Kaposi's sarcoma (KS) is a common AIDS-defining disease that occurs more frequently among HIV-infected homosexual men than among other groups of HIV-infected individuals . This finding is consistent with reports of a high prevalence of infection with the KS-associated herpesvirus, also referred to as human herpesvirus 8 (HHV-8) [2–4]. In fact, infection with HHV-8 is strongly predictive of the development of KS in HIV-infected individuals [5–8].
The decline in the proportional incidence of KS as an AIDS-defining disease started before the introduction of dual-combination or highly active antiretroviral therapy (HAART) [9,10]. However, longitudinal studies of HIV-infected individuals have shown that this proportional decrease may occur even in the presence of an increasing incidence rate of KS, due to a higher increase in the incidence of other AIDS-defining diseases [11–13]. A decrease in cumulative KS rates after AIDS diagnosis and a later occurrence of KS have also been reported .
The decrease in KS might be due either to a shorter incubation period with respect to other AIDS-defining diseases [11,14–16], and thus an earlier peak and decrease in incidence, or to a reduced prevalence and/or virulence of a cofactor, such as HHV-8 , although this latter hypothesis has not been tested.
To evaluate the temporal trend of KS and the dynamics underlying its potential decrease, we analysed the data prospectively collected among homosexual men enrolled in the Italian Seroconversion Study. Specifically, we investigated the effect of the year of HIV seroconversion and of calendar year on the risk of developing KS. Furthermore, to evaluate whether the prevalence of cofactors for the development of KS changed over time, we studied the temporal trend of HHV-8 seroprevalence.
The analysis was performed on the cohort of individuals of the Italian Seroconversion Study, an open cohort study on the natural history of HIV infection. For this study, participants were homosexual men with known dates of HIV seroconversion (defined as the midpoint in time between the last negative HIV test and the first positive test within 2 years). For the Italian Seroconversion Study, clinical information was routinely collected from participants, who were advised to undergo clinical and laboratory examination every 6 months, including measurements of CD4 T-cell counts. Detailed information on the study has been provided elsewhere [17–19]. Information on antiretroviral therapy was collected for each individual. Specifically, in our cohort, double therapy was introduced in October 1995, whereas HAART, including protease inhibitors, was introduced in August 1996.
Diagnosis of KS was confirmed by histological examination in all cases diagnosed up to 1987; in cases diagnosed afterwards, either histological examination or presumptive criteria were used, in accordance with the 1987 AIDS case definition .
Seropositivity was determined by detecting antibodies to HHV-8 lytic antigens, using an immunofluorescence assay with body cavity B-cell lymphoma cell line (obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, National Institutes of Health, from M. McGrath and D. Ganem). A detailed description of the methodology used has been reported elsewhere [8,21,22].
HHV-8 seroprevalence by year of HIV seroconversion was estimated among a subgroup of 200 participants (51.7% of the study population) who seroconverted for HIV between 1984 and 1994 and for whom at least one serum specimen was available after seroconversion. On average, the serum sample was available within approximately 1 year of seroconversion (the median lag from seroconversion was 0.7 years and the interquartile range was 0.4–1.6 years). For 38 (19.0%) of these individuals, the period tested for HHV-8 coincided with seroconversion for HIV. Selection criteria have been discussed previously . No differences were found in the general characteristics or in disease progression between those with tested sera and those without.
The distribution of HIV seroconversion by calendar year was computed, and changes over time were assessed estimating the incidence of KS in different calendar periods. To this end, the contribution in terms of person–years without KS was calculated for each individual and for each of the following calendar periods: 1984–1989, 1990–1992, 1993–1995, and 1996–1997. The proportion of cases of the first AIDS-defining disease represented by KS was estimated for the same calendar periods. To determine a possible cohort effect, the risk of developing KS in participants who seroconverted for HIV before 1990 (the median year of seroconversion) was compared with that for those who seroconverted later, using the Kaplan–Meier method. KS-free time was estimated considering the time interval between the date of HIV seroconversion and the first diagnosis of KS, or the date of death, or the date of the end of the study (31 December 1997).
