HPTN 067 ADAPT: A Phase 2 randomized open label trial of daily, twice weekly, and sex event PrEP dosing. Presented by Robert Grant, MD, MPH on behalf of the HPTN 067/ADAPT Study Team

Background Oral FTC/TDF PrEP is effective for preventing HIV acquisition.1 Full protection after rectal exposure with use of 4+ tabs/week.2 Full protection after vaginal exposure likely requires more PrEP use.3 Sex is often planned, and plans change over time.4 PrEP provides benefit when used during seasons of risk.5 Such strategic PrEP use has been observed in MSM.2 Measurement of adherence is challenging, especially when dynamic.6 Recommending PrEP dosing before and after sex was effective among MSM.7 Study Premise: Adapting PrEP regimens to match patterns of sex could increase strategic PrEP use and minimize medication costs and side effects. Showing our study in context of released of results from other PrEP studies. Cape Town enrollment began in September 2011, follow up ended in June 2013. Grant NEJM 2010, Baeten NEJM 2012, Thigpen NEJM 2012, Choopanya Lancet 2013; Grant Lancet Infec. Dis. 2014, Liu JAMA Int. Med. 2015; 3. Grant AIDS 2015, Cottrell JID 2016; 4. van Griensven JIAS 2010, 5. Hojilla AIDS and Behavior 2015, Grant Lancet 2016 6. Mutua PLoS One 2012, Kibengo PLoS One 2013; 7. Molina NEJM 2015.

self-administered dosing HPTN 067 Design FTC/TDF Randomized D T E Daily- One tablet/day Time driven- 1 tablet/2x week with a post sex boost Event driven- 1 tablet pre-sex and 1 tablet post-sex No more than 2 tablets daily or 7 tablets/week Final Study Visit Week 34 6 weeks DOT period 24 weeks self-administered dosing 4 weeks off drug Women (incl. TGW) & MSM After 6 weeks of DOT to estimate steady state drug levels, participants were randomly assigned to one of three unblinded PrEP dosing regimens for 24 weeks of self-administered dosing as follows: Daily (D) Time Driven: Twice weekly with a post-intercourse boost (T) Event-driven: Before and after intercourse (E) Pills were dispensed from an electronic dispensing Wisepill device that recorded each opening Participants were contacted weekly by phone or in person to review Wisepill data and sex events Emphasize that collection of reported pill taking and sex events was not combined in the interview to reduce social desirability bias. Final study visit at 34 weeks, 4 weeks after ending self-administered dosing Key informant interviews and focus groups PrEP use and sex

Silom Community Clinic 178 HIV-uninfected at risk MSM/TGW Bangkok, Thailand Completed March 2014 Site PI: Tim Holtz Harlem Prevention Center 179 HIV-uninfected at risk MSM/TGW NYC (Harlem), USA Completed Dec 2014 Site PI: Sharon Mannheimer Emavundleni 178 HIV-uninfected at risk WSM Cape Town, South Africa Completed June 2013 Site PI: Linda Gail Bekker

Definition of Primary Outcome: Coverage Coverage of sex events for all arms: >1 pill taken in the 4 days before sex >1 pill taken in the 24 hours after sex Coverage Definition is the same for all 3 arms. Only vaginal and anal sex acts will be considered; oral sex acts will not be included Periods of product hold (prior to acquisition of HIV) are included The date and time of the sex act will be compared to the data and time of pill use as reported on item 2 of the Weekly Interview Log. A sex act will be considered “covered” if the following two conditions are met:   i) At least one pill is taken during the 96 hours before the act ii) A pill is taken within 24 hours after the act Note that the same pill can cover a post-exposure dose for one event and a pre-exposure dose for a different event. >1 tablet >1 tablet Sex Bekker IAS2015, Vancouver, 2015

