readily fragment lengths found in identified. Fragment lengths are H37Rv. The data exhibit 2% to given in kilobasepairs (kb). 3% error for between gel comparisons, but only approximately 1% error on average for within gel comparisons. Alignment and Residual Error for H37Rv Lanes By optimally aligning (i.e., minimizing the sum of proportional errors) pairs of H37Rv lanes, we find a distribution of scaling factors and of residual error. Table 1 shows 6,641 pairwise comparisons of 91 replicate lanes (members of each pair are taken from different gels). The mean value of s, the scaling factor, for these 6,641 alignments is 0.0212, and the standard deviation of s is 0.0189. The reduction in error due to alignment is approximately twofold, to approximately 1% with a standard deviation also of approximately 1% (Table 1); therefore, a search range, S, of 0.10 will allow for virtually every incidence of scaling error in the data. (Assuming normally distributed scaling and noting that 0.10 lies more than four standard deviations about the mean scaling error, we conclude that scaling error will not be compensated in fewer than 1 out of 10,000 independent pairwise comparisons.) Employing a deviation tolerance, [Delta], of 0.045 should ensure a sensitivity very close to 100% for matching individual bands, after alignment. (Assuming normally distributed differences between replicate band measurements and a deviation tolerance of 4.5%, which is approximately 3.5 standard deviations above the mean fragment length error after alignment, one should falsely conclude two identical bands do not match with an approximate probability of 0.0002.) These parameter values, together with the number of incremental searches, I, set to 100, empirically gave results agreeing closely with visual inspection by experienced researchers who matched entire lanes. Similarly, one may use analysis of within gel lane error to determine the parameter values for the ACM to match lanes from the same gel. Table 1. Pairwise comparisons (n=6,641) of lanes across gels characterize unaligned proportional error and residual error ----------------------------------------------------- Unaligned Aligned pairwise pairwise proportional proportional Mean error error H37Rv kilo ------------ ---------------- band bases mean s.d. mean s.d. b(sup *)w(sub b) r(b) r(sup a)(b) ---------------------------------------------------- 1 5.029 0.0253 0.0196 0.0099 0.0079 2 4.853 0.0249 0.0202 0.0084 0.0075 3 3.533 0.0253 0.0303 0.0109 0.0203 4 2.814 0.0237 0.0283 0.0105 0.0182 5 2.153 0.0202 0.0186 0.0086 0.0125 6 1.892 0.0210 0.0204 0.0100 0.0085 7 1.800 0.0226 0.0220 0.0092 0.0082 8 1.684 0.0227 0.0193 0.0083 0.0083 9 1.541 0.0221 0.0198 0.0079 0.0081 10 1.397 0.0225 0.0200 0.0093 0.0078 11 1.314 0.0231 0.0200 0.0111 0.0090 12 0.0936 0.0281 0.0249 0.0142 0.0157 mean 0.0235 0.0219 0.0099 0.0110 -------------------------------------------------- (sup *)Symbols as in Appendix. Both the range of scaling factors and the threshold deviations are derived from empirical investigation of the San Francisco data. An adjustment for larger error in measurement is included for the more rare larger fragment length bands. In applying the ACM to San Francisco bacterial fingerprints, [Delta] is allowed to increase at a rate of 0.005/kbp above a value of 7 kbp. San Francisco Bacterial Genetic Fingerprint Comparisons To evaluate the performance of the ACM, we analyzed (by computer-assisted visual inspection and by ACM) 125 lanes from bacterial isolates obtained in the first half of 1996. We evaluated ACM's performance by first determining whether a 1996 lane matched all bands to lanes in visually defined clusters (from previous years), matched other 1996 lanes, or did not find any identical matches at all. The automated matching of the 1996 lanes agreed nearly perfectly with the visual analysis; the few conflicting results were due to inconsistencies in the existing data (e.g., two bands of nearly identical size being edited sometimes as one band and sometimes as two). Humans often compensate for such inconsistencies, whereas a computational method would have to have such capabilities explicitly built in. Using ACM, we analyzed all 890,445 pairwise comparisons of isolates from 1,335 TB cases in San Francisco from 1991 to mid-1996. The autoclusters defined from the pairwise comparisons agreed closely with the clusters defined by computer-assisted visual inspection (Table 2). An example of one autocluster is shown in Figure 3. Without additional alignment, noise in fragment length measurements makes it difficult to determine if these putative clusters include individual patterns which, although similar, are distinct (outliers) or if the cluster contains identifiable subgroups of patterns (subclusters). Table 2. Preliminary autoclustering agrees closely with results obtained by computer-assisted visual inspection (CAVI) ----------------------------------------------------- Clustered Not clustered by CAVI by CAVI Total ----------------------------------------------------- Autoclustered 540 27 567 Not autoclustered 49 719 768 Total 589 746 1,335 ------------------------------------------------------- We then further analyzed autoclusters and identified more precisely nonidentical genotypes. Refinement of clusters begins with defining a consensus pattern for the cluster (consisting of the collection of mean fragment lengths for each band). Then the fragment lengths for each putative member of the autocluster are aligned to the consensus pattern. After alignment, a pattern can be easily identified as an outlier (Figure 3). If the analysis of a refined, aligned cluster shows multiple outliers, realignment of subsets of lanes can be used to reveal subclusters of identical patterns. In Figure 4 the distributions of fragment lengths before (a,b) and after (c,d) an initial alignment are presented for an autocluster of 84 2-banders from San Francisco. Given that the aligned bands are clearly split into two distributions, we split the autocluster into two subclusters. A set of 26 fingerprints (group 1) is aligned to its assembled mean-value lane (Figure 4 e,f), as is a set of 58 fingerprints (group 2, Figure 4 g,h). The contrast between the original fragment length data and the two well-aligned groups of fingerprints shows that the higher fragment length band is shifted between the two groups and no clear outliers exist after alignment. Figure 5 shows that alignment greatly improves a difficult-to-resolve clustering issue among four lanes. Preliminary investigation with a polymorphic GC-rich sequence (PGRS) fingerprinting method shows that the two subclusters exhibit distinct fingerprints, which further validates the increased specificity in IS6110 fingerprints. Of 81 PGRS genotyped autoclustered 2-banders, 63 fall into eight visually defined clusters, the remaining being unique PGRS patterns among the members of the IS6110 2-banded autocluster. Each cluster consists of isolates that all fall into one or the other IS6110-refined subcluster. Conclusions We have developed and validated a systematic approach to pairwise comparison and clustering of identical patterns in a large data set of DNA fragment-based genotypes. Incorporating a control pattern in each experiment allows the nature and magnitude of error in DNA fragment length measurements to be determined. An analysis of measurement error provides parameter values to use with algorithms that accommodate these errors. Relative scaling of entire lanes, an important characteristic of the error generated in quantitating fragment lengths from RFLP patterns used to type M. tuberculosis isolates, arises in part from aligning two images of a gel, one for internal lane size standards and one for data fragments; the image of internal standards and the image of data fragments are registered by three marks. Error in registration of the two images occurs, leading to the scaling effect in the fragment lengths reported. We strongly suggest that software for analyzing internal lane size standards and data fragments from separate images permit (and encourage) use of more than three registration marks. While internal standards compensate for idiosyncrasies in lane-specific fragment mobility, the limitations imposed by poor registration methods can result in increased scaling error. We have demonstrated that allowing for scaling, as in ACM, greatly assists in automating matching; incorporating alignment of pairs of lanes into the method has provided fully automated lane matching that agrees closely with results of the well-established method of computer-assisted visual comparison. We successfully address the nontransitivity of pairwise identity and once again use scaling of lanes to ensure the reliability of automated clustering of identical patterns. Alignment of DNA fragment patterns removes noise in clusters of fingerprints, showing further specificity in genotyping. This is analogous to the experimental approach of rerunning similar patterns on a single gel to reduce intergel noise. Mathematical transformation of the fragment length data yields similar information for far less cost in labor and materials. Further automated clustering within sets of aligned patterns could exploit the fact that we assume putative homology of fragments. As the aligned patterns have the same number of fragments, the residual error presents a multidimensional clustering problem; each pattern may deviate from the mean-value pattern in any of the fragment lengths. This clustering may prove more straightforward than the more general problem of clustering among patterns differing in numbers of fragments. [fig 4] Figure 4. Histograms of the fragment lengths for 84 two-banded patterns connected by identity (autoclustered with in-house software) exhibit enough spread in values to make detecting outliers and band shifts difficult (a,b). Aligning the 84 lanes to the mean-value lane for this collection reveals that the lanes do not align well, but instead shows bimodal distributions for the fragment lengths (c,d). Dividing the 84 fingerprints into two sets and separating the distinct distributions detected when aligning all 84 fingerprints show that 26 fingerprints align well to their mean-value lane (e,f), and the remaining 58 also align well to their respective mean value lane (g,h). The smaller fragment length fragment does not appear shifted between the two sets of 2-banders(comparing e to g), but the larger fragment is clearly shifted (comparing f to h). Numerous commercial packages of computer software are available to compare and match DNA fragment patterns. The availability of these systems fosters unquestioning application of turnkey pattern matching and clustering methods, which are not always fully documented or validated (for fragment-based genotypes) in the scientific literature. While possibly acceptable for small studies when visually validated, this approach to data analysis is risky for large studies. The methods for matching DNA fragment patterns presented in this paper should be an adjunct to software packages that quantitate fragments. We provide a systematic approach to analysis of fragment length estimates, appropriate whether the numerical data are generated by hand using a ruler and arithmetic or are output by a multithousand dollar gel analyzer. This approach can thus be used by molecular epidemiologists working with both large and small budgets around the world. We anticipate that our methods can be incorporated into existing commercial software packages with broad distribution and encourage similar documentation in peer-reviewed journals of other methods provided in software packages. In addition, our focus on the use of fragment length data, as opposed to the comparison of actual images, will fostercomparison of data generated in different laboratories that use different proprietary software. We are working on a World-Wide Web–based system to facilitate IS6110 genotype data sharing. The availability of a precise and validated method to count the number of matching fragments in a pairwise fashion among very large numbers of patterns now permits an assessment of the importance and usefulness of approaches that exploit fingerprint similarity. In conjunction with the growing understanding of the underlying biology of IS6110 instability and the relevant statistical issues, this may greatly expand the scope of TB molecular epidemiology. A general question arises when comparing molecular fingerprints: how many bands need to match to indicate a close biologic relationship? Aside from the issues of defining biologic "closeness"—too often ignored in epidemiologic studies—technical issues are relevant to ACM. The usefulness of the matched band information output by ACM depends on at least several factors: the underlying band size distribution from which fingerprint bands are sampled, the independence of sampled bands, measurement error (both scaling and independent band error), the length of the gel, and the number of fragments. At one extreme, where error is large and few bands are observed, even in the case of a perfect match, statistical analysis may fail to reject coincidental band matches. By using computer simulation and sets of assumptions regarding band size distributions, one may learn about the role of coincidental band matching. We are actively researching these issues for IS6110 fingerprint comparisons. Furthermore, we intend to provide a general computational framework in which one may assess error in laboratory measurement, the appropriateness of ACM for analyzing data (and appropriate parameter values to use with the method), and the role of coincidence in band matching. Information regarding computer programs for various tasks, including ACM matching itself, will be made available on the Internet at http://molepi.stanford.edu/hugh/acm. fig 5] Figure 5. Prior to alignment of two sets of 2-banders, lanes are difficult to cluster(lanes a-d are from the distributions in Figures 4a and 4b). Subsequent to alignment, lanes are much easier to cluster (lanes a' and b' are specific examples from the distributions in Figures 4e and 4f; lanes c' and d' likewise correspond to Figures 4g and 4h). Fragment lengths are given in kilobasepairs (kb). A study of M. tuberculosis isolates from northern Tanzania demonstrates the utility of partial matching (3). The study brings into focus difficulties inherent in employing one-parameter tolerance for DNA fragment-based genotype matching, a technical issue effectively addressed by ACM. Gillespie et al. (7) also call attention to the poor specificity of low copy number IS6110- based fingerprints, exacerbated by the use of the Dice coefficient. We are pursuing alternative similarity measures that use the numbers of matching fragments identified by ACM and are tailored to the needs of epidemiologic investigations. Dendogram clustering methods often provided in software targeted at DNA-fragment genotype management and analysis could in some instances fail to reconstruct the correct relationships among infectious organism isolates, even when presented perfect clock-like genetic distances. This may result in part from the fact that fast-evolving markers are characterized for isolates sampled over a period; samples are not contemporaneous. In conjunction with our efforts to define similarity measures, we are also working to modify phylogenetic inference tools used in clinical and molecular epidemiologic settings to better handle the data typically analyzed in molecular fingerprint management and analysis software. Appendix Analysis of Error in Data To characterize error in fragment length measurement, pairs of 12 band 7H37Rv lanes are compared; the difference between fragment lengths of each band is calculated as follows. Let w(sub i,b) be the measured fragment length of band b of lane i. In general, let B be the number of bands. To compare lanes i and j, we can calculate the absolute difference between the measured lengths, d(sub i,j)(b) = |w(sub i,b) - w(sub j,b)|. Let lower case DELTA(sub i,j) be an indicator that equals 1 when lanes i and j are from the same electrophoresis gel and 0 otherwise. Let n be the total number of replicate lanes. The mean absolute difference for a fragment over all pairwise comparisons of lanes from different gels is found by, (eq) d(b)(overscore)=1(over)t(sup i=n-1)EPSILON(sub i=1)(sup j=n)EPSILON(sub j>i) d(sub i,j)times (1-lower case delta(sup i,j)(sup 3) where (eq) t= EPSILON(sup i=n-1)(sub i=t) EPSILON(sup i=n-1) (sub i=t)(1-lower case delta(sup i,j) is the number of pairs of lanes from different gels. Similarly, we calculate the proportional difference between measurements for a fragment, (eq)r(sub i,j)(b)=|w(sub i,b)- w(sub j,b)| over (w(sub i,b)+ w(sub j,b))/2(sup 2) and its mean for comparisons between lanes from different gels, (eq) r(b)(overscore)=1(over)t(sup i=n-1)EPSILON(sub i=1)(sup j=n)EPSILON(sub j>i) r(sub i,j)times (1-lower case delta(sup i,j)(sub -) Calculating fragment measurement errors for the (eq) u= EPSILON(sup i=n-1)(sub i=1) EPSILON(sup j=n-1)(sub j>i)times lower case delta(sup i,j) pairwise comparisons of lanes from the same gel are also performed. Align-and-Count Matching Algorithm The Align-and-Count Method (ACM) for counting the matching bands between two lanes is defined as follows. Consider two lanes, lane A with m bands, w(sub A,x) , x = 1,...,m, and lane B with n bands, wB,y , y = 1,...,n. We count the mutually closest measured fragment lengths within a proportional deviation factor, [delta] , over a range of alignments (Figure 1). Alignments are searched by scaling the fragment lengths. Multiplying the fragment lengths in a lane by a scaling factor reflects the phenomenon that error in fragment lengths is proportional to fragment length and is positively correlated for fragments in a lane. Define (eq)Match [w(sub A,x),w(sub B,y)]={1 over 0 if |w (sub B,y)- w (sub A,x)|10 years) on small rodent abundance are scarce. However, in Sweden such data are available from the area with cyclic rodent populations (2, 4-8). We used data from Grimsö (59°40' N, 15°25' E), where small rodents have been collected every autumn since 1973 (6-8). Snap-trapping was performed where possible in six adjacent 5x5-km areas, each with four 1-ha plots. Of 24 permanent 1-ha plots, 20 were trapped. At each trapping, approximately 940 traps were set during three consecutive days, amounting to approximately 2,800 "trap-nights." Species, date, and location of trapped animals were recorded (4, 6-8). As an index of bank vole abundance, we calculated the number of animals trapped per 100 trap nights (i.e., density). Bank vole abundance fluctuated in cycles with 3- or 4-year intervals between density peaks (Figure 1). [fig 1] Figure 1. Bank vole abundance in Grimsö, 1973-150;1994. Untransformed data. Incidence of Disease We used the national census to calculate (according to counties and age groups) annual disease incidences per 100,000 population. Because of the lack of long-term records on noncyclic small rodent populations, we analyzed data from the counties with cyclic rodent populations: Norrbotten, Västerbotten, Västernorrland, Jämtland, Gävleborg, Kopparberg, Värmland, Örebro, Västmanland, Uppsala, Stockholm, Södermanland, Göteborg, -lvsborg, Jönköping, and Kronoberg (2-3). We excluded Stockholm and Uppsala counties as they are predominantly urban and on the border between counties with cyclic and noncyclic rodents. Myocarditis Myocarditis is a clinical condition in which the heart is infiltrated by inflammatory cells. Diagnosis is very difficult during the acute stage and is often made by postmortem microscopic investigation of the myocardium. All causes of death in Sweden are registered by the Swedish Cause-of-Death Statistics, Statistics Sweden, S-115 81 Stockholm, Sweden. Our study included all deaths between 1970 and 1986, in the age group 11 to 46 years, caused by myocarditis (ICD 8 code 422), if the diagnosis was given as either the direct (primary) cause of death or as the first out of six recorded contributing causes of death. The age group 11 to 46 years was chosen because congenital or arteriosclerotic cardiovascular disease is less common in this age group. In 201 (92%) of 218 cases, the diagnoses were based on clinical or forensic autopsies. The remaining diagnoses were based on clinical examination before death. We did not use data collected after 1986, as the disease classification changed from ICD 8 to ICD 9 in 1987, making comparison with the period after 1986 difficult. Guillain-Barré Syndrome In different parts of Europe (9), reported incidence of GBS varies considerably depending on study method. Our data, from the hospital discharge registry, Swedish National Board of Health and Welfare, S-106 30 Stockholm, Sweden, are considered adequate for epidemiologic surveillance of GBS (9). We selected patients hospitalized with the diagnosis GBS (ICD 8 code 354) and gathered information on age, sex, county of residence, county of hospitalization, diagnosis, date of admission and date of discharge. We only analyzed data of GBS patients 46 years of age and younger, as the diagnoses in older patients frequently are obscured by nonspecific illnesses and other problems (10). Insulin-Dependent Diabetes Mellitus IDDM was studied in Medelpad (part of Väternorrland County), an area with cyclic rodent populations. The patients were identified through registries of patients with diabetes at the only hospital and at 16 out of the 17 health-care centers, as well as through common registries of diagnosis at one health-care center. The World Health Organization (WHO) diagnostic criteria were applied. We analyzed case data in all age groups because precision in the diagnosis of IDDM is not age related. Statistical Analysis A cross-correlation function (CCF) (11) can indicate any direct or delayed dependence between two different time series. A CCF graph is a plot of positive and negative correlation coefficients between pair-wise values from the two series. The correlations have been calculated between the two series for different time shifts (lags), with lag 0 meaning no time shift, lag -1 meaning the first series has been shifted backwards one time unit, and lag 1 meaning the first series has been shifted forward by one time unit. We used CCFs, calculated in Statistical Package for the Social Sciences (SPSS) (12), to indicate any temporal association between the incidence in cyclic versus noncyclic areas of GBS and of deaths from myocarditis, respectively. We also used CCFs to indicate any direct (at lag 0) or delayed dependence (at negative lags) of the different disease incidences on vole density. At negative lags, the vole time series is the leading indicator that is shifted "backwards" along the time scale, relative to the disease series. Thus, the correlation coefficients at negative lags are in focus, as they indicate any disease incidence dependence on past vole densities, i.e., on densities 1, 2, or 3 years ago. At positive lags, the disease time series is shifted "backwards" relative to the vole time series. The correlation coefficients at positive lags may be large because of the similarities between the two series, but we see no causal relationship between vole density and past disease incidence. Log transformation of the time series was used because of the variable amplitudes of the fluctuations in the disease and vole series. In the case of IDDM, we also ran one CCF with the disease and vole time series difference-transformed by one year. Findings Myocarditis A total of 218 patients, ages 11 to 46 years, died of acute myocarditis in 1970 to 1986; 148 in counties with cyclic rodent density and 70 in counties with noncyclic rodent density. Sex ratio was 2:1 (146 male, 72 female patients). Myocarditis death incidences in cyclic and noncyclic areas (Figure 2) were not correlated as revealed by cross-correlation of the disease series (CCF not shown; n = 17 computable 0-order correlations). Cross-correlation of the annual incidence of myocarditis deaths in 1970 to 1986 in the cyclic area with bank vole abundance (Figure 3) showed that the incidence of myocarditis was most highly correlated with vole abundance in the previous year, according to the high positive and significant correlation at lag -1 in the CCF (Figures 4, 5) (r = 0.635, p < 0.05, n = 13). [fig 2] [fig3] Figure 2. Incidence of death from Figure 3. Myocarditis deaths, myocarditis, 1970–1986. 1974–1986 relative to bank vole Untransformed data. abundance 1 year previously (vole data from 1973-1985). Untransformed data. [fig 4] [fig 5] Figure 4. Cross-correlation Figure 5. Myocarditis deaths, function of incidence of death from 1974–1986, relative to bank vole myocarditis with bank vole abundance 1 year previously (vole abundance, 1973–1986. Time series data from 1973-1985). Log are log transformed; n = 14 transformed data. r = 0.635, n = 13. computable 0-order correlations. Lines represent + 2 SE. The standard error is based on the assumption that the series are not cross-correlated and one of the series is white noise. Guillain-Barré Syndrome In five cyclic counties and two noncyclic counties, where complete data were available, 258 GBS patients (300 times (4), this rodent is likely to have a high and highly variable potential over time to transmit any etiologic agent to humans. In addition, bank voles are known to enter buildings, thus transporting disease risk closer to humans. Three novel virus isolates from bank voles, resembling picornaviruses in size and morphology, have been found. The first isolate was named Ljungan after the Ljungan river in Medelpad County, Sweden, where the animals were trapped. The second and third isolate originated from animals trapped in Västerbotten County. The amino acid sequences of predicted Ljungan virus capsid proteins were closely related (approximately 70% similarity) to the human pathogen echovirus 22. Partial 5' noncoding region sequence of Ljungan virus was most closely related to cardioviruses. Two additional isolates were serologically and molecularly related to the prototype (Niklasson et al., unpub. observation). Echovirus 22 is a known human pathogen, and other known cardioviruses can induce not only myocarditis but also neurologic diseases and IDDM in several species of animals. We hope to elucidate the role of Ljungan virus as a human pathogen by serologic assays for Ljungan virus, diagnostic polymerase chain reaction based on generated sequence data, and specific antisera for viral antigen detection. Studies are also under way to identify and determine the role of other viruses of these small rodents in inducing myocarditis, GBS, and IDDM in humans. Acknowledgments We thank Sara Sjöstedt for statistical help and Stiftelsen Olle Engkvist, Byggmästare and by the Swedish Environment Protection Agency (via the National Environmental Monitoring Programme) for financial support for the vole trapping. Address for correspondence: Bo Niklasson, Swedish Institute for Infectious Disease Control, S-105 21 Stockholm, Sweden; fax: 46-8-735-66-15; e-mail BO.NIKLASSON@SMI.KI.SE References 1. Niklasson B, Le Duc J. Epidemiology of nephropathia epidemica in Sweden. J Infect Dis 1987;155:269-76. 2. Hansson L, Henttonen H. Gradients in density variations of small rodents: the importance of latitude and snow cover. Oecologia 1985; 67:394-402. 3. Hansson L, Henttonen H. Rodent dynamics as community processes. Trends in Ecology & Evolution 1988;3:195-200. 4. Hörnfeldt B. Delayed density dependence as a determinant of vole cycles. Ecology 1994;75:791-806. 5. Niklasson B, Hörnfeldt B, Lundkvist Ĺ, Björsten S, LeDuc J. Temporal dynamics of Puumala virus antibody prevalence in voles and of nephropathia epidemica incidence in humans. Am J Trop Med Hyg 1995;53:134-40. 6. Hörnfeldt B. Smsdäggdjursinventering i PMK:s referensomrĺden - rapport frĺn verksamheten 1992. Solna, Sweden: Swedish Environmental Protection Agency; 1994. Report 4294. 7. Lindström ER, Andrén H, Angelstam P, Cederlund G, Hörnfeldt B, Jäderberg L, Lemnell P-A, Martinsson B, Sköld K, Swenson JE. Disease reveals the predator: sarcoptic mange, red fox predation and prey populations. Ecology 1994;75:1042-9. 8. Lindström ER, Hörnfeldt B. Vole cycles, snow depth and fox predation. Oikos 1994;70:156-60. 9. Jiang GX, de Pedro-Cuesta J, Fredriksson G. Guillain-Barré syndrome in South-West Stockholm, 1973-1991. Quality of registered hospital diagnosis and incidence. Acta Neurol Scand 1995;91:109-17. 10. Winner SJ, Evans JG. Age-specific incidence of Guillain-BarrT syndrome in Oxfordshire. Quarterly Journal of Medicine, New Series 1990;77:1297-1304. 11. Box GEP, Jenkins GM. Time series analysis, forecasting and control. San Francisco: Holden Day; 1970. 12. SPSS Base System Syntax Reference Guide, Release 6.0. Carey (NC): SPSS Inc.; 1993. 13. Myhrman G. En njursjukdom med egenartad symptombild. (In Swedish). Nord Med Tidskr 1934;7:793-94. 14. Gajdusek DC. Virus haemorrhagic fevers-special reference to haemorrhagic fevers with renal syndrome (epidemic haemorrhagic fever). J Pediatr 1962;60:841-57. 15. WesslTn L, Pshlson C, Friman G, Fohlman J, Lindquist O, Johansson C. Myocarditis caused by Chlamydia pneumonie (TWAR) and sudden unexpected death in a Swedish elite orienteer. Lancet 1992;340:427-8. 16. Hörnfeldt B. Synchronous population fluctuations in voles, small game, owls and tularemia in northern Sweden. Oecologia 1978; 32:141-152. 17. Hörnfeldt B, Löfgren O, Carlsson B-G. Cycles in voles and small game in relation to variations in plant production indices in northern Sweden. Oecologia 1986;68:496-502. Emerging Infectious Diseases National Center for Infectious Diseases Centers for Disease Control and Prevention Atlanta, GA URL: http://www.cdc.gov/ncidod/EID/vol4no2/niklason.htm Please note that figures and equations are not available in ASCII format; their placement within the text is noted by [fig] and [eq], respectively. Greek symbols are spelled out. The following codes are used: (ft) for footnote; (sup) for superscript; (sub) for subscript; /= for greater than or equal to. ------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------- [Emerging Infectious Diseases] [Volume 4 No. 2 / April - June 1998] Perspectives Multidrug-Resistant Mycobacterium tuberculosis: Molecular Perspectives Ashok Rattan, Awdhesh Kalia, and Nishat Ahmad All India Institute of Medical Sciences, Ansari Nagar, New Delhi, India -------------------------------------------------- Multidrug-resistant strains of Mycobacterium tuberculosis seriously threaten tuberculosis (TB) control and prevention efforts. Molecular studies of the mechanism of action of antitubercular drugs have elucidated the genetic basis of drug resistance in M. tuberculosis. Drug resistance in M. tuberculosis is attributed primarily to the accumulation of mutations in the drug target genes; these mutations lead either to an altered target (e.g., RNA polymerase and catalase-peroxidase in rifampicin and isoniazid resistance, respectively) or to a change in titration of the drug (e.g., InhA in isoniazid resistance). Development of specific mechanism–based inhibitors and techniques to rapidly detect multidrug resistance will require further studies addressing the drug and drug-target interaction. In the last decade, tuberculosis (TB) has reemerged as one of the leading causes of death (nearly 3 million deaths annually) (1). The estimated 8.8 million new cases every year correspond to 52,000 deaths per week or more than 7,000 each day, which translates into more than 1,000 new cases every hour, every day (2,3). These death rates, however, only partially depict the global TB threat; more than 80% of TB patients are in the economically productive age of 15 to 49 years. The emergence of AIDS and decline of socioeconomic standards contribute to the disease's resurgence in industrialized countries (4). In most developing countries, although the disease has always been endemic, its severity has increased because of the global HIV pandemic and extensive social restructuring due to rapid industrialization and conflicts. A major public health problem worldwide, TB is now a global emergency (Figure 1). [fig 1] Figure 1. Global incidence of tuberculosis. Of the estimated 8.8 million cases worldwide, more than 40% of the cases are in Southeast Asia; India has approximately 53.3% of those cases. A, Americas; Afr, Africa; WP, Western Pacific; E, Europe; M, Eastern Mediterranean; and SEA, Southeast Asia; Ind, Indonesia; B, Bangladesh; Thai, Thailand; My, Myanmar. *Others include Bhutan, 0.05%; Nepal, 1.2%; Maldives, 0.001%; Sri Lanka, 1%; DPR Korea, 1.2%. (Data from reference 2). Short-course chemotherapy forms the backbone of antitubercular chemotherapy (5). Proper prescriptions and patient compliance almost always cure. In fact, TB incidence was steadily declining in most industrialized countries, until the trend was reversed (6). Further contributing to the increased death rate is the emergence of new strains of M. tuberculosis resistant to some or all current antitubercular drugs. The resistance is attributed primarily to improper prescriptions or patient noncompliance and is often a corollary to HIV infection (7-9). Multidrug-resistant TB (MDRTB), associated with high death rates of 50% to 80%, spans a relatively short time (4 to 16 weeks) from diagnosis to death (10). Delayed recognition of drug resistance, which results in delayed initiation of effective therapy, is one of the major factors contributing to MDRTB outbreaks, especially in health-care facilities (11,12). In most countries, MDRTB has increased in incidence and interferes with TB control programs, particularly in developing countries, where prevalence rates are as high as 48% (13,14). The high infection and death rates pose an urgent challenge to rapidly detect cases. In the past few years, genetic and molecular insights have unraveled the mechanisms involved in the acquisition of drug resistance by Mycobacterium tuberculosis (MTB), concomitant with the development of various molecular strategies to rapidly detect MDRTB. In this review, we examine the status of the mechanisms of resistance to antitubercular drugs. MDRTB and the Mechanisms of Resistance Currently TB is treated with an initial intensive 2-month regime comprising multiple antibiotics—rifampicin (RIF), isoniazid (INH), pyrazinamide (PZA), and ethambutol (EMB) or streptomycin (SM)—to ensure that mutants resistant to a single drug do not emerge (15). The next 4 months, only RIF and INH are administered to eliminate any persisting tubercle bacilli. INH and RIF, the two most potent antituberculous drugs, kill more than 99% of tubercule bacilli within 2 months of initiation of therapy (16,17). Along with these two drugs, PZA, with a high sterilizing effect, appears to act on semidormant bacilli not affected by any other antitubercular drugs (18). Using these drugs in conjunction with each other reduces antitubercular therapy from 18 months to 6 months. Therefore, the emergence of strains resistant to either of these drugs causes major concern, as it leaves only drugs that are far less effective, have more toxic side effects, and result in higher death rates, especially among HIV-infected persons. The phrase "MDR state" in mycobacteriology refers to simultaneous resistance to at least RIF and INH (19) (with or without resistance to other drugs). Genetic and molecular analysis of drug resistance in MTB suggests that resistance is usually acquired by the bacilli either