Abstract

Abstract Immunodeficiency, be it congenital, therapeutic, or infectious in origin, increases the risk of certain, but not all, types of cancer. A common feature of these cancers is that specific infectious agents appear to be important in their etiology, not only in immunodeficient subjects but also in the general population. People with acquired immunodeficiency syndrome (AIDS) are at an increased risk of Kaposi's sarcoma, non-Hodgkin's lymphoma, Hodgkin's disease, squamous cell carcinoma of the conjunctiva, and childhood leiomyosarcoma. It is striking that most of these cancers have been associated with specific human herpesvirus (HHV) infections: HHV-8 with Kaposi's sarcoma and the closely related Epstein-Barr virus with non-Hodgkin's lymphoma, Hodgkin's disease, and possibly also with childhood leiomyosarcoma. Moreover, similar associations between these viruses and cancer have been found, albeit inconsistently, in people who are not immunosuppressed. Further research is needed to establish whether the risk of other cancers is also increased in people with AIDS, although, if so, the cancers are likely to be rare or to have comparatively small associated relative risks. Existing evidence suggests that there may be no marked increase in the risk of two common cancers that are known to be caused by infectious agents—hepatocellular carcinoma and invasive carcinoma of the uterine cervix. The apparent lack of an increase in invasive cervical cancer is unexpected and needs further investigation, especially since the prevalence of cervical infection with human papillomaviruses and of low-grade preneoplastic changes in the cervical epithelium is increased in women with AIDS. With the prospect of improved survival in people with AIDS, the effect of immunosuppression on cancer is likely to become an increasingly important issue. The study of cancer in immunodeficient populations offers a unique opportunity to investigate the role of the immune system in controlling the development, growth, and dissemination of tumors. Such studies have already contributed substantially to knowledge about the role of infectious agents in human cancer. This article reviews the epidemiologic evidence about the effect of immunosuppression on cancer risk. Immunodeficiency and Cancer: The Evidence Before the Acquired Immunodeficiency Syndrome (AIDS) Epidemic It has long been thought that the immune system plays a vital role in the etiology of cancer. In 1965, Nobel laureate Sir McFarlane Burnet argued that immunosurveillance was a central mechanism by which tumor development was kept in check and predicted that individuals who were immunosuppressed would be at an increased risk of cancer (1). It was already known that children with rare congenital defects of their immune system, such as X-linked gammaglobulinemia or ataxia telangiectasia, were at increased risk of lymphoma, but the number of children with such defects was exceedingly small and so it was not possible to tell whether they were also at an increased risk of other types of cancer (2). Since the 1970s, with the increasing use of immunosuppressive drugs in relation to tissue transplantation, it has been possible to investigate Burnet's hypothesis in detail. Studies of individuals on long-term immunosuppressive drug therapy have shown that such people were at an increased risk of certain, but not all, types of cancer. The most marked increases were for non-Hodgkin's lymphoma (the most commonly reported tumor in most studies of immunosuppressed transplant recipients), Kaposi's sarcoma, hepatocellular carcinoma, and squamous cell carcinoma of the skin, including lip and vulval cancers (3,4). The magnitude of the increase in the relative risk of these tumors was very large indeed. For example, there was about a 100-fold increase in the risk of Kaposi's sarcoma and a 10-fold or greater increase in the risk of other cancers. Some studies (4) reported an increase in Hodgkin's disease, and others (3) reported an increase in cervical cancer, but it is not clear whether this was for premalignant epithelial changes in the cervix and/or for invasive cervical cancer (3). The findings in transplant recipients suggested that immunosuppression led to the selective development of cancers that were caused by infections. The first specific infectious agent implicated as causing cancer in transplant recipients was the Epstein-Barr virus (EBV), which has been consistently identified in transplant-related lymphomas (5). Human herpesvirus (HHV)-8 has been associated with Kaposi's sarcoma, and hepatitis viruses B and C (HBV and HCV, respectively) have been associated with