The content on this page is being archived for historic and reference purposes only. The content, links, and pdfs are no longer maintained and might be outdated.
Epidemiologic Notes and Reports Burkitt's Lymphoma -- Winchester, Virginia
Three cases of Burkitt's lymphoma (BL) have occurred since 1971 in young boys living in one residential section of Winchester, Virginia. Onset of illness in the first 2 cases (ages 9 and 15) occurred simultaneously in August 1971 (1); the third patient (age 8) first became ill in July 1975.
The first 2 patients lived 2 houses apart and the third one-half mile away. None of the 3 patients or their families was acquainted or shared common community activities, although all 3 boys had attended the same grade school at different times. All 3 patients were Caucasian.
Manifestations of disease were similar in all 3 cases. Initial symptoms in each case involved persistent sore throat progressing to peritonsillar masses in the first 2 cases and to cervical adenopathy in the third. Histologic diagnosis of BL was made on biopsy of these masses and was confirmed on pathologic review at the National Cancer Institute. Marrow aspiration was suggestive of leukemic transformation in the third case, but not in the first 2. In each case there was clinical evidence of central nervous system (CNS) tumor involvement. Despite intensive chemotherapy in all 3 cases, and radiation therapy in 2, disease was unremitting and led to death 1, 4, and 7 months after initial diagnosis. Autopsy findings in the first 2 cases showed widespread tumor involving the CNS, as well as cervical and abdominal lymph nodes. In the third case, autopsy showed residual tumor only at the original site of occurrence (the neck). Serum antibody against Epstein Barr virus (EBV) was present at low titers in 2 cases.
Epidemiologic investigation revealed no clues to suggest a possible common etiology for these cases beyond their closeness in time and place of occurrence. Although the summers of 1971 and 1975 had somewhat greater rainfall than other years, no evidence was found to link the cases to increased mosquito or other insect exposures. No unusual local patterns of leukemia, lymphoma, or infectious mononucleosis incidence were found, and no other persons with childhood cancer were identified as having been in contact with the patients or being in their particular neighborhood.
Reported by: N McWilliams, MD, Medical College of Virginia; W Hatfield, MD, Lord Fairfax Health District; R Jackson, MD, State Epidemiologist, Virginia State Dept of Health; and Cancer and Birth Defects Div and Field Services Div, Bur of Epidemiology, CDC.
editorial note: Burkitt's lymphoma is extraordinarily rare outside certain parts of central Africa and New Guinea where it constitutes the most common tumor of childhood (2). Epidemiologic surveys of childhood cancer in the metropolitan areas of Atlanta, Georgia, and Houston, Texas, suggest that the annual incidence of BL in the United States is in the range of about 1 case per million children, or an expected incidence of about 1 case every 200 years for a town the size of Winchester. In this context, the occurrence of 3 cases over 5 years in one neighborhood is distinctly unusual.
The etiologic significance of such case clustering is not clear. While early studies of African BL suggested that time-space case clustering was a common feature of the disease (3), more recent epidemiologic work has cast doubt on the idea (4). Current evidence suggests that the etiology of African BL may be related to EBV infection in a host whose immunologic state has been severely affected by constant and severe malarial infections (5). The present cases, however, provide no obvious clinical or epidemiologic clues regarding their particular etiology. references
Editorial Note 1997: Public concern can quickly rise when persons perceive an excess of cancer in their local community. Such situations are not infrequent. Each deserves prompt public health attention to address community concerns and to explore possible etiologic clues. However, it is difficult to know when field investigations are needed and how far they should proceed. The problem is neither new nor unique to cancer because similar concerns arise about other chronic or noninfectious diseases. Cases can cluster simply by chance, and random case distributions probably account for most community case clusters. It is possible that certain clusters may have community-based causes, possibly resulting from particular patterns of infectious disease occurrence or from environmental exposures. Although methodologic problems greatly limit the exploration of such possibilities, carefully designed field investigations in selected situations should be considered (1,2).
Reports of such investigations about cancer clusters have appeared in MMWR. The two reports reprinted in this issue illustrate the diversity of analytic approaches that have been used, as well as the uncertainty of conclusions. Both of these investigations involved collaborative efforts between CDC (the Chronic Diseases Division and the Field Services Division) and local/state medical and public health authorities. In the case of acute childhood leukemia in Columbus, Ohio (3), concern was for disease frequency in the city as a whole after eight cases were diagnosed during a 3-month period -- twice the maximum number seen before. However, after further investigation, no features were observed that distinguished these cases from others, and a statistical analysis of time-space closeness indicated no evidence of unusual clustering.
The appearance of Burkitt's tumor in Winchester, Virginia, presented a different problem (4). This particular cancer is rare outside central Africa and New Guinea, and the two initial cases occurred in children whose diagnoses were simultaneous and who lived only two doors apart. The third patient, diagnosed 4 years later, lived nearby. Although no other links were found among the three cases, the rareness of the tumor, the time-space closeness of the first two cases, and the recurrent pattern suggested by the third case required consideration of alternate explanations to that of a chance event.