by alteration of the drug target through mutation (20) or by titration of the drug through overproduction of the target (21). MDRTB results primarily from accumulation of mutations in individual drug target genes (Table). The probability of resistance is very high for less effective antitubercular drugs such as thiacetazone, ethionamide, capreomycin, cycloserine, and viomycin [10(sub -3)]; intermediate for drugs such as INH, SM, EMB, kanamycin, and p-amino salicylic acid [10(sub -6)]; and lowest for RIF [10(sub -8)] (22,23). Consequently, the probability of a mutation is directly proportional to the bacterial load. A bacillary load of 109 will contain several mutants resistant to any one antitubercular drug (24). Because the mutations conferring drug resistance are chromosomal, the likelihood of a mutant being simultaneously resistant to two or more drugs is the product of individual probabilities; thus the probability of MDR is multiplicative. Resistance to a drug does not confer any selective advantage to the bacterium unless it is exposed to that drug (19). Under such circumstances, the sensitive strains are killed and the drug-resistant mutants flourish. When the patient is exposed to a second course of drug therapy with yet another drug, mutants resistant to the new drug are selected, and the patient may eventually have bacilli resistant to two or more drugs. Serial selection of drug resistance, thus, is the predominant mechanism for the development of MDR strains; the patients with MDR strains constitute a pool of chronic infections, which propagate primary MDR resistance. In addition to accumulation of mutations in the individual drug target genes, the permeability barrier imposed by the MTB cell wall can also contribute to the development of low-level drug resistance. Studies addressing resistance to SM have found evidence of such a two-step mechanism for the development of drug resistance (119,120). Table. Gene loci involved in conferring drug-resistance in Mycobacterium tuberculosis. -------------------------------------------------------------------------------- Reported frequency in resistant strains(sup a) Drug Gene Product (%) Reference -------------------------------------------------------------------------------- Rifampicin rpoB B-subunit of RNA >95 45-48,68-71 polymerase Isoniazid katG Catalase-peroxidase 60-70 39-48 oxyR-ahpC Alky hydro-reductase ~20 36 INH-Ethionamide inhA Enoyl-ACP reductase <10 46-48 Streptomycin rpsL Ribosomal protein S12 60 46-48 rrs 16s rRNA <10 113-117 Fluoroquinolone gyrA DNA gyrase >90 107 Pyrazinamide pncA Amidase 70-100 92-94 Ethambutol embCAB EmbCAB 69 88 -------------------------------------------------------------------------------- (sup a)Mutation frequencies are as determined by sequencing and polymerase chain reaction-single strand conformational polymorphism (PCR-SSCP) analysis. Resistance to INH INH (isonicotinic acid hydrazide, 4-pyridinecarboxylic acid hydrazide), highly active against the MTB complex (M. tuberculosis, M. bovis, M. africanum, and M. microti), has very low MICs (0.02 µg/ml to 0.06 µg/ml) (25). The mechanism of action of INH, as well as mechanisms conferring INH resistance, are complex and not completely understood (Figure 2). However, evidence suggests that INH inhibits the biosynthesis of cell wall mycolic acids (long-chain alpha-branched ß-hydroxylated fatty acids), thereby making the mycobacteria susceptible to reactive oxygen radicals and other environmental factors. Activation of INH to an unstable electrophilic intermediate requires the enzyme catalase-peroxidase (KatG, coded by katG) and an electron sink (hydrogen peroxide) (26), although hydrazine formed after INH spontaneously decomposes may also mediate activation of INH (27). Nevertheless, KatG is the only enzyme capable of activating INH, and consequently, KatG mutant MTB strains are invariably INH resistant. [fig 2] Figure 2. Mechanism of action of isoniazid (INH); acquisition of resistance and combating oxidative stress. DPR, divergent promoter region. Early studies by Middlebrook demonstrated that INH resistance was associated with loss of catalase activity (28). Genetic studies demonstrated that transformation of INH-resistant M. smegmatis and MTB strains with a functional KatG restored INH susceptibility and put forth the hypothesis that katG deletion may cause INH resistance in MTB (29,30). However, in the absence of a peroxide-inducible genetic response, mediated in most bacteria by the transcription factor OxyR (31), KatG is the only peroxide-inducible MTB protein (32). Consequently, MTB resistance to INH is paradoxical; it has to sacrifice KatG function. MTB's ability to adapt to the loss of KatG function and combat organic peroxides is remarkable. Studies conducted by Sherman et al. demonstrated that all KatG mutant MTB strains overexpressed a 22-kD protein at levels significantly higher than INH-sensitive strains (33). Sequence analysis confirmed that this protein was similar to the earlier reported MTB AhpC protein. AhpC can detoxify organic peroxides and is homologous to other bacterial and eukaryotic proteins with alkyl hydroperoxidase and thioredoxin-dependent peroxidase activities (34,35). The 5' regions (39 to 81 bp upstream from the ahpC start codon) of each AhpC-upregulated (and katG mutant) isolate contained mutations that could increase promoter activity; it was proposed that compensatory mutations in the ahpC promoters were selected in katG mutant strains to combat oxidative stress (33). Subsequent studies using immunoblotting experiments demonstrated the consistency of AhpC upregulation among clinical isolates with complete deletion of katG (36,37). katG mutant isolates with variable residual KatG activity did not have this strict linear relationship. Characterization of the oxyR-ahpC region further demonstrated that mutations responsible for AhpC upregulation occurred at low frequencies and were primarily G>C to A>T transitions localized in the oxyR-ahpC intervening region (36). Although the sequence alterations in the oxyR-ahpC region were predominantly restricted to INH-resistant isolates, not all alterations detectably increased the AhpC levels. The apparent rarity of AhpC upregulation among INH-resistant and katG mutant isolates could be attributed partially to the rare occurrence of MTB strains with complete katG deletion (38-41,44). Alternatively, among katG mutant isolates, selection of AhpC upregulatory mutations may be subject to the selective pressure exerted by residual catalase-peroxidase activity (36). However, AhpC upregulation was not observed among MTB isolates with katG315 codon mutations, which reportedly lead to more than a 20-fold decrease in KatG activity and confer high MICs against INH (>90 µg/ml) (42, 43). This inconsistency and rarity of AhpC upregulation among katG mutant INH-resistant isolates indicates a more complex relationship between the two and underlines the need for in-depth studies to determine precisely the conditions regulating AhpC expression. Clinical studies to validate the paradigm of katG deletions and INH resistance showed that complete deletion rarely occurred (38-41). We constructed a 35-mer oligonucleotide probe specific for katG gene. Southern hybridization demonstrated the presence of katG in all INH-resistant isolates, precluding complete deletion of katG gene as a dominant mechanism for INH resistance (44). Previous studies using polymerase chain reaction (PCR) amplification had also established these findings; sequence analysis of katG from INH-resistant strains showed randomly distributed mutations, including point mutations and deletions and insertions of up to 1 to 3 bases (38-41). These mutations could disrupt the katG gene, leading to the production of an inactive gene product or a gene product with compromised peroxidative activity. PCR amplification of the katG gene followed by single strand conformational polymorphism (SSCP) detected mobility shifts supporting the presence of these mutations and thereby INH resistance. Our analysis of the katG gene by PCR-SSCP resulted in the amplification of the 237 bp fragment of the katG gene and demonstrated a 67.3% (n = 19) correlation between mutations in the katG gene and INH resistance (45). The results were consistent with those from earlier studies indicating that katG gene mutations had a correlation rate of less than 60% to 70% with INH resistance (46-48). Sequence analysis of INH-resistant strains demonstrating altered SSCP patterns showed that the most common mutation was G>T transversion in codon 463 (42). In this G>T change, Leu is substituted for Arg, and the restriction site for NciI and MspI is lost (40). Polymorphism in the katG locus can then be easily detected by restriction digestion. Recent kinetic and spectroscopic studies have demonstrated striking similarities between KatG from wild-type strains and the R463L mutant isolates (49). Both enzymes had similar visible and electron-paramagnetic-resonance spectra and similar ability to oxidize INH and inactivate InhA. Further, when the INH-resistant katG-defective strains of M. smegmatis with wild-type katG or the R463L katG were transformed, INH susceptibility was restored to about the same extent (50). These similarities do not support the contention that the R463L mutation of katG allows discrimination against INH as a substrate and thereby confers resistance to INH. Although the exact role of the R463L mutation of katG requires further scrutiny, this mutation may be a frequent polymorphism and may not affect INH susceptibility. Other common mutations resulting in an attenuated KatG have been identified primarily as missense mutations that result in single amino acid substitutions (46-48). While the data point towards mutations in the katG gene as the dominant mechanism for INH resistance, they also point to other factors that could mediate MTB acquisition of resistance to INH. Mutations in the oxyR regulon, from which AhpC is divergently transcribed, could explain the acquisition of INH resistance in the remaining INH-resistant isolates (33,51). OxyR confers high-level intrinsic resistance to INH in Escherichia coli and Salmonella Typhimurium; mutations in the oxyR or AhpC restore INH susceptibility in these species (51). The MTB oxyR regulon is much smaller than in M. leprae and other mycobacteria—because of two important deletions of 29 bp and 372 bp (32,52). In addition to these deletions, the oxyR regulon carries many frame shift mutations, which result in low expression of this regulon and eventually lead to low-level expression of AhpC (consistent with the finding of low-level expression of AhpC in INH-sensitive strains vs. INH-resistant strains) (33). A related member of the genus resistant to INH, M. leprae, however, has a complete oxyR-ahpC region that is transcriptionally fully active and may play a role in the detoxification of active INH intermediates (52). By analogy, therefore, the loss of the OxyR function, in conjunction with its putative effects on ahpC expression, could explain the exquisite specificity of INH for the MTB complex. However, evidence from recent studies does not indicate a direct role for oxyR or the ahpC genes in determining susceptibility to INH (36,37). Polymorphisms in oxyR do not have any preferential predisposition and exist among both INH-resistant and -susceptible isolates with about the same frequency (36). The relationship of AhpC overexpression to INH resistance is more complex. Earlier observations based on transformation of M. smegmatis strains suggested a possible involvement of AhpC overexpression in acquiring INH resistance (53). Transformation of M. smegmatis isolates with multicopy constructs of ahpC led to almost a fivefold increase in the MIC for INH. However, an increasing body of evidence precludes any direct role of AhpC in determining INH susceptibility among MTB isolates. MTB transformants bearing multicopy constructs of ahpC did not demonstrate significant increase in the MIC for INH, thus any direct role for AhpC in acquisition of INH resistance was ruled out (37). Efforts to determine the factors involved in resistance to INH led to the discovery of the inhA locus, which was proposed as the primary target for coresistance to INH and ethionamide (54). This locus is composed of two open reading frames (ORFs), designated orf1 and inhA, separated by a 21-bp noncoding region. InhA, an enoyl-ACP reductase (55), more than 40% homologous to the EnvM protein, catalyzes an early step in fatty acid synthesis among enterobacteria. Like EnvM, InhA activity is also thought to use NAD(H) as cofactor. INH susceptibility could result from incorporation of iso-NAD, which is formed as a consequence of the action of KatG on INH, and thus hinders the enzymatic activity of InhA and blocking fatty acid synthesis (56). A T>G transversion, observed in few of the resistant strains, at position 280 in the inhA gene, results in the ser94 to ala94 replacement (54). This replacement, thought to alter the binding affinity of InhA to NAD(H), ultimately results in INH resistance (57). Alternatively, because of mutations in the putative promoter region, hyperexpression of InhA could result in INH resistance. Studies conducted in clinical settings to provide corroborating evidence of mutations in the inhA locus and INH resistance have shown approximately 10% correlation (46-48). Analysis of 37 INH-resistant isolates by Kapur et al. demonstrated no ser94-ala94 substitution in the resistant isolates. Only one isolate had a missense mutation: ATC>ACC at position 47, resulting in substitution of Ile16 by Thr16. Morris et al. also demonstrated the lack of mutations in the inhA gene among 42 INH-resistant MTB isolates. However, five of the INH-resistant isolates showed single nucleotide mutations in the putative inhA regulatory region upstream of orf1. Subsequent biochemical characterization of InhA function demonstrated that it catalyzed the reduction of 2-trans-octenoyl-acyl carrier protein and also that protein of enoyl CoA esters (58-59), thereby acting at the final step in chain elongation in fatty acid synthesis (58). This observation contradicted earlier biochemical evidence suggesting that an enzyme involved in the synthesis of an unsaturated 24-carbon fatty acid was the target for activated INH (60,61). Thus, the targets identified biochemically and by complementation of M. smegmatis are different. Lipid pulse labeling experiments demonstrated that the lipid biosynthetic response of M. smegmatis and MTB after exposure with INH were different (62), indicating a different mechanism of action for the INH intermediate in the two species. Transformation of M. smegmatis with single-copy alleles of mutant inhA loci did not result in significant resistance to INH, indicating the presence of a different promoter in M. smegmatis. Further, the inability of multicopy vector constructs bearing the inhA gene to significantly increase the MIC for INH provided substantiating evidence for the limited involvement of this locus in mediating INH resistance among MTB isolates. These data, along with clinical evidence, preclude the likelihood that inhA is the primary target for the activated form of INH. Functional characterization of inhA mutations, occurring with katG mutations (as observed in isolates with very high MICs) (46) in relation to lipid metabolism of INH-resistant isolates, could perhaps resolve this discrepancy and delineate the roles of the respective loci in the mechanism of action of INH and subsequent acquisition of drug resistance. In summary, mutations in the katG and the inhA genes are associated with approximately 70% to 80% of INH-resistant MTB isolates; molecular mechanisms operating in the remaining isolates are still unknown. The role of the MTB cell wall as an important permeability barrier needs to be explored in greater detail, particularly with reference to INH resistance (56). Resistance to RIF RIF, first introduced in 1972 as an antitubercular drug, is extremely effective against MTB. It has MICs of 0.1 µg to 0.2 µg (16,63). Because of its high bactericidal action, RIF, along with INH, forms the backbone of short-course chemotherapy (5). Although rare, resistance to RIF is increasing because of widespread application and results in selection of mutants resistant to other components of short-course chemotherapy. In this context, resistance to RIF can be assumed to be a surrogate marker for MDRTB (19). RIF had long been believed to target the mycobacterial RNA polymerase and thereby kill the organism by interfering in the transcription process (64). Using purified RNA polymerase from M. smegmatis, strain mc(sup 2)155, Levin and Hatfull demonstrated that RIF specifically inhibited the elongation of full-length transcripts and had virtually no effect on the initiation of transcription (65). RNA polymerase, a complex oligomer composed of four different subunits (alpha,ß,ß'and sigma, encoded by rpoA, rpoB, rpoC, and rpoD, respectively), is highly conserved among bacterial species (66). Characterization of the rpoB gene in E. coli demonstrated that RIF specifically interacted with the ß subunit of RNA polymerase, thereby hindering transcription, and that mutations in the rpoB locus conferred conformational changes leading to defective binding of the drug and consequently resistance (67). Subsequently, the rpoB locus from MTB was characterized and mutations conferring the resistant trait were identified (Figure 3; 68-71). Most mutations were determined to be restricted to an 81-bp core region and are dominated by single nucleotide changes, resulting in single amino acid substitutions, although inframe deletions and insertions also occur at lower frequencies. Changes in the codons Ser531 and His526 have been documented in more than 70% of the RIF-resistant isolates. A very small number of mutations in RIF-resistant isolates do not map in this 81-bp core region; it is speculated that additional mechanisms, including RIF permeability and mutations in alternate subunits of RNA polymerase, may also be involved in conferring the resistance phenotype. [fig 3] Figure 3. Single amino acid substitutions in the 81 bp core-region of the rpoB gene responsible for conferring rifampicin (RIF) resistance (Insertions and deletions that confer the RIF-resistance phenotype are not depicted). Amino acids are represented with single letter abbreviations. Changes in codon Ser531 and His526 account for more than 70% of the mutations with RIF resistance (depicted in shaded ellipses). The consistency of mutations in the rpoB locus and the RIF- resistant phenotype (>95%) has marked clinical implications. Because it may act as a surrogate marker for MDRTB, RIF resistance has prompted development of various diagnostic tests to improve the sensitivity of mutation detection. Although automated sequencing has been unambiguously applied to characterize mutations associated with RIF resistance, a number of other techniques such as PCR-SSCP (41,45-48,121), dideoxy fingerprinting (72), heminested PCR (73), PCR heteroduplex analysis (70), and line probe hybridization (74,75) have been successfully applied to detecting these mutations. Such novel strategies to detect drug-resistant MTB isolates have been described elsewhere (76). PCR-SSCP analysis for detection of mutations responsible for conferring drug-resistance is increasingly useful. In particular, the development of nonisotopic PCR-SSCP analysis has simplified the procedure, enhancing its utility in routine laboratories (41,45). However, results obtained with SSCP analysis should be interpreted with caution as the technique only detects mutations and gives no information on the nature of associated mutation. For example, silent mutations in the rpoB gene have been identified that give altered mobility patterns on SSCP analysis but have no association with RIF resistance, which underlines the need for caution in interpreting results and phenotypic or genotypic correlation (77). Resistance to EMB EMB [dextro-2,2'-(ethyldiimino)-di-1onol], synthetic compound with profound antimycobacterial effects (78), is a first-line anti-MTB drug with a broad spectrum of activity, unlike INH. EMB is also advocated in disseminated M. avium complex infections, particularly in HIV-infected persons (79). Until recently, EMB's mechanism of action and the genetic basis for resistance to it were largely obscure. Specificity of EMB for mycobacterial species, however, indicated that its target may have been involved in the construction of the outer cell wall. Synergy resulting from coadministration of EMB and other drugs gave further evidence for the involvement of EMB in obstructing the formation of cell wall. The synergistic effect was explained as a consequence of increased permeability of the mycobacterial cell wall leading to increased drug uptake (80,81). Indeed, earlier studies of Takayama and colleagues demonstrated that administration of EMB led to rapid cessation of mycolic acid transfer to the cell wall and equally rapid accumulation of trehalose mono- and di-mycolates (82,83). Mycolic acids attach to the 5'-hydroxyl groups of D-arabinose residues of arabinogalactan and form mycolyl-arabinogalactan-peptidoglycan complex in the cell wall. Disruption of the arabinogalactan synthesis inhibits the formation of this complex and may lead to increased permeability of the cell wall. Subsequently, it was demonstrated that EMB specifically inhibited arabinogalactan synthesis (84). A breakthrough was achieved in defining the precise cellular target for EMB with the isolation and identification of ß-D-arabinofuronosyl-1-monophosphoryl decaprenol (DPA), which accumulates rapidly (less than 2 minutes) on exposure of EMB- sensitive cells to EMB (86). DPA is an arabinosyl donor; cell-free assay systems developed for DPA established that it was one of the major intermediates of arabinan synthesis. It was later shown that EMB specifically inhibited arabinosyl transfer, suggesting that arabinosyl transferase was the primary cellular target for EMB (Figure 4). [fig 4] Figure 4. Mechanism of action of ethambutol (adapted from 84-88). EMB interacts with the EmbCAB proteins encoded by the embC, embA, and embB genes, leading to inactivation of arabinogalactan synthesis. Mutations in the embB locus cause alterations in EmbB, possibly leading to an altered target for EMB. Alternatively, hyperexpression of the EmbCAB proteins could lead to EMB resistance. Inlet box: Organization of the emb operon in Mycobacterium tuberculosis(MTB). Identification of arabinosyl transferase as the primary target for EMB helped unravel the genetic basis for EMB resistance. Using target overexpression by a plasmid vector, Belanger et al. cloned the emb locus from an EMB-resistant strain of M. avium (86). Transformation of this emb locus conferred resistance to M. smegmatis mc(sup 2)155 strain and also demonstrated that the level of resistance conferred depended on the copy number of the gene, which was consistent with the notion of drug resistance due to target overexpression. Site-directed mutagenesis and overlapping clone analysis localized a 9.8-kb EMB resistance locus, subsequently shown to be ubiquitous among mycobacteria. Sequence analysis of this locus revealed three complete ORFs—designated embR, embA, and embB. The embR ORF is separated by a 178 bp divergent promoter region from the embA and embB ORFs. Characterization of the embR ORF showed that the region was strongly homologous with a family of transcriptional activators of Streptomyces and thus could play a role in modulating the expression of embA and embB. Importantly, the embB ORF lacks a potential ribosome binding site and is thus translationally coupled to embA, which suggests that a heterodimeric enzyme complex may be the target for EMB. Mapping studies further demonstrated that both embA and embB, along with the divergent promoter region, were essential to EMB resistance. In contrast to the organization of the emb locus in M. avium, molecular genetic approaches applied to MTB revealed a highly conserved 14-kb region comprising three homologous ORFs designated embC, embA, and embB preceded by a predicted coding region and by orfX (which encodes a putative protein belonging to the short chain alcohol dehydrogenase family) (87). Primer extension analysis of the emb region supported the notion of its organization as an operon and further indicated the polycistronic nature of its transcripts. The emb genes are translationally coupled the absence of any untranslated intercistronic region between the emb genes so indicated). However, the presence of a secondary stem loop structure between the embA and the embB genes indicates that the embB gene in MTB could be differentially regulated. The embCAB proteins are believed to be integral membrane proteins, consistent with their role in the synthesis of various arabinan-linkage motifs of the arabinogalactan and lipoarabinomannan (86,87). Identification of the embCAB genes prompted a detailed analysis of the molecular mechanisms responsible for conferring resistance to EMB in MTB isolates. Preliminary studies documented among EMB-resistant isolates missense substitutions in the conserved embB codon 306 that coded for methionine; their role in conferring resistance to EMB was confirmed by gene tranfer assays (87). Recent analysis of the embCAB region has confirmed the predominance of embB Met306 substitutions among EMB-resistant clinical isolates of MTB (approximately 89% among EMB-resistant isolates with single amino acid substitutions) (88). Sequence analysis of 118 clinical isolates of MTB showed five mutants of the embB codon 306, all leading to substitution of Met with Val, Leu, or Ile. MTB strains with Met306Leu and Met306Val substitutions demonstrated a higher MIC for EMB (40 µg/ml) than those for organisms with Met306Ile substitutions (20 µg/ml). The embB codon 306 may contain important structure-function information; structural alterations in this codon may have a detrimental effect on the interaction of EMB and EmbB, thereby resulting in a EMB-resistant phenotype. Sequence alterations in the embCAB region correlate with approximately 70% of EMB-resistant strains. Overexpression of the EmbB protein has been documented to mediate resistance in M. smegmatis (87), and a homologous mechanism may operate in MTB, perhaps accounting for the remaining 30% of the EMB-resistant isolates. A full understanding of the mechanisms for acquisition of EMB resistance among these isolates requires further studies. Resistance to PZA PZA, a structural analog of nicotinamide, was shown to have considerable anti-MTB activity in 1952, but it became an important component of short-course chemotherapy only in the mid- 1980s. PZA, active against semidormant bacilli not affected by any other drug, has strong synergy with INH and RIF and shortens the chemotherapeutic schedule for antitubercular treatment from 9 to 12 months to 6 months (15). Depending on the assay system and conditions applied, MICs of PZA vary from 8 µg/ml to 60 µg/ml. However, even at very high MICs, PZA has no significant bactericidal effect and is primarily considered a "sterilizing drug" (18). Activity of PZA is highly specific for MTB; PZA has scant or no effect on other mycobacteria, including M. bovis, which demonstrate high-level intrinsic resistance to PZA (89). Naturally resistant strains of M. bovis lack the enzyme Pzase, which hydrolyzes PZA to pyrizinoic acid, the presumed active form of PZA (90,91). PZA in this context is similar to INH; it is transported as a neutral species into the cell, where it is converted into its active form. This notion was strengthened by evidence provided by in vitro studies that demonstrated the susceptibility of PZA-resistant MTB and M. bovis to pyrizinoic acid. MTB Pzase has both pyrazinamidase and nicotinamidase activities (90). Using sequence information of E. coli nicotinamidase, Scorpio and Zhang isolated the mycobacterial pncA gene, which codes for the amidase (92). Characterization of the pncA gene from M. bovis isolates identified a single point mutation that results in the substitution of His to Asp at position 57. This substitution results in the production of an ineffective Pzase in M. bovis strains. Point mutations in the pncA gene of PZA-resistant MTB strains were also identified. Substitution of Cys138 with Ser, Gln141 with Pro, and Asp63 with His and deletion G nucleotide at positions 162 and 288 resulted in a defective Pzase. Transformation of Pzase-resistant strains with functional construct of MTB pncA gene restored susceptibility to PZA, providing further evidence that mutations in the pncA gene were responsible in conferring the resistant phenotype. Subsequent characterization of the pncA gene from clinical isolates of MTB confirmed these findings (93,94). Mutations including missense alterations, nucleotide insertions or deletions, and termination mutations have been found in the pncA gene from PZA-resistant MTB isolates. These sequence alterations are interspersed along the entire length of the pncA gene, demonstrate limited degree of clustering, and vary in frequency from 70% to 100% (93,94). The absence of correlating mutations in the pncA gene from PZA- resistant MTB isolates indicates that perhaps at least one additional mechanism mediates resistance to PZA. The cellular target for PZA, however, has not been identified, although the apparent similarity of PZA to nicotinamide suggests that enzymes involved in pyridine nucleotide biosynthesis are probable targets. Implication of the pncA gene in conferring PZA-resistant phenotype has profound clinical applications. Application of PCR-SSCP for detection of mutations in the pncA gene could help circumvent the difficulties in determining PZA susceptibilities (96) and rapidly discriminate between MTB and M. bovis (96). Resistance to Fluoroquinolones (FQ) FQs as antimycobacterial agents were first described in 1984 and have primarily been used as therapeutic alternatives in MDRTB cases (97). DNA gyrase (Gyr), a member of the type II DNA topoisomerases (98), is the primary target for FQ action. Gyr introduces negative supercoils in closed circular DNA molecules and is a heterotetramer (A2B2), coded by gyrA and gyrB respectively (99,100). Quinolone sensitivity is determined by the GyrA protein, which contains the cleavage/religation activity (100), while GyrB contains the intrinsic coumarin-sensitive ATPase activity (101). FQs, synthetic derivatives of nalidixic acid, act by inhibiting DNA supercoiling and relaxation activity of Gyr without affecting the ATPase activity (102) and enhance the rate of DNA cleavage by Gyr. Quinolone-mediated cleavage of double-stranded DNA results in a 4 bp 5' overhangs on either strand, to which GyrA subunits become attached covalently by O4 phosphotyrosine bond (103). Gyr catalyzes the cutting of DNA, denaturation of the overhang, and strand separation. The exact mechanism of inhibition of Gyr activity with respect to quinolones remains unknown. However, quinolone drugs bind with a greater affinity to single-stranded DNA than double-stranded DNA and possibly do not bind to Gyr at all (104). Consequently, by binding to the single-stranded DNA, the quinolones may inhibit religation, thereby imposing an effective transcriptional block (105), culminating in cellular death. However, questions about the specific interaction of quinolones and the Gyr/DNA complex remain unsolved (106). Cloning and expression of the MTB gyrA and gyrB genes allowed mapping of mutations that confer resistance to FQs (107). Mutations were found to be clustered in a small region in GyrA that is close, approximately 40 residues amino-terminal, in the linear amino acid sequence to the active site tyrosine, Tyr122 (E. coli numbering) (108). Other single amino substitutions, for residues 88 to 94, were also identified in ciprofloxacin-resistant MTB isolates (Figure 5). Because polymorphism encountered at codon 95 (Ser95>Thr95) occurred in both resistant and susceptible isolates, it may not be involved in acquiring the FQ-resistant phenotype. Alternative mechanisms to gyrA mutations, including changes in cell wall permeability and active quinolone efflux pumping, have also been proposed and could account for the low-level resistance among MTB isolates. fig 5] Figure 5. Single-amino acid substitutions responsible for conferring resistance to fluoroquinolones (FQ). Mutation in the Ser95 codon (shown in stippled box), observed in both FQ-sensitive and FQ-resistant isolates, rules out its role in acquisition of resistance. Newer FQ derivatives such as sparfloxacin have shown greater anti-MTB potency (MIC = 0.2 µg/ml) than ciprofloxacin and ofloxacin, giving hope for better therapeutic alternatives for MDRTB. However, FQ susceptibility in the treated patient population must be continuously monitored to prevent low-level FQ-resistant strains from acquiring additional mutations that lead to high-level resistance (109). Resistance to Streptomycin and Other Inhibitors Of Protein Synthesis Various drugs exert their antibacterial effects by inhibiting the protein transitional machinery. Among these, aminoglycosides, macrolides, tetracyclines, and basic peptides like viomycin and capreomycin are active against mycobacteria (110). SM, one of the oldest drugs known to be active against MTB, disrupts the decoding of aminoacyl-tRNA and thus inhibits mRNA translation or causes inefficient translation (111). One of the most common mechanisms for acquisition of resistance to SM is acetylation of the drug by aminoglycoside-modifying enzymes (111,112). However, this mechanism is not found in MTB. Instead, resistance to SM is attributed, at least partially, to two distinct classes of mutations including point mutations in S12 ribosomal protein, encoded by rpsL gene (113), and mutations in the rrs operon encoding the 16S rRNA (114). Point mutations in the rpsL gene result in single amino acid substitutions (114-117) that affect higher order structures of 16S rRNA and thereby confer SM resistance. Mapping of the mutations in the rpsL gene demonstrated that they primarily affected one of the two critical lysine residues at positions 43 and 88 and led to the substitution with either arginine at 88 or arginine and threonine at position 43 (115). An SM-resistant isolate (>60 µg/ml) showed an A>G transversion at position 904 in the 16S rRNA with an additional single A>C transversion in the rpsL gene, which resulted in the substitution of Lys-Gln at position 88 (115). Because each of the corresponding mutations in the small subunit rRNA or the ribosomal protein S12 confer the resistant phenotype in E. coli, these mutations mediated ribosomal drug resistance and were responsible for conferring high-level SM resistance. Mutations in the rpsL gene accounted for more than two thirds of SM-resistant cases. The genesis of SM resistance in some of the SM-resistant isolates is due to point mutations in the 16S rRNA. Mutations in the rrs locus have been mapped to two regions, the 530 loop and the 915 region. Within the 530 loop, C>T transitions at 491, 512, and 516, in addition to the A>C transversion at position 513, are consistent with the SM-resistant phenotype (114) pseudoknot formation within the MTB 16S rRNA. Base pairing between residue 524-526 (of the 530 region of the hairpin loop) and residue 504-507 (of the adjacent 510 region bulge loop) (118) results in SM resistance in clinical isolates of MTB (114). Further, G-U