In people with AIDS, Kaposi's sarcoma (KS) has epidemiologic features that suggest that it may be etiologically related to a sexually transmissible agent. In Western countries, KS is much more common in homosexual men with AIDS than in other risk groups (1) and KS is associated with various indices of sexual behavior (2). Somewhat inconsistent evidence suggests that those homosexual men reporting insertive oroanal contact may be more likely to develop KS (2-4).
Since the discovery of Kaposi's sarcoma-associated herpesvirus (KSHV) (5), also known as human herpesvirus 8, this gamma herpesvirus has been found in >95% of fresh KS tissue samples but rarely in control tissue in Western countries (6). Detection of KSHV in peripheral blood mononuclear cells, or antibodies to KSHV, predict future development of KS (7-9), although the magnitude of the association between seropositivity and development of KS has not been described. These data suggest that KSHV is the likely cause of KS.
Few data are available on the means of transmission of KSHV. Among HIV-seropositive people in Western countries, the prevalence of antibodies to KSHV is much higher in homosexual men than others (10-13), and among homosexual men prevalence is higher in those reporting a greater number of sexual partners (14). In this report, we describe the association between reported history of specific sexual behaviors and sexually transmissible diseases (STDs) and KSHV infection in a cross-sectional study of homosexual men with AIDS in Sydney, Australia. In addition, we report on the magnitude of the prospective relation between KSHV infection and KS development in this population.
MATERIALS AND METHODS
Study Subjects and Questionnaire Data
This study is based on a previously described study of risk factors for KS in homosexual men newly diagnosed with AIDS in Sydney, Australia, between 1991 and 1993. The methods of sampling, collection of behavioral data, and characteristics of the study population have been described previously (4). Briefly, doctors who notified a person with newly diagnosed AIDS between February 1991 and March 1993 in Sydney, Australia, were invited to refer the patient for interview. For those patients who agreed to participate, information about sexual practices without condoms with regular and casual partners since 1980, and about history of STDs (i.e., nonspecific urethritis, gonorrhea, syphilis, hepatitis B, herpes, genital and anal warts, and amoebic infection), was sought by an interviewer-administered questionnaire. Casual partners were defined as sexual partners with whom the study subject had had sex on only a few occasions, or over a period of ≤1 week.
Only those study subjects who had banked serum specimens available were included in this analysis. Serum was available if blood had been taken as part of routine clinical care, and was stored at −20°C. The current analysis has two components: a cross-sectional analysis of risk factors for KSHV infection, and a prospective study of the relation between KSHV infection and development of KS. The study design is summarized in Figure 1.
Cross-Sectional Analysis of Risk Factors for KSHV Infection
The association between KSHV seropositivity and specific sexual practices with both casual and regular partners, and with self-reported and serologic evidence of sexually transmissible infections was quantified using OR and 95% CIs. Dose-response analyses were performed for numbers of casual partners with whom specific sexual acts were reported. Multivariate analyses of risk of seropositivity based on sexual behaviors with casual partners were performed using unconditional logistic regression using the statistical package SPIDA (Statistical Computing Laboratory Pty Ltd, Eastwood, NSW, Australia). Both insertive and receptive anogenital intercourse were included in the multivariate model, but because of the collinearity between insertive and receptive forms of orogenital and oroanal sexual behavior, only insertive oroanal and receptive orogenital contact were included.
Despite the previously reported presence of KSHV DNA in >95% of fresh KS tissue samples (6), we found that some men with KS were KSHV-seronegative at the time of specimen collection. Thus, a second analysis was performed in an attempt to better define KSHV infection. In this analysis, men with KS were considered to be KSHV-infected, and were compared with KSHV-seronegative men who did not develop KS during the study follow-up, who were considered to be KSHV uninfected.
Prospective Analysis of KS Development
Patients were observed until death, KS development, or October 1996 by medical record review. Those men who developed KS within 3 months of specimen collection were excluded from the prospective analysis of KS risk, given that clinical suspicion of KS development may have lead to serum collection before a diagnosis of KS was recorded. In 16 study subjects with KS, multiple blood specimens were tested. In these study subjects, the result of the first blood specimen was used for the survival analysis, unless the person seroconverted from negative to positive, in which case the first positive test was taken as the entry point into the survival analysis. The rate of KS diagnosis in those who were KSHV-seropositive was compared with the rate in those who were KSHV-seronegative using Kaplan-Meier plots and log-rank estimates of relative risk.
