More than 170 000 men have been diagnosed with AIDS in Europe , and as a consequence of the HIV epidemic the incidence of many types of cancers is likely to increase. The existing evidence, however, only suffices to support the indirect causal role of HIV in Kaposi's sarcoma (KS) and non-Hodgkin's lymphoma (NHL) . In addition to KS and NHL, a nearly 10-fold increased risk of Hodgkin's disease (HD) in HIV-infected individuals has been consistently reported [2–5]. For a few other neoplasms, such as anal cancer , oropharyngeal cancer, lung cancer, testicular cancer, non-melanoma skin cancer and cancer of the penis, there have been reports of a weak association with HIV infection, whereas for hepatocellular cancer no evidence of an elevated incidence in HIV-positive individuals has emerged .
Overall, infections are estimated to account for approximately 9% of all cancer cases diagnosed in developed countries (0.5% of these cases are attributable to HIV-associated KS and NHL) . The investigation of the relationship between infectious agents, suppression of the immune system and cancer occurrence is of particular interest now that so many people are living with HIV infection. The present study was thus conducted to help elucidate the full spectrum of cancers in men with, or at risk of, HIV infection in southern Europe.
Materials and methods
In this investigation, longitudinal data from the DMI-2 HIV clinical database, South-eastern France, the Italian HIV Seroconversion Study (ISS), and the clinical database of former intravenous drug users (IDU) admitted to the San Patrignano Community, Northern Italy, were combined. The aim was to assess cancer risk among men with, or at risk of, HIV infection. The analysis focused on men of 20–49 years of age who were IDU or who acquired HIV infection through homosexual or heterosexual intercourse. A detailed description of the study methodology has already been published .
The DMI-2 HIV clinical database
The DMI-2 is a national multicentric longitudinal database of all HIV-positive individuals who have access to hospital care in France . In this study, 2976 men followed between 1 January 1988 and 31 January 1998 at Nice University Hospital were evaluated. At baseline, information collected included sociodemographic characteristics, HIV exposure category, date of first HIV positive test (when available) and clinical conditions. Follow-up visits were scheduled at least every 6 months. Person-years at risk of cancer were computed from the first visit until the date of death, cancer diagnosis, or the last follow-up. Cancer diagnoses were histologically confirmed by the Pathology Department of Nice University Hospital or the collaborating centres. For patients who missed the follow-up visits for more than 1 year, information on AIDS and on vital status was actively elicited either from the French National AIDS registry, clinical records, or the census bureau of the town of residence.
The ISS cohort
A total of 1062 men who had had a documented HIV seronegative test, followed by a positive, confirmed test, were included in this investigation . The maximum accepted lag between the two HIV tests was 2 years, and the midpoint between the last HIV negative test result and the first HIV positive test result was used to estimate the seroconversion timepoint. Person-years at risk of cancer were computed from the seroconversion date to the date of death, cancer diagnosis, or a cut-off date (i.e. 30 June 1997). Pathologists from the collaborating centres histologically confirmed cancer diagnoses. To reduce follow-up losses, a linkage with the Italian National AIDS Registry was carried on for AIDS-free individuals. Further information on vital status was also obtained through the census bureau of the town of residence.
The San Patrignano Community clinical database
A total of 4347 formerly IDU men, admitted between 1982 and 1997 to the San Patrignano Community, Rimini, North Italy, (a residential institution for the cessation of drug use) constitute the third group under study. These men injected heroin for at least one year, and spent at least 6 months in the San Patrignano Community. Person-years at risk of developing cancer were computed from the date of admission to San Patrignano Community to the date of death, cancer diagnosis, the date of leaving San Patrignano Community, or to 30 September 1997. Collaborating pathologists from the Medical Centre of San Patrignano Community (mostly at Rimini General Hospital) histologically confirmed all cancer diagnoses.
To avoid the inclusion of prevalent cancer cases, men with a cancer diagnosis made within 2 months of enrolment were not considered. Therefore, 27 men with KS, 11 men with NHL, one man with prostate cancer, and one man with gastric cancer were excluded from this study.
