Sexually Transmitted Diseases:
Failure of Serosorting to Protect African American Men Who Have Sex With Men From HIV Infection
Golden, Matthew R. MD, MPH*,†; Dombrowski, Julia C. MD, MPH*,†; Kerani, Roxanne P. PhD*,†; Stekler, Joanne D. MD, MPH*,†
From the *Public Health Seattle–King County, Seattle, WA; †Center for AIDS and STD, University of Washington, Seattle, WA
Conflicts of Interest and Source of Funding: none declared.
Correspondence: Matthew Golden, MD, Harborview Medical Center, Seattle, WA. E-mail: firstname.lastname@example.org.
Received for publication February 14, 2012, and accepted March 23, 2012.
Background: Serosorting is the practice of choosing sex partners or selectively using condoms based on a sex partner's perceived HIV status. The extent to which serosorting protects African American (AA) and Hispanic men who have sex with men (MSM) is unknown.
Methods: We analyzed data collected from MSM sexually transmitted diseases clinic patients in Seattle, WA, 2001–2010. Men were asked about the HIV status of their anal sex partners in the prior year and about their condom use with partners by partner HIV status. We defined serosorters as MSM who had unprotected anal intercourse (UAI) only with partners of the same HIV status, and compared the risk of testing HIV positive among serosorters and men who reported having UAI with partners of opposite or unknown HIV status (ie, nonconcordant UAI). We used generalized estimating equations to evaluate the association of serosorting with testing HIV positive.
Results: A total of 6694 MSM without a prior HIV diagnosis were tested during 13,657 visits; 274 men tested HIV positive. Serosorting was associated with a lower risk of testing HIV positive than nonconcordant UAI among white MSM (2.1 vs. 4.5%, odds ratio [OR]: 0.45, 95% confidence interval [CI]: 0.34–0.61), but not AA MSM (6.8 vs. 6.0%, OR: 1.1, 95% CI: 0.57–2.2). Among Hispanics, the risk of testing HIV positive was lower among serosorters than men engaging in nonconcordant UAI, though this was not significant (4.1 vs. 6.0%, OR: 0.67, 95% CI: 0.36–1.2).
Conclusions: In at least some AA MSM populations, serosorting does not seem to be protective against HIV infection.
Black men who have sex with men (MSM) experience the highest rate of HIV infection of any large, readily identifiable population in the United States. In 2009, 47% of all newly diagnosed cases of HIV infection in MSM occurred in blacks; almost 24% of all persons diagnosed with HIV infection in areas of the United States with named-based HIV reporting occurred in African American MSM, a group that includes approximately 0.26% of the US population.1–4 The reasons African American MSM experience such a disproportionately high risk of HIV infection relative to white MSM are not well defined.5,6
Serosorting, the practice of choosing partners or selectively using condoms based on a sex partner's perceived HIV status, is common among MSM, including African American MSM.6–9 Prior studies suggest that serosorting provides partial protection against HIV acquisition.7,10–12 However, the extent to which serosorting is protective is critically dependant on the population's HIV testing frequency; in a population that tests infrequently, the practice may increase HIV risk.13–15 Few data exist on the extent to which serosorting protects African American MSM.7 We evaluated whether the protective impact of serosorting varies by race. Based on prior reports suggesting that undiagnosed HIV infection is more common among African American than white MSM,16–18 we hypothesized that serosorting would be less protective in African American MSM than white MSM.