Interest in cancer case clusters has been evident in the medical literature since the late 19th century. Leukemia has received particular attention, perhaps because clusters suggest infectious disease outbreaks and because white cell elevations are associated with infection. The focus on leukemia intensified in the 1960s when tumor virologists, after demonstrating that leukemia viruses exist in other species, initiated an extensive search for such viruses in humans. During the 1960s and 1970s, CDC played a prominent role in that search by working closely with the National Cancer Institute in field investigations of leukemia case clusters. One of the earliest of these investigations involved eight cases of childhood leukemia in the Chicago suburb of Niles, Illinois (5). Seven of those cases were associated with one particular school where a parallel pattern of rheumatic-like illness had simultaneously appeared. The neighborhood was newly created, a situation later recalled when observations concerning a childhood leukemia cluster in the United Kingdom (UK) suggested that risk might be heightened in newly established communities (6).
Epidemiologic work continued at CDC into the 1980s, with several field investigations conducted each year in cooperation with various local and state health departments (7). Most studies involved time-space clusters of childhood leukemia cases in residential communities; however, some work involved adult leukemias, cancers of other types in both adults and children, multiple-case families, associations between human and animal cancers (i.e., pets and farm animals), and cancers occurring within acquaintance networks (e.g., former school mates). In addition, occasional situations were studied in which case clusters involved congenital malformations or chronic neurologic disease. Over time, emphasis gradually moved away from infectious disease hypotheses and increasingly focused on environmental exposures (e.g., hazardous waste sites, water pollution, and ionizing radiation).
In recent years, interest has been revived about clusters of childhood leukemia in residential communities. Much of the impetus has come from studies in the UK where a cluster of five cases occurred in a small town affecting children of men employed at a nearby nuclear fuel-reprocessing plant. Extensive epidemiologic studies confirmed this local increase in incidence and suggested a possible relation to increased paternal exposure to low levels of ionizing radiation in the workplace before the children were conceived (8).
Although studies in other human and experimental settings have not confirmed this association, nationwide studies of childhood leukemia in the UK have suggested that risk may be increased in newly settled towns, such as the one near the nuclear reprocessing plant. This observation has led to the hypothesis that infectious disease patterns among children in such new towns may be less stable than in more settled communities and may, on occasion, be reflected in unusual patterns of childhood leukemia as a rare sequel to certain infections (6). Again, this concept recalls the earlier Niles, Illinois, experience.
Despite frequent attention over the years to individual cancer case clusters and despite the various hypotheses generated, there has been no instance yet where a biologic cause for clustering has been convincingly demonstrated. In the rare instance when causation has been proven from studies of small groups of cases (9), virtually all have involved rare tumors in occupational settings where exposures were high and where reasonable evidence existed to establish individual exposure levels. The most striking example involved vinyl chloride monomer exposures causing hepatic angiosarcomas in vinyl chloride polymerization workers (10,11). No comparable set of data can be cited for exposures hypothesized in residential settings.
There are clear reasons why community cluster studies are generally inconclusive. When known carcinogens are in question (either chemicals or ionizing radiation), exposures are usually very low. Many situations, however, focus on exposures, such as nonionizing radiation, for which firm evidence of carcinogenicity is lacking at any dose. Studies also are limited by the long and irregular latency of cancer, the clinical nonspecificity of cancer cases (e.g., our inability to assign specific causes to individual cases), and the relative rareness of disease at any one point in time, resulting in very small numbers on which to base epidemiologic analyses. Nonetheless, public health departments must respond in some manner when cancer case clusters come to their attention. The first step is to respond quickly and openly so that communication with community residents is effective and concerns are not neglected. The second step is to confirm the accuracy of reported diagnoses and to compare the number of observed cases with the number to be expected. Often investigations need go no further. If more work is needed, however, the next step is either to interview affected patients or their families in search of common life features or to design and conduct a more formal epidemiologic study, usually a case-control design. Here the work may not be worth the effort unless a reasonable etiologic hypothesis exists or community concern is exceptionally high. The methodologic difficulties become all too apparent, especially the weakness of such studies, limited as they are by very small numbers. Environmental studies also should be approached with caution because of their weakness in establishing clear case-exposure relations and because of their great expense.
Despite these analytic limitations, cancer case clustering deserves research attention. With the possibility of molecular marker techniques, which in the future may help determine the etiologies of individual cancer cases, it may be possible to conduct useful studies of selected clusters, whether the hypotheses are infectious or environmental. Still, the rarity of nonchance clusters must always be considered. Area-wide statistical techniques designed to measure the degree to which cases may cluster in time and space have failed to document any consistent tendency, except perhaps in childhood cancers (12-14). If further knowledge is to develop in this difficult area of public health practice, it is among childhood cancers that work should focus (15). 1997 Editorial Note by Clark W Heath, Jr, MD, Vice President for Epidemiology and Surveillance Research, American Cancer Society, Atlanta, Georgia, and former Director, Division of Chronic Diseases, Center for Disease Control. Glyn G Caldwell, MD, Clinical Coordinator, Indiana Medical Review Organization, Indianapolis, and former Chief, Cancer Branch, Bureau of Epidemiology, Center for Disease Control.
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 firstname.lastname@example.org.
Page converted: 09/19/98
This page last reviewed 5/2/01