Further efficacy analyses from the iPrEx study

Simon Collins, HIV i-Base

Robert Grant from the Gladstone Institute of Virology and Immunology San Francisco presented follow-up results from the international iPrEx study. [1]

The initial analysis published in the NEJM in November 2010 reported a 44% reduction (36 vs 64 infections; 95%CI 15 to 63%; p=0.005) from use of daily oral tenofovir/FTC compared to placebo in 2499 young sexually active MSM at high risk of catching HIV. However, all infections in the active arm of the study were associated with undetectable (91%) or low (9%) drug levels in blood (compared to 51% of uninfected matched cases) indicating higher protection rates (~95%) with actual use. [2]

Results presented at CROI 2011 from an additional final four months of blinded randomised arms included 12 new infections in the active vs 19 in the placebo groups (p=0.002). During follow-up three months after drug/placebo discontinuation there were 4 new infections in the active arm and 2 in the placebo group (p=NS).

By all analyses, these final data retained similar significant protection rates to the initial published results. Protection was seen in all subgroups (by age, ethnicity, region, schooling, alcohol use, circumcision) but greater protection was seen in people reporting higher risks (i.e. unprotected anal intercourse, p=0.03).

Side effects were reported at similar rates in both arms and are summarised in Table 1. Differences in nausea and weight loss were seen in the first four weeks that subsequently resolved to placebo levels. No differences were seen in laboratory safety markers including phosphate, electrolytes, AST, ALT, amylase, glucose etc. While encouraging, these safety data need to be interpreted with consideration of the minimal adherence in each arm suggested by the PK sub-study.

Table 1: Adverse events reported in iPrEx; n (%)
Tenofovir / FTC placebo P
Depression 43 (3%) 62 (5%) 0.07
Grade 3/4 events 151 (12%) 164 (13%) p=0.51
Death 1 (<1%) 4 (<1%) p=0.18
Serious AE 60 (5%) 67 (5%) p=0.57
Diarrhoea 46 (4%) 56 (4%) p=0.36
Headache 56 (4%) 41 (3%) p=0.10
Nausea 20 (2%) 9 (<1%) p=0.04
Weight decrease 27 (2%) 14 (1%) p=0.04
Creatinine elevation 25 (2%) 14 (1%) p=0.08
Confirmed Cr increase   (0.4%) 5 0 p=0.06

Drug resistance was not detected in any of the cases of seroconversion that occurred during the study, but resistance to FTC was found in three people who were seronegative at baseline but later found to be recently infected by PCR viral load.

Self-reported risk reductions in behaviour changes in terms of partner reduction and increased condom use should perhaps be considered cautiously given the low correlation between reported and actual adherence.

However, the positive results reported from the study should increase the likelihood for actual adherence in the future. Of interest, despite this optimism, Dr Grant suggested that actual adherence should not be assumed, even knowing when protection is proven.

Participants in iPrEx study are being given the option to enroll in a follow-up study of open label. tenofovir/FTC.

Bone mineral density (BMD) changes in HIV-negative men using tenofovir/FTC PrEP

Two studies were presented looking at the impact on bone mineral density during PrEP. These are important both for the opportunity to evaluate the safety of PrEP but also to look at the previously observed side effects of tenofovir separate from an additional impact of both HIV and ageing.

Albert Liu from University of California, San Francisco presented baseline (n=200) and longitudinal (n=184) results from HIV-negative men followed from 2005-2007 in a randomised tenofovir vs placebo study run in San Francisco. [3]

Demographics and baseline characteristics included: median age 40 years (range 19-60); 77% white, 7% Latino or Hispanic, 5% African American, and 5% Asian or Pacific Islander. Approximately 20% were smokers, 40% with moderate/heavy alcohol use, 15% corticosteroid use and 60% used multivitamin, calcium of vitamin D supplements. At baseline a higher proportion of men were found to have low BMD (defined as z-score > –2.0) than would have been expected (~10%; 20 vs 4.8 cases; p<0.001).

DEXA scans performed at baseline, 9 or 12 months, and 24 months showed –1.1% net decrease in mean BMD in the tenofovir vs placebo group at the femoral neck (95%CI 0.4 to 1.9%, p = 0.004) and a –0.8% net decline at the total hip (95%CI 0.3 to 1.3%, p = 0.003). At the spine (L2 to L4) there was a trend towards a –0.7% difference (95%CI –0.1 to 1.5%, p = 0.11). These results were similar after adjustment for baseline BMD, BMI, race, age, and creatinine clearance.

Declines mainly occurred during the first 12 to 15 months of treatment and were generally small. More significant losses (>3% BDM) occurred more frequently in the tenofovir group for hip and femoral neck (both p=0.02) but this was no longer significant at any location when using a >5% cut off (p=1.0, 0.13 and 0.72; for hip, neck and spine respectively).

The ten reported fractures (six in the active and four in the placebo group; p=0.75) were all trauma-related and judged unrelated to study drug.

In a second study, Kathleen Mulligan from University of California, San Francisco presented hip and spine BMD results measured every 24 weeks in approximately 500 people enrolled in a substudy of iPrEx. [4]

Participant were enrolled from sites in Peru (n=221), Thailand (n=95), the US (n=71), South Africa (n=61) and Brazil (n=55) and resulted in broader ethnicity: 18% were Caucasian, 13% black, 20% Asian, 49% mixed race; 52% were Hispanic or Latino. This was a younger population: 48% were under 25 years old and therefore likely to still be experiencing bone mass growth.