wobble base pairing between residues 522-501 stabilizes the pseudoknot formation and thereby confers resistance to SM. It can thus be concluded that SM resistance in MTB stems from alterations of the drug target and not by drug modification. However, no mutations in the rpsL and the rrs genes are detected in a significant number of SM-resistant isolates (46,48). Curiously, intrinsically SM-resistant strains of M. gordonae, M. szulgae, and M. avium do not show any alterations in the rpsL or the rrs genes, suggesting a probable third factor in conferring SM resistance. Earlier studies have documented the inhibitory effect of SM on protein synthesis in vitro to the same extent as observed in wild-type MTB strains. The same inhibitory effect was not observed on whole cells, suggesting the probable role of cell wall permeability barrier in conferring SM resistance (119). More recently, it has been demonstrated that membrane-active substances augmented the MIC for SM in strains with alterations in the rrs genes, thus providing further evidence for a probable role of the MTB-permeability barrier in mediating resistance to SM (120). Resistance to Other Drugs Related aminoglycosides such as kanamycin, amikacin, and paromomycin demonstrate no obvious cross-resistance to SM and thus are alternatives in cases of SM resistance. Viomycin and capreomycins are bacteriostatic agents that act by binding to the 50S ribosomal subunit and inhibit the translocation reaction (111). Although cross-resistance between viomycin and capreomycin does occur, the exact mechanism for acquisition of drug resistance is not known. Conclusions Molecular insights suggest that accumulation of mutations in the individual drug target genes is the primary mechanism of MDRTB. Morris and colleagues' investigation of the molecular mechanisms of drug resistance in MDR strains found that 25 of 44 SM-resistant strains had mutations in the rpsL gene, while five others had rrs gene perturbations (48). The rpoB gene had mutations in 28 of 29 RIF-resistant strains. Mutation in the katG gene was seen in 20 of the 42 INH-resistant stains, while five had inhA gene mutations. Of the 20 MDRTB strains, 11 had mutations in genetic markers associated with resistance to each of these three drugs. Similarly, Heym et al. reported that resistance to antitubercular agents in their collection of strains resulted from alterations to chromosomal genes encoding the drug targets; they excluded the possibility that MDRTB stemmed from acquisition of genes for novel resistance determinants (46). MDR appeared to result from the stepwise acquisition of new mutations in the genes for different drug targets. In all cases exactly the same mutations or combination of mutations were observed, regardless of the patient's HIV status. Thus, the origin of MDRTB is due more to treatment difficulties, including noncompliance and administration of inadequate treatment regime, and not to the emergence of novel resistance mechanisms; this is reassuring for the future of short-course chemotherapy. Administration of directly observed combination chemotherapy (or Directly Observed Treatment Short-Course [DOTS]) appears to be the most effective way to ensure a decrease in primary resistance, acquired resistance, and relapses (3). DOTS has been successfully implemented in diverse geographic areas including Tanzania, Guinea, China, Bangladesh, New York City, and Peru, which reported more than a 90% cure rate (3). Nearly 70 countries have adapted DOTS as a part of their national TB control programs and achieve good cure rates. Successful implementation of DOTS in the coming decades requires not only a concerted effort from various funding agencies but also a strong social and political commitment. Apart from strategic interventions based on strong political will, grass-roots action will have to be strengthened mainly at the primary health-care level to check the unlimited upsurge of this preventable fatal disease. Basic research will have to be continually updated to prevent the drug-resistant strains from becoming an unmanageable clinical paradigm. DOTS currently is our only option to reverse the global TB epidemic and prevent MDRTB. The inability to detect resistance early, however, is one of the major factors involved in the genesis and control of MDRTB; this invariably results in prolonged exposure to drugs that are virtually ineffective. One of the major consequences of unraveling the genetic basis of drug resistance in MTB is the development of various molecular strategies to rapidly detect MDRTB (76). However, the sheer multiplicity of gene loci to be investigated for diagnosis of MDRTB renders most of the approaches mentioned above as tedious and resource-intensive for a routine laboratory service program, particularly in developing countries like India, with limited resources and high disease incidence. Resistance to most anti-MTB drugs, with the exception of RIF, cannot be attributed to a single locus in substantial percentage (>90%), which is perhaps the greatest deterrent in the development of single amplification–based methods for rapid detection of resistance. Working out the exact biochemical details of drug-drug target interaction acquires considerable attention in the era of MDRTB, because only then will more rational structure- and mechanism-based approaches to inhibitor design be possible. Clearly, a concerted global effort is required to defeat TB resurgence. Acknowledgments We are indebted to Dr. Ellen Jo Baron and Prof. Eric C. Bottger for their suggestions and encouragement; Dr. James M. Musser for providing recent data on ethambutol resistance; and Swathi Arur, Nadeem Hasan, and Koninika Ray for help in preparing this manuscript. Work in our laboratory is supported by financial grants from the Department of Biotechnology, Government of India. Dr. Ashok Rattan is an additional professor, Department of Microbiology, All India Institute of Medical Sciences, New Delhi, India. His research focuses on drug resistance and application of cost-effective methods for surveillance of MDRTB strains. Address for Correspondence: Ashok Rattan or Awdhesh Kalia, TB and STD Section, Department of Microbiology, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110 029, India; fax: 91-11-686-2663; e-mail: rattan@medinst.ernet.in or awdhesh@aiims.ernet.in. References 1. Bloom BR, Murray CJL. Tuberculosis: Commentary on a reemergent killer. Science 1992;257:1055-64. 2. World Health Organization. Bridging the gaps: the world health report. Geneva: The Organization; 1995. 3. World Health Organization report on TB epidemic. Global TB programme. Geneva: The Organization; 1997. 4. Barnes P, Blotch AB, Davidson BT, Snyder Jr DE. Tuberculosis in patients with immuno-deficiency virus infection. N Engl J Med 1991;324:1644-50. 5. Kochi A, Vareldzis B, Styblo K. Multi-Drug resistant tuberculosis and control. Res Microbiol 1993;144:104-10. 6. Bell RT. Tuberculosis of the 1990s: the quiet public health threat. Pa Med 1992;95:24-5. 7. Snider DE Jr, Roper WL. The new tuberculosis. N Engl J Med 1992;326:703-5. 8. Freiden TE, Sterling T, Pablos-Mendez A, Kilburn JO, Cauthen JO, Dooley SW. The emergence of drug-resistant tuberculosis in New York city. N Engl J Med 1993;328:521-6. 9. Nosocomial transmission of multi-drug resistant tuberculosis among human immuno-deficiency virus infected patients—Florida and New York, 1988-1991. MMWR Morb Mortal Wkly Rep 1991;40:585-91. 10. Dooley SW, Jarvis WR, Martone WJ, Snyder DE Jr. Multi-Drug resistant tuberculosis [editorial]. Ann Intern Med 1992;117:257-8. 11. Edlin BR, Tokers JI, Greeko MH, Crawford JT, Williams J, Sordillo EM, et al. An outbreak of multi-drug resistant tuberculosis among hospitalized patients with the Acquired Immuno-Deficiency syndrome. N Engl J Med 1992;326:1514-21. 12. Pearson ML, Jareb JA, Freiden TR, Crawford JT, Davis BJ, Dooley SN, et al. Nosocomial transmission of multi-drug resistant tuberculosis—a risk to patients and health care workers. Ann Intern Med 1992;117:191-6. 13. Iseman MD, Sbarbaro JA. The increasing prevalence of resistance to antituberculosis chemotherapeutic agents: implications for global tuberculosis control. Curr Clin Top Infect Dis 1992;12:188-204. 14. Cohn DL, Flavia B, Raviglione MC. Drug-resistant tuberculosis: review of the worldwide situation and the WHO/IUATLD global surveillance project. Clin Infect Dis 1997;24:S121-30. 15. Initial therapy for tuberculosis in the era of multi-drug resistance: recommendations of the advisory council for the elimination of tuberculosis. MMWR Morb Mortal Wkly Rep 1993;42 (RR-7). 16. Mitchison DA. Mechanism of drug action in short-course chemotherapy. Bulletin International Union Against Tuberculosis 1985;65:30-7. 17. Iseman MD, Madsen LA. Drug-resistant tuberculosis. Clin Chest Med 1989;10:341-53. 18. Heifets LB, Lindohlm-Levy PJ. Pyrazinamide sterilizing activity in vitro against semidormant Mycobacterium tuberculosis populations. American Review of Respiratory Diseases 1992;145:1223-5. 19. Vareldzis BP, Grosset J, de Kantor I, Crofton J, Laszlo A, Felten M, et al. Drug-resistant tuberculosis: laboratory issues. World Health Organization recommendations. Tubercle and Lung Diseases 1994;75:1-7. 20. Spratt BG. Resistance to antibiotics mediated by target alterations. Science 1994;264:388-93. 21. Davis J. Inactivation of antibiotics and the dissemination of resistance genes. Science 1994;264:375-82. 22. Shimao T. Drug-resistance in tuberculosis control. Tubercle 1987;68(suppl):5-15. 23. Crofton J. The assessment and treatment of drug-resistance problems in tuberculosis. Journal of the Irish Medical Association 1970;63:75-8. 24. Grange JM. Drug-resistance and tuberculosis elimination. Bulletin International Union Against Tuberculosis and Lung Diseases 1990;65:57. 25. Youatt J. A review of the action of isoniazid. American Review of Respiratory Diseases 1969;99:729-49. 26. Shoeb HA, Bowman BU Jr, Ottolenghi AC, Merola AJ. Peroxidase-mediated oxidation of isoniazid. Antimicrob Agents Chemother 1985;27:399-403. 27. Maggliozzo RS, Marcinkeviciene JA. Evidence for isoniazid oxidation by oxypressors mycobaterial catalase-oxidase. Journal American Chemical Society 1996;118:11303-4. 28. Middlebrook G. Isoniazid-resistance and catalase activity of tubercle bacilli. American Review of Tuberculosis 1954;69:471-2. 29. Zhang Y, Heym B, Allen B, Young D, Cole S. The catalase-peroxidase gene and isoniazid resistance of Mycobacterium tuberculosis. Nature 1992;358:591-3. 30. Zhang Y, Garbe T, Young D. Transformation with katG restores isoniazid-sensitivity in Mycobacterium tuberculosis isolates resistant to a range of drug concentrations. Mol Microbiol 1993;8:521-4. 31. Demple B, Halbrook J. Inducible repair of oxidative DNA damage in Escherichia coli. Nature 1983;304:466. 32. Sherman DR, Sabo PJ, Hickey MJ, Arain TM, Mahairas GG, Yuan Y, et al. Disparate responses to oxidative stress in saprophytic and pathogenic mycobacteria. Proc Natl Acad Sci U S A 1995;92:6625-9. 33. Sherman DR, Mdluli K, Hickey MJ, Arain TM, Morris SL, Barry CE III, Stover CK. Compensatory ahpC gene expression in isoniazid-resistant Mycobacterium tuberculosis. Science 1996;272:1641-3. 34. Chae HZ, Robinson K, Leslie B, Church GB, Storz G, Rhee SG. Cloning and sequencing of thiol-specific antioxidant from mammalian brain, alkyl hydro-peroxide reductase and thiolspecific anti-oxidant define a large family of antioxidant enzymes. Proc Natl Acad Sci U S A 1994;91:7017-21. 35. Chae HZ, Chung SJ, Rhee SG. Thioredoxin-dependent-peroxide reductase from yeast. J Biol Chem 1994;269:276-0. 36. Sreevatsan S, Pan X, Zhang Y, Deretic V, Muser JM. Analysis of the oxyR-ahpC region in isoniazid-resistant and -susceptible Mycobacterium tuberculosis complex organims recovered from diseased humans and animals in diverse localities. Antimicrob Agents Chemother 1997;41:600-6. 37. Heym B, Stavropoulos E, Honore N, Domenech P, Saint-Joanis B, Wilson TM, et al. Effects of overexpression of the alkyl hydroperoxide reductase AhpC on the virulence and isoniazid resistance of Mycobacterium tuberculosis. Infect Immun 1997;65:1395-401. 38. Altamarino M, Marostenmaki J, Wong A, Fitzgerald M, Black WA, Smith JA. Mutations in the catalase-peroxidase gene from isoniazid-resistant Mycobacterium tuberculosis isolates. J Infect Dis 1994;160:1162-5. 39. Stoeckle MY, Guan L, Riegler N, Weitzman I, Kreiswirth B, Kornblum J, et al. Catalase-peroxidase gene sequences in isoniazid-sensitive and -resistant strains of Mycobacterium tuberculosis from New York City. J Infect Dis 1993;168:1063-5. 40. Cockerill FR III, Uhi JR, Temesgen Z, Zhang Y, Stockman L, Roberts GD, et al. Rapid identification of a point mutation of the Mycobacterium tuberculosis catalase-peroxidase (katG) gene associated with isoniazid resistance. J Infect Dis 1995;171:240-5. 41. Pretorius GS, Van Helden PD, Sergel F, Eisenach KD, Victor TC, et al. Mutations in katG gene sequences in isoniazid-resistant clinical isolates of Mycobacterium tuberculosis are rare. Antimicrob Agents Chemother 1995;39:2276-81. 42. Heym B, Alzavi PM, Honore N, Cole ST. Missense mutations in the catalase-peroxidase gene, katG, are associated with isoniazid resistance in Mycobacterium tuberculosis. Mol Microbiol 1995;15:235-45. 43. Rouse DA, Devito JA, Li Z, Byer M, Morris SL. Site-directed mutagenesis of the katG gene of Mycobacterium tuberculosis: effects on catalase-peroxidase activities and isoniazid resistance. Mol Microbiol 1996;22:583-92. 44. Jaber M, Rattan A, Kumar R. Presence of katG gene in isoniazid-resistant strains of Mycobacterium tuberculosis. J Clin Pathol 1996;49:945-7. 45. Kalia A, Ahmad N, Rattan A. Diagnosis of multi-drug resistant tuberculosis: comparison of traditional, radiometric and molecular methods [abstract]. In: Abstracts of the 20th International Congress of Chemotherapy; 29 Jun-3 Jul 1997; Sydney, Australia. Sydney: International Society of Chemotherapy; 1997. p. 211. 46. Heym B, Honore N, Truffot-Pernot C, Banerjee A, Schurra C, Jacobs WR Jr, et al. Implications of multidrug resistance for the future of short-course chemotherapy of tuberculosis: a molecular study. Lancet 1994;344:293-8. 47. Kapur V, Li LL, Hamrick MR, Plikaytis BB, Shinnick TM, Telenti A, et al. Rapid Mycobacterium species assignment and unambiguous identification of mutations associated with antibiotic resistance in Mycobacterium tuberculosis by automated DNA sequencing. Arch Pathol Lab Med 1995;119:131-8. 48. Morris SL, Bai Gh, Suffys P, Portillo-Gomez L, Fairchok M, Rouse D. Molecular mechanisms of multidrug resistance in clinical isolates of Mycobacterium tuberculosis. J Infect Dis 1995;171:954-60. 49. Johnsson K, Froland WA, Schultz PG. Overexpression, purification and characterization of the catalase-peroxidase, katG from Mycobacterium tuberculosis. J Biol Chem 1997;272:2834-40. 50. Rouse DA, Li Z, Baig M, Morris SL. Characterization of the katG and inhA genes of isoniazid resistant clinical isolates of Mycobacterium tuberculosis. Antimicrob Agents Chemother 1995;30:2472-7. 51. Rosner JL. Susceptibility of oxyR regulon mutants of Escherichia coli and Salmonella typhimurium to isoniazid. Antimicrob Agents Chemother 1993;37:2251-3. 52. Deretic V, Philipp W, Dhandyuthapani S, Mudd MH, Curcic R, Garbe T, et al. Mycobacterium tuberculosis is a natural mutant with an inactivated oxidative stress regulatory gene: implications for sensitivity to isoniazid. Mol Microbiol 1995;17:889-900. 53. Wilson TM, Collins DM. ahpC, a gene involved in isoniazid resistance of Mycobacterium tuberculosis complex. Mol Microbiol 1996;19:1025-34. 54. Banerjee A, Dubnau E, Quemard A, Balasubramanian V, Um KS, Wilson T, et al. inhA, a gene encoding a target for isoniazid and ethionamide in Mycobacterium tuberculosis. Science 1994;263:227-30. 55. Bergler H, Wallner P, Ebeling A, Leitinger B, Fuchshlschler S, Aschauer H, et al. Protein EnvM is the NADH-dependent-enoyl-ACP-reductase (Fab1) of Escherichia coli. J Biol Chem 1994;269:5493-6. 56. Cole ST. Mycobacterium tuberculosis: drug-resistance mechanisms. Trends Microbiol 1994;2:411-5. 57. Dessen A, Quemard A, Blanchard JS, Jacobs WR Jr, Sacchettini JC. Crystal structure and function of the isoniazid target of Mycobacterium tuberculosis. Science 1995;267:1638-41. 58. Quemard A, Sacchettini JC, Dessen A, Jacobs WR Jr, Blanchard JS, et al. Enzymatic characterization of the target for isoniazid in Mycobacterium tuberculosis. Biochemistry 1995;34:8235-41. 59. Johnsson K, King DS, Schultz PG. Studies on the mechanism of action of isoniazid and ethionamide in chemotherapy of tuberculosis. Journal of American Chemical Society 1995;117:5009-10. 60. Takayama K, Schoenes HK, Armstrong EL, Boyle KW. Site of inhibitory action of isoniazid in the synthesis of mycolic acids in Mycobacterium tuberculosis. J Lipid Res 1975;16:308-17. 61. Davidson LA, Takayma K. Isoniazid inhibition of the synthesis of mono-saturated long chain fatty acids in Mycobacterium tuberculosis H37Ra. Antimicrob Agents Chemother 1979;16:104-5. 62. Mdluli K, Sherman DR, Hickey MJ, Kreiswirth BN, Morris S, Stover CK, Barry LE III, et al. Biochemical and genetic data suggest that inhA is not the primary target for activated isoniazid in Mycobacterium tuberculosis. J Infect Dis 1996;174:1085-90. 63. Woodley CL, Kilburn JO, David HL, Silcox VA. Susceptibility of mycobacteria to rifampin. Antimicrob Agents Chemother 1972;2:245-9. 64. Ovchinnikov YA, Monastyrskaya GS, Gubanov VV, Lipkin VM, Sverdlov ED, Kiver IF, et al. Primary structure of Escherichia coli RNA polymerase nucleotide substitution in the ß-subunit gene of rifampicin resistant rpoB255 mutant. Mol Gen Genet 1981;84:536-8. 65. Levin ME, Hatfull GF. Mycobacterium smegmatis RNA polymerase: DNA supercoiling, action of rifampicin and mechanism of rifampicin resistance. Mol Microbiol 1993;8:277-85. 66. Ovchinnikov YA, Monastryskaya, GS, Gubanov VV, Lipkin VM, Sverdlov ED, Kiver IF, et al. The primary structure of Escherichia coli RNA polymerase. Nucleotide sequence of rpoB gene and amino-acid sequence of the ß-subunit. Eur J Biochem 1981;116:621-9. 67. Jin D, Gross C. Mapping and sequencing of mutations in the Escherichia coli rpoB gene that leads to rifampicin resistance. J Mol Biol 1988;202:45-58. 68. Telenti A, Imboden P, Marchesi F. Lowrie D, Cole S, Colston MJ, et al. Detection of rifampicin-resistance mutations in Mycobacterium tuberculosis. Lancet 1993;341:647-50. 69. Telenti A, Imboden P, Marchesi F, Schidheini T, Bodmer T. Direct, automated detection of rifampicin-resistant Mycobacterium tuberculosis by polymerase chain reaction and single-strand conformation polymorphism analysis. Antimicrob Agents Chemother 1993;37:2054-8. 70. Williams DL, Waguespack C, Eisenach K, Crawford JT, Portaels M, Salfinger M, et al. Characterization of rifampicin-resistance in pathogenic mycobacteria. Antimicrob Agents Chemother 1994;38:2380-6. 71. Kapur V, Li LL, Iordanescu S, Hamrick MR, Wanger A, Kreisworth RN, et al. Characterization by automated DNA sequencing of mutations in the gene (rpoB) encoding the RNA polymerase ß-subunit in rifampicin-resistant Mycobacterium tuberculosis strains from New York City and Texas. J Clin Microbiol 1994;32:1095-8. 72. Felmlee TA, Liu Q, Whelen AC, Williams D, Sommer SS, Persing DH. Genotypic detection of Mycobacterium tuberculosis rifampin resistance: comparison of single strand conformation polymorphism and dideoxy fingerprinting. J Clin Microbiol 1995;33:1617-23. 73. Whelen AC, Felmlee TA, Hunt JM, Williams DL, Roberts GD, Stockman L, Persing DH. Direct genotype detection of Mycobacterium tuberculosis rifampicin resistance in clinical specimens by using single-tube heminested PCR. J Clin Microbiol 1995;33:556-61. 74. De Benhouwer, Lhiang HZ, Jannes G, Mijis W, Machtelinckx L, Rossau H, et al. Rapid detection of rifampin resistance in sputum and biopsy samples from tuberculosis patients by PCR and line probe assay. Tubercle and Lung Diseases 1995;76:425-30. 75. Cooksey RC, Morlock GP, Glickman S, Crawford JT. Evaluation of a line probe assay kit for characterization of rpoB mutations in rifampin resistant Mycobacterium tuberculosis isolates from New York City. J Clin Microbiol 1997;35:1281-3. 76. Telenti A, Persing DH. Novel strategies for the detection of drug resistance in Mycobacterium tuberculosis. Res Microbiol 1996;147:73-9. 77. Kim BJ, Kim SY, Park B-H, Liu M-A, Park I-K, Bai GH, et al. Mutations in the rpoB gene in Mycobacterium tuberculosis that ineterfere with PCR-single strand conformation polymorphism analysis for rifampin susceptibility testing. J Clin Microbiol 1997;35:492-4. 78. Thomas JP, Baughan CO, Wilkinson RG, Shephard RG. A new synthetic compound with anti-tuberculous activity in mice: ethambutol (dextro-2,2'-[ethylenediimino]-di-1-butonol). American Review of Respiratory Diseases 1961;83:891-3. 79. Masur H. Recommendations on prophylaxis and therapy for disseminated Mycobacterium avium complex disease in patients infected with HIV virus. N Engl J Med 1993;329:828-33. 80. Rastoggi N, Goh KS. Action of 1-isonicotinyl-2-palmitoyl hydrazine against the Mycobacterium avium complex and enhancement of its activity by m-flurophenyl alanine. Antimicrob Agents Chemother 1990;34:2061-4. 81. Rastoggi N, Goh KS, Labrausse V. Activity of clathiromycin compared with those of other drugs against Mycobacterium paratuberculosis and further enhancement of its extracellular and intracellular activities by etham-butol. Antimicrob Agents Chemother 1992;36:2843-6. 82. Takayama K, Armstrong EL, Kunugi KA, Kilburn JO. Inhibition by ethambutol of mycolic acid transfer into the cell wall of Mycobacterium smegmatis. Antimicrob Agents Chemother 1979;16:240-2. 83. Kilburn JO, Takayama K. Effects of ethambutol on accumulation and secretion of trehalose mycolates and free mycolic acid in Mycobacterium smegmatis. Antimicrob Agents Chemother 1981;20:401-4. 84. Takayama K, Kilburn JO. Inhibition of synthesis of arabinogalactan by ethambutol in Mycobacterium smegmatis. Antimicrob Agents Chemother 1989;33:1493-9. 85. Wolucka BA, McNeil MR, de Hoffman E, Chojnaki T, Brennan PJ. Recognition of the lipid intermediate for arabinogalactan/arabinomanan biosynthesis and its relation to the mode of action ethambutol on mycobacteria. J Biol Chem 1994;269:23328-35. 86. Belanger AE, Besra GS, Ford ME, Mikusova K, Belisle JT, Brennan PJ, Inamine JM. The embAB genes of Mycobacterium avium encode an arabinosyl transferase involved in cell wall arabinan biosynthesis that is the target for the antimycobacterial drug ethambutol. Proc Natl Acad Sci U S A 1996;93:11919-24. 87. Telenti A, Philipp WJ, Sreevatsan S, Bernasconi C, Stockbauer KE, Weites B, et al. The emb operon, a gene cluster of Mycobacetrium tuberculosis involved in resistance to ethambutol. Nat Med 1997;3:567-70. 88. Sreevatsan S, Stockbauer KE, Pan X, Kreisworth BM, Moghazeh SL, Jacobs WR Jr, et al. Ethambutol resistance in Mycobacterium tuberculosis: critical role of embB mutations. Antimicrob Agents Chemother 1997;41:1677-81. 89. Konno K, Nagayama H, Oka S. Nicotinamidase in mycobacteria: a method for distinguishing bovine type tubercle bacilli from other mycobacteria. Nature 1959;184:1743-4. 90. Konno K, Feldman FM, McDermot W. Pyrazinamide susceptibility and amidase activity of tubercle bacilli. American Review of Respiratory Diseases 1967;95:461-7. 91. Mackaness GB. The intracellular activation of pyrazinamide and nicotinamide. American Review of Tuberculosis 1953;74:718-28. 92. Scorpio A, Zhang Y. Mutations in pncA, a gene encoding pyrazinamidase/nicotinamidase, cause resistance to the antituberculous drug pyrazinamide in tubercle bacillus. Nat Med 1996;2:662-7. 93. Sreevatsan S, Pan X, Zhang Y, Kreisworth BN, Musser JM. Mutations associated with pyrazinamide resistance in pncA of Mycobacterium tuberculosis complex organisms. Antimicrob Agents Chemother 1997;41:636-40. 94. Scorpio A, Lindholm-Levy P, Heifets L, Gilman R, Siddiqi S, Cynamon M, Zhang Y, et al. Characterization of pncA mutations in pyrazinamide resistant Mycobacterium tuberculosis. Antimicrob Agents Chemother 1997;41:540-2. 95. Hewlett D, Horn DL, Alfalfa C. Drug resistant tuberculosis: inconsistent results of pyrazinamide susceptibility testing. JAMA 1995;273:916-7. 96. Scorpio A, Collins D, Whipple D, Cave D, Bates J, Zhang Y. Rapid differentiation of bovine and human tubercule bacilli based on a characteristic mutation in the bovine pyrazinamidase gene. J Clin Microbiol 1997;35:106-10. 97. Gay JD, de Young DR, Roberts GD. In vitro activities of norfloxacin and ciprofloxacin against Mycobacterium tuberculosis, M. avium complex, M chelonei, M. forfuitre, and M. kansasii. Antimicrob Agents Chemother 1984;26:94-6. 98. Gellert M, Mizuuchi K, O'Dea MH, Nash HA. DNA gyrase: an enzyme that introduces superhelical turns into DNA. Proc Natl Acad Sci U S A 1976;73:3872-5. 99. Kirchausen T, Wang JC, Harrison SC. Purification of the subunits of Escherichia coli DNA gyrase and reconstitution of enzyme activity. Proc Natl Acad Sci U S A 1978;75:1773-7. 100. Gellert M, O'Dea MH, Itoh T, Tomizava J. Novobiocin and coumeromycin inhibit DNA supercoiling catalyzed by DNA gyrase. Proc Natl Acad Sci U S A 1976;73:4474-8. 101. Sugino A, Peebles CL, Kreuzer KN, Cozzarelli NR. Mechanism of action of Nalidixic acid: purification of Escherichia coli nalA gene product and its relationship to DNA gyrase and a novel nicking closing enzyme. Proc Natl Acad Sci U S A 1977;74:4667-71. 102. Kirkegaard K, Wand JC. Mapping the topography of DNA wrapped around gyrase by nucleolytic and chemical probing of complexes of unique DNA sequences. Cell 1981;23:721-9. 103. Shen LL, Pernet AG. Mechanism of inhibition of DNA gyrase by analogues of nalidixic acid: the target of the drugs is the DNA gyrase. Proc Natl Acad Sci U S A 1985;82:307-11. 104. Wilmont CJR, Critchlow SE, Eperon IC, Maxwell A. The complex of DNA gyrase and quinolone drugs with DNA forms a barrier to transcription by RNA polymerase. J Mol Biol 1994;242:351-63. 105. Lewis RJ, Tsai FTF, Wigley DB. Molecular mechanism of drug inhibition by DNA gyrase. Bioessays 1996;18:661-71. 106. Takiff HE, Salazar L, Guerrero C, Philipp W, Huang WM, Kreisworth B, et al. Cloning and nucleotide sequencing of Mycobacterium tuberculosis gyrA and gyrB genes and detection of quinolone resistant mutations. Antimicrob Agents Chemother 1994;38:773-80. 107. Rees RJ, Maxwell A. DNA gyrase. Structure and function. Crit Rev Biochem Mol Biol 1991;26:335-75. 108. Revel V, Cambau E, Jarlier E, Sougakoff W. Characterization of mutations in Mycobacterium smegmatis involved in resistance to fluoroquinolones. Antimicrob Agents Chemother 1994;38:1991-6. 109. Inderlied CB. Antimycobacterial agents: in vitro susceptibility testing, spectrums of activity, mechanisms of action and resistance, and assays for activity in biological fluids. In: Lorain V, editor. Antibiotics in laboratory medicine. Baltimore: Williams and Wilkins, Baltimore; 1991. p. 134-197. 110. Benveinste R, Davies J. Mechanism of antibiotic resistance in bacteria. Annu Rev Biochem 1973;42:471-506. 111. Davies J, Wright JD. Bacterial resistance to aminoglycoside antibiotics. Trends Microbiol 1997;5:234-9. 112. Douglass J, Steyn LM. A ribosomal gene mutation in streptomycin-resistant Mycobacterium tuberculosis isolates. J Infect Dis 1993;167:1505-6. 113. Finken M, Kirschner P, Meier, A, Wrede A, Bottger EC. Molecular basis of streptomycin-resistance in Mycobacterium tuberculosis: alteration of the ribosomal protein S12 gene and point mutations within a functional 16S rRNA pseudoknot. Mol Microbiol 1993;9:1239-46. 114. Meier A, Kirschner P, Bange FC, Vogel U, Botger EC. Genetic alteration in streptomycin-resistance in Mycobacterium tuberculosis: mapping of mutations conferring resistance. Antimicrob Agents Chemother 1994;38:228-33. 115. Nair J, Rouse DA, Bai GH, Morris SL. The rpsL gene and streptomycin resistance in single and multi-drug resistant strains of Mycobacterium tuberculosis. Mol Microbiol 1993;10:521-4. 116. Honore N, Cole ST. Streptomycin resistance in myco-bacteria. Anitmicrob Agents Chemother 1994;38:238-42. 117. Woes CR, Gutell RR. Evidence for several higher order structural elements in ribosomal rRNA. Proc Natl Acad Sci U S A 1989;86:3119-22. 118. Shaila MS, Gopinathan RP, Ramakrishnan T. Protein synthesis Mycobacterium tuberculosis H37Rv and the effect of streptomycin in streptomycin susceptible and resistant strains. Antimicrob Agents Chemother 1973;4:205-13. 119. Meier A, Sander P, Schaper KJ, Scholz M, Bottger EC. Correlation of molecular resistance mechanisms and phenotypic resistance levels in streptomycin-resistant Mycobacterium tuberculosis. Antimicrob Agents Chemother1996;40:2452-4. 120. Orita M, Suzuki Y, Sekiya T, Hayashi K. Rapid and sensitive detection of point mutations and DNA polymorphisms using the polymerase chain reaction. Genomics 1989;5:875-9. Emerging Infectious Diseases National Center for Infectious Diseases Centers for Disease Control and Prevention Atlanta, GA URL: http://www.cdc.gov/ncidod/EID/vol4no2/rattan.htm Please note that figures and equations are not available in ASCII format; their placement within the text is noted by [fig] and [eq], respectively. Greek symbols are spelled out. The following codes are used: (ft) for footnote; (sup) for superscript; (sub) for subscript; /= for greater than or equal to. ------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------- [Emerging Infectious Diseases] [Volume 4 No. 2 / April - June 1998] Synopses Hospitalizations for Unexplained Illnesses among U.S. Veterans of the Persian Gulf War James D. Knoke and Gregory C. Gray Naval Health Research Center, San Diego, California, USA -------------------------------------------------- Persian Gulf War veterans have reported a variety of symptoms, many of which have not led to conventional diagnoses. We ascertained all active-duty U.S. military personnel deployed to the Persian Gulf War (552,111) and all Gulf War era military personnel not deployed (1,479,751) and compared their postwar hospitalization records (until 1 April 1996) for one or more of 77 diagnoses under the International Classification of Diseases (ICD-9) system. The diagnoses were assembled by the Emerging Infections Program, Centers for Disease Control and Prevention, and are here termed "unexplained illnesses." Deployed veterans were found to have a slightly higher risk of hospitalization for unexplained illness than the nondeployed. Most of the excess hospitalizations for the deployed were due to the diagnosis "illness of unknown cause" (ICD-9 code 799.9), and most occurred in participants of the Comprehensive Clinical Evaluation Program who were admitted for evaluation only. When the effect of participation in this program was removed, the deployed had a slightly lower risk than the nondeployed. These findings suggest that active-duty Gulf War veterans did not have excess unexplained illnesses resulting in hospitalization in the 4.67-year period following deployment. The Persian Gulf War was one of the briefest full-scale conflicts in U.S. history. For a 2-month period of fighting ending in March 1991, nearly 700,000 U.S. service members were deployed to the Persian Gulf region. Since returning from the war, many veterans have reported unexplained symptoms (1-5), prompting allegations of a new disease or diseases (2,3,5,6). Numerous expert panels and research projects have examined illness and death among Gulf War veterans (4,7,8). However, with the exception of self-reported symptoms (8-14), no consistent pattern of increased illness or death has been reported (10,14-18). The U.S. Department of Defense (DoD) conducted an epidemiologic comparison of the postwar DoD hospitalizations of service members deployed to the Gulf War and service members of the same era not deployed (15). In the 2-year period after the war, no consistent increase in the overall risk for hospitalization (or specific risk for hospitalization for various broad diagnostic categories) was found for those deployed. This study relied upon diagnoses based on the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9) system (19). Medical providers may not consistently classify a new or poorly recognized syndrome, and consequently a true difference in hospitalization risk could be spread across numerous diagnostic categories and remain undetected. The present study compares the postwar DoD hospitalizations for diagnoses consistent with an unexplained illness of Persian Gulf War veterans and their nondeployed peers. Analytic Approach Study Population The study population consisted of active-duty service members (Army, Air Force, Navy, Marine Corps, and Coast Guard) who were either deployed to the Persian Gulf War for 1 or more days during the Gulf War deployment period (8 August 1990 through 31 July 1991) or were not deployed but were on active duty for at least part of the Gulf War deployment period. All deployed (n = 552,111) and nondeployed (n = 1,479,751) service members who remained on active duty at the end of the period were included in the study population. The study was restricted to active-duty personnel because they are rarely hospitalized outside DoD facilities. Members of the U.S. Reserve and National Guard forces were not studied because only a fraction of their postdeployment hospitalizations were in DoD facilities. Study Outcome The study outcome was determined by 77 ICD-9 diagnoses assembled by the Emerging Infections Program, Centers for Disease Control and Prevention, to monitor death certificates for unexplained deaths (20). These diagnoses span several of the 17 major categories delineated in ICD-9 and include selected diagnoses from diseases of the blood, nervous system, circulatory system, respiratory system, and digestive system; infectious and parasitic diseases; and symptoms, signs, and ill-defined conditions. These diagnoses primarily relate to nonspecific infections and other ill-defined conditions, and for convenience they are termed "unexplained illnesses" in this study. All study population admissions to U.S. military hospitals worldwide (reported to the DoD computerized hospitalizations database by 1 October 1996), subsequent to the Gulf War deployment period and before 1 April 1996, were evaluated for unexplained illnesses among as many as eight ICD-9 diagnoses coded for each admission. For another analysis, only the first coded diagnosis (the principal diagnosis) was ascertained. The DoD ICD-9 coding guidelines define the principal diagnosis as "the condition established, after study, to be responsible for the admission." Hospitalizations prior to the conclusion of the deployment period were not evaluated because access to care for the deployed differed markedly from that for the nondeployed during this time. Outpatient visits were not studied because they are not computerized centrally by DoD. Data Demographic variables available for use as covariates included age, race/ethnicity, occupation, rank, salary, branch of service, length of service, marital status, and gender. Time-dependent variables were evaluated as of 31 July 1990. A previous study (15) found hospitalization for any reason in the 12 months preceding the Gulf War deployment period to be an important predictor of postwar hospitalization. This indicator variable may be a surrogate for baseline health status and was also used as a covariate. Demographic data, including deployment status, were obtained from the Defense Manpower Data Center, Seaside, California. Hospitalization information was obtained from the Data Processing