Secondary Logo

Journal Logo


Brief Report: Anal Intercourse, HIV-1 Risk, and Efficacy in a Trial of a Dapivirine Vaginal Ring for HIV-1 Prevention

Peebles, Kathryn MPHa; van der Straten, Ariane PhDb; Palanee-Phillips, Thesla MMed Sci, PhDc; Reddy, Krishnaveni MBChBc; Hillier, Sharon L. PhDd; Hendrix, Craig W. MDe; Harkoo, Ishana MBChBf; Gati Mirembe, Brenda MBChB, MScg; Jeenarain, Nitesha BPharmh; Baeten, Jared M. MD, PhDa,i,j; Brown, Elizabeth R. ScDk,l, on behalf of the MTN-020/ASPIRE Study Team

Author Information
JAIDS Journal of Acquired Immune Deficiency Syndromes: March 1, 2020 - Volume 83 - Issue 3 - p 197-201
doi: 10.1097/QAI.0000000000002253



Women in sub-Saharan Africa face disproportionately high HIV-1 risk, with particularly elevated risk among young women and adolescents.1 Receptive anal intercourse (RAI) has long been recognized as a potentially important HIV-1 transmission route among women.2 Approximately 5%–17% of women of the general-risk population report engaging in RAI in the prior 3 months in Eastern and Southern Africa,3–5 with each act conferring risk estimated to be 4- to 16-fold higher than an act of receptive vaginal intercourse (RVI).6–8 However, although RAI confers greater per-act risk, the product of site-specific per-act risk and relative proportion of RAI and RVI acts determines their respective contribution to the likelihood of HIV-1 acquisition.

Vaginal microbicides in development provide higher levels of drug to the lower genital tract than oral dosing of licensed products, with limited cross-compartmental transfer of drug to the rectum.9,10 Thus, RAI could hypothetically undermine the impact of vaginal microbicides,11,12 because vaginal products would not offer protection against HIV-1 transmission in acts of RAI. Two clinical trials of a dapivirine vaginal ring estimated HIV-1 risk reduction of 27%13 and 31%14 in intent-to-treat analyses, with evidence that product non-adherence contributed to a diluted estimate of efficacy13,15; however, it is unknown whether RAI also contributed to dilution of trial efficacy estimates.

An understanding of RAI practices and their potential impact on ring effectiveness is needed to inform ring implementation and HIV-1 risk among prospective ring users. We describe RAI behavior and correlates among the MTN-020/ASPIRE trial participants and investigate the association between RAI and HIV-1 risk and whether RAI may have contributed to efficacy dilution of the ring.


We conducted a secondary analysis of data from the ASPIRE clinical trial of a dapivirine vaginal ring for HIV-1 prevention; detailed trial methods have been published previously.13 Briefly, 2629 healthy, sexually active, HIV-1-negative women living in Malawi, South Africa, Uganda, and Zimbabwe enrolled in the trial from 2012 to 2014. Enrolled women were randomized 1:1 to receive either an intravaginal ring containing 25 mg of dapivirine or a placebo intravaginal ring that was changed each month. Women were instructed to use the ring continuously over a 28-day period, with monthly follow-up visits for HIV-1 serologic testing and dispensation of new rings.

At enrollment, participants reported vaginal and anal sex act frequency over the prior 3 months and condom use at their most recent sex act in standardized face-to-face interviews. RAI was assessed with the question, “In the past 3 months, how many times have you had anal sex? By anal sex, we mean when a man puts his penis inside your anus.” This item was translated and back-translated to ensure that its meaning was well understood by participants, following lessons learned in the VOICE trial in which some participants interpreted RAI items as referring to vaginal sex from behind.16 To encourage accurate reporting of sensitive behaviors, all women also completed audio computer-assisted self-interviews (ACASI) at baseline and the month 3 follow-up visit, which included items that assessed RAI frequency and condom use, alcohol consumption in the prior 3 months, and transactional sex in the prior year. Both baseline demographic data and follow-up RVI frequency (over the prior 7 days) were collected in face-to-face surveys only. At baseline, women were evaluated for syphilis, Chlamydia trachomatis, Neisseria gonorrhoeae, Trichomonas vaginalis, and bacterial vaginosis (defined as Nugent score >617).

