1. Long term follow up of PROUD
Evidence for high continued HIV exposure and durable effectiveness of PrEP
E.White, D.Dunn, R.Gilson, A.Sullivan, A.Clarke, I.Reeves, G.Schembri, N.Mackie, C.Dewsnapp, C.Lacey, V.Apea, M.Brady, J.Fox, S.Taylor, J.Rooney, M.Gafos, N.Gill, S.McCormack,
and the PROUD study group
IAS, July 2017, Paris, France 1 1

2. PROUD (Nov12-Nov16)
HIV negative, G/MSM/TGW reporting condomless anal sex last/next 90days; 18+; clinic attendee; willing to take a pill every day
Enrolled 544
Risk reduction includes Truvada NOW Randomised 275
Risk reduction includes Truvada AFTER 12M Randomised 269
Follow 3 monthly
Endpoint: HIV infection in first 12 months
Impact on sexual behaviour and other STIs 2 2

3. Background PROUD reported early due to
unexpectedly high HIV incidence in the no-PrEP group (DEF)
very high effectiveness in the PrEP group (IMM)
Significantly more non-condom anal sex partners when the participant was receptive (bottom)
21% IMM with 10 or more compared to 12% DEF
No major differences in STIs between arms, except
trend towards lower rectal chlamydia in DEF compared to IMM

4. For this analysis Definitions
HIV infection if Ag/Ab negative at enrolment, then a reactive test confirmed by detection of virus
STI infection (yes or no, if yes the number) for syphilis, gonorrhoea or chlamydia by location since last visit
Stopped PrEP if last prescription was before 31May2016
Deferred follow-up time from enrolment to
HIV test reactive and infection subsequently confirmed
12 month visit when offered PrEP
Earlier visit after amended protocol when offered PrEP
Post-deferred follow-up time through to
Last HIV test in the study

5. PrEP starts over calendar time
Deferred phase
Post-deferred phase
Nov2012-Apr2014: enrolment
Nov2012-Oct2014: immediate or deferred access to PrEP – note early blips in the summer of 2013 due to the two ppts who enrolled twice
Nov2014-Oct2016: all participants had access to PrEP
Apr2014-Oct2016: planning close-out

6. PrEP prescriptions Number of participants IMM DEF Randomised 275 269
At least one PrEP prescription, N(%)
275 (100%)
206 (100%)
Prescription in last 6 months, N(%)
180 (65%)
146 (71%)
At least 80% days covered by PrEP*, N(%)
244 (93%)
183 (92%)
At least 60% days covered by PrEP*, N(%)
253 (97%)
194 (97%)
We can see that 63 (23%) participants didn’t return to receive PrEP in the DEF arm (21 had acquired HIV).
However, in both trial arms around 2/3s of the participants that initiated PrEP were continuing to collect prescriptions in the 6 months before the end of the study.
We can also see that during the time that participants were taking PrEP nearly all had at least 60% and 80% of days covered by PrEP.
*Up to last prescription 6 6

7. Participants continuing to collect PrEP prescriptions
Participants continuing to collect PrEP prescriptions*, by time since initiation
We also wanted to ascertain the time that participants remained on PrEP since this should be helpful in informing a PrEP program.
We ignored interruptions and assumed that anyone with their last prescription before June 2016 had stopped.
What we can see from this figure is that a high proportion are remaining on drug at 3 years after initiating. 85% at 1 year, 74% at 2 years and 63% at 3 years.**Look for exact proportions in log files
*Defined as any prescription since 31/05/2016 7 7

8. HIV incidence in the deferred and post-deferred phases
Deferred phase
Post-deferred phase
IMM
DEF
HIV infections (N) 4 21 5 1
Total person years (PY)
254
223
424
356
Incidence rates
(per 100PY)
1.6
9.4
1.2
0.3
Rate ratio
(90% CI)
6.0
(2.5 – 16.2)
0.2
( )
Rate difference
7.9
( )
-0.9
( )
I have no added the “exact” confidence intervals for the rate ratio but I would like to ask David what the most appropriate is for the rate difference. 8 8

9. HIV tests and PrEP prescriptions for the 10 who acquired HIV after starting
Here we illustrate the 10 cases in which the individuals had initiated PrEP and have subsequently seroconverted. We can see for each participants the HIV tests in addition to their prescriptions during the trial.
A few patterns emerge from this:
The first case was infected at baseline (he and his clinician agreed this was the most likely scenario)
Four were diagnosed after a period of lost to follow-up longer than 6 months (numbers 2, 4, 7 and 9).
Three continued to visit clinic but decided to discontinue PrEP for a number of reasons such as 3, 5 and 10.
In 2 cases (6 and 8) it is plausible that they were taking PrEP at the time of exposure.
Across all these different reasons the only seroconversion with high level resistance is the baseline case.
Two RTs missing –
1. Number 8 have very low viral load and went straight into a trial for individuals with primary HIV infection and onto trt – there are stored samples.
2. Number 10 was diagnosed elsewhere and we haven’t been able to obtain a resistance test for them (yet). 9 9

10. STI incidence in the deferred and post-deferred phases
Deferred Phase
Post-deferred Phase
Rate
(N/100 pyrs)
IMM
DEF
Rectal GC
35.3
(81/229)
33.0
(67/203)
31.4 (129/411)
32.6 (116/356)
Rectal CT
33.6
(77/229)
21.7
(44/203)
33.1 (136/411)
29.8 (106/356)
Syphilis
19.4
(46/237)
13.2
(28/212)
31.1
(132/424)
25.9
(93/359)
P=0.03 for IMM vs DEF rectal CT in deferred phase
P= 0.12 for DEF rectal CT between the two phases Ellen to check
NB incidence during the deferred phase even lower in those without a history of rectal CT in the year before enrolling 16.9, jumping up to 27.9 in the post-deferred phase.
Lacey et al BASHH June 2017 10 10