To examine the effect of calendar time on the incidence of KS as first AIDS-defining disease, we applied competing-risk models . This enabled us to study trends in the incubation period over calendar time simultaneously for both KS and non-KS first AIDS-defining events. AIDS-free time was estimated as the interval from HIV seroconversion to the first AIDS diagnosis, or to death, or to the end of the study. Data were duplicated for each patient, including either the first KS-AIDS-defining event as end-point or the first non-KS AIDS-defining event as failure event. In particular, calendar period was entered as a categorical time-dependent variable, considering the same calendar periods mentioned above. Relative hazards (RH) were adjusted for year of seroconversion, which was entered as a dichotomous covariate (i.e. before 1990 versus 1990 or later), and for other possible confounders, such as age at HIV seroconversion and lag between the last HIV-negative test and the first positive test. Competing-risks models were repeated considering the effect of calendar period on the incidence rate of both KS and other non-KS AIDS-defining diseases (also considering KS occurring after the first diagnosis of AIDS).
To verify whether changes over time in KS incidence were due to the introduction of new antriretroviral therapies, their effect was evaluated in separate competing-risks models. Treatment variable was entered as a dichotomous time-dependent covariate considering an ‘intention-to-treat approach’ (mono-, dual- or triple-therapy versus no therapy).
As HHV-8 is an important cofactor for KS, the competitive-risk models were also repeated entering in the models this cofactor among individuals with known HHV-8 serology. In particular, we evaluated whether the RH of HHV-8 as predictor for KS changed over time (i.e. participants who seroconverted for HIV before 1990 were compared with those who seroconverted later).
Forty-eight cases of KS were observed during the follow-up, and in 35 of these cases (72.9%), KS was the first AIDS-defining disease.
Rates and proportional incidence of KS by calendar period
The incidence of KS by calendar period increased through 1995, followed by a dramatic decrease in 1996–1997 (Table 1); however, P-values (obtained by using a two-tailed Fisher exact test comparing each time period versus 1984–1989 as reference) were not statistically significant. These findings did not change when comparing individuals who seroconverted before 1990 with those who seroconverted later (data not shown).
The proportional incidence of KS as first AIDS-defining disease was 33.3% among AIDS cases diagnosed in 1984–1989, 24.3% among those diagnosed in 1993–1995, and 16.7% among those diagnosed in 1996–1997. These findings did not significantly change after excluding the other AIDS-defining diseases included in the 1993 revised AIDS case definition (KS would account for 25.0% of cases in 1993–1995 and 18.2% in 1996–1997).
The median CD4 count at KS diagnosis decreased from 368 in 1984–1989 to 70 in 1993–1995; it then increased to 163 in 1996–1997.
Risk of developing KS by period of HIV seroconversion: analysis of the cohort effect
The risk of developing KS by period of HIV seroconversion is shown in Figure 1. No statistically significant difference was observed between the two groups (P-value = 0.812). In particular, the cumulative incidence of KS at 5 years from seroconversion was 6.0% [95% confidence interval (CI), 2.6–9.5%] for those who seroconverted before 1990 and 6.8% (95% CI, 2.4–11.1%) for those who seroconverted later.
Crude and adjusted RHs of KS or other AIDS-defining events were estimated for different calendar periods applying competing-risks models (Table 2). The risk of KS was reduced by approximately 60% for persons followed between 1996 and 1997, compared with those followed between 1984 and 1989; however, this reduction was not statistically significant. When adjusting for period of HIV seroconversion (i.e., before 1990 versus 1990 and later), the RH for the 1996–1997 period tended to decrease. In the same model, crude and adjusted RH of non-KS AIDS-defining events over the same time-periods were greater than 1, but not statistically significant. However, during the last period considered (1996–1997), the risk of non-KS AIDS-defining events tended to decrease towards the unity. Similar results were obtained when also including in the competitive-risks models KS cases and other AIDS-defining events occurring after AIDS diagnosis (results not shown).
To evaluate whether the decline of KS over time was due to new combination therapy, a separate competing-risks model was performed. The risk of KS was approximately 50% lower for those who received dual therapy compared to those who did not receive therapy, whereas the effect of antiretroviral treatment with protease inhibitors was not estimable given that none of those who received triple therapy developed AIDS (results not shown). The risk of developing KS did not change by period of HIV seroconversion. The RH was 0.92 (95% CI, 0.41–2.03) for those who seroconverted in 1990 or later compared with those who seroconverted earlier (data not shown).
