Objective Three decades into the HIV epidemic and with the advancement of HIV treatments, condom and non-condom-based anal intercourse among gay men in resource-rich countries needs to be re-assessed.
Methods The proportions of men engaging in a range of anal intercourse practices were estimated from the ongoing cross-sectional Gay Community Periodic Surveys in six states in Australia from 2007 to 2009. Comparisons were made between HIV-negative men, HIV-positive men with an undetectable viral load and those with a detectable viral load.
Results Condoms play a key role in gay men's anal intercourse practices: 33.8% of HIV-negative men, 25.1% of HIV-positive men with an undetectable viral load and 22.5% of those with a detectable viral load reported consistent condom use with all male partners in the 6 months before the survey. Among HIV-negative men, the second largest group were men who had unprotected anal intercourse (UAI) only in the context of HIV-negative seroconcordant regular relationships. Among HIV-positive men, the second largest group was men who had UAI in casual encounters preceded by HIV status disclosure to some, but not all, casual partners.
Conclusions A minority, yet sizeable proportion, of men consistently engaged in a number of UAI practices in specific contexts, suggesting they have adopted deliberate HIV risk-reduction strategies. While it is important that HIV behavioural prevention continues to reinforce condom use, it needs to address both the challenges and opportunities of the substantial uptake of non-condom-based risk-reduction strategies.
- behavioural interv
- men who have sex with men (MSM)
- prevention of sexual transmission
- risk behaviours
- sexual behaviour
- sexual practices
- viral load
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
- behavioural interv
- men who have sex with men (MSM)
- prevention of sexual transmission
- risk behaviours
- sexual behaviour
- sexual practices
- viral load
Among homosexually active men in resource-rich countries, a number of strategies have been identified that reduce but do not eliminate the risk of HIV transmission during unprotected anal intercourse (UAI).1–9 The most notable strategies include negotiated safety,10–12 serosorting,4 13–15 strategic positioning2 4 and withdrawal before ejaculation into the anus.4 16 More recently, UAI with an HIV-positive partner is regarded by some gay men as less risky if the partner's viral load is undetectable rather than detectable.17–19
To date, there is little consensus over whether it is possible that some UAI practices are deliberate and planned HIV risk-reduction strategies. Also, in the literature, UAI risk-reduction strategies have been given various definitions, mostly inferred from descriptive behavioural patterns. The Swiss study is the only one that evaluated both intentions to use and adoption of UAI risk-reduction strategies.20 In that study, no significant differences were found in the key characteristics of men who reported consistent condom use; unprotected anal intercourse with casual partners (UAIC) employing additional serosorting, strategic positioning or withdrawal strategies; or UAIC with no such strategies. The Swiss findings suggest that the use of risk-reduction strategies during UAIC may be an opportunistic response rather than a planned strategy.20 Two San Francisco studies in 2004 and 2008 found that the use of UAI risk-reduction strategies differed by HIV status.21 22 However, UAI risk-reduction strategies in relation to relationship types and viral load levels have not been systematically investigated.
Using data from the ongoing Gay Community Periodic Surveys across Australia from 2007 to 2009, we estimated the proportions of men who consistently used condoms for anal intercourse and men who had any UAI. Comparisons are made between HIV-negative men, HIV-positive men with an undetectable viral load and those with a detectable viral load. We explored three research questions. First, what proportions of men are 100% condom users? Second, do 100% condom users and men having UAI differ in terms of relationship type (regular or casual), HIV status (their own and their partners'), and latest viral load level (if HIV positive)? Finally, do these different patterns of anal practices imply the deliberate adoption of HIV risk-reduction strategies?
Gay Community Periodic Surveys have consistently used a convenience time–location sampling method across six states in Australia since 1998.23–25 The research protocol has been approved by the Human Research Ethics Committee of the University of New South Wales. Participants complete an anonymous questionnaire on site. The average response rate is approximately 70%.26 (Detailed descriptions of recruitment sites, response rates and corresponding questionnaires can be found in each state report, accessible via http://www.nchsr.arts.unsw.edu.au/publications/).
From 2007 to 2009, there were 19 648 valid responses. Men who did not know their HIV status or who had never tested for HIV were excluded, as were HIV-positive men who did not report their latest viral load levels (n=3273, 16.7% excluded). The remaining 16 375 responses were analysed.
We developed a hierarchy of nine mutually exclusive anal intercourse practices, with contextual differentiation between regular relationships and casual encounters as a proxy of partner familiarity, and ordered the practices from the lowest to the highest risk reduction of HIV transmission.7 27 28 Each participant was placed in the potentially highest category of risk behaviour they reported in the 6 months before the survey:
no male sex partners;
no anal intercourse with any male partner;
consistent condom use with all male partners;
Unprotected anal intercourse with one's regular male partner (UAIR) and no UAI with casual male partners (UAIC):
seroconcordant UAIR: all UAIR was with a regular partner who had the same HIV status as the participant;
serononconcordant UAIR: all UAIR was with a regular partner whose HIV status was unknown;
serodiscordant UAIR: all UAIR was with a regular partner who had a different HIV status from that of the participant;
UAIC, with or without UAIR:
UAIC with consistent HIV status disclosure: UAIC was preceded by disclosure of HIV status to every casual partner;
UAIC with some HIV status disclosure: UAIC was preceded by disclosure of HIV status to some, but not all, casual partners; and
UAIC with no HIV status disclosure: UAIC without disclosure of HIV status to any casual partners.
