Skip Navigation LinksSkip Navigation Links
Centers for Disease Control and Prevention
Safer Healthier People
Blue White
Blue White
bottom curve
CDC Home Search Health Topics A-Z spacer spacer
Blue curve MMWR spacer

Clinical Sepsis and Death in a Newborn Nursery Associated with Contaminated Parenteral Medications -- Brazil, 1996

In October 1996, a total of 35 newborn infants died in a 26-bed nursery of a 200-bed hospital in Roraima, Brazil; these deaths represented a significant increase over the baseline mortality rate in the nursery (6.0 versus 1.7 per 100 live births; pless than 0.01). Twenty of the deaths were attributed to sepsis. Fatal episodes of sepsis began 24-72 hours after birth. Although an investigation by the Roraima Health Department resulted in an improvement of infection control, increased episodes of fever and clinical sepsis persisted. As a result, in November 1996, the Secretary of Health of Roraima, Brazil Ministry of Health, requested that CDC assist in the investigation. This report summarizes this investigation, which implicated locally produced intravenous (IV) solutions as the source of the outbreak and underscores the need to assure proper quality control of parenteral medications and the importance of nosocomial infection surveillance.

In November 1996, CDC conducted a cohort study to identify risk factors for the development of fever. A case was defined as fever of 100.4 F (38 C) without a recognized cause in any neonate who was admitted to the hospital nursery on October 1, 5, 15, or 25 (these dates were chosen to represent the entire month) and who received antimicrobial therapy for sepsis. Six of the 66 patients admitted to the nursery on these days met the case definition. When case- and non-case-patients were compared, neither sex, gestational age, birthweight, nor Apgar score at 5 minutes were associated with the development of fever. In comparison, Apgar scores at 1 minute of less than 8 (five of 23 versus one of 43; p=0.01) and insertion of a peripheral IV catheter with receipt of parenteral medications (six of nine versus zero of 57; pless than 0.01; relative risk {RR}=20) were associated with the development of fever. Because this cohort study strongly implicated the insertion of peripheral IVs and receipt of parenteral medications as risk factors for the development of fever, a second cohort study was conducted to determine whether IVs were associated with the development of fever before the time of the outbreak. This second cohort included all patients admitted to the hospital nursery on June 1, 5, 15, or 25, dates preceding the time of the increase in death rate. None of the 55 patients admitted to the nursery during the June dates developed fever, including 11 patients who had been exposed to parenteral medications. The attack rate for fever following exposure to parenteral medications was significantly higher on the four dates in October compared with the four dates in June (six of nine versus zero of 11; pless than 0.01; RR=undefined).

To identify risk factors for death attributed to clinical sepsis, CDC conducted a case-control study on infants admitted to the nursery during October 1996. The case definition was expanded to include death following the onset of fever and clinical sepsis. Twenty infants admitted to the nursery during October met the case definition; 40 birthdate-matched patients were included as controls. Case-patients were more likely than controls to have a lower birthweight (mean: 2.3 kg versus 3.3 kg; p=0.01), lower Apgar score at 1 minute (mean: 6.5 versus 7.7; p=0.01) or 5 minutes (mean: 8.0 versus 8.8; p=0.01), lower gestational age (mean: 33.8 versus 38.8 weeks; p=0.01), or a peripheral IV and to have received parenteral medications (20 of 20 versus zero of 40; p=0.01). Various parenteral medications (i.e., glucose, aminophylline, calcium gluconate, penicillin, sodium chloride, potassium, and sodium bicarbonate) were administered to case-patients; only glucose, aminophylline, and bidistilled water (used to reconstitute medications) were administered to all case-patients. All case-patients developed fever only after they were exposed to the parenteral medications. Thirteen of 14 blood cultures taken from case-patients less than or equal to 2 days after the onset of fever were negative for bacterial growth; one blood culture was positive for Klebsiella pneumoniae.

Samples of parenteral fluids and medications used in the nursery were examined for bacterial and/or endotoxin contamination. Endotoxin was measured using the limulus amoebocyte assay. All cultures of these solutions were negative for bacterial growth. However, six of 13 unopened vials of bidistilled water for injection and 12 of 15 unopened vials of 25% glucose, manufactured by Hipolabor Farmaceutica Ltda. (Sabara, Minas Gerais, Brazil), had elevated endotoxin levels of 0.8-5.8 endotoxin units (EU)/mL (mean: 3.8 EU/mL) and 0.8-1.9 EU/mL (mean: 1.2 EU/mL), respectively. The United States Pharmacopeia (U.S.P.) endotoxin limit on water for injection is 0.25 EU/mL and for glucose (5%-70%) is 0.5 EU/mL. Caked amorphous-like material and bacterial cells were observed by scanning electron microscopy in samples of bidistilled water containing elevated levels of endotoxin.

Reported by: A Wanderley, Roraima Health Dept; C Wanderley, Hospital e Maternidade Nossa Senhora de Nazare, Boa Vista, Roraima; Ministry of Health, Brazil. Hospital Infections Program, National Centers for Infectious Diseases; and EIS officers, CDC.

Editorial Note

Editorial Note: The findings in the investigation described in this report implicated insertion of a peripheral IV line and receipt of parenteral medications as resulting in clinical sepsis and death during the outbreak. Laboratory results documented endotoxin contamination of unopened vials of parenteral medications administered to infants, suggesting intrinsic contamination of these products.