This study is a secondary data analysis of information recorded in the Public Health—-Seattle & King County (PHSKC) Sexually Transmitted Disease (STD) clinic electronic medical record. Clinicians and staff collected all behavioral information used in the study as part of routine clinical risk assessments, and all HIV testing was done as part of standard clinical care. The unit of analysis throughout the study is the clinic visit, meaning that some individuals contributed >1 visit to the analysis. We have previously described our data collection methods.3,11
Study Population, Data Collection, and HIV Testing
The study population was comprised of MSM evaluated in the PHSKC STD clinic between October 1, 2001 and December 31, 2010. We excluded visits from the analysis if data on a patient's sexual behavior were insufficiently complete to allow us to classify the patient as a serosorter or not a serosorter, or if clinicians neglected to record information on the patient's HIV status and testing history. Because clinicians did not routinely obtain complete sexual histories during follow-up visits for specific problems occurring within 60 days of a prior clinic visit, follow-up visit data were likewise excluded from the analysis. Clinicians and HIV testing staff used a structured form to record patient histories. These histories were obtained before HIV testing. Clinicians routinely asked patients if they had sex with men, women, or both men and women in the preceding year. We defined MSM to include men who reported any male sex partners in that time frame. Clinicians and staff asked MSM whether they had anal intercourse with partners who were HIV positive, HIV negative, or of unknown HIV status, and about how often patients used condoms (always, usually, sometimes, or never) with partners of each HIV status category; these questions used a 12-month time frame, and separately collected aggregate information on insertive and receptive sex partners. All patients in the clinic completed a registration form that asked them to designate their race and whether they were Hispanic; these were separate questions. Patients are able to designate >1 race. We classified white Hispanics and persons who indicated that they were Hispanic but declined to define a race as Hispanic. Because <30 Asian and Alaska Native/Native American MSM tested HIV positive during the study period, we elected not to present findings on these groups. Similarly, we excluded persons with multiple races because they represent a heterogeneous group.
Clinicians recommended HIV testing to all patients who reported that they had not previously tested HIV positive. Depending on the study year, we tested patients who agreed to a blood draw using 3 different HIV EIAs (Vironostika HIV-1 Microelisa System [bioMerieux, Durham, NC], Genetic Systems rLAV EIA or Genetic Systems HIV-1/HIV-2 Plus O EIA [both Bio-Rad, Redmond, WA]). We offered OraQuick rapid HIV antibody tests (Orasure technologies Inc., Bethlehem, PA) to MSM patients at high risk for HIV infection; the clinic defines MSM without a prior HIV diagnosis to be at high risk for HIV infection, if they have 1 or more of the following risk factors: (1) bacterial STD in the prior year, (2) methamphetamine or popper use in the prior year, (3) ≥10 sex partners in the prior your, or (4) any unprotected anal sex with an HIV-positive partner or partner of unknown HIV status.19 Oraquick tests were performed on fingerstick blood from 2001 to 2007 and from 2008 to 2010; rapid tests were performed on saliva in 2007 and for some time in 2008. In addition, starting in 2003, we performed pooled-HIV RNA testing on all MSM who tested for HIV and who agreed to a venipuncture.20 We considered men to have acute HIV, if they were HIV RNA positive but tested negative using an EIA.
To assess how effective serosorting might be in preventing HIV, we classified men's visits into 4 groups: (1) visits during which men reported unprotected anal intercourse (UAI) only with partners of the same HIV status (serosorters), (2) visits during which men reported UAI with partners of discordant HIV status or partners of unknown HIV status (nonconcordant UAI), (3) visits during which men reported only protected anal sex, and (4) visits during which men denied having anal sex. We used generalized estimating equations (GEE) to evaluate the risk of testing HIV positive by serosorting categorization, stratified by race/ethnicity; this analytic approach adjusts for the fact that some subjects contributed >1 visit to the data set, and that not all observations were independent. We also used GEE to evaluate the interaction of race/ethnicity with serosorting in a model using HIV as an endpoint, and to compare the sexual behavior of African American Hispanic, and white MSM, including MSM with a prior HIV diagnosis. These analyses were undertaken to identify factors that might help explain a differential impact of serosorting by race/ethnicity. Analyses comparing medians used the Wilcoxin rank-sum test. We performed all analyses using the SAS system 9.1 (SAS Institute, Cary, NC).
The University of Washington Division of Human Subjects approved all study procedures.
Between October 2001 and December 31, 2010, a total of 10,620 non-Hispanic white, African American or Hispanic MSM patients attended the clinic during 22,370 new problem visits. Data were sufficiently complete to classify patients based on serosorting behaviors for 10,074 (95%) different men and 20,735 (93%) clinic visits. This population of visits included 1382 men with a prior HIV diagnosis who attended 2763 visits and 17,972 visits by 8950 men without a prior HIV diagnosis. (258 men had visits both before and after testing HIV positive.) A total of 6694 MSM without a prior HIV diagnosis tested for HIV during 13,657 separate clinic visits.