All study subjects gave consent to serologic testing of stored blood specimens as part of the initial study protocol. The blood specimen used for this analysis was the closest available prior to the date of AIDS diagnosis. All serum testing was performed without knowledge of KS status. In addition to antibodies to KSHV, antibodies to two other agents with known routes of transmission were tested. These included an organism transmitted primarily by the fecal-oral route, hepatitis A virus (HAV), and the sexually transmitted herpes simplex type 2 (HSV2).
Two previously described assays were used to detect antibodies to latent and lytic cycle products of KSHV (10,13). The first is an immunofluorescence assay (IFA) based on the latency-associated nuclear antigen (LANA) encoded by ORF 73 (15) and produced by the Epstein-Barr virus (EBV)-negative KSHV-infected cell line BCP-1 (10). Slides fixed with whole cells were incubated with patient sera diluted 1:160, washed and labeled with goat anti-human fluorescein isothiocyanate (FITC) immunoglobulin G (IgG) at 1:100 (Molecular Probes, Eugene, OR, U.S.A.) and IgM at 1:50 (Sigma, St Louis, MO, U.S.A.). A specular nuclear staining pattern was scored positive as previously described (10). The second assay is a Western blot assay to detect antibodies to ORF 65.2 truncated fusion antigen of KSHV, partially encoding a viral capsid protein expressed during lytic replication (13). Sera were diluted 1:100 and labeled with alkaline phosphatase goat anti-human IgG, IgM and IgA (Sigma) at a dilution of 1:2500. Nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indolylphosphate p-toluidine salt (NBT/BCIP, GIBCO-BRL, Grand Island, NY, U.S.A.) was used for development.
Serum samples were tested for the presence of total HAV antibodies by competitive inhibition enzyme immunoassay (MUREX Diagnostics, Dartford, U.K.). Initially reactive samples were retested in duplicate using the same assay. Those with OD values less than or equal to the calculated cut-off were considered positive.
Antibodies to HSV2 were detected by an indirect enzyme immunoassay previously described (16). Briefly, purified HSV2 specific glycoprotein G (gpG-2; 92- and 140-kd forms) antigen was harvested from HSV2 (B4327UR-infected BHK-21 cells) following affinity chromatography with gel-bound helix pomatia lectin and bound to wells of a 96-well microtiter plate. An IgG-specific alkaline phosphatase-conjugated second antibody was used to detect bound human anti-HSV2 specific IgG antibodies. Specimens with equivocal results were further tested by an IgG specific Western blot (17) incorporating the HSV2 gpG-2 antigens utilizing an amplified substrate detection procedure. Those samples demonstrating a repeat reactive enzyme-linked immunosorbent assay (ELISA) and/or demonstrating a characteristic immunoblot reactivity pattern to HSV2 specific gpG-2 (92-kd form) were considered seropositive.
Banked sera were available in 130 of 202 originally recruited study subjects with AIDS (51 of 71 men who developed KS (72%) and 79 of 131 men who did not develop KS (60%); p = .10 ). Those without stored sera were not significantly different with respect to age at AIDS diagnosis or reported sexual behaviors.
Of the 130 study subjects, the prevalence of antibodies to KSHV by LANA IFA, ORF65 WB, and either assay was 36 (28%), 63 (48%), and 68 (52%), respectively. Of 14 men who had KS at the time of specimen collection, 9 (64%) were positive by LANA IFA, 11 (79%) by ORF65 WB, and 11 (79%) by either assay. Of the 79 who never developed KS, 10 (13%) were positive by LANA IFA, 24 (30%) by ORF65 WB, and 27 (34%) by either. The OR for the association between KSHV seropositivity and KS using serum samples taken nearest to AIDS diagnosis was 7.2 (95% CI, 3.0-17) for LANA seropositivity and 7.5 (95% CI, 3.3-16.7) for ORF65 seropositivity.
Cross-Sectional Analysis of Risk Factors for KSHV Infection
KSHV-seropositive patients were more likely to report a history of most STDs than seronegative patients, although this reached significance only for the association between anal warts and LANA (OR 3.2; 95% CI, 1.0-9.8). For both KSHV assays, the median number of casual sexual partners was nonsignificantly higher in seropositives than seronegatives (337 and 276, p = .628 for LANA, and 410 and 264, p = .281 for ORF65, respectively).