Center, Fort Detrick, Frederick, Maryland. Comprehensive Clinical Evaluation Program data were obtained from the Deployment Surveillance Team, Falls Church, Virginia. Statistical Analysis Frequencies of selected diagnoses and causes of death were calculated. The Cox proportional hazards survival analysis model (21) was used to obtain the risk ratio (RR) and 95% confidence interval (CI) of deployment status (deployed relative to nondeployed) for an event consisting of hospitalization with an unexplained illness, adjusting for the covariates. Follow-up time was computed from 1 August 1991 until hospitalization in any DoD hospital worldwide with at least one unexplained illness, separation from the service, or until 31 March 1996, whichever occurred first. All data management and statistical calculations were performed with the Statistical Analysis System (22). Findings Frequent Unexplained Illnesses Our study population consisted of 25,495 first hospitalizations (with at least one unexplained illness among the eight possible diagnoses), 6,672 in the deployed and 18,823 in the nondeployed. For these hospitalizations, the 10 most frequent first unexplained illnesses among all eight possible diagnoses were tabulated (Table 1). Eight of these diagnoses occurred with similar proportional distributions between the two groups. The diagnosis "nonspecific abnormal findings in the amniotic fluid" accounted for a higher proportion of hospitalizations among the nondeployed than among the deployed. This is consistent with a higher proportion of women among the nondeployed (12.7%) than among the deployed (6.1%). The tenth most frequent diagnosis, "illness of unknown cause," accounted for a considerably greater proportion of hospitalizations among the deployed (8.3%) than among the nondeployed (1.8%). Table 1. Frequencies of the 10 most common diagnoses for first hospitalizations with an unexplained illness, considering all eight coded diagnosesa ------------------------------------------------------------------------------ Frequency Percentage of all diagnoses ------------------------ -------------------- ICD-9 Non- Non- Code Description Total Deployed deployed Deployed deployed ----------------------------------------------------------------------------------- 008.8 Unspecified 2286 581 1705 8.7 9.1 intestinal infections 047.9 Unspecified viral 946 236 710 3.5 3.8 meningitis 079.9 Unspecified viral 3540 896 2644 13.4 14.1 infections 465.9 Acute URIb, 1999 527 1472 7.9 7.8 unspecified site 486 Pneumonia, organism 2840 719 2121 10.8 11.3 unspecified 780.6 Pyrexia of unknown 961 233 728 3.5 3.9 origin 785.6 Enlarged lymph 1587 423 1164 6.3 6.2 nodes 786.0 Dyspnea 2736 703 2033 10.5 10.8 792.3 Nonspecific abnormal amniotic 1317 188 1129 2.8 6.0 fluid 799.9 Illness of unknown 892 555 337 8.3 1.8 cause Subtotal 19,104 5061 14,043 75.9 74.6 Other 6391 1611 4780 24.1 25.4 Total 25,495 6672 18,823 100.0 100.0 ------------------------------------------------------------------------------------- (sup a)Only the first unexplained illness coded for each hospitalization was tabulated. (sup b)URI, upper respiratory infection. An unexplained illness was the principal diagnosis in 13,490 first hospitalizations, 3,525 from the deployed and 9,965 from the nondeployed. The proportional distributions of the 10 diagnoses reported in Table 1 showed some rearrangements when only the principal diagnosis was considered (Table 2). The percentages of "acute upper respiratory infection, unspecified site"; "nonspecific abnormal amniotic fluid"; and "illness of unknown cause" were dramatically lower as the principal diagnosis than they were as any diagnosis. However, "illness of unknown cause" remained the only diagnosis that accounted for an appreciably larger proportion of admissions among the deployed (2.0%) than among the nondeployed (0.3%). Screening of Covariates The demographic characteristics of the deployed and nondeployed groups have been reported (15). All available covariates were included in a preliminary Cox proportional hazards model to assess their effect on the risk for hospitalization for an unexplained illness (Table 3). All available covariates (one variable at a time) were also included in a series of preliminary models for this same purpose. These two approaches to screening covariates gave somewhat different results, probably because of multicolinearity among the covariates. The second approach was less useful than the first, so it is not further reported here. Given the other covariates, age, marital status, and length of service were only minimally related to risk (possibly because of colinearity) and were not included in subsequent model analyses. The covariates retained included race (coded as white; black; or other, including unknown); rank (coded as enlisted or warrant or commissioned officer, including unknown); salary (coded as less than $1,000; $1,000 to $1,399; or at least $1,400 a month, including unknown); and branch of service (coded as Army or other). The military system has a large number of occupational categories; however, only "health-care worker" appeared to have an appreciably higher risk for hospitalization for an unexplained illness than other occupational categories. Consequently, occupation was simply coded as health-care worker or other (including unknown). Prewar hospitalization status was coded as "yes" if one was hospitalized for any reason during the 12 months before 1 August 1990 and as "no" otherwise. All of these other covariates, as well as gender (with unknowns included with men), had a highly statistically significant effect on the risk for hospitalization for an unexplained illness and were included in all subsequent model analyses. Other possible covariates, whose effect on risk was either less highly significant or nonsignificant, were not included so as to minimize unnecessary computation and variance inflation. Survival Analysis, Using All Eight Diagnoses A model analysis including deployment status and the selected covariates showed that deployment status was significantly associated with hospitalization for an unexplained illness when all eight possible diagnoses were used (Table 4). Separate model analyses for the two deployment status groups indicated that the parameter estimates were similar, and the effect of the covariates on the probability of hospitalization was essentially independent of deployment status. Table 2. Frequencies of the 10 most common diagnoses for first hospitalization with an unexplained illness, considering only the principal diagnosis -------------------------------------------------------------------------------------- Percentage of all Frequency diagnoses ------------------------ -------------------- ICD-9 Non- Non- Code Description Total Deployed deployed Deployed deployed -------------------------------------------------------------------------------------- 008.8 Unspecified intestinal 1807 467 1340 13.2 13.5 infections 047.9 Unspecified viral 912 228 684 6.5 6.9 meningitis 079.9 Unspecified viral 2620 661 1959 18.7 19.8 infections 465.9 Acute URI, unspecified 487 151 336 4.3 3.4 site 486 Pneumonia, organism 2350 615 1735 17.5 17.4 unspecified 780.6 Pyrexia of unknown 382 93 289 2.6 2.9 origin 785.6 Enlarged lymph nodes 1000 274 726 7.8 7.3 786.0 Dyspnea 1176 302 874 8.6 8.8 792.3 Nonspecific abnormal 5 1 4 0.0 0.0 amniotic fluid 799.9 Illness of unknown cause 101 72 29 2.0 0.3 Subtotal 10,840 2864 7976 81.2 80.0 Other 2650 661 1989 18.8 20.0 Total 13,490 3525 9965 100.0 100.0 --------------------------------------------------------------------------------------- The probabilities under the separate models of hospitalization with an unexplained illness, at the mean values of the included covariates, are presented as a function of follow-up time in Figure 1. These probabilities for the two groups were virtually coincidental and linear over time, until late 1994 when the deployed group's probability increased. This shift in the historical track for the deployed group suggests that the hazards for the two groups were not proportional over time and that additional analyses may be indicated. The Comprehensive Clinical Evaluation Program In May 1994, DoD announced a Comprehensive Clinical Evaluation Program (CCEP), offering thorough clinical examinations and evaluations to Gulf War veterans who sought them. This program was implemented in June 1994. The divergence in probability curves (Figure 1) in the last quarter of 1994 prompted us to investigate whether the introduction of CCEP may have affected the probability of hospitalization. Referring to any hospitalization for unexplained illness after 1 June 1994 of a CCEP participant as a CCEP hospitalization, we found that 837 first unexplained illness hospitalizations in the deployed were CCEP hospitalizations, as were 55 in the nondeployed. (For the purpose of CCEP participation, deployment status was self-determined.) Many of the nondeployed (according to Defense Manpower Data Center data) CCEP participants may have been in the Persian Gulf region after 1 August 1991. Of these 892 hospitalizations, 59% of the first diagnoses for unexplained illnesses were "illness of unknown cause," the most nonspecific diagnosis in ICD-9. Furthermore, with 128 DoD hospitals operating worldwide in 1995, 60% of these hospitalizations were in six facilities-three Army hospitals and three Air Force hospitals-and fewer than 1% were in any Navy facility. Table 3. Frequencies and risk ratios under the Cox proportional hazards model for all hospitalizations with at least one unexplained illness, all covariates, 1 August 1991 to 1 April 1996 ------------------------------------------------------------------------------------ 95% Non- Risk Confidence Variable Deployed deployed ratio(sup a) interval ------------------------------------------------------------------------------------- Not deployed 1,479,751 Deployed 552,111 1.08 1.05-1.11 Not hospitalized before war(sup b) 513,486 1,375,169 Hospitalized before war 38,625 104,582 1.73 1.66-1.79 Not married 258,821 602,553 Married 291,009 875,309 1.05 1.02-1.08 Unknown marital status 2282 1889 0.64 0.41-1.00 Male 517,223 1,291,323 Female 33,690 188,273 2.13 2.06-2.20 Unknown gender 1198 155 0.60 0.20-1.75 White 383,704 1,110,949 Black 130,249 286,427 1.05 1.02-1.08 Hispanic 12,900 26,696 0.91 0.83-1.00 Other 23,592 54,128 0.91 0.85-0.97 Unknown race/ethnicity 1666 1551 1.08 0.69-1.68 Age in years: 17-21 142,547 343,535 22-25 146,652 306,272 0.94 0.90-0.97 26-31 138,173 373,698 0.95 0.90-1.00 >/=32 123,353 455,302 1.05 0.99-1.12 Unknown 1386 944 1.04 0.57-1.88 Infantry, gun crews, seamanship 140,742 273,079 Electronic equipment repair 45,263 159,195 0.99 0.93-1.04 Communications/intelligence 55,665 120,108 1.00 0.95-1.06 Health care 28,398 107,591 1.46 1.39-1.53 Other technical 11,633 33,246 1.02 0.93-1.11 Administration 62,194 255,576 0.96 0.93-1.02 Construction/related trades 19,424 47,800 1.11 1.06-1.17 Trainees, undesignated 6421 64,625 1.02 0.95-1.10 Unknown job category 18,147 80,911 1.18 1.11-1.25 Enlisted 489,034 1,238,790 Warrant Officer 7760 13,555 0.79 0.69-0.89 Commissioned Officer 53,121 227,381 0.70 0.65-0.75 Unknown status 2196 25 0.43 0.20-0.93 Salary/month: <$1000 122,161 328,187 $1000-$1399 269,500 573,995 0.81 0.77-0.85 $1400-$2099 75,523 252,923 0.68 0.63-0.72 $2100-$3199 60,528 200,295 0.73 0.68-0.79 >/=$3200 22,203 124,326 0.74 0.66-0.83 Unknown 2196 25 - Army 258,858 459,644 Navy 141,604 423,138 0.61 0.59-0.64 Marine Corps 83,377 113,467 0.66 0.63-0.70 Air Force 67,942 447,125 0.76 0.73-0.78 Coast Guard 330 36,377 0.42 0.37-0.48 Service in months: < 22 124,629 262,260 22-54 162,123 269,727 1.05 1.00-1.11 55-126 133,608 363,138 0.97 0.92-1.02 >/=127 131,751 584,626 1.12 1.07-1.18 ----------------------------------------------------------------------------------------- (sup a)Risk ratios are relative to the first group in each variable category. When the algorithm failed to converge, because of lack of cases in the category, a dash (-) is given. (sup b)In the 12 months preceding 1 August 1990. Table 4. Frequencies and risk ratios under the Cox proportional hazards model for all hospitalizations with at least one unexplained illness, selected demographic variables, 1 August 1991 to 1 April 1996 -------------------------------------------------------------------------------------------- 95% Non- Risk Confidence Variable Deployed deployed ratio(sup a) interval -------------------------------------------------------------------------------------------- Not deployed - 1,479,751 Deployed 552,111 - 1.06 1.03-1.09 Not hospitalized before war(sup b,c) 513,486 1,375,169 Hospitalized before war 38,625 104,582 1.71 1.65-1.78 Male(sup c) 518,421 1,291,478 Female 33,690 188,273 2.11 2.04-2.18 White 383,704 1,110,949 Black 130,249 286,427 1.04 1.01-1.08 Other race/ethnicity(sup c) 38,158 82,375 0.91 0.86-0.96 All other occupations(sup c) 523,713 1,372,160 Health-care worker 28,398 107,591 1.44 1.38-1.50 Enlisted 489,034 1,238,790 Officer(sup c) 63,077 240,961 0.66 0.62-0.70 Salary/month: <$1,000 122,161 328,187 $1,000-$1,399 269,500 573,995 0.77 0.74-0.79 >/=$1,400(sup c) 160,450 577,569 0.82 0.78-0.87 Army 258,858 459,644 Other branches 293,253 1,202,107 0.60 0.58-0.62 ---------------------------------------------------------------------------------------------- (sup a)Risk ratios are relative to the first category for each variable. (sup b)In the 12 months preceding 1 August 1990. (sup c)Includes unknown. These unusual results prompted us to make inquiries at some of the DoD hospitals reporting large numbers of hospitalizations for "illness of unknown cause." We learned that several of the larger Army and Air Force hospitals established special wards for CCEP participants reaching phase 2 of the evaluation process, admitted them for several days, and performed extensive evaluations (including invasive procedures and sleep studies) during hospitalization. These facilities also gave at least some of these participants a diagnosis of "illness of unknown cause." This coding practice for evaluation admissions may have resulted from a memo dated 25 August 1994 from the Headquarters, U.S. Army Medical Command at Fort Sam Houston, San Antonio, Texas, which directed Army facilities to code CCEP participants with "unexplained complaints with no confirmed diagnosis" as 799.9. The practice of admitting CCEP participants was discontinued by mid-1995. The probability of hospitalization curves tended to become parallel again about mid-1995 (Figure 1). Thus, we inferred that a substantial majority of the CCEP hospitalizations were primarily for evaluation. Independent evaluation supported this inference. Since DoD computerized hospital records do not include cause of hospitalization or any other indication of whether a patient may have been hospitalized primarily for evaluation, we randomly selected for chart review 50 CCEP hospitalizations from each of the Army and Air Force facilities with the greatest numbers of CCEP hospitalizations. Both of these facilities were located in the same metropolitan area, San Antonio, Texas. The selected CCEP hospitalizations were independently evaluated by two clinicians, one from each facility. Seventy-nine charts were reviewed, 44 from the Air Force facility and 35 from the Army facility. The other 21 records were located in satellite facilities or otherwise not readily accessible. The two clinicians agreed that 77 of these hospitalizations were for evaluation only and would not have been considered hospitalizations had the CCEP not been in effect. One of the clinicians thought that two of the hospitalizations were for clinical management and would have occurred regardless of CCEP. [fig 1] Figure 1. Probability of hospitalization for unexplained illness, deployed and nondeployed veterans. Survival Analysis, Using All Eight Diagnoses, Censoring CCEP Participants We repeated the survival analyses (reported in Table 4; Figure 1), censoring all CCEP participants on 1 June 1994. This change resulted in 5,835 first hospitalizations for an unexplained illness in the deployed group and 18,768 in the nondeployed. No important differences were observed in the effects of the covariates between the models censoring CCEP participants on 1 June 1994 and those reported in Table 4. However, when CCEP participants were censored, the risk for hospitalization for an unexplained illness was lower in the deployed than in the nondeployed (RR = 0.93, CI = 0.91 to 0.96), and the probability of hospitalization for an unexplained illness was generally lower over time for the deployed than for the nondeployed (Figure 2). Also, the proportional hazards assumption now appeared reasonable. CCEP participants may truly have been at increased risk for hospitalization for an unexplained illness, in spite of many CCEP hospitalizations having been for evaluation only. Consequently, the true risk for the deployed is likely to be inter- mediate to that depicted in Figures 1 and 2. However, our record review, which concluded that the preponderance of CCEP hospitalizations were for evaluation only, suggested that the results are more closely depicted in Figure 2 than Figure 1. [fig 2] Figure 2. Probability of hospitalization for unexplained illness, deployed and nondeployed veterans, censoring comprehensive clinical evaluation program participants on 1 June 1994. Survival Analyses, Using Principal Diagnosis Only Model analysis of hospitalizations for which the principal diagnosis was an unexplained illness showed that deployment status was not a statistically significant predictor (RR = 0.99, CI = 0.95 to 1.03). When CCEP participants were censored on 1 June 1994, model analysis results were similar and showed that the deployed were slightly less likely than the nondeployed to be hospitalized (RR = 0.93, CI = 0.89 to 0.97). Model analysis results for hospitalization for an unexplained illness using only the principal diagnosis thus were similar to the results when all eight diagnoses were employed, except that there were many fewer admissions with "illness of unknown cause" as a principal diagnosis than as a secondary diagnosis. Deaths During hospitalization with an unexplained illness 348 veterans died. Of those who died, 86 had been deployed, and 262 had not. Of the 348 deaths, only 62 had an unexplained illness indicated as the underlying cause. Of these 62, 17 had been deployed and 45 had not. Deployed veterans with an unexplained illness as the underlying cause of death included four with "unspecified septicemia"; six with "pneumonia, organism unspecified"; three with "other diseases of the lung"; and one each with "other primary cardiomyopathies," "thrombotic microangiopathy," "unexplained death," and "respiratory failure." Conclusions Because they live in close quarters and are required to travel, military personnel commonly acquire diseases endemic to the regions they visit and may serve as reservoirs and vectors for disease transmission when they return home. A number of military deployment-related epidemics have occurred in recent years. Persian Gulf War veterans have been diagnosed with a new manifestation of leishmaniasis (23). Veterans of the peace-keeping effort in Somalia have had increased hospitalization rates for malaria (24), deployed Navy personnel have brought nonendemic strains of HIV virus into the United States (25), and Russian soldiers have been implicated in the mass epidemics of diphtheria in Eastern Europe (26). These and other observations make surveillance for unexplained illnesses among U.S. military personnel an important federal public health issue. We designed this study to screen for unexplained illnesses using a collection of ICD-9 diagnoses (20). Since unexplained illnesses generally have no specific ICD-9 diagnoses, a nonspecific diagnosis from this collection is likely to be used by hospital coders when an unexplained illness is encountered. The DoD worldwide hospitalization data subsequent to the Persian Gulf War were analyzed using the Cox proportional hazards model for evidence of unexplained illnesses. Analyses were performed both with all eight coded diagnoses and with only the principal diagnosis. Model analysis showed that Gulf War veterans were more likely than their peers to be hospitalized for unexplained illness. After 4.67 years of follow-up, the cumulative probability of hospitalization for unexplained illness was 0.0185 in the deployed and 0.0166 in the nondeployed (Figure 1). This increased hospitalization risk of 11% for the deployed was a consequence of the recruiting for free clinical evaluations beginning in June 1994, with most of the resulting CCEP hospitalizations being for medical evaluation and not for clinical management. When CCEP participants were censored on 1 June 1994, deployed Gulf War veterans were not at greater risk than those not deployed. The slightly lower hospitalization risk for the deployed than for the nondeployed (Figure 2) is consistent with a healthy service member effect; that is, those selected for deployment are, on average, slightly healthier than those not selected. This study had a number of limitations. Some miscoding of Gulf War deployment status was suggested by the finding that some nondeployed veterans were evaluated under the CCEP as being Gulf War veterans. However, veterans who did not serve in the Gulf region until after 31 July 1991 were eligible for the CCEP, and this apparent discrepancy was less than 6%. Also, many personnel separated from the service during the period of follow-up; 59.8% of the deployed and 54.3% of the nondeployed had separated by 1 April 1996. However, the fact that separating veterans receive thorough medical screening and have the potential of receiving lifelong disability benefits motivates them to be thorough in their reporting of illness before separation. The broad categorization of deployment status may have masked illness due to time- and geography-specific exposures, and illnesses not serious enough to require hospitalization have not been captured by our analyses. Finally, the collection of diagnoses used as the outcome measure was not designed specifically for our purposes, but it does have the advantage of prior development by other researchers (20). This study also had its strengths. The large group sizes, along with the availability of numerous important covariates (demographic variables and prewar hospitalization), offer unusually high statistical power for the detection of differences in the hospitalization risk between groups. Additionally, discharge diagnoses are thoroughly recorded and edited by DoD hospitals, and active-duty personnel have few opportunities to be hospitalized outside the DoD system, which ensures high-quality data with few missing values. In summary, these analyses show a slightly greater hospitalization risk for unexplained illness among deployed Gulf War veterans than among those not deployed. However, the excess hospitalizations for the deployed are attributable to participation in CCEP; most of these hospitalizations were for medical evaluation, not clinical management. Consequently, before initiation of CCEP, active-duty Gulf War veterans were not at increased risk for hospitalization for unexplained illness. Acknowledgments We thank Drs. Edwin C. Matthews and Gregg T. Anders for completing the record review of the 79 randomly selected CCEP hospitalizations. We also thank Michael A. Dove and Wayne F. Woo for their assistance in obtaining the deployment status and covariate data. This is Naval Health Research Center report no. 96-35, supported by the Department of Defense (Health Affairs) and the Naval Medical Research and Development Command, Department of the Navy, Bethesda, Maryland, under work unit no. 63738DP4464.001-6423. James D. Knoke is statistician, Emerging Illness Division, Health Sciences and Epidemiology Department, Naval Health Research Center, San Diego. His work concentrates on methodology and applications in clinical trials and epidemiology. He has special interests in the areas of survival analysis, longitudinal analysis, multivariate analysis, and nonparametric methods. He serves as an investigator in a number of epidemiologic studies among Persian Gulf War veterans. Captain Gregory C. Gray is head, Emerging Illness Division, Health Sciences and Epidemiology Department, Naval Health Research Center, San Diego. He is a medical epidemiologist, specializing in general preventive medicine and public health. He has experience conducting epidemiologic investigations among military populations, particularly those involving respiratory pathogens. He serves as principal investigator for a number of large epidemiologic studies among Persian Gulf War veterans. Address for correspondence: James D. Knoke, Naval Health Research Center, P.O. Box 85122, San Diego, CA 92186, USA; fax: 619-553-7601; e-mail: knoke@nhrc.navy.mil. References 1. Cotton P. Gulf War symptoms remain puzzling. JAMA 1992;268:2619. 2. Why are we sick? Army Times 1994 Apr 25:12-23. 3. Cowley G, Hager M, Liu M. Tracking the second storm. Newsweek 1994 May 16:56-7. Persian Gulf Veterans Coordinating 4. Board. Unexplained illnesses among Desert Storm veterans: a search for causes, treatment, and cooperation. Arch Intern Med 1995;155:262-8. 5. Brown D. Diagnosis unknown: Gulf War syndrome. The Washington Post 1994 Jul 24;Sect. A:1 (col. 2-4). 6. Nicolson GL, Rosenberg-Nicolson NL. Doxycycline treatment and Desert Storm. JAMA 1995;273:618-9. 7. Research Working Group of the Persian Gulf Veterans Coordinating Board. Federally Sponsored Research on Persian Gulf Veterans' Illnesses for 1995. Washington: Department of Veterans Affairs; 1996. Annual Report to Congress. 8. Institute of Medicine. Health consequences of service during the Persian Gulf War: recommendations for research and information systems. Washington: National Academy Press; 1996. 9. Unexplained illness among Persian Gulf veterans in an Air National Guard unit: Preliminary report-August 1990-March 1995. MMWR Morb Mortal Wkly Rep 1995;44:443-7. 10. DeFraites RF, Wanat ER, Norwood AE, Williams S, Cowan D, Callahan D. Investigation of a suspected outbreak of an unknown disease among veterans of Operation Desert Shield/Storm, 123rd Army Reserve Command, Fort Benjamin Harrison, Indiana, April 1992. Washington: Epidemiology Consultant Service, Division of Preventive Medicine, Walter Reed Army Institute of Research; 1992. 11. Presidential Advisory Committee on Gulf War Veterans' Illnesses. Final Report, Dec 1996. Washington: U.S. Government Printing Office; 1996. 12. Defense Science Board. Report of the Defense Science Board Task Force on Persian Gulf War Health Effects. Washington: Office of the Under Secretary of Defense for Acquisition and Technology; 1994. 13. National Institutes of Health Technology Assessment Workshop Panel. The Persian Gulf experience and health. JAMA 1994;272:391-6. 14. Kaiser KS, Hawksworth AW, Gray GC. A comparison of self-reported symptoms among active-duty Seabees: Gulf War veterans versus era controls. In: Proceedings of the 123rd Annual Meeting of the American Public Health Association; 1995 Oct 31; San Diego, California. Washington: American Public Health Association; 1995. 15. Gray GC, Coate BD, Anderson CM, Kang HK, Berg SW, Wignall FS, et al. The postwar hospitalization experience of U.S. Persian Gulf War veterans compared with other veterans of the same era. N Engl J Med 1996;335:1505-13. 16. Kang HK, Bullman TA. Mortality among U.S. veterans of the Persian Gulf War. N Engl J Med 1996;335:1498-504. 17. Writer JV, DeFraites RF, Brundage JF. Comparative mortality among U.S. military personnel in the Persian Gulf region and worldwide during Operations Desert Shield and Desert Storm. JAMA 1996;275:118-21. 18. Cowan DN, DeFraites RF, Wishik SM, Goldenbaum MB, Wishik SM. The risk of birth defects among children of Gulf War veterans. N Engl J Med 1997;336:1650-6. 19. The International Classification of Diseases, 9th Revision, Clinical Modification. 3rd ed. Washington: U.S. Department of Health and Human Services; 1991. 20. Perkins BA, Flood JM, Danila R, Holman RC, Reingold AL, Klug LA, et al. Unexplained deaths due to possibly infectious causes in the United States: defining the problem and designing surveillance and laboratory approaches. Emerg Infect Dis 1996;2:47-53. 21. Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Time Data. New York: John Wiley & Sons; 1980. 22. SAS Institute Inc. SAS Language Reference. Version 6. 1st ed. Cary (NC): SAS Institute Inc.; 1990. 23. Magill AJ, Grogl M, Gasser RA Jr, Sun W, Oster CN. Visceral infection caused by Leishmania tropica in veterans of Operation Desert Storm. N Engl J Med 1993;328:1383-7. 24. Wallace MR, Sharp TW, Romajzl PJ, Batchelor RA, Thornton SA, Longer CF, et al. Malaria in Mogadishu, Somalia. Clin Infect Dis 1993;17:510-1. 25. Brodine SK, Mascola JR, Weiss PJ, Ito SI, Porter KR, Artenstein AW, et al. Detection of diverse HIV-1 genetic subtypes in the USA. Lancet 1995;346:1198-9. 26. Maurice J. Russian chaos breeds diphtheria outbreak. Science 1995;267:1416-7. Emerging Infectious Diseases National Center for Infectious Diseases Centers for Disease Control and Prevention Atlanta, GA URL: http://www.cdc.gov/ncidod/EID/vol4no2/knoke.htm Please note that figures and equations are not available in ASCII format; their placement within the text is noted by [fig] and [eq], respectively. Greek symbols are spelled out. The following codes are used: (ft) for footnote; (sup) for superscript; (sub) for subscript; /= for greater than or equal to. ------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------- [Emerging Infectious Diseases] [Volume 4 No. 2 / April - June 1998] Synopses Agricultural Use of Burkholderia (Pseudomonas) cepacia: A Threat to Human Health? Alison Holmes,* John Govan,† and Richard Goldstein‡ *Imperial College School of Medicine, Hammersmith Hospital, London, United Kingdom; †University of Edinburgh Medical School, Edinburgh, United Kingdom; and ‡Maxwell Finland Laboratory for Infectious Diseases, BostonUniversity School of Medicine, Boston, Massachusetts, USA -------------------------------------------------- In the past 2 decades, Burkholderia cepacia has emerged as a human pathogen causing numerous outbreaks, particularly among cystic fibrosis (CF) patients. One highly transmissible strain has spread across North America and Britain, and another between hospitalized CF and non-CF patients. Meanwhile, the organism has been developed as a biopesticide for protecting crops against fungal diseases and has potential as a bioremediation agent for breaking down recalcitrant herbicides and pesticides. However, B. cepacia is inherently resistant to multiple antibiotics; selection of strains "safe" for environmental application is not at present possible phenotypically or genotypically; molecular epidemiology and phylogenetic studies demonstrate that highly transmissible strains emerge randomly; and the organism has a capacity for rapid mutation and adaptation (facilitated by numerous insertion sequences), and a large, complex genome divided into separate chromosomes. Therefore, the widespread agricultural use of B. cepacia should be approached with caution. Burkholderia (previously known as Pseudomonas) cepacia, a nutritionally versatile, gram-negative organism, was first described in 1949 by Walter Burkholder of Cornell University, as the phytopathogen responsible for a bacterial rot of onions (1) (Figure 1). Ironically, B. cepacia is now being considered by agricultural microbiologists as an agent to promote crop growth. [fig] Figure 1. B. cepacia causes an onion rot known as slippery skin (1). The onions shown were inoculated with three strains of B. cepacia. Rot occurred in onion1 left), which was inoculated with strain originally isolated from onions. Rot did not occur with environmental isolates tested or with strains from CF lung. B. cepacia is inherently resistant to multiple antibiotics, can metabolize diverse substrates, and is found in soil and in moist environments. The organism has a particular predilection for the lung in patients with cystic fibrosis (CF) and has emerged as an important opportunistic human pathogen in hospitalized and immunocompromised patients (2,3). B. cepacia as a Human Pathogen In CF Patients CF is a recessively transmitted genetic disease that occurs in approximately 1 in 2,500 Caucasians (carrier frequency of 1 in 25). The condition is characterized by a generalized dysfunction of the exocrine glands, giving rise to a broad spectrum of clinical syndromes, especially malabsorption due to pancreatic insufficiency and chronic progressive lung disease giving rise to bronchiectasis. B. cepacia is associated with increased illness and death among CF patients. In the early 1980s, the organism emerged as a major threat, causing superinfection in as many as 40% of patients in some CF centers (4-6). While in some patients indolent pulmonary infection occurs with only gradual deterioration in lung function similar to that associated with Pseudomonas aeruginosa, approximately 35% of B. cepacia–infected patients contract accelerated pulmonary deterioration or fulminant, necrotizing pneumonia with rapidly fatal bacteremia (3,7-10), sometimes referred to as "cepacia syndrome" (7). Unlike infection with P. aeruginosa, B. cepacia infection significantly increases death rates among CF patients (11) at all levels of lung function. Over the last 20 years, CF survival rates have increased as a result of improved treatment, the median length of survival increasing from early childhood to more than 29 years. As a result, approximately one third of CF patients are adults. Pulmonary infection causes the most illness and ultimately more than 90% of CF-related deaths. B. cepacia's emergence as a pathogen coincided with social and medical grouping of CF patients in specialized units, clinics, and social groups. Studies subsequently demonstrated that social contact between CF patients was an important factor in transmission of B. cepacia (12). The threat of B. cepacia infection led to severe control measures, affecting not only the treatment but also the social and family lives of CF patients. CF centers adopted stringent infection control policies, assuming that all B. cepacia isolates were highly transmissible (13,14). CF summer camps in North America were closed, and many lung transplant centers ceased to accept B. cepacia–infected CF patients as transplant candidates. Newly formed national and international associations for CF adults and CF families (providing conferences, holidays, and support groups) addressed the issue of B. cepacia transmission by abandoning many activities and adopting exclusion and segregation measures (13). Numerous CF-associated B. cepacia epidemics have now been described, and the epidemic strains have been characterized (15-18). One particular highly transmissible strain, epidemically spread within and between CF centers on both sides of the Atlantic, carries the cblA gene (18). This gene encodes for the major structural subunit of unique mucin binding cable pili (4). These enormously long pili (radiating 2 to 4 microns) are peritrichously arranged and are intertwined to form cablelike structures that adhere to CF mucin (4,18) (Figure 2). This cblA+ strain has spread across Canada and has now been isolated in 50% of CF centers in the United Kingdom (19). Another strain of B. cepacia has spread among CF centers in four regions of France (20). [fig] Figure 2. Transmission electron micrograph of Toronto/Edinburgh epidemic clone of B. cepacia expressing CF mucous-binding Cbi adhesin pili. High resolution was achieved by using a JEOL 100CX electron microscope and negative staining. Reprinted with permission from Richard Goldstein and Journal of Bacteriology (J Bacteriol 1995;177:1039-52). However, it has become clear that transmissibility varies markedly from strain to strain, and that most strains are not involved in epidemics but appear to be independently acquired and show no evidence of transmission. For example, in 8 years no cases of transmission were detected at the University of North Carolina CF center, despite clinical and social contact between patients and the absence of stringent infection control measures (21). Independent acquisition of B. cepacia with no evidence of transmission between CF patients was also reported from Denmark (22). Lack of transmission of some strains has also been observed between siblings with CF (23). In Patients Without CF B. cepacia was first reported as a human pathogen causing endocarditis in the 1950s. Since then the organism has caused numerous catheter-associated urinary tract infections, wound infections, and intravenous catheterassociated bacteremias. In 1971, it was reported as the causative organism of "foot rot" in U.S. troops on swamp training exercises in northern Florida; it was also isolated from troops serving in the Mekong Delta, Vietnam (24). In the 1980s the number of nosocomial infections with B. cepacia increased markedly, with deaths particularly associated with lung infections (25). Numerous small focal hospital outbreaks involving non-CF patients have usually been due to a contaminated common source, such as disinfectant preparations or intravenous solutions (2,26,27). The organism's unusually broad metabolic capabilities, which enable it to survive and proliferate in water-based environments, probably account for its propensity to cause nosocomial outbreaks (2,27-29). Because of its resistance to multiple antibiotics, once acquired, the organism can be difficult to treat. B. cepacia is rarely found in the non-CF patient (30); however, when it is found the organism can spread to and from CF patients. Transmission between CF and non-CF patients has been associated with serious illness and death (31;Holmes et al., unpub. data) and presents a greater nosocomial threat than previously recognized. In addition, although B. cepacia is not a common commensal organism, hospitalized patients without CF may harbor it and pose an infection threat to vulnerable CF patients. This possible source of infection may account for the known association between hospitalization and of B. cepacia infection (32). B. cepacia may be underreported because selective media for B. cepacia, which greatly increase the yield (33,34), are rarely used except in specimens from CF patients. B. cepacia as an Agricultural Agent While emerging as a human pathogen, B. cepacia has attracted intense interest from the agricultural industry as a possible biologic control agent. The organism, which has been shown to have remarkable potential as an agent for both biodegradation and biocontrol, is being considered as a plant-growth-promoting rhizobacterium (23,35-43). B. cepacia has extraordinary metabolic versatility and can degrade chlorinated aromatic substrates (toxic compounds found in complex pesticides and herbicides, some with carcinogenic potential) for use as carbon energy sources. One important toxic compound degraded by B. cepacia is 2,4,5 chlorophenoxy acetic acid (2,4,5-T), a potent herbicide that is not easily biodegradable and persists for long periods in the environment (37). B. cepacia can also antagonize and repress many soilborne plant pathogens. It can prevent leaf and stem blight caused by the fungus Alternaria by inhibiting spore germination. Economically important crop diseases such as blight due to A. solani and the blight caused by A. brassicae and A. brassicola, which affects the oil-producing plants rape and canola, can be controlled by B. cepacia. The organism is also being used to prevent the blight of ginseng plants due to A. panax (41) and is effective against the fungus Aphanomyces euteiches, which causes root rot in peas, alfalfa, and snap beans (39,40). It can prevent Pythium diseases of cucumber and peas, and Rhizoctonia solani stem rot of poinsettia (42). To prevent these plant diseases, B. cepacia provides a seemingly environmentally friendly alternative to potent and toxic fungicides, which cannot be broken down in the environment. The forestry industry also sustains large economic losses from the pathogenic effects of fungi such as Fusarium, Pythium, Rhizoctonia, Cylindrocarpum, and Botrytis. These widespread fungal pathogens cause seedling loss in nurseries and may kill or stunt transplanted seedlings. A strain of B. cepacia has been developed as a successful seed and root inoculant, which can suppress these fungi on a variety of conifers (43). Numerous patents are being sought for specific agricultural applications of different strains of B. cepacia. The ecologic and economic benefits could be enormous if the organism's antifungal activity is used to enhance crop yields and reduce the need for pesticides and its ability to degrade complex herbicides and pesticides is harnessed for bioremediation. Molecular Epidemiology of B. cepacia Evolution of Highly Transmissible Strains B. cepacia isolates are closely related (a panmictic population structure) (18,31,44) and epidemic isolates are scattered throughout (18,31), as demonstrated by rrn-based phylogenic trees, which include large numbers of environmental and clinical isolates. Such a phylogeny indicates that in B. cepacia strains, genetic changes conferring high transmissibility are occurring at random, or, given the right epidemiologic circumstances, random environmental strains are innately highly transmissible. In addition, the widespread geographic distribution of different epidemic strains of B. cepacia would also suggest that highly transmissible strains are emerging independently and randomly. Taxonomy Isolates presumptively identified as B. cepacia belong to at least five genomovars (23,45,46). This group of phenotypically similar subpopulations is referred to as the B. cepacia complex (6,46). After multiphasic taxonomic analysis, the species names B. multivorans and B. vietnamiensis have been proposed for genomovars II and V, respectively (45). The pathogenic and epidemic potential of each of these subpopulations of the B. cepacia complex requires further examination. Although it appears that strains associated with cepacia syndrome belong to genomovar III (46), isolates belonging to each of these five genomovars have been cultured from CF patients. The Unusual Genome B. cepacia's capacity to propagate as an environmental microbe and as an opportunistic pathogen may be due to its possession of an unusually large (more than four times that of H. influenzae, two times that of E. coli, and half again as large as P. aeruginosa), complex, and variable genome. The genome contains numerous insertion sequences and is divided into one to four circular replicons (47,48). A few other bacterial species of agricultural and medical importance also have multiple chromosomes: Brucella melitensis, B. abortus, B. suis, B. ovis, Rhodobacter sphaeroides, and Agrobacterium tumefaciens (49-51). This unusual genomic arrangement may account for B. cepacia's nutritional versatility and adaptability. Such a division of genomic content would allow for high levels of homologous and illegitimate recombination. The resultant chromosomal rearrangements and associated mutations could provide a basis for spontaneous "pulsed" evolutionary spurts, such as that seen from soil to the CF lung, suited for rapid adaptation to radical changes in environmental growth conditions. Selection of Strains for Agricultural Use Because evolution of highly transmissible strains occurs randomly and transmission of B. cepacia between CF and non-CF populations can cause severe illness and death, the deliberate widespread environmental application of strains of this organism should be considered carefully. Although B. cepacia does not appear to survive on dry surfaces for more than 1 week, it can survive for many months in water. The agricultural application of B. cepacia will lead to environmental and water contamination and increased human exposure. In addition to medical, metabolic, and taxonomic issues, increasing knowledge of B. cepacia raises important ecological issues, including the evolution of pathogenicity and multiresistant environmental bacteria through horizontal gene transfer. The agricultural use of B. cepacia risks this hazard of horizontal gene transfer between the strains applied and existing soil organisms and the subsequent emergence of pathogenic, highly resistant organisms. Recently, in an attempt to assess the human risk associated with the use of rhizophere isolates as field inoculants, two clinical isolates of different strain types from two CF patients and two agricultural isolates from the rhizopheres of rice and maize were examined (38). The clinical isolates had identical 16S ribosomal rDNA sequences, but differences were seen between the soil isolates and clinical isolates. The results were presented as evidence of evolutionary divergence of the rhizophere isolates from their clinical counterparts. Alternatively, in the light of the multiple replicon model, this difference in 16S rrn is most likely due to the multiple replicons carrying varying sequences of the 16s gene. Furthermore, evidence based on four isolates alone is inadequate to support the safe application of rhizophere isolates. We have seen diversity among 16s rrn that is unrelated to source in a large collection of B. cepacia isolates (Holmes, Truong, Geisselsoder, Goldstein, unpub. data). The possibility of highly virulent strains of B. cepacia with the potential to survive intracellularly exists if the organism acquires virulence genes from B. pseudomallei, a very closely related pseudomonad. B. pseudomallei is a soil saprophyte that gives rise to melioidosis, a life-threatening tropical disease seen mostly in Southeast Asia. The disease has a broad clinical spectrum and can remain dormant for years before giving rise to sepsis and death. The pattern of disease produced by B. pseudomallei may warn of a spectrum of clinical consequences from B. cepacia acquisition. The transfer of genetic material between these two closely related organisms in the environment is highly probable, with the subsequent emergence of a hybrid pathogen. This speculation is supported by recent detection of insertion sequences within B. pseudomallei that have been identified in B. cepacia, including an isolate belonging to the highly transmissible transatlantic epidemic lineage (18,52). With the potential emergence of diseases related to new and developing practices in the food and agricultural world, it seems prudent that communication and information sharing between medical microbiologists, agricultural microbiologists, and public health professionals be optimized and promoted. Meanwhile, it is impossible to identify, phenotypically or genotypically, strains of B. cepacia that could safely be used in agriculture. Even if environmental strains incapable of human infection could be identified, their potential to evolve into human pathogens remains. Current molecular genetic evidence indicates that the deliberate environmental distribution of the organism might pose a hazard to human health, regardless of which particular strains are selected. Until more is known about the organism and the risks from environmental application, a moratorium should be called on the widespread use of B. cepacia in agriculture. This work was supported in part by a research grant award to RG from the National Institutes of Health (NIDDK RO1 DK50838). Address for correspondence: Alison Holmes, Department of Infectious Diseases, Hammersmith Hospital, Du Cane Rd, London W12 0NN, United Kingdom; fax: 44-181-383-3394; e-mail: aholmes@rpms.ac.uk. References 1. Burkholder W. Sour skin, a bacterial rot of onion bulbs. Phytopathology 1950;40:115-8. 2. Goldmann D, Klinger J. Pseudomonas cepacia: Biology, mechanisms of virulence, epidemiology. J Pediatr 1986;108:806-12. 3. Thomassen MJ, Demko C, Klinger, J, Stern, R. Pseudomonas cepacia colonization among patients with cystic fibrosis—a new opportunist. Am Rev Respir Dis 1985;131:791-6. 4. Sajjan U, Sun L, Goldstein R, Forstner J. Cable (Cbl) type II pili of cystic fibrosis-associated Burkholderia (Pseudomonas) cepacia: nucleotide sequence of the cblA major subunit pilin gene and novel morphology of the assembled appendage fibers. J Bacteriol 1995;177:1030-8. 5. Pseudomonas cepacia—more than a harmless commensal? [editorial]. Lancet 1992;329:1385-6. 6. Govan J, Deretic V. Microbial pathogenesis in cystic fibrosis: mucoid Pseudomonas aeruginosa and Burkholderia cepacia. Microbiological Reviews 1996;60:539-74. 7. Isles A, Maclusky I, Corey M, Gold R, Prober C, Fleming P, et al. Pseudomonas cepacia infection in cystic fibrosis: an emerging problem. J Pediatr 1984;104:206-10. 8. Tablan O, Chorba T, Schidlow D, White J, Hardy K, Gilligan P, et al. Pseudomonas cepacia colonization in patients with cystic fibrosis: risk factors and clinical outcome. J Pediatr 1985;107:382-7. 9. Thomassen MJ, Demko C, Doershuk C, Stern R, Klinger J. Pseudomonas cepacia: decrease in colonization in patients with cystic fibrosis. Am Rev Respir Dis 1986;134:669-71. 10. Tablan O, Martone W, Coershuk D, Stern R, Thomassen MJ, Klinger J, et al. Colonization of the respiratory tract with Pseudomonas cepacia in cystic fibrosis. Chest 1987;91:527-32. 11. Corey M, Farewell V. Determinants of mortality from cystic fibrosis in Canada. Am J Epidemiol 1996;143:1007-17. 12. Govan J, Brown P, Maddison J, Doherty C, Nelson J, Dodd M, et al. Evidence for transmission of Pseudomonas cepacia by social contact in cystic fibrosis. Lancet 1993;342:15-9. 13. Walters S, Smith E. Pseudomonas cepacia in cystic fibrosis: transmissibility and its implications. Lancet 1993;342:3-4. 14. Hearst J, Elliott K. Identifying the killer in cystic fibrosis. Understanding the genetic defects underlying CF is only half the battle. Identifying the specific bacterium infecting CF patients is just as important. Nature Medicine 1995;1(No. 7):626-7. 15. LiPuma J, Mortensen J, Dasen, S, Stull T. Ribotype analysis of Pseudomonas cepacia from cystic fibrosis treatment centers. J Pediatrics 1988;113:859-62. 16. Johnson W, Tyler S, Rozee K. Linkage analysis of geographic and clinical clusters in Pseudomonas cepacia infections by multilocus enzyme electrophoresis and ribotyping. J Clin Microbiol 1994;32:924-30. 17. Mahenthiralingam E, Simpson D, Speert D. Identification and characterisation of a novel DNA marker associated with epidemic strains recovered from patients with cystic fibrosis. J Clin Microbiol 1997;35:808-16. 18. Sun L, Jiang R-Z, Steinbach S, Holmes A, Campanelli C, Forstner J, et al. The emergence of a highly transmissible lineage of cbl+ Pseudomonas (Burkholderia) cepacia causing CF centre epidemics in North America and Britain. Nat Med 1995;1:661-6. 19. Pitt T, Kaufmann M, Patel P, Benge L, Gaskin S, Livermore D. Type characterization and antibiotic susceptibility of Burkholderia (pseudomonas) cepacia isolated from patients with cystic fibrosis in the United Kingdom and Republic of Ireland. J Med Microbiol 1996;44:203-10. 20. Segonds C, Bingen, B, Couetdic G, Mathy S, Brahimy N, Narty N, et al . Genotypic Analysis of Burkholderia cepacia isolates from 13 French cystic fibrosis centers. J Clin Microbiol 1997;35:2055-60. 21. Steinbach S, Sun L, Jiang R, Flume P, Gilligan P, Egan T, et al. Transmissibility of Pseudomonas cepacia infections in clinic patients and lung-transplant recipients with cystic fibrosis. N Engl J Med 1994;331:981-7. 22. Ryley H, Ojeniyi B, Hoiby N, Weeks J. Lack of evidence of cross-infection by Burkholderia cepacia among Danish cystic fibrosis patients. Eur J Clin Microbiol Infect Dis 1996;15:755-9. 23. Govan J, Hughes J, Vandamme P. Burkholderia cepacia: medical, taxonomic and ecological issues. J Med Microbiol 1996;45:395-407. 24. Taplin D, Bassett D, Mertz P. Foot lesions associated with Pseudomonas cepacia. Lancet 1971;568-71. 25. Jarvis W, Olson D, Tablan O, Martone J. The epidemiology of nosocomial Pseudomonas cepacia infections: endemic infections. Eur J Epidemiol 1987;3:233-6. 26. Anderson D, Kuhns J, Vasil M, Gerding D, Janoff E. DNA fingerprinting by pulsed field gel electrophoresis and ribotyping to distinguish Pseudomonas cepacia isolates from a nosocomial outbreak. J Clin Microbiol 1991;29:648-9. 27. Pegues D, Carson L, Anderson R, Norgard M, Argent T, Jarvis W, et al. Outbreak of Pseudomonas cepacia in oncology patients. Clin Infect Dis 1993;16:407-14. 28. Dixon R, Ksalow D, Mackel C, Fulkerson C, Mollison G. Aqueous quaternary ammonium antiseptics and disinfectants—use and misuse. JAMA 1977;236:2415-7. 29. Sobel J, Hashman N, Reinherz G, Merzbach D. Nosocomial Pseudomonas cepacia infection associated with chlorhexidine contamination. Am J Med 1982;73:183-6. 30. Hyde J, Humphreys H. Absence of Burkholderia cepacia from the respiratory tract of non-cystic fibrosis patients. Eur J Clin Microbiol Infect Dis 1997;16:253-4. 31. Holmes A, Nolan R, Finley R, Riley M, Sun L, Goldstein R. An epidemic clone of Burkholderia (Pseudomonas) cepacia transmitted between cystic fibrosis and non-cystic fibrosis patients. In Abstracts of the 9th Annual North American Cystic Fibrosis Conference; 1995; Dallas, Texas. Dallas (TX): Cystic Fibrosis Foundation; 1995, abstract 228. 32. Tablan O, Chorba T, Schidlow D, White J, Hardy K, Gilligan P, et al. Pseudomonas cepacia colonization in patients with cystic fibrosis: risk factors and clinical outcome. J Pediatr 1984;107:382-7. 33. Carson L, Tablan O, Cusick L, Jarvis W, Favero M, Bland L. Comparative evaluation of selective media for isolation of Pseudomonas cepacia from cystic fibrosis patients and environmental source. J Clin Microbiol 1988;26:2096-100. 34. Gilligan P, Gage P, Bradshaw L, Schidlow D, DeCicco B. Isolation medium for the recovery of Pseudomonas cepacia from respiratory secretions of patients with cystic fibrosis. J Clin Microbiol 1985;22:5-8. 35. McLoughlin T, Quinn J, Bettermann A, Bookland R. Pseudomonas cepacia suppression of sunflower wilt fungus and role of antifungal compounds in controlling disease. Appl Environ Microbiol 1992;58:1760-3. 36. Homma Y, Sato Z, Hirayama F, Kanno K, Shirahama H, Suzui T. Production of antibiotics by Pseudomonas cepacia as an agent for biological control of soilborne pathogens. Soil Biology and Biochemistry 1989;21:723-8. 37. Sangodkar U, Chapman P, Chakrabarty A. Cloning, physical mapping and expression of chromosomal genes specifying degradation of the herbicide 2,4,5-T by Pseudomonas cepacia AC1100. Gene 1988;71:267-77. 38. Tabacchioni S, Visca P, Chiarini L, Bevivino A, Di Serio C, Francelli S, et al. Molecular characterization of rhizosphere and clinical isolates of Burkholderia cepacia. Res Microbiol 1995;146:531-42. 39. Bowers J, Parke J. Epidemiology of Pythium damping-off and Aphanomyces root rot of peas after seed treatment with bacterial agents for biological control. Phytopathology 1993;83:1466-73. 40. King E, Parke J. Population density of the biocontrol agent Burkholderia cepacia AMMDR1 on four pea cultivars. Soil Biology and Biochemistry 1996;28:306-12. 41. Joy A, Parke J. Biocontrol of Alternaria leaf blight on American ginseng by Burkholderia cepacia AMMD. In: Bailey WG, Whitehead C, Procter JTA, Kyle JT, editors. Challenges of the 21st century. Proceedings of the International Ginseng Conference; 1994; Vancouver, British Columbia, Canada. Burnaby, B.C., Canada: Simon Fraser Univ.; 1994. 42. Cartwright D, Chilton C, Benson D. Pyrrolnitrin and phenazine production by Pseudomonas cepacia strain 5.5B, a biocontrol agent of Rhizoctonia solani. Appl Microbiol Biotechnol 1995;43:211-6. 43. Reddy M. Status on commercial development of Burkholderia cepacia for biological control of fungal pathogens and growth enhancement of conifer seedlings for a global market. Washington: US Forest service general technical report PNW; 1997. Report no. 389. p. 235-44. 44. Wise M, Shimkets L, McArthur J. Genetic structure of a lotic population of Burkholderia (Pseudomonas) cepacia. Appl Environ Microbiol 1995;61:1791-8. 45. Ursing JB, Rosello-Mora RA, Garcia-Valdes E, Lalcut J. A pragmatic approach to the nomenclature of phenotypically similar genomic groups. Int J Syst Bacteriol 1995;45:604. 46. Vandamme P, Holmes B, Vancanneyt M, Coenye T, Hoste B, Coopman R, et al. Occurrence of multiple genomovars of Burkholderia cepacia in cystic fibrosis patients and proposal of Burkholderia multivorans sp nov. Int J Syst Bacteriol 1997;47:1188-200. 47. Cheng H, Lessie T. Multiple replicons constituting the genome of Pseudomonas cepacia 17616. J Bacteriol 1994;176:4034-2. 48. Holmes A, Zhou H, Sun L, Goldstein R. Multiple chromosome variability amongst clinical and environmental isolates of Burkholderia (Pseudomonas) cepacia; Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) 1996; San Francisco. San Francisco: ICAAC; 1995; Abstract C51. 49. Allardet-Servent A, Michaux-Charachon S, Jumas-Bilak E, Karayan L, Ramuz M. Presence of one linear and one circular chromosome in the Agrobacterium tumefaciens C58 genome. J Bacteriol 1993;175:7869-74. 50. Suwanto A, Kaplan S. Physical and genetic mapping of the Rhodobacter sphaeroides 2.4.1 genome: genome size, fragment identification, and gene localisation. J Bacteriol 1989;171:5840-9. 51. Michaux S, Paillisson J, Charles-Nurit M, Bourg G, Allardet-Servent A, Ramuz M. Presence of two independent chromosomes in the Brucella melitensis 16M genome. J Bacteriol 1993;175:701-5. 52. Mark K, Titball R. The detection of insertion sequences within the human pathogen Burkholderia pseudomallei which have been identified previously in Burkholderia cepacia. FEMS Microbiol Letts. In press 1998. Emerging Infectious Diseases National Center for Infectious Diseases Centers for Disease Control and Prevention Atlanta, GA URL: http://www.cdc.gov/ncidod/EID/vol4no2/holmes.htm Please note that figures and equations are not available in ASCII format; their placement within the text is noted by [fig] and [eq], respectively. Greek symbols are spelled out. The following codes are used: (ft) for footnote; (sup) for superscript; (sub) for subscript; /= for greater than or equal to. ------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------- [Emerging Infectious Diseases] [Volume 4 No. 2 / April - June 1998] Synopses Haemophilus influenzae Invasive Disease in the United States, 1994-1995: Near Disappearance of a Vaccine-Preventable Childhood Disease Kristine M. Bisgard, Annie Kao, John Leake, Peter M. Strebel, Bradley A. Perkins, and Melinda Wharton Centers for Disease Control and Prevention, Atlanta, Georgia, USA --------------------------------------------------------------------------- We analyzed national Haemophilus influenzae (Hi) surveillance data from 1994 and 1995 to describe the epidemiology of Hi invasive disease among persons of all ages. Serotype data were available for 376 (56%) of 669 reported Hi cases among children aged 4 years or younger; 184 (49%) were H. influenzae type b (Hib). Among children aged 4 or younger, incidence (per 100,000) of all Hi invasive disease was 1.8 in 1994 and 1.6 (p < 0.05) in 1995. Children aged 5 months or younger had the highest average annual incidence rate of Hib invasive disease (2.2 per 100,000); children aged 6 to 11 months had the next highest rate (1.2 per 100,000) (p < 0.05). Of 181 children with Hib invasive disease whose age in months was known, 85 (47%) were too young (aged 5 months or younger) to have completed a primary series with an Hib-containing vaccine. Of the 83 children with known vaccination status who were eligible to receive a primary series (aged 6 months or older), 52 (63%) were undervaccinated, and the remaining 31 (37%) had completed a primary series in which vaccine failed. Among persons aged 5 years or older with Hi invasive disease, the lowest average annual incidence was among those 20 to 39 years of age (0.15 per 100,000), and the highest was among those aged 80 years or older (2.26 per 100,000). Among persons aged 5 years or older, serotype data were available for 1,372 (71%) of the 1,940 Hi invasive disease cases; 159 (28%) of the 568 Hi cases with known serotype were due to Hib. Before the first Haemophilus influenzae type b (Hib) polysaccharide vaccine was introduced in 1985, Hib was the most common cause of bacterial meningitis in children under 5 years of age (approximately 12,000 cases per year, most in children younger than 18 months [1]); approximately 5% of affected children died. Neurologic sequelae developed in 15% to 30% of the surviving children (1). An additional estimated 7,500 cases of other invasive Hib infections also occurred annually in young children (1). The cumulative risk for Hib invasive disease before the age of 5 was one in 200 children, similar to the risk for poliomyelitis during the 1950s (1,2). The incidence of Hib invasive disease among children aged 4 years or younger has declined by 98% since the introduction of Hib conjugate vaccines, initially licensed in 1989 for use in children aged 15 months or older, then for infants beginning at 2 months of age in 1990 (3-6). One goal of the Childhood Immunization Initiative (7) was to eliminate invasive Hib disease among children aged 4 years or younger by 1996; however, approximately 300 cases of H. influenzae (Hi) invasive disease per year continue to be reported (3). Moreover, Hi invasive disease is an important cause of illness and death in immunocompromised persons and adults with certain underlying medical conditions (8-11). The average incidence of Hi invasive disease from 1988 to 1990 was estimated at 1.7 cases per 100,000; 50% of cases examined in this study were due to Hib (8). Prior studies of Hi invasive disease among adults found case-fatality rates of 26% to 36% (8-11). In this article, we summarize the epidemiology of Hi and Hib invasive disease among children aged 4 years or younger and persons aged 5 years or older in the United States during 1994 and 1995. Data Collection We used three sources of surveillance data to document the number of reported Hi invasive disease cases with onset in 1994 and 1995. Cases were reported to the Centers for Disease Control and Prevention (CDC). Two sources were passive surveillance systems: the National Notifiable Diseases Surveillance System (NNDSS) and the National Bacterial Meningitis and Bacteremia Reporting System (NBMBRS); the third source was the active, laboratory-based surveillance sites. The NNDSS case reports of Hi invasive disease are transmitted electronically from state health departments to CDC as part of the weekly notifiable disease reporting system. These reports contain only demographic information. Supplemental information, including Hi serotype, infection site, outcome, and Hib vaccination status and specific vaccine received among children aged 4 years or younger is transmitted electronically or mailed to CDC on NBMBRS reporting forms. During 1994, 10.5 million persons were under active, laboratory-based surveillance for Hi invasive disease: these included residents of the state of Oklahoma; three counties in the San Francisco Bay area of California; eight counties in the metropolitan area of Atlanta, Georgia; and four counties in Tennessee (3). In 1995, 12.8 million persons were under surveillance; an additional county in Tennessee was added, and Maryland was included in place of Oklahoma. Each site was contacted every 2 weeks to identify Hi invasive disease cases; surveillance personnel then collected detailed data on these cases. Data on cases from the active surveillance sites were also mailed to CDC on a form similar to the NBMBRS report form. Information from the three data sources was combined, and duplicates (cases with identical date of birth, onset, county of residence, and demographic data) were eliminated. After the end of each calendar year, each state health department was contacted by telephone to send information on Hi invasive disease cases among children aged 4 years or younger not previously reported to CDC and to provide missing critical supplementary information, such as vaccinations received. Our analysis included confirmed and probable cases. A confirmed case of Hi invasive disease was defined as an illness clinically compatible with invasive disease (such as meningitis, pneumonia, cellulitis, epiglottitis, peritonitis, pericarditis, septic arthritis, empyema, and abscesses) with isolation of Hi from a normally sterile site (e.g., cerebrospinal fluid or blood). A probable case was defined as an illness clinically compatible with detection of Hib antigen in cerebrospinal fluid (12). Up to three infection sites could be recorded from a list including primary bacteremia, meningitis, pneumonia, cellulitis, epiglottitis, peritonitis, pericarditis, septic arthritis, or another specified site. Completion of a primary vaccination series was defined according to recommendations from the Advisory Committee on Immunization Practices (4). Hib vaccine doses administered fewer than 14 days before the onset of disease were excluded. Data Analysis U.S. census population estimates for 1994 and 1995 were used to calculate age- and race-specific average annual incidence rates (calculated by [events/2] ÷ [population/2]) for all Hi invasive disease and for documented Hib disease (13). The difference between two rates was regarded statistically significant at the 0.05 level if it exceeded the following: 2 x [R1(sup 2)/N1 + R2(sup 2)/N2](sup 1/2) , where R1 is the rate corresponding to N1 events and R2 is the rate corresponding to N2 events (14). County average annual incidence rates of Hi invasive disease were calculated by using 1990 county population estimates and were mapped with Atlas Mapmaker (15). For Hi invasive disease cases among children aged 4 years or younger, we looked at the proportion due to Hib, the Hib vaccination history, and the age-specific and geographic distribution of residual Hib cases. A chi-square test of independence was used to test the association between age category and outcome. Findings In 1994 and 1995, 1,277 and 1,332 cases of invasive Hi disease were reported, respectively. Although the overall incidence of invasive Hi disease remained constant, the incidence of invasive Hi disease among children aged 4 years or younger was lower in 1995 (305 cases; 1.56 per 100,000) than in 1994 (364 cases; 1.84 per 100,000) (p < 0.05). During 1994 and 1995, children 5 months or younger had the highest reported disease incidence of Hi invasive disease (8.02 per 100,000 children), Hib invasive disease (2.20 per 100,000), and Hi invasive disease of unknown serotype (3.57 per 100,000) (Table 1). The incidence of Hi invasive disease, Hib invasive disease, and Hi invasive disease of unknown serotype was significantly greater (p < 0.05) among children aged 5 months or younger than among children 6 to 11 months of age, who had the next highest rates; a high rate of Hi invasive disease was also observed among adults aged 80 years or older. The lowest disease rates were found in persons 5 to 19 and 20 to 39 years of age. Hi isolates were serotyped for only 36% of 2,609 reported cases. A higher percentage of isolates were serotyped among children aged 4 years or younger (56%) than among older persons (29%) (Table 2). Among Hi cases in children aged 4 years or younger with reported serotype information, Hib was the most common single serotype (184 type b of 376 serotyped isolates); Hib represented 49% of invasive disease and 65% of meningitis. The average annual incidence for Hib invasive disease was 0.47 per 100,000 children aged 4 years or younger. Among persons aged 5 years or older, Hib accounted for 28% of serotyped isolates. Among the 433 Hi case-patients younger than 1 year of age for whom age in months was reported during the study period, most (72%) were aged 5 months or younger; 29% were younger than 1 month of age (Figure 1). Similarly, of 184 children aged 4 years or younger with Hib invasive disease, most (133, 72%) cases were in children younger than 1 year of age. Table 1. Number of cases and average annual incidence rates of Hi and Hib invasive disease and Hi invasive disease of unknown serotype, by age(sup a), United States, 1994 - 1995 Hi Hib Hi serotype unknown ------------------- -------------------- -------------------------- Age group No.(sup b)Incid.(sup c)No.(sup b)Incid.(sup c)No.(sup b)Incid.(sup c ---------------------------------------------------------------------------- 0-5 mo 310 8.02 85 2.20 138 3.57 6-11 mo 123 3.18 45 1.16 45 0.16 1-4 yr 219 0.69 51 0.16 99 0.31 5-19 yr 179 0.16 23 0.02 108 0.10 20-39 yr 249 0.15 25 0.02 166 0.10 40-59 yr 405 0.33 42 0.03 280 0.23 60-79 yr 715 1.01 51 0.07 539 0.76 80 + yr 361 2.26 18 0.11 152 0.95 --------------------------------------------------------------------------- (sup a)Does not include 31 cases missing age and 17 patients aged <1 year but missing age in months. (sup b)Number of cases over the 2-year period. (sup c)Incidence per 100,000 population. The highest average annual incidence rates (5.1 per 100,000) of Hi invasive disease among case-patients aged 4 years or younger were in American Indians; comparisons with other race or ethnic groups were statistically significant (p /= 0.05) (Table 3). Similar race and ethnic differences were noted for Hib invasive disease incidence but not for Hi invasive disease of unknown serotypes. Among case-patients aged 5 years or older, differences between race or ethnic groups were less marked. The incidence rates among male and female case-patients aged 5 years or older and among case-patients aged 4 years or younger were equivalent. Infection site data were available for 1,556 (60%) case-patients (Table 4). Bacteremia was the most frequently reported site of Hi invasive disease for both children aged 4 years or younger (62%) and persons aged 5 years or older (51%). The second most frequently reported site of invasive disease among children aged 4 years or younger was meningitis (44%), and among persons aged 5 years or older it was pneumonia (36%). Table 2. Serotype information for reported Hi invasive disease (n=2609) and meningitis cases (n=380), stratified by age, United States, 1994 - 1995 -------------------------------------------------------- No. (%) of children No. (%) of persons aged 0-4 years aged >/=5 years Invasive Invasive Serotype disease Meningitis disease Meningitis -------------------------------------------------------- Serotype b 184 (27) 101 (47) 159 (8) 41 (25) Other(sup a) 114 (17) 39 (18) 171 (9) 26 (15) Non-typeable 78 (12) 16 (8) 238 (12) 25 (15) Unknown 293 (44) 57 (27) 1,372(71) 75 (45) Total 669 (100) 213 (100) 1,940(100) 167 (100) -------------------------------------------------------- (sup a)Includes serotypes a, c, d, e, f. Outcome data were available for 1,380 (53%) of 2,609 case-patients with Hi invasive disease. Among case-patients with known outcome, the overall case-fatality ratio (CFR) was 14%. The CFR differed by age; the CFRs among case-patients aged 0 to 1 year, 1 to 29 years, 30 to 59 years, and 60 years of age or older were 10%, 6%, 13%, and 24%, respectively (p < 0.001). During 1994 and 1995, a moderate seasonal pattern of Hi and Hib invasive disease was observed, with more cases occurring during the winter months (Figure 2); this pattern was similar for children 4 years of age or younger (data not shown). [fig 1] Figure 1. Reported Hi invasive disease cases by age and serotype, children aged <1 year,a United States, 1994-1995. (sup a)N=433; excluding 17 children aged <1 year missing age in months. Table 3. Number of cases and average annual incidence rates of Hi and Hib invasive disease and Hi invasive disease of unknown serotype, by race and ethnicity,a United States, 1994 - 1995 Persons aged >/=5 Children aged 0-4 year years ------------------------------------------------- Hi: unknown All Hi Hib serotypes All Hi ------------------- ------------------ -------------- --------------- Race/ethnicity No.