1. Residual dapivirine ring levels indicate higher adherence to vaginal ring is associated with HIV-1 protection Elizabeth R. Brown, ScD Fred Hutchinson Cancer Research Center and University of Washington, Seattle, WA USA On behalf of the MTN-020/ASPIRE study team

2. MTN-020/ASPIRE MTN-020/ASPIRE was a multi-center, randomized, double-blind, placebo-controlled phase III trial of a vaginal matrix ring containing the non-nucleoside reverse transcriptase inhibitor dapivirine. 2629 women followed for 12-33 months 4280 person-years of follow-up were accrued 168 HIV seroconversions observed ( Placebo: 97; Active: 71) Primary analysis: intent-to-treat Effectiveness=27% (1, 46), p=0.046 Secondary analyses showed higher effectiveness in subgroups with higher adherence, including women ≥ 21 years of age

3. Schedule of sample collection Monthly follow-up w/ HIV-1 testing (min 12 / max 33 months) Monthly follow-up included HIV testing and provision of a new ring containing 25 mg of dapivirine

4. Schedule of sample collection Monthly follow-up w/ HIV-1 testing (min 12 / max 33 months) Monthly follow-up included HIV testing and provision of a new ring containing 25 mg of dapivirine Plasma was archived quarterly (for dapivirine and HIV RNA testing)

5. Schedule of sample collection Monthly follow-up w/ HIV-1 testing (min 12 / max 33 months) Monthly follow-up included HIV testing and provision of a new ring containing 25 mg of dapivirine Plasma was archived quarterly (for dapivirine and HIV RNA testing) Starting one year after study initiation, returned rings were archived and sent to a central lab for testing of remaining dapivirine levels (thus, nearly all participants have rings for visits at Month 12+ but some do not have prior to Month 12)

6. Distribution of residual ring levels

7. Cutpoints ≥23.5: non-adherent* *Well supported to represent distribution from unused rings. 23.5 mg cutpoint <23.5: mix of low to high adherence

8. Cutpoints ≥23.5: non-adherent* ≥22 mg: non/low adherence <22: mix of medium to high adherence *Well supported to represent distribution from unused rings. 23.5 mg cutpoint <23.5: mix of low to high adherence 22 mg cutpoint

9. Assessing relationship between adherence and HIV risk Risk reduction estimated from time-varying Cox model (1-HR) – Time-varying exposure: Adherence based on residual dapivirine levels in returned rings in previous 3 months – Outcome: Time to first detection of HIV Analyses limited to a subcohort including month 12 visits and beyond (when residual ring follow-up available for all women) Comparisons against placebo arm All analyses adjusted for site and age

10. Results Placebo arm Non- Adherent (≥23.5 mg) Low to high (<23.5mg) Person-years1089360741 % f-u (in arm) 100%33%67% Infections501314 Incidence (per 100 p-y) 4.63.61.9 Risk reduction (relative to placebo) --31% (-28, 63) p=0.24 56% (20, 76) p=0.007

11. Results Placebo arm Non- Adherent (≥23.5 mg) Low to high (<23.5mg) Medium to high (<22 mg) Person-years1089360741458 % f-u (in arm) 100%33%67%42% Infections5013147 Incidence (per 100 p-y) 4.63.61.91.5 Risk reduction (relative to placebo) --31% (-28, 63) p=0.24 56% (20, 76) p=0.007 65% (22,84) p=0.01

12. Length-adjusted analyses We also conducted analyses that took into account the amount of time a woman had the ring. Clinic visits scheduled every 28 days Visit windows +/- 2 weeks around that date, respected women’s schedules Drug released continuously while worn A woman who used the ring off-and-on for 6 weeks might look similar to one who used the ring consistently for 3 weeks. Analyses taking into account duration of use could give a more precise understanding of use These "length-adjusted" analyses consider adherence in four categories: no use and thirds of use (bottom, middle, and top)

13. Adherence vs. HIV protection: Ring data at time of detection Proportion of f/u 20%27% Incidence /100 p-y (# infections) 4.4 (9)2.3 (6)1.2 (3)2.3 (6) Risk reduction 49% (95% CI -19-78) Risk reduction 75% (95% CI 18-92) Risk reduction 47% (95% CI -25-77) Risk reduction 20% (95% CI -65-61) 100% 4.7 (47) Placebo No useBottom thirdTop third

14. HIV infection Monthly follow-up w/ HIV-1 testing (min 12 / max 33 months) HIV HIV- For women seroconverting to HIV, three approaches were taken to assign adherence at infection: The visit at which HIV was first detected

15. HIV infection Monthly follow-up w/ HIV-1 testing (min 12 / max 33 months) HIV HIV- For women seroconverting to HIV, three approaches were taken to assign adherence at infection: The visit at which HIV was first detected The visit with the lowest adherence for the two visits prior to HIV detection

16. HIV infection For women seroconverting to HIV, three approaches were taken to assign adherence at infection: The visit at which HIV was first detected The visit with the lowest adherence for the two visits prior to HIV detection The visit with the lowest adherence for the three visits prior to HIV detection Monthly follow-up w/ HIV-1 testing (min 12 / max 33 months) HIV HIV-

