The effectiveness of a serosorting strategy for HIV prevention depends on the accuracy of individuals’ serostatus disclosures. We modeled the risks of sexual transmission of HIV under various circumstances differing by the type of disclosures made. Accounting for rates of unrecognized HIV infection, treatment status and differences in infectivity by stage of infection, we found that serosorting can increase the transmission risk for some groups.
University of California, San Diego, California, USA.
Received 3 December, 2006
Revised 22 February, 2007
Accepted 23 February, 2007
Recommendations for preventing sexual transmissions of HIV emphasize measures that might reduce infection risk, such as consistently using latex condoms or serosorting [1,2]. In serosorting, partners make decisions about how they have sex on the basis of their assumed or reported HIV serostatus. If other prevention methods, such as latex condoms, are not employed with every sex partner but only with those having a discordant serostatus, then knowing the true HIV status is imperative. Without lying, some HIV-infected individuals may mistakenly believe they are not infected and disclose as ‘HIV negative’ . Because the effectiveness of serosorting depends upon accurate disclosure, we decided to model the risk of transmission from ‘HIV-positive’ versus ‘HIV-negative’ disclosers to someone who is not infected and not using condoms. Our model accounts for the prevalence of unrecognized infections, treatment status and differences in infectivity by stage of infection.
Individuals who incorrectly disclose as ‘HIV negative’ are either chronically or recently infected. Although standard antibody tests diagnose chronic infection with a very low false-negative rate, acute infections cannot be diagnosed by standard HIV testing until after detectable antibodies have been produced . Individuals may erroneously believe themselves to be uninfected after a negative antibody test result, and disclose themselves as ‘HIV negative’ to sexual partners when, in fact, they are highly contagious . Many factors affect the incidence of HIV within groups [6–9]; and, for some groups, the density rate of new infections (and therefore the number of individuals with recent infections) can be quite high .
During the period of recent HIV infection, individuals typically have a much higher viral burden than they do for most of the time that they are infected, and the viral load has been shown to correlate with infectiousness [9,11,12]. During acute infection, which lasts approximately 6–8 weeks , infectiousness is probably greater than at any other time. Infectiousness, however, remains elevated even after the acute period for up to 25 months . As the proportion of recently infected potential sex partners in a population increases, the effectiveness of disclosure for preventing HIV transmissions by serosorting decreases. Therefore, the same serosorting strategy will not have the same effect on the transmission risk for individuals from low-risk populations as individuals from high-risk populations. As a result of differences in the proportion of potential partners who are recently infected, the risk of acquiring HIV from one sexual exposure with one randomly selected ‘HIV-negative’ discloser from a high-risk population may actually be greater than the risk associated with a randomly selected ‘HIV-positive’ discloser (Table 1).
These provocative results are based upon the simple modeling of empirical data when they exist , epidemiological patterns found in the United States for disease and treatment rates , and conservatively biased estimates of transmission rates, which do not factor the relative frequency of anal versus vaginal sex, the rates of sexually transmitted infections, or recreational drug use, any of which would probably increase the rate of transmission above the estimates we do use [6,7,12]. The limitations of our model include that, whereas clade B virus predominates in the United States where men who have sex with men comprise the majority of infected individuals, the only source of empirical data for calculating rates of sexual transmission is a heterosexual cohort in Africa with non-clade B virus . Few in this cohort were on effective therapy, so the rate for transmission from a source with chronic, treated HIV was very conservatively estimated to be one-seventh the rate from an untreated source on the basis of an expected reduction in viral load and its correlation with infectiousness . In addition, we assumed that newly diagnosed individuals with recent infection will not participate in high-risk activities before antiretroviral therapy effectively controls their viral load, and we assigned them to the ‘on treatment’ group of ‘HIV-positive’ disclosers, but this may not always be so.
Our conservative calculations show that serosorting based on disclosure is not likely to be an effective prevention strategy when the prevalence of recently infected ‘HIV-negative’ disclosers comprises approximately 4% of the potential sex partner population. This is a realistic estimate based on current data . MacKellar et al.  found that among young men who have sex with men who had disclosed as ‘HIV negative’, 6% had unknowingly seroconverted in the previous 6 months. Using 6% as our estimated rate, the total cumulative risk of transmission per sexual exposure for someone who serosorts with the group of ‘HIV-negative’ disclosers would be 576.5 in a million compared with a risk of 395 in a million by sorting with the ‘HIV-positive’ disclosers group. In populations such as these, HIV-uninfected individuals who try to serosort may be more likely to become HIV infected than if they had not tried to serosort in the first place.