The release of endotoxin into the circulatory system is the initiating event of sepsis associated with gram-negative organisms. Subsequently, reactions may range from no detectable response to the onset of profound shock and death (1,2). Such reactions are highly dependent on the body mass of the patient (3). Because the minimal pyrogenic dose of endotoxin is 5 EU/kg (4), 2-3 mL of the contaminated bidistilled water (mean level of contamination: 3.8 EU/mL) would have been sufficient to evoke a pyrogenic reaction in an average 4 lbs, 8 oz (2000 g) infant. As a result, IV administration of these endotoxin-contaminated fluids (bidistilled water and/or 25% glucose) explained the increased number of febrile reactions detected during this outbreak. All infants receiving parenteral medications were receiving bidistilled water and glucose. Attack rates of 70% among these infants suggest that not all lots of bidistilled water and glucose were contaminated.

Exposure of the infants to parenteral fluids contaminated with endotoxin also may have been sufficient to cause the increased number of deaths during October. Previous studies on both humans and animals have demonstrated that endotoxin is capable of inducing clinical sepsis and death (5). Infants with low birthweight and gestational age were probably at increased risk for death because of the smaller amount of endotoxin required to cause serious pyrogenic reactions.

An investigation at the hospital described in this report previously had detected several breakdowns in aseptic technique and infection-control practices. Blood cultures collected and processed during October (mean: 8 days after the onset of fever) revealed the presence of bloodstream infection (BSI) in several infants. Exposure to endotoxins may have resulted in higher BSI rates by prolonging the exposure of infants to peripherally inserted IVs and breaks in aseptic technique during their manipulation. Despite a continued increase in episodes of unexplained fever, BSIs and deaths decreased in November after an improvement in infection-control practices.

Unopened vials of contaminated medication were undamaged and had no evidence of tampering, suggesting that contamination most likely occurred during the manufacturing process. Without appropriate manufacturing processes, endotoxin can contaminate solutions and reagents (6). Many gram-negative organisms, which can release endotoxin, require few nutrients and can grow in distilled water at 39.2 F (4 C). In addition, endotoxins can survive exposure to steam autoclaving, organic solvents, acids, ethanol, and sterilizing liquids. Only dry heat (greater than or equal to 482 F {greater than or equal to 250 C} for 30 minutes or greater than or equal to 356 F {greater than or equal to 180 C} for 3 hours) can assure the elimination of endotoxin (7).

The manufacturing plant in Minas Gerais, Brazil, where the medications implicated in this outbreak were made, was closed when inadequate quality-control testing was observed by the Secretary of Health of Minas Gerais. Although all Brazilian state secretaries of health were notified of the closure, no nationally coordinated product recall was performed. Based in part on the findings of this investigation, the Secretary of Health of Minas Gerais decided not to allow reopening of the manufacturing facility until quality-control measures were improved.

The routine surveillance of nosocomial infections at the hospital level is essential for the early detection and control of such epidemics. Clusters of pyrogenic reactions should always lead to an evaluation of possible product contamination. Surveillance also is important at a nationwide level. In this outbreak, contaminated medications were distributed widely throughout Brazil, and similar episodes of sepsis among neonates were reported from other nurseries around the country.


  1. Bone RC. The pathogenesis of sepsis. Ann Intern Med 1991;115:457-69.

  2. Danner RL, Elin RJ, Hosseini JM, Wesley RA, Reilly JM, Parillo JE. Endotoxemia in human septic shock. Chest 1991;99:169-75.

  3. Burrell R. Human responses to bacterial endotoxin. Circulatory Shock 1994;43:137-53.

  4. Greisman SE, Hornick RB. Comparative pyrogenic reactivity of rabbit and man to bacterial endotoxin. Proc Soc Exp Biol Med 1969;131:1154-8.

  5. Natanson C, Danner RL, Elin RJ, et al. Role of endotoxemia in cardiovascular dysfunction and mortality: Escherichia coli and Staphylococcus aureus challenges in a canine model of human septic shock. J Clin Invest 1989;83:243-51.

  6. Bennett IL, Beeson PB. The properties and biologic effects of bacterial pyrogens. Medicine 1950;29:365-400.

  7. Tsuji K, Harrison SJ. Dry-heat destruction of lipopolysaccharide: dry-heat destruction kinetics. Appl Environ Microbiol 1978;36:710-4.

Disclaimer   All MMWR HTML versions of articles are electronic conversions from ASCII text into HTML. This conversion may have resulted in character translation or format errors in the HTML version. Users should not rely on this HTML document, but are referred to the electronic PDF version and/or the original MMWR paper copy for the official text, figures, and tables. An original paper copy of this issue can be obtained from the Superintendent of Documents, U.S. Government Printing Office (GPO), Washington, DC 20402-9371; telephone: (202) 512-1800. Contact GPO for current prices.

**Questions or messages regarding errors in formatting should be addressed to

Page converted: 10/05/98


Safer, Healthier People

Morbidity and Mortality Weekly Report
Centers for Disease Control and Prevention
1600 Clifton Rd, MailStop E-90, Atlanta, GA 30333, U.S.A


Department of Health
and Human Services

This page last reviewed 5/2/01