In the DMI-2 database and the ISS cohort, HIV-positive men were classified either as IDU, homosexuals or heterosexuals according to the criteria adopted by the European Centre for the Epidemiological Monitoring of AIDS .
A person-year analysis was conducted: for each type of cancer observed, the expected number was computed based on incidence rates registered, between 1988 and 1992, among men aged 20–49 years, in the population-based cancer registries of Italy or France . Data were derived from all of Italy's 13 and all of France's eight population-based cancer registries, because men enrolled in these three studies came from several different Italian or French areas, not all covered by cancer registration. The average annual, age-standardized incidence rates were multiplied by the number of person-years of observation to obtain the expected number of cancer cases. Skin cancers were not considered in the computation of the total number of expected cancers. Standardized incidence ratios (SIR) were obtained by dividing the observed number of cases by the expected number: the 95% confidence intervals (CI) of these SIR were determined using the Poisson distribution for the observed cases .
Selected characteristics at baseline of the 8385 men included in this analysis are listed in Table 1. The majority of the 5281 HIV-positive men were IDU (64.8%), 25.8% were homosexual and 9.4% heterosexual. The median age was higher among HIV-positive men enrolled at Nice University Hospital (31.7 years) than in HIV-positive men enrolled in the two Italian studies, and in HIV-negative IDU admitted at the San Patrignano Community (26.6 years). At enrolment, no AIDS cases were registered among men in the ISS, nor among those admitted at the San Patrignano Community. On the contrary, 12.6% of HIV-positive men enrolled in the DMI-2 study at Nice University Hospital had AIDS at baseline (Table 1). The 5281 HIV-positive men were followed-up for 19 609 person-years: men in the HIV seroincident cohort (i.e. the ISS) had a higher median follow-up time (i.e. 7.1 years) than men in the two HIV seroprevalent cohorts. The 3104 HIV-negative IDU were followed-up for 7957 person-years (Table 1). During the follow-up period, 956 deaths were observed among HIV-positive men (48.8 per 1000 person-years) and 11 deaths (1.4 per 1000 person-years) occurred among HIV-negative IDU.
Three hundred and sixty-three AIDS-associated cancers and 25 non-AIDS-associated cancers were diagnosed among HIV-positive men; eight cancers occurred among HIV-negative IDU. SIR according to cancer type, or site, are listed in Table 2.
Statistically significant elevated SIR were recorded, among HIV-positive men, for KS (SIR = 689), NHL (SIR = 103), HD (SIR = 8.7), hepatocellular carcinoma (two cases observed, SIR = 11.0), and cancer of the salivary glands (two cases observed, SIR = 33.6). A non-statistically significant threefold higher risk was recorded for lung cancer, whereas the overall SIR for non-AIDS-defining cancers was 1.9 (95% CI: 1.2–2.8) (Table 2).
HIV-negative IDU admitted to the San Patrignano Community had, overall, twice the risk of cancer (SIR = 2.3, 95% CI: 1.0–4.5) than men in the general population of Italy (Table 2). This excess risk was not associated with KS or NHL, but it was largely attributable to an increased risk of lung cancer (SIR = 14.3, 95% CI: 1.4–52.8) and for HD (Table 2).
Among HIV-infected men, SIR were also separately computed for IDU and homosexuals (because of small numbers, heterosexuals were excluded from this analysis) (Table 3). As expected, KS risk was higher among homosexuals (SIR = 2055) than among IDU (SIR = 147). Significantly increased risks for HD were recorded in both groups (SIR = 8.8 in IDU; SIR = 11.2 in homosexuals), whereas lung cancer (SIR = 6.2) and hepatocellular carcinoma (SIR = 24.3) risks significantly elevated only among IDU. For all non-AIDS-associated cancers, a SIR of 2.4 (95% CI: 1.4–3.9) was noted in IDU, and a SIR of 1.2 (95% CI: 0.4–2.6) in homosexuals (Table 3).
