Abstract
Blood plasma HIV-RNA load (BPVL) is the strongest predictor of HIV-1 transmission during sex. Unprotected anal intercourse (UAI) is the highest risk activity for transmission among men who have sex with men (MSM). Awareness of BPVL may influence rates of UAI. We assessed whether optimism towards antiretroviral therapy (ART) and/or biomedical factors influenced sexual activities with regular partners. Questionnaires were administered to 109 HIV-positive MSM participating in a cross-sectional study of BPVL and seminal viral load. The survey assessed HIV transmission beliefs and sexual practices with regular male partners in the past three months. Sixty-nine of 109 (63.3%) had been in a regular relationship and 42 reported having had anal sex. Unprotected receptive anal intercourse without ejaculation (URAI − e) was associated with awareness that their most recent BPVL was detectable (>50 RNA copies/mL) and not taking ART. Receptive UAI with ejaculation (URAI + e) was associated with not taking ART, having a sexually transmissible infection and having an HIV-positive partner; the latter was also associated with insertive UAI with ejaculation (UIAI + e). Treatment optimism was not associated with UAI. In this cohort, sexual practices were based more upon knowledge of biomedical factors rather than attitudes regarding transmission risks.
INTRODUCTION
In general, HIV-positive men who have sex with men (MSM) have unprotected anal intercourse (UAI) more frequently with other HIV-positive partners rather than HIV-negative or HIV-unknown partners. 1–4 In one study, however, a significant proportion of men (42.9%) reported UAI with casual male partners who were HIV negative or serostatus unknown; the majority of UAI with HIV-unknown partners may have resulted from the assumption that those partners were also HIV positive. 4 Individuals in sero-concordant relationships report higher rates of UAI than those in discordant relationships, interpreted as negotiated safety strategies to minimize HIV transmission. 2,5 These studies, however, do not examine if cognitive or biomedical issues contributed to UAI particularly for those in concordant positive and non-concordant relationships.
In a recent study of UAI between HIV-positive men in this cohort and their casual male partners (HIV negative or unknown serostatus), we found that cognitive factors (treatment optimism and sexual beliefs) and biomedical factors (recent blood plasma HIV-RNA load [BPVL] and current sexually transmissible infection [STI]) were associated with different types of UAI. 6 Cognitive variables were predominately associated with UAI in that study. The current analysis was undertaken to determine whether similar variables predicted different types of UAI with regular partners in the same cohort.
METHODS
Study design and subjects
The study is based on a cross-sectional cohort of 109 HIV-positive MSM 7 attending the Albion Street Centre, a metropolitan HIV/AIDS clinic in Sydney. The study was approved by the Institutional Ethics Committee. Eligible patients were HIV-positive MSM ≥18 years of age attending for routine care. Subjects were stratified into two groups: one group receiving antiretroviral therapy (ART) (36 were prescribed nucleoside reverse transcriptase inhibitor containing regimens while 45 had protease inhibitor containing regimens) and one untreated group (28 subjects).
Treated patients were required to have had three consecutive BPVL <50 copies/mL in the last 12 months, the most recent being obtained in the three months before enrolment, and to have taken three antiretroviral medications continuously for >12 months. Patients were excluded if they had genital symptoms, symptomatic genital herpes, a laboratory-proven STI or an active opportunistic infection. Subjects were screened for urethral and rectal chlamydia (polymerase chain reaction), pharyngeal and rectal gonorrhoea (culture) and syphilis (serology). Blood and semen were obtained at the same visit to determine BPVL and seminal viral load, respectively.
Questionnaires
At the end of the consultation, patients completed four self-administered questionnaires: an eight-item version of the Optimism-Scepticism (O-S) scale, 8 a modification of a previously developed Sexual Practices Questionnaire, 9 a Sexual Beliefs (S-B) scale and a clinical and demographics questionnaire. 6 Subjects designated sexual partners as ‘regular’ or ‘casual’ when completing the questionnaires. Responses could not be blinded as these were correlated with the patient's clinical data.
Statistics
Univariate analyses were undertaken testing associations for four types of UAI with regular partners: unprotected insertive anal intercourse with ejaculation (UIAI + e) and without ejaculation (UIAI − e), and unprotected receptive anal intercourse with ejaculation (URAI + e) and without ejaculation (URAI − e). Mann-Whitney U test was used for ordinal variables (Likert-type scales) and non-normal factors, and Pearson's chi-square test was used for categorical variables. Biopsychosocial associations with these sexual behaviours were examined using SPSS version 10 (Chicago, IL, USA).
