Human Herpesvirus 8 (HHV-8) Infection in HIV/AIDS Patients From Santos, Brazil: Seroprevalence and Associated Factors

Goal: The goal of this study was to evaluate the seroprevalence of human herpesvirus 8 (HHV-8) infection among HIV-infected individuals from Brazil and the associated risk factors.

Study: A cross-sectional survey was carried out with 497 HIV/AIDS outpatients attending the local AIDS Reference Center in Santos (southeastern Brazil) between February 1997 and January 1998 had serum samples screened for anti-HHV-8 antibodies. Patients were considered seropositive whenever reactivity was observed in at least 1 of 3 tests (immunofluorescence assays for latent nuclear and lytic antigens and orf65 recombinant antigen enzyme-linked immunosorbent assay).

Results: Overall HHV-8 seroprevalence was 13.9% (95% confidence interval [CI], 10.9-17.6). HHV-8 coinfection was significantly more frequent in men (18.7%; 95% CI, 14.1-23.4) than in women (7.8%; 95% CI, 4.2-11.3) (P <0.001). According to the mode of HIV acquisition among males, seroprevalence of HHV-8 infection was significantly higher in men who have sex with men when compared with the other groups (32.4% vs. 10.0%, P <0.001). Multivariate logistic regression revealed HHV-8 infection among men to be independently associated with sexual orientation (adjusted odds ratio [AOR], 5.5 for homosexuals; AOR, 2.8 for bisexuals). No significant risk factor for HHV-8 infection could be demonstrated for HIV-infected women in this cohort,

Conclusions: This study provides further evidence that men who have sex with men are at higher risk of HHV-8 infection and shows that the epidemiologic pattern of this infection among HIV/AIDS patients from Santos, Brazil, is similar to that described in other countries with a low incidence of Kaposi's sarcoma.

EPIDEMIOLOGIC EVIDENCE SUGGESTING that acquired immunodeficiency syndrome (AIDS)-associated Kaposi's sarcoma (AIDS-KS) is a sexually transmitted disease led scientists to discover, in 1994, a new herpesvirus named Kaposi's sarcoma-associated herpesvirus or human herpesvirus 8 (HHV-8).¹ Since then, HHV-8 has been detected by polymerase chain reaction (PCR) in almost all KS tissues from both AIDS-associated and non-AIDS-associated KS.^2-4 Furthermore, studies showing an association between seropositivity to HHV-8 infection and risk of developing KS, as well as the demonstration of HHV-8 DNA in peripheral blood mononuclear cells (PBMC) and of serum anti-HHV-8 antibodies in HIV-infected individuals preceding the development of KS, further strengthened the epidemiologic evidence for the etiologic role of HHV-8 in KS.^5-7

Although molecular and serologic assays have been recently used to unravel the epidemiology of HHV-8 infection, the importance of the different modes of transmission of this viral agent still remains uncertain. Unlike other human herpesviruses, HHV-8 is not ubiquitous and low seropositivity rates are usually found in the United States ^8,9 and in many parts of Europe⁸ and Asia,¹⁰ whereas intermediate rates are found in Mediterranean countries ^9-11 and the highest prevalences are found in Central Africa.^8,9,12,13 Previous studies on the seroprevalence of anti-HHV-8 antibodies showed not only higher rates among male homosexuals,^9,14,15 but also demonstrated an association between seropositivity and higher number of sexual partners in this population group,¹⁶ reinforcing prior epidemiologic suspicions concerning the sexual transmission of KS.¹⁷ However, the available studies disagree about the association between particular sexual practices and risk of HHV-8 transmission.^18-20