Some main original results can be summarized from this pool of longitudinal data on cancer risk among men with, or at risk of, HIV infection in Italy and France. Primarily, the nearly twofold overall excess risk of non-AIDS-defining cancers for HIV-positive men and for HIV-negative IDU men; second, the consistent excess risk of HD; third, the suggestion that the risk of hepatocellular carcinoma and carcinoma of the salivary glands may also increase in men with HIV infection; and, finally, the suggestion that lung cancer risk may be enhanced in IDU men regardless of HIV infection.
Before comparing these results with existing evidence on the spectrum of HIV-associated cancers, it is worth outlining some general limitations and strengths of the present research. As already illustrated by other authors [14,15], investigations based on HIV seroprevalent cohorts or on HIV seroincident cohorts like those used in this study have various types of biases. Because of selection bias, the IDU admitted to the San Patrignano Community and those enrolled in the ISS or in the DMI-2 studies cannot be considered representative of the IDU in Italy or France. This applies equally to heterosexual and homosexual men. Another limiting factor in the present investigation, as in other longitudinal investigations of HIV-associated cancers , was the small number of observed cases of many cancer types, particularly when SIR were stratified according to HIV exposure category. The different length of time of HIV infection in the HIV seroincident cohort and in the two HIV seroprevalent cohorts could have biased the estimation of cancer risk. However, no differences emerged when the overall risk of non-AIDS-defining cancers in IDU and homosexual men was separately assessed in each of the three cohorts. Among HIV-infected men the overall mortality was nearly 35-fold higher than among HIV-negative men, and this is likely to have led to an underestimation of the real cancer risk associated with HIV infection. Finally, the number of expected cancer cases was computed using incidence rates representative of the general populations of Italy or France , because men enrolled in these three studies often lived in areas of Italy or France not covered by cancer registration schemes.
On the other hand, the use of three different data sources for this longitudinal investigation allowed for cancer risk analysis in more than 5000 HIV-positive men followed for approximately 20 000 person-years in a well-defined geographical area (i.e. southern Europe), in which such an extensive investigation on HIV and cancer has never been conducted. In this regard, it is noteworthy that results were consistent across the three different data sources.
In addition to the well-known high risk of KS (we registered an excess of nearly 2000 times among homosexuals and of nearly 150 times among IDU) and of NHL (approximately 100-fold), HIV-infected men in this study had twice the risk of non-AIDS-associated cancers than men in the general population of Italy or France. This magnitude in the excess of non-AIDS-associated cancers among HIV-infected individuals confirms previous reports from the United States  and Australia , while adding new elements. In fact, the present study assessed risk only among men, and analysed IDU and homosexuals separately. Moreover, a twofold cancer risk excess was also registered among HIV-negative IDU, a population group for which cancer risk has rarely been assessed. Among HIV-positive men, few types of neoplasms contributed to enhance the risk of cancer, namely HD, hepatocellular carcinoma, carcinoma of the salivary glands, and lung cancer.
The nearly ninefold excess of HD that emerged from this study concurs with the excess (between five- and 10-fold) consistently reported among HIV-infected individuals by longitudinal [3,17] and record-linkage [4,5] investigations. This study is the first to show that HD risk is significantly elevated in both HIV-positive IDU and homosexual men. In addition to epidemiological findings, recent evidence suggests that HD risk is greater in people with advanced immunodeficiency , thus strengthening the biological plausibility of the association between HIV and HD. Infection with Epstein–Barr virus has been hypothesized to play a role in the aetiology of HD in HIV-infected individuals , although Epstein–Barr virus also appears to favour the occurrence of HD in the immunocompetent host . Our finding of a non-statistically significant excess risk of HD among HIV-negative IDU seems to concur with this hypothesis.
