Malignant Lymphomas of African Children


Clinical, pathologic, and epidemiologic observations of malignant tumors of children in sub-Sahara Africa suggest alternative theories of causation, and give insight into environmental influences that may play a large role in the etiology and form of malignant lymphatic tumors and cerebral neoplasms of infants. Malignant lymphomas account for nearly half of the cancers of young children, but there are striking geographic differences among the several types. In northern hemispheres the malignant lymphomas are commonly lymphoblastic leukemias, but in central, sub-Sahara Africa and in New Guinea they are almost always solid, multicentric tumors, the so-called Burkitt type of lymphoma. In Brazil, and probably in South Africa as well, the ratios are intermediate, and in both areas hybrid forms have been conspicuous. Indeed, the pattern of Burkitt tumor in North America is notably different from that in Africa as, for example, in the frequency of leukemia and of jaw tumors. The Burkitt tumor has been intensively investigated (1) during the past dozen years, largely because its distribution in Africa resembles the geographic pattern of arbovirus infections; these studies led to the suggestion that the tumor might be caused by a mosquito-borne virus (2). A miscellany of agents have since been recovered from African tumors and one, a herpes virus, is still being exhaustively studied because of its intimate association with the tumor. Whatever the role of this virus may prove to be, it is now quite evident that it is just as intimately associated with various other diseases, which include infectious mononucleosis, cancers of the postnasal space, and Hodgkin's disease. The elevated antibody levels to the virus, which attracted much attention when they were first associated with the Burkitt tumor, are also present in these and other circumstances; elevated antibody levels are now thought to be an expression of impaired cellular immunity, rather than indicative of a casual relationship (3). Furthermore, the Epstein-Barr (EB) virus is distributed worldwide and, therefore, offers no explanation of the uneven distribution of the Burkitt tumors. In addition there is little in the known natural history of EB virus infection to distinguish it from other herpes viruses, especially the cytomegalo viruses and virus III, which is a passenger of the Brown-Pearce tumor of rabbits. On the other hand, a herpes virus is believed to be the cause of a lymphatic malignancy in primates, and herpes viruses have been related to cancers of the cervix and uterus; continued exploration of the role of all these agents is surely in order. 15 An environmental factor that plays a significant role in determining the geographic pattern of the lymphomas is tropical malaria-holoor hyper-endemic malaria. The association of malaria of this kind, in which repeated infections during the first year of life induce a severe and uninterrupted struggle with the parasite was originally recognized epidemiologically (4) and has been repeatedly confirmed. Initially, it was simply postulated that the pre-existing malaria transformed acute leukemia into a restricted, localized, subacute, or chronic disease. Both processes, the malaria and the leukemia, compete within the reticuloendothelial system; it was presumed that the leukemia cell is greatly disadvantaged by the malaria and survives and grows only in distant loci, such as the jaw bones and ovaries, the favorite sites of the African tumors. This view of pathogenesis also provided an explanation of the somewhat later peak-age incidence of the Burkitt tumor and its greater response to chemotherapy. The concept was further reinforced by the observation that myeloid leukemia at times behaves in a more chronic fashion in Africa than in European children, and that chloromatous deposits are more frequently seen in Africa than elsewhere (5). If the Burkitt tumor bears such a relationship to acute, lymphoblastic leukemia, then the reciprocal ratios of the two are easily understood and the lessons of the Oxford survey, which revealed that childhood leukemia is initiated during gestation and that its prolonged latency is characterized by extreme susceptibility to various infections (6), may be pertinent to the Burkitt tumor as well. The first of these two observations, the origin during gestation, is critical to the proposition that the Burkitt lymphoma is caused by a postnatal virus infection, such as that associated with EB virus. The second observation is a reminder of deficiencies in the developing countries of adequate vital statistics. We cannot measure the mortality among the young that is due to holoendemic malaria, or measure the relationship between deaths during infancy that are caused by malaria and by leukemia. We found no answers to these questions during a recent survey that spanned central Africa, but we did benefit from the firmly established and expertly directed cancer registries in Ibadan, Kampala, and Nairobi. Selected, combined notifications from these three registries are shown in Table 1 in parallel with the 1962-1965 notifications from England and Wales. Table 1 (which assumes the population at risk is unknown for the African series) recognizes two types of lymphatic neoplasm, two types of solid tumors, and three age 16 Medical Sciences: Stewart et al. TABLE 1. The relative frequency of different types of cancer in African and European children Age at notification, years 0-4 5-9 10-14 All ages Diagnosis A B A B A B A B Leukemia 23 671 37 375 38 258 98 1304 (7) (19) (29) (15) Lymphomas 131 137 493 156 288 221 912 514 (186) (614) (253) (345) Neural 36 629 27 404 16 314 75 1347 tumors (10) (13) (10) (11) Other 302 640 142 231 185 352 629 1223 cancers (92) (119) (102) (100) Leukemia 154 808 530 531 326 479 101


    1 Figures and Tables

    Download Full PDF Version (Non-Commercial Use)