In Brazil, although 237,588 cases of AIDS have been reported as of March 2002,²¹ only limited data concerning HHV-8 infection are available.^22-26 The aim of the current study is thus to estimate the prevalence of HHV-8 infection in a cohort of HIV/AIDS patients from Santos, southeastern Brazil, and to investigate potentially associated risk factors for this herpesvirus infection. Santos has approximately 412,000 inhabitants and the fifth highest AIDS incidence rate in the country (51.5 in 100,000 population in 1997).²¹ This survey is part of a seroepidemiologic study of bloodborne and sexually transmitted infections in HIV patients.^27,28

Materials and Methods

Study Population

The study was carried out in a cohort of HIV-infected subjects seen at the Santos AIDS Reference Center (CRAIDS). This institution provides outpatient medical care following the Ministry of Health's HIV/AIDS Program Guidelines.²⁹ Patients were enrolled either at admission to the center or during regular medical visits. Between February 1997 and January 1998, 497 patients were included after signing an informed consent, corresponding to one third of the total number of registered patients at CRAIDS at the onset of the study. Sociodemographic data (age, race, sex, educational background) and information on intravenous drug use (IDU) and sexual practices (lifetime number of sexual partners, history of anal and oral intercourse, condom use, previous sexually transmitted diseases [STDs], and sexual partnership information) were collected by means of a standardized questionnaire. Information on sexual orientation among males was collected regardless of the informed mode of HIV acquisition. Patients' medical records were reviewed to search for information on staging of HIV disease (as defined by the Centers for Disease Control and Prevention [CDC]).³⁰ The study was approved by the Ethics Committee of the School of Medicine, University of São Paulo.

Serologic Assays for Detection of Anti-Human Herpesvirus 8 Antibodies

Serum samples at 1:40 dilution were screened for antibodies to HHV-8 latency-associated nuclear antigen (IFA-Lana) and lytic phase antigens (IFA-Lytic) through immunofluorescence assays (IFA) using the BCBL-1 cell line.^31,32 Both cell lines were supplied by Dr. Niel Constantine from the Institute for Human Virology, Baltimore, Maryland. Viral lytic cycle was induced by incubating BCBL-1 cells with 20 ng/mL of 12-o-tetradecanoylphorbol-13-acetate (TPA) for 96 hours. Punctuate nuclear staining in untreated BCBL-1 cells was considered as a positive result for antibodies to LANA, whereas entire cell fluorescence in approximately 20% of TPA-treated cells without staining of Ramos cells was considered as a positive result for antibodies to the lytic-phase antigens. In case unspecific staining was observed, sera were retested after absorption with a 10% Hep2 extract for 30 minutes at room temperature. Antibodies to the HHV-8 capsid protein were detected by an in-house enzyme-linked immunosorbent assay (ELISA) using as antigen an orf65 recombinant protein, as previously described.⁸ Positive and negative control sera were run in parallel to the studied samples in every test. These tests have been previously validated by our laboratory team, demonstrating anti-Lana, anti-Lytic, and antiorf65Ag antibodies in 71.4%, 95.2%, and 92.8% of KS patients, and in 0%, 2.4%, and 6.1% of asymptomatic blood donors from São Paulo, respectively.³³ This diagnostic performance was similar to that reported by other researchers in these populations.^8,34

HHV-8 infection-seropositivity was defined as seroreactivity to viral antigens in at least 1 of the used serologic assays (IFA-Lana, IFA-Lytic, or ELISA).

Serologic Investigation of Other Blood or Sexually Transmitted Agents

To investigate bloodborne or sexually transmitted coinfections, blood samples were further screened for serologic markers for hepatitis B virus (HBV) (HBs Ag and anti-HBc), hepatitis C virus (HCV) (anti-HCV), human T-cell lymphotrophic virus (HTLV) (HTLV-I/II, and syphilis using standard immunoenzymatic assays (Enzymum-test, Boehringer-Mannheim for HBV and HCV; Murex HTLV-I+II GE80/81, Murex Diagnostics for HTLV-I/II; and TP-Hemagen for anti-Treponema IgG). HTLV-I/II infection was confirmed by Western blotting (GLD HTLV Blot 2.4; Genelabs Diagnostics).

