Des Jarlais, Don C. PhD*; Hagan, Holly PhD†; Arasteh, Kamyar PhD*; McKnight, Courtney MPH*; Perlman, David MD‡; Friedman, Samuel R. PhD†
GENITAL HERPES SIMPLEX VIRUS 2 (HSV-2) has been associated with increased transmission of human immunodeficiency virus (HIV) in a number of studies.1,2 Two recent meta-analyses found average risk estimates from 2 to 3 for acquiring HIV among persons with prevalent HSV-2 infection when compared with persons without HSV-2 infection.3,4 HSV-2 is believed to increase susceptibility to HIV through genital lesions that provide portals for HIV entry, and to increase transmissibility of HIV through increasing HIV viral shedding. Freeman and colleagues suggested that, “in populations with high HSV-2 prevalence, this virus may contribute a substantial proportion of the attributable risk for HIV infection.”3
HSV-2 seroprevalence has been assessed in a modest number of studies of persons who use illicit drugs (primarily heroin and cocaine).5–7 HSV-2 seroprevalence ranged from 44% to 65% among subjects in these studies. Higher HSV-2 prevalence was associated with older age, non-white ethnic status, being female, being a male who engaged in sex with other males (MSM), having multiple sex partners, and having a history of other sexually transmitted infections. In the same studies, HIV seroprevalence ranged from 4% to 37%. One study suggested a possible relationship between HSV-2 seroprevalence and HIV seroprevalence. In the study of young drug users in Baltimore by Plitt et al.,7 HSV-2 seroprevalence was associated with HIV prevalence among male but not female drug users. All of these studies, however, included both injecting and noninjecting drug users. As injecting drug users may have become infected with HIV through sharing drug injection equipment, it is difficult to assess the extent to which HSV-2 might be contributing to sexual HIV transmission without separate analyses of injecting and noninjecting drug users.
We report here on the risk for HIV and HSV-2 seroprevalence among noninjecting heroin and cocaine users in New York City. We found a high HIV prevalence, a very high HSV-2 prevalence, and a consistent association between HIV and HSV-2 prevalence. Noninjecting drug users in New York City may exemplify the potential contribution of HSV-2 to sexual transmission of HIV in industrialized countries.
The data reported here is part of an ongoing series of studies of drug users entering the Beth Israel Medical Center drug detoxification program in New York City. The methods were previously described in detail,8,9 so only a summary will be presented here. The Beth Israel detoxification program serves the city as a whole; approximately half of its patients live in Manhattan, one quarter in Brooklyn, one-fifth in the Bronx, and the rest (5%) elsewhere. The program is quite large, with 5000–7000 admissions per year. Almost all patients enter the program on a voluntary basis, because they can no longer manage their drug use.
Both injecting and noninjecting drug users are eligible to participate in the study. For this report, however, we used data only from persons who reported using heroin and/or cocaine in the 6 months before the interview and that they had never injected an illicit drug. The period of data collection for this report was from fall 2005 to spring 2006. Patients in the detoxification program were recruited in a manner to minimize both staff selecting possible subjects and volunteering bias. Within gender, patients are admitted to specific wards of the program on a nonselective basis (open beds). Research staff visited the general admission wards of the program in a preset order. Research staff examined all intake records of a specific ward and constructed a list of eligible patients admitted within the past 3 days. All of the patients on the list for the specific ward were then asked to participate in the study. (Patients admitted to the ward more than 3 days before constructing the list or after the list was constructed were not asked to participate, nor were patients admitted to other wards during the 3-day period.) Among patients approached by our interviewers, willingness to participate was over 95%. After all of the patients admitted to a specific ward in the 3-day period had been asked to participate and interviews had been conducted with those who agreed to participate, the interviewer moved to the next ward in the preset order. Data collection was continuous throughout the study period.
A structured questionnaire covering demographics, drug use, injection, sexual risk behavior, and use of HIV prevention services was administered by a trained interviewer. Risk behaviors were assessed for the 6 months before the interview using audio computer-assisted self-interviewing. After completion of the interview, the participant was referred to an HIV counselor for pretest counseling and serum collection. A separate informed consent was obtained for HIV testing. HIV testing was conducted at the New York City Department of Health Laboratory using repeated enzyme-linked immunosorbent assays (ELISA) testing with Western blot confirmation. HSV-2 testing was conducted by Bio-Reference Laboratories using the HerpeSelect HSV-1 and HSV-2 ELISA (Focus Technologies, Cypress, CA).