Analyses here include 2562 (98.0%) participants who responded to RAI survey items at baseline and/or month 3. RAI exposure was classified according to ACASI-reported behavior in all analyses. We assessed the association between baseline participant characteristics and reporting any RAI in the previous 3 months using Cochran–Mantel–Haenszel χ2 tests stratified by study site for categorical variables and a t test from linear models adjusted for site for continuous variables. We estimated the proportion of all sex acts that were RAI using enrollment reports of the number of RAI and RVI acts over the prior 3 months. We used a Cox proportional hazards model stratified by study site and with an interaction term for RAI and study arm to estimate the association between any RAI and HIV-1 acquisition and to evaluate potential ring effect measure modification associated with RAI. We additionally estimated the association of any RAI and HIV-1 risk within study arm, and ring efficacy within groups of women defined by reporting any versus no RAI. Models were adjusted for variables identified as potential confounders (baseline report of transactional sex, education level, and use of hormonal injectable contraception). In the primary analysis, we defined RAI exposure as self-reported RAI at baseline and/or month 3. In sensitivity analyses, we characterized RAI exposure as a continuous variable of the total number of RAI acts reported at baseline and month 3. In addition, we conducted sensitivity analyses excluding subpopulations with enrollment characteristics predictive of low adherence (2 study sites and women younger than 25)13 to increase the sensitivity of the analysis to detect an interaction effect between RAI and ring efficacy. All analyses were conducted in R

Institutional review boards from each study site approved the study protocol. All participants provided written informed consent.


In interviewer-administered surveys at study enrollment, 54 (2.1%) women reported engaging in RAI in the previous 3 months, whereas 334 (13.1%) women reported RAI in the previous 3 months via ACASI. A total of 468 (18.3%) women reported RAI at baseline and/or month 3, with 136 (5.3%) women reporting RAI at both time points. Among women reporting RAI at baseline, 14.3% of their total sex acts over 3 months were RAI; among all participants, 1.5% of all sex acts reported at baseline were RAI (see Table S1, Supplemental Digital Content, Combining baseline and month 3, approximately 3-quarters of women who engaged in RAI reported 3 or fewer RAI acts in the previous 3 months. Self-reported condom use was lower for vaginal than anal sex acts (57.4% and 62.9% condom use at the most recent vaginal and anal sex act at baseline, respectively, P < 0.001). Baseline RAI prevalence was highest among women who reported engaging in transactional sex (22.6%; Table 1). RAI was also more common among women with less than secondary education and women who used hormonal injection contraceptive methods (Table 1). RAI was not associated with baseline vaginal STI infection nor baseline bacterial vaginosis (Table 1).

Association of Baseline Participant Characteristics and Engaging in Receptive Any Anal Intercourse in the Previous Three Months. Values are n (%) Except Where Noted

In models adjusted for randomization arm, RAI at baseline and/or month 3 was not associated with HIV-1 risk (aHR: 0.93, 95% CI: 0.57 to 1.54; P = 0.71); this association did not differ by study arm (P-value for interaction = 0.83). Within the placebo arm, HIV-1 incidence among women who did not engage in RAI was lower, at 4.3 per 100 woman-years (95% CI: 3.4 to 5.3; 70 events/1642 woman-years), relative to 5.2 per 100 woman-years (95% CI: 3.4 to 7.7; 21 events/401 woman-years) among women who reported any acts of RAI. Differences in incidence rates were primarily driven by confounding by study site; after stratification on site and adjustment for potential confounders, incidence was not significantly different between groups (aHR: 0.97, 95% CI: 0.58 to 1.62, P = 0.87) (Fig. 1A). Similarly, in the dapivirine arm, incidence was 3.1 per 100 woman-years (95% CI: 2.4 to 4.0; 52 events/1663 woman-years) among those who engaged in only vaginal sex and 4.2 (95% CI: 2.6 to 6.5; 16 events/381 woman-years) among those who also engaged in any acts of RAI (aHR: 0.94, 95% CI: 0.52 to 1.68, P = 0.80) (Fig. 1B). In a sensitivity analysis in which RAI exposure was modeled continuously as the number of total RAI acts reported at baseline and month 3, adjusted hazard ratios were similar to those in the primary analysis, and did not differ by study arm (P-value for interaction = 0.15).

