Serosorting does help prevent HIV – up to a point

This article is more than 12 years old. Click here for more recent articles on this topic

A meta-analysis of HIV-negative gay men’s sexual behaviour and HIV incidence rate in four HIV prevention studies, presented earlier this month at the 19th Conference on Retroviruses and Opportunistic Infections (CROI), has found that attempting to ‘serosort’ by restricting unprotected sex to partners known to be HIV negative does have efficacy as an HIV prevention strategy, when compared with using no strategy at all.

Serosorting is, however, considerably less effective in reducing the chances of acquiring HIV than four other strategies: 100% condom use, monogamy, only having insertive sex, or ‘seropositioning’ (only taking the bottom role with partners known not to have HIV and being top with partners of positive or unknown status). Interestingly, 100% condom use was the least effective of these other four strategies.

‘Seroadaptive’ behaviours include any method of attempting to reduce the risk of HIV acquisition or transmission by altering one’s sexual behaviour according to the HIV status of partners. The term ‘serosorting’ has been used in various different ways. Most commonly, it means restricting unprotected anal sex to partners known to have the same HIV status as yourself. When unprotected sex between HIV-negative men is confined to a primary relationship, with condoms used in all other encounters, this has been called ‘negotiated safety’.



Choosing sexual partners of the same HIV status, or restricting condomless sex to partners of the same HIV status. As a risk reduction strategy, the drawback for HIV-negative people is that they can only be certain of their HIV status when they last took a test, whereas HIV-positive people can be confident they know their status


How well something works (in a research study). See also ‘effectiveness’.


A sexual risk reduction strategy for gay men having anal sex without a condom. Different sexual positions are adopted according to HIV status: the HIV negative man takes the insertive role (‘top’) and the HIV-positive man the receptive role (‘bottom’).



When the statistical data from all studies which relate to a particular research question and conform to a pre-determined selection criteria are pooled and analysed together.


Insertive anal intercourse refers to the act of penetration during anal intercourse. The insertive partner is the ‘top’. 

While some studies have found serosorting in HIV-negative men to be effective, others have not. Attempted serosorting by HIV-negative people has an inherent drawback that serosorting by HIV-positive people lacks: people can only be certain of their status up to the first time they risk exposure to HIV after their last negative HIV test. Research indicates that a large minority of people in high-risk communities who assume they are HIV negative in fact have HIV, and that a large proportion of men who ‘know’ their partner’s HIV status have, in fact, tried to guess it.  

The meta-analysis

Nonetheless, though serosorting is fallible, a recent meta-analysis of studies presented at CROI found that serosorting halved the likelihood of acquiring HIV compared to having no strategy at all.

The study pooled behavioural data and HIV incidence rates from four different studies in gay men:

  • The HIVNET 001 Vaccine Preparedness Study (VPS), an observational study that took place in eight cities in the US between 1995 and 1997.
  • VAX 004, the first phase III efficacy trial of a candidate HIV vaccine, which took place at 61 sites in the US, Canada and the Netherlands between 1998 and 2001.
  • The EXPLORE study, a randomised controlled trial of a behavioural HIV-prevention intervention that took place in six US cities between 1999 and 2003.

  • The STEP study, a phase III trial of another candidate vaccine, which took place in North and South America and Australia between 2004 and 2007.

There were a total of 12,705 HIV-negative gay men from North America included in these trials, of whom 663 (5%) acquired HIV.

By analysing the results of sexual behaviour questionnaires, the researchers in the present study divided respondents' sexual behaviour into a risk hierarchy, running from what they hypothesised would be the safest option to the most risky. 'Hierarchy' means that only those not practising the first option were assessed for their use of the second option, only those not practising the second option assessed for their use of the third, and so on.

The options were:

  • No unprotected anal sex: either no anal sex at all or 100% condom use (47% of the group)

  • Monogamy: has unprotected anal sex, but only within a monogamous, seroconcordant relationship (11%)

  • Top only: only ever takes the insertive role in unprotected anal sex, regardless of condom use or partner’s serostatus (10%)
  • Serosorting: has unprotected anal sex but only with partners thought to be HIV negative (8%)
  • Seropositioning: unprotected receptive anal sex only with negative partners, always insertive with partners of positive or unknown HIV status (3%)
  • Risky sex: has unprotected sex with no risk reduction strategy (21%).

Throughout the studies, adoption of these strategies was fluid and inconsistent: only 23% of men in the studies maintained one specific strategy throughout the study.


The four study groups were in some ways very similar and in others very different. The average age was 34, with little variation between studies, and the groups were predominantly white (78%), with only 6% African American.

In other ways they were very different: EXPLORE was recruited from a ‘high-risk’ group of gay men and STEP from a largely low-risk group. One in six of the men had been celibate or monogamous in the last six months, but this varied from 8% in EXPLORE to 51% in STEP; 60% in EXPLORE but only 11% in STEP had had six partners or more in the last six months; 13% of the men in EXPLORE had used methamphetamine but only 4% in STEP.

All risk reduction strategies helped to curb HIV infections, but some much more than others – and the safety hierarchy wasn’t what researchers had expected.

HIV annual incidence in people with no safer-sex strategy was 2.95%. In serosorters, it was 1.44% (a 51% reduction in HIV incidence).

After this, the next riskiest thing was 100% condom use/no anal sex: the seroconversion rate in this group was 0.76% a year (74% reduction). Seropositioning was about as safe: the incidence rate was 0.73% (75% reduction).

Men who were ‘top only’ only had an HIV incidence rate of 0.4% (86% reduction). And the safest option of all was old-fashioned monogamy; in monogamous men the annual HIV incidence rate was only 0.25%, a 91.5% reduction in HIV risk.

Studies of serosorting have in the past suffered from small numbers and wide confidence intervals. Although the validity of the risk categories and the methodology of individual behavioural questionnaires may be challenged, this is a large enough meta-analysis to more reliably establish the risk of different methods. The 74% reduction seen with 100% condom use is not incompatible with figures arrived at in the relatively few studies of condom efficacy in gay men: what perhaps is more of a surprise is that having a trusted, monogamous relationship or being an exclusive top are so much more protective than this.


Vallabhaneni S et al. Seroadaptive behavior: association with seroconversion among HIV- MSM. Nineteenth Conference on Retroviruses and Opportunistic Infections (CROI), Seattle, abstract 140, 2012. The abstract is available on the official conference website.