Statistical Analysis

Data were entered using Epi Info version 6.0 (CDC, Atlanta, GA) and analyzed using STATA version 6.0 software. Seroprevalence rates of HHV-8 infection and their respective 95% confidence intervals (95% CIs) were calculated using normal approximation to the binomial distribution for each mode of HIV acquisition and age group. HHV-8 seropositivity was considered as the dependent variable. Association with independent variables was measured in univariate and multivariate analysis using chi-squared or chi-squared for trend tests (for both schooling and lifetime number of STD episodes). Adjusted OR (AOR) by age in univariate and by age and schooling in multivariate analysis and their respective 95% CIs were calculated. Multivariate analysis was carried out based on logistic regression models, including all P <0.25 variables in univariate analysis.³⁵ Statistical significance was set at P <0.05.

Results

Sociodemographic Characteristics and Modes of Exposure to HIV

The median age in this cohort was 35.0 years (range, 16-68 years) and 278 participants (55.9%) were male. Approximately half (49.1%) reported less than 8 years of schooling. A total of 72.2% of patients were classified as white; 27.8% were classified as nonwhite, which included blacks, mulattos, and Asians; no Native Americans were included. Only 9.4% of men and 8.6% of women reported consistent condom use, whereas 62.2% of men and 37.4% of women reported a history of STDs. As far as HIV disease progression is concerned, 24.2% were classified as being in stage A, 32.2% in stage B, and 43.6% in stage C according to CDC classification.³⁰

The large majority (83. 6%) of HIV-infected women reported HIV acquisition by heterosexual exposure and 16.4% informed sexual partnership with bisexual males. The mode of HIV acquisition among men was believed to have been the result of sexual exposure exclusively in 69.4% of subjects (64 homosexuals, 44 bisexuals, and 85 heterosexuals).

Human Herpesvirus 8 Seroprevalence

Sixty-nine of 497 participants were seropositive for HHV-8 infection, yielding an overall prevalence of 13.9% (95% CI, 10.9-17.6). However, seroprevalence differed significantly according to sex: 18.7% (95% CI, 14.1-23.4) among men and 7.8% (95% CI, 4.2-11.3) among women (P <0.001). Although women reporting IDU had 2-fold HHV-8 seropositivity (14.3%) when compared with women who acquired HIV by heterosexual contact (7.2%), and none of those who acquired HIV by blood transfusions had anti-HHV-8 antibodies, the prevalence of HHV-8 seropositivity did not differ significantly among these 3 modes of HIV acquisition (P = 0.29). Among males, on the other hand, HHV-8 infection was significantly more frequent in men who have sex with men (MSM) when compared with the other groups (32.4% vs. 10.0%, P <0.001) (Table 1).

Table 1

Risk Factors for Human Herpesvirus 8 Infection

In women, HHV-8 infection was not associated with the studied variables (Table 2). Furthermore, HHV-8 female seropositivity was not associated with AIDS diagnosis (AOR 0.8 for women with AIDS; 95% CI, 0.3-2.3) or with lowest CD4+ cell count below 200/mm³ (AOR 0.7; 95% CI, 0.2-2.2). Logistic regression analysis, with ORs adjusted for age, schooling, and all others variables in the final model, showed no statistically significant association between HHV-8 seropositivity and the variables tested, including race (AOR, 2.4; 95% CI, 0.8-6.9; P = 0.13), condom use (AOR, 0.3; 95% CI, 0.1-1.2; P = 0.11), and syphilis (AOR, 1.9; 95% CI, 0.5-6.8; P = 0.34).