HIV-1 incidence rates and Kaplan–Meier curves of cumulative HIV-1 incidence, stratified by (A) any RAI reported at baseline or month 3 among those assigned to placebo ring; (B) any RAI reported at baseline or month 3 among those assigned to dapivirine ring; (C) treatment assignment among those who reported any RAI at baseline or month 3; and (D) treatment assignment among those who reported no RAI at baseline nor month 3. Kaplan–Meier curves in (A), (B) are adjusted for site.

Among all women, the dapivirine ring reduced the risk of HIV-1 acquisition by 25% (95% CI: −3 to 55). In comparison, among women reporting exclusively RVI at baseline and month 3, risk reduction was 27% (95% CI: −5 to 49; Fig. 1D) after stratification for study site, whereas among women who reported any RAI, the reduction in risk of HIV-1 acquisition was attenuated at 18% (95% CI: −57 to 57; Fig. 1C) (P-value for interaction = 0.77). In sensitivity analyses in which groups with low adherence were excluded, the interaction effect between RAI and the dapivirine ring remained statistically nonsignificant (analysis excluding 2 study sites: P = 0.63; analysis excluding women younger than 25 years of age: P = 0.74; see Table S2, Supplemental Digital Content,


In the ASPIRE trial of a dapivirine vaginal ring for HIV-1 prevention, RAI was not associated with HIV-1 acquisition, nor was the protective effect of the ring significantly reduced among women who reported any RAI. Approximately one in 5 women reported RAI at baseline and/or month 3, although notably, a much smaller proportion (approximately one in 20) reported RAI at both time points, suggesting that RAI is an occasional rather than consistent sexual practice among this cohort of African women. Nearly 3-quarters of the women reporting any RAI reported an average of one RAI act per month or less. Because the vast majority of sex acts among trial participants were reported to be RVI, the total risk contributed by RAI may have been sufficiently small as to not result in a sizeable nor statistically significant association with HIV-1 risk, nor significant effect modification of ring efficacy. Although our primary analysis assessed effect modification by randomization arm, results were robust to sensitivity analyses in which we excluded groups of women with relatively lower adherence. Our study also collected RAI data at 2 time points only, limiting the precision of our estimates of the association of RAI with HIV-1 risk and ring efficacy. Nonetheless, among women reporting only vaginal sex at baseline and month 3, we estimated that ring efficacy was 27%, only modestly higher than the overall efficacy of 25% estimated among all women included in these analyses. This suggests that any efficacy dilution in the ASPIRE trial because of RAI may be minimal, supporting the hypothesis that most efficacy dilution is attributable to ring non-adherence.15

Mathematical modeling may provide greater clarity about the relative contribution of RAI to reduced ring efficacy, and the impact that such efficacy dilution would have on HIV-1 prevention. The frequency of RAI in the ASPIRE cohort was lower than has been reported in other cohorts3; the potential for RAI to undermine the effectiveness of the ring at the population level may be higher in other settings where RAI is more common. In addition, despite use of ACASI to collect RAI behaviors, RAI remains a taboo behavior19 and may be underreported in our cohort. Should the ring become widely available in the future, geographically-specific representative estimates of RAI prevalence and frequency among women at risk of HIV-1 acquisition are needed to inform ring implementation strategies, including the need for counselling on prevention options, and to evaluate the potential for RAI to reduce the population-level impact of the ring.