Temporal trend of HHV-8 seroprevalence
The seroprevalence of HHV-8 did not change significantly over time (χ2 for trend, P-value = 0.762) (Fig. 2). Some fluctuation was observed, possibly due to differing numbers of newly HIV-infected men tested for HHV-8 each year. HHV-8 seroprevalence remained almost stable when comparing those who seroconverted for HIV before 1990 with those who seroconverted later (66.7 and 62.0%, respectively).
Competitive risks models were performed among those with data on HHV-8 serology to evaluate whether the RH of HHV-8 as predictor of KS changed over time, in particular, for those who seroconverted for HIV in 1990 or later versus before 1990. The crude RH of KS for HHV-8 positive individuals was 12.97 (95% CI, 1.74–96.69); as expected, no association was found for non-KS end-points (RH, 1.04; 95% CI, 0.61–1.76). The RH of KS for HHV-8 remained of the same magnitude and also statistically significant (RH, 13.15; 95% CI, 1.76–97.99) when adjusting for HIV seroconversion before 1990 compared with later seroconversion.
In our study population, the incidence of KS increased through 1995, as a possible consequence of the increasing maturity of infection in the cohort, whereas the proportional incidence of KS showed a slight decrease over time. These findings are consistent with those of other studies conducted in different areas of the industrialized world [11,13], which have suggested that the upward trend of KS rates may be explained by progressive immunosuppression .
The reasons for the decrease in the proportional incidence of KS have yet to be defined. One possible explanation is that homosexual men who developed AIDS at the beginning of the epidemic were among the most sexually active and were more likely to be co-infected with HHV-8, and that HIV then spread to less sexually active homosexual men, who were at lower risk of KS . An alternative hypothesis is that KS has a shorter incubation period than other AIDS-defining diseases and thus tends to occur earlier in the course of HIV disease; under certain conditions, such as a bulk of new HIV infections occurring in a limited time period, the incidence of KS would peak before that of other opportunistic diseases and would thus decrease earlier .
Our data suggest that differences in the incubation time are likely to explain the findings of most studies better than temporal changes in the prevalence of a cofactor. In fact, HHV-8 seroprevalence remained high throughout the study period. However, as already observed among persons with AIDS reported to the Italian AIDS Registry , the decrease in the proportional incidence of KS in our cohort was not as steep as in studies conducted in the United States . This may be due to the different dynamics of the HIV epidemic among homosexual men, characterised by a lower peak and by a persistent but relatively lower viral circulation.
The reason why KS tended to be diagnosed at lower CD4 T-cell counts beginning in 1990 is not known. A study conducted among persons with AIDS in Australia reported a similar finding and suggested a tendency toward a later occurrence of KS .
This is one of the few studies that correlates the trend of KS with that of HHV-8 seroprevalence. Moreover, only a few studies have reported changes in the occurrence of HHV-8 over time among HIV-infected homosexual men. A study conducted in the United States provided evidence of concurrent epidemics of HHV-8 and HIV among homosexual men in the early 1980s . These results appear to confirm those of a Danish study that showed a decrease in HHV-8 incidence over time, suggesting that this infection may have been actively spread simultaneously with HIV and through the same modalities . Although our data do not allow incidence to be estimated at the population level, they do suggest that HHV-8 continues to spread among homosexual men who still engage in at-risk practices for HIV infection.
The decrease in the risk of KS after 1995 is likely to be due to the increasing use of effective combination therapy and HAART. Our findings are concordant with those of other studies  and are likely to be attributable to the efficacy of antiretroviral treatment via HHV-8-specific or non-specific immune effects . The decrease in the proportional incidence of KS in 1996–1997 suggests a stronger effect of antiretroviral combination therapy on KS, which requires further evaluation. CD4 T-cell count at diagnosis of KS tended to increase after 1995, as a possible consequence of combined antiretroviral treatment.
Before drawing conclusions, some limitations and possible biases should be mentioned. First, our study population may not be representative of all HIV-infected homosexual men in Italy, and the local situation may vary by geographical area. For this reason, the dynamics of KS and HHV-8 infection could not be generalized to the entire country. However, KS temporal trends were similar to those observed in the National AIDS Registry (25). Second, HHV-8 seroprevalence was not estimated on the total study population and a selection bias may have occurred. However, all the men with an available sample after HIV seroconversion were tested, and the characteristics of those who were tested did not significantly differ from those who were not tested. Third, concerns have been expressed about the sensitivity and specificity of HHV-8 assays, although the method used in our study has been shown to have a high prognostic value . Unfortunately, accurate estimates of sensitivity and specificity of HHV-8 serological assays are not available because of the lack of a gold standard.
