THE OCCURRENCE OF sexually transmitted diseases (STDs) in persons with HIV infection is prima facie evidence of continued risk taking. Ideally, the cooccurrence of HIV and STDs should be vanishingly rare, but the frequency of coinfection is well documented by studies that have sought to assess the influence of STDs on HIV transmission.1 In general, these studies have reported an important role for ulcerative and nonulcerative STDs in facilitating HIV transmission. The biologic basis for such facilitation, including increased shedding of HIV in the presence of STDs, is under active investigation.2 Ironically, advances in antiretroviral therapy may foster the acquisition of STDs and the potential transmission of HIV because these advances may influence some persons to use less-safe practices.3–6
Assessing the impact of STDs on HIV transmission is complex because of the variety of epidemiologic circumstances under which coinfection may occur. Persons with HIV face a heterogeneous STD prevalence that will influence their risk of acquiring an STD and the risk of their transmitting HIV to a partner with STD. In this study, we have focused on the potential for decreasing HIV transmission by the treatment of STDs in dually infected persons. We used data from eight STD clinic sites across the United States to examine the frequency of STDs among clinic attendees who are HIV positive, and to assess the potential impact of STD treatment on HIV transmission.
Sexually transmitted disease clinics in New York City, New Orleans, Birmingham, Los Angeles, Colorado Springs, Denver, Chicago, and Miami provided data on HIV-positive attendees. The period and length of data collection varied among the clinics depending on the availability of computerized records, but in general included 1 year to 2 years of data from the early to mid 1990s. Using a standard protocol (except where noted in the data presentation), each site provided information on age, sex, ethnicity, sexual orientation, reason for visit, results, and dates of HIV and STD tests performed within the specified interval. Only one site (New Orleans) was able to provide information on seroconverters (i.e., persons with a positive HIV test result who had had a documented negative test result within the previous 6 months). The data were processed using the Statistical Analysis System.7
Data regarding more than 20 STDs were provided. We grouped these conditions into three categories: (1) genital ulcer disease (including primarily syphilis, herpes simplex virus type 2, and chancroid); nonulcerative STDs (primarily chlamydial and gonococcal infection); and any STD (i.e., persons with any of these diagnoses). Analysis by specific microbiologic entity was not attempted because of the small numbers at most sites and the lack of uniform data regarding treatment efficacy, particularly for viral STDs. Based on an exhaustive review of the current literature,1 we used a relative risk (Rt) of 3.0 for the increase in transmission of HIV from coinfected persons (STD+/HIV+) compared to persons with HIV in the absence of another STD (STD−/HIV+). This relative risk was used for both genital ulcer disease and nonulcerative STDs.
We assessed the prevalence of coinfection with HIV and STDs for the group as a whole and for persons with specific attributes (Tables 1 and 2). Because this mechanism for case acquisition was not a sample of individuals but a sample of clinic estimates, for the calculations described in the Analytic Methods (Table 3), we used the prevalence in groups cross-classified by clinic site (8), age (2), sex (2), ethnicity (3), and sexual orientation (3). Of the total of 288 possible cells, 63 were empty and the remainder were used to calculate mean prevalences for the variables of interest. For example, to calculate the prevalence of STD+/HIV+ women, we used the mean of the 85 cells in which women appeared.
We used the prevented fraction (PF)8,9 to assess the impact of STD treatment on HIV transmission in this setting. The PF is usually defined as (D0−D1)/D0, where D0 is the risk of disease in the absence of the intervention, and D1 is the overall risk of disease in the presence of the the intervention. It may be described-in a manner analogous to the attributable fraction-as the proportion of prevented cases that can be attributed to an intervention. In this example, HIV transmission by a dually infected person is the outcome we wish to prevent and STD treatment is the intervention.
The analytic process requires two steps. First, we treat the potential HIV transmitter for other STDs. The PF associated with that treatment may be written as PFstd = pe(1 − RR), where RR (<1) is the relative risk of failing to recover if given treatment compared to those not given treatment, and pe is the proportion of the population exposed to the intervention. In a perfect intervention, pe = 1.0 and RR = 0, so that 100% of the outcomes are attributable to the intervention. But because the intervention is imperfect, we have assigned a value of 0.9 for population exposure to the intervention and 0.1 for the relative risk for the efficacy of treatment, thereby estimating that the average PF for STD treatment is 0.8.
