Recent data demonstrate that many men who have sex with men (MSM) do not adhere to consistent condom use or abstinence from anal sex.1-5 For some HIV-negative MSM who choose not to use condoms for anal sex with their seroconcordant primary partners, "negotiated safety" (NS) has emerged as an alternative risk reduction strategy. In 1993, Kippax and colleagues6 first used the term to describe seroconcordant HIV-negative primary relationships among men who have unprotected anal sex (UA) together but not with partners outside the primary relationship. The major tenets of this type of relationship include confirmation of HIV-negative status by both partners through mutual HIV testing and establishment of specific agreements or rules prohibiting UA outside the relationship.6-10 Those rules may include agreeing to be monogamous or only allowing comparatively lower risk sexual activities with others (eg, unprotected oral sex or protected anal sex).11 By forming such agreements, HIV-negative men practicing NS may be at relatively low risk of infection from their primary partner despite engaging in UA together.10 An important comparison to consider in evaluating NS and the risk of HIV transmission within a primary relationship is sexual behavior with casual partners among men practicing NS and men choosing another risk reduction strategy-having no UA with primary partners (NUAP) by either always using condoms or avoiding anal sex entirely. Men practicing NUAP minimize the risk of transmitting HIV within the relationship, regardless of risk behavior with casual partners and whether or not such behavior is disclosed to the primary partner.
Despite its potential to reduce HIV risk, NS may present some difficulties for those who attempt to apply it successfully to their relationships. Some men could have trouble adhering to established rules regarding sex with others despite having agreed to them. In theory, risk behavior with casual partners should be lower among men practicing NS compared with men practicing NUAP and other men who are not in relationships because of the presence of specific agreements that attempt to minimize risk with casual partners. Risk within NS relationships may be elevated when NS agreements are incomplete or not fully adhered to by both partners, however. To reduce the risk of HIV transmission within the relationship as a result of activity with casual partners, NS agreements should also include the additional safeguard of partners mutually agreeing to disclose high-risk sex that may occur outside the relationship.
Prior studies have not described the presence of disclosure agreements within NS relationships, which may limit the interpretation of the findings. We investigated characteristics of men practicing NS, how they compared with those of with men adopting other strategies to manage risk within relationships, the details of their NS and disclosure of risk behavior agreements, and to what extent men adhered to their agreements. The study reported here is the first to describe characteristics of men practicing NS and adherence to NS in a diverse community-based sample within the United States (San Francisco Bay Area) in which the incidence of HIV among MSM seems to be increasing.12,13
Recruitment and Data Collection
Data were collected from a cross-sectional survey of 554 MSM in late 2000 and early 2001. Men were recruited through participant referral, from street locations, and from venues (eg, bars, dance clubs, gyms, community agencies) in the San Francisco Bay Area, including Oakland and the Tenderloin, Mission, and Castro Districts in San Francisco. Recruitment locations were identified through focus groups, pilot testing, and previous studies. African-American and Latino MSM venues were oversampled to increase representation of these groups. Eligible men were at least 18 years old, lived or worked in the San Francisco Bay Area, and either identified as gay/bisexual or reported having had sex with a man in the prior year. Participants were scheduled to complete interviewer-administered questionnaires at the San Francisco Health Department or community agencies. Participants gave informed consent and received $25. Interviewers were trained to conduct interviews about sensitive topics and informed participants that no personal identifying information was linked to study data or retained after participation.
Negotiated safety relationships were defined as those in which HIV-negative men were in a current seroconcordant primary relationship of at least 6 months'duration with another man, had UA together within the past 3 months, and had rules prohibiting UA with others outside the relationship, although NS men may have violated such rules. To avoid including people within the HIV-antibody test window period who might have been unsure of their HIV-negative status because of behavior before their relationship began, we restricted the NS category to men who had been together for at least 6 months, as done in previous studies.7,9,10,14 The definition of UA included insertive or receptive intercourse with or without ejaculation. A current primary relationship was defined as one in which the partner was a man the participant lived with or saw frequently and to whom the participant felt a special emotional commitment. HIV status was determined by self-report, which has recently been shown to be highly consistent with actual testing in a study of MSM that included men from San Francisco.15 Seroconcordance was determined by participant report of a partner's HIV status. The presence of NS agreements was established through an open-ended question regarding specific rules that both partners had agreed to regarding sexual behavior outside their relationship. NS was defined by a response indicating prohibitions against UA outside the relationship and a report of less than 100% condom use for anal sex with a primary partner in the prior 3 months.
