Biomedical HIV Prevention Trials A Sisyphean Task

1
Biomedical HIV Prevention Trials(A Sisyphean
Task)

• Ronald Gray
• Johns Hopkins University
• Bloomberg School of Public Health

2
Objectives

• Review trials of
• Bacterial and viral STD control
• Microbicides
• PrEP
• Vaccines
• Male circumcision

3
Approach

• Review observational/biologic evidence
• Summarize trials using forest plots
  http//www.cc-ims.net/revman
• Issues with trials and interpretation of findings
• Irritating comments

4
Overview

• 29 completed biomedical prevention trials with an
  HIV endpoint
• 4 showed significant efficacy in
  intention-to-treat (ITT) analyses
• 3 were male circumcision trials
• 5 trials suggested a possibility of harm
• We need to re-evaluate strategies for prevention
  and for trials

5
Control of STIs for HIV Prevention

• STIs (HSV-2) and STI symptoms (GUD), are
  associated with increased risk of HIV acquisition
  (epidemiologic synergy)
• How much is causal and how much is confounding?
• Observational studies cannot disaggregate
  biologic STI cofactor effects from confounding

6
Behavioral confounding
Biologic cofactor effect
HIV acquisition
STIs
Sexual risk behaviors Co-infected partners
Differential Infectivity and temporal sequence
HIV Less infectious
STIs More infectious
Time
7
8 Trials of STI Control for HIV Prevention

• Control of Curable STIs
• Syndromic management or presumptive therapy
• 5 community randomized trials
• Grosskurth Lancet 2005, Wawer Lancet 1999, Gray
  Am J Ob Gynecol 2001, Kamali Lancet 2003, Gregson
  PLos 2007
• 1 individually randomized trial
• Kaul JAMA 2004
• HSV-2 suppression in HIV-negative participants
• 2 individually randomized trials of acyclovir
• Watson Jones NEJM 2007, Celum Lancet 2008

8
Trials of STI Control for Prevention of HIV
Acquisition
7 negative trials One RCT showed efficacy in an
atypically low HIV incidence/prevalence setting
(Mwanza)
9
Why were STI Control RCTs largely negative?(Gray
and Wawer, Lancet 20083712064)

• Population Attributable fraction of HIV due to
  STIs
• STIs play a modest role in HIV acquisition at a
  population level?
• Trials were not powered to detect modest effects
• Control of STIs is difficult no trial controlled
  all STIs
• Curable STI treatment insufficient to reduce
  population prevalence?
• HSV-2 suppression with acyclovir insufficient to
  reduce genital inflammation and GUD?
• If we cannot control STIs in trials, how can we
  do so in programs?

10
HSV-2 Suppression in HIV co-infected persons to
prevent transmission

• 4 RCTs with Intermediate end points
• HIV shedding, genital and plasma viral load
• Ouedraogo AIDS 2006, Zuckerman JID 2007,NaGOT
  nejm 2007, Baeten JID 2008, Zuckerman AIDS 2009
• One RCT with a HIV end point
• (Celum et al, IAS, Abstract WELBC101)

11
Proportion of genital samples with detectable
HIV
14 reduction in detectable genital HIV
12
Effects on Genital Viral Load
-0.26 log10 cps/mL decrease in genital viral load
13
Effects on Plasma Viral Load
-0.30 log10 cps/mL decrease in plasma viral load
14
HSV-2 Suppression in HIV Co-infected Partners
in Serodiscordant couples (Celum et al, IAS,
Abstract WELBC101.)

• 3408 HIV-serodiscordant couples
• Co-infected HIV partners treated with acyclovir
  400mg bid
• Primary endpoint HIV transmission
• Results
• HIV transmission HR 0.92 (0.60-1.41)ns
• HSV-2 GUD HR 0.27 (0.20-0.36) lt0.001
• Plasma viral load -0.25 log10 cps/mLlt0.001

15
Thomas Huxley The greatest tragedy of Science
is the slaying of a beautiful hypothesis by an
ugly fact

• Trials fail to support the hypothesis that
  control of classical STIs can prevent HIV
  infection
• Is it time to reassess the STI/HIV hypothesis?