most hepatocellular carcinomas (6). In addition, human papillomavirus (HPV) types 16/18 have been associated with cervical cancer, and HPV types 5/8 appear to be responsible for some skin cancers in the immunosuppressed (7). The clinical course of the cancers that occur in immunosuppressed transplant recipients tends to be more aggressive than in the general population. It has also been noted that the cessation of immunosuppressive therapy can halt or even reverse tumor growth (8). Furthermore, similar risk factors seemed to determine who developed the cancer, irrespective of immune status. For example, the transplant recipients who developed skin cancer tended to be fair skinned and have excessive lifetime exposures to the sun (9). By the late 1970s, before the AIDS epidemic, it was widely believed that immunosuppression led to the selective development of cancers that were caused by infectious agents. However, because the immunodeficient patients tended to be ill for other reasons, it was not always clear whether the cancers that occurred among them were due to the immunodeficiency itself or to the underlying medical conditions, related exposures, or even to the immunosuppressive drugs themselves. For example, the increased risk of hepatocellular carcinoma seen in transplant recipients may have been due to the high prevalence of HBV or HCV infection in this group, and the high rates of cervical cancer may have been due to more frequent screening for that cancer in transplant recipients than in the general population. Cancers Associated With AIDS The AIDS epidemic has provided an unprecedented opportunity to study the effects of immunosuppression on cancer. The fact that large numbers of people throughout the world are infected with the human immunodeficiency virus (HIV) permits the study of the effects of immunosuppression on cancer risk on a scale and in populations that had not been studied before. Furthermore, the reason for the immunodeficiency in people with AIDS was not the same as that for the populations studied previously. As will be discussed below, many of the increases in cancer risk found in people with AIDS are similar to the findings in immunodeficient children and in transplant recipients, suggesting that it is the impairment of immune function, rather than other factors, that is leading to the appearance of these tumors. Cancers Definitely Increased in People With AIDS Although many cancers have been reported to be increased in people with AIDS, for only five cancers is the evidence sufficiently strong and consistent that it is possible to conclude that there is a definite increase in risk. These cancers are listed in Table 1. The evidence for three of those five cancers—Kaposi's sarcoma, non-Hodgkin's lymphoma, and squamous cell carcinoma of the conjunctiva—is well established and has been reviewed in depth elsewhere (6,10). Recent evidence for the other two cancers—Hodgkin's disease and childhood leiomyosarcoma—also suggests a definite increase in risk associated with HIV infection (11-13). The relative risks for HIV-seropositive compared with HIV-seronegative people for the five cancers listed in Table 1 tend to be very large—generally 10-fold or greater. Furthermore, each of those cancers is believed to be caused by a specific infectious agent: HHV-8 for Kaposi's sarcoma, HPV 16/18 for conjunctival cancers (although the evidence for this is not consistent), and EBV for non-Hodgkin's lymphoma and Hodgkin's disease and possibly, also, for leiomyosarcoma in children. Some, but not all, of the cancers listed in Table 1 have been associated with other forms of immunodeficiency. Kaposi's sarcoma and non-Hodgkin's lymphoma have been strongly associated with both immunosuppressive drug therapy and with AIDS. Hodgkin's disease has also been associated with both types of immunosuppression, although the relative risks are not as large as those for the other two cancers (4). Some of the associations found in people with AIDS that have not been reported in transplant recipients can probably be explained by the fact that HIV infection affects a much broader range of people living in different parts of the world. For example, squamous cell carcinoma of the conjunctiva is common in equatorial Africa, where HIV infection is common, but where tissue transplantation is rare. Also, leiomyosarcoma appears to occur exclusively as a rare complication of HIV infection in children, and few children have been given long-term immunosuppressive drug therapy. Other Cancers in People With AIDS The available evidence suggests that people with AIDS are not experiencing large increases in the risk