A meta-analysis of 17 studies of HIV pre-exposure prophylaxis (PrEP) in gay men and other men who have sex with men (MSM) has found that, while PrEP protected them from HIV, the proportion diagnosed with gonorrhoea, chlamydia or syphilis increased significantly in the period between starting PrEP and follow-up, with an average length of time on PrEP at follow-up of six months.
The headline figures come from eight studies that recorded both STI diagnoses at both baseline and during follow-up. On average, after starting PrEP, there was:
- an increase of 24% in diagnoses of any of these three STIs
- a 39% increase in rectal STIs
- a 59% increase in rectal chlamydia.
All of these increases were statistically significant.
The studies included in the analysis
The meta-analysis was conducted by Australian researchers, but eleven of the 17 studies were conducted in the USA and only two, at least partly, in lower-income settings. Only open-label studies were included, meaning that all participants knew they were taking PrEP. One, the PROUD study, was a randomised controlled study (RCT); another, iPrEx-OLE, was an open-label extension of an RCT; the others were cohort studies of PrEP rollout programmes or smaller demonstration projects for specific populations.
Thirteen of the studies measured sexual risk behaviour in the follow-up period, eleven measured STI diagnoses, but only seven measured both. Only eight had STI diagnoses measured both at baseline (covering the previous three to twelve months) and for a similar period at follow-up. These eight were used for the comparison. Two studies (of the Kaiser Permanente PrEP programme in northern California) included some of the same subjects but reported on different outcomes, so were treated as separate studies.
Results: risk behaviour
The researchers could not calculate an average overall change in sexual risk behaviour (unlike changes in STI diagnoses) because the measures of risk used in studies differed too widely from each other.
Seven studies reported an increase in condomless anal intercourse between baseline and follow-up, with four increases being statistically significant: ‘condomless anal sex’ was defined in different ways ranging from proportion saying that they had had condomless sex to the proportion saying they never used condoms.
Five studies reported an increase in condomless receptive anal intercourse (CRAI), one of them significant, and one a significant decrease. One found no increase in the overall number or proportion of people who had CRAI but did find a significant increase in the proportion who had CRAI who had more than ten sex partners in the previous three months.
Four studies found an increase in the overall number of sex partners, one significant, and one reported a significant decrease. Three studies found an increase in the proportion reporting condomless sex with HIV-positive partners, one of them significant.
More studies reported increases in occasions of condomless sex than in the overall number of partners or in the proportion of men who had condomless sex.
More results on STIs
The researchers looked at studies with results published between the beginning of 2014 and August 2017. There was a greater and significant increase of 47% in post-PrEP STI diagnoses in studies still running at the beginning of 2016, compared with ones that stopped before that. Only studies where the follow-up time was more than 12 months reported an increase in STIs, of 48%; the increase was only 8% in studies with shorter follow-up times. This difference did not reach statistical significance (p = 0.12). However, taken alongside the data on the date of the studies, it suggests STI risk increased over time, both within a study, and between different studies.
The meta-analysis averages out individual features of studies, and there was actually a wide variation in the data on STIs. Out of the eight studies that compared baseline with follow-up STI diagnoses, one observed a 60% decrease in STIs, though this was the smallest study and was not statistically significant. Three saw no change, and two saw increases (of 35% and 39% respectively) that were not by themselves statistically significant. One study found an increase in STIs in the first six months but a decrease, almost to baseline levels, in the following six months.
This meant that only two of the eight studies actually saw a sustained statistically significant increase in STIs between baseline and follow-up. One was the Kaiser Permanente rollout, which saw a 1.48 odds ratio (48%) increase in STIs (95% confidence interval, 1.18 to 1.85). This was the longest-lasting and the second-largest study, with 972 participants, meaning it had the biggest statistical ‘weight’.
The other study was a complete outlier. In this, the VIC-PrEP rollout study in Australia, STI diagnoses doubled, with a highly significant 2.98 (198%) odds ratio increase (95% confidence interval, 1.42-6.51). This study had the second-lowest statistical ‘weight’, with 317 participants but the increase is real enough and, possibly significantly, it was the most recent study.
It is interesting that this increase was observed in Australia. The researchers themselves draw attention to the “unprecedented rate of enrolment” into PrEP in that country and, as another recent study reports, Australia now has a higher proportion of its MSM population on PrEP than any other country, and has seen HIV diagnoses fall concomitantly. But it has also seen rates of condomless sex increase significantly since PrEP started, even among MSM not taking PrEP.
Conclusions and implications
Taken alongside these last two studies, the increases in STIs in the other studies which, by themselves, were not statistically significant, added up to a consistent trend that was significant. This is a good example of how meta-analysis can add power to a finding.
This study shows several things. It shows that, averaged across all the studies, starting PrEP has been accompanied by a statistically significant increase in STI diagnoses. It shows that this increase is greater in rectal STIs. It shows that the increase in STIs has become more pronounced in the last couple of years.
It also shows that in many of the studies, starting PrEP has also been accompanied by increases in what should perhaps now be called STI risk behaviour – given that the increases in STIs are accompanied by falls in HIV infection.
The researchers say that the data seem to show increases in the number of partners and/or the number of occasions in which condoms are not used, rather than the proportion of men who don’t use condoms. In other words, PrEP does not seem to be associated with a sudden abandonment of condoms, or in men who always use them stopping use. However, men who are already inconsistent condom users seem to be increasingly willing to take a chance on not using them. This is what would be expected if PrEP is preferentially used by those already at the highest risk of HIV.
The researchers comment that they could not establish the degree of correlation between starting PrEP and increasing ‘risk behaviour’ given that the measures of risk behaviour are so varied and are somewhat out of date. No study, for instance, asked about the actual or perceived viral loads of HIV-positive partners, and none asked specifically about group sex, though PROUD did subdivide risk by number of partners.
They also comment that it is now more important to chart how risk of both STIs and of HIV changes over time in individuals, as they move in and out of ‘seasons of risk’, rather than to document average behaviour over cohorts.
There is no evidence from this study that increases in STI testing may lead to more infections being treated and therefore an eventual fall in STIs. The STI increases were not associated with numbers of tests and therefore are more likely to reflect increased risk than increased diagnosis.
The authors conclude that “Responses to emerging trends in risk compensation need to be balanced against the considerable HIV transmission averted and the long-term prevention impact of greater PrEP coverage.”
Traeger MW et a;. Effects of pre-exposure prophylaxis for the prevention of human immunodeficiency virus infection on sexual risk behaviour in men who have sex with men: a systematic review and meta-analysis.Clinical Infectious Diseases, March 2018. doi: 10.1093/cid/ciy182. See abstract here.