Table 2

Univariate analysis aimed at investigating potential risk factors for HHV-8 coinfection among men revealed a positive association with higher education (AOR, 2.3; 95% CI, 1.1-4.1) for men with more than 8 years of schooling, which corresponds to elementary education in Brazil. Infection with HHV-8 was also shown to be significantly more frequent among MSM, including homosexuals (AOR, 6.0; 95% CI, 2.7-13.2) and bisexuals (AOR, 3.3; 95% CI, 1.4-8.3). Moreover, anal receptive intercourse and sexual partnership with foreigners were also positively associated with HHV-8 seropositivity (AOR, 2.8; 95% CI, 1.4-5.5; and 2.4; 95% CI, 1.2-5.0, respectively), whereas IDU and HCV seropositivity were shown to be protective factors (AOR, 0.4; 95% CI, 0.2-0.8; and AOR, 0.3; 95% CI, 0.1-0.6, respectively) (Table 3). In addition, HHV-8 seropositivity did not differ according to HIV disease staging (AOR, 1.2 for men with AIDS diagnosis; 95% CI, 0.6-2.4) and to lowest CD4+ cell count below 200/mm³ (AOR, 0.8; 95% CI, 0.4-1.6). In multivariate analysis, sexual orientation was independently associated with HHV-8 infection among men in this cohort of HIV/AIDS outpatients (Table 4). Variables receptive anogenital intercourse, sexual partnership with foreigners, and sex with commercial sex workers included in the final logistic regression model were not associated with HHV-8 seropositivity, whereas HCV seropositivity was shown to be inversely associated with HHV-8 infection. History of IDU was not included in the final logistic regression model, because we considered both this variable and HCV seropositivity as correlates of risk for blood-transmitted infection. In fact, 88.9% of IDU were HCV-seropositive in this cohort (P <0.001).

Table 3

Table 4

Discussion

The overall prevalence of HHV-8 infection found in this cohort of HIV/AIDS patients from Santos is similar to those reported from HIV-infected populations in northwestern Europe and North America, where KS is a rare disease and is frequently associated with HIV infection.^{8,9,14,15,31,34} The fact that HIV-infected individuals are at higher risk of being coinfected with HHV-8 is expected, because both viral agents share common modes of transmission. Previous studies have reported a low frequency of HHV-8 infection among Brazilian blood donors from different regions. Based on the IFA-Lana assay, HHV-8 seroprevalence among blood donors varied from 4.0% to 7.4%.^24,26,36 In a previous study published by our laboratory team, using the same assay, no HHV-8 seropositivity was found among 275 blood donors.²² In contrast, higher prevalences of HHV-8 infection have been demonstrated in the Brazilian HIV-infected population, varying from 15% to 26%.^22-26

The overall prevalence of anti-HHV-8 antibodies among women in our cohort was significantly lower than among men. When different modes of HIV acquisition are compared in males, HHV-8 seroprevalence rates are higher among MSM when compared with heterosexuals, IDU, and blood transfusion recipients. These results are in agreement with reports from countries where the prevalence of HHV-8 infection in the general population is low,^8,23 suggesting that blood transmission of this herpesvirus is rare.

No significant association between HHV-8 infection and age was observed in our cohort, suggesting that this herpesvirus has been recently introduced in the studied population, as evidenced by others.¹⁸ Conversely, in Italy, where higher rates of endemic KS and HHV-8 infection are recognized in the general population, increasing anti-HHV-8 antibody prevalences with older age have been reported.³⁷

As far as HIV-infected women in this cohort are concerned, both univariate and multivariate analyses failed to identify associated risk factors for HHV-8 infection, probably as a result of the small number of subjects in each subgroup, what should not be interpreted as evidence against sexual transmission of HHV-8 among women. Further investigation involving larger population samples, providing stronger statistical power, is required to evaluate potential risk factors associated with HHV-8 transmission in women.

Regarding males, HHV-8 seroprevalences were not significantly different in terms of educational background. However, men with longer schooling (more than 10 years) had higher frequencies of HHV-8 infection. Homosexuality may have been a confounder in this regard, because 42.9% of men with longer schooling (>10 years) among the studied population also reported homosexual behavior (P <0.001).