16
Heresy (Gray and Wawer, Lancet 20083712064)

• STI control is a priority, but its effects on the
  HIV epidemic may be limited.
• Should we change STI/HIV prevention policies?
• Evaluate other non-STI causes of genital tract
  inflammation?
• Genital microbiome pro-inflammatory pathogens
• 16S rRNA pyrosequencing

17
Microbicides

• Urgent need for female controlled prevention led
  to 11 completed microbicide trials for HIV
  prevention
• Surfactants
• Nonoxynol 9 (N-9), Savvy
• Blocking Agents
• PRO 2000, Cellulose Sulphate (CS), Carraguard
• Acidifying Agents
• BufferGel

18
N-9 Trials with an HIV Endpoint
4 trials, all negative One borderline
protective One borderline harm
19
N-9 Trials and Vaginal Epithelial Disruption
(Wilkinson et al. Lancet Inf Dis 20022613)
18 increased risk of vaginal epithelial lesions
20
Newer microbicides with an HIV Endpoint
7 trials, all negative in ITT analyses Pro 2000
30 efficacy in as treated analysis
21
Newer microbicides and epithelial
disruption(Poynten et al AIDS 2009231245)
No increase in lesions with Cellulose Sulphate or
Carraguard
22
Issues with microbicide trials

• Screening for HIV inhibition
• Done in the absence of seminal plasma
• Addition of seminal plasma reduces viral
  inhibition
• HIV (Neurath et al BMC Inf Dis 20066150)
• HSV-2 (Patel et al JID 20071961394)
• Need to screen products in the presence of
  seminal and cervical secretions

23
Other Issues with Microbicide Trials

• Enrolled high risk populations (8/11 RCTs)
• Frequent intercourse and frequent product use
• Increase risks of vaginal lesions?
• Reduce compliance with intercourse related
  methods?
• High pregnancy rates
• Protocol required interruption of use reduced
  power
• Need testing of safety in pregnancy prior to
  trials?
• and/or
• Require effective contraception in trials
• Lower than expected HIV incidence
• Compromised power in West African SAVVY RCTs
• Need to establish incidence prior to trial
  initiation

24
Albert Einstein
Insanity is doing the same thing over and over
again and expecting different results.

• Need to learn from problems with completed RCTs
  and not repeat them
• Future RCTs will assess antiretroviral
  microbicides (Hladik PLos Med 20085e167)
• Need more research on mucosal barriers

25
Pre-exposure prophylaxis (PrEP) Trials

• 7 PrEP ongoing trials assessing TDF or Truvada
• Animal data promising
• One Phase 2 RCT (under powered) (Peterson et
  al Plos Clin Trials 2007e27)
• HIV RR 0.35 (0.03-1.93)

26
Vaccine Trials

• 2 trials of Vaxgen (recombinant gp 120 to elicit
  humoral immunity)
• 2 trials of MRKAd5 (gag, pol and nef with
  adenovirus 5 vector, to elicit CTL responses).
• One completed (STEP trial),
• One stopped early (Phambuli)

27
HIV Vaccine Trials
3 completed trials, all negative One suggested
possible harm
28
STEP MRKAd5 Trial Issues

• Women
• 1135 high risk women enrolled,
• Too few HIV infections for efficacy estimates
  (under powered)
• Men
• HIV infections increased in men with prior Ad5
  immunity (HR 2.0)
• Especially if they were uncircumcised (HR
  4.0)
• Did the vaccine activate HIV-specific T cells in
  foreskin mucosa? (Uberla PLOs 200841880, Corey
  AIDS 2009233)

29
ELISpot for Vaccine Screening

• The MRKAd5 vaccine elicited HIV-specific ELISpot
  responses in phase 1 studies
• Corey et al AIDS 2009233 no obvious correlate
  between immune responses to the vaccination and
  the acquisition of HIV infection.
• Should ELISpot be used to screen CTL vaccine
  candidates?

