Human Herpesvirus 8 Infection Among Adolescents in the REACH Cohort

Objective  Human herpesvirus 8 (HHV-8) infection is common among children in areas where Kaposi sarcoma is endemic. Human herpesvirus 8 is uncommon in children but prevalent in adults at risk for human immunodeficiency virus (HIV) infection in the United States, including men who have sex with men (MSM) and women who engage in high-risk sexual behavior. We examined the prevalence and predictors of HHV-8 infection among adolescents with or at high risk for acquiring HIV infection.

Design  Cross-sectional analysis.

Setting  National study of HIV infection among adolescents in primary care.

Participants  A total of 537 young adults practicing high-risk sexual behavior, of which 403 were women and 134 were men; among the 134 men, 75% were MSM.

Interventions  Detailed questionnaires and testing for serum antibodies to HHV-8.

Outcome Measure  Detection of serum antibodies to HHV-8.

Results  Sixty (11.2%) of 537 young adults were HHV-8 seropositive, including 20 MSM (19.6%), 2 male heterosexuals (6.5%), and 27 female heterosexuals (8.2%). The prevalence of HHV-8 in HIV-positive MSM (17/74 [23.0%]) was twice as high as that in HIV-negative MSM (3/28 [10.7%]) (P = .18), but no characteristic predicted HHV-8 infection among MSM. In multivariate analysis, history of gonorrhea (odds ratio [OR], 2.8; 95% confidence interval [CI], 1.4-5.7; P<.01), history of having sex with women (OR, 2.4; 95% CI, 1.1-5.3; P = .03), and African American race (OR, 3.4; 95% CI, 1.1-10.0; P = .03) were associated with HHV-8 infection among women.

Conclusions  Human herpesvirus 8 is common among US adolescents practicing high-risk sexual behaviors. Sexual identity, race, and sexual behavior may influence the risk of infection with HHV-8 in women.

Human herpesvirus 8 (HHV-8) (also known as Kaposi sarcoma [KS]–associated herpesvirus) was first identified in 1994.¹ In the United States and western Europe, the virus is likely acquired through sexual contact and is largely restricted to men who have sex with men (MSM)² and women at high risk for sexually transmitted or parenteral infections.³ Infection is rare among US children⁴ but ranges from 30% to 100% in African children, who typically acquire HHV-8 before the onset of puberty but after the perinatal period.^5- 7 In many parts of Africa, KS is a tumor endemic among children with and without human immunodeficiency virus (HIV) infection,^8- 10 but most HHV-8 infections worldwide are asymptomatic.

Human herpesvirus 8 is shed frequently in the oropharynx.¹¹ Contact with saliva during sex is associated with incident HHV-8 infection in adult MSM,^11- 14 whereas household contact with saliva may be a mode of HHV-8 transmission in African children.¹⁵ Children contract other HHV infections through contact with saliva, including herpes simplex virus type 1 (HSV-1),¹⁶ cytomegalovirus,¹⁷ and Epstein-Barr virus,¹⁸ which are all ubiquitous worldwide. If some or all HHV-8 transmission is explained by contact with saliva, it remains unclear why HHV-8 infection is not more prevalent in most parts of the developed world.

We hypothesized that HIV-positive adolescents and young adults and those practicing high-risk sexual behavior may be at substantial risk of HHV-8 infection because of homosexual contact and multiple sex partners. We therefore examined the seroprevalence and predictors of HHV-8 infection among such individuals.

The REACH (Reaching for Excellence in Adolescent Care and Health) project recruited adolescents aged 13 to 18 years from 15 clinical sites in 13 US cities to evaluate the biomedical and behavioral characteristics of HIV infection.¹⁹ Eligible adolescents either had acquired HIV sexually or parenterally or were HIV negative but with similar age and risk behaviors. Participants or their guardians provided informed consent, as approved by each site's institutional review board and human experimentation guidelines of the US Department of Health and Human Services. Subjects underwent physical examination, questionnaire collection, and phlebotomy at 3-month intervals for HIV-positive and at 6-month intervals for HIV-negative persons.