Associations between HSV-2 and HIV were examined with prevalence ratios and with a multivariate Poisson regression with a robust error variance.10 For the female subjects, where all HIV seropositives were also HSV-2 seropositive, the P value was calculated with an exact binomial test, and the risk ratio was estimated using a correction of 0.5.11 The SAS statistical software12 was used.
Table 1 presents selected demographic and drug use characteristics of the sample. The sample had a mean age of 41, was 74% male, and 5% white, 65% black, and 29% Latino/a. Smoking crack cocaine, intranasal use of heroin, and intranasal use of powder cocaine were the most commonly reported types of drug use for the 6 months before the interview. Multiple forms of cocaine and heroin use were quite common, 45% of the subjects reported 2 or more methods of using cocaine and heroin (without injecting).
Table 2 shows HIV and HSV-2 seroprevalence for the entire sample and within demographic subgroups (gender, age, race/ethnicity). HIV prevalence was 19% (95% CI 15%–24%) and HSV-2 seroprevalence was 60% (95% CI 55%–64%) for the sample as a whole. There were no statistically significant differences in HIV prevalence by demographic subgroup. Older subjects, female subjects, and black subjects were more likely to be HSV-2 seropositive.
We examined the relationship between HSV-2 and HIV among these subjects using prevalence ratios and a modified multivariate Poisson regression approach. The unadjusted risk (prevalence) ratio for HSV-2 seropositive status as a predictor of HIV seropositive status was 1.9 (95% CI 1.21–2.98) (Table 3). In a multivariate analysis, adjusting for age, gender, and race/ethnicity as potentially confounding variables, the adjusted risk ratio was 1.87 (95% CI 1.18–2.97), almost identical to the unadjusted risk ratio. We also examined the consistency of the unadjusted risk ratios of HSV-2 as a risk factor for HIV among the demographic subgroups. The risk ratios among the demographic subgroups clustered around 2.0, with the exception of the risk ratio among the females. All 26 of the HIV seropositive females were also HSV-2 seropositive. This is significantly different from expectation (P <0.003; by exact binomial test comparing to the 82% prevalence rate among HIV seronegative females). We used the Cochran-Mantel-Haenszel test with a correction of 0.5 for the zero cell count to estimate a risk ratio (RR = 10.01, 95% CI = 0.62–161.43).
Because the Focus ELISA we used may lack specificity when the index value is less than 3.5, we examined the number of HSV-2 samples with index values less than 3.5 and redid the calculations after excluding these samples. Sixteen percent of our HSV-2 positive samples (11% of all samples) were in this range. When these possibly nonspecific samples were excluded from the analyses, the HSV-2 prevalence was 56%, the HIV prevalence was 17%, and the risk ratio was 1.7. These values are not meaningfully different from the 60%, 19%, and 1.9 values when the potentially nonspecific specimens were included.
Among the 59 subjects who reported male-with-male sexual activity in the 5 years before the interview, HIV prevalence was 31%, and HSV-2 seroprevalence was 59%, and the odds ratio for HSV-2 as a risk factor for HIV was 2.25, very similar to the odds ratios within other demographic subgroups. Almost half (28 of 59, 47%) of the subjects who reported male-with-male sexual activity in the 5 years before the interview also reported vaginal and/or anal intercourse with females in the 6 months before the interview; 34% reported intercourse with a primary female partner, 27% reported intercourse with casual partners, and 7% reported exchanging sex for money or drugs with female partners (with the MSM subject providing sexual services and receiving drugs or money in return).
We examined possible age relationships with HSV-2 and HIV seroprevalence among both males and females. Among the males, there was a clear age gradient for HSV-2, with 20% prevalence among those 30 and younger, 41% prevalence among those 31–40, 62% among those 41–50, and 63% among those more than 50 (P <0.01 by Cochran-Armitage for trend). There was also a significant age gradient for HSV-2 among the females. Five of the 9 (55%) females aged 30 and younger were HSV-2 seropositive, but among those aged 31–40 HSV-2, 88% were HSV-2 seropositive, among those aged 41–50, 87% were HSV-2 seropositive, and among those aged 51 and older, 100% (7 of 7) were HSV-2 seropositive (P <0.05 by Cochran-Armitage test for trend). There were no significant age trends for HIV among either females or males. (Data not presented, available from the first author.)