of most types of cancer. Cancer trends for men living in San Francisco, where HIV prevalence is relatively high, show marked increases over time in the incidence of Kaposi's sarcoma and non-Hodgkin's lymphoma only (14). Furthermore, in a case-control study of about 1000 people with all types of cancer in South Africa, the only cancers that were significantly increased in HIV-seropositive people compared with HIV-seronegative people were Kaposi's sarcoma and non-Hodgkin's lymphoma (15). Results broadly similar to those from South Africa were found in a case-control study of 250 people in Rwanda (16). Likewise, record linkage of data from AIDS registries and cancer registries in the United States and in Australia has tended to find increased risks mainly for the cancers listed in Table 1 (11,17). In the Australian study linking information from an AIDS registry to cancer registry data, Grulich et al. (11) reported a relative risk of 5.8 (95% confidence interval [CI] = 1.2-17) for multiple myeloma. There have been a number of case reports of plasmacytomas occurring in people with AIDS (6). It seems likely, therefore, that some form of plasmacytoma may be a rare consequence of HIV-associated immunosuppression, although, at this stage the evidence is not as firm as it is for the five cancers listed in Table 1. Some researchers have linked multiple myeloma to infection with EBV, and others have linked it to infection with HHV-8 (6,18). Perhaps the most fascinating results are the consistent reports that people with AIDS do not appear to be at an increased risk for two common cancers that are known to be caused by infectious agents: invasive cervical cancer and hepatocellular cancer (6). Cervical Cancer Cervical cancer is the most common cancer among women in Africa, and although HIV infection is highly prevalent in central and southern Africa, no epidemic of cervical cancer has been observed. By contrast, there are marked epidemics of AIDS-related Kaposi's sarcoma in these African countries (19,20). Fig. 1 summarizes the relative risk of invasive cervical cancer in HIV-seropositive subjects compared with HIV-seronegative subjects from two published studies (15,16) and preliminary results from another study (21) conducted in three African countries, which together total 363 women with the tumor. The relative risk for invasive cervical cancer is not elevated in any study. The overall relative risk is 0.8 (95% CI = 0.5-1.4), thus arguing against the possibility of a large increase in risk. It should be noted that the relative risk estimates are adjusted by age, but not by sexual practices, and the failure to adjust by sexual practices may even lead to a slight overestimate of the relative risk. Although some of the HIV-seropositive subjects may not be severely immunosuppressed, such a bias would act to reduce the relative risk, but not to eliminate an association altogether. There are very few data about invasive cervical cancer in North America or Europe, chiefly because of the low prevalence of HIV infection in women in those countries, but also because women in the West tend to have regular Pap smears and so preclinical cervical neoplasia is generally treated early. However, the few studies in the West that have looked at this question have also failed to find an increase in the risk of invasive cervical cancer for HIV-infected women (6,22). In contrast to the apparent lack of an increase in invasive cervical cancer, HIV infection is associated with a definite increase in the expression of HPV by cervical cells and in the prevalence of apparently premalignant cervical lesions. Numerous studies have shown markedly higher prevalences of cervical HPV infection, by either single or multiple types, among HIV-infected women. There is also a corresponding increase in the prevalence of low-grade cervical lesions; however, few women with high-grade cervical lesions or with in situ cervical cancer have been found in these studies (6). At this stage, there is no obvious explanation as to why HIV infection appears to increase the risk of low-grade cervical lesions, but not the risk of invasive cervical cancer. It has been suggested that HIV-infected women in Africa may die of other causes before they have time to develop invasive cervical cancer. However, cervical cancer is so common in Africa that there should be many women who already had premalignant cervical lesions and even in situ cancer prior to HIV infection. If immunosuppression associated with HIV infection hastened the clinical course of disease in such women, it is surprising that an epidemic of cervical cancer has not been seen in Africa and