A combined whole-virus enzyme immunoassay with immunofluorescence assay detected serum HHV-8 antibodies from the last study visit available.²⁰ Antigens derived from purified viral lysate (Advanced Biotechnologies Inc, Columbia, Md) are plated on 96-well plates. After application of human serum, bound antibodies are detected by horseradish peroxidase–conjugated goat anti–human IgG Fc (Chemicon, Temecula, Calif) and TMB (3,3′,5,5′-tetramethylbenzidine) substrate (KPL Inc, Gaithersburg, Md). Optical densities are determined in a standard enzyme-linked immunosorbent assay plate reader. Samples with intermediate enzyme immunoassay values are rerun with a standard immunofluorescence assay for antibodies to latent and lytic antigens in body cavity–based B-cell lymphoma 1 cells. This testing strategy was found to be 88% sensitive and 97% specific for the identification of HHV-8 infection.

Laboratory tests for gonorrhea, chlamydia, and syphilis were performed biannually. For HIV-positive participants, the AIDS Clinical Trials Group laboratories determined T-cell lymphocyte counts by flow cytometry, and HIV plasma RNA levels were quantified by nucleic acid sequence-based amplification (Organon Technika, Westchester, Pa).

For each potential predictor of HHV-8 infection, we selected data from the period closest to the date of HHV-8 antibody assessment, with the exception of some characteristics collected only at baseline.

The analysis was restricted to participants previously identified as having a high HHV-8 prevalence among adults, that is, MSM (defined by self-identification as being homosexual or bisexual or reporting any history of anal receptive sex) and all women (who by inclusion in the cohort were at risk for or had HIV infection). Men reporting exclusively heterosexual partners and behaviors were excluded.

Statistical analysis was performed using SAS for Windows Version 8.2 software (SAS Institute Inc, Cary, NC). Associations with HHV-8 status were evaluated using univariate and multivariate logistic regression. Multivariate model selection proceeded by including all factors univariately associated with HHV-8 at P<.1 and eliminating factors until all remaining predictors were significant at P≤.05.

Of the 550 adolescents enrolled in the REACH project, 537 (97.6%) had serum samples available for HHV-8 testing. Participants had a median follow-up time of 33 months (range, 0-57 months) prior to the HHV-8 serologic assessment. Of the 537 participants, 134 (25.0%) were men and 403 (75.0%) were women. Among the 134 men, 102 (76.1%) were classified as MSM.

The median age of participants at the time of HHV-8 serologic assessment was 19 years (range, 18-21 years), with most (71.1%) identifying themselves as African American (Table 1). A greater proportion of women study participants were African American when compared with MSM (75.4% of women vs 53.9% of MSM; P<.001).

Most (82.1%) of the women identified themselves as heterosexual (Table 1), whereas 47.8% of the men identified themselves as homosexual. Men who have sex with men were more likely to report a history of sex with a commercial sex worker (20.6% of MSM vs 7.9% of women [P<.01]) but had similar exposure to HIV-positive sex partners (24.5% of MSM vs 19.6% of women [P = .34]). Anal sex was common among both MSM and women. At the time of HHV-8 serologic testing, 12.7% of women and 8.8% of MSM had a concurrent diagnosis of chlamydia; 5.2% of women and 3.9% of MSM, of gonorrhea; and 2.2% of women and 3.9% of MSM, of syphilis. Intravenous drug use was uncommon among the study youth, reported by 0.5% of women and 6.9% of MSM.

At the time of enrollment, 64.3% of women and 70.6% of MSM were HIV-positive, and an additional 2 MSM acquired HIV during the study. The median CD4 count was 601 cells/μL (interquartile range, 398-839 cells/μL), and the median HIV plasma RNA level was 3.9 log copies/mL. Only 7 participants reported having an AIDS-defining illness²¹ and no participant had KS.