RESULTS
Study subjects
All 109 patients enrolled in the study completed the questionnaires. The cohort had a mean age of 41 ± 9 years, a mean CD4 count of 595 cells/µL (range 150–1560) and had been HIV positive for a median of nine years (range 0–23). Approximately half (46.5%) completed tertiary education, 23.8% completed high school, 22.8% completed technical education and 6.9% completed some high school. Subjects reporting a regular partner in the past three months formed the basis for our analysis. Sixty-nine of the 109 subjects (63.3%) reported having been in a regular relationship in the past three months. Twenty reported that their regular partner was HIV positive, 42 HIV negative and seven did not know their regular partner's HIV status. There were only four subjects who did not disclose their HIV-positive status to partners.
Scales
As previously reported, 6 internal consistency for the modified O-S scale was 0.79 for the complete scale, while that for the S-B scale was 0.82. Both scales demonstrated high internal consistency and reliability in the current sample. There was a moderate level of correlation (rho = 0.55, P < 0.01) between the S-B and O-S scales.
Sexual behaviours
Forty-two of the 109 (38.5%) reported any anal intercourse with their regular male partner. Within the group reporting anal sex, seven subjects (16.6%) reported URAI − e, 14 (33.3%) reported URAI + e, six (14.2%) reported UIAI − e and 12 (28.5%) reported UIAI + e. Unlike previous studies, 10,11 treatment optimism did not correlate with UAI with regular partners in this study. Factors significantly associated with each of these sexual behaviours are discussed below.
URAI − e
Awareness that the most recent BPVL was detectable (4 of 10 subjects, 40%) cf. undetectable (1 of 23 subjects, 4%) (χ 2 = 6.9, P = 0.021) and currently not receiving ART (5 of 11 subjects, 45%) cf. receiving ART (2 of 28 subjects, 7%) (χ 2 = 7.87, P = 0.012) were significantly associated with URAI − e. Disclosure of subjects' HIV-positive status showed a trend with URAI − e (χ 2 = 5.623, P = 0.07).
URAI + e
Not receiving ART (7 of 11 subjects, 64%) cf. receiving ART (6 of 28 subjects, 21%) (χ 2 = 6.3, P = 0.022) having an HIV-positive partner (10/16 subjects, 62%) cf. an HIV-negative partner (3 of 22 subjects, 14%) (χ 2 = 9.8, P = 0.004) and currently having an STI (5 of 7 subjects, 71%) cf. no STI (9 of 35 subjects, 26%) (χ 2 = 5.5. P = 0.031) were significantly associated with URAI + e. There was a tendency for individuals who scored higher on the S-B scale to engage in this behaviour (mean = 10.07) cf. those who did not (mean = 8.46); however, the association was not significant (P = 0.09).
UIAI − e
No factors were significantly associated with UIAI − e.
UIAI + e
Finally, having an HIV-positive partner (11/16 subjects, 70%) cf. HIV-negative partner (1 of 22 subjects, 5%) (χ 2 = 17.7, P<0.001) was significantly associated with UIAI + e. Having an HIV-positive partner (11/16 subjects, 70%) cf. an HIV-unknown partner (none of three subjects) (χ 2 = 4.9, P = 0.058) was of borderline significance.