Comparison of White and Black MSM
Comparing data collected during visits by white and African American MSM without a prior HIV diagnosis, African American men were significantly younger and reported having fewer sex partners (Table 1). The anal sexual repertoire of African American and white MSM also differed; black men more often reported that they only engaged in insertive anal sex, and less often indicated that they only had receptive anal sex or were versatile (ie, engaged in both insertive and receptive anal sex). Serosorting categorization was similar between white and African American MSM, with men of both races reporting behaviors consistent with serosorting during approximately one-third of clinic visits. Men in both groups reported never having previously HIV tested during 5% of clinic visits, and the median time since last HIV test did not differ across groups. However, despite having lower levels of high-risk sexual behavior and similar testing histories, African American MSM were significantly more likely to test HIV positive (4.7% vs. 2.5%, P < 0.0001).
The association of serosorting in the past year with the risk of testing HIV positive differed by race (Table 2). Among whites, serosorting was associated with a significantly lower risk of testing HIV positive than having nonconcordant UAI (2.1% vs. 4.5%, P < 0.0001). In contrast, the risk of testing HIV positive among blacks was similar among serosorters and men who reported having nonconcordant UAI (6.8% vs. 6%, P = 0.72). There was a significant interaction between race and serosorting status in a GEE model (P = 0.02), indicating that serosorting was associated with a significant, approximately 50% reduction in HIV risk among whites, but was not protective among blacks.
Comparison of White and Hispanic MSM
Hispanic MSM were significantly younger, reported having fewer sex partners, and were less likely to report not having any anal sex than white MSM (Table 1). They more often reported never having previously tested for HIV infection and were more likely to newly test HIV positive. We did not observe differences between Hispanic and white non-Hispanics in anal sexual repertoire or disclosure of HIV status. The overall pattern of association between serosorting behaviors and HIV test results in Hispanic men paralleled those observed in whites; serosorting men experienced an intermediate level of risk between having only protected anal sex and nonconcordant UAI. There was no significant interaction between ethnicity and serosorting in a GEE model using HIV test results as an outcome.
Restricting the analysis to men who reported testing HIV negative in the year before clinic visit did not affect the overall pattern of association between serosorting and HIV risk for any racial or ethnic group (data not shown).
Differences in Risk Behavior and HIV Testing and Care by Race/Ethnicity Among Men who Newly Diagnosed HIV and Previously Diagnosed HIV
We compared sexual risk behaviors, serostatus disclosure frequency, HIV testing histories, and HIV treatment histories among men of different races and ethnicities to identify possible explanations for why serosorting might have a differential impact on HIV risk among men of different races and ethnicities. Compared with white MSM, African American MSM with newly diagnosed HIV were younger, reported having fewer sex partners, more often had only insertive anal sex, less often engaged in unprotected receptive anal intercourse with a partner of unknown discordant HIV status, and had similar HIV testing histories (Table 3). African American MSM with newly diagnosed HIV infection were also significantly more likely than white MSM to be classified as serosorters (42% vs. 26%). Among men with previously diagnosed HIV infection, African American MSM reported having significantly fewer sex partners than white MSM, and less often reported having nonconcordant UAI (Table 4). Although African American MSM with a prior HIV diagnosis reported taking antiretroviral therapy and always telling partners about their status during a smaller proportion of their visits than white MSM, these differences were not statistically significant.
Hispanic MSM with newly diagnosed HIV infection were younger than white MSM with newly diagnosed HIV but were otherwise similar with respect characteristics and behaviors reported in Table 3. Among MSM with a prior HIV diagnosis, Hispanics were younger, had fewer partners, less often reported serosorting, and more often reported always using condoms than white MSM.
We found that serosorting was common among white, African American and Hispanic MSM in our STD clinic but that compared with nonconcordant UAI, it was only clearly associated with a lower risk of HIV infection among whites. Among blacks, the risk of testing HIV positive among MSM classified as serosorters was slightly higher than it was among men who reported having nonconcordant UAI, and 42% of all African American MSM with newly diagnosed HIV infection in our clinic were classified as serosorters.