30
Vaccines A Need for New Directions

• Systemic immunity rarely occurs in nature and may
  not be achievable
• Rapid viral replication in the mucosa and rapid
  systemic dissemination (Li et al. Science
  20093231726)
• Should priority shift to mucosal vaccines which
  elicit an immune response at the portal of entry?

31
Male Circumcision (MC) for HIV Prevention in Men

• Observational studies
• MC associated with reduced HIV prevalence and
  incidence in men
• Meta-analysis (Weiss Geneva, 2005)
• General populations
• RR 0.57 (0.47-0.70)
• High risk populations
• RR 0.31 (0.23-0.42)

32
Three Trials of Circumcision for Prevention of
HIV Acquisition in Men
All trials showed efficacy at interim analyses
All were stopped early Overall efficacy 57
33
Other STIs and Circumcision in Men

• GUD
• RR 0.53 (0.43-0.64) Gray et al Lancet 2007
• HSV-2
• RR 0.72 (0.56-0.92) Tobian et al NEJM 2009
• HPV
• RR 0.65 (0.46-0.90) Tobian et al NEJM 2009
• RR 0.66 (0.51-0.86) Auvert et al JID 2009
• Trichomonas, Gonorrhea, Chlamydia
• No statistically significant effects (Sobngwi et
  al STI 2009, Mehta et al JID 2009)
• Pro-inflammatory anaerobes
• Marked reduction following circumcision (Price et
  al CROI 2009)

34
Biologic Plausibility that MC Prevents HIV

• Foreskin contains HIV target cells
• Dendritic cells
• CD4 and CD8 cells, activated with
  inflammation (Johnston et al AIDS 2009)
• Foreskin vulnerable to trauma and GUD
• Moist subpreputial space may
• Enhance HIV survival
• Provide an environment for inflammatory anaerobes

35
Male Circumcision for HIV Prevention in Women
Gray et al AIDS 2000142371
Male Viral load

• Baeten et al IAS 2009, Abstract LBPEC06)

36
Trial of circumcision of HIV men and
transmission to female partners in Rakai, Uganda.
(Wawer et al Lancet, 2009 www.thelancet.com 373)

• 922 HIV men randomized to immediate or delayed
  circumcision
• HIV men in serodiscordant relationships with
  uninfected women assessed for male-to-female HIV
  transmission
• Cannot deny HIV men circumcision because of
  unintended social consequences (stigmatization)

37
Male-to-Female HIV Transmission, Rakai Trial
(Wawer et al Lancet, 2009)
Adj HR 1.49 (0.62-3.57)
38
Female HIV infection 0-6 months, by resumption of
sex and wound healing
Early resumption of sex may increase HIV
transmission
39
HIV viral load and circumcision in HIV men
(Wawer et al IAS 2009, Abstract CDC119)
Stress of MC increases plasma viral load may
increase infectivity?
40
Circumcision and female genital tract infections
(Gray et Am J Obstet Gynecol 2008199

• Female partners of HIV-negative intervention arm
  men have reduced
• GUD RR 0.78 (0.63-0.97)
• Trichomonas RR 0.52 (0.05-0.98)
• BV RR 0.60 (0.38-0.94)
• Severe BV RR 0.39 (0.24-0.64)

41
Prevention trials are difficult and expensive.

• Trial design
• Pretrial incidence estimates
• Compliance with intercourse-related methods
  (e.g., lower risk populations)
• Require contraception if intervention is
  interrupted by pregnancy
• Improve screening of candidate interventions
• Microbicides
• Vaccines
• Fewer trials
• more rigor and quality?
• It could improve our batting average!

42
Einstein has the last word
If we knew what we were doing, it would not be
called research, would it?
A person who never made a mistake never tried
anything new.
He might have added that we should learn from our
mistakes