Consistent with our study are results indicating that the prevalence of HHV-8 antibodies is higher among MSM than among heterosexual men.^14,15,37 In Brazil, 1 smaller study has reported 18% of HHV-8 infection among HIV-infected homosexual men,²³ whereas others have found higher rates of HHV-8 seropositivity among MSM, ranging from 30% to 87%.^24,26 High rates of HHV-8 in MSM and the fact that HIV-infected individuals who develop KS are mostly homosexuals strongly indicate that HHV-8 is transmitted through homosexual practices. This evidence is further supported by the observation that sporadic cases of KS in HIV-negative homosexuals occur more frequently than in the general population.³⁸

Although our results demonstrate the importance of sexual transmission of HHV-8, we failed to show an association between HHV-8 seropositivity and specific sexual practices among males, including orogenital sex, sexual partnership with foreigners or with commercial sex workers, lifetime number of sexual partners, lifetime number of STDs episodes, and condom use. In contrast to previous studies,^18,39 receptive anogenital intercourse was not shown to be independently associated with HHV-8 infection in our final regression model, although it was associated with HHV-8 seropositivity in univariate analysis. Study limitations might explain the lack of association between HHV-8 seropositivity and the investigated sexual practices. Because our study was based on current sexual behavior, it may not have elicited responses that sufficiently reflect past sexual practices. Moreover, sexual behavior may have been influenced by HIV/AIDS diagnosis, and thus current sexual behavior of individuals from the cohort may not reflect their practices in the relevant period. Additionally, self-reported information on IDU or on sexual practices considered socially unacceptable in certain sociocultural contexts may be biased and therefore underestimated. Finally, individual behavioral risk factors that are associated with HHV-8 infection may not have been fully captured by the investigated variables.

An important issue to be addressed is the routes of HHV-8 acquisition in the studied cohort. We thus investigated the association between HHV-8 seropositivity and serologic evidence of exposure to other blood- or sexually transmitted infections (HBV, HCV, HTLV, and syphilis). HCV seropositivity was shown to be inversely associated with HHV-8 infection in our final regression model. This evidence supports the hypothesis that these viruses circulate by different routes in the studied population and suggest that HHV-8 acquisition through blood is not important.

Our finding that HHV-8 seropositivity was not related to the stage of HIV-1 disease and CD4+ cell counts is consistent with results of previous studies.^8,40 Unfortunately, although almost half of the patients from this cohort were in CDC stage C (43.6%), no information on history of KS was available. However, the Brazilian nationwide prevalence of KS is low, and from 1988 to 1998, only 0.9% of AIDS reported cases had KS.⁴¹

Finally, current limitations on HHV-8 serodiagnosis should be discussed. Studies on the prevalence of HHV-8 infection have been hindered by the absence of a well-defined, standardized algorithm for the serologic diagnosis of this viral disease.⁴² Although different available tests for detection of anti-HHV-8 antibodies present relatively high sensitivities when applied to patients with KS, when dealing with asymptomatic infection, tests have been reported to exhibit uncertain accuracy,^43,44 in part because they identify antibodies to antigens expressed at different time points during the viral cycle.^8,45,46 Therefore, in an attempt to avoid false-negative results, we have defined HHV-8 seropositivity based on reactivity in at least 1 of 3 different serodiagnostic techniques.

In conclusion, the current study provides evidence that HIV-positive MSM are at greater risk of HHV-8 coinfection and that the epidemiology of this infection in Santos, Brazil, is similar to that described in other countries that present a low incidence of KS. Understanding the potential transmission of HHV-8 through unprotected sex reinforces the need to promote safer sexual practices. Identifying population groups at higher risk of acquiring HHV-8 infection and the main route of viral transmission will certainly be useful in the establishment of adequate preventive strategies against the dissemination of HHV-8 infection in Santos.