Our data did not identify reasons for why serosorting might be less effective in African American. Like several previously published studies,21 we found that African American MSM had fewer sex partners than white MSM. Blacks were also less likely to have a versatile anal sexual repertoire, a pattern of behavior that mathematical models suggest should be protective at a population level.22,23 Some, but not all, prior studies have suggested that African American MSM may less frequently disclose their HIV status or know a partner's HIV status.6,24,25 We observed somewhat lower levels of HIV status disclosure among African American than white MSM, though African American MSM in our study were not more likely than white MSM to report having UAI with partners of unknown status. Finally, we did not observe lower levels of HIV testing among African American MSM. However, our study, which was conducted in a clinic that performs HIV tests, may not provide an accurate picture of HIV testing in the wider population. Prior reports, including studies conducted in Seattle, have found that late HIV diagnoses and undiagnosed HIV infection are more common in black than white MSM.16–18 We suspect that inadequate knowledge of HIV status in a population with race assortative sexual mixing26,27 is the primary explanation for why serosorting was not protective among African American MSM in our clinic. If that is correct, serosorting could be made safer through increased HIV testing. The US Centers for Disease Control and Prevention is currently attempting to increase HIV testing among black MSM through the Expanded Testing Initiative.28
Our findings are somewhat at odds with a study by Marks et al, which observed a lower level of HIV infection among African American MSM who reported limiting UAI to partners they believed to be HIV negative compared with men who engaged in nonconcordant UAI.7 That study used audio computer-assisted self-interviews to collect sexually history data, whereas our data were collected by clinicians as part of routine care. Our clinicians did not ask patients questions using a consistent script, which may have led to confusion among patients, diminishing our ability to observe an association between serosorting and the risk of testing HIV positive. On the other hand, the protective effect of serosorting was evident among whites and perhaps Hispanics, suggesting that if inconsistencies in how questions were asked resulted in error, that error affected blacks differentially. Alternatively, disparate findings between the 2 studies may reflect true differences in the populations studied, or imprecision in findings resulting from the relatively small number of HIV diagnoses in both studies.
Our study has several important limitations. First, although our analysis included data from >2500 clinic visits and HIV tests in African American and Hispanic MSM, the number of new HIV diagnoses was relatively small, limiting our power to assess the association of serosorting behavior with HIV. Second, our study evaluated the association of sexual behavior with HIV risk and did not measure participants' purposeful decision to serosort. Third, as mentioned earlier, our data were collected as part of routine medical care, not a research study, and variability in how clinicians asked questions may have introduced an error into our measurements of behavior. Fourth, our behavioral data are based on self report and may be subject to social desirability bias, which may affect men of different races and ethnicities in different ways. Finally, our findings may not be generalizable to other populations.
We found that serosorting behaviors, which are associated with a significant decrease in the risk of HIV infection among white MSM in our clinic, may not be protective against HIV among African American MSM. Although our findings are not definitive, we believe that they should prompt additional caution related to the promotion of serosorting among African American MSM and highlight the need to increase HIV testing coverage and frequency in this critically important population.
1. CDC. HIV/AIDS Surveillance Report, 2009. Atlanta, GA: Department of Health and Human Services, Centers for Disease Control and Prevention, 2011.
2. Glick SN, Golden MR. Persistence of racial differences in attitudes toward homosexuality in the United States. J Acquir Immun Defic Syndr 2010; 55:516–523.
3. Anderson JE, Mosher WD, Chandra A. Measuring HIV risk in the U.S. population aged 15–44: Results from cycle 6 of the National Survey of Family Growth. Adv Data 2006; 377, 1–27.
4. Laumann EO, Gagnon JH, Michael RT, et al.. The social organization of sexuality: Sexual practices in the United States. Chicago, IL: University of Chicago Press, 1994.
5. Liau A, Millett G, Marks G. Meta-analytic examination of online sex-seeking and sexual risk behavior among men who have sex with men. Sex Transm Dis 2006; 33:576–584.
6. Oster AM, Wiegand RE, Sionean C, et al.. Understanding disparities in HIV infection between black and white MSM in the United States. AIDS 2011; 25:1103–1112.