that an increase in the risk of invasive cervical cancer has not been found in association with HIV infection. The possibility that there is little or no increase in the risk of invasive cervical cancer among HIV-infected women is worth considering, and, if so, this observation offers an important clue to the pathogenesis of cervical cancer. In women who are chronically infected with HPV, the immune system may play an important role in determining what type of cervical lesion ultimately develops. Low-grade cervical lesions mainly reflect HPV infection, and only a small proportion of women with these lesions go on to develop invasive disease. The HPV infections that occur in association with HIV infection may have little immunologic control, and this possibility may favor the persistence of low-grade cervical lesions that do not progress. By contrast, the typical, chronic HPV infections in women who are not immunosuppressed and have strong immunologic control may favor progression to invasive disease. Hepatocellular Carcinoma It is curious that there appears to be no increase in the risk of hepatocellular carcinoma in people with AIDS. Fig. 2 summarizes the results from two case-control studies in Africa (15,16), preliminary results from a case-control study in Uganda (21), and from one cohort study of patients with hemophilia in the U.K. (23). In no study is there an increase in the relative risk of hepatocellular carcinoma in HIV-seropositive subjects compared with HIV-seronegative subjects. The overall relative risk is 0.8 (95% CI = 0.5-1.5). Hepatocellular carcinoma is sometimes difficult to diagnose in the absence of modern techniques, and it might be argued that this absence contributes to the absence of an association with HIV infection in Africa. Nevertheless, the study of U.K. hemophiliacs found a 17-fold overall increase in the risk of liver cancer (because of the high prevalence of HBV and HCV in hemophiliacs), but no difference in cancer risk between HIV-seropositive subjects and HIV-seronegative subjects, although numbers were small. Also, there was no evidence of a rise in hepatocellular cancer incidence in populations in the United States with a high prevalence of HIV infection (6,14). Therefore, it is worth considering that, as with cervical cancer, the mechanism of hepatic carcinogenesis by HBV and HCV may be dependent on a strong immunologic response by the host or that it takes so long for the cancers to develop that people die of other causes. Other Cancers There may be other cancers whose incidence is increased in association with HIV infection but, if so, they are probably rare and the associated relative risks are not likely to be large. Increases in oral, testicular, skin, brain, lung, breast, and thyroid cancers have been suggested (9). Different researchers have tended to report excesses of different types of cancer, and there was no consistent elevation in the relative risk for any particular tumor. With the prospect of improved survival for HIV-infected people, it will become increasingly important to know more about the risk of cancer in these subjects. Further research is needed to clarify which other tumors are and which are not increased in people with AIDS and, if so, the magnitude of the associated relative risk. In particular, there is a need for further record linkage studies in populations where HIV prevalence is low and for further case-control studies in populations where HIV prevalence is high. In the meantime, because the numbers of other cancers reported in any single study tend to be small, it would be valuable to combine the results from existing studies. Clinical and Biologic Features of the Cancers Ever since the first cases of Kaposi's sarcoma were described in homosexual men, heralding the beginning of the HIV epidemic, it was clear that the clinical characteristics of the disease differed markedly from those of the usually indolent condition, generally affecting the skin of the lower limbs of elderly men, that had been described previously (24). Although identical histologically, HIV-associated Kaposi's sarcoma often presents with multiple lesions affecting both the skin and internal organs, and survival drops from 10-15 years for HIV-seronegative individuals to about 14-18 months for HIV-seropositive individuals (25,26). Such an aggressive presentation has occasionally been seen in HIV-seronegative children in Uganda (27,28). Immunosuppression also affects the ease with which HHV-8 can be detected by polymerase chain reaction (PCR) in the peripheral blood of Kaposi's sarcoma patients. The frequency of detection via PCR increases, while the CD4 