Adherence to NS agreements was examined based on self-reported sexual behavior in the past 3 months. In an attempt to reduce socially desirable responding, the sexual behavior items appeared in a separate section of the assessment instrument and were administered before the questions on relationship strategies. We determined whether men engaged in UA with a man other than their NS primary partner or whether they broke another rule established as part of their NS agreement. Men were also asked if they had an agreement with their NS partner to disclose rule breaking and if they had broken these rules within the past 12 months or since the rules were established if less than 12 months ago. Information on oral sex with casual partners was collected, but this behavior alone did not violate NS agreements unless specifically mentioned as part of the agreement.
In contrast to those practicing NS who were required to have reported UA with their primary partner, men who were currently in a primary relationship with a man and had no UA with this primary partner in the past 3 months were categorized as NUAP. This included those who had no anal sex with their primary partner or had 100% condom use for anal sex with their primary partner. Questions regarding relationship risk reduction strategies were not asked of men who were not in relationships.
Of the men who reported that they were HIV-negative, we identified those who were in current seroconcordant primary relationships with other men for at least 6 months. In this subgroup, we identified the men who reported NS or NUAP risk reduction strategies. Of those in NS relationships, we examined adherence to NS rules and the percentage of men who had disclosure agreements with their partners. The NS, NUAP, and not in primary relationship groups were compared on the basis of behavioral (eg, anal sex with casual partners, self-report of sexually transmitted infection [STI] in past 12 months) and demographic (age, education, race/ethnicity, and income) variables using the χ2, Fisher exact test (2-tailed), or Mantel-Haenszel χ2 test for trend. Multivariate models consisting of demographic variables were constructed to assess factors potentially associated with type of relationship. Age was included as a continuous variable; other demographic variables were included as categoric variables. Those variables found to be significant at P < 0.05 were included in the multivariate models.
Prevalence and Demographics
Among 340 HIV-negative men, 265 (78%) reported anal intercourse with a male partner in the prior 3 months and 143 (42%) reported UA. Regarding relationship status, 203 (60%) were not currently in a primary relationship, 34 (10%) were in a relationship with an HIV-positive or unknown HIV status man, and 103 (30%) were in a seroconcordant relationship of any length. Seventy-six of these 103 (22% of all HIV-negative men) were in a seroconcordant primary relationship of at least 6 months' duration with another man. None of these reported >1 simultaneous primary relationship. Of these 76 men, NS was more common than the NUAP strategy (38 [50%] vs. 30 [39%]). Additionally, 8 (11%) men practiced UA within the relationship but had no rule prohibiting UA with other partners. Within the NUAP group, 17 (57%) practiced 100% condom use for anal sex with their primary partner and 13 (43%) reported no recent anal sex together. Among those in a relationship of at least 6 months' duration, younger participants and those with a higher income were more likely to practice NS than NUAP (Table 1). The relationship for age remained significant in a multivariate model that included age and income, such that older men had reduced odds of practicing NS versus NUAP (adjusted odds ratio [AOR] = 0.94 per year, 95% confidence interval [CI]: 0.89-1.00). There were no significant differences in choosing either the NS or NUAP strategy by race/ethnicity, income, or education level. No significant sociodemographic differences were found in multivariate analysis between the NS group and those not in a relationship. In comparing the NUAP group with men not in a primary relationship by means of multivariate analysis, older men (AOR = 1.05 per year, 95% CI: 1.01-1.10) and Latinos versus whites (AOR = 3.50, 95% CI: 1.15-10.72) were more likely to report NUAP.
Risk Behavior With Casual Partners
Both groups of men in relationships were similarly likely to state that they allowed sex with other partners outside the relationship (58% for NS vs. 57% for NUAP). Men practicing NS were more likely to have had anal sex with other partners than were NUAP men (47% vs. 20%; P = 0.02). Anal sex with nonprimary partners was highest (70%) among men not in a primary relationship. Sample size limited power to detect differences regarding other specific risk behaviors. There was a tendency toward higher risk behavior outside the relationship among the NS group, however. Eighteen percent of men in NS relationships reported having had 1 of 4 STIs (gonorrhea, Chlamydia, syphilis, or nonspecific urethritis) in the prior 12 months compared with 3% of those practicing NUAP (P = 0.07) and 7% of men not in primary relationships (P = 0.06). Among men practicing NS, 18% reported UA with a casual partner in the prior 3 months, a rate not significantly different than for the NUAP group (7%) or the not in a relationship group (31%).