In addition, positioning and withdrawal during UAI were examined. First, we assessed whether men practised insertive-only or receptive-only UAI (see supplementary appendix 1 for a detailed description of positioning during UAI by three groups of men, available online only). In this paper, HIV-negative men who had insertive-only UAI or HIV-positive men who had receptive-only UAI were defined as practising consistent seropositioning. Second, if participants consistently avoided ejaculation inside of partners as well as from partners during UAI, they were defined as having UAI with consistent withdrawal. These two practices are not mutually exclusive, and each was evaluated in combination with the six UAI practices (ie, a two by six stratification).
Of the 16 375 participants, 90.3% reported being HIV negative at the time of the survey, 6.8% reported the latest viral load as undetectable, and 2.9% reported the latest viral load as detectable. Men were largely recruited from gay community social venues and events (81%), with the majority self-identifying as gay or homosexual (89.7%). As shown in table 1, HIV-negative men were younger and more likely to be recruited from social sites, better educated and full-time employed. On the other hand, HIV-positive men were older and more likely to be recruited from clinic sites, of Anglo-Australian background, identify as gay or homosexual and have predominantly gay friends. Apart from being older, men with an undetectable viral load were similar to those with a detectable viral load.
Anal intercourse practices
Of the 14 785 HIV-negative men, 51.7% reported no UAI. As shown in table 2, the largest group was consistent condom users (33.8%). The second largest group engaged in seroconcordant UAIR exclusively (n=3022, 20.4%). The majority of these men (n=1794) reported verbal agreements with their regular partner that allowed UAI within their relationship but no sex (60.7%), no anal sex (6.5%) or no UAI (32.8%) with other men. That is, these men had adopted a ‘negotiated safety’ strategy (59.4%, 95% CI 57.6% to 61.1%). Among HIV-negative men who reported any UAIC (19.8%), disclosure of HIV status before sex did not affect whether they proceeded to UAIC.
Of the 1109 HIV-positive men with an undetectable viral load, 42.6% reported no UAI. As shown in table 2, the largest group was consistent condom users (25.1%). This was followed by men reporting UAIC after disclosing their HIV-positive status to some (19.6%) and to all (13.9%) of their casual partners. A slightly higher proportion of men with an undetectable viral load reported exclusive seroconcordant UAIR (9.6%) than exclusive serodiscordant UAIR (4.3%).
Of the 481 HIV-positive men with a detectable viral load, 35.8% reported no UAI. As shown in table 2, the largest group was, again, consistent condom users (22.5%). Similar to men with an undetectable viral load, this was followed by men reporting UAIC after disclosing their HIV-positive status to some (22.2%) or all (18.9%) of their casual partners. More men with a detectable viral load reported exclusive seroconcordant UAIR (7.9%) than exclusive serodiscordant UAIR (3.7%).
In comparison, a higher proportion of HIV-negative men reported no anal intercourse, consistent condom use and exclusive seroconcordant or serononconcordant UAIR. A higher proportion of HIV-positive men, on the other hand, reported serodiscordant UAIR and UAIC after consistent or partial HIV status disclosure. Reporting no sex partners and having UAIC without disclosure were similar across three groups.
UAI practices by consistent seropositioning and by consistent withdrawal
Of all participants, 10.9% reported consistent seropositioning and 11.5% reported consistent withdrawal with all male partners. The point prevalence estimates of consistent seropositioning and withdrawal are shown in table 3, stratified by UAI practices.
The most distinctive patterns in table 3 were among men who reported exclusive UAIR and no UAIC. Compared with HIV-negative men and regardless of viral load levels, a higher proportion of HIV-positive men reported receptive-only UAI in the context of a serodiscordant regular relationship (ie, strategic positioning), whereas a lower proportion of HIV-positive men reported receptive-only UAI in the context of a seroconcordant regular relationship. Similar patterns apply for consistent withdrawal.
For HIV-positive men, variation in viral load levels made little difference in condom and non-condom-based anal intercourse practices. As shown in table 3, the only marked difference was that a higher proportion of men with a detectable viral load adopted consistent withdrawal during UAIC after disclosure to some casual partners.