Overall, 60 (11.2%) of 537 participants were HHV-8 seropositive, including 20 (19.6%) of 102 MSM, 2 (6.5%) of 31 heterosexual men (data not shown), and 38 (9.4%) of 403 women (χ² test, P<.01). The Figure depicts the prevalence of HHV-8 infection by demographic group and HIV status (26 [10.0%] of 259 HIV-positive women vs 12 [8.3%] of 144 HIV-negative women; odds ratio [OR], 1.2; 95% confidence interval [CI], 0.6-2.5; P = .58; 1 [7.1%] of 14 vs 1 [5.9%] of 17 heterosexual males; OR, 1.2; 95% CI, 0.1-23.8; P = .89; 17 [23.0%] of 74 vs 3 [10.7%] of 28 MSM; OR, 2.5; 95% CI, 0.7-9.4; P = .18). There was no significant difference in the prevalence of HHV-8 infection among HIV-negative women and HIV-negative MSM (8.3% vs 10.7%, P = .68). In a univariate analysis, only 3 factors were significantly associated with an increased risk of HHV-8 infection among women, including a history of gonorrhea (OR, 2.8; 95% CI, 1.4-5.7; P = .002) and sex with other women (WSW) (OR, 2.4; 95% CI, 1.1-5.3; P = .03) (Table 2). In a multivariate model restricted to female participants, these factors remained significantly associated with an elevated risk of HHV-8 infection. No factors were significantly associated with HHV-8 infection among MSM.

In subset analyses of 331 HIV-positive subjects separated by sex, neither HIV RNA level (modeled continuously, OR, 1.4; 95% CI, 0.9-2.1; P = .12 among women, and OR, 0.8; 95% CI, 0.4-1.4; P = .40 among MSM) nor highly active antiretroviral treatment (HAART) use (OR, 0.5; 95% CI, 0.1-1.2; P = .15 among women, and OR, 1.5; 95% CI, 0.5-4.8; P = .46 among MSM) was associated with HHV-8 infection.

A decade after the initial identification of HHV-8, many questions still remain about its epidemiology.²² The restriction of HHV-8 infection to specific geographic and demographic “pockets” is perplexing given the evidence for the transmission of HHV-8 through saliva and the comparative ubiquity of other HHVs.¹² We found antibodies to HHV-8 in 11.2% of adolescents with HIV infection and/or high-risk sexual and substance use behavior. Our results provide several important insights into the epidemiology of HHV-8 in adolescents.

In the United States and western Europe, 2% to 10% of the general population have serum antibodies to HHV-8,²³ but little is known about the timing of HHV-8 acquisition. One previous study²⁴ found a low rate of infection among MSM aged 16 to 22 years in the United States, which was comparable to the rate in a matched population of young heterosexual males and females.

We found the highest rates of HHV-8 infection in HIV-positive adolescent MSM, but this rate is lower than the 30% to 75% prevalence previously observed in populations of adult HIV-positive MSM.^14,25 In a cross-sectional study, the relationship between the timing of HHV-8 and HIV infection could not be assessed. The lower prevalence of HHV-8 in HIV-positive adolescent MSM compared with adult MSM suggests that they acquire HIV prior to HHV-8. This may be clinically relevant because the incidence of KS appears to be higher among MSM who acquire HHV-8 after HIV infection compared with those first infected with HHV-8.²⁶

The rarity of KS among women in the pre-HAART era led many to speculate that HHV-8 infection was uncommon among women.²⁷ Recent data, however, suggest rates that are 2 to 4 times higher than the rate among the general population of the United States and similar to that among HIV-negative MSM.^3,28,29 We identified 2 novel risk factors for HHV-8 infection among young women.