No association was found between reported sexual activities with regular partners and KSHV serostatus and most OR were nonsignificantly <1 (data not shown). In contrast, most sexual behaviors with casual partners were reported nonsignificantly more frequently by the KSHV-seropositive. The sexual behavior most closely related to risk of seropositivity, although still nonsignificant, was insertive oroanal contact with casual partners (Table 1). No association between frequency of insertive oroanal contact with casual partners and KSHV seropositivity was found (data not shown). Although not statistically significant, insertive oroanal contact remained the most strongly associated activity with KSHV seropositivity when controlling for insertive and receptive anogenital intercourse and receptive orogenital contact on multivariate analysis (OR 2.1; 95% CI, 0.82-5.3 for LANA, and OR 2.2; 95% CI, 0.92-5.1 for ORF65, respectively).
The seroprevalence of HSV2 and HAV was 64% and 54%, respectively. HSV2 serology was significantly related to LANA and ORF65 serology (Table 2). None of the 16 study subjects with multiple serum samples showed evidence of simultaneous seroconversion to KSHV (on either assay) and HSV2. HAV serology was significantly related to ORF65 and nonsignificantly related to LANA (Table 2).
In all, 51 men with KS (including 10 KSHV-seronegative men, 7 of whom had blood collected >6 months prior to KS) were compared with 52 men who were seronegative for both KSHV assays and never developed KS. Excluded from Analysis 2 were 27 men who were seropositive on one or both tests but did not develop KS. The median number of casual sexual partners since 1980 was 500 in men with KS and 225 in KSHV-uninfected men (p < .01). The only sexual behavior significantly related to KS status was insertive oroanal contact with casual partners (OR 2.8; 95% CI. 1.2-6.7). This showed a dose-response trend; persons reporting between 1 and 60 partners (n = 16) and those reporting >60 partners (n = 17) were 2.4 and 3.4 times as likely to have KS and KSHV infection, respectively, than men reporting no oroanal insertive contact with casual partners (p for trend = .02). Insertive oroanal intercourse remained significantly associated with KS status when controlling for insertive and receptive anogenital intercourse and receptive orogenital contact on multivariate analysis (OR 2.7; 95% CI, 1.0-7.0).
HSV2 serology was related to KSHV infection in analysis 2 (OR 2.6; 95% CI, 1.0-6.7), but HAV serology was not (OR 1.6; 95% CI, 0.69-3.8).
Risk Factors for HSV2 and HAV Infection
Those seropositive for HSV2 and HAV had significantly more casual sexual partners than those seronegative for each test (median, 457 and 128, p < .01 for HSV2; median, 462 and 194, p = .02 for HAV, respectively). For HSV2 and HAV, a range of sexual behaviors with casual but not regular partners was significantly related to seropositivity (Table 1). Risk of HAV was nonsignificantly associated with casual insertive oroanal contact (OR 1.8; 95% CI, 0.84-4.0). Multivariate analysis identified no sexual behaviors significantly associated with risk of HSV2 or HAV infection, but most sexual behaviors with casual partners were nonsignificantly associated.
Prospective Analysis of KS Development
Fourteen men were excluded from the survival analysis because they had KS within 3 months of specimen collection, and 1 man was excluded for lack of follow-up information. Thus, the rate of KS development was examined in 115 men. In the 37 who developed KS, blood was collected a median of 840 days prior to KS onset. KSHV seropositivity was predictive of future development of KS. The RR associated with seropositivity to LANA was 2.1 (95% CI, 1.1-4.1), and with ORF65 was 4.3 (95% CI, 2.0-9.2). The RR associated with seropositivity by either LANA or ORF65 was 4.4 (95% CI, 1.9-10.2; Fig. 2). Of 57 men initially seropositive to either test, 30 (53%) developed KS, compared with only 7 of 58 (12%) initially seronegative men.
This study provides suggestive evidence that KSHV may be a sexually transmitted infection among homosexual Australian men. KSHV-seropositive men were significantly more likely to be HSV2- and HAV-seropositive, and to report a history of anal warts. They were nonsignificantly more likely to report a history of other STDs and to have more casual sexual partners, and to report a variety of sexual practices with casual partners.
Those seropositive to either KSHV assay were more than four times as likely to develop KS subsequently than those seronegative to both, and 53% of study subjects with serologic evidence of KSHV developed KS during the course of AIDS. Given that many competing causes of death exist within this population, this may underestimate the likelihood of KS development among KSHV-positive AIDS patients who survive for long periods with immune deficiency. Most of the KSHV-seronegative men who developed KS had blood drawn >1 year prior to KS onset and may have been infected with KSHV in the interim.