Second, we examine the effect of STD treatment of the dually infected person on HIV transmission to their sexual partners. If ps is the proportion of persons who are STD+/HIV+, 1−ps is the proportion of persons who are STD−/HIV+, x is the transmission rate, and the overall risk of transmission in the absence of an intervention will be (Rt)(ps)(x) + (1−ps)(x). As previously defined, the risk of transmission in the presence of an intervention is x. Substituting in the prior formulation for the prevented fraction (D0−D1)/D0, we get, with algebraic manipulation, EQUATION which is the well-known formula for the attributable fraction.10 Therefore, in the first step, the calculated PFstd is the proportion of diminution in STD among HIV-positive persons that can be attributed to treatment. In the second step, the calculated PFt is equivalent to the attributable fraction and is the maximal achievable decrease in HIV transmission that can result from treating STDs in HIV-positive persons. It follows that the actual achievable decrease is simply the product of these two prevented fractions: (PFstd)(PFt).
To estimate the overall PF [(PFstd)(PFt)] from these data, we perform these calculations as a weighted average of the mix of age, sex, ethnicity, sexual orientation, and STD prevalence at each of the contributing sites. For this purpose, the actual transmission rate, x, drops from the equations. To calculate the actual number of cases that would be prevented by STD treatment of HIV-positive persons, however, we used average literature-based transmission probabilities in the absence of STD: 0.001 from female to male, 0.002 from male to female, and 0.01 from male to male.11,12 Further refinements, such as the variation in RR and PF that may occur with individual STDs, are not included in this analysis.
The eight sites differed substantially in their contribution to the overall sample (Table 1) and in demographic makeup and sexual orientation (Figure 1). For example, black persons constituted 61.5% of the total group of 4516, but their proportion varied from a low of 18.7% in Colorado Springs to a high of 95.3% in Chicago. Similarly, the proportion of bisexual persons varied from 0% in Birmingham to 48.4% in Denver. In another marked contrast, more than half of the small Birmingham cohort was younger than 30 years, compared with less than 10% of those from New York City. More than half of the patients from Los Angeles were Hispanic, though overall 16.9% of the entire cohort was of Hispanic ethnicity.
Sexually Transmitted Infections
In the overall cohort, the proportion infected with STDs varied by sociodemographic characteristics and sexual orientation (Table 1). The prevalence of genital ulcer disease varied widely among subgroups, from a low of 6.0% among whites to a high of 23.4% among heterosexuals. The largest discrepancy within a given characteristic occurred for nonulcerative STDs among Hispanics (14.8%) compared with whites (33.1%). The combination of heterogeneity in the distribution of demographic and sexual orientation by sites and the prevalence of STDs within subgroups led to considerable diversity in the frequency of these conditions at the eight sites (Table 2). Genital ulcer disease occurred in only 4.9% of the HIV-positive attendees in Colorado Springs, but in 45.0% of the small cohort of attendees in Birmingham and in 38.5% and 34.9% of the larger groups in Chicago and New York. Overall, the highest proportion of STD+/HIV+ persons was documented in Chicago (67.6%) and the lowest in Miami (13.9%). In the special group of 76 seroconverters who were followed up in New Orleans, 78.9% were infected with an STD at the time of their last negative HIV test result. This proportion was 28.9% at the time of seroconversion, within 6 months of the previous negative test (Table 2).
Potential Reductions in HIV Transmission
In the overall population, the maximal and actual achievable reductions are substantial for all of the subgroups examined (Table 3). On average, approximately 33% of of the maximum potential HIV transmission could be eliminated. With a prevented fraction of 0.8, approximately a 27% reduction in transmission might actually be achieved. Because of the heterogeneity in demographic composition and STD prevalence at the participating sites, their maximal and actual reductions vary considerably (Figure 2). Though the cohort is small, a reduction of 38.1% of the potential HIV transmission from these clients could be achieved at the Colorado Springs site. In contrast, a reduction of 10.0% might be possible at the Los Angeles site. Using standard HIV transmission probabilities (see Methods) for the probability of transmission from a single sexual contact, the number of new infections that would result if each of the 4516 HIV-infected persons had a single unprotected sexual contact with an uninfected person would be 28 in the absence of STD treatment. With STD treatment, we estimate that 16 transmissions would occur. These estimates are the result of averaging many factors (e.g., number of partners, number of encounters, types of exposure) that can affect transmission.