We also examined heterosexual activity (vaginal or anal intercourse with a member of the opposite sex) and “unsafe sex” (vaginal or anal intercourse without always using a condom) by gender and by HIV and HSV-2 serostatus. The HIV seropositive and the HSV-2 seropositive subjects were likely to have been infected with these viruses well before the 6-month period before the interview, so that the data on sexual behavior should not be interpreted as indicating how the subjects may have become infected. Rather the sexual behaviors indicate the percentages of uninfected subjects who are at risk of becoming infected and the percentages of infected subjects who are at risk of transmitting the viruses to uninfect others. Results are presented in Table 4. A majority of the subjects were sexually active, particularly with primary partners of the opposite sex. A substantial percentage of the subjects also reported engaging in unprotected vaginal or anal intercourse, most often with their primary partners. Many men reported sex with casual partners of the opposite sex and females reported engaging in commercial sex. The percentages of subjects reporting unsafe sex in these types of relationships tended to be low but not negligible. Among both men and women, there was a consistent pattern that subjects who were HIV seropositive were less likely to be sexually active and less likely to report engaging in unsafe sex when compared with subjects who were HIV seronegative. This pattern did not hold for HSV-2 serostatus, where there were no consistent differences in behavior between seropositives and seronegatives.
The HIV prevalence—19%—that we observed in this sample of noninjecting cocaine and heroin users in New York City was quite high. It is as high as the HIV prevalence among current injecting drug users in the city.13 (That article contains additional data on characteristics of noninjecting drug users in New York City.) HIV prevalence among injecting drug users in this study was also higher than the prevalence among injecting drug users 89 of the 96 largest metropolitan areas in the United States.14 [Almost two-thirds of the HIV seropostives in this study reported that they had received antiretroviral treatment, which would have reduced loss of death among the HIV seropositives (Des Jarlais, unpublished data).] HSV-2 prevalence in this sample was quite high (60%), toward the upper end of the 44%–65% range observed in previous studies of drug users. The odds ratios for HSV-2 as a risk factor for HIV were approximately 2.0 in all of the demographic subgroups except females (where it was considerably higher).
Although it is difficult to estimate the size of any population of heroin and cocaine users, current estimates of the numbers of noninjecting heroin and cocaine users in New York City indicate that the numbers of injecting and noninjecting users are approximately equal.13,15 Estimates of injecting drug users in the city are between 77,000 and 194,000.16 The population of noninjecting heroin and cocaine uses is thus clearly of public health importance.
It is important to note that noninjecting cocaine and heroin users in New York do not confine their sexual relationships to their drug-using peers. They also have sexual relationships with injecting drug users and with persons who do not use heroin and cocaine. Thus, both HSV-2 and HIV are likely to be transmitted from others to noninjecting heroin and cocaine users and from these users to other groups.
A full understanding of the importance of HSV-2 infection to HIV transmission would require longitudinal data on rates of HSV-2 and HIV acquisition and transmission, and sexual mixing patterns among noninjecting and injecting drug users and nondrug users in relation to HSV-2 and HIV serostatus. This cross-sectional study clearly does not include such comprehensive data. Nevertheless, it does provide some insight into the potential effect of HSV-2 on the dynamics of HIV transmission among noninjecting cocaine and heroin users in New York City.
The inferences are clearest for the female noninjecting drug users. HIV prevalence was high (23%) and HSV-2 prevalence was very high (86%) among the females in this study. That all of the HIV seropositive females were also HSV-2 seropositive suggests that the great majority of HIV seropositive females would have been infected with HSV-2 before becoming infected with HIV. HSV-2 infection would then have greatly increased the likelihood that these females would become infected with HIV.3,4 Large percentages of HIV seronegative/HSV-2 seropositive female subjects reported engaging in sexual risk behavior—54% reported unprotected intercourse with a primary partner and 25% reported unprotected intercourse with commercial partners. These female cocaine and heroin users would appear to be at very high risk for transmitting HSV-2 to others and for acquiring HIV infection themselves.
We should note that the most recent meta-analysis3 found unexplained “significant heterogeneity” in the studies of HIV incidence among “high-risk females” with prevalent HSV-2 infection. (The adjusted relative risk estimates ranged from a nonsignificant 0.517 to a highly significant 6.3).18 Difficulties in controlling for other sexually transmitted diseases that may facilitate HIV acquisition and for fully adjusting for differences in sexual behaviors may underlie this heterogeneity. With an HIV prevalence of 23%, the females in this study would almost certainly qualify as “high risk.” Our cross-sectional data are not comparable to the longitudinal data from the cohort studies in the meta-analyses, but our findings are certainly more consistent with a moderate to strong effect of HSV-2 infection on HIV acquisition in high-risk females than the absence of an effect.