11. Clinic burden

12. Strengths and weaknesses
The end was difficult to plan without access to PrEP in the NHS
Study schedule for STIs followed standard of care and this changed from 6m to 3m in 2014
Participants on PrEP more likely to attend
Matching against national HIV infection dataset held by PHE identified 1 new infection
Able to use routine STI clinic data, at least in enrolling clinic

13. Conclusions
Three years after initiating PrEP, 60% of participants were still collecting drug (may over-estimate if stock-piling for the future)
The ongoing high rates of bacterial STIs in those attending confirmed that they needed PrEP
The reduction in HIV incidence was sustained and confirmed the very high adherence in this population and durable effectiveness of PrEP
The most likely reason for the HIV infections seen was the lack of PrEP at the time of exposure

14. Acknowledgements (1) PROUD Study participants MRC CTU at UCL
Sarah Banbury, Liz Brodnicki, Christina Chung, Yolanda Collaco-Moraes, Monica Desai,
David Dolling, David Dunn, Mitzy Gafos, Adam Gregory, Sajad Khan, Brendan Mauger, Sheena McCormack, Yinka Sowunmi, Ellen White, Gemma Wood
HIV & STI Dept, Public Health England
Monica Desai, Sarika Desai, Noel Gill, Anthony Nardone, GUMCAD team, HIV team
Clinics
Vanessa Apea (Barts Health NHS Trust), Christine Bowman & Claire Dewsnap (Sheffield
Teaching Hospitals NHS Foundation Trust), Michael Brady (Kings College Hospital NHS Foundation
Trust), Amanda Clarke & Martin Fisher (Claude Nichol Centre), Julie Fox (Guy’s and St Thomas’s
NHS Foundation Trust), Richard Gilson (The Mortimer Market Centre), Charles Lacey (York Hospitals
NHS Foundation Trust), Nicola Mackie (St Mary’s Hospital), Alan McOwan (56 Dean Street),
Iain Reeves (Homerton University Hospital NHS Foundation Trust), Gabriel Schembri (Manchester
Centre for Sexual Health), Ann Sullivan (John Hunter Clinic for Sexual Health), Steve Taylor, David White
(Heart of England NHS Foundation Trust)
Ask Liz to check

15. Acknowledgements (2) PROUD
Trial Steering Committee Independent members: Mike Adler (Co-Chair), Gus Cairns (Co-Chair), Dan Clutterbuck, Rob Cookson, Claire Foreman, Stephen Nicholson, Tariq Sadiq, Matthew Williams Investigator members: Brian Gazzard, Noel Gill, Anne Johnson, Sheena McCormack, Andrew Phillips Gilead: Matt Bosse, Rich Clarke, Sonia Gupta, Jim Rooney, Murad Ruf University of Liverpool: Saye Khoo Independent Data Monitoring Committee: Anton Pozniak, Simon Collins, Fiona Lampe Community Engagement Group Community: Yusef Azad (NAT), Gus Cairns (NAM), Rob Cookson (LGF), Tom Doyle (Mesmac), Justin Harbottle (THT), Marion Wadibia (NAZ), Matthew Hodson (GMFA), Cary James (THT), Roger Pebody (NAM) Clinics: Anthony Bains, Alan McOwan (Lead), MRC CTU at UCL: Sheena McCormack, Mitzy Gafos, Annabelle South Social Science Advisory Group Interviewers: Caroline Rae, Gill Bell, Michael Rayment, Sonali Wayal, Will Nutland, Mitzy Gafos Advisors: Ingrid Young, Ford Hickson, Lisa McDaid, Marsha Rosengarten, Nicolas Lorente, Agata Pacho, Elizabeth Poliquin, Anthony Nardone, Catherine Dodds, Adam Bourne, David Dolling, Sheena McCormack, Rob Horne
Ask Liz to check

16. Change in receptive anal intercourse without a condom
Number of partner categories 1
2-4
5-9
10-19
20+
Spare in case people ask about numbers of partners.
Caveat is the missing data which can be seen from the numbers at the bottom. There were big efforts to bring everyone back and offer PrEP at m12, but after that less effort went into retention due to lack of resources. Reasonable to assume that those that stayed on PrEP and kept coming back through m24 were at higher risk and having more partners than those that did not attend clinic, so this is likely to overestimate the proportions that increased their risk between m12-24 and baseline-m24.
N = 16 16

17. Exit questionnaire Has your condom use changed since taking PrEP?
No, I didn't generally use condoms before 98
32%
No, I use condoms the same as before 61
20%
Yes, I use condoms less
135
44%
Yes, I use condoms more 6 2%
Other 7
While taking PrEP do you ever use condoms? No 79
26%
Yes
226
74%
When my partners wants me to
165
When I'm worried about other STIs 84
When I'm concerned about HIV 39
When with partner other than regular partner 45
When I've forgotten to take some PrEP tablets 28 10
Are you more or less concerned about catching other STIs since taking PrEP?
More 89
29%
Less 24 8%
Not changed
187
61% 5
 2%
**Ellen to add number that completed this and check 17 17

18. Distribution of follow-up (yrs)
Median 2.6 years IQR

19. PHE matching
Of 32 known HIV infections in PROUD, 29 were in the national dataset held by PHE
2 diagnosed in PROUD in 2016 so likely reporting delay
1 diagnosed in PROUD in July 2014
One additional HIV infection (DEF) identified in PHE dataset and confirmed via clinic
Spare if needed to explain about the additional infections identified