count declines (29). HIV-associated non-Hodgkin's lymphomas, although histologically heterogeneous, often exhibit pleomorphic features encompassing a range of overlapping subtypes (including Burkitt's lymphoma) and can be characterized by a number of features, including an aggressive clinical course. High-grade disease is common and extranodal sites are often involved, with lesions in the central nervous system being virtually unknown, except in the immunosuppressed. Similarly, Hodgkin's disease among HIV-infected individuals is clinically unusual, generally presenting at a late stage with extranodal dissemination common. The predominant histologic subtypes are mixed cellularity and lymphocyte-depleted, which are relatively rare in the HIV-uninfected population. Whether these differences in the presentation and the clinical course between HIV-infected and uninfected individuals reflect differences in disease etiology or the host response is unclear, although the latter may be more likely. Certainly, infected individuals have a higher viral load of EBV when they are immunosuppressed and possibly also of HHV-8, although the evidence is less clear (29,30). Other Risk Factors An individual's risk of developing a specific cancer is affected by various factors, including age, sex, and exposure to environmental factors. As yet, there is only limited information about the characteristics and exposures of HIV-infected people who develop cancer, but such people tend to have characteristics and exposures similar to those of HIV-uninfected people with a similar cancer. The main determinant of Kaposi's sarcoma risk is infection with HHV-8, and this applies both to people with and without AIDS (31). More is known about the personal characteristics of people with AIDS-related Kaposi's sarcoma than of people without AIDS. The main behavioral risk factor for AIDS-related Kaposi's sarcoma in western populations is sex between men (32). This fact suggests that HHV-8 may be transmitted by sexual contact between men, and there is now some evidence that this is so (33,34). Whether sex between men is also a risk factor for HIV-negative Kaposi's sarcoma is unknown. Certainly, Kaposi's sarcoma was far more common in men than in women before the AIDS epidemic; thus, it is possible that HHV-8 was transmitted sexually between men in the past, even though its prevalence in most countries was probably low. The recent spread of HIV in the population may have led to an increase in the spread of HHV-8 infection (because immunosuppression may facilitate transmission of HHV-8), resulting in an increase in Kaposi's sarcoma even in people who are not infected by HIV. For example, Kaposi's sarcoma is being increasingly reported in HIV-seronegative homosexual men in New York and in HIV-seronegative children in Africa (6,27). Sex between men is not the only behavioral risk factor for Kaposi's sarcoma, since the disease also occurs in women and in children. The incidence of Kaposi's sarcoma varied markedly across the world before the AIDS epidemic and was particularly frequent in Africa and infrequent in the U.K. (Table 2). The geographic variation in Kaposi's sarcoma before the AIDS epidemic reflects the geographic variation in the prevalence of HHV-8 infection and is similar to the geographic variation in the proportion of women with AIDS who also had Kaposi's sarcoma (Table 2) (31,32,35-37). This variation suggests that the main determinant of Kaposi's sarcoma risk in different countries is the prevalence of HHV-8 infection. Burkitt's lymphoma is associated with EBV infection both in people with and without AIDS. Burkitt's lymphoma is also far more common in males than in females and shows a characteristic peak at age 10-19 years, both in people with AIDS and in the general population (38). EBV infection has also been associated with Hodgkin's disease, both in HIV-seropositive subjects and in HIV-seronegative subjects (39,40). For non-Burkitt's, non-Hodgkin's lymphomas, the proportion of people with AIDS who have these lymphomas increases with age, is greater in whites than in blacks, and is greater in men than in women, similar to the pattern of risk seen in the general population (38). EBV has been found in about half the AIDS-related non-Burkitt's non-Hodgkin's lymphomas and in virtually all the AIDS-related lymphomas of the central nervous system (5). EBV has also been found, although not consistently, in subjects with non-Burkitt's non-Hodgkin's lymphoma in the general population. Although the evidence to date is not strong, it appears that cancers that occur in the immunosuppressed population have a similar etiology to