(sup b)Incid.(sup c)No.(sup b)Incid.(sup c)No.(sup b)Incid.(sup c) ------------------------------------------------------------------------------------- Non-Hispanic White 338 1.34 102 0.40 150 0.59 1,269 0.35 Black 122 2.08 32 0.55 52 0.89 258 0.45 American Indian 18 5.10 10 2.83 3 0.84 12 0.37 Asian or Pacific Islander 13 0.86 3 0.20 6 0.40 21 0.13 Hispanic 94 1.49 7 0.11 31 0.49 70 0.15 ------------------------------------------------------------------------------------- (sup a)Does not include 84 (13%) of children aged /= 4 years, and 310 (16%) of persons aged >/= 5 years for whom race and ethnicity data were missing. (sup b)Number of cases over the 2-year period. (sup c)Per 100,000 population. To determine the remaining cases of Hib invasive disease, data on the number of Hib vaccine doses received were reviewed for the 184 reported cases among children aged 4 years or younger by the age (in months) of the case-patient (Table 5). Eighty-five (47%) of the 181 children aged 4 years or younger with known age in months were too young (5 months or younger) to have completed a primary series with the most commonly used Hib vaccines (three-dose primary series). Of the 149 children aged 4 years or younger with known vaccination status, 83 were eligible (aged 6 months or older) to receive a primary series; of these, 52 (63%) were undervaccinated. Thirty-one (37%) of the 83 children had completed a primary series with either three doses of HbOC (Wyeth-Lederle Laboratories, Pearl River, NY) or PRP-T (Pasteur Mérieux Connaught, Lyon, France), or with two doses of PRP-OMP (Merck, Inc., West Point, PA), or one dose with any Hib-containing vaccine on or after 15 months of age. Manufacturer and vaccine lot information was available for 22 of the 31 children who had received a primary series. No single lot predominated, and no child received vaccine from the same lot for all doses. Table 4. Reported infection site(sup a) among Hi invasive disease cases, (sup b)United States, 1994 - 1995 Number (%) ---------------------------------- Children aged Persons aged Site 0 - 4 years(sup c) >/= 5 years(sup d) ---------------------------------------------------- Bacteremia(sup e) 299 (62) 551 (51) Meningitis 213 (44) 167 (16) Pneumonia 42 (9) 390 (36) Cellulitis 19 (4) 7 (1) Epiglottitis 8 (2) 19 (2) ---------------------------------------------------- (sup a)Includes case-patients with multiple infection sites, such as meningitis and bacteremia. (sup b)Case-patients with known infections site(n=1,556) (sup c)N=485. (sup d)N=1,071. (sup e)Bacteremia is reported following positive blood culture results, with or without foci. [fig 2] Figure 2. Reported Hi invasive disease cases by serotype and month of onset,a United States, 1994-1995. (sup a)N=2,609; excluding 10 cases with unknown month of onset. Table 5. Children aged /=15 months of age. (sup c)Cumulative recommended doses for PRP-T and HbOC vaccines. Overall, 49 states and 673 (21%) of the 3,137 U.S. counties reported at least one Hi invasive disease case with disease onset during the 2-year period; of these, 327 (49%) counties (from 47 states) reported two or more Hi cases (Figure 3A). Forty-seven states and 339 (11%) of counties reported Hi invasive disease cases among children aged 4 years or younger, and 42 states and 137 (4%) counties reported Hib cases during the 2-year period; 25 (18%) of these counties reported two or more Hib cases (range 2 to 8 cases). In these 25 counties, the incidence for Hib invasive disease was 0.41 to 12.19 per 100,000 children aged 4 years or younger (Figure 3B). Among the 14 counties with two Hib cases, the median time between cases was 139 days (range 14 to 334 days). In three of the 11 counties with three or more Hib invasive disease cases, children were too young (aged 5 months or younger) to have completed a primary series with an Hib vaccine; in the remaining eight counties, 20 (67%) of 30 children had been eligible for vaccination (aged 6 months or older) and only one child had completed a primary series. In these 11 counties, the incidence for Hib invasive disease was 0.41 to 4.31 per 100,000 children aged 4 years or younger. [fig 3] Figure 3. A. Average annual incidence of Hi invasive disease among persons of all ages, by county, United States, 1994 - 1995. B. Reported number of Hib invasive disease cases among children aged 90%) found a carriage level of 8%, a rate similar to that in the prevaccine era (24). However, a 1992 study in the Atlanta, Georgia, metropolitan area documented a low (0.2%) carriage level among children with a 75% Hib vaccination coverage (25). In the United States, the reduction in the number of reported cases has been greater than expected when the efficacy of Hib conjugate vaccines and the coverage of a primary series alone are taken into account; this is likely due to the reduction of pharyngeal carriage in vaccinated children, which protects nonvaccinated persons in the community (i.e., herd immunity) (18,26). High carriage levels are unlikely in most areas of the United States; the difference in the results of the two studies (24,25) may be due to the use of different Hib conjugate vaccines, as well as to different host and environmental factors. In European countries that use Hib conjugate vaccines, increasing vaccine use has been associated with diminishing rates of both Hib pharyngeal carriage and invasive Hib disease (26). With the marked decline among children, the occurrence of Hib invasive disease among older persons (aged 5 years or older) has achieved greater prominence. Although serotype data were missing from most cases, reported cases among both children (n = 184) and adults (n = 159) indicate persistent circulation of Hib. Reported Hib invasive disease cases and nasopharyngeal carriage among children have decreased; however, because these data were from passive reporting sources it is not clear if a similar decrease in invasive Hib disease in older children and adults has occurred nationally. Because of underreporting of adult Hib invasive disease cases nationally, 1994 to 1995 incidence rates cannot be directly compared with rates found in prior studies (8-11). In our study, 28% of serotyped isolates were Hib; however, among adults in the Atlanta, Georgia, metropolitan active surveillance site, Hi invasive disease cases due to Hib declined in 1990 to 1991 from 50% to 0% in 1995 to 1996 (8,27). Nevertheless, because previous studies have found that most Hi and Hib invasive disease occurred among immunocompromised patients or patients with underlying conditions (e.g., chronic lung disease, splenectomy, leukemia, HIV infection, sickle-cell disease), health-care providers should consider Hib vaccination for patients at high risk (6). Because of continued occurrence of Hib invasive disease among older persons and infants too young to be vaccinated, additional strategies may be needed to eliminate Hib invasive disease among children aged 4 years or younger. In addition, ongoing surveillance for Hi invasive disease is needed to monitor incidence among older persons; case investigations may include assessments for underlying immunosuppressive conditions or ascertainment of other cofactors such as prior or concomitant viral infections. Continued decline in incidence of Hib invasive disease in the United States requires that children aged 4 years or younger be appropriately vaccinated to protect them and other children and infants in the community. Intensive efforts to increase vaccination levels among children in areas with coverage rates lower than 90% are needed to reach the elimination goal. Children should receive a primary Hib vaccination series at ages 2, 4, and 6 months with either HbOC or PRP-T, or at ages 2 and 4 months with PRP-OMP (6). A booster dose with any of the four licensed Hib conjugate vaccines should be given at 12 to 15 months of age (6). Among children who were eligible for vaccination, timely completion of a primary series might have prevented 63% of reported Hib invasive disease cases in 1994 and 1995. Few (28) Hib cases among children 4 years or younger were reported in those who had completed a primary series with Hib vaccine, which suggests that vaccine failure was uncommon. However, because of passive reporting, vaccination status and vaccine failure may be underestimated. Hib invasive disease in children too young to have completed a primary series will continue to be a barrier for elimination unless Hib nasopharyngeal carriage rates in the population (including adults) approach zero. Because the highest Hib invasive disease incidence occurs among infants, all infants should be appropriately vaccinated, beginning at 2 months of age. In a case-control study for undervaccination among Hib invasive disease cases among children aged 2 to 18 months in the United States, risk factors for undervaccination included low maternal education and having a single mother (29); special outreach efforts may be needed to vaccinate children at risk. Other programmatic strategies to increase vaccination rates include linkage of the Special Supplemental Nutrition Program for Women, Infants and Children (WIC) to vaccination assessment and referral (30), implementation of immunization registries and reminder/recall systems to encourage parents to vaccinate their children in a timely fashion (28,31,32), and monitoring clinic coverage levels and providing feedback to clinic staff (33). In addition to high vaccination coverage of other children in the community to help reduce carriage, breast-feeding of infants should be promoted because it is protective against invasive Hi disease in this age group (21). For children aged 4 years or younger with Hib invasive disease who have completed a primary Hib vaccination series, case investigations may include examining reasons for vaccine failure, such as measuring antibodies to the polyribosylribitol phosphate (PRP) capsule of Hib; undetectable levels suggest lack of immune response (18). To monitor the changing epidemiology of Hib invasive disease and look for an increase or decline in incidence among children and adults, serotype information for all Hi invasive disease cases is essential and will generate data needed to examine new strategies for elimination. Surveillance is necessary for detecting pockets of continuing transmission and identifying areas in need of improved vaccine coverage (3). Each case of Hib invasive disease is a sentinel for circulation of Hib in a community. In addition, Hib invasive disease cases among children aged 4 years or younger may be a marker for low community coverage. Therefore, community immunization policies, practices, and coverage levels should be reviewed, and recommendations should be made, if needed, to improve Hib vaccination coverage levels. Acknowledgments The authors thank Wendy Wattigney for her review of the manuscript and assistance with statistical analysis and Shalani Desai for her meticulous assistance in data cleaning. Dr. Bisgard is a medical epidemiologist in CDC's National Immunization Program. Her research interests include vaccine effectiveness, uses of new vaccines, and outbreak control of vaccine-preventable diseases; her areas of expertise include vaccine preventable diseases (Hib, pertussis, and diphtheria) and infectious disease epidemiology. Address for correspondence: Kristine M. Bisgard, Centers for Disease Control and Prevention, 1600 Clifton Road, Mail Stop E61, Atlanta, GA 30333, USA; fax: 404-639-8616; e-mail: kmb6@cdc.gov. References 1. Cochi SL, Broome CV, Hightower AW. Immunization of U.S. children with Haemophilus influenzae type b polysaccharide vaccine: a cost-effectiveness model of strategy assessment. JAMA 1985;253:521-9. 2. Strebel PM, Sutter RW, Cochi SL, Biellik RJ, Brink EW, Kew OM, et al. Epidemiology of poliomyelitis in the United States one decade after the last reported case of indigenous wild virus-associated disease. Clin Infect Dis 1992;14:568-79. 3. Centers for Disease Control and Prevention. Progress towards elimination of Haemophilus influenzae type b disease among infants and children-United States, 1987-1995. MMWR Morb Mortal Wkly Rep 1996;45:901-6. 4. Adams WG, Deaver KA, Cochi SL, Plikaytis BE, Bell ER, Broome CV, et al. Decline of childhood Haemophilus influenzae type b (Hib) disease in the Hib vaccine era. JAMA 1993;269:221-6. 5. Centers for Disease Control. Haemophilus b conjugate vaccines for prevention of Haemophilus influenzae type b disease among infants and children two months of age and older: recommendations of the Immunization Practices Advisory Committee (ACIP). MMWR Morb Mortal Wkly Rep 1991;40(RR-1):1-6. 6. Centers for Disease Control and Prevention. Recommendations for use of Haemophilus b conjugate vaccines and a combined diphtheria, tetanus, pertussis and Haemophilus type b vaccine: Recommendations of the Immunization Practices Advisory Committee (ACIP). MMWR Morb Mortal Wkly Rep 1993;42(RR-13):1-15. 7. Centers for Disease Control and Prevention. Reported vaccine-preventable disease-United States, 1993, and the Childhood Immunization Initiative. MMWR Morb Mortal Wkly Rep 1994;43:57-60. 8. Farley MM, Stephen DS, Brachman PS, Harvey C, Smith JD, Wenger JD, et al. Invasive Haemophilus influenzae disease in adults. A prospective, population-based surveillance. Ann Intern Med 1992;116:806-12. 9. Takala AK, Eskola J, van Alpen L. Spectrum of invasive Haemophilus influenzae type b disease in adults. Arch Intern Med 1990;150:2573-6. 10. Steinhart R, Reingold AL, Taylor F, Anderson G, Wenger JD. Invasive Haemophilus influenzae infections in men with HIV infection. JAMA 1992;268:3350-2. 11. Kostman JR, Sherry BL, Filgner CL, Egaas S, Sheeran P, Baken L, et al. Invasive Haemophilus influenzae infections in older children and adults in Seattle. Clin Infect Dis 1993;17:389-96. 12. Centers for Disease Control and Prevention. Case definitions for infectious conditions under public health surveillance. MMWR Morb Mortal Wkly Rep 1997;46(RR-10):15. 13. Resident population-estimates by age, sex, and Hispanic origin. Washington: U.S. Dept. of Commerce, Bureau of the Census, Population Division; 1995 July. Appendix A. 14. Technical notes: nature and sources of data. Hyattsville (MD): Centers for Disease Control and Prevention, National Center for Health Statistics; 1992. Monthly vital statistics report 41:21. 15. Atlas GIS Software (computer program). Arlington (VA): ATLAS Software, Claritas, Inc. 16. Centers for Disease Control and Prevention. Vaccination coverage levels of 2-year-old children United States, January-March 1994. MMWR Morb Mortal Wkly Rep 1995;44:142-50. 17. Centers for Disease Control and Prevention. National, state, and urban area vaccination coverage levels among children aged 19-35 months United States, January-December 1995. MMWR Morb Mortal Wkly Rep 1997;46:176-82. 18. Ward J, Leiberman JM, Cochi SL. Haemophilus influenzae vaccines. Vaccines. 2nd ed. In: Plotkin SA, Mortimer EA, editors. Philadelphia: WB Saunders Co. 1994;337-86. 19. Murphy TV, Osterholm MT, Pierson LM, White KE, Breedlove LVN, Siebert GB, et al. Prospective surveillance of Haemophilus influenzae type b disease in Dallas County, Texas and in Minnesota. Pediatr 1987;79:173-80. 20. Ward JI, Lum MKW, Hall DB, Silimperi DR, Bender TR. Invasive Haemophilus influenzae type b disease in Alaska: background epidemiology for a vaccine efficacy trial. J Infect Dis 1986;153:17-26. 21. Cochi SL, Fleming DW, Hightower AW, Limpakarnjanarat K, Facklam RR, Smith JD, et al. Primary invasive Haemophilus influenzae type b disease: a population-based assessment of risk factors. J Pediatr 1986;108:887-96. 22. Granoff DM, Basdon M. Haemophilus influenzae infections in Fresno County, California: a prospective study of the effects of age, race, and contact with a case on incidence of disease. J Infect Dis 1980;141:40-6. 23. Standaert SM, Lefkowitz LB, Horan JM, Hutcheson RH, Schaffner W. The reporting of communicable diseases: a controlled study of Neisseria meningitidis and Haemophilus influenzae infections. Clin Infect Dis 1995;20:30-6. 24. Galil K, Singleton R, Levine O, Fitzgerald M, Ajello G, Bulkow L, Parkinson A. High prevalence of Haemophilus influenzae type b (Hib) carriage among Alaska Natives despite widespread use of Hib conjugate vaccine. In: Abstracts of the 35th Infectious Disease Society of America; San Francisco, California; 1997 Sep 13-16; Abstract 421. Alexandria (VA): Infectious Disease Society of America; 1997. 25. Mohle-Boetani JC, Ajello G, Breneman E, Deaver KA, Harvey C, Plikaytis BD, et al. Carriage of Haemophilus influenzae type b in children after widespread vaccination with conjugate Haemophilus influenzae type b vaccines. Pediatr Infect Dis J 1993;12:589-93. 26. Barbour ML. Conjugate vaccines and the carriage of Haemophilus influenzae type b. Emerg Infect Dis 1996;3:176-82. 27. Farley MM, Baughman W, Robinson K, Facklam R, Schuchat A, Wenger JD. Elimination of invasive Haemophilus influenzae type b (Hib) disease in adults in Atlanta through immunization of children with conjugate Hib vaccine. In: Abstracts of the 37th Interscience Conference on Antimicrobial Agents and Chemotherapy; Toronto, Canada. Abstract K-150. Washington: American Society for Microbiology, Washington; 1997. p. 352. 28. Jafari H, Adams W, Deaver K, Plikaytis B, Wenger J. Efficacy of Haemophilus influenzae Type b (Hib) conjugate vaccine, risk factors for invasive Hib disease and under-vaccination in the United States. Abstracts of the Interscience Conference on Antimicrobial Agents and Chemotherapy; San Francisco, California; 1995; Abstract # G12. Washington: American Society for Microbiology; 1995. p. 160. 29. Centers for Disease Control and Prevention. Recommendations of the Advisory Committee on Immunization Practices: programmatic strategies to increase vaccination coverage by age 2 years-linkage of vaccination and WIC services. MMWR Morb Mortal Wkly Rep 1996;45:217-8. 30. Campbell JR, Szilagyi PG, Rodewald LE, Doane C, Roghmann KJ. Patient-specific reminder letters and pediatric well-child-care show rates. Clin Pediatr (Phila) 1994;May:268-72. 31. Abramson JS, O'Shea MO, Ratledge DL, Lawless MR, Givner LB. Development of a vaccine tracking system to improve the rate of age-appropriate primary immunization in children of lower socioeconomic status. J Pediatr 1995;126:583-6. 32. Dini EF, Linkins RW, Chaney M. Effectiveness of computer-generated telephone messages in increasing clinic visits. Arch Pediatr Adolesc Med 1995;149:902-5. 33. Centers for Disease Control and Prevention. Recommendations of the Advisory Committee on Immunization Practices: programmatic strategies to increase vaccination rates assessment and feedback of provider-based vaccination coverage information. MMWR Morb Mortal Wkly Rep 1996;45:219-20. Emerging Infectious Diseases National Center for Infectious Diseases Centers for Disease Control and Prevention Atlanta, GA URL: http://www.cdc.gov/ncidod/EID/vol4no2/bisgard.htm Please note that figures and equations are not available in ASCII format; their placement within the text is noted by [fig] and [eq], respectively. Greek symbols are spelled out. The following codes are used: (ft) for footnote; (sup) for superscript; (sub) for subscript; /= for greater than or equal to. ------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------- [Emerging Infectious Diseases] [Volume 4 No. 2 / April - June 1998] Synopses Multiple-Drug Resistant Enterococci: The Nature of the Problem and an Agenda for the Future Mark M. Huycke,*† Daniel F. Sahm,‡ and Michael S. Gilmore† *University of Oklahoma Health Sciences Center, Oklahoma, USA; †Department of Veterans Affairs Medical Center, Oklahoma City, Oklahoma, USA; and ‡MLR Pharmaceutical Services, Inc., Reston, Virginia, USA -------------------------------------------------- Enterococci, leading causes of nosocomial bacteremia, surgical wound infection, and urinary tract infection, are becoming resistant to many and sometimes all standard therapies. New rapid surveillance methods are highlighting the importance of examining enterococcal isolates at the species level. Most enterococcal infections are caused by Enterococcus faecalis, which are more likely to express traits related to overt virulence but—for the moment—also more likely to retain sensitivity to at least one effective antibiotic. The remaining infections are mostly caused by E. faecium, a species virtually devoid of known overt pathogenic traits but more likely to be resistant to even antibiotics of last resort. Effective control of multiple-drug resistant enterococci will require 1) better understanding of the interaction between enterococci, the hospital environment, and humans, 2) prudent antibiotic use, 3) better contact isolation in hospitals and other patient care environments, and 4) improved surveillance. Equally important is renewed vigor in the search for additional drugs, accompanied by the evolution of new therapeutic paradigms less vulnerable to the cycle of drug introduction and drug resistance. The past few years have witnessed increasing interest in enterococci. Until recently, these ordinary bowel commensals languished as misclassified streptococci, commonly perceived "with the exception of endocarditis and rare cases of meningitis … [as] not … a major cause of serious infection" (1). In the last decade, however, enterococci have become recognized as leading causes of nosocomial bacteremia, surgical wound infection, and urinary tract infection (2,3). Two types of enterococci cause infections: 1) those originating from patients' native flora, which are unlikely to possess resistance beyond that intrinsic to the genus and are unlikely to be spread from bed to bed, and 2) isolates that possess multiple antibiotic resistance traits and are capable of nosocomial transmission. The therapeutic challenge of multiple-drug resistant (MDR) enterococci—those strains with significant resistance to two or more antibiotics, often including, but not limited to, vancomycin—has brought their role as important nosocomial pathogens into sharper focus. The accretion and spread of antibiotic resistance determinants among enterococci, to the point where some clinical isolates are resistant to all standard therapies, highlight both the vulnerability of our present armament as well as the looming prospect of a "postantibiotic era" (4). This review focuses on the magnitude and nature of the problem posed by enterococci in general, and MDR enterococci in particular. For many points, only representative citations are provided. Habitat and Microbiology Enterococci normally inhabit the bowel. They are found in the intestine of nearly all animals, from cockroaches to humans. Enterococci are readily recovered outdoors from vegetation and surface water, probably because of contamination by animal excrement or untreated sewage (5). In humans, typical concentrations of enterococci in stool are up to 10(sup 8) CFU per gram (6). Although the oral cavity and vaginal tract can become colonized, enterococci are recovered from these sites in fewer than 20% of cases. The predominant species inhabiting the intestine varies. In Europe, the United States, and the Far East, Enterococcus faecalis predominates in some instances and E. faecium in others (6). Ecologic or microbial factors promoting intestinal colonization are obscure. Of 14 or more enterococcal species (7), only E. faecalis and E. faecium commonly colonize and infect humans in detectable numbers. E. faecalis is isolated from approximately 80% of human infections, and E. faecium from most of the rest. Infections to other enterococcal species are rare. Enterococci are exceedingly hardy. They tolerate a wide variety of growth conditions, including temperatures of 10°C to 45°C, and hypotonic, hypertonic, acidic, or alkaline environments. Sodium azide and concentrated bile salts, which inhibit or kill most microorganisms, are tolerated by enterococci and used as selective agents in agar-based media. As facultative organisms, enterococci grow under reduced or oxygenated conditions. Enterococci are usually considered strict fermenters because they lack a Kreb's cycle and respiratory chain (8). E. faecalis is an exception since exogenous hemin can be used to produce d, b, and o type cytochromes (9,10). In a survey of 134 enterococci and related streptococci, only E. faecalis and Lactococcus lactis expressed cytochrome-like respiration (11). Cytochromes provide a growth advantage to E. faecalis during aerobic growth (9). E. faecalis cytochromes are only expressed under aerobic conditions in the presence of exogenous hemin (9,10,12) and, therefore, may promote the colonization of inappropriate sites. Enterococci are intrinsically resistant to many antibiotics. Unlike acquired resistance and virulence traits, which are usually transposon or plasmid encoded, intrinsic resistance is based in chromosomal genes, which typically are nontransferrable. Penicillin, ampicillin, piperacillin, imipenem, and vancomycin are among the few antibiotics that show consistent inhibitory, but not bactericidal, activity against E. faecalis. E. faecium are less susceptible to ß-lactam antibiotics than E. faecalis because the penicillin-binding proteins of the former have markedly lower affinities for the antibiotics (13). The first reports of strains highly resistant to penicillin began to appear in the 1980s (14,15). Enterococci often acquire antibiotic resistance through exchange of resistance-encoding genes carried on conjugative transposons, pheromone-responsive plasmids, and other broad-host-range plasmids (6). The past two decades have witnessed the rapid emergence of MDR enterococci. High-level gentamicin resistance occurred in 1979 (16) and was quickly followed by numerous reports of nosocomial infection in the 1980s (17). Simultaneously, sporadic outbreaks of nosocomial E. faecalis and E. faecium infection appeared with penicillin resistance due to ß-lactamase production (18); however, such isolates remain rare. Finally, MDR enterococci that had lost susceptibility to vancomycin were reported in Europe (19,20) and the United States (21). Among several phenotypes for vancomycin-resistant enterococci, VanA (resistance to vancomycin and teicoplanin) and VanB (resistance to vancomycin alone) are most common (22). In the United States, VanA and VanB account for approximately 60% and 40% of vancomycin-resistant enterococci (VRE) isolates, respectively (23). Inducible genes encoding these phenotypes alter cell wall synthesis and render strains resistant to glycopeptides (22). VanA and VanB types of resistance are primarily found among enterococci isolated from clinical, veterinary, and food specimens (24), but not other common intestinal or environmental bacteria. In the laboratory, however, these genes can be transferred from enterococci to other bacteria (22). For example, Staphylococcus aureus has been rendered vancomycin-resistant through apparent transfer of resistance from E. faecalis on the surface of membrane filters and on the skin of hairless obese mice (25), which indicates that there is no biologic barrier to the emergence of vancomycin-resistant S. aureus. Clinical isolates of highly vancomycin-resistant S. aureus have yet to be identified, although strains with reduced susceptibility to vancomycin have appeared (26). The mechanism of resistance for these strains remains undetermined but does not appear to involve genes associated with VanA or VanB phenotypes. Epidemiology Enterococci account for approximately 110,000 urinary tract infections, 25,000 cases of bacteremia, 40,000 wound infections, and 1,100 cases of endocarditis annually in the United States (2,27,28). Most infections occur in hospitals. Although several studies have suggested an increase in nosocomial infection rates for enterococci in recent years, National Nosocomial Infections Surveillance system data show little change in the percentage of enterococcal bloodstream (12% vs. 7%), surgical site (15% vs. 11%), and urinary tract (14% vs. 14%) infections over the past 2 decades (3,29). Adequate surveillance data prior to 1980 are not available. Enterococcal infection deaths have also been difficult to ascertain because severe comorbid illnesses are common; however, enterococcal sepsis is implicated in 7% to 50% of fatal cases (6). Several case-control and historical cohort studies show that death risk associated with antibiotic-resistant enterococcal bacteremia is severalfold higher than death risk associated with susceptible enterococcal bacteremia (30). This trend will likely increase as MDR isolates become more prevalent. Colonization and infection with MDR enterococci occur worldwide. Early reports showed that in the United States, the percentage of nosocomial infections caused by VRE increased more than 20-fold (from 0.3% to 7.9%) between 1989 and 1993, indicating rapid dissemination. New database technologies, such as The Surveillance Network (TSN) Database-USA, now permit the assessment of resistance profiles according to species. TSN Database electronically collects and compiles data daily from more than 100 U.S. clinical laboratories, identifies potential laboratory testing errors, and detects emergence of resistance profiles and mechanisms that pose a public health threat (e.g., vancomycin-resistant staphylococci). Data collected by the TSN Database between 1995 and September 1, 1997 were analyzed to determine whether the earlier increase in vancomycin resistance was unique to vancomycin, whether it represented a continuing trend, and whether speciation is quantifiably important in analyzing this trend. E. faecalis resistance to ampicillin and vancomycin is uncommon Figure 1); little change in resistance prevalence occurred from 1995 to 1997. In contrast, E. faecium vancomycin and ampicillin resistance increased alarmingly. In 1997, 771 (52%) of 1,482 of E. faecium isolates exhibited vancomycin resistance, and 1,220 (83%) of 1,474 isolates exhibited ampicillin resistance (Figure 1). E. faecium resistance notwithstanding, E. faecalis remained by far the most commonly encountered of the two enterococcal species in TSN Database. E. faecalis to E. faecium total isolates were approximately 4:1 (Figure 1), blood isolates 3:1, and urine isolates 5:1. This observation underscores important differences in the survival strategies and likelihood of therapeutic success, critical factors usually obscured by lumping the organisms together as Enterococcus species or enterococci. Widespread emergence and dissemination of ampicillin and vancomycin resistance in E. faecalis would significantly confound the current therapeutic dilemma. There is little reason to suspect that vancomycin and ampicillin resistances only provide selective advantage for the species faecium and not faecalis. The relative absence of these resistances in E. faecalis may simply reflect a momentary lack of penetrance and equilibration of the traits. Because of these important differences between the two species, meaningful surveillance of enterococcal resistance must include species identification. [fig 1] Figure 1. Epidemiology of enterococcal infection based on 15,203 susceptibility results obtained by The Surveillance Network (TSN) Database-USA, 1995 to Sep 1, 1997. The increase in total numbers between 1995 and 1996 represents additional reporting centers coming on line. Numbers for 1997 represent total collected for the partial year to Sep 1, 1997. Although exact modes of nosocomial transmission for MDR enterococci are difficult to prove, molecular microbiologic and epidemiologic evidence strongly suggest spread between patients, probably on the hands of health-care providers or medical devices, and between hospitals by patients with prolonged intestinal colonization. At least 16 outbreaks of MDR enterococci have been reported since 1989 (31); all but two were due to E. faecium. This disparity, particularly in view of the higher numbers of clinical E. faecalis isolates, may reflect a reporting bias due to the novelty of the combinations of resistance that occur in E. faecium. When isolates from outbreaks of MDR enterococci have been analyzed by genetic fingerprints, more than half involve clonally related isolates (18,32). Prior treatment with antibiotics is common in nearly all patients colonized or infected with MDR enterococci (33-35). Clindamycin, cephalosporin, aztreonam, ciprofloxacin, aminoglycoside, and metronidazole use is equally or more often associated with colonization or infection with MDR enterococci than vancomycin use. Other risk factors include prolonged hospitalization; high severity of illness score; intraabdominal surgery; renal insufficiency; enteral tube feedings; and exposure to specific hospital units, nurses, or contaminated objects and surfaces within patient-care areas. Infection Control Controlling the spread of MDR enterococci among inpatients is difficult. We know relatively little about the biology of enterococcal transmission or the specific microbial factors favoring colonization by exogenous enterococcal strains. Nevertheless, VRE infection control policies, which could apply to MDR enterococci, were recently published by the Hospital Infection Control Practices Advisory Committee (36). Control methods include routine screening for vancomycin resistance among clinical isolates, active surveillance for VRE in intensive care units, contact isolation to minimize person-to-person transmission, and vancomycin restriction. These measures to limit VRE spread, however, have failed on occasion (35). Not all hospitals can or are willing to perform active surveillance. Because more patients are typically colonized with VRE (3% to 47%) than are infected (35,37,38), and because intestinal colonization can be prolonged, passive surveillance by routine cultures allows colonized inpatients to go unidentified and serve as point sources for continued spread of VRE. Even if all colonized inpatients are successfully identified, VRE may be spread by health-care workers through either inadequate hand washing (39) or through contact with items such as bedrails, sinks, faucets, and doorknobs (enterococci can persist for weeks on environmental surfaces) (40). Decontamination efforts must be rigorous. The Hospital Infection Control Practices Advisory Committee strongly recommended restricting oral and parenteral vancomycin to control VRE (36). However, limiting use of vancomycin while ignoring widespread use of other broad spectrum antibiotics likely will not lead to maximal control of VRE or of MDR enterococci. Antibiotics may promote colonization and infection with MDR enterococci by at least