With this method, we estimate that about 27% of HIV transmissions from persons who are dually infected with HIV and an STD could be averted through adequate treatment of the STD, independent of behavioral change. Estimates of this proportion varied from 10.0% to 38.1%, depending on the mix of clients and conditions at the eight participating sites. Because we do not deal with the potential for increased susceptibility of those with an STD to acquire HIV, these figures may underestimate the impact of treatment of STDs in general in high-prevalence settings, because an intervention for reducing HIV transmission may be underestimated.
Although our results are encouraging, they should be viewed with caution because our method provides only a general approach and would require some important refinements. A more comprehensive approach would require consideration of the role of treatment of STDs in the HIV-negative partners of those with HIV infection, and the parallel risk of needle sharing in the population under consideration.
In addition, specific sexually transmitted conditions differ in the degree to which they facilitate transmission of HIV. In the published studies, the relative risk for HIV transmission associated with nonulcerative genital infection varies from 1.9 to 6.3, and for genital ulcer disease the risk varies from 3.3 to 11.3.1,12 Further, the diagnostic tools available and the treatment for these conditions differ in their efficacy. In fact, the natural course and treatment for some STDs (e.g., neurosyphilis, herpes) may be adversely influenced by the presence of infection with HIV.13,14 The complex relationship between HIV and herpes simplex virus type 2 is illustrative; drugs that inhibit the latter condition may have beneficial effects on the clinical and virologic course of HIV infection.15,16 Conversely, the upregulation of HIV expression induced by herpes simplex virus in vitro may be reflected clinically in increased plasma HIV RNA levels.17,18
Sexually transmitted diseases differ in the extent to which they can be recognized and treated effectively. In this type of analysis, the relative risk of HIV transmission in the presence of coinfection with an STD, set at an average value of 3.0, and the PFstd, set at 0.8, should vary with the individual STD. Because clinical settings also vary from ideal to nonexistent, some regard for the clinical setting must also be factored into a more generalizable approach. However, data on which to base estimates of relative risk, prevented fraction, and clinic adequacy are not readily available for some STDs, and we have chosen not to attempt such specificity (even though the computational effort involved is not a barrier).
Generalizability of these results is also hindered by not knowing how representative this group is of the general population of HIV- infected persons. It is likely that a significant portion of current HIV-positive persons are not involved in risky activity, but the population-based frequency of those who are and are not is not known. Highly active antiretroviral therapy has become widespread in the United States since these data were collected and may have paradoxical modulating effects on transmission. Highly active antiretroviral therapy may decrease HIV risk by reducing HIV shedding in genital secretions, but may simultaneously increase risky sexual behaviors and STD acquisition.
Even if our cohort represents a relatively small fraction of the overall group with HIV infection, it is likely to be a fraction highly involved in HIV transmission. From the standpoint of controlling the further spread of HIV and other STDs in communities, dually infected persons may be a critical group for targeting epidemiologic services because their sexual partners and associates are all likely to be part of a milieu in which transmission of STDs, and possibly HIV, is taking place. Experience with seroconverters in New Orleans lends some credibility to this concept because before seroconversion, this population had the highest STD prevalence of any group.
The theoretical results offered here should also be viewed in light of empirical results that have emerged from two attempts to alter HIV transmission through treatment of STDs.19,20 These community-level, randomized controlled trials, which differed in intervention-design approach, measurement, and preexisting HIV prevalence, reported differing results, and are now the subject of considerable debate. Their applicability to the settings described in this report (i.e., STD clinics in the United States) is uncertain because they used intermittent mass treatment19 or syndromic management of symptomatic STDs, whereas in STD clinics in the United States the intervention of interest is the optimal application of STD treatment within such clinic settings. Therefore, the theoretical results from this study and from others21 await empirical confirmation.
In view of the limits of generalizability, we offer the seemingly quantitative estimate of a 27% reduction in HIV transmission from dually infected persons more as a qualitative assessment of potential importance of STD control in this setting. Certainly, the high STD prevalence among HIV-positive persons seen at these clinics documents the existence of a group at continuing high risk for transmission of both HIV and other STDs (and possibly of blood-borne infections). Our estimate suggests that effective treatment of their STDs and interventions targeted to their social milieu should be implemented and empirically evaluated.