the cancers that occur in the general population. Infectious agents are consistently found in a large proportion of tumors in the immunosuppressed subjects but generally in a smaller proportion of similar tumors in immunocompetent people. The question is whether the same infectious agents may be involved in the etiology of the same type of cancers both in the immunosuppressed and in the general population. Present diagnostic techniques are not sensitive enough to identify the infectious agents in a large group of immunocompetent subjects. Conclusions Studies of cancer in people with AIDS have provided unique opportunities to examine the role of the immune system in human cancer. It is clear that immunodeficiency, whether congenital, drug-induced, or related to HIV infection, increases the risk of certain, but not of all, types of cancers. There is strong and consistent evidence that the immunosuppression associated with HIV infection increases the risk of Kaposi's sarcoma, non-Hodgkin's lymphoma, Hodgkin's disease, squamous cell carcinoma of the conjunctiva, and leiomyosarcoma in children. Most of these cancers are known to be caused by specific herpesviruses. The individuals who develop immunodeficiency-related cancers seem to have characteristics similar to those of immunocompetent individuals who develop the same type of cancer. Most other cancers show no increase in risk associated with HIV infection, although relatively small increases for rare tumors cannot be excluded. The available evidence suggests that invasive cervical cancer and hepatocellular cancer, both of which are known to be caused by infectious agents, do not appear to be increased markedly in people with AIDS. Understanding why immunosuppression increases the risk of certain, but not all, cancers that are known to be caused by infectious agents may lead to important insights into the carcinogenic process. In addition, understanding why certain viruses can be found in association with tumors in HIV-seropositive subjects, but not in similar tumors in HIV-seronegative subjects, may aid our understanding of the role of these infections in the etiology of cancer in the general population. With the prospect of improved survival in HIV-infected people, it is difficult to predict what will happen to cancer incidence in that population. Improved survival should, nevertheless, focus attention on the risk of HIV-associated cancer and may lead to further understanding of the role of infectious agents and the immune system in cancer. Table 1. Cancers that are definitely increased among human immunodeficiency virus (HIV)-infected people Cancer Approximate relative risk in HIV-seropositive individuals Infectious agent* Kaposi's sarcoma 10 000 HHV-8 Non-Hodgkin's lymphoma 50 EBV Hodgkin's disease 10 EBV Squamous cell carcinoma  of the conjunctiva 10 Possibly HPV 16/18 Childhood leiomyosarcoma Unclear Possibly EBV Cancer Approximate relative risk in HIV-seropositive individuals Infectious agent* Kaposi's sarcoma 10 000 HHV-8 Non-Hodgkin's lymphoma 50 EBV Hodgkin's disease 10 EBV Squamous cell carcinoma  of the conjunctiva 10 Possibly HPV 16/18 Childhood leiomyosarcoma Unclear Possibly EBV * HHV = human herpesvirus; EBV = Epstein-Barr virus; HPV = human papillomavirus. Open in new tab Table 1. Cancers that are definitely increased among human immunodeficiency virus (HIV)-infected people Cancer Approximate relative risk in HIV-seropositive individuals Infectious agent* Kaposi's sarcoma 10 000 HHV-8 Non-Hodgkin's lymphoma 50 EBV Hodgkin's disease 10 EBV Squamous cell carcinoma  of the conjunctiva 10 Possibly HPV 16/18 Childhood leiomyosarcoma Unclear Possibly EBV Cancer Approximate relative risk in HIV-seropositive individuals Infectious agent* Kaposi's sarcoma 10 000 HHV-8 Non-Hodgkin's lymphoma 50 EBV Hodgkin's disease 10 EBV Squamous cell carcinoma  of the conjunctiva 10 Possibly HPV 16/18 Childhood leiomyosarcoma Unclear Possibly EBV * HHV = human herpesvirus; EBV = Epstein-Barr virus; HPV = human papillomavirus. Open in new tab Table 2. Geographic variation in the incidence of Kaposi's sarcoma before the acquired immunodeficiency syndrome (AIDS) epidemic, prevalence of human herpesvirus-8 (HHV-8) infection, and proportion of women with AIDS who have Kaposi's sarcoma* Region or country Kaposi's sarcoma incidence per million before the AIDS epidemic Approximate prevalence of HHV-8 in human immunodeficiency virus-seronegative individuals, % Proportion of women with AIDS who also have Kaposi's sarcoma, % U.K.  0.1 0 0 United States  1.8 0 1 Italy  6.6 4 2 Uganda 50   50 10 Region or country Kaposi's sarcoma incidence per million before