Negotiated Safety Relationship Rules
Men in NS relationships (n = 38) reported a variety of rules (perceived to be mutually agreed on with partners) for their relationships, which encompassed disallowing UA with casual partners. Fifty percent (19 of 38 men) stated their NS-specific rule as not allowing any type of sex with men outside the relationship, but 26% of these participants (5 of 19 men) reported sexual behavior in the prior 3 months that broke that rule. Another 42% (16 of 38 men) of those in NS relationships had a rule specifically prohibiting UA outside the relationship, but 31% of these individuals (5 of 16 men) had engaged in UA with another man in the past 3 months. Finally, 8% (3 of 38 men) in the NS group had a rule specifying no anal sex (protected or unprotected) with others, but 1 of these 3 men did engage in anal sex with another man. Overall, 29% (11 of 38 men) of those in NS relationships broke their NS-specific rule based on self-reported sexual behavior in the prior 3 months, including 7 who reported UA, which carries the highest risk.
Of the men in NS relationships, 74% (28 of 38 men) reported that they had an agreement with their primary partner to disclose rule breaking to each other. If this type of agreement is factored into the analysis of NS, we find that 23 of the 38 men (61%) met the strictest criteria of having an agreement to disclose rule breaking as well as reporting that they had not broken any of the rules. Among those with a disclosure agreement, 5 (18%) broke a rule, whereas 6 (60%) of those without such an agreement broke a rule.
Of HIV-negative men in seroconcordant relationships of at least 6 months' duration, half practiced NS despite there having been no community-level campaigns promoting NS as a prevention strategy in the San Francisco Bay Area. Nearly 2 in 5 men with NS agreements failed to adhere to rules and to have an agreement to disclose rule breaking, however. NS has been evaluated in prior studies based on sexual behavior outside the primary relationship and by comparison with men in primary relationships governed by different rules.7,9,10,16 Kippax et al10 found that for HIV-negative MSM in seroconcordant relationships, having an NS agreement was associated with a reduced risk of UA outside the relationship compared with men in relationships not having such an agreement. Another study found that UA with casual (nonprimary) partners among men in NS relationships was not higher than among men in relationships who agreed to consistently use condoms with both primary and casual partners.9 Our finding that 18% of men in NS relationships had UA outside the relationship in the prior 3 months appears somewhat higher than the levels of 6.6% to 11% found in other studies based on 6 months of behavior.7,9,10 Nearly 1 in 5 of our NS participants recently engaged in high-risk sexual behavior with others that is expressly prohibited by NS agreements, and nearly one third broke the specific rule they agreed to. Because men practicing NS have specific agreements prohibiting UA with other partners, we expected that they would report a lower prevalence of UA with casual partners compared with other HIV-negative men not currently in primary relationships, but our data did not support this expectation and showed similar levels of UA among both groups.
This study allowed for determination of the prevalence of NS among a diverse sample of MSM. Although we were able to recruit a very diverse sample (over two thirds of whom were men of color) successfully, participants were only recruited from the San Francisco Bay Area, which may not generalize to the broader population of MSM in this and other areas. The relatively small number of participants from the full study sample who ultimately qualified for this analysis limited power to detect behavioral differences with casual partners by relationship risk reduction strategy type. Although we attempted to minimize social desirability bias by asking sexual behavior questions before relationship questions, the data were collected in a face-to-face interview and some men in NS relationships may have been reluctant to admit to rule breaking directly, because the rule-breaking question was asked after the question about what their rules were. Although we strove for consistency with prior publications in our definition of NS, more restrictive definitions were not used, such as those proffered in some prevention materials, which involve up to 13 specific steps, including regular communication about the agreement and an immediate return to condom use if either partner expresses such a desire.17 It is likely that additional questions regarding further steps in establishing a safe agreement would have found even fewer men practicing NS and fewer adhering to a strict NS definition.
Our results raise important concerns about NS and suggest several potential prevention strategies to enhance likelihood of NS success. Because adequate education regarding the complexity of NS agreements includes a long list of procedures to follow, there is a danger that some men may decide to follow only some of the recommendations and believe the measures they have already adopted are sufficient for protection when they are not.8 Nevertheless, it is important to acknowledge that some men attempt NS and may not be willing to use condoms consistently with their primary partners. With these men, clinicians, counselors, and health educators should discuss how NS agreements may fail and suggest specific efforts regarding communication with partners about NS. Such discussions should include clarifying all the tenets of a safer NS agreement and how to deal with agreement failure, as has been outlined in some NS health promotion materials.17-20 Testing for STIs should also be strongly encouraged before men cease using condoms with primary partners and in instances when NS agreements are broken. Given that we found young men are more likely to be in NS relationships and other data demonstrate that young men are at elevated risk of HIV infection, including from their primary partners, specific efforts may be warranted to educate young men about potential pitfalls in NS agreements.21-24
Educating men about NS involves considerably more complexity than a recommendation to use condoms with all partners and is less easily adaptable to social marketing or other widely used models of community health promotion, particularly those geared toward low-literacy populations. Until further data are collected from studies of US MSM regarding their NS relationships, including factors associated with agreement rule breaking, caution should be used in endorsing NS as an alternative risk reduction strategy to condom use for US HIV-negative men in seroconcordant relationships.