DISCUSSION
This study provides new information regarding factors associated with UAI among HIV-1 positive MSM and their regular partners. Cognitive factors (e.g. ART optimism) have been associated with UAI between HIV-positive men and their partners. In a large Australian community sample, men who engaged in UAI in the past six months with both casual and regular partners were significantly more optimistic about ART than those who reported no UAI. 10 In a recent meta-analysis, cognitive factors (e.g. beliefs that ART prevent transmission) were again shown to be associated with UAI, while biomedical factors (e.g. undetectable BPVL) were not. 12
Other studies have demonstrated an association between biomedical factors and UAI in HIV-positive MSM where BPVL and CD4 counts were shown to have different effects on UAI rates with casual and regular partners. 13 In particular, ART-related immunological and virological improvements were associated with UAI, particularly with casual partners. There was increased likelihood of UAI with casual partners when BPVL became undetectable after initiating ART, while in bi-variate analyses higher CD4 counts without ART and increasing CD4 counts with ART were associated with UAI with casual partners. Notably, UAI with regular partners was associated with having an undetectable BPVL while not receiving ART. Access to BPVL results was an important factor in this association. The results also indicated that UAI occurred more frequently with casual partners at higher BPVL but an association was not found with regular partners. 13
In a later study, 11 both biomedical (undetectable BPVL) and cognitive factors (ART optimism) were associated with UAI in serodiscordant relationships. ‘Strategic positioning’ – a factor with inherent cognitive and behavioural components – also contributed to UAI. In those serodiscordant relationships where the HIV-positive partner had an undetectable BPVL and UAI was practised, three-quarters of HIV-negative partners avoided URAI + e and similar proportions of HIV-positive partners avoided UIAI + e. 11
We found that biomedical factors rather than cognitive factors are primarily associated with UAI with regular male partners, in contrast to a previous report conducted with the same cohort, where predominantly cognitive factors were associated with UAI with casual male partners. 6 In that study, treatment optimism, sexual beliefs or a combination of both were associated with UIAI − e, URAI + e and UIAI + e with other HIV-positive partners. The cognitive factors were also associated with UIAI + e and UIAI − e with HIV-unknown partners and with URAI − e with HIV-negative partners. Biomedical factors (past or current STI and/or recent BPVL) were associated with URAI − e and UIAI − e with HIV-positive partners, and URAI + e and UIAI − e with HIV-unknown partners. 6
In the current study, it is of interest that untreated subjects who were aware that their most recent BPVL was detectable and who were not currently receiving ART reported URAI − e with their regular partners. URAI − e also had borderline association with the partner knowing the subject was HIV-positive. This finding suggests the practice of ‘strategic positioning,’ as URAI is perceived as a lower risk for transmission to the HIV-negative partner.
Having an HIV-positive partner rather than an HIV-negative partner was significantly associated with URAI + e and UIAI + e. There was a borderline association (P = 0.058) between having an HIV-positive partner rather than an HIV-unknown status partner and UIAI + e. Once again, subjects appeared to use biomedical factors (knowledge of their partner's HIV status, i.e. sero-sorting) to minimize transmission. Strategic positioning is consistent with the finding that untreated subjects were significantly more likely to report URAI + e, a behaviour that is perceived by many as a lower risk of transmission. Finally, currently having an STI was associated with URAI + e; this probably reflects recent sexual activity, as individuals were screened for STIs at the time when the questionnaires were administered, and did not know the diagnosis.
Treatment optimism had no significant association with UAI with regular partners, which contrasts with its association with UAI with casual partners in the same cohort 6 and other studies that have demonstrated an association with UAI with regular partners. 10,11 Only one cognitive variable was associated with UAI in our study: i.e. higher S-B scores were marginally associated with URAI + e (P = 0.09).
The methodology of this study may have contributed to the finding that biomedical rather than cognitive factors were associated with UAI with regular partners. Subjects' willingness to provide a semen sample may have selected for individuals who were already making sexual risk decisions with regular partners based on biomedical factors. Such individuals may have been keen to enrol in the study to see if their decisions regarding UAI were backed up by seminal viral load results. The methodology does not, however, account for our previous results where both cognitive and biomedical factors were associated with UAI with casual partners. 6
The study has several limitations including the cross-sectional design, the small sample size and the necessity to provide a semen sample that may reduce the generalizability of the findings as highly motivated patients, well-versed in risk reduction may have self-selected for enrolment. Moreover, because the questionnaires formed part of a larger study of viral loads, subjects' responses were not totally anonymous and this may have biased their reporting of UAI and disclosure of HIV-positive serostatus to partners. Notably, however, four subjects reported not having disclosed their HIV-positive status to their partners. Recall bias is always a potential problem in retrospective questionnaires such as ours.
CONCLUSION
Decisions about UAI with regular partners appear to have more association with biomedical factors rather than cognitive variables; i.e. HIV-positive men used more concrete evidence of reduced infectivity to support UAI activity with their regular partners than they did with their casual partners. 6 The findings suggest a difference in the factors used by HIV-1 positive MSM in sexual decision-making depending on the nature of the relationship. The study suggests that it may be prudent for clinicians to ensure that HIV-positive patients have access to biomedical information that may assist in their sexual decision-making with regular partners. Clinicians should continue to counsel about the risks of UAI and the person's individual and public health responsibilities around disclosure and safer sex. Further research on larger samples is needed to clarify these trends.
Footnotes
ACKNOWLEDGEMENTS
The authors would like to thank the staff and clients of the Albion Street Centre for their participation in this study, and Dr Tonia Rihs for her assistance with the study. Supported in part by a grant from the New South Wales Health Department as part of its health promotion programme (Ref: AA483). Preliminary findings were presented at the 4th International AIDS Society Conference, Sydney, 22–25 July 2007, Australia (Abstract CDC059).