two mechanisms. First, many broad spectrum antibiotics have little or no anti-enterococcal activity, and administration commonly leads to overgrowth of susceptible (or resistant) enterococci at sites at risk for infection. Second, most antibiotics substantially reduce the normal resistance of the intestinal tract to colonization by exogenous organisms (41). Colonization resistance results primarily from the "limiting action" of the normal anaerobic flora, and to a lesser extent from an intact mucosa, gastric acid secretion, intestinal motility, and intestinal-associated immunity (41). Antibiotic-induced alterations in the protective flora of the intestine provide large footholds for colonization with exogenous pathogens such as MDR enterococci (41). Antibiotic restriction programs would be more effective if they included prudent prescribing of all antibiotics, not just single agents such as vancomycin. This approach substantially decreased intestinal colonization with VRE in one hospital pharmacy that restricted vancomycin, cefotaxime, and clindamycin (42). At a minimum, a successful program for control of MDR enterococci requires effective passive and active surveillance to identify colonized and infected patients, absolute adherence to contact isolation by health-care workers, rigorous decontamination of patient-contact areas and judicious use or restriction of vancomycin and other broad spectrum antibiotics. Therapeutic Approaches Suitable antibiotics often are not available to treat MDR enterococcal infections, e.g., endocarditis or bacteremia, in the presence of neutropenia. Combinations of penicillin with vancomycin, ciprofloxacin with ampicillin, or novobiocin with doxycycline, among others, have been used (43) but can be unpredictable and remain clinically unproven. In one report chloramphenicol successfully treated chloramphenicol-susceptible infections (44), but these findings await confirmation in controlled trials. Promising new antibiotics for MDR enterococcal infection under investigation include fluoroquinolones, streptogramins, oxazolidinones, semisynthetic glycopeptides, and glycylcyclines. Clinafloxacin, a fluoroquinolone with improved potency against enterococci compared with ciprofloxacin, has excellent activity against VRE, appears bactericidal in vitro, and has been effective in treatment of enterococcal infections in a murine model (45). Although single-step resistance to clinafloxacin could not be detected in vitro, multistep resistance is readily achieved. Should this agent gain approval for treatment of enterococcal infections, selection for resistance may limit its effectiveness. Quinupristin/dalfopristin (Synercid) is a combination of streptogramins A and B that inhibits protein synthesis and has a narrower spectrum of activity against enterococci than clinafloxacin (46). Many, but not all, E. faecium isolates with VanA and VanB phenotypes are susceptible (47); however, E. faecalis is uniformly resistant, and superinfection has been reported during therapy (48). In addition, quinupristin/dalfopristin is bacteriostatic only, potentially allowing emergence of resistance (49). For these reasons the drug may have only a limited role in treating MDR enterococcal infections. Novel oxazolidinones and glycylcyclines have also shown potent activity against enterococci, including MDR enterococci (50,51), but await further testing. The substantial drawback of the broad spectrum approach is that the more organisms affected (both protective commensals as well as pathogens), the more opportunities for resistance to evolve. Broad spectrum antibiotics permit empiric therapy in the absence of a specific diagnosis and generate a more substantial return on investment in the short term. However, broad spectrum antibiotics affect not only disease-causing organisms but also commensals present in numbers large enough to generate resistance by otherwise rare mutations or genetic exchange events. As long as market forces favor development of broad spectrum therapeutics, a cycle of drug introduction followed by emergence of resistance undoubtedly will continue. Targeted Therapeutics In contrast to the historical reliance on broad spectrum antibiotic therapy, the continuing development and introduction of rapid diagnostic techniques (52) may allow a more focused approach to infectious disease therapy. Any of the myriad microbial-host interactions that subvert the host response or damage tissues during an infection represent potential therapeutic targets. However, many key interactions in disease pathogenesis are specific to the organism involved—a characteristic that is both a strength and a weakness. Because of the specificity of these interactions, rapid and accurate diagnosis is required. However, therapeutics aimed only at interaction between host and a specific pathogen should leave the diverse commensal flora essentially unaffected. As a result, the targeted population would be restricted to the relatively small numbers of disease-producing bacteria and would not likely reach the numbers or diversity required to make development of resistance a statistical probability. The current spectrum of approaches to identify new antiinfective compounds has two extremes: 1) screening vast libraries of compounds to identify substances that by chance inhibit a microbial property and 2) detailed study of interactions between host and parasite to identify critical events leading to host tissue damage or compromise (53). With a long-term view toward new therapeutic approaches as well as optimal use of existing therapies, we and others have begun examining in detail the interactions between enterococci and host (6). A major obstacle is that enterococci also form part of the commensal or autochthonous flora; as such, they are nearly devoid of traits traditionally associated with overt pathogens and have subtle interactions with the host. Using inocula with as few as 10 organisms, we have developed sensitive biologic systems for examining the host-parasite interactions (54). Although E. faecium strains are resistant to vancomycin and ampicillin more often than E. faecalis strains, the relative proportion of infections caused by these species has not dramatically changed in recent years (Figure 1). Since both organisms are frequently isolated from the commensal flora, this bias suggests that E. faecalis traits confer a greater degree of intrinsic virulence, for example, cytolysin production, pheromone-responsive plasmid transfer (and accompanying production of aggregation substance), extracellular superoxide production, and a newly identified surface protein tentatively termed Esp (5,56,57) (Figure 2). These properties provide logical points of departure for developing new targeted therapeutic approaches to enterococcal disease; examination of more subtle interactions between E. faecium and host will follow as an understanding of enterococcal biology evolves. [fig 2] Figure 2. Virulence traits and their association with enterococcal species. Targeting the E. faecalis Cytolysin Cytolysin is disproportionately expressed by E. faecalis strains associated with disease (5,55,56). This cytolysin causes rupture of a variety of target membranes, including bacterial cells, erythrocytes, and other mammalian cells, with activity observed as a hemolytic zone on some types of blood agar. Cytolysin contributes to the toxicity or lethality of infection in several infection models and is associated with a fivefold increased risk of sudden death from nosocomial bacteremia (54,56-59). Cytolysin also contributes to the appearance of enterococci in a murine bacteremia model (Figure 3; 45,60), an observation consistent with the disproportionate representation of cytolytic strains among human blood isolates (56,62). [fig 3] Figure 3. Cytolysin favors the appearance of circulating enterococci. In this experiment, 10(sup)7 CFU of E. faecalis, either cytolytic FA2-2(pAM714) (60) or noncytolytic FA2-2(pAM771) (64), were intraperitoneally injected (45) into groups of five BalbC mice. Viable bacteria in liver, spleen, and the bloodstream were enumerated 48 hrs following injection, and significance assessed by Student's t-test. (P. Coburn, L.E. Hancock, and M.S. Gilmore, in preparation). Beginning with E.W. Todd in 1934 (63) and culminating in a recent study (64), the E. faecalis cytolysin has now been characterized as a unique, extensively modified bacterial toxin (Figure 4). The cytolysin maturation pathway is ideally designed for therapeutic targeting because the two structural subunits are activated by an extracellular protease, an event that is accessible and potentially inhibitable by a novel therapeutic. Moreover, the activator protease, CylA, belongs to the subtilisin class of serine proteases (64), whose structure-function relationships and inhibitor design we are beginning to understand. Investigations are in progress to design and test inhibitors of extracellular cytolysin activation to determine whether a reduction by several logs in the levels of circulating enterococci can be attained as would be predicted by the observed behavior of cytolysin mutants (Figure 3). [fig 4] Figure 4. Cytolysin is expressed and processed through a complex maturation pathway (64). The cytolysin precursors, CylL(sub L) and CylL(sub S), are ribosomally synthesized. The putative modification protein, CylM, is required for the expression of CylL(sub L) and CylL(sub S) in an activatable form, and the secreted forms, CylL(sub L) and CylL(sub S) were recently shown to possess the amino acid lanthionine as the result of posttranslational modification (64). CylL(sub L) and CylL(sub S) both are secreted by CylB (65), which is accompanied by an initial proteolytic trimming event (64) converting each to CylL(sub L)' and CylL(sub S)', respectively. Once secreted, CylL(sub L)' and CylL(sub S)' are both functionally inactive until six amino acids are removed from each amino terminus. This final step in maturation is catalyzed by CylA (64), a subtilisin-type serine protease. Since this final catalytic event is essential, occurs extracellularly, and is catalyzed by a class of enzyme for which a substantial body of structural information exists, it represents an ideal therapeutic target. As shown in Figure 3, inhibition of cytolysin by mutation (or potentially by therapeutic intervention) results in a reduction by several orders of magnitude in the number of circulating organisms. An inhibitor of cytolysin activation, accompanied by appropriate rapid diagnostics, would be of potential value in treating bacteremia caused by cytolytic strains of E. faecalis without affecting commensal flora. Development of resistance should be exceedingly improbable because of the small number of bacteria targeted and because unlike antibiotics, cytolysin inhibitors would not act directly on the organism itself. Other Enterococcal Targets Several laboratories are using information on the E. faecalis genome and genomes of other pathogens to identify therapeutic targets (66) and facilitate studies on pathogenesis for the remaining 60% of noncytolytic enterococcal infections. The genome of an E. faecalis strain that caused multiple hospital infections (56) was sampled at high frequency by sequence analysis. Several sequences appeared to have a role in host-parasite interaction. The gene specifying Esp encodes an apparent surface protein of unusual repeating structure (67). Although a role for this protein in enterococcal infection has yet to be determined, its distribution among clinical and commensal strains is tantalizing: 29 of 30 strains with this gene were recovered from patients with bacteremia or endocarditis; one of 34 isolates obtained from healthy volunteers possessed Esp. The core of this large protein (inferred mass of 202 kDa) consists of a series of 82 amino acid repeats encoded by highly conserved tandem 246 base pair repeats. Lack of divergence in repeat sequences suggests that recombination occurs at high frequency, perhaps during infection. Moreover, the number of repeats observed in homologous genes from different E. faecalis isolates is 3 to 9 (67). This gene is flanked by a sequence similar to the transposase of IS905. None of 24 clinical or laboratory E. faecium isolates had this gene (67; V. Shankar, G. Lindahl, and M. Gilmore, unpub. data). A second promising lead involves a series of genes encoding products highly related to enzymes involved in O-antigen synthesis in gram-negative bacteria (68). Preliminary evidence suggests that in E. faecalis these genes contribute to cell wall carbohydrate synthesis and that this carbohydrate relates to persistence in vivo. A knockout in one of these genes results in a strain with normal in vitro growth, but after subcutaneous injection, the mutant was more readily cleared than the wild type parental strain (68). One of the genes studied was present in all E. faecalis strains examined, whereas another occurs only in E. faecalis strains that share a periodate-susceptible epitope (68). Collectively, these data indicate that enzymes for synthesis of E. faecalis surface carbohydrates are important for persistence in vivo and may represent a useful therapeutic target. Taking a different approach, Arduino et al. (69,70) identified a protease-resistant, periodate susceptible substance associated with some strains of E. faecium, but not E. faecalis, which conferred resistane to phagocytosis in vitro. The relationship between the putative carbohydrate of E. faecalis under study above and the inhibitory substance of E. faecium remains to be determined. It may be found that many enterococci produce such carbohydrates at biologically significant levels in vivo, but only some strains of E. faecium do so in vitro. Finally, recent observations indicate that nearly all E. faecalis strains, and only a few E. faecium strains, generate substantial extracellular superoxide. When E. faecalis isolates from patients with endocarditis and bacteremia were compared with isolates from healthy volunteers (71), on average, extracellular superoxide production was 60% higher among blood isolates than commensal strains. These data raised several questions: Do E. faecalis that produce larger amounts of extracellular superoxide possess greater metabolic flexibility, facilitating adaptation to nonintestinal infection sites? Does free radical production lead to host cell damage, allowing release of normally sequestered nutrients (e.g., hemin) that might promote enhanced E. faecalis growth through cytochrome formation? Might antioxidants modulate colonization or invasive infection? Answers to these questions may provide new insights into the transition from intestinal colonization to infection and may suggest new preventive strategies. Obstacles to Further Development Although important insights into enterococcal biology and pathogenesis are being gleaned from a reverse genetic approach, a paucity of information still exists on how enterococci colonize the intestinal tract and cause infection. For example, do E. faecalis or E. faecium colonize the colon through specific interactions with ligands on human epithelial cells or intestinal mucin? Do MDR enterococci possess alternate binding activities that enable them to colonize the intestinal tract at new sites without competing with the indigenous enterococci? Do probiotics have a role in restoring colonization resistance to an intestinal ecology altered by broad spectrum antibiotics? Is enough being done to combat the emergence of highly resistant nosocomial pathogens? To effectively compete, industry remains highly responsive to market opportunities. Research in the public sector has been slow to respond, and as a result, our understanding of the biology of enterococcal infection is inadequate. Reasons for the modest public sector response include the following. 1) The emergence of resistant enterococci coincided with a reduction of public support for non-AIDS related infectious disease research. 2) The pathogenesis of nosocomial infection deviates from paradigms established for obligate pathogens. 3) The research infrastructure is relatively small because of the low importance traditionally attached to enterococci as etiologic agents of human disease and the deemphasis on antibiotic resistance research in the 1980s. Conclusions Historically, substantial resources have been invested in developing an in-depth understanding of the molecular biology of model organisms. During the 1960s and 1970s, when gram-negative organisms were leading causes of hospital- and community-acquired infections and gram-positive organisms were typically sensitive to existing antibiotics (72), a sizable investment in gram-negative model organisms was appropriate. However, with the emergence of gram-positive organisms as leading causes of both hospital- and community-acquired infection in the 1990s, a reevaluation of public research priorities is warranted. Since antibiotic use became widespread 50 years ago, bacteria have steadily and routinely developed resistance. Control of the emergence of resistance will depend on new approaches to prudent antibiotic use in hospitals and clinics, based in part on improved surveillance for MDR enterococci and on better systems to encourage staff adherence to contact isolation procedures. Equally important will be development of new drugs with narrower spectra of activity aimed at known and potentially new targets and the evolution of market conditions that favor their use. Portions of the work described were supported by Veterans Administration Merit Review Program, grants from the Public Health Service (EY08289 and AI41108), and an unrestricted award from Research to Prevent Blindness, Inc. Dr. Huycke is an associate professor in the Infectious Diseases Section, Department of Medicine, Oklahoma University Health Sciences Center. He is interested in enterococcal pathogenesis as it relates to extracellular superoxide production by E. faecalis. Address for correspondence: Michael Gilmore, Department of Microbiology and Immunology, University of Oklahoma Health Sciences Center, PO Box 26901, Oklahoma City, OK 73190, USA; fax: 405-271-8128; e-mail: mgilmore@aardvark.ouknor.edu. References 1. Kaye D. Enterococci: biologic and epidemiologic characteristics and in vitro susceptibility. Arch Intern Med 1982;142:2006-9. 2. Emori TG, Gaynes RP. An overview of nosocomial infections, including the role of the microbiology laboratory. Clin Microbiol Rev 1993;6:428-42. 3. Jarvis WR, Gaynes RP, Horan TC, Emori TG, Stroud LA, Archibald LK, et al. Semiannual report: aggregated data from the National Nosocomial Infections Surveillance (NNIS) system. CDC, 1996:1-27. 4. Cohen ML. Epidemiology of drug resistance: implications for a post-antimicrobial era. Science 1992;257:1050-5. 5. Jett BD, Huycke MM, Gilmore MS. Virulence of enterococci. Clin Microbiol Rev 1994;7:462-78. 6. Rice EW, Messer JW, Johnson CH, Reasoner DJ. Occurrence of high-level aminoglycoside resistance in environmental isolates of enterococci. Appl Environ Microbiol 1995;61:374-6. 7. Devriese LA, Pot B, Collins MD. Phenotypic identification of the genus Enterococcus and differentiation of phylogenetically distinct enterococcal species and species groups. J Appl Bacteriol 1993;75:399-408. 8. Willett HP. Energy metabolism. In: Joklik WK, Willett HP, Amos DB, Wilfert CM, editors. Zinsser microbiology. 20th ed. East Norwalk (CT): Appleton & Lange; 1992. p. 53-75. 9. Ritchey TW, Seeley HW. Cytochromes in Streptococcus faecalis var. zymogenes grown in a haematin-containing medium. J Gen Microbiol 1974;85:220-8. 10. Pritchard GG, Wimpenny JWT. Cytochrome formation, oxygen-induced proton extrusion and respiratory activity in Streptococcus faecalis var. zymogenes grown in the presence of haematin. J Gen Microbiol 1978;104:15-22. 11. Ritchey TW, Seeley HW Jr. Distribution of cytochrome-like respiration in streptococci. J Gen Microbiol 1976;93:195-203. 12. Bryan-Jones DG, Whittenbury R. Haematin-dependent oxidative phosphorylation in Streptococcus faecalis. J Gen Microbiol 1969;58:247-60. 13. Williamson R, Le Bougu,nec C, Gutmann L, Horaud T. One or two low affinity penicillin-binding proteins may be responsible for the range of susceptibility of Enterococcus faecium to benzylpenicillin. J Gen Microbiol 1985;131:1933-40. 14. Bush LM, Calmon J, Cherney CL, Wendeler M, Pitsakis P, Poupard J, Levison ME, Johnson CC. High-level penicillin resistance among isolates of enterococci: implications for treatment of enterococcal infections. Ann Intern Med 1989;110:515-20. 15. Sapico FL, Canawati HN, Ginunas VJ, Gilmore DS, Montgomerie JZ, Tuddenham WJ, et al. Enterococci highly resistant to penicillin and ampicillin: an emerging clinical problem? J Clin Microbiol 1989;27:2091-5. 16. Horodniceanu T, Bougueleret L, El-Solh N, Bieth G, Delbos F. High-level, plasmid-borne resistance to gentamicin in Streptococcus faecalis subsp zymogenes. Antimicrob Agents Chemother 1979;16:686-9. 17. Zervos MJ, Kauffman CA, Therasse PM, Bergman AG, Mikesell TS, Schaberg DR. Nosocomial infection by gentamicin-resistant Streptococcus faecalis: an epidemiologic study. Ann Intern Med 1987;106:687-91. 18. Murray BE, Singh KV, Markowitz SM, Lopardo HA, Patterson JE, Zervos MJ, et al. Evidence for clonal spread of a single strain of ß-lactamase-producing Enterococcus (Streptococcus) faecalis to six hospitals in five states. J Infect Dis 1991;163:780-5. 19. Uttley AHC, Collins CH, Naidoo J, George RC. Vancomycin-resistant enterococci. Lancet 1988;1:57-8. 20. Leclercq R, Derlot E, Duval J, Courvalin P. Plasmid-mediated resistance to vancomycin and teicoplanin in Enterococcus faecium. N Engl J Med 1988;319:157-61. 21. Sahm DF, Kissinger J, Gilmore MS, Murray PR, Mulder R, Solliday J, et al. In vitro susceptibility studies of vancomycin-resistant Enterococcus faecalis. Antimicrob Agents Chemother1989;33:1588-91. 22. Arthur M, Courvalin P. Genetics and mechanisms of glycopeptide resistance in enterococci. Antimicrob Agents Chemother 1993;37:1563-71. 23. Clark NC, Cooksey RC, Hill BC, Swenson JM, Tenover FC. Characterization of glycopeptide-resistant enterococci from U.S. hospitals. Antimicrob Agents Chemother 1993;37:2311-7. 24. Klare I, Heier H, Claus H, Reissbrodt R, Witte W. VanA-mediated high-level glycopeptide resistance in Enterococcus faecium from animal husbandry. FEMS Microbiol Lett 1995;125:165-72. 25. Noble WC, Virani Z, Cree RGA. Co-transfer of vancomycin and other resistance genes from Enterococcus faecalis NCTC 12201 to Staphylococcus aureus. FEMS Microbiol Lett 1992;93:195-8. 26. Hiramatsu K, Hanaki H, Ino T, Yabuta K, Oguri T, Tenover FC. Methicillin-resistant Staphylococcus aureus clinical strain with reduced vancomycin susceptibility. J Antimicrob Chemother 1997;40:135-146. 27. Haley RW, Culver DH, White JW, Meade WM, Emori TG, Munn VP, et al. The efficacy of infection surveillance and control programs in preventing nosocomial infections in US hospitals. Am J Epidemiol 1985;121:182-205. 28. Harris SL. Definitions and demographic characteristics. In: Kaye D, editor. Infective endocarditis. New York: Raven Press, Ltd.; 1992. p. 1-18. 29. Hughes JM, Culver DH, W, Morgan WM, Munn VP, Mosser JL, Emori TG. Nosocomial infection surveillance, 1980-1982. MMWR Morb Mortal Wkly Rep 1983;32:1SS-16SS. 30. Edmond MB, Ober JF, Dawson JD, Weinbaum DL, Wenzel RP. Vancomycin-resistant enterococcal bacteremia: natural history and attributable mortality. Clin Infect Dis 1996;23:1234-9. 31. Rhinehart E, Smith NE, Wennersten C, Gorss E, Freeman J, Eliopoulos GM, et al. Rapid dissemination of ß-lactamase-producing, aminoglycoside-resistant Enterococcus faecalis among patients and staff on an infant-toddler surgical ward. N Engl J Med 1990;26:1814-8. 32. Chow JW, Kuritza A, Shlaes DM, Green M, Sahm DF, Zervos MJ. Clonal spread of vancomycin-resistant Enterococcus faecium between patients in three hospitals in two states. J Clin Microbiol 1993;31:1609-11. 33. Montecalvo MA, Horowitz H, Gedris C, Carbonaro C, Tenover FC, Issah A, et al. Outbreak of vancomycin-, ampicillin-, and aminoglycoside-resistant Enterococcus faecium bacteremia in an adult oncology unit. Antimicrob Agents Chemother 1994;38:1363-7. 34. Livornese LL, Dias S, Samel C, Romanowski B, Taylor S, May P, et al. Hospital-acquired infection with vancomycin-resistant Enterococcus faecium transmitted by electronic thermometers. Ann Intern Med 1992;117:112-6. 35. Handwerger S, Raucher B, Altarac D, Monka J, Marchione S, Singh KV, et al. Nosocomial outbreak due to Enterococcus faecium highly resistant to vancomycin, penicillin, and gentamicin. Clin Infect Dis 1993;16:750-5. 36. Centers for Disease Control and Prevention. Recommendations for preventing the spread of vancomycin resistance: recommendations of the Hospital Infection Control Practices Advisory Committee (HICPAC). MMWR Morb Mortal Wkly Rep 1995;44(No. RR-12):1-13. 37. Morris JG Jr, Shay DK, Hebden JN, McCarter RJ Jr, Perdue BE, Jarvis W, et al. Enterococci resistant to multiple antimicrobial agents, including vancomycin. Ann Intern Med 1995;123:250-9. 38. Edmond MB, Ober JF, Weinbaum DL, Pfaller MA, Hwang T, Sanford MD, et al. Vancomycin-resistant Enterococcus faecium bacteremia: risk factors for infection. Clin Infect Dis 1995;20:1126-33. 39. Goldman D, Larson E. Hand-washing and nosocomial infections. N Engl J Med 1992;327:120-2. 40. Noskin GA, Stosor V, Cooper I, Peterson LR. Recovery of vancomycin-resistant enterococci on fingertips and environmental surfaces. Infect Control Hosp Epidemiol 1995;16:577-81. 41. Vollaard EJ, Clasener HAL. Colonization resistance. Antimicrob Agents Chemother 1994;38:409-14. 42. Quale J, Landman D, Saurina G, Atwood E, DiTore V, Patel K. Manipulation of a hospital antimicrobial formulary to control an outbreak of vancomycin-resistant enterococci. Clin Infect Dis 1996;23:1020-5. 43. Caron F, Pestel M, Kitzis M-D, Lemeland JF, Humbert G, Gutmann L. Comparison of different ß-lactam-glycopeptide-gentamicin combinations for an experimental endocarditis caused by a highly ß-lactam-resistant and highly glycopeptide-resistant isolate of Enterococcus faecium. J Infect Dis 1995;171:106-12. 44. Norris AH, Reilly JP, Edelstein PH, Brennan PJ, Schuster MG. Chloramphenicol for the treatment of vancomycin-resistant enterococcal infections. Clin Infect Dis 1995;20:1137-44. 45. Cohen MA, Yoder SL, Huband MD, Roland GE, Courtney CL. In vitro and in vivo activities of clinafloxacin, CI-990 (PD 131112), and PD 138312 versus enterococci. Antimicrob Agents Chemother 1995;39:2123-7. 46. Aumercier M, Bouhallab S, Capmau M-L, LeGoffic F. RP 59500: A proposed mechanism for its bactericidal activity. J Antimicrob Chemother 1992;30(Suppl A):9-14. 47. Collins LA, Malanoski GJ, Eliopoulos GM, Wennersten CB, Ferraro MJ, Moellering RC Jr. In vitro activity of RP59500, an injectable streptogramin antibiotic, against vancomycin-resistant gram-positive organisms. Antimicrob Agents Chemother 1993;37:598-601. 48. Chow JW, Davidson A, Sanford III E, Zervos MJ. Superinfection with Enterococcus faecalis during quinupristin/dalfopristin therapy. Clin Infect Dis 1997;24:91-2. 49. Chow JW, Donahedian SM, Zervos MJ. Emergence of increased resistance to quinupristin/dalfopristin during therapy for Enterococcus faecium bacteremia. Clin Infect Dis 1997;24:90-1. 50. Eliopoulos GM, Wennersten CB, Cole G, Moellering RC. In vitro activities of two glycylcyclines against gram-positive bacteria. Antimicrob Agents Chemother 1994;38:534-41. 51. Jones RN, Johnson DM, Erwin ME. In vitro antimicrobial activities and spectra of U-100592 and U-100766, two novel fluorinated oxazolidinones. Antimicrob Agents Chemother 1996;40:720-6. 52. Hillyard DR. The molecular approach to microbial diagnosis. Am J Clin Pathol 1994;101:S18-S21. 53. Strohl WR. Biotechnology of Antibiotics. 2nd ed: Drugs and the Pharmaceutical Sciences 82, 1997. 54. Jett BD, Jensen HG, Nordquist RE, Gilmore MS. Contribution of the pAD1-encoded cytolysin to the severity of experimental Enterococcus faecalis endophthalmitis. Infect Immun 1992;60:2445-52. 55. Ike Y, Hashimoto H, Clewell DB. High incidence of hemolysin production by Enterococcus (Streptococcus) faecalis strains associated with human parenteral infections. J Clin Microbiol 1987;25:1524-8. 56. Huycke MM, Spiegel CA, Gilmore MS. Bacteremia caused by hemolytic, high-level gentamicin-resistant Enterococcus faecalis. Antimicrob Agents Chemother 1991;35:1626-34. 57. Jett BD, Jensen HG, Atkuri R, Gilmore MS. Evaluation of therapeutic measures for treating endophthalmitis cause by isogenic toxin producing and toxin non-producing Enterococcus faecalis strains. Invest Ophthalmol Vis Sci 1995;36:9-15. 58. Ike Y, Hashimoto H, Clewell DB. Hemolysin of Streptococcus faecalis subspecies zymogenes contributes to virulence in mice. Infect Immun 1984;45:528-30. 59. Chow JW, Thal LA, Perri MB, Vazquez JA, Donabedian SM, Clewell DB, et al. Plasmid-associated hemolysin and aggregation substance production contributes to virulence in experimental enterococcal endocarditis. Antimicrob Agents Chemother 1993;37:2474-7. 60. Ike Y, Clewell DB. Genetic analysis of pAD1 pheromone response in Streptococcus faecalis using transposon Tn917 as an insertional mutagen. J Bacteriol 1984;158:777-83. 61. Ike Y, Clewell DB, Segarra RA, Gilmore MS. Genetic analysis of the pAD1 hemolysin/bacteriocin determinant in Enterococcus faecalis: Tn917 insertional mutagenesis and cloning. J Bacteriol 1990;172:155-63. 62. Huycke MM, Gilmore MS. Frequency of aggregation substance and cytolysin genes among enterococcal endocarditis isolates. Plasmid 1995;34:152-6. 63. Todd EW. A comparative serological study of streptolysins derived from human and from animal infections, with notes on pneumococcal haemolysin, tetanolysin and staphylococcus toxin. Journal of Pathol ogy and Bacteriology 1934;39:299-321. 64. Booth MC, Bogie CP, Sahl H-G, Siezen RJ, Hatter KL, Gilmore MS. Structural analysis and proteolytic activation of Enterococcus faecalis cytolysin, a novel lantibiotic. Mol Microbiol 1996;21:1175-84. 65. Gilmore MS, Segarra RA, Booth MC. An hlyB-type function is required for expression of the Enterococcus faecalis hemolysin/bacteriocin. Infect Immun 1990;58:3914-23. 66. Katz L, Chu DT, Reich K. Bacterial genomics and the search for novel antibiotics. In: Plattner JJ, editor. Annual Reports in Medicinal Chemistry. Vol. 32. New York: Academic Press, Inc., 1997. p. 121-30. 67. Shankar V, Gilmore MS. Structure and expression of a novel surface protein of Enterococcus faecalis. In: Abstracts of the 97th General Meeting of the American Society for Microbioloty; 4-8 May 1997;Miami Beach, Florida. Washington: The Society; 1997. 68. Hancock LE, Gilmore MS. The contribution of a cell wall associated carbohydrate to the in vivo survival of Enterococcus faecalis in a murine model of infection. In: Abstracts of the 97th General Meeting of the American Society for Microbioloty. 4-8 May 1997;Miami Beach, Florida. Washington: The Society; 1997. 69. Arduino RC, Murray BE, Rakita RM. Roles of antibodies and complement in phagocytic killing of enterococci. Infect Immun 1994;62:987-93. 70. Arduino RC, Palaz-Jacques K, Murray BE, Rakita RM. Resistance of Enterococcus faecium to neutrophil-mediated phagocytosis. Infect Immun 1994;62:5587-94. 