In conclusion, although KS rates increased through 1995, the proportional incidence decreased, due to the higher increase in rates of other AIDS-defining diseases. The finding that the risk of developing KS after HIV seroconversion remained stable over time is consistent with the stable trend of HHV-8 seroprevalence. The dramatic decrease in KS rates after 1995 coincides with the introduction of combination treatment or HAART.
The authors wish to thank Mark Kanieff for editing the text and Manuela Zazzara for data input and secretarial support.
1. Beral V. Epidemiology of Kaposi's sarcoma.
Cancer Surv 1991, 10: 5 –22.
2. Chang Y, Cesarman E, Pessin MS. et al. Identification of herpesvirus-like DNA sequences in AIDS-associated Kaposi's sarcoma.
Science 1994, 266: 1865 –1869.
3. Kedes DH, Operskalski E, Busch M, Kohn R, Flood J, Ganem D. The seroepidemiology of human herpesvirus 8 (Kaposi's sarcoma-associated herpesvirus): distribution of infection in KS risk groups and evidence for sexual transmission.
Nature Med 1996, 2: 918 –924.
4. Lennette ET, Blackbourn DJ, Levy JA. Antibodies to human herpesvirus type 8 in the general population and in Kaposi's sarcoma patients.
Lancet 1996, 348: 858 –861.
5. Whitby D, Howard MR, Tenant-Flowers M. et al. Detection of Kaposi's sarcoma associated herpesvirus in peripheral blood of HIV-infected individuals and progression to Kaposi's sarcoma.
Lancet 1995, 346: 799 –802.
6. Martin JN, Ganem DE, Osmond DH, Page-Shafer KA, Macrae D, Kedes DH. Sexual transmission and the natural history of human herpesvirus 8 (HHV-8) infection.
N Engl J Med 1998, 338: 948 –954.
7. Renwick N, Halaby T, Weverling GJ. et al. Seroconversion for human herpesvirus 8 during HIV infection is highly predictive of Kaposi's sarcoma.
AIDS 1998, 12: 2481 –2488.
8. Rezza G, Andreoni M, Dorrucci M. et al. Human herpesvirus 8 seropositivity and risk of developing Kaposi's sarcoma and other AIDS-related diseases among individuals with known dates of HIV seroconversion.
J Natl Cancer Inst 1999, 91: 1468 –1474.
9. Beral V, Peterman TA, Berkelman RL, Jaffe HW. Kaposi's sarcoma among patients with AIDS: a sexually transmitted infection?
Lancet 1990, 335, i: 123 –128.
10. Franceschi S, al Maso L, Lo Re A, Serraino D, La Vecchia C. Trends of Kaposi's sarcoma at AIDS diagnosis in Europe and the United States, 1987–94.
Br J Cancer 1997, 76: 114 –117.
11. Lifson AR, Darrow WW, Hessol NA. et al. Kaposi's sarcoma in a cohort of homosexual and bisexual men.
Am J Epidemiol 1990, 131: 221 –231.
12. Katz MH, Hessol NA, Buchbinder SP, Hirozawa A, O'Malley P, Holmberg SD. Temporal trends of opportunistic infections and malignancies in homosexual men with AIDS.
J Infect Dis 1994, 170: 198 –202.
13. Veugelers PJ, Strathdee SA, Moss AR. et al. Is the human immunodeficiency-related Kaposi's sarcoma epidemic coming to an end? Insights from the Tricontinental Seroconverter Study.
Epidemiology 1995, 6: 382 –386.
14. Dore GJ, Li Y, Grulich AE. et al. Declining incidence and later occurrence of Kaposi's sarcoma among persons with AIDS in Australia: the Australian AIDS cohort.
AIDS 1996, 10: 1401 –1406.
15. Hermans P, Lundgren J, Sommereijns B. et al. Epidemiology of AIDS-related Kaposi's sarcoma in Europe over 10 years.
AIDS 1996, 10: 911 –917.
16. Fusaro RE, Bacchetti P, Jewell NP. A competing risks analysis of presenting AIDS diagnoses trends.
Biometrics 1996, 52: 211 –225.