Defining an individual as KSHV-infected for risk factor analyses was problematic in the absence of a serologic assay that is close to 100% sensitive and specific. In addition to the serologic comparisons, in the second risk factor analysis, we defined 10 men with KS who were seronegative on both assays as KSHV-infected. Of these 10 men, most had specimens collected >6 months prior to KS development, and KSHV infection may have occurred after the time of specimen collection. Possible false-positivity of the serologic tests for KSHV was addressed by excluding from the second analysis those who were seropositive by one or both KSHV assays but did not develop KS. In this analysis, men with KS had more than twice as many casual sexual partners as uninfected men, and were significantly more likely to report insertive oroanal sexual contact with casual partners. Thus, relations with behavioral sexual variables were stronger, consistent with this being a more precise definition of KSHV infection if KSHV is truly sexually transmissible.
Serum was only available on 64% of study subjects who were enrolled in the initial survey of people with AIDS (4). However, those with and without sera did not differ with respect to age at AIDS diagnosis or sexual behavior variables. It is likely, however, that recall of STDs and sexual behaviors over a period of >10 years was imprecise. Assuming nondifferential misclassification, the true associations are likely to be stronger than those we have reported. Recall bias is unlikely to explain the behavioral associations found, in that the sexually transmitted hypothesis of KS was not widely appreciated at the time when study subjects were interviewed (i.e., between 1991 and 1993). The relatively limited power of this study, and the use of prevalent rather than incident cases of KSHV infection may explain why the relations between KSHV serology and sexual behavioral variables were consistently suggestive of sexual transmission but did not reach statistical significance for most variables.
The significant association between KSHV infection and HSV2 serology in this study is likely to reflect similar risk factors for infection for the two viruses. Although it is possible that serologic cross-reactivity occurs, no simultaneous seroconversion to KSHV and HSV2 was determined among any of the patients with multiple serum samples. Examination of the amino acid sequences for KSHV's LANA and ORF65 and HSV's glycoprotein G-2 identified no more apparent shared linear epitopes of >5 amino acids.
Several studies have found an association between self-reported history of oroanal contact and KS risk (3,4,18,19) but others have not confirmed this (20-25). Only one study looked for but was unable to find KSHV DNA in feces of men with AIDS-KS (7); however, KSHV DNA has been reported in rectal biopsy specimens (26). KSHV DNA is detectable in prostatic tissue (27) and semen (28-30). HAV is believed to be transmitted by the fecal-oral route in homosexual men (31). Our findings of an association between HAV and KSHV ORF65 serology and a nonsignificant association between KSHV serology and oroanal contact with casual partners suggests that KSHV may be transmitted by the fecal-oral route. However, these associations were not strong and larger studies are needed to further characterize the precise sexual mode of transmission of KSHV.
Acknowledgment: The Australian National Centre in HIV Epidemiology and Clinical Research is supported by the Commonwealth Department of Health and Family Services through the Australian National Council on AIDS and Related Diseases and its Research Advisory Committee.
1. Beral V, Peterman TA, Berkelman RL, Jaffe HW. Kaposi's sarcoma among persons with AIDS: a sexually transmitted infection? Lancet
2. IARC working group on the evaluation of carcinogenic risks to humans. IARC Monographs on the evaluation of carcinogenic risks to humans,
vol 67. Human immunodeficiency viruses and human T cell lymphotrophic viruses. Lyon: IARC, 1996.
3. Beral V, Bull D, Darby S, et al. Risk of Kaposi's sarcoma and sexual practices associated with faecal contact in homosexual or bisexual men with AIDS. Lancet
4. Grulich A, Hendry O, Luo K, Bodsworth N, Cooper D, Kaldor J. Risk of Kaposi's sarcoma and oroanal sexual contact. Am J Epidemiol
5. Chang Y, Cesarman E, Pessin MS, et al. Identification of herpesvirus-like DNA sequences in AIDS-associated Kaposi's sarcoma. Science
6. Olsen SJ, Moore PS. Kaposi's sarcoma-associated herpesvirus (KSHV/HHV8) and the etiology of Kaposi's sarcoma. In Freidman H, Medreczky P, Bendinelli M. Molecular immunology of herpesviruses.
New York: Plenum Publishing, 1998.