Our study adds to a limited body of research on RAI in sub-Saharan African settings. A single study in South African community and clinic settings observed higher RAI prevalence among younger and unmarried individuals and those with a history of sexually transmitted infections or transactional sex.20 Similarly, RAI was associated with several risk factors for HIV-1 acquisition in this cohort, including transactional sex, suggesting potential populations for targeted communication regarding RAI-associated risk, and among prospective ring users, the lack of protective effect of the ring against RAI. For such women, the availability of alternative methods that protect against HIV-1 acquisition in all compartments, such as oral PrEP or condoms, will be particularly important. Although previous research in South Africa reported similar condom use rates for both vaginal and anal sex acts,3 ASPIRE participants reported higher rates of condom use during acts of RAI than acts of RVI, potentially mitigating RAI-associated risk in this population and contributing to the observed lack of association between RAI and HIV-1 acquisition. These higher rates of condom use also suggest that women are appropriately evaluating the higher risk associated with RAI, in contrast to previous qualitative research in which women reported lower perceived HIV-1 risk from RAI relative to RVI.21 Similar to previous research, we observed strikingly different self-reports of RAI behaviors in face-to-face interviews, in which only about 1 in 50 women reported any RAI, relative to ACASI,3 emphasizing the need to offer confidential data collection methods to ensure more accurate reporting of sensitive behaviors in future research in this area.

In summary, RAI was not a prominent risk factor for HIV-1 nor was the protective effect of the ring significantly lower among women who engaged in RAI. These findings suggest that, in populations with similar prevalence and frequency of RAI as in this cohort, RAI may not appreciably reduce the population-level impact of the ring, thus supporting use of the ring for HIV-1 prevention in women.


The authors express their many thanks to the women who contributed to this work through their dedicated participation in MTN-020/ASPIRE. Special thanks to Lydia Soto-Torres, MD, MPH for her insightful feedback on drafts of this manuscript. The International Partnership for Microbicides developed the dapivirine ring and supplied rings for this trial.

MTN-020/ASPIRE Study Team: Study Team Leadership: Jared Baeten, University of Washington (Protocol Chair); Thesla Palanee-Phillips, Wits Reproductive Health and HIV Institute (Protocol Co-chair); Elizabeth Brown, Fred Hutchinson Cancer Research Center (Protocol Statistician); Lydia Soto-Torres, US National Institute of Allergy and Infectious Diseases (Medical Officer); Katie Schwartz, FHI 360 (Clinical Research Manager). Study sites and site Investigators of Record: Malawi, Blantyre site (Johns Hopkins University, Queen Elizabeth Hospital): Bonus Makanani; Malawi, Lilongwe site (University of North Carolina, Chapel Hill): Francis Martinson; South Africa, Cape Town site (University of Cape Town): Linda-Gail Bekker; South Africa, Durban—Botha's Hill, Chatsworth, Isipingo, Tongaat, Umkomaas, Verulam sites (South African Medical Research Council): Vaneshree Govender; Samantha Siva, Zakir Gaffoor, Logashvari Naidoo, Arendevi Pather, and Nitesha Jeenarain; South Africa, Durban, eThekwini site (Center for the AIDS Programme for Research in South Africa): Gonasagrie Nair; South Africa, Johannesburg site (Wits RHI): Thesla Palanee-Phillips; Uganda, Kampala site (John Hopkins University, Makerere University): Flavia Matovu; Zimbabwe, Chitungwiza, Seke South and Zengeza sites (University of Zimbabwe, University of California San Francisco): Nyaradzo Mgodi; Zimbabwe, Harare, Spilhaus site (University of Zimbabwe, University of California San Francisco): Felix Mhlanga.

Data management was provided by The Statistical Center for HIV/AIDS Research & Prevention (Fred Hutchinson Cancer Research Center, Seattle, WA) and site laboratory oversight was provided by the Microbicide Trials Network Laboratory Center (Pittsburgh, PA).