17. Rezza G, Dorrucci M, Pezzotti P, Lazzarin A, Angarano G, Sinicco A. et al
. The seroconversion study on the natural history of HIV infection.
In HIV Epidemiology: Models and Methods
. Edited by Nicolosi A. New York: Raven Press; 1994: 279 –289.
18. Mariotto AB, Mariotti S, Pezzotti P, Rezza G, Verdecchia A. Estimation of the acquired immunodeficiency syndrome incubation period in intravenous drug users: a comparison with male homosexuals.
Am J Epidemiol 1992, 135: 428 –437.
19. Dorrucci M, Balducci M, Pezzotti P, Sinicco A, Alberici F, Rezza G. Temporal changes in the rate of progression to death among Italians with known date of HIV seroconversion: estimates of the population effect of treatment.
:Italian HIV Seroconversion Study (ISS).
J Acquir Immune Defic Syndr 1999, 22: 65 –70.
20. Centers for Disease Control. Revision of the CDC surveillance case definition for acquired immunodeficiency syndrome.
MMWR 1987, 36: 1S. 1S.
21. Andreoni M, El-Sawaf G, Rezza G. High seroprevalence of antibodies to human herpesvirus-8 in Egyptian children: evidence of non-sexual transmission.
J Natl Cancer Inst 1999, 91: 465 –469.
22. Monini P, Carlini F, Sturzl M. et al. α-Interferon inhibits HHV-8 reactivation in primary effusion lymphoma cells and reduces HHV-8 load in cultured PBMC.
J Virol 1999, 73: 4029 –4041.
23. Andersen PK, Borgan Ø, Gill RD, Keiding N. Statistical Models Based on Counting Processes.
New York: Springer; 1993.
24. Munoz A, Schrager LK, Bacellar H. et al. Trends in the incidence of outcomes defining acquired immunodeficiency syndrome (AIDS) in the Multicenter AIDS Cohort Study: 1985–1991.
Am J Epidemiol 1993, 137: 423 –438.
25. Pezzotti P, Serraino D, Rezza G. et al. The spectrum of AIDS-defining diseases: temporal trends in Italy prior to the use of highly active anti-retroviral therapies, 1982–1996.
Int J Epidemiol 1999, 28: 975 –981.
26. O'Brien TR, Kedes D, Ganem D. et al. Evidence for concurrent epidemics of Human Herpesvirus 8 and Human Immunodeficiency Virus type 1 in US homosexual men: rates, risk factors, and relationship to Kaposi's sarcoma.
J Infect Dis 1999, 180: 1010 –1017.
27. Melbye M, Cook PM, Hjalgrim H. et al. Risk factors for Kaposi's sarcoma- associated herpesvirus (KSHV/HHV-8) seropositivity in a cohort of homosexual men, 1981–1996.
Int J Cancer 1998, 77: 543 –548.
28. Ledergerber B, Talenti A, Egger M. Risk of HIV related Kaposi's sarcoma and non-Hodgkin's lymphoma with potent antiretroviral therapy: prospective cohort study.
BMJ 1999, 319: 23 –24.
29. Lebbé C, Blum L, Pellet C. et al. Clinical and biological impact of antiretroviral therapy with protease inhibitors on HIV-related Kaposi's sarcoma.
AIDS 1998, 12: F45 –F49.
The Italian Seroconversion Study also includes the following investigators: A. Sinicco, R. Sciandra (Amedeo di Savoia Hospital, Turin); F. Alberici (Ospedale Civile, Piacenza); F. Castelli (University of Brescia); A. De Luca, L. Ortona (Catholic University, Rome); M.A. Ursitti (Santa Maria Nuova Hospital, Reggio Emilia); G. Mazzarello (University of Genova); A. Lazzarin (IRCCS San Raffaele Hospital, Milan); M. Zaccarelli (IRCCS L. Spallanzani Hospital, Rome); U. Tirelli (IRCCS CRO, Aviano); R.M. Rosso, F. Aiuti (University of Rome ‘La Sapienza'); M. Barbanera (Ospedali Riuniti, Livorno); G. Angarano (Infectious Disease Clinic, University of Bari); R. Pristerà (Infectious Disease Division, Ospedale Regionale, Bolzano).