In conclusion, the effectiveness of serosorting on the basis of mutual disclosure of perceived HIV status is a flawed strategy for reducing sexual transmissions of HIV when it does not consider the prevalence of recent HIV infections in specific populations. Importantly, the individuals at greatest risk of HIV infection predictably belong to these very groups having the greatest proportions of recently infected people. By ignoring the increased potential for HIV transmission by recently infected individuals, serosorting may paradoxically increase the number of new HIV infections in certain populations.
Sponsorship: This work was supported by National Institutes of Health (NIH) funded grants AI27670, AI043638, 2 P30 MH62512-06, K23 (AI55276), the Adult AIDS Clinical Trials Group funded by the National Institute of Allergy and Infectious Diseases, and the AACTG Central Group grant (U01AI38858), the UCSD Center for AIDS Research (AI 36214), AI29164, AI047745, the AIDS training grant AI07384 from the NIH, and the Research Center for AIDS and HIV Infection of the San Diego Veterans Affairs Healthcare System.
Conflicts of interest: None.
1. Varghese B, Maher JE, Peterman TA, Branson BM, Skeketee RW. Reducing the risk of sexual HIV transmission: quantifying the per-act risk for HIV on the basis of choice of partner, sex act, and condom use. Sex Transm Dis 2002; 29:38–43.
2. Parsons JT, Scrimshaw EW, Wolitski RJ, Halkitis PN, Purcell DW, Hoff CC, Gomez CA. Sexual harm reduction practices of HIV-seropositive gay and bisexual men: serosorting, strategic positioning, and withdrawal before ejaculation. AIDS 2005; 19(Suppl. 1):S13–S25.
3. MacKellar DA, Valleroy LA, Behel S, Secura GM, Bingham T, Celentano DD, et al
. Unintentional HIV exposures from young men who have sex with men who disclose being HIV-negative. AIDS 2006; 20:1637–1644.
5. Pilcher CD, Tien HC, Eron JJ, Vernazza PL, Leu SY, Stewart PW, et al
. Brief but efficient: acute HIV infection and the sexual transmission of HIV. J Infect Dis 2004; 189:1785–1792.
6. Centers for Disease Control and Prevention. Methamphetamine use and HIV risk behaviors among heterosexual men–preliminary results from five northern California counties, December 2001–November 2003
. MMWR Morb Mortal Wkly Rep
7. Cohen MS. Sexually transmitted diseases enhance HIV transmission: no longer a hypothesis. Lancet 1998; 351(Suppl. 3):5–7.
8. Katz MH, Schwarcz SK, Kellogg TA, Klausner JD, Dilley JW, Gibson S, McFarland W. Impact of highly active antiretroviral treatment on HIV seroincidence among men who have sex with men: San Francisco. Am J Public Health 2002; 92:388–394.
9. Quinn TC, Wawer MJ, Sewankambo N, Serwadda D, Li C, Wabwire-Mangen F, et al
. Viral load and heterosexual transmission of human immunodeficiency virus type 1. Rakai Project Study Group. N Engl J Med 2000; 342:921–929.
10. Weinstock H, Sweeney S, Satten GA, Gwinn M. HIV seroincidence and risk factors among patients repeatedly tested for HIV attending sexually transmitted disease clinics in the United States, 1991 to 1996. STD Clinic HIV Seroincidence Study Group. J Acquir Immune Defic Syndr Hum Retrovirol 1998; 19:506–512.
11. Wawer MJ, Gray RJ, Sewankambo NK, Serwadda D, Li X, Laeyendecker O, et al
. Rates of HIV-1 transmission per coital act, by stage of HIV-1 infection, in Rakai, Uganda. J Infect Dis 2005; 191:1403–1409.
12. Royce RA, Sena A, Cates W Jr, Cohen M. Sexual transmission of HIV. N Engl J Med 1997; 336:1072–1078.
13. Little SJ, McLean AR, Spina CA, Richman DD, Havlir DV. Viral dynamics of acute HIV-1 infection. J Exp Med 1999; 190:841–850.
14. Centers for Disease Control and Prevention. AIDS cases, by geographic area of residence and metropolitan statistical area of residence, 2004
. Atlanta, GA: Centers for Disease Control and Prevention; 2006.
15. Doll LS, Holtgrave DR. The HIV/AIDS Prevention Research Synthesis Project: implications for federal HIV prevention policy. J Acquir Immune Defic Syndr 2002; 30(Suppl. 1):S130–S133.