71. Huycke MM, Joyce W, Wack MF. Augmented production of extracellular superoxide production by blood isolates of Enterococcus faecalis. J Infect Dis 1996;173:743-6 72. Schwartz MN. Hospital-acquired infections: diseases with increasingly limited therapies. Proc Natl Acad Sci U S A 1994;91:2420-7. Emerging Infectious Diseases National Center for Infectious Diseases Centers for Disease Control and Prevention Atlanta, GA URL: http://www.cdc.gov/ncidod/EID/vol4no2/huycke.htm Please note that figures and equations are not available in ASCII format; their placement within the text is noted by [fig] and [eq], respectively. Greek symbols are spelled out. The following codes are used: (ft) for footnote; (sup) for superscript; (sub) for subscript; /= for greater than or equal to. ------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------- [Emerging Infectious Diseases] [Volume 4 No. 2 / April - June 1998] Synopses Enteroaggregative Escherichia coli James P. Nataro,* Theodore Steiner,† and Richard L. Guerrant† *University of Maryland School of Medicine, Baltimore, Maryland, USA; and †University of Virginia School of Medicine, Charlottesville, Virginia, USA -------------------------------------------------- Enteroaggregative Escherichia coli (EAEC), an increasingly recognized cause of diarrhea in children in developing countries, has been particularly associated with persistent diarrhea (more than 14 days), a major cause of illness and death. Recent outbreaks implicate EAEC as a cause of foodborne illness in industrialized countries. The pathogenesis of EAEC infection is not well understood, but a model can be proposed in which EAEC adhere to the intestinal mucosa and elaborate enterotoxins and cytotoxins, which result in secretory diarrhea and mucosal damage. EAEC's ability to stimulate the release of inflammatory mediators may also play a role in intestinal illness. Since first described in 1987, enteroaggregative Escherichia coli (EAEC) have been recognized increasingly as agents of diarrhea in developing countries and (more recently) in industrialized countries. We review the data supporting the pathogenicity of EAEC, discuss clinical and epidemiologic features, and summarize the current status of EAEC pathogenesis research. History E. coli have been implicated as agents of diarrheal disease since the 1920s (1). In 1979, Cravioto et al. reported that most enteropathogenic E. coli (EPEC) adhered to HEp-2 cells in culture and that the adherence phenotype could be used to differentiate EPEC. Subsequently, many E. coli not of EPEC serogroups were also found to adhere to semiconfluent HEp-2 cells, but the adherence phenotype was clearly different from that induced by EPEC (2,3). Whereas the adherence pattern of EPEC was localized, the non-EPEC pattern was diffuse. The diffuse adherent strains were later subdivided into two further phenotypic subcategories: aggregative and true "diffuse" (4). Only the aggregative were associated with pediatric diarrhea in Santiago, Chile (4), prompting the investigators to propose two independent categories: EAEC and diffusely adherent E. coli. At the same time, non-EPEC, HEp-2 adherent strains (termed enteroadherent E. coli) were associated with pediatric and traveler's diarrhea in Mexico (5,6). These enteroadherent strains were later found to belong to the EAEC or diffusely adherent E. coli categories (7). Thus, EAEC are now defined as E. coli that do not secrete heat-labile or heat-stable enterotoxins and adhere to HEp-2 cells in an aggregative (AA) pattern (Figure 1). This definition may encompass both pathogenic and nonpathogenic clones that share a factor(s) conferring a common phenotype. [fig 1] Figure 1. Aggregative pattern of adherence in the HEp-2 assay after 3 hours incubation of bacteria with HEp-2 cells. Note bacteria on the surface of the HEp-2 cells as well as on the glass substratum. Pathogenesis Histopathology EAEC strains adhere to the mucosal epithelium of gnotobiotic piglets in a thick mucus-containing blanket, irrespective of whether or not the strain induces diarrhea in this animal (8; Nataro, unpub. obs.; Figure 2). In these studies, the intestinal epithelium displayed pitting of goblet cells, suggesting stimulation of mucus secretion. EAEC strains adhered to sections of pediatric small bowel mucosa in an in vitro organ culture (IVOC) model (9); the bacteria were also embedded within a mucus-containing biofilm. Two further observations support a role of mucus in EAEC pathogenesis: EAEC bound rabbit mucus avidly in an in vitro model (10) and volunteers fed EAEC strains excrete mucoid stools (11). The formation of a heavy mucus biofilm may contribute to EAEC diarrheagenicity and, perhaps, to its ability to cause persistent colonization and diarrhea. [fig 2] Figure 2. Biofilm (arrow), containing aggregating bacteria and mucus, adhering to the mucosa of a gnotobiotic piglet fed enteroaggregative E. coli strain 042 and sacrificed after 3 days. This piglet did not contract diarrhea. (Reprinted with permission of James Nataro and Clinical Microbiology Reviews. Clin Microbiol Rev 1998;11:142-201.) In addition to forming the characteristic mucus biofilm, many EAEC strains induce cytotoxic effects on the intestinal mucosa. Infection with EAEC strains in rabbit and rat ileal loop models resulted in a destructive lesion demonstrable by light microscopy (7). EAEC induced shortening of the villi, hemorrhagic necrosis of the villous tips, and a mild inflammatory response with edema and mononuclear infiltration of the submucosa. Transmission electron microscopy showed normal microvillar architecture and no invasion of enterocytes; both light and electron microscopy showed adherent bacteria without the attaching and effacing lesion, which is characteristic of EPEC. Mucosal damage has also been demonstrated in ileum specimens taken after patients died of EAEC persistent diarrhea during an outbreak in the malnutrition ward of Mexico City Hospital (12). This effect was reproduced in the IVOC model; EAEC induced exfoliation of enterocytes from the mucosal surface of pediatric intestinal biopsies (9). Indeed, EAEC strains elicit cytotoxic effects on T84 cells (human intestinal carcinoma) in vitro (13; Figure 3). In the in vitro model, EAEC induced the microvillar membrane to vesiculate and the cells from the monolayer to exfoliate. This effect was accompanied by increased vacuole formation and separation of the nucleus from the surrounding cytoplasm. Damage was most prominent in areas where EAEC were adhering to the cells. In both the T84 and IVOC systems, the toxic effects required genes encoding the adherence fimbriae (14), as well as other genes on the 65MDa plasmid. Adherence Adherence of EAEC to the intestinal mucosa has been well studied. A flexible, bundle-forming EAEC fimbrial structure 2 nm to 3 nm in diameter, designated aggregative adherence fimbriae I (AAF/I) (15), mediates HEp-2 adherence and human erythrocyte hemagglutination in EAEC strain 17-2. The genes for AAF/I are organized as two separate gene clusters on the 60 MDa plasmid separated by 9 kb of intervening DNA (16-18). Region 1 genes encode the fimbrial structure itself; nucleotide sequence analysis of the Region 1 cluster suggests that AAF/I is related to the Dr family of adhesins (18). Region 2 genes encode a transcriptional activator of AAF/I expression (AggR), which is a member of the AraC family of DNA binding proteins (17). The AAF/I fimbriae are bundle-forming fimbriae but do not show homology with members of the so-called type 4 class of fimbriae (19). [fig 3] Figure 3. Effects of enteroaggregative E. coli strain 042 on T84 cells in culture. Polarized T84 monolayers were washed and inoculated with 10(sup 6) CFU of 042 and allowed to coincubate at 37°C for 6 hrs. Transmission electron microscopy reveals adherence of bacteria to the apical surface of the T84 cells without internalization. The apical brush border is effaced; cells are ballooning and will eventually be extruded from the monolayer. (J.P. Nataro and S. Sears, unpub. data). Bar, 1 mm. A second fimbrial structure, designated AAF/II, has been identified (14); it is distinct from but related to AAF/I. Insertional mutants in AAF/II genes no longer adhered to colonic mucosa in IVOC. DNA probe analysis suggested that AAF/I and AAF/II were each present in only a minority of EAEC strains and thus, as is the case with enterotoxigenic E. coli, intestinal colonization of EAEC appears to be mediated by one or more of several different fimbrial antigens. In some EAEC strains, AA may be due to factors other than the AAF fimbriae (20,21). An afimbrial outer membrane protein or a 38 kDa outer membrane protein may be responsible for AA in some strains (20,21). EAEC ST-Like Toxin Savarino et al. have identified an open reading frame encoding a 4,100 Da homologue of the heat-stable enterotoxin designated EAST1 (22,23), which is a 38-amino acid protein featuring four cysteine residues, instead of six (which are characteristic of E. coli ST). A role for EAST1 in human disease has not been demonstrated, although EAST1 clones yield net increases in short circuit current in the rabbit mucosal Ussing chamber model (22). Of the four strains given to volunteers, one that induced diarrhea, as well as one that did not, secreted EAST1 (11). EAST1 has been found in approximately 40% of EAEC strains and in a similar proportion of nonpathogenic E. coli strains (24). Other E. coli categories, most notably EHEC, elaborate EAST1 with a higher frequency than EAEC (24). Invasiveness Some EAEC strains may invade intestinal epithelial cells in vitro (25). However, human intestinal explants do not internalize EAEC, and clinical evidence for or against a role for invasiveness is lacking (9). Other EAEC Toxins In an outbreak of EAEC diarrhea in Mexico, Eslava et al. identified (in serum from patients in the outbreak) an approximately 108 kDa protein that induces exfoliation of enterocytes in the rat ileal loop model. DNA sequence analysis of the plasmid-borne gene (C. Eslava, J. Czeczulin, F. Navarro-Garcia, I. Henderson, A. Cravioto, J.P. Nataro, unpub. obs.) encoding this protein suggests that the toxin is a member of the family of proteins (26) called autotransporters because their secretion through the bacterial outer membrane is mediated by the carboxy terminus of the molecule. The 108 kDa toxin (termed Pet for plasmid-encoded toxin) also induces enterotoxic activity from rat intestinal tissue mounted in the Ussing chamber (F. Navarro-Garcia, C. Eslava, J.P. Nataro, A. Cravioto, unpub. obs.). A 120 kDa EAEC supernatant protein had been described that elicited rises in intracellular calcium in HEp-2 cells (27). Whether or not this protein is related to the protein described by Eslava et al. remains to be determined. Intestinal Inflammation An ongoing study of infant diarrhea in Fortaleza, Brazil (28-30), has shown that children with EAEC infection have intestinal inflammation, as measured by the presence of inflammatory markers in the stool (31). These markers include lactoferrin, a stable neutrophil product and sensitive marker for fecal neutrophils; interleukin-8 (IL-8), a neutrophil chemokine; and interleukin-1ß (IL-1ß), a polyfunctional proinflammatory cytokine. Fecal lactoferrin, IL-8, and IL-1ß were elevated in children with persistent diarrhea due to EAEC but not in controls (children in the same cohort without stool pathogens and free of diarrhea for 3 weeks before and after the time of the stool sample in prospective surveillance). Fecal lactoferrin and IL-1b were also elevated in children with EAEC but no diarrhea for 3 weeks. The origin of these inflammatory factors remains unknown, although they may be the result of direct stimulation of cytokine release from the epithelium. While studying cytokine release in vitro, Steiner et al. observed IL-8 release from Caco-2 intestinal epithelial cells exposed to cell-free filtrates of EAEC strains 042 or 17-2. This release was apparently mediated by a heat-stable, high molecular weight, chromosomally encoded protein (31). Lipopolysaccharide does not appear responsible for this effect as it is not blocked by polymyxin B and is not elicited by normal E. coli flora. The relationship between IL-8 release and the symptoms of EAEC infection is not known, especially since severe infiltration by neutrophils has not yet been reported in EAEC histopathologic specimens. However, a hypothetical role for IL-8 release in EAEC diarrhea can be envisioned. IL-8 released from epithelial cells in response to EAEC could act as the first step in a secretory cascade by recruiting neutrophils, since neutrophils in the intestinal epithelium release 5'-adenosine monophosphate, which is converted by a 5'-nucleotidase on the apical surface of enterocytes to adenosine, an agonist for chloride secretion (32). A similar mechanism of neutrophil chemotaxis and cytokine release may also be involved in the diarrhea produced by other microbial products, such as Clostridium difficile toxin A (33,34). Strain Heterogeneity EAEC strains exhibit considerable heterogeneity. Studies at the University of Maryland found that at a dose of 10(sup 10) CFU with bicarbonate, strain 042 (which produces AAF/II fimbriae, EAST1, and the 108 kDa Pet toxin) elicited loose stools in 4 of the 5 adult volunteers (11). Three other strains, all of which expressed AAF/I and one of which secreted EAST1, did not produce any intestinal symptoms. These data suggest that virulence is likely to be heterogeneous and that Pet may play an important role. A Model of EAEC Pathogenesis Extrapolating from the observations described above, we propose a three-stage model of EAEC pathogenesis. Stage I involves initial adherence to the intestinal mucosa and the mucus layer; this initial adherence is apparently mediated by the AAF fimbriae. Stage II comprises enhanced mucus production, apparently leading to the deposit of a thick mucus-containing biofilm encrusted with EAEC. The blanket may promote persistent colonization and nutrient malabsorption. Stage III, suggested from histopathologic and molecular evidence, includes the elaboration of toxins or inflammation, which result in damage to the mucosa and intestinal secretion. Malnourished hosts may be unable to repair the mucosal damage and may thus become prone to the persistent diarrhea syndrome. The site of EAEC infection in the human intestine has yet to be clearly demonstrated. IVOC experiments have shown that EAEC strains can adhere to both small and large bowel mucosa (9). Tissue specimens in these studies were derived from pediatric patients and were not formalin-fixed before they were incubated with EAEC strains. These features may explain the discrepant results obtained by other investigators (35,36). The short incubation period observed in some humans challenged with strain 042 (as short as 8 hours) is also consistent with involvement of the small bowel in diarrheagenicity (11). Clinical Features The clinical features of EAEC diarrhea are increasingly well defined in outbreaks, sporadic cases, and the volunteer model. Typical illness is manifested by a watery, mucoid, secretory diarrhea with low grade fever and little to no vomiting (37,38). However, up to one-third of patients with EAEC diarrhea have had grossly bloody stools (39); this outcome may be strain dependent. In volunteers infected with EAEC strain 042, the diarrhea was mucoid, of low volume, and without occult blood or fecal leukocytes; all patients remained afebrile. In such volunteers, the incubation period of the illness was 8 to 18 hours (11). The duration of EAEC diarrhea is its most striking feature. In a community surveillance study in Anapur-Palla in northern India, the mean duration of EAEC-associated diarrhea cases in children under 3 years of age was 17 days, longer than that associated with any other pathogen (37). In this study, of the 41 cases of diarrhea, fever was found in 12%, gross blood in 12%, vomiting in 7%, and persistence of the episode more than 14 days in 44%. A large percentage of patients excreting EAEC have detectable fecal lactoferrin and supranormal levels of IL-8 in the stool (31). Although this observation suggests that EAEC infection may be accompanied by mucosal inflammation, most patients lack overt clinical evidence of inflammation. Diagnosis EAEC colonization is detected by isolating E. coli from stool samples and demonstrating the AA pattern in the HEp-2 assay. Analysis of small bowel aspirates has not increased yield (40). Implication of EAEC as the cause of the patient's disease must be done cautiously, given the high rate of asymptomatic colonization in many populations (4,30,41-43). If no other organism is implicated in the patient's illness and EAEC is isolated repeatedly, it may be considered to be a probable cause of the patient's illness. Despite various methods of performing the HEp-2 assay (4,44,45), comparative studies suggest that the technique first described by Cravioto et al. (45,46) (i.e., a single 3-hour incubation of bacteria with cells, without a change in medium during the assay) best discriminates among the three adherence patterns. Because AAF adhesins are expressed maximally in static L-broth cultures at 37°C (15), we incubate all HEp-2 assay inocula in this manner. A DNA fragment probe has proven highly specific in detecting EAEC strains. This probe was developed by Baudry et al. (47), who tested fragments derived empirically from the plasmids of strains 17-2 and 042. A 1.0 kb plasmid-derived Sau3a fragment hybridized with 56 (89%) of 63 EAEC strains (defined by HEp-2 assay); of 376 strains representing normal flora and other diarrheagenic categories, only 2 hybridized with the probe. The correlation of the EAEC probe-positivity with AA varies by location. In some studies, the correlation achieves the 89% sensitivity reported by Baudry et al. (47,48), while in other studies, the sensitivity may be substantially lower (40). The epidemiologic significance of probe-positive versus probe-negative strains has not been determined. The nucleotide sequence of the AA probe represents a cryptic open reading frame adjacent to the plasmid replicon (J.P. Nataro, unpub. obs.). A polymerase chain reaction assay using primers derived from the AA probe sequence shows similar sensitivity and specificity (49). Epidemiology Volunteer studies and outbreak investigations have shown that at least some EAEC strains are human pathogens. A growing number of studies have supported the association of EAEC with diarrhea in developing countries; the increasing number of such reports and the rising proportion of diarrheal cases in which EAEC are implicated suggest that EAEC are important emerging agents of pediatric diarrhea. EAEC in Sporadic Diarrhea The earliest epidemiologic reports showed that EAEC were most prominently associated with persistent (lasting >/=14 days) cases of pediatric diarrhea (Table 1). In some of these studies, EAEC cultured from the stool during the first few days of diarrhea were predictive of a longer illness (37,43). The importance of EAEC in diarrheal disease appears to vary geographically. On the Indian subcontinent, six studies (which included hospitalized patients with persistent diarrhea [50], outpatients visiting health clinics [51], and cases of sporadic diarrhea detected on household surveillance [37]) have demonstrated the importance of EAEC in pediatric diarrhea (37,38,48,50,51,53). EAEC and persistent diarrhea syndrome have been consistently associated (28,29,30,40). In Brazil, EAEC have been implicated in up to 68% of persistent diarrhea cases (40), which represent a disproportionate share of deaths due to diarrhea. EAEC have also been implicated as causes of sporadic diarrhea in Mexico, Chile, Bangladesh, and Iran (4,39,42,43). Table 1. Epidemiologic studies associating enteroaggregative Escherichia coli (EAEC) with diarrhea ( p < 0.05) ---------------------------------------------------------------------------------- Source or EAEC in cases reference Syndrome Site vs controls Comments ----------------------------------------------------------------------------------- 4 All Santiago, Chile 84(33%)/253 vs Community and diarrhea 20(15%)/134(sup a) hospital-based. First description of EAEC 37 Persistent Anapur-Palla, 18(30%)/61 vs. Household surveillance of diarrhea India 20(10%)/201(sup b) ruralchildren 1 log kill(sup a,b) (interquartile range) O strains ------------------------- ----------------------- Illness serotype studied Pool 1(sup c) Pool 2(sup c) Pool 1 Pool 2 --------------------------------------------------------------------------------- GBS(sup d) 19 7 1 (14) 0 (0) 0.53 0.39 (0.18-0.63) (0.08-0.52) Enteritis 19 11 0 (0) 0 (0) 0.37 0.48 (0.09-0.47) (0.29-0.50) GBS non-19 10 3 (30) 3 (30) 0.70 0.43 (0.44-1.18) (0.03-1.32) Enteritis non-19 16 8 (50) 7 (44) 0.90 0.59 (0.24-2.00) (0.35-2.53) All 19 18 1 (6) 0 (0) 0.38 0.45 (0.17-0.58) (0.23-0.52) All non-19 26 11 (42) 10 (38) 0.70 0.54 (0.26-1.86) (0.17-2.27) --------------------------------------------------------------------------------- (sup a)Bacterial suspensions were incubated in either NHS or heat inactivated NHS at 37°C for 1 hr as described in text, and net CFU (NHS minus HINHS) determined. Values greater than 1 log(sub 10) kill were considered to denote a serum-susceptible strain. (sup b)Comparison of O:19 strains with non-O:19 strains: Pool 1, odds ratio = 12.5, p = 0.008; Pool 2, odds ratio = undefined, p = 0.003. (sup c)Pool 1 consists of C. jejuni-antibody positive serum from five healthy adults; Pool 2 consists of C. jejuni-antibody negative serum from two healthy adults. (sup d)GBS = Guillain-Barré syndrome. (sup 125)I-C3 Binding Assays Eight strains from Group 1 (four randomly selected from each subgroup) and eight from Group 2 (four randomly selected from each subgroup) were grown on blood agar plates as described above. C. fetus strains 23D and 23B again served as controls and the assays were conducted (21). In brief, bacteria from each plate were harvested in 1.5 ml of HBSS and were centrifuged at 8,000 g for 15 minutes; the pellet was resuspended in 0.5 ml HBSS and adjusted to OD(sub 450)0 = 3.0. For each strain studied, a suspension of 4 µl (sup 125)I-C3 (with 20,000 cpm to 65,000 cpm) and 2.5 µl of either NHS or HINHS from pool 2 and 100 µl of the bacteria-HBSS suspension were incubated at 37oC for 15 minutes. (sup 125)I-C3 was prepared in the laboratory of one of the authors (RGW) (21). The suspensions were then centrifuged twice at 175xg for 5 minutes, the pellet was resuspended in HBSS, and the supernatant was discarded. The bottom 5 mm (containing the pellet) of each tube was clipped, and emissions were determined in a gamma counter. The counts in the negative control mixtures containing HINHS were subtracted from the NHS counts. An assay was considered valid only if the net pellet counts (NHS minus HINHS) for control strain 23B were at least four times higher than for 23D. To control for nonspecific binding, the counts for each strain studied were expressed as the ratio of net counts relative to the serum susceptible control (23B). Each strain was assayed two to four times. Mean binding of serum-susceptible control strain 23B was 497 cpm, whereas mean binding of serum-resistant control strain 23D was 54. The mean ratio of C3-binding for strain 23D to 23B of 0.114 was as expected (21). In contrast, the mean ratio for C3 binding to C. jejuni strains in comparison with strain 23B was 0.022 to 0.464 (mean 0.216). The eight O:19 strains were not significantly different from the eight non-O:19 strains in their ability to bind (sup 125)I-C3 (Table 3). Table 3. Comparison of (sup 125)I-C3 binding to Campylobacter jejuni strains -------------------------------------------------------------- Measure Serotype of C. jejuni strain -------------------------------------------------------------- O:19 non-O:19 (n = 8) (n = 8) Counts per minute(sup a) 85 +/-51 98 +/- 48 Mean binding ratio(sup b) 0.187 +/- 0.245 +/- 0.116 0.130 -------------------------------------------------------------- (sup a)Bacterial cells were incubated with pool 2 normal human serum or heat inactivated normal human serum and (sup 125)I-C3 (mean 40,416 [range 20,234-63,503]. cpm) at 37°C for 15 minutes. The cells were centrifuged and washed; net (NHS minus HINHS) 125I binding was measured as counts per minute. p-value = 0.20 (paired T-test, 1-tailed). (sup b)The mean binding ratio is an average of the ratio of the net counts for each study strain relative to the net counts of the serum-susceptible control (23B). p-value = 0.31 (paired T-test, 1-tailed). Conclusion This study of GBS patients in the United States and Germany confirms the observation made in Japan that serotype O:19 is overrepresented among patients with C. jejuni-induced GBS. Of 298 randomly collected Campylobacter isolates from patients with enteritis in the United States, only 3% were serotype O:19 (19). A similarly low prevalence of O:19 strains is found in all parts of the world, including North and South America, Asia, and Europe (5,6,19,23). Although specific serotyping surveys have not been done in Germany, it is unlikely that serotype O:19 is more frequent among German C. jejuni isolates. Thus, the GBS-associated strains in our study were more than 11 times as likely to belong to this serotype (p = 0.03). Although the association among GBS and C. jejuni serotype O:19 was not as marked as in Japan (where more than 80% of GBS-associated isolates are O:19), this serotype clearly is overrepresented among GBS-associated strains in other countries. We conclude that the association of O:19 strains with GBS is not just a local phenomenon in Japan but likely reflects a fundamental characteristic of O:19 strains. O-serotype 2 and heat-labile serotype 4, which were common among the GBS strains, are commonly represented among infected persons in the United States (23). Despite the frequency of Campylobacter infections in GBS patients, such strains are difficult to obtain for several reasons. First, in most C. jejuni-infected patients, stools are clear before neurologic symptoms begin. Second, most neurologists do not culture stool samples when GBS is first diagnosed. And finally, even if a stool culture is ordered and Campylobacter is present, few microbiology laboratories save their isolates; by the time the case is reported, the strain has been discarded. Thus, these seven strains represent one of the largest collections of GBS-associated strains described. Additionally, these are likely to be representative of the population of GBS-associated C. jejuni strains. Unless serotype O:19 strains persist in stools longer, are more easily cultured, or are less likely to be discarded by microbiology laboratories (and no data support any of these possibilities), these strains probably are not different in any systematic way from other GBS-associated isolates. Most C. jejuni strains are susceptible to killing by human serum (22), but because we studied highly selected strains (most either serotype O:19 or from GBS patients), a high proportion of strains in this investigation were resistant. In this context, the finding that C. jejuni strains from GBS patients were no more likely than the strains from patients with uncomplicated enteritis to be serum resistant is not surprising. Even when the analysis was limited only to serotype O:19 strains, no differences were found between GBS- and enteritis-associated strains. However, serotype O:19 strains were substantially less serum-susceptible than strains of other serotypes and were 12 times as likely to be serum-resistant. Although serotype O:19 represents only a small percentage of C. jejuni strains from patients with uncomplicated enteritis in the United States or Japan, it is the most common serotype identified in GBS patients in both locales. The relatively small variation in serum susceptibility of the O:19 strains is consistent with the close genetic relationship observed among these strains (24). To better understand the basis for the relative serum-resistance among the O:19 strains, we compared their ability to bind C3 in relation to non-O:19 strains. Because C3 binding occurs after activation by either the alternative or classical complement pathways, it is a screen for differences in these early steps. The serum-resistance of C. fetus is explained by the inability of C3 to bind to the cell surface (25). In contrast, C3 binding is normal in serum-resistant Salmonella, but the C5-9 membrane attack complex does not insert properly (26). In the present study, the lack of substantial C3 binding differences among O:19 and non-O:19 C. jejuni strains suggests that the early steps in complement activation are similar among strains. However, clear differences in serum-susceptibility bespeak either rapid inactivation of C3 or reduced assembly or function of the membrane attack complex in the more resistant strains. Future studies should address this point. The increasing awareness of the importance of C. jejuni infection in triggering GBS is another example of how previously well-described diseases have emerged as sequelae of acute infectious illnesses. This study attempts to begin to characterize the nature of this association; however, there is much to learn about how an acute gastrointestinal infection results in ascending paralysis. One fact is quite clear: many more people are infected with C. jejuni than contract GBS subsequently. Perhaps some persons are predisposed to contracting GBS after infection with campylobacters that might cause only uncomplicated enteritis in another patient. Conversely, as we have suggested in this paper, some strains may be more likely than others to trigger GBS. No associations between human leukocyte antigen (HLA) types and GBS have been found (27,28). However, in Great Britain and Japan, an association between HLA type and C. jejuni-associated GBS has been suggested (9,29). Perhaps some combination of familial susceptibility, HLA type, strain serotype, or other host or strain characteristics together play a role in the pathogenesis of C. jejuni-induced GBS. The relative serum-resistance of O:19 strains correlates with mechanism. Furthermore, the relevance of these in vitro assays to the susceptibility of organisms in vivo cannot be known with certainty. We speculate that the relative insensitivity of these strains to the lytic effects of complement allows them to trigger a heightened specific immunologic response. We further speculate that this heightened immunologic response leads to injury of peripheral nerve structures. Since only a small fraction of infections caused by C. jejuni O:19 lead to GBS (estimated incidence 1 in 158) (30), additional factors also must be involved in vivo. Acknowledgments We thank Charlotte Patton for serotyping the strains and Manfred Kist for supplying the strains from Germany. This work was supported by grant NIH-1-K08-NS01709 of the NINDS, NIH (BMA); NIH grant KO4-AI01036 (RGW); and the Medical Research Service of the Department of Veterans Affairs (MJB). Ban Mishu Allos is an assistant professor of medicine in the Division of Infectious Diseases, Department of Medicine and the Department of Preventive Medicine at Vanderbilt University School of Medicine in Nashville, Tennessee. Her clinical work relates to HIV and other infectious diseases; her research has focused on foodborne illness in general and Campylobacter jejuni as a cause of Guillain-Barré syndrome in particular. Address for correspondence: Ban Mishu Allos, A-3310 MCN, Division of Infectious Diseases, Vanderbilt University School of Medicine, Nashville, TN 37232, USA; fax: 615-343-6160; e-mail: Ban.mishu.allos@mcmail.vanderbilt.edu. References 1. Hughes RAC. Guillain-Barré Syndrome. London: Springer-Verlag; 1990. 2. Asbury AK, Johnson PC. Pathology of peripheral nerves. Philadelphia: WB Saunders; 1978. 3. Winer JB, Hughes RAC, Anderson MJ, Jones DM, Kangro H, Watkins RFP. A prospective study of acute idiopathic neuropathy II. Antecedent events. J Neurol Neurosurg Psychiatry 1988;51:613-8. 4. Mishu B, Ilyas AA, Koski CL, Vriesendorp F, Cook SD, Mithen FA, et al. Serologic evidence of preceding Cam-pylobacter jejuni infection in patients with the Guillain-Barré syndrome. Ann Intern Med 1993;118:947-53. 5. Kuroki S, Saida T, Nukina M, Haruta T, Yoshioka M, Kobayashi Y, et al. Campylobacter jejuni strains from patients with Guillain-Barré syndrome belong mostly to Penner serogroup 19 and contain B-N-acetylglucosamine residues. Ann Neurol 1993;33:243-7. 6. Rees JH, Soudain SE, Gregson NA, Hughes RAC. Campylobacter jejuni infection and Guillain-Barré syndrome. N Engl J Med 1995;333:1374-9. 7. Allos BM. Campylobacter jejuni infection as a cause of the Guillain-Barré syndrome. Infect Dis Clin North Am 1998;12:173-84. 8. Ho TW, Mishu B, Li CY, Gao CY, Cornblath DR, Griffin JW, et al. Guillain-Barré syndrome in northern China: relationship to Campylobacter jejuni infection and anti-glycolipid antibodies. Brain 1995;118:597-605. 9. Yuki N, Taki T, Takahashi M, Saito K, Yoshino H, Tai T, et al. Molecular mimicry between GQ1b ganglioside and lipopolysaccharide of Campylobacter jejuni isolated from patients with Fisher's syndrome. Ann Neurol 1994;36:791-3. 10. Tauxe RV. Epidemiology of Campylobacter jejuni infections in the United States and other industrialized nations. In: Nachamkin I, Blaser MJ, Tompkins LS, editors. Campylobacter jejuni-current strategy and future trends. Washington: American Society for Microbiology; 1992. p. 9-19. 11. Allos BM, Blaser MJ. Potential role of lipopolysaccharides of Campylobacter jejuni in the development of Guillain-Barré syndrome. Journal of Endotoxin Research 1995;2:237-8. 12. Rees JH, Vaughan RW, Kondeatis E, Hughes RAC. HLA-class II alleles in Guillain-Barré syndrome and Miller Fisher syndrome and their association with preceding Campylobacter jejuni infection. J Neuroimmunol 1995;62:53-7. 13. Fujimoto S, Yuki N, Itoh T, Amako K. Specific serotype of Campylobacter jejuni associated with Guillain-Barré syndrome. J Infect Dis 1992;165:183. 14. Lastovica AJ, Goddard EA, Argent AC. Guillain-Barré syndrome in South Africa associated with Campylobacter jejuni O:41 strains. J Infect Dis 1997;(Suppl 2):S139-S43. 15. Enders U, Karch H, Toyka KV, Heesemann J, Hartung HP. Campylobacter jejuni and Guillain-Barré syndrome. Ann Neurol 1994;35:249. 16. Blaser MJ, Perez-Perez G, Smith PF, Patton CM, Tenover FC, Lastovica AJ, et al. Extraintestinal Campylobacter jejuni and Campylobacter coli infections: host factors and strain characteristics. J Infect Dis 1986;153:552-9. 17. Penner JL, Hennessy JN. Passive hemagglutination technique for serotyping Campylobacter fetus subsp. jejuni on the basis of soluble heat-stable antigens. J Clin Microbiol 1980;2:378-83. 18. Lior H, Woodward DL, Edgar JA, Laroche LJ, Gill P. Serotyping of Campylobacter jejuni by slide agglutination based on heat-labile antigenic factors. J Clin Microbiol 1982;15:761-8. 19. Patton CM, Nicholson, MA, Ostroff SM, Ries AA, Wachsmuth IK, Tauxe RV. Common somatic O and heat-labile serotypes among Campylobacter strains from sporadic infections in the United States. J Clin Microbiol 1993;31:1525-30. 20. Blaser MJ, Smith PF, Hopkins JA, Heinzer I, Bryner JH, Wang WLL. Pathogenesis of Campylobacter fetus infections. Serum resistance associated with high molecular weight surface proteins. J Infect Dis 1987;155:696-709. 21. Gonzales-Valencia G, Perez-Perez GI, Washburn RG, Blaser MJ. Susceptibility of Helicobacter pylori to the bactericidal activity of human serum. Helicobacter 1996;1:28-33. 22. Blaser MJ, Smith PF, Kohler PF. Susceptibility of Campylobacter isolates to the bactericidal activity of human serum. J Infect Dis 1985;151:227-35. 23. Patton CM, Wachsmuth IK. Typing schemes-are current methods useful? In: Nachamkin I, Blaser MJ, Tompkins LS, editor. Campylobacter jejuni-current strategy and future trends. Washington: American Society for Microbiology; 1992. p. 110-28. 24. Fujimoto S, Allos BM, Patton CM, Blaser MJB. Restriction fragment length polymorphism analysis and random amplified polymorphic DNA analysis of Campylobacter jejuni strains isolated from patients with Guillain-Barré syndrome. J Infect Dis 1997;176:1105-8. 25. Blaser MJ, Smith PF, Repine JE, Joiner KA. Pathogenesis of Campylobacter fetus infection. J Clin Invest 1988;81:1434-44. 26. Joiner KA, Hammer CH, Brown EJ, Cole RJ, Frank MM. Studies on the mechanism of bacterial resistance to complement-mediated killing. I. Terminal complement components are deposited and released from Salmonella minnesota S218 without causing bacterial death. J Exp Med 1982;155:809-19. 27. Hillert J, Osterman PO, Olerup O. No association with HLA-DR, -DQ, or -DP alleles in Guillain-Barré syndrome. J Neuroimmunol 1991;31:67-72. 28. Winer JB, Briggs D, Welsh K, Hughes RAC. HLA antigens in the Guillain-Barré syndrome. J Neuroimmunol 1988;18:13-6. 29. Yuki N, Fujimoto S, Yamada S, Tsujino Y, Kinoshito A, Itoh T. Serotype of Campylobacter jejuni, HLA, and the Guillain-Barré syndrome. Muscle Nerve 1992;968-9. 30. Allos BM, Blaser MJ. Campylobacter jejuni infection and the Guillain-Barré syndrome: mechanisms and implications. International Journal of Medical Microbiology, Virology, Parasitology, and Infectious Diseases 1994;281:544-8. Emerging Infectious Diseases National Center for Infectious Diseases Centers for Disease Control and Prevention Atlanta, GA URL: http://www.cdc.gov/ncidod/EID/vol4no2/allos.htm Please note that figures and equations are not available in ASCII format; their placement within the text is noted by [fig] and [eq], respectively. Greek symbols are spelled out. The following codes are used: (ft) for footnote; (sup) for superscript; (sub) for subscript; /= for greater than or equal to. ------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------- [Emerging Infectious Diseases] [Volume 4 No. 2 / April - June 1998] Dispatches An Apparently New Virus (Family Paramyxoviridae) Infectious for Pigs, Humans, and Fruit Bats Adrian W. Philbey,* Peter D. Kirkland,* Anthony D. Ross,* Rodney J. Davis,* Anne B. Gleeson,* Robert J. Love,† Peter W. Daniels,‡ Allan R. Gould,‡ and Alex D. Hyatt‡ *Elizabeth Macarthur Agricultural Institute, Camden, New South Wales, Australia; †University of Sydney, Camden, New South Wales, Australia; and ‡CSIRO Australian Animal Health Laboratory, Geelong, Victoria, Australia -------------------------------------------------- We isolated an apparently new virus in the family Paramyxoviridae from stillborn piglets with deformities at a piggery in New South Wales, Australia. In 1997, the pregnancy rate and litter size at the piggery decreased markedly, while the proportion of mummified fetuses increased. We found serologic evidence of infection in pigs at the affected piggery and two associated piggeries, in humans exposed to infected pigs, and in fruit bats. Menangle virus is proposed as a common name for this agent, should further studies confirm that it is a newly recognized virus. Viruses in the family Paramyxoviridae have been associated with new diseases in a variety of species, including humans, throughout the world. Morbilliviruses related to canine distemper virus have caused disease outbreaks in seals, dolphins, porpoises, and lions (1-6). Equine morbillivirus (EMV) (also called Hendra virus or bat paramyxovirus), responsible for the deaths of horses and humans with respiratory or neurologic disease in Australia (7,8), is thought to have originated from fruit bats (Pteropus sp.) (9-11). La Piedad Michoacan paramyxovirus (blue eye paramyxovirus) causes encephalitis and death in piglets and reproductive disease in adult pigs in Mexico (12). We have isolated an apparently new virus in the family Paramyxoviridae from stillborn piglets