Because infection with hepatitis C virus is one of the primary causes of hepatocellular carcinoma, an increase in hepatocellular carcinoma in people with HIV infection is plausible. However, no increase in the incidence of hepatocellular carcinoma has been reported so far in the United States , nor in HIV-infected individuals in Europe [4,21], or in Australia . Two cases of hepatocellular carcinoma were diagnosed among the 5281 HIV-positive men followed-up in this study (both were recorded among French IDU co-infected with hepatitis C virus) against 0.22 expected diagnoses, thus our observation needs to be confirmed through investigations with longer follow-up periods.
Some clinical reports suggest that lung cancer is more frequent among HIV-positive individuals , but evidence linking HIV infection and lung cancer is inconsistent . A slight increase in lung cancer was noted in 1994 among never-married men in the United States , and, more recently, in Italy  and in Australia . Interestingly, the data from the present study show a nearly threefold, non-statistically significant, excess risk among HIV-infected men, but they suggest that the excess risk is restricted to IDU, regardless of their HIV serostatus. Behavioural risk factors unrelated to HIV infection (e.g. heavy cigarette smoking) are likely to explain such increased risk.
A statistically significant excess of cancers of the salivary glands (based on two observed cases only) was neither expected nor previously noted. Whether this excess represents a spurious association, or whether it could be linked with other cofactors (e.g. other viral infections) needs to be considered.
This investigation confirms that HIV-positive men below 50 years of age in southern Europe have twice the risk of non-AIDS-associated cancers than men in the general population, and that HD risk is strongly enhanced in HIV-infected men. Implications that hepatocellular carcinoma is more frequent among HIV-positive IDU, and that lung cancer risk is increased in IDU independently from HIV infection are new, and provide interesting lines for further research.
The authors would like to thank Ms C Gisbert and Ms N Oran for their invaluable support in the management of the DMI-2 database at the CISIH-Nice University Hospital, France; the staff of Centro Operativo AIDS, Istituto Superiore di Sanità, Rome; the staff of Centro Medico, Comunità di San Patrignano, Rimini, Italy; and Mr J. Canwell for text editing.
1. European Centre for the Epidemiological Monitoring of AIDS. HIV/AIDS Surveillance in Europe
. Quarterly Report No. 57, 1998.
2. International Agency for Research on Cancer. Human immunodeficiency viruses and human T-cell lymphotropic viruses.
IARC Monograph on the Evaluation of Carcinogenic Risk to Humans, Vol. 67. Lyon, 1996. pp. 103–114.
3. Serraino D, Pezzotti P, Dorrucci M, Alliegro MB, Sinicco A, Rezza G. Cancer incidence in a cohort of human immunodeficiency virus seroconverters.
Cancer 1997, 79: 1004 –1008.
4. Franceschi S, Dal Maso L, Arniani S. et al
. Risk of cancer other than Kaposi's sarcoma and non-Hodgkin's lymphoma in persons with AIDS in Italy.
Br J Cancer 1998, 78: 966 –970.
5. Goedert JJ, Coté TR, Virgo P. et al
. Spectrum of AIDS-associated malignant disorders.
Lancet 1998, 351: 1833 –1839.
6. Melbye M, Coté TR, Kessler L, Gail M, Biggar RJ, and the AIDS/Cancer Working Group. High incidence of anal cancer among AIDS patients.
Lancet 1994, 343: 636 –639.
7. Newton R, Beral V, Weiss R. Human immunodeficiency virus infection and cancer.
In:Cancer surveys, Vol. 33: Infections and human cancer.
Oxford: Cold Spring Harbor Laboratory Press; 1999. pp. 237– 262.
8. Pisani P, Parkin DM, Munoz N, Ferlay J. Cancer and infection: estimates of the attributable fraction in 1995.
Cancer Epidemiol, Biomarkers Prevention 1997, 6: 387 –400.
9. Serraino D, Carrieri P, Pradier C. et al
. Risk of invasive cervical cancer among women with, or at risk for, HIV infection.
Int J Cancer 1999, 82: 334 –337.
10. Pradier C, Pesce A, Taillan B. et al
. Reducing the incidence of Pneumocystis carinii pneumonia: a persisting challenge.