7. 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
8. Gao SJ, Kingsley L, Hoover DR, et al. Seroconversion to antibodies against KSHV-related latent nuclear antigens before the development of Kaposi's sarcoma. N Engl J Med
9. Moore P, Kingsley LA, Holmberg SD, et al. Kaposi's sarcoma-associated herpesvirus infection prior to onset of Kaposi's sarcoma. AIDS
10. Gao SJ, Kingsley L, Li M, et al. KSHV antibodies among Americans, Italians, and Ugandans with and without Kaposi's sarcoma. Nat Med
11. Kedes DH, Operskalski E, Busch M, Kohn R, Flood J, Ganem D. The sero-epidemiology of KSHV (KSHV): distribution of infection in Kaposi's sarcoma risk groups and evidence for sexual transmission. Nat Med
12. Lennette ET, Blackbourn DJ, Levy JA. Antibodies to KSHV in the general population and in Kaposi's sarcoma patients. Lancet
13. Simpson GR, Schulz TF, Whitby D, et al. Prevalence of Kaposi's sarcoma-associated herpesvirus infection measured by antibodies to recombinant capsid protein and latent immunofluorescence antigen. Lancet
14. Martin JN, Ganem DE, Osmond DH, Page-Shafer KA, Macrae D, Kedes DH. Sexual transmission and the natural history of human herpesvirus 8 infection. N Engl J Med
15. Rainbow L, Platt GM, Simpson GR, et al. The 222 to 234 kilodalton nuclear protein (latent nuclear antigen) of Kaposi's sarcoma associated-herpesvirus (human herpesvirus eight) is encoded by ORF73 and a component of the latency associated nuclear antigen. J Virol
16. Ho DWT, Field PR, Sjogren-Jansson E, Jeansson S, Cunningham AL. Indirect ELISA for the detection of IgG and IgM antibodies with glycoprotein G (gG-2). J Virol Methods
17. Ho DWT, Field PR, Irving WL, Pakham DR, Cunningham AL. Detection of immunoglobulin M antibodies to glycoprotein G-2 by Western blot (immunoblot) for diagnosis of initial herpes simplex type 2 genital infections. J Clin Microbiol
18. Darrow WW, Peterman TA, Jaffe HW, Rogers MF, Curran JW, Beral V. Kaposi's sarcoma and exposure to faeces [letter]. Lancet
19. Jacobson LP, Muñoz A, Fox R, et al. Incidence of Kaposi's sarcoma in a cohort of homosexual men infected with the HIV type 1. J Acquir Immune Defic Syndr
20. Elford J, Tindall B, Sharkey T. Kaposi's sarcoma and insertive rimming [letter]. Lancet
21. Lifson AR, Darrow WW, Hessol NA, et al. Kaposi's sarcoma in a cohort of homosexual and bisexual men. Am J Epidemiol
22. Page-Bodkin K, Tappero J, Samuel M, Winkelstein W. Kaposi's sarcoma and faecal-oral exposure [letter]. Lancet
23. van Griensven GJ, Boucher EC, Coutinho RA. Oro-anal sex and the occurrence of Kaposi's sarcoma [letter]. Genitourin Med
24. Kaldor JM, Tindall B, Williamson P, Elford J, Cooper DA. Factors associated with Kaposi's sarcoma in a cohort of homosexual and bisexual men. J Acquir Immune Defic Syndr
25. Archibald CP, Schechter MT, Le TN, Craib KJP, Montaner JSG, O Shaughnessy MV. Evidence for a sexually transmitted cofactor for AIDS-related Kaposi's sarcoma in a cohort of homosexual men. Epidemiol
26. Thomas JA, Brookes LA, McGowan I, Weller I, Crawford D. HHV8 DNA in normal gastrointestinal mucosa from HIV seropositive people [letter]. Lancet
27. Corbellino M, Poirel L, Bestetti G, et al. Restricted tissue distribution of extra-lesional Kaposi's sarcoma-associated herpesvirus-like DNA sequences in AIDS patients with Kaposi's sarcoma. AIDS Res Hum Retroviruses
28. Gupta P, Singh MK, Rinaldo C, et al. Detection of KSHV DNA in semen of homosexual men with Kaposi's sarcoma. AIDS
29. Howard MR, Whitby D, Bahadur G, et al. Detection of HHV8 DNA in semen from HIV-infected individuals but not healthy semen donors. AIDS
30. Huang YQ, Li JJ, Poiesz BJ, Kaplan MH, Friedman-Kien AE. Detection of the herpesvirus-like DNA sequences in matched specimens of semen and blood from patients with AIDS-related Kaposi's sarcoma by polymerase chain reaction in situ hybridization. Am J Pathol
31. Corey L, Holmes KK. Sexual transmission of hepatitis A in homosexual men. N Engl J Med 1980;302:435-8.