Coverage of Sex Events – MSM/TGW in Bangkok As Bob described, the primary outcomes in this trial were coverage of sex events, number of tablets required and taken, and self-reported side effects. At our site, PrEP coverages were similar in arms D and T, at 85% vs 84%, with no statistically significant difference between these arms, with a p=0.79. Both coverages were statistically significantly greater than in arm E at 74%, which is still arguably quite high for a non-daily regimen. In this arm E, the missing post-sex dose was largely responsible for the lower coverage seen. The global p value for the 3 arms was 0.19, reflecting the nearly identical coverage of sex events between the daily arm, and the time-driven arm. Maoji updated 2016/06/07: Coverage (yes vs. no) Time vs Daily: 0.94( 0.58, 1.51), P-value=0.7942 Event vs Daily: 0.48( 0.26, 0.89), P-value=0.0197 Time vs Event: 1.95( 1.04, 3.65), P-value=0.0372 Overall P-value=0.1850 Daily/Time p = 0.79, Daily/Event p = 0.02, Time/Event p = 0.04, global p = 0.19 Holtz, IAS 2015, Vancouver

Tenofovir diphosphate in PBMCs: % with TFVDP >= 5. 2 fmol/106 cells Tenofovir diphosphate in PBMCs: % with TFVDP >= 5.2 fmol/106 cells* Bangkok MSM/TGW Participants who report sex in last 7 days with detectable TFVDP in PBMC (>=5.2 fmol/10^6 cells) Daily (D) Time-driven (T) Event-driven (E) Week 10 31/31 (100%) 29/29 (100%) 30/30 (100%) Week 18 28/29 (96.6%) 24/26 (92.3%) Week 30 22/23 (95.7%) 18/19 (94.7%) 13/14 (92.9%) Maoji updated 2016/06/07: Time vs Daily: 1.90 (0.18, 20.4), P-value=0.5953 Event vs Daily: 0.55 (0.09, 3.26), P-value=0.5113 Time vs Event: 3.45 (0.37, 32.4), P-value=0.2791 Overall P-value=0.5213 *Indicative of at least 2 tablets per week. Time/Daily p = 0.60, Event/Daily p = 0.51, Time/Event p=0.28

FTC/TDF Pills by Arm MSM/TGW in BKK HPTN 067/ADAPT As in the other sites, the number of tablets required, shown in blue, was calculated by the number required by each dosing regimen over study follow-up based on the number of reported sex acts, or daily for the D arm. The number of required tablets differed significantly by arms - daily required the most, E required the fewest tablets. The number of tablets actually taken (shown in red) were significantly lower in the T and E arms compared to daily. Sometimes participants took more tablets than required. The E arm participants were told to take a tablet if they thought they might have sex. Required tablets: p < 0.001 for all comparisons (D/T, D/E, and T/E) Tablets actually taken: p < 0.001 for all comparisons (D/T, D/E, and T/E)

Neuro and GI Symptoms / Side Effects MSM/TGW in Bangkok HPTN 067/ADAPT Side Effect reported on Structured Interview Daily Time Event p value % PPTs who experienced any neurologic side effects 48% 46% 54% 0.64 % PPTs who experienced any GI side effects 45% 34% 41% 0.46 The neurologic and GI side effect proportions shown here are EVER occurring during the follow up period. Reported side effects were not uncommon, but were generally mild, and not substantially different in the daily and non-daily arms. CHANGE: Self reported GI: (such as nausea, vomiting, diarrhea, gas, bloating and abdominal pain) Self reported Neuro (such as headache, dizziness, lightheadedness) Jim – which p value?

Coverage of Sex Events – Women in Cape Town Maoji updated 2016/06/07: Coverage (yes vs. no) Time vs Daily: 0.43( 0.26, 0.70), P-value =0.0007 Event vs Daily: 0.36( 0.22, 0.60), P-value<.0001 Time vs Event: 1.18( 0.78, 1.77), P-value=0.4361 Overall P-value=0.0006 Sex event defined as vaginal or anal intercourse Time/Daily p = 0.0007, Event/Daily p < 0.0001, Time/Event p = 0.43 Bekker IAS2015, Vancouver, 2015