The situation among the male subjects is less dramatic, but still of importance given the much larger number of male drug users. HIV prevalence was 17%, HSV-2 prevalence was 50%, and the odds ratio for HSV-2 as a risk factor for HIV was 1.69. All of these are lower than the comparable figures for the females, consistent with a lower likelihood of sexual acquisition of HIV and HSV-2 among males. The lower HSV-2 seroprevalence among the males may also be a reflection of male drug users being more likely to have nondrug-using sexual partners.19 These nondrug-using partners would be much less likely to be infected with either HIV or HSV-2, and also much less likely to be coinfected.20
The subjects who reported male-with-male sex (in the 5 years before the interview) deserve some comment. The modest number of such subjects in this study (59) greatly limits drawing conclusions. However, the 31% HIV seroprevalence in this subgroup was certainly quite high, and 72% of those who were HIV seropositive were also HSV-2 seropositive. Almost half the subjects in this study who reported male with male activity also reported sexual activity with female partners in the 6 months before the interview. Thus, this subgroup may serve as an important bridge for HIV transmission from other MSM to women.
In addition to the cross-sectional design, 2 other limitations of this study should be noted. The subjects had voluntarily entered drug abuse treatment, and we do not know if the results would generalize to heroin and cocaine users whose level of use was less than that of persons entering treatment. Second, the behavioral data were based on self-reports. Because of social desirability factors, some subjects might have injected drugs and denied this in the interview. However, if subjects had become infected with HIV through unreported drug injection, the likely effect would have been to reduce the association between HSV-2 and HIV.
Caution is needed in estimating the population attributable risk percentage for HIV attributable to HSV-2 in our noninjecting heroin and cocaine users for several reasons. Sexual transmission among our subjects is undoubtedly predominantly through heterosexual activity, and none of the longitudinal studies used to generate risk estimates in the 2 meta-analyses3,4 were of heterosexual populations in industrialized countries, and there was “significant heterogeneity” in the studies of “high-risk” females in developing countries.
Nevertheless, the biologic mechanisms through which HSV-2 infection may facilitate HIV transmission are likely to be similar in our population and in other populations, and it would be very useful for public health planning to have some estimate of the proportion of new HIV infections attributable to HSV-2 (the population attributable risk percentage) for our population of noninjecting drug users. We therefore used what we consider a conservative 2.0 relative risk (taken from the meta-analyses) and used the conventional formula population attributable risk percentage = 100 × [Px × (RR − 1)]/(1 + [Px × (RR − 1)], where Px is the percentage of HSV-2 seropositives among HIV seronegative noninjecting drug users and RR is the relative risk of acquiring HIV among HSV-2 seropostives when compared with HSV-2 seronegatives.
For the females in this study, with an 82% HSV-2 seroprevalence among HIV seronegatives the population attributable risk percentage for new HIV infections would be approximately 45%; for the males in this study, with a 48% HSV-2 seroprevalence among the HIV seronegatives, the population attributable risk percentage for new HIV infections would be approximately 32%, and for the total sample, with an 56% HSV-2 seroprevalence among the HIV seronegatives, the population attributable risk percentage for new HIV infections for would be approximately 36%. Note that these population attributable risk percentages are also conservative in that they do not include consideration of any effects of HSV-2 on increasing transmissibility of HIV among dually infected persons.
From a public health perspective, the data in this report show causes for both optimism and for concern. The optimistic aspect of the data is the greatly lower rates of risk behavior among the HIV seropositives. Although there is probably a social desirability component in these self-reports, the reports are consistent with the high rates of HIV testing among heroin and cocaine users in New York, the evidence for reduced risk behavior among HIV-positive injecting drug users in New York,21 and the evidence for reduced sexual risk behavior among many different HIV seropositive populations.22
A more pessimistic aspect of the data is the percentages of subjects who were HSV-2 seropositive and engaging in high-risk sexual behaviors. These persons are both at considerable risk of transmitting HSV-2 to others and of becoming infected with HIV if they should happen to engage in sex with an HIV infected partner. The HSV-2 seropositive females who reported unprotected sex with commercial partners would be the most dramatic example.
There are a variety of interventions to reduce the risk of HIV acquisition among persons with HSV-2 infection. These include behavioral interventions (recognizing HSV-2 outbreaks and avoiding sex during outbreaks) and suppressive antiviral therapy.1 We would suggest that such interventions be integrated into comprehensive HIV prevention programs for HSV-2 seropositive heroin and cocaine users, particularly for female noninjecting heroin and cocaine users. Such interventions may be relatively expensive, but are likely to be cost-saving when compared with treating HIV infection.23 The much lower rates of risk behavior among the HIV seropositives are an important indication that these noninjecting drug users are capable of substantial behavior change.