AIDS 1997, 11: 823 –824.
11. Rezza G, for the Italian Seroconversion Study. Determinants of progression to AIDS in HIV-infected individuals: an update from the Italian Seroconversion Study.
J Acquired Immun Defic Syndr Hum Retrovirol 1998, 17 (Suppl 1) : S13 –S16.
12. Parkin DM, Whelan SL, Ferlay J, Raymond L, Young J. Cancer incidence in five continents, Vol. VII. IARC Scientific Publication No. 143.
IARC, Lyon; 1997.
13. Breslow NE, Day NE. Statistical methods in cancer research, Vol. II. The design and analysis of cohort studies. IARC Scientific Publication No. 82.
Lyon; 1987. pp. 48– 79.
14. Munoz A, Sabin CA, Phillips AN. The incubation periods of AIDS.
AIDS 1997, 11 (Suppl. A) : S69 –S76.
15. Alcabes P, Pezzotti P, Phillips AN, Rezza G, Vlahov D. Long-term perspective on the prevalent-cohort biases in studies of human immunodeficiency virus progression.
Am J Epidemiol 1997, 146: 543 –551.
16. Grulich AE, Wan X, Law MG, Coates M, Kaldor JM. Risk of cancer in people with AIDS.
AIDS 1999, 13: 839 –843.
17. Hessol NA, Katz MH, Liu IY. et al
. Increased incidence of Hodgkin's disease in homosexual men with HIV infection.
Ann Intern Med 1992, 117: 309 –311.
18. Tirelli U, Errante D, Dolcetti R. et al
. Hodgkin's disease and human immunodeficiency virus infection: clinicopathologic and virologic features of 114 patients from the Italian Cooperative Group on AIDS and Tumours.
J Clin Oncol 1995, 13: 1758 –1767.
19. Mueller N, Evans A, Harris NL. et al
. Hodgkin's disease and Epstein–Barr virus: altered antibody pattern before diagnosis.
N Engl J Med 1989, 320: 689 –695.
20. Rabkin CS, Biggar RJ, Horn JW. Increasing incidence of cancer associated with the human immunodeficiency virus epidemic.
Int J Cancer 1991, 47: 692 –696.
21. Darby SC, Ewart DW, Giangrande PLF, Dolin PJ, Spooner RJD, Rizza CR, on behalf of the UK haemophilia Centre Directors Organisation. Mortality before and after HIV infection in the complete UK population of hemophiliacs.
Nature 1995, 377: 79 –82.
22. Parker MS, Leveno DM, Campbell TJ, Worrell JA, Carrozza SE. AIDS-related bronchogenic carcinoma: fact or fiction?
Chest 1998, 113: 154 –161.
23. Rabkin CS, Yellin F. Cancer incidence in a population with a high prevalence of infection with human immunodeficiency virus type I.
J Natl Cancer Inst 1994, 86: 1711 –1716.
The HIV Italian Seroconversion Study (ISS) Group also includes the following members: G. Tarantini (Centro Venereo, IRCCS, Maggilore Hospital, Milan); G. Angarano (University of Bari); A. Lazzarin (San Raffaele Hospital, Milan); F. Aiuti (University of Rome); B. Salassa (Amedeo di Savoia Hospital, Turin); F. Castelli (University of Brescia); P. Viale (Ospedale Civile, Piacenza); A. Canessa (University of Genoa); M. Barbanera (Livorno Hospital); E. Ricchi (University of Bologna); L. Ortona (Catholic University, Rome); R. Pristerà (General Hospital, Bolzano); S. Gafà (S. Maria Nuova Hospital, Reggio Emilia); and U. Tirelli (Centro di Riferimento Oncologico, Aviano).
Other members of the DMI-2 study group at Nice University Hospital are: P.M. Roger and P. Dellamonica (Department of Infectious Disease); J.P. Cassuto (Department of Haematology); F. Sanderson and J.G. Fuzibet (Department of Internal Medicine); and J.F. Michiels (Department of Pathology).