17. Adherence vs. HIV protection: Ring data two months prior to detection Proportion of f/u 20%27% Incidence /100 p-y (# infections) 4.9 (10)2.7 (7)1.6 (4)1.2 (3) Risk reduction 75% (95% CI 18-92) Risk reduction 66% (95% CI 6-88) Risk reduction 38% (95% CI -39-72) Risk reduction 11% (95% CI -78-55) 100% 4.7 (47) Placebo No useBottom thirdTop third

18. Adherence vs. HIV protection: Ring data three months prior to detection Proportion of f/u 20%27% Incidence /100 p-y (# infections) 4.9 (10)3.1 (8)1.9 (5)0.4 (1) Risk reduction 92% (95% CI 38-99) Risk reduction 58% (95% CI -7-83) Risk reduction 29% (95% CI -52- 66) Risk reduction 11% (95% CI -78-55) 100% 4.7 (47) Placebo No useBottom thirdTop third

19. Summary Across multiple analyses, there is a statistically significant relationship between ring use and HIV protection These analyses provide evidence suggesting a dose-response relationship between ring use and HIV acquisition Results suggest ring use is associated with at least 56% and potentially >75% protection when used consistently

20. Limitations Returned rings not tested at all visits for entire cohort – leading to this subcohort Unknown potential confounding – although we did adjust for two significant predictors of adherence and risk: age and site Misclassification remains due to manufacturing, assay and biological variability Does not identify a threshold for HIV protection

21. MTN-020/ASPIRE Study Team MTN-020/ASPIRE leadership: Jared M. Baeten (protocol chair), Thesla Palanee-Phillips (protocol co-chair), Elizabeth R. Brown (protocol statistician), Katie Schwartz (FHI 360 senior clinical research manager), Lydia E. Soto-Torres (DAIDS medical officer) Study sites: Malawi: Blantyre site (Malawi College of Medicine-John Hopkins University Research Project): Bonus Makanani, Taha E. Taha Malawi: Lilongwe site (University of North Carolina Project): Francis Martinson South Africa: Cape Town site (University of Cape Town): Linda-Gail Bekker South Africa: Durban eThekwini site (Centre for AIDS Programme of Research in South Africa): Gonasagrie Nair South Africa: Durban – Botha’s Hill, Chatsworth, Isipingo, Tongaat, Umkomaas, Verulam sites (South African Medical Research Council): Vaneshree Govender, Samantha Siva, Nitesha Jeenarain, Zakir Gaffoor, Arendevi Pather, Logashvari Naidoo, Gita Ramjee South Africa: Johannesburg site (Wits Reproductive Health and HIV Institute): Thesla Palanee-Phillips Uganda: Kampala site (Makerere University-Johns Hopkins University Research Collaboration): Flavia Matovu Kiweewa, Clemensia Nakabiito Zimbabwe: Chitungwiza-Seke South, Chitungwiza-Zengeza, Harare-Spilhaus sites (University of Zimbabwe-University of California San Francisco Collaborative Research Program): Nyaradzo M. Mgodi, Felix Mhlanga, Zvavahera M. Chirenje Microbicides Trials Network Leadership and Operations Center (University of Pittsburgh, Magee-Womens Research Institute, University of Washington, FHI 360, Population Council, RTI International): Sharon Hillier, Ian McGowan, Katherine Bunge, Beth Galaska, Cindy Jacobson, Judith Jones, Ashley Mayo, Barbara S. Mensch. Elizabeth T. Montgomery, Patrick Ndase, Rachel Scheckter, Devika Singh, Kristine Torjesen, Ariane van der Straten, Rhonda White Microbicides Trials Network Laboratory Center (Magee-Womens Research Institute, University of Pittsburgh, Johns Hopkins University): Craig W. Hendrix, Edward Livant, Mark A. Marzinke, John W. Mellors, Urvi M. Parikh Microbicides Trials Network Statistical and Data Management Center (Fred Hutchinson Cancer Research Center): Elizabeth R. Brrown, Jennifer Berthiaume, Marla Husnik, Karen Patterson, Barbra A. Richardson, Daniel W. Szydlo US National Institutes of Health: Nahida Chakhtoura, Donna Germuga, Cynthia I. Grossman, Lydia E. Soto-Torres International Partnership for Microbicides: Zeda Rosenberg, Annalene Nel MTN-020/ASPIRE participants and their communities; MTN-020 Community Working Group; MTN-020 Study Monitoring Committee; DAIDS MNDSMB The Microbicide Trials Network is funded by the National Institute of Allergy and Infectious Diseases (UM1AI068633, UM1AI068615, UM1AI106707), with co-funding from the Eunice Kennedy Shriver National Institute of Child Health and Human Development and the National Institute of Mental Health, all components of the U.S. National Institutes of Health. We are grateful to Dr. Roberta Black at NIAID for her oversight. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.