1. Celum C, Robinson N, Cohen M. Potential effect of HIV type 1 antiretroviral and herpes simplex virus type 2 antivirual therapy on transmission and acquisition of HIV type 1 infection. J Infect Dis 2005; 191:S107–S114.
2. Strick L, Wald A, Celum C. Management of herpes simplex virus type 2 infection in HIV type1-infected persons. Clin Infect Dis 2006; 43:347–356.
3. Freeman E, Weiss H, Glynn J, et al. Herpes simplex virus 2 infection increases HIV acquisition in men and women: Systematic review and meta-analysis of longitudinal studies. AIDS 2006; 20:73–83.
4. Wald A, Link K. Risk of human immunodeficiency virus infection in herpes simplex virus type 2-seropositive persons: A meta-analysis. J Infect Dis 2002; 185:45–52.
5. Hwang L, Ross M, Zack C, et al. Prevalence of sexually transmitted infections and associated risk factors among populations of drug users. Clin Infect Dis 2000; 31:920–926.
6. Ross M, Hwang L, Zack C, et al. Sexual risk behaviors and STIs in drug abuse treatment populations whose drug of choice is crack cocaine. Int J STD AIDS 2002; 13:769–774.
7. Plitt S, Sherman S, Strathdee S, et al. Herpes simplex virus 2 and syphilis among drug users in Baltimore, Maryland. Sex Transm Infect 2005; 81:248–253.
8. Des Jarlais DC, Friedman SR, Novick DM, et al. HIV-1 infection among intravenous drug users in Manhattan, New York City, from 1977 through 1987. JAMA 1989; 261:1008–1012.
9. Maslow CB, Friedman SR, Perlis TE, et al. Changes in HIV seroprevalence and related behaviors among male injection drug users who do and do not have sex with men: New York City, 1990–1999. Am J Public Health 2002; 92:382–384.
10. Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol 2004; 159:702–706.
11. Agresti A. An Introduction to Categorical Data Analysis. New York: Wiley, 1996.
12. SAS. SAS/STAT® User’s Guide, Version 9.1. Version 8 ed. Cary, North Carolina: SAS Institute, 2004.
13. Des Jarlais D, Arasteh K, Perlis T, et al. Convergence of HIV seroprevalence among injecting and non-injecting drug users in New York City: A new stage in a very large HIV epidemic. AIDS 2007; 21:231–235.
14. Friedman S, Lieb S, Tempalski B, et al. HIV among injection drug users in large US metropolitan areas, 1998. J Urban Health 2005; 82:434–445.
15. Golub A, Johnson B. The new heroin users among Manhattan arrestees: Variations by race/ethnicity and mode of consumption. J Psychoactive Drugs 2005; 37:51–62.
16. New York City Department of Health and Mental Hygiene Website: www.nyc.gov
. New York: NYC DOH, 2007.
17. Reynolds S, Risbud A, Shepherd M, et al. Recent herpes simplex virus type 2 infection and the risk of human immunodeficiency virus type 1 acquisition in India. J Infect Dis 2003; 187:1513–1521.
18. Kaul R, Kimani J, Nagelkerke N, et al. Monthly antibiotic chemoprophylaxis and incidence of sexually transmitted infections and HIV-1 infection in Kenyan sex workers: A randomized controlled trial. JAMA 2004; 291:2555–2562.
19. Ross M, Wodak A, Gold J. Sexual behavior in injecting drug users. J Psychol Human Sex 1992; 5:89–104.
20. Friedman SR, Flom PL, Kottiri BJ, et al. Drug use patterns and infection with sexually transmissible agents among young adults in a high-risk neighborhood in New York City. Addiction 2003; 98:159–169.
21. Des Jarlais DC, Perlis T, Arasteh K, et al. “Informed altruism” and “partner restriction” in the reduction of HIV infection in injecting drug users entering detoxification treatment in New York City, 1990–2001. J Acquir Immune Defic Syndr 2004; 35:158–166.
22. Marks G, Crepaz N, Senterfitt JW, et al. Meta-analysis of high-risk sexual behavior in persons aware and unaware they are infected with HIV in the United States: Implications for HIV prevention programs. J Acquir Immune Defic Syndr 2005; 39:446–453.
23. Schackman B, Gebo K, Walensky R, et al. The lifetime cost of current human immunodeficiency virus care in the United States. Med Care 2006; 44:990–997.