TFVDF in PBMCs: % with TFVDP >= 5. 2 fmol/106 cells PBMC TFVDF in PBMCs: % with TFVDP >= 5.2 fmol/106 cells PBMC* - Cape Town Participants who report sex in last 7 days with detectable TFV-DP in PBMC (>=5.2 fmol/10^6 cells) Daily (D) Time-driven (T) Event-driven (E) Week 10 33/40 (82.5%) 16/23 (69.6%) 25/37 (67.6%) Week 18 29/39 (74.4%) 16/25 (64.0%) 10/30 (33.3%) Week 30 19/29 (65.5%) 13/24 (54.2%) 12/31 (38.7%) Maoji updated 2016/06/07: Modeling detectable (>=5.2 fmol/10^6 cells) Time vs Daily: 0.56 (0.25, 1.25), P-value=0.1568 Event vs Daily: 0.31 (0.15, 0.64), P-value=0.0017 Time vs Event: 1.81 (0.84, 3.90), P-value=0.1307 Overall P-value=0.0080 *Indicative of at least 2 tablets per week. Time/Daily p = 0.16, Event/Daily p = 0.002, Time/Event p=0.13

Coverage of Sex Events – MSM/TGW in NYC Maoji updated 2016/06/06: Coverage (yes vs. no) Time vs Daily: 0.45( 0.24, 0.85), P-value=0.0137 Event vs Daily: 0.55( 0.35, 0.89), P-value=0.0146 Time vs Event: 0.80( 0.45, 1.45), P-value=0.4682 Overall P-value=0.0265 Time/Daily and Event/Daily p = 0.01; Time/Event p =0.47 Mannheimer, IAS 2015, Vancouver

TFVDF in DBS: % with TFVDP >= 326 FMOLE/punch DBS* - Harlem Participants who report sex in last 7 days with detectable TFV-DP in DBS (>=326 fmole/punch) Daily (D) Time-driven (T) Event-driven (E) Week 10 13/23 (56.5%) 8/23 (34.8%) 5/27 (18.5%) Week 18 11/27 (40.7%) 10/27 (37.0%) 3/21 (14.3%) Week 30 9/18 (50.0%) 3/18 (16.7%) Maoji updated 2016/06/07: Time vs Daily: 0.47(0.19, 1.19), P-value=0.11 Event vs Daily: 0.21 (0.07, 0.61), P-value=0.004 Time vs Event: 2.23 (0.79, 6.35), P-value=0.13 Overall P-value=0.0142 *Indicative of at least 2 tablets per week. Time/Daily p = 0.11, Event/Daily p = 0.004, Time/Event p=0.13

Barriers and Facilititators to Event Driven Dosing: Thai MSM Facilitators: Barriers: Preference for fewer doses (less concerns about side effects) Complicated Regimen eg: No more than 1 dose in a 2 hour window Ability to plan for sex Inability to forecast sex Having to carry all time Have control over planning for sex with sexual partners Have infrequent sex Chemnasiri IAS 2015, Vancouver, BC

Barriers and Facilititators to event driven dosing: Thai MSM Chemnasiri IAS 2015, Vancouver, BC

HPTN 067 Summary of Plasma Drug Detection in participants reporting sex in the past 7 days

Limitations HPTN ADAPT event-driven dosing was complicated by instructions that … 2 Tablets were not to be taken at a time, There was 2 hour target window for post-sex dosing. Participants were informed that daily dosing was proven to be effective, and that non-daily dosing was unproven. This information likely undermined motivation to use non-daily regimens.

Conclusions Adherence to oral PrEP is feasible in diverse groups. A recommendation for daily PrEP dosing led to… Highest coverage, Highest adherence, Highest PrEP drug concentrations, Higher pill burden, No difference in side effects. Non-daily dosing in Thai MSM led to… Comparable PrEP coverage if time-driven dosing, Slightly less coverage if event-driven dosing.

HPTN 067 Participants and the men and women they represent The HPTN 067/ADAPT Study Team acknowledges key support and contributions of the following: HPTN 067 Participants and the men and women they represent Team members at Emavundleni, Silom Clinic and the Harlem Prevention Center. HPTN 067 Protocol Team (including those at LC, LOC, SDMC, PAB and DAIDS)