Although >60% were active (weight bearing exercise), alcohol (80%) and tobacco (43%) use was common. Mean body mass index was 23.5 (SE 0.2) kg/m2. At baseline, total age and race adjusted z-score for the group which would be expected to be 1.0 was negatively shifted to the left. Low bone mineral density (z-score < –1.0/–2.0) was seen in 36%/12% people in the spine and 18%/2% in the hip, with no difference between the active and placebo groups. Unfortunately there are limited studies of bone disease in otherwise healthy young men to verify the normative data used in bone studies, however the same data are used when evaluating low BMD related to HIV, where similar or greater rates of low BMD are commonly reported.

During the study, bone mineral density tended to increase in the placebo arm and decrease in the FTC/TDF arm, resulting in modest (–0.7 to –1.0%) but statistically significant differences between the groups by week 24 (see Table 2). These are lower than the –2 to –4% seen in HIV-positive people starting tenofovir-containing HAART. Lower patient numbers at later timepoints will require later analyses.

While these are encouraging data, it was also pointed out that it would not be expected to see differences in young healthy men, although limited drug exposure due to low treatment adherence, will also have diluted any reductions caused by tenofovir.

Table 2: Mean (SE) percent change in bone mineral density from enrollment
FTC/TDF Placebo Difference (95%CI) p value
Total hip
week 24 –0.31 (0.13) +0.34 (0.13) –0.65(–1.03 to –0.28) 0.001
week 48 –0.05 (0.22) +0.90 (0.22) –0.95(–1.56 to –0.35) 0.002
week 72 +0.27 (0.28) +0.49 (0.28) –0.22(–1.00 to 0.56) 0.581
week 24 –0.66 (0.20) +0.29 (0.20) –0.95(–1.51 to –0.39) 0.001
week 48 –0.41 (0.24) +0.13 (0.24) –0.54(–1.20 to 0.12) 0.106
week 72 –0.97 (0.32) –0.10 (0.32) –0.87(–1.75 to –0.01) 0.052

Pharmacokinetic substudy: drug levels in iPrEx

The nested case-control pharmacokinetic substudy in iPrEx matched samples from 34 newly infected participants to 43 controls who remained HIV-negative. The group tested for tenofovir diphosphate and FTC-triphosphate (the active moieties of tenofovir and FTC) in blood plasma and stored PBMCs.

The results were important in demonstrating a strong correlation between presence of active drug and risk of infection/protection (present in 9% of cases vs 51% of controls, p<0.001). Importantly, this was able to show >95% concordance between paired plasma and PBMC samples.

In the presentation at CROI, Peter Anderson presented pharmacokinetic results that were independently associated with detectable drug levels. [5]

This included study site region (97% for US vs 50% for non-US participants, p<0.0001), age (73% in 25 year olds vs 44% in under 25 year olds, p<0.001) recent sexual activity (73% with recent high risk URAI vs 59% without UAI vs 35% no activity within 12 weeks, p=0,003).

The open-label phase of the iPrEx study will include real-time drug level and adherence analysis that will help increase the understanding of drug concentration associated with protection.

The drug level test used in these studies was sensitive enough to drug use within the previous two weeks if the person was already at study state levels, but this period would have been shorter for sporadic adherence.

The relationship between drug levels and degree of protection are a crucial aspect of how and whether PrEP will find a role in prevention strategies.

Intracellular drug levels following single doses of tenofovir/FTC were reported from seven HIV-negative volunteers (5 women, 2 men) in a poster from the same study group. [6]

Tenofovir-DP levels were as expected in plasma but reduced in PBMCs: they were only 15% of those expected at steady state dosing and were only 35% of the levels seen in primate studies. FTC-TP levels achieved 30% of those expected at steady state but were similar to single-dose primate studies.

The clinical results from iPrEx may not have required the optimum levels achieved at steady state from daily dosing but current analysis are unable to show this.

It would seem to be important to prioritise this aspect of protection in animal studies to further inform the likely protection in humans from weekly or twice-weekly dosing.


Unless mentioned otherwise, all references are to the Programme and Abstracts of the 18th Conference on Retroviruses and Opportunistic Infections, 28 February– – 2 March 2011, Boston.

Webcasts are available at the following link:

  1. Grant R et al. Pre-exposure Chemoprophylaxis for Prevention of HIV among Trans-women and MSM: iPREx Study. 18th CROI, 2011. Oral abstract 92.
  2. Grant RM et al. Preexposure chemoprophylaxis for HIV prevention in men who have sex with men. NEJM. 23 November 2010 (10.1056/NEJMoa1011205). Free access:
  3. Liu A et al. BMD loss in HIV-negative men participating in a TDF PrEP clinical trial in San Francisco. 18th CROI, 2011. Oral abstract 93.
  4. Mulligan K et al. Effects of FTC/TDF on bone mineral density in seronegative men from 4 continents: DEXA results of the global iPrEx study.18th CROI, 2011.  Oral abstract 94LB.
  5. Anderson P et al. Interpreting Detection Rates of Intracellular FTC-TP and TFV-DP: The iPrEx Trial. 18th CROI, 2011. Oral abstract 96LB.
  6. Anderson P et al. Single-dose pharmacokinetic profile of intracellular TFV-DP and FTC-TP in HIV-negative volunteers. 18th CROI, 2011. Poster abstract 641,

Links to other websites are current at date of posting but not maintained.