7. Marks G, Millett GA, Bingham T, et al.. Prevalence and protective value of serosorting and strategic positioning among Black and Latino men who have sex with men. Sex Transm Dis 2010; 37:325–327.
8. McFarland W, Chen YH, Raymond HF, et al.. HIV seroadaptation among individuals, within sexual dyads, and by sexual episodes, men who have sex with men, San Francisco, 2008. AIDS Care 2011; 23:261–268.
9. Lattimore S, Thornton A, Delpech V, et al.. Changing patterns of sexual risk behavior among London gay men: 1998–2008. Sex Transm Dis 2011; 38:221–229.
10. Golden MR, Stekler J, Hughes JP, et al.. HIV serosorting in men who have sex with men: Is it safe? J Acquir Immun Defic Syndr 2008; 49:212–218.
11. Jin F, Crawford J, Prestage GP, et al.. Unprotected anal intercourse, risk reduction behaviours, and subsequent HIV infection in a cohort of homosexual men. AIDS 2009; 23:243–252.
12. Philip SS, Yu X, Donnell D, et al.. Serosorting is associated with a decreased risk of HIV seroconversion in the EXPLORE Study Cohort. PLoS One 2010; 5.
13. Cassels S, Menza TW, Goodreau SM, et al.. HIV serosorting as a harm reduction strategy: Evidence from Seattle, Washington. AIDS 2009; 23:2497–2506.
14. Wilson DP, Regan DG, Heymer KJ, et al.. Serosorting may increase the risk of HIV acquisition among men who have sex with men. Sex Transm Dis 2010; 37:13–17.
15. Heymer KJ, Wilson DP. Available evidence does not support serosorting as an HIV risk reduction strategy. AIDS 2010; 24:935–936, author reply 936–938.
16. Nelson KM, Thiede H, Hawes SE, et al.. Why the wait? Delayed HIV diagnosis among men who have sex with men. J Urban Health 2010; 87:642–655.
17. MacKellar DA, Valleroy LA, Secura GM, et al.. Unrecognized HIV infection, risk behaviors, and perceptions of risk among young men who have sex with men: Opportunities for advancing HIV prevention in the third decade of HIV/AIDS. J Acquir Immun Defic Syndr 2005; 38:603–614.
18. Prevalence and awareness of HIV infection among men who have sex with men — 21 cities, United States, 2008. MMWR Morb Mortal Wkly Rep 2010; 59:1201–1207.
19. Menza TW, Hughes JP, Celum CL, et al.. Prediction of HIV acquisition among men who have sex with men. Sex Transm Dis 2009; 36:547–555.
20. Stekler JD, Swenson PD, Coombs RW, et al.. HIV testing in a high-incidence population: Is antibody testing alone good enough? Clin Infect Dis 2009; 49:444–453.
21. Millett GA, Flores SA, Peterson JL, et al.. Explaining disparities in HIV infection among black and white men who have sex with men: A meta-analysis of HIV risk behaviors. AIDS 2007; 21:2083–2091.
22. Goodreau SM, Goicochea LP, Sanchez J. Sexual role and transmission of HIV Type 1 among men who have sex with men, in Peru. J Infect Dis 2005; 191(suppl 1):S147–S158.
23. Goodreau SM, Golden MR. Biological and demographic causes of high HIV and sexually transmitted disease prevalence in men who have sex with men. Sex Transm Infect 2007; 83:458–462.
24. Wei C, Raymond HF, Guadamuz TE, et al.. Racial/Ethnic differences in seroadaptive and serodisclosure behaviors among men who have sex with men. AIDS Behav 2011; 15:22–29.
25. Bird JD, Fingerhut DD, McKirnan DJ. Ethnic differences in HIV-disclosure and sexual risk. AIDS Care 2011; 23:444–448.
26. Raymond HF, McFarland W. Racial mixing and HIV risk among men who have sex with men. AIDS Behav 2009; 13:630–637.
27. Berry M, Raymond HF, McFarland W. Same race and older partner selection may explain higher HIV prevalence among black men who have sex with men. AIDS 2007; 21:2349–2350.
© Copyright 2012 American Sexually Transmitted Diseases Association
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Data is temporarily unavailable. Please try again soon.