The results from the Ipergay study of intermittent pre-exposure prophylaxis (PrEP) were published in the New England Journal of Medicine (NEJM) on 1 December, World AIDS Day. The journal-published results are little changed from those presented at the CROI conference last February by principal investigator Jean-Michel Molina but the researchers make a number of additional comments and are notably cautious about not over-interpreting a study that only had 400 participants – very small for a prevention study – and an average of nine months’ follow-up.
Despite this, Ipergay represents a major innovation in PrEP. It is the first, and so far only, randomised study to show that an intermittent regimen of PrEP, that users only take when anticipating sex, can be just as effective as daily PrEP.
To summarise the results: in Ipergay 400 gay men and transgender women were randomised to either take Truvada (tenofovir/emtricitabine) or a placebo.
The regimen that participants were told to take was a double dose (two pills) 24 to two hours in advance of anticipated sex, and then a pill on both of the two days following sex. If they continued to have sex, they were to keep taking one pill a day until two days after the last sex. If they restarted sex, they had to take a double dose if there had been more than a week since the previous sex.
The use of a placebo was somewhat controversial in Ipergay as some activists felt that the effectiveness of PrEP had already been sufficiently demonstrated by the iPrEx study of 2010. However effectiveness in iPrEx was only 42% overall, and the writers comment that “The use of a placebo was deemed to be justified because of the inconsistent efficacy of PrEP in previous trials and the moderate efficacy in the iPrEx trial.”
There were 19 HIV infections during the whole Ipergay trial period but three of those were determined to have happened shortly before participants started taking PrEP/placebo. There were 14 infections in men allocated to placebo and two in men allocated to PrEP.
In fact neither of these men had taken PrEP, or hardly any, in the two-month period before they tested positive (in Ipergay, clinic visits were every eight weeks). One new detail added in the NEJM piece is that participants were asked to return unused pills and on the visit they were diagnosed, these two participants returned respectively 60 and 58 out of the 60 pills given to them two months previously.
The 14/2 difference in infections represents an 86% (95% confidence interval: 40%-98%) reduction in the risk of HIV infection in the trial participants allocated to Truvada. HIV incidence in the trial was 6.6% a year in placebo recipients and 0.91% in Truvada recipients. The 6.6% is more than double the 3% incidence anticipated before the trial, though not as high as the 9% seen in the deferred-PrEP arm in PROUD.
We learn a little more about the study participants from the report. Their average age was 34.5, with one in seven participants aged 18 to 24 and another one in seven 25 to 29. The majority (62%) were aged 30 to 49. Only a quarter said they were in a relationship but remarkably – and this is presumably one reason for volunteering for the study – 30% of those in a relationship said their partner was HIV-positive (slightly more in the Truvada arm).
Half the participants were from Paris and 11% from Montréal in Canada, with the others from Lyon, Nice, Tourcoing and Nantes in France. They had high levels of recreational drug use (44.5% in the last year had used ecstasy, methamphetamine, amphetamine, cocaine, GHB or GBL) and 25% had had more than five alcoholic drinks per day in the past month.
The men had had a median number of eight sexual partners in the past two months – similar to PROUD, where participants averaged ten in the previous three months. This was the one behavioural indicator that changed slightly during the study: it fell to 7.5 partners per two months in the placebo arm and not the Truvada arm and this was statistically significant but nonetheless likely to be a chance finding.
Apart from this slight drop in the number of partners in participants on placebo, sexual behaviour did not change during the study, with 70% throughout and in both arms having condomless anal sex in the previous two months, and two-thirds of those having it receptively.
Adherence was as described before: while it is challenging to measure or even define ‘adherence’ to an intermittent regimen, 86% of participants on Truvada had levels of tenofovir in their body consistent with them taking at least one dose a week. Interestingly so did eight men in the placebo arm; three of these had taken post-exposure prophylaxis (PEP) but presumably the other five had taken “informal PrEP”.
Only 43% in self-report, however, said they had taken PrEP in accordance with the study regimen last time they had sex, while 28% had not taken it at all.
On average, participants took PrEP half the time, i.e. 15 days per month; 17% took PrEP more than 26 days a month, i.e. essentially daily, while 31% took it for eleven days or fewer, i.e. less than 2.5 days per week, a level that, according to data from the iPrEx study, would not be enough to protect them from HIV if it represented a steady level of PrEP taking.
There was, however, tremendous variability in individual participants’ adherence ‘careers’ throughout the study; while some took it steadily and others scarcely at all, many participants stopped or restarted PrEP irregularly, many with gaps of several months off PrEP. Since irregular adherence (and the fact that 29% of sex was not covered by PrEP at all) was not associated with HIV infection, this presumably means that participants were correctly judging their levels of risk and only taking PrEP when they judged it as high.
Fourteen per cent of participants took PEP during Ipergay (16% on Truvada and 12% on placebo), compared with an average of 18% in PROUD.
There were no serious drug-related adverse events in the study though one participant had to stop Truvada due to a suspected reaction with the anticoagulant drug dabigatran. However there were significantly more gastro-intestinal events such as nausea, vomiting, abdominal pain or diarrhoea in the Truvada group than the placebo group (14% versus 5%).
Creatinine levels, an indication of kidney performance, were raised in 18% of participants on Truvada versus 10% on placebo. Two participants on Truvada had drops in creatinine clearance to below 60mls/minute, definitive of mild kidney dysfunction, but these were transient and did not last.
In their comments the Ipergay researchers are notably cautious about their findings. They comment that although the effectiveness of PrEP seen in Ipergay was nearly the highest to date, “the short follow-up for our study may have increased the likelihood of an exaggerated estimate of efficacy due in part to high initial adherence.”
They add that “given that participants took a median of 15 pills a month, the results…cannot be extrapolated to [gay men] who have less frequent sexual intercourse and thus would be taking tenofovir/emtricitabine on a more intermittent regimen.” By definition men who had sex less often would have lower HIV incidence and so a much bigger study would be needed to establish whether intermittent PrEP was effective in them.
Molina J-M et al. On-demand preexposure prophylaxis in man at high risk for HIV-1 infection. NEJM early online publication, DOI: 10.1056/NEJMoa1506273. 2015. See full text here.