the AIDS epidemic Approximate prevalence of HHV-8 in human immunodeficiency virus-seronegative individuals, % Proportion of women with AIDS who also have Kaposi's sarcoma, % U.K.  0.1 0 0 United States  1.8 0 1 Italy  6.6 4 2 Uganda 50   50 10 * See text for references. Open in new tab Table 2. Geographic variation in the incidence of Kaposi's sarcoma before the acquired immunodeficiency syndrome (AIDS) epidemic, prevalence of human herpesvirus-8 (HHV-8) infection, and proportion of women with AIDS who have Kaposi's sarcoma* Region or country Kaposi's sarcoma incidence per million before the AIDS epidemic Approximate prevalence of HHV-8 in human immunodeficiency virus-seronegative individuals, % Proportion of women with AIDS who also have Kaposi's sarcoma, % U.K.  0.1 0 0 United States  1.8 0 1 Italy  6.6 4 2 Uganda 50   50 10 Region or country Kaposi's sarcoma incidence per million before the AIDS epidemic Approximate prevalence of HHV-8 in human immunodeficiency virus-seronegative individuals, % Proportion of women with AIDS who also have Kaposi's sarcoma, % U.K.  0.1 0 0 United States  1.8 0 1 Italy  6.6 4 2 Uganda 50   50 10 * See text for references. Open in new tab Fig. 1. Open in new tabDownload slide Summary of results from studies investigating the risk of invasive cervical cancer in human immunodeficiency virus (HIV)-seropositive subjects compared with HIV-seronegative subjects. CI = confidence interval, * = preliminary results, and + = relative risk in HIV-seropositive subjects compared with HIV-seronegative subjects. Fig. 2. Open in new tabDownload slide Summary of results from studies investigating the risk of hepatocellular carcinoma in human immunodeficiency virus (HIV)-seropositive subjects compared with HIV-seronegative subjects. CI = confidence interval, * = preliminary results, and + = relative risk in HIV-seropositive subjects compared with HIV-seronegative subjects. This research was supported by the Imperial Cancer Research Fund, U.K. References (1) Burnet M. Somatic mutation and chronic disease. Br Med J 1965 ; 1 : 338 -42. (2) Mueller BU, Pizzo PA. Cancer in children with primary or secondary immunodeficiencies. J Pediatr 1995 ; 126 : 1 -10. (3) Kinlen LJ, Sheil AG, Peto J, Doll R. Collaborative United Kingdom-Australasian study of cancer in patients treated with immunosuppressive drugs. Br Med J 1979 ; 2 : 1461 -6. (4) Birkeland SA, Storm HH, Lamm LU, Barlow L, Blohme I, Forsberg B, et al. Cancer risk after renal transplantation in the Nordic countries, 1964-1986. Int J Cancer 1995 ; 60 : 183 -9. (5) Thomas JA, Crawford DH. Epstein-Barr virus associated B-cell lymphomas in AIDS and after organ transplantation. Lancet 1989 ; 1 : 1075 -6. (6) IARC monograph on the evaluation of carcinogenic risks to humans. Volume 67, Human immunodeficiency viruses and human T-cell lymphotropic viruses. Lyon, France: IARC, 1996. (7) Barr BB, Benton EC, McLaren K, Bunney MH, Smith IW, Blessing K, et al. Human papilloma virus infection and skin cancer in renal allograft recipients. Lancet 1989 ; 1 : 124 -9. (8) Starzl TE, Naleshik MA, Porter KA, Ho M, Iwatsuki S, Griffith BP, et al. Reversibility of lymphomas and lymphoproliferative lesions developing under cyclosporin-steroid therapy. Lancet 1984 ; 1 : 583 -7. (9) Bavinck JN, De Boer A, Vermeer BJ, Hartevelt MM, van der Woude FJ, Claas FH, et al. Sunlight, keratotic skin lesions and skin cancer in renal transplant recipients. Br J Dermatol 1993 ; 129 : 242 -9. (10) Newton R. A review of the aetiology of squamous cell carcinoma of the conjunctiva. Br J Cancer 1996 ; 74 : 1511 -3. (11) Grulich A, Wan X, Law M, Coates M, Kaldor J. Rates of non-AIDS defining cancers in people with AIDS. J Acquir Immune Defic Syndr Hum Retrovirol 1997 ; 14 : A18 . (12) Rabkin C. Epidemiology of malignancies other than Kaposi's sarcoma and non-Hodgkin's lymphoma in HIV infection. J Acquir Immune Defic Syndr Hum Retrovirol 1997 ; 14 : A12 . (13) Pollock BH, Jenson HB, McClain KL, Leach CT, Joshi VV, Murphy SB. Risk factors for HIV-related malignancies in children. J Acquir Immune Defic Syndr Hum Retrovirol 1997 ; 14 : A17 . (14) Rabkin CS, Biggar RJ, Horm JW. Increasing incidence of cancers associated with the human immunodeficiency virus epidemic. Int J Cancer 1991 ; 47 : 692 -6. (15) Sitas F, Bezwoda WR, Levin V, Ruff P, Kew MC, Hale MJ, et al. Association between human immunodeficiency virus type 1 infection and cancer in the black population of Johannesburg and Soweto, South Africa. Br J Cancer 1997 ; 75 : 1704 -7. (16) Newton R, Grulich A, Beral V, Sindikubwabo B, Ngilimana PJ, Nganyira A, et al. Cancer and HIV infection in Rwanda [letter]. Lancet 1995 ; 345 : 1378 -9. (17) Cote TR, O'Brien TR, Ward JW, Wilson SE, Blattner WA. AIDS and cancer registry linkage: measurement and enhancement of registry completeness. The National AIDS/Cancer Match Study Group. Prev Med 1995 ; 24 : 375 -7. (18) Rettig MB, Ma HJ, Vescio RA, Pold M, Schiller G, Belson D, et al. Kaposi's sarcoma-associated herpesvirus infection of bone marrow dendritic cells from multiple myeloma patients. Science 1997 ; 276 : 1851 -4. (19) Wabinga HR, Parkin DM, Wabwire-Mangen F, Mugerwa JW. Cancer in Kampala, Uganda, in 1989-91: changes in incidence in the era of AIDS. Int J Cancer 1993 ; 54 : 26 -36. (20) Bassett MT, Chokunonga E, Mauchaza B, Levy L, Ferlay J, Parkin DM. Cancer in the African population of Harare, Zimbabwe, 1990-1992. Int J Cancer 1995 ; 63 : 29 -36. (21) Newton R, Ziegler JL, Beral V, on behalf of the Uganda Kaposi's sarcoma study group. A review of the aetiology of squamous cell carcinoma of the conjunctiva: with preliminary results of a case-control study from Kampala, Uganda. J Acquir Immune Defic Syndr Hum Retrovirol 1997 ; 14 ; A17 . (22) Rabkin CS, Biggar RJ, Baptiste MS, Abe T, Kohler BA, Nasca PC. Cancer incidence trends in women at high risk of human immunodeficiency virus (HIV) infection. Int J Cancer 1993 ; 55 : 208 -12. (23) Darby SC, Ewart DW, Giangrande PL, Dolin PJ, Spooner RJ, Rizza CR. Mortality before and after HIV infection in the complete UK population of haemophiliacs. U.K. Haemophilia Centre Directors' Organisation. Nature 1995 ; 377 : 79 -82. (24) Dorffel J. Histogenesis of multiple idiopathic hemorrhagic sarcoma of Kaposi. Arch Dermatol Syph 1932 ; 26 : 608 -34. (25) Tappero JW, Conant MA, Wolfe SF, Berger TG. Kaposi's sarcoma. Epidemiology, pathogenesis, histology, clinical spectrum, staging criteria and therapy. J Am Acad Dermatol 1993 ; 28 : 371 -95. (26) Casabona J, Salas T, Salinas R. Trends and survival in AIDS-associated malignancies. Eur J Cancer 1993 ; 29A : 877 -81. (27) Ziegler JL, Katongole-Mbidde E. Kaposi's sarcoma in childhood: an analysis of 100 cases from Uganda and relationship to HIV infection. Int J Cancer 1996 ; 65 : 200 -3. (28) Mbulaiteye SM, Ziegler JL, Katongole-Mbidde E, Tindyebwa D, Marrum L, Newton R, et al. Kaposi's sarcoma study group. Risk factors for childhood Kaposi's sarcoma in Uganda: a case-control study. Abstract 7. J Acquir Immune Defic Syndr Hum Retrovirol 1997 ; 14 : A18 . (29) Whitby D, Howard MR, Tenant-Flowers M, Brink NS, Copas A, Boshoff C, et al. Detection of Kaposi sarcoma associated herpesvirus in peripheral blood of HIV-infected individuals and progression to Kaposi's sarcoma. Lancet 1995 ; 346 : 799 -802. (30) Rickinson A. The role of herpesviruses in immune deficiency-associated lymphomas. Abstract S17. J Acquir Immune Defic Syndr Hum Retrovirol 1997 ; 14 : A15 . (31) Gao SJ, Kingsley L, Li M, Zheng W, Parravicini C, Ziegler J, et al. KSHV antibodies among Americans, Italians and Ugandans with and without Kaposi's sarcoma. Nat Med 1996 ; 2 : 925 -8. (32) Beral V. Epidemiology of Kaposi's sarcoma. Cancer Surv 1991 ; 10 : 5 -22. (33) Melbye M, Cook PM, Hjalgrim H, Begtrup K, Simpson GR, Biggar RJ, et al. Risk factors for HHV-8 seropositivity and progression to Kaposi's sarcoma in a cohort of homosexual men, 1981-96. J Acquir Immune Defic Syndr Hum Retrovirol 1997 ; 14 : A16 . (34) Grulich A, Olsen S, Hendry O, Luo K, Cooper D, Zheng W, et al. Route of transmission of Kaposi's sarcoma associated herpesvirus. J Acquir Immune Defic Syndr Hum Retrovirol 1997 ; 14 : A16 . (35) Geddes M, Franceschi S, Barchielli A, Falcini F, Carli S, Cocconi G, et al. Kaposi's sarcoma in Italy before and after the AIDS epidemic. Br J Cancer 1994 ; 69 : 333 -6. (36) Templeton AC, Buxton E, Bianchi A. Cancer in Kyadondo County, Uganda, 1968-70. J Natl Cancer Inst 1972 ; 48 : 865 -74. (37) Simpson GR, Schulz TF, Whitby D, Cook PM, Boshoff C, Rainbow L, et al. Prevalence of Kaposi's sarcoma associated herpesvirus infection measured by antibodies to recombinant capsid protein and latent immunofluorescence antigen. Lancet 1996 ; 348 : 1133 -8. (38) Beral V, Peterman T, Berkelman R, Jaffe H. AIDS-associated non-Hodgkin lymphoma. Lancet 1991 ; 337 : 805 -9. (39) Moran CA, Tuur S, Angritt P, Reid AH, O'Leary TJ. Epstein-Barr virus in Hodgkin's disease from patients with human immunodeficiency virus infection. Mod Pathol 1992 ; 5 : 85 -8. (40) Levine PH, Pallesen G, Ebbesen P, Harris N, Evans AS, Mueller N. Evaluation of Epstein-Barr virus antibody patterns and detection of viral markers in the biopsies of patients with Hodgkin's disease. Int J Cancer 1994 ; 59 : 48 -50. Oxford University Press Oxford University Press