1. Fleming DT, Wasserheit JN. From epidemiological synergy to public health policy and practice: the contribution of other sexually transmitted diseases to sexual transmission of HIV infection. Sex Transm Infect 1999; 75:3–17.
2. Cohen MS. Sexually transmitted diseases enhance HIV transmission: no longer an hypothesis. Lancet 1998; 351(suppl 3):5–7.
3. DiClemente R. Evidence for an association between treatment with protease inhibitors and risky sex among persons living with HIV/AIDS (abstract no. 3263). Presented at the APHA Meeting. Washington, DC, November 15–18, 1998.
4. Page-Shafer KA, McFarland W, Kohn R, et al. Increases in unsafe sex and rectal gonorrhea among men who have sex with men- San Francisco, California, 1994–1997. MMWR 1999; 48:45–48.
5. Woods WJ, Dilley JW, Liharsh T, Sbatino J, Adler B, Rinaldi J. Name-based reporting of HIV-positive test results as a deterrent to testing. Am J Public Health 1999; 89:1097–1100.
6. Centers for Disease Control and Prevention. Resurgent bacterial sexually transmitted disease among men who have sex with men-King County, Washington, 1997–1999, MMWR 1999; 48:773–777.
7. SAS Institute Inc. SAS Language and Procedures: Usage, Version 6. 1st ed. Cary, NC: SAS Institute, 1989.
8. Miettinen OS. Proportion of disease caused or prevented by a given exposure, trait or intervention. Am J Epidemiol 1974; 99:325–332.
9. Garguillo PM, Rothenberg RB, Wilson HG. Confidence intervals, hypothesis tests and sample sizes for the prevented fraction in cross-sectional studies. Stat Med 1995; 14:51–72.
10. Levin M. The occurrence of lung cancer in man. Acta Univ Intern Cancer 1953; 9:531–41.
11. Royce RA, Sena A, Cates W, Cohen MS. Sexual transmission of HIV. N Engl J Med 1997; 336:1072–1078.
12. Mastro TD, de Vincenzi I. Probabilities of sexual HIV-1 transmission. AIDS 1996; 10(suppl A):S75-S82.
13. Augenbraun M, Feldman J, Chirgwin K, et al. Increased genital shedding of herpes simplex virus type 2 in HIV-seropositive women. Ann Intern Med 1995; 123:845–847.
14. Schmid GP. Treatment of chancroid, 1997. Clin Infect Dis 1999; 28(suppl 1):S14-S20.
15. Ioannidis J, Collier A, Cooper D, et al. Clinical efficacy of high-dose acyclovir in patients with human immunodeficiency virus infection: a meta-analysis of randomized individual patient data. J Infect Dis 1998; 178:349–359.
16. Schacker TW, Lu HL, Zeh J, Shaughnessy M, Corey L. HSV suppression is associated with a significant decrease in plasma levels of HIV RNA (abstract no. 260). Paper presented at: 5th Conference on Retroviruses and Opportunistic Infection; February 1–5, 1998; Chicago, IL.
17. Mole L, Ripich S, Margolis D, Holodniy M. The impact of active herpes simplex virus infection on human immunodeficiency virus load. J Infect Dis 1997; 176:766–770.
18. Golden MP, Sunyoung J, Hammer SM. Activation of human immunodeficiency virus by herpes simplex virus. J Infect Dis 1992; 166:494–499.
19. Wawer MJ, Sweankambo NK, Serwadda D, et al, for the Rakai Project Study Group. Control of sexually transmitted diseases for AIDS prevention in Uganda: a randomised community trial. Lancet 1999; 353:525–535.
20. Grosskurth H, Mosha F, Todd J, et al. Impact of improved treatment of sexually transmitted diseases on HIV infection in rural Tanzania: randomised controlled trial. Lancet. 1995; 346:530–536.
21. Robinson NH, Mulder DW, Auvert B, Hayes R. Proportion of HIV infections attributable to other sexually transmitted diseases in a rural Ugandan population: simulation model estimates. Int J Epidemiol 1997; 26:180–188.