After decades of the HIV epidemic in Australia, consistent condom use for anal intercourse remains the most common HIV prevention strategy for both HIV-negative and HIV-positive gay men. Approximately one-third of all HIV-negative men and close to a quarter of all HIV-positive men always used condoms for anal intercourse within a 6-month period. These estimates are comparable to those reported from the USA.21 22
Approximately 60% of HIV-positive men reported UAI within a 6-month period, higher than that of HIV-negative men (approximately 50%). Our estimate of UAI among HIV-positive men in Australia is similar to that reported in San Francisco (61.0–66.7%).21 22
To investigate whether some UAI practices are adopted as deliberate HIV risk-reduction strategies, our study is the first to examine systematically the specific contexts of such practices, combining relationship types with HIV status concordance between respondents and their male sex partners. We regard UAI practices as deliberate HIV risk-reduction strategies if they occur consistently in specific contexts with no UAI occurring in other contexts. On the other hand, we consider UAI practices as opportunistic if they occur inconsistently or randomly across different settings. Based on the behavioural patterns described in this paper, we propose that a number of UAI practices are likely to be deliberate HIV risk-reduction strategies and practised by sizeable proportions of men. Although we have no data on intentions, it is highly unlikely that such consistent behavioural patterns are opportunistic. Rather, they indicate the deliberate adoption of strategies to reduce the risk of HIV transmission to self or partners.20
First, approximately 20% of HIV-negative men restricted UAI to their known or assumed HIV-negative seroconcordant regular partner, indicating serosorting within a regular relationship. In particular, close to 60% of these men further reported a matching verbal agreement, suggesting the continued popularity of the negotiated safety strategy among Australian HIV-negative gay men. A smaller proportion of HIV-positive men (9%) engaged in HIV-positive seroconcordant UAIR exclusively, again indicating serosorting within a regular relationship.
Second, approximately 15% of HIV-positive men engaged in UAIC preceded by consistent HIV positivity disclosure to all casual partners. Given that HIV status disclosure in casual encounters tends to be reciprocal,29 this, to some degree, suggests the adoption of HIV-positive seroconcordant serosorting as a deliberate strategy to prevent HIV transmission in casual encounters. We did not, however, ask participants directly about their casual partners' HIV statuses. A smaller proportion of HIV-negative men (5%) behaved similarly with respect to UAIC, suggesting that they engage in HIV-negative seroconcordant serosorting in casual encounters.
Third, seropositioning and withdrawal before ejaculation, if practised consistently during UAI, are deliberate risk-reduction strategies. In line with earlier research, however, our data suggest that they remain minority practices, with each being reported by approximately 10% of gay men overall.2 30 Compared with the popularity of serosorting in regular relationships among HIV-negative men and serosorting in casual encounters among HIV-positive men, our findings suggest that these two practices remain supplementary risk-reduction strategies. It is also noteworthy that in the context of known or assumed serodiscordance, a minority of HIV-positive men tend to practise receptive-only UAI consistently (ie, strategic positioning, 2%) or withdrawal (1%), to protect their HIV-negative regular partners from HIV infection.
There are a number of limitations to our study. First, our sample represents gay men who are highly socially and sexually involved with other gay men and are at a comparatively higher risk of HIV transmission than gay men in general.25 It is also likely that some individuals might have participated in the anonymous surveys repeatedly. Of note, the behavioural measures were constructed post-hoc and participants were not asked about their intentions to engage in particular practices. Furthermore, participants were asked to report their own and their regular partner's HIV status without confirmatory serological testing. HIV status disclosure to casual partners was not assessed for each partner (or by episode) and casual partners' HIV statuses were not recorded. Therefore, we could not accurately evaluate serosorting in casual encounters. Finally, causal inferences cannot be drawn from our cross-sectional data.
Our findings highlight the need for continuing reinforcement of condom use for anal intercourse, as it is still the main practice of gay men. Also, none of the assessed UAI risk-reduction strategies is as effective as consistent condom use in preventing HIV transmission.7 27 31–33 However, given that a notable minority of men engage in a range of UAI practices with different levels of HIV risk, HIV behavioural prevention needs to recognise gay men's willingness to practise UAI and their acceptance of some HIV risk.2 34–37 HIV prevention now faces both opportunities and challenges in this regard. Instead of being counterproductive, supporting some well-documented UAI risk-reduction strategies, such as negotiated safety in the earlier years and serosorting more recently, may add value to HIV prevention.38 39 However, the focus on establishing HIV seroconcordance, in particular, may well deepen divisions within a population that continues to stigmatise and discriminate against HIV-positive men.34 40 41
Consistent condom use for anal intercourse remains the most common HIV prevention strategy for both HIV-negative and HIV-positive gay men in Australia.
Approximately 60% of HIV-positive gay men and 50% of HIV-negative gay men reported any UAI in a 6-month period in Australia.
UAI in the context of known or perceived seroconcordance (ie, serosorting) has become a relatively common HIV risk-reduction strategy among Australian gay men.
The authors would like to thank their key community partners including the Australian Federation of AIDS Organisations (AFAO), the National Association of People Living with HIV/AIDS (NAPWA) and the AIDS councils and organisations of people living with HIV in all participating states/territories.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Download Supplementary Data (PDF) - Manuscript file of format pdf
Funding The Gay Community Periodic Surveys are funded by the departments of health in each participating state or territory in Australia. The National Centre in HIV Social Research and The Kirby Institute receive project funding from the Australian government Department of Health and Ageing.
Competing interests None.
Ethics approval Ethics approval was received from the Human Research Ethics Committee at the University of New South Wales.
Provenance and peer review Not commissioned; externally peer reviewed.