We found a high HHV-8 prevalence among WSW but did not specifically ask about their behaviors. Previous studies offer potential explanations for the higher prevalence in this group. First, WSW in this study infrequently reported exclusively female sex partners, and WSW are more likely to have sex with bisexual males.^30- 32 Second, some WSW practice riskier sexual behavior, including concurrent substance abuse with sex.³² Third, exposure of the oral cavity to genital or oral secretions during sex is nearly universal among WSW³³ and the use of protective barriers is uncommon.³⁴ Human herpesvirus 8 infects epithelial cells, which may be the site of primary infection.^11,35 Limited studies in women find HHV-8 in oral and genital secretions.^36- 39 Whether HHV-8 shed in saliva is transmissible through other mucosal sites, such as the genital epithelium, has yet to be determined. Taken together, it is possible that sexual mixing between WSW and other demographic groups with high HHV-8 prevalence may allow transmission to female partners during oro-oral or orogenital sex.

The association between HHV-8 and gonorrhea has been reported in both men^40,41 and women,⁴² but the small number of participants in our study infected with both HHV-8 and gonorrhea precluded a comprehensive analysis. Sexual networks could also help explain the relationship between HHV-8 prevalence and gonorrhea in women because previous studies⁴³ have found that persons infected with Neisseria gonorrhoeae have larger, more demographically discordant sexual networks.

Our observed association between African American race and HHV-8 infection has been corroborated by 2 other cohorts of adult women^3,28 and, to our knowledge, has not been described in men.⁴⁴ Studies of sexual networks among African American adolescents show that most report their typical sex partner to be older and sex with these partners to be often unprotected.⁴⁵ This “disassortive mixing” by age increases the risk of sexually transmitted infections in adolescents⁴⁶ and of HSV-2 infection in adults.⁴⁷

The findings of this study must be taken in the context of the included study population. Specifically, the relatively small number of MSM limits the ability to detect predictors of HHV-8 infection among this group. Conversely, the overrepresentation of female or African American youths may have allowed us to identify previously uncharacterized risks for HHV-8 in these groups. Finally, the lack of information on behaviors specifically associated with the exchange of saliva is a limitation of previous studies and our current investigation.

In sum, we found HHV-8 infection to be common among US adolescents practicing high-risk sexual behavior and identified novel risk factors for HHV-8 infection in females. The incidence of KS in the United States and western Europe has declined dramatically in the HAART era,⁴⁸ with little change in the prevalence of HHV-8 infection.⁴⁹ The long-term sequelae of infection with HHV-8, a known oncogenic virus, are still undetermined, and specific therapy for HHV-8 infection is lacking. Consequently, prevention of HHV-8 infection remains important, and future studies should examine the mode of HHV-8 transmission and acquisition.

Correspondence: Corey Casper, MD, MPH, Program in Infectious Diseases, Fred Hutchinson Cancer Research Center, 1100 Fairview Ave N, Mail Stop D3-100, Seattle, WA 98109 (ccasper@u.washington.edu).

Accepted for Publication: April 27, 2006.

Author Contributions:Study concept and design: Casper, Wald, Corey, and Moscicki. Acquisition of data: Morrow and Moscicki. Analysis and interpretation of data: Casper, Meier, Wald, Corey, and Moscicki. Drafting of the manuscript: Casper, Meier, and Moscicki. Critical revision of the manuscript for important intellectual content: Meier, Wald, Morrow, Corey, and Moscicki. Statistical analysis: Casper and Meier. Obtained funding: Corey and Moscicki. Administrative, technical, and material support: Casper, Morrow, Corey, and Moscicki. Study supervision: Wald, Morrow, and Corey.

Funding/Support: This study was supported in part by National Institutes of Health (NIH) grant U01-HD32830 from the National Institute of Child Health and Human Development with cofunding from the National Institutes on Drug Abuse, of Allergy and Infectious Diseases, and of Mental Health; NIH grants K23 AI54162 and U19 AI31448 from the National Institute of Allergy and Infectious Diseases; and a Clinical Scientist Development Award from the Doris Duke Charitable Foundation.

Acknowledgment: We thank the investigators and staff of the Adolescent Medicine HIV/AIDS Research Network (1994-2001) (listed in the Journal of Adolescent Health [2001;29(suppl):5-60]), and the adolescents and young adults who participated in the REACH project, for their valuable contributions.