The authors thank Eric Vittinghoff for assistance with data analysis and Jonas Abella, Oscar Macias, and Jesus Perez for their work in recruiting and interviewing participants.
1. 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 Morb Mortal Wkly Rep.
2. Kellogg T, McFarland W, Katz M. Recent increases in HIV seroconversion among repeat anonymous testers in San Francisco. AIDS.
3. del Romero J, Castilla J, Garcia S, et al. Time trend in incidence of HIV seroconversion among homosexual men repeatedly tested in Madrid, 1988-2000. AIDS.
4. Hogg RS, Weber AE, Chan K, et al. Increasing incidence of HIV infections among young gay and bisexual men in Vancouver. AIDS.
5. Ostrow DE, Fox KJ, Chmiel JS, et al. Attitudes towards highly active antiretroviral therapy are associated with sexual risk
taking among HIV-infected and uninfected homosexual men. AIDS.
6. Kippax S, Crawford J, Davis M, et al. Sustaining safe sex: a longitudinal study of a sample of homosexual men. AIDS.
7. Davidovich U, deWit JB, Stroebe W. Assessing sexual risk
behaviour of young gay men
in primary relationships
: the incorporation of negotiated safety
and negotiated safety
8. Elford J, Bolding G, Maguire M, et al. Sexual risk
behaviour among gay men
in a relationship. AIDS.
9. Crawford JM, Rodden P, Kippax S, et al. Negotiated safety
and other agreements between men in relationships: risk practice redefined. Int J STD AIDS.
10. Kippax S, Noble J, Prestage G, et al. Sexual negotiation in the AIDS era: negotiated safety
11. Vittinghoff E, Douglas J, Judson F, et al. Per-contact risk of human immunodeficiency virus transmission between male sexual partners. Am J Epidemiol.
12. Chen SY, Gibson S, Katz MH, et al. Continuing increases in sexual risk
behavior and sexually transmitted diseases among men who have sex with men: San Francisco, CA, 1999-2001. Am J Public Health.
13. San Francisco Department of Public Health. HIV/AIDS Epidemiology Annual Report. San Francisco HIV Seroepidemiology and AIDS Surveillance Section.
14. Horsburgh CR, Oy CY, Jason J, et al. Duration of human immunodeficiency virus infection before detection of antibody. Lancet.
15. Osmond DH, Catania J, Pollack L, et al. Obtaining HIV test results with a home collection test kit in a community telephone sample. J Acquir Immune Defic Syndr.
16. Ekstrand M, Stall R, Kegeles S, et al. Safer sex among gay men
: what is the ultimate goal? AIDS.
17. Billington A. Thinking It Through: A New Approach to Sex, Relationships and HIV for Gay Men.
London: HPS Camden and Islington Community Health Services, National Health Service Trust; 1997. Available at: http://www.thinking-it-through.co.uk/text10.htm
18. Information and Services of Santa Clara County. Is it Safe to Have Unsafe Sex With Your Partner?
San Jose: AIDS Resources, Information and Services of Santa Clara County; 2001. Available at: http://www.aris.org/pages/boyfriend.htm
19. Hildon A. Agreements in Relationships: The Essential Guide for Gay Men in Relationships Who Want to Stop Using Condoms When They Have Sex With Each Other.
London: HPS Camden and Islington Community Health Services, National Health Service Trust. 2001. Available at: http://www.freedoms.org.uk/advice/air/air01.htm
21. Mansergh G, Marks G. Age and risk of HIV infection in men who have sex with men. AIDS.
22. Ekstrand ML, Stall RD, Paul JP, et al. Gay men
report high rates of unprotected anal sex with partners of unknown or discordant HIV status. AIDS.
23. Valleroy LA, MacKellar DA, Karon JM, et al. HIV prevalence and associated risks in young men who have sex with men. Young Men's Survey Study Group. JAMA.
24. Xiridou M, Geskus R, De Wit J, et al. The contribution of steady and casual partnerships to the incidence of HIV infection among homosexual men in Amsterdam. AIDS.