The trial profile is shown in . A total of 7,274 men were screened of whom 1,151 (15.8%) were HIV-infected and eligible; of these, 922 (80.1%) consented and enrolled. Among 474 HIV-infected men randomized to the intervention arm, 374 (78.9%) were currently married or in a consensual union: 291 (77.8%) of their female partners consented and enrolled, of whom 122 (42.0%) were HIV-negative at the time of enrolment. Of these HIV-negative women, 93 (32.0%) enrolled concurrently with their husbands and 92 had at least one follow up visit over 24 months. Among 448 HIV-positive men randomized to the control arm, 348 (77.6%) were in a current marriage/consensual union; 239 (68.7%) of these female partners consented and enrolled, of whom 100 (42.0%) were HIV-negative at the time of enrolment. Of these women, 70 (29.3%) enrolled concurrently with their husband and 67 had at least one follow up visit over 24 months. The 92 intervention arm couples and 67 control arm couples with concurrent female and male enrolment and at least one female follow up visit constitute the primary population for determining male-to-female HIV transmission.
An additional 29 HIV-negative female partners of intervention arm men and 30 HIV-negative partners of control arm men entered the study six or more months after their husband's enrollment. These women were excluded from the primary male-to-female HIV transmission analysis since, unless women enrolled at the same time as their partner, we did not know their HIV status at the time of their husband’s enrolment. Thus, we could not determine which HIV-infected late-enrolling women had seroconverted since their husband’s enrolment, and this could thus result in bias if HIV transmission in the first six months differed by study arm. The couples with delayed female enrollment were assessed in secondary analyses.
shows the enrollment characteristics of HIV infected men and concurrently enrolled HIV-uninfected partners. There were no significant differences between arms in male characteristics or behaviors at enrollment. Female partners of men in both arms were comparable with respect to numbers of sexual partners in the past year, alcohol use and STI symptoms. However, intervention arm female partners were somewhat younger (p = 0.067) and less likely to report condom use in the past year (p = 0.017). At enrollment, 97.7% of intervention arm and 94.1% of control arm men had received their HIV results and post-test counseling. Among female partners, 68.8% in the intervention arm and 74.3% in the control arm accepted HIV results and post-test counseling at time of enrolment, and an additional 16.4% of intervention arm and 16.2% of control arm women reported they had previously received their results (i.e., 85.2% intervention and 90.5% control females had received HIV results). All participants received intensive HIV prevention education. Female retention rates were comparable in both arms at the 6, 12 and 24 month follow up visits. ()
Enrollment characteristics of HIV-infected men and HIV-negative women in couples enrolled concurrently.
Female Partner Retention Rates
shows the Kaplan-Meier cumulative probabilities of female HIV acquisition in couples with concurrent male and female enrolment. Over the 24 month follow up, the cumulative probability of female HIV acquisition was 21.7% (95%CI 12.7–33.4%) in the intervention arm and 13.4% (95%CI 6.7–25.8%) in the control arm (unadjusted HR=1.58, 95%CI: 0.68–3.66, p=0.287). After adjustment for differences in enrollment characteristics by Cox proportional hazards regression, the adjusted HR was 1.49, 95% CI 0.62–3.57, p = 0.368). When female partners who enrolled six months or more after their husband are included, the cumulative probability of infection was 17.4% in the intervention arm and 15.8% in the control arm (HR = 1.22, 95%CI 0.59–2.54, p = 0.65). There were no male crossovers among couples in the primary analysis. There were three crossovers among men whose female partner had delayed enrolment, but none of these crossover men transmitted to their partners.
Cumulative probability of Female HIV acquisition.
In a subanalysis (not specified in the protocol), we assessed whether HIV transmission in intervention arm couples was associated with the timing of resumption of intercourse relative to wound healing (). HIV acquisition, observed at six months, occurred in 27.8% (5/18) of women in intervention arm couples who resumed sex early, compared to 9.5% (6/63) of women in couples with delayed resumption of sex (RR = 2.92, 95%CI 1.01–8.46, p = 0.06). The proportion of women acquiring HIV by 6 months in intervention arm couples who delayed sex (9.5%) was comparable to the proportion of newly HIV-infected control arm women ( 7.9%[(5/63]; p = 1.0). However, the rate of female HIV acquisition at 6 months in intervention arm couples with early post-surgical resumption of sex (27.8%) was significantly higher than in control arm women (RR = 3.50. 1.14–10.76, p = 0.038).
Proportions of women with observed HIV acquisition at the 6 month follow-up visit in the control arm, and in the intervention arm by timing of resumption of intercourse in relation to post surgical wound healing.
There were no significant differences in HIV transmission between study arms by enrollment covariates, nor by female-reported sexual risk behaviors during follow up (data not shown). Among women whose partner’s enrollment viral load was <50,000 cps/mL, the cumulative probability of HIV acquisition was 15.7% (11/70) in the intervention arm and 10.6% (5/47) in the control arm (HR = 1.48, 95%CI 0.55–3.98, p = 0.43). Among couples with a male enrollment viral load > 50,000 cps/mL, female cumulative HIV acquisition was 27.3% (6/22) in the intervention arm and 15.0% (3/20) in the control arm (HR = 1.82, 95%CI 0.52–6.32, p = 0.34)
There were no statistically significant differences in female-reported number of sexual partners, condom use, or use of alcohol with sex during follow up (). In the intervention arm, 75.3% (70/93) of HIV-infected men disclosed their serostatus to their female partner; and in the control arm, 77.1% (54/70) of men disclosed their serostatus (p = 0.38).
Women’s sexual behaviors during follow up, by study arm
The proportions of follow up visits at which female partners reported STI symptoms or had laboratory diagnosed BV in the intervention and control arms, respectively, were: GUD, 16.4% (37/225) versus 16.1% (26/161), p=0.95; vaginal discharge, 36.4% (82/225) versus 32.3% (52/161), p=0.50; dysuria, 15.5% (35/225) versus 14.9% (24/161), p=0.89; and BV, 55.8% (121/217) versus 51.9% (83/160), p=0.54. Trichomonas was detected in 6.5% (9/138) of follow up visits in intervention arm women and 15.2% (17/112) of visits in control arm women (PRR = 0.43, 95%CI 0.18–1.02), which was of borderline statistical significance (p = 0.056).
Among 25 couples in which the female partner seroconverted during the trial, sequence data for both partners were available for 13 pairs. In all 13 couples, the genetic distance of the viral sequences between partners was < 0.5% , which was less than two standard deviations below the median distance of sequences between unrelated individuals in Rakai, indicating probable HIV acquisition within the partnership.(13
We assessed pre- and postoperative HIV VL in 89 ART naïve control arm participants receiving MC as a service. Among 80 men with detectable VL prior to surgery, the mean VL log10 cps/mL was 4.30 (SD 0.83) preoperatively, and 4.50 (SD 0.74) at the fourth postoperative week, a mean increase in intra-individual VL of 0.20 log10 cps/mL (p = 0.002). All 9 men with undetectable VL load prior to surgery remained undetectable at week four. In 25 control arm men who had initiated ART prior to circumcision, we observed no increase in VL in the 21 (84.0%) who had an undetectable VL prior to surgery, nor in the 4 men who had detectable preoperative VL.