1. UNAIDS. The Gap Report. Geneva, Switzerland: UNAIDS; 2016.
2. Voeller B. AIDS and heterosexual anal intercourse. Arch Sex Behav. 1991;20:233–276.
3. Owen BN, Elmes J, Silhol R, et al. How common and frequent is heterosexual anal intercourse among South Africans? A systematic review and meta-analysis. J Int AIDS Soc. 2017;20:21162.
4. Gorbach PM, Mensch BS, Husnik M, et al. Effect of computer-assisted interviewing on self-reported sexual behavior data in a microbicide clinical trial. AIDS Behav. 2013;17:790–800.
5. Marrazzo JM, Ramjee G, Richardson BA, et al. Tenofovir-based preexposure prophylaxis for HIV infection among African women. New Engl J Med. 2015;372:509–518.
6. Patel P, Borkowf CB, Brooks JT, et al. Estimating per-act HIV transmission risk: a systematic review. AIDS. 2014;28:1509–1519.
7. Boily MC, Baggaley RF, Wang L, et al. Heterosexual risk of HIV-1 infection per sexual act: systematic review and meta-analysis of observational studies. Lancet Infect Dis. 2009;9:118–129.
8. Baggaley RF, Owen BN, Silhol R, et al. Does per-act HIV-1 transmission risk through anal sex vary by gender? An updated systematic review and meta-analysis. Am J Reprod Immunol. 2018;80:e13039.
9. Murphy DJ, Desjardins D, Dereuddre-Bosquet N, et al. Pre-clinical development of a combination microbicide vaginal ring containing dapivirine and darunavir. J Antimicrob Chemother. 2014;69:2477–2488.
10. Justman JE, Nair GL, Hendrix CW, et al. Pharmacokinetics and pharmacodynamics of tenofovir reduced-glycerin 1% gel in the rectal and vaginal compartments in women: a cross-compartmental study with directly observed dosing. J Acquir Immune Defic Syndr. 2018;78:175–182.
11. Masse BR, Boily MC, Dimitrov D, et al. Efficacy dilution in randomized placebo-controlled vaginal microbicide trials. Emerg Themes Epidemiol. 2009;6:5.
12. McGowan I, Taylor DJ. Heterosexual anal intercourse has the potential to cause a significant loss of power in vaginal microbicide effectiveness studies. Sex Transm Dis. 2010;37:361–364.
13. Baeten JM, Palanee-Phillips T, Brown ER, et al. Use of a vaginal ring containing dapivirine for HIV-1 prevention in women. New Engl J Med. 2016;375:2121–2132.
14. Nel A, van Niekerk N, Kapiga S, et al. Safety and efficacy of a dapivirine vaginal ring for HIV prevention in women. New Engl J Med. 2016;375:2133–2143.
15. Brown E, Palanee-Phillips T, Marzinke M, et al. Residual dapivirine ring levels indicate higher adherence to vaginal ring is associated with HIV-1 prevention. 2016. Paper presented at: AIDS 2016; 19 July 2016; Durban, South Africa.
16. Duby Z, Hartmann M, Mahaka I, et al. Lost in translation: language, terminology, and understanding of penile-anal intercourse in an HIV prevention trial in South Africa, Uganda, and Zimbabwe. J Sex Res. 2016;53:1096–1106.
17. Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J Clin Microbiol. 1991;29:297–301.
18. R: A Language and Environment for Statistical Computing [computer Program]. Vienna, Austria: R Foundation for Statistical Computing; 2008.
19. Duby Z, Colvin C. Conceptualizations of heterosexual anal sex and HIV risk in five east African communities. J Sex Res. 2014;51:863–873.
20. Kalichman SC, Simbayi LC, Cain D, et al. Heterosexual anal intercourse among community and clinical settings in Cape Town, South Africa. Sex Transm Infect. 2009;85:411–415.
21. Duby Z, Hartmann M, Montgomery ET, et al. Condoms, lubricants and rectal cleansing: practices associated with heterosexual penile-anal intercourse Amongst participants in an HIV prevention trial in South Africa, Uganda and Zimbabwe. AIDS Behav. 2016;20:754–762.

dapivirine; vaginal ring; HIV-1 prevention; anal intercourse

Supplemental Digital Content

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.