Meta-analysis finds high but variable STI rates in PrEP studies – is PrEP the cause or a potential solution?

A new meta-analysis of 20 studies and roll-out projects of HIV pre-exposure prophylaxis (PrEP) in gay and bisexual men confirms very high rates of diagnosis of sexually transmitted infections (STIs) in study participants and PrEP takers.

Dr Ricardo Werner and colleagues from the Berlin Institute of Health analyse STI rates in the largest dataset of PrEP studies and rollout programmes so far. In their discussion section they also look at the evidence for and against PrEP having a causative role in STI increases, or whether these are driven more by other factors such as increased testing.

The meta-analysis finds that the rates of diagnosis for any STI in participants in the studies ranges from 33 to 100% – meaning, in the latter case, that there were more STI diagnoses over the course of a year than there were participants in the study.

Thus there is considerable heterogeneity in the STI incidence recorded by different studies. Some of the studies they refer to also show heterogeneity between individuals, with a minority of participants accounting for a majority of STI infections.

The authors add that the evidence for whether increases in STIs happen before or after PrEP initiation is very mixed, and suggests that at least some of the high rates of STIs seen in people taking PrEP is due to high rates of STI testing.

The authors conclude that the incidence rates of STIs among gay men who engage in high-risk behaviour are high and raise particular concerns with regard to gonorrhoea and hepatitis C. They add that “by offering access to structures that provide regular STI monitoring and prompt treatment, PrEP may not only decrease HIV incidence but also have beneficial effects in decreasing the burden of STIs.”

The meta-analysis

Werner and colleagues analysed STI rates in 24 papers concerning 20 different studies of PrEP in gay and bisexual men. These studies included randomised controlled studies such as iPrEx, PROUD and Ipergay and their open-label extensions, as well as studies targeting particular populations such as the PrEPare study for teenagers. They also included country-specific demonstration studies such as PrEP Brasil, AmPrEP in the Netherlands and the Australian demo projects, and also PrEP rollout programmes such as the Kaiser Permanente PrEP rollout in northern California and Prévenir in France.

A total of 11,918 gay and bisexual men were included and the data included about the same number of person-years of follow-up (11,686).

The annual diagnosis rates for any STI ranged, as mentioned above, from 33 to 100%, with the average diagnosis rate among the highest-quality studies being 84%.

Of the three most common bacterial STIs (syphilis, gonorrhoea and chlamydia), syphilis is the one where increased STI testing may have the smallest impact on diagnoses. This is because it can be picked up in a blood test and is therefore already tested for more consistently; in contrast, gonorrhoea and chlamydia diagnoses are often dependent on rectal and throat swabs that are less consistently performed, at least in some countries.

The average diagnosis rate of syphilis in PrEP-takers in this meta-analysis was 9.2% overall and 9.5% in the highest-quality studies.

These rates are obviously much higher than in the general population, where only one in 10,000 people a year (0.0097%) is diagnosed with syphilis. But, while higher, it is not of a different order of magnitude than the rate seen in gay and bisexual men attending STI clinics, which in London in 2016 was 4.4%.

Bacterial STIs have a specific age profile in both men and women, with low rates seen in very young people just starting sex, but the highest rates in people in their late teens and twenties, when they are most sexually active. In this meta-analysis the lowest rate of syphilis was in a study in 15 to 17 year olds (1.8%), but the highest rate was in its companion study in 18 to 22 year olds (15%).

The rate of gonorrhoea varied from 12 to 43%, with an average rate of 27% and a rate in the most rigorously controlled and largest studies of 40%.

Chlamydia was diagnosed at very similar rates: the range was 14 to 48% in different studies, with an average rate of 30%, and 42% in the most rigorously controlled and largest studies.

One reason the larger studies had higher rates is because STI rates tended to be higher in the cohort studies and rollout programmes than they were in the earlier randomised controlled studies, where people didn’t necessarily know they were on PrEP and which were done at a time when STI rates were somewhat lower.

Gonorrhoea and chlamydia were also classified by body site. In the most rigorous analysis, 4% had urethral gonorrhoea and 9% urethral chlamydia; rectal infections were more common, with rates of 17% and 33% respectively for rectal infections.

Five studies reported on the incidence of hepatitis C. This ranged from zero to 1.9%, with an average annual rate of 1.3%. This is very high, given that annual hepatitis C incidence in HIV-positive gay men is only 0.78% and in HIV-negative men in general is only 0.04%.

There is thus no doubt that there are high rates of STIs in people taking PrEP. However, the evidence as to whether PrEP leads to people acquiring more STIs is much more ambiguous. In particular, the picture is confounded by the fact that PrEP leads to people taking STI tests and getting diagnosed and treated more often.

In their discussion section, the authors summarise this evidence briefly: the following section elaborates on this, adding in findings from some of the studies they reference.

Discussion – does PrEP raise STI rates?

The first meta-analysis (Kojima) to compare STI rates in studies of gay men taking PrEP and studies of gay men not on PrEP caused considerable concern and controversy when researchers reported that men using PrEP were 25 times more likely to acquire gonorrhoea and 47 times more likely to acquire syphilis than gay men not on PrEP.

However, the authors themselves commented that “PrEP studies recruited MSM [men who have sex with men] with high-risk sexual behaviour, whereas MSM in studies not using PrEP may have had different baseline risk behaviour.” Also, the studies they included differed in the populations included, how often they were tested, and which tests were used.

Their findings were also further challenged by other researchers (Harawa). They pointed out that more than half of the data on gay men in the entire meta-analysis came from Project EXPLORE, an interventional study that was explicitly designed to increase condom use. The researchers did not test participants routinely after its first year, likely leading to STIs being underestimated. If this study was excluded from the analysis, Harawa and colleagues found, gay men on PrEP were 2.8 times more likely to get an STI than those not on PrEP – still a considerably raised risk, but ten times less than the estimate in the original.

Another study presented in 2016, called SPARK (Golub) found evidence that increased testing rates were contributing to at least some of the rise in STI diagnoses in gay men on PrEP. In SPARK, from 68 to 83% of the STIs diagnosed were asymptomatic. The researchers estimated that 24% of STIs would have been missed even if participants had screened for STIs every six months, as the US Centers for Disease Control and Prevention recommend, instead of every three months as in the trial.

A Canadian study published in 2018 (Nguyen – data collected 2010-2015) compared STI rates in the same men before and after starting PrEP. After adjusting for more testing on PrEP, it found that there was a 39% raised risk of STI diagnosis following PrEP, but that this was not statistically significant.

A study presented at CROI in 2017 (Montaño), based on data collected in Seattle in 2014-2016, found that while chlamydia rates did increase after men started on PrEP, rates of syphilis and gonorrhoea increased in the year before men started on PrEP. Rates of gonorrhoea increased no further after starting PrEP, while rates of syphilis declined.

In March 2018 the a meta-analysis of eight PrEP studies (Traeger – data collected between 2014 and 2017), found that the risk of acquiring an STI increased by 25% in the first six months after starting PrEP; the risk of rectal STIs increased by 39% and of rectal chlamydia by 59%. There was also evidence that the increase was greater in more recent studies, with STIs increasing by 47% in the first six months after starting PrEP in studies terminating after 2016. However, there were only two out of the eight studies included in the analysis where the increase in STIs was statistically significant.

The lead author of this meta-analysis said, in a later presentation at the International AIDS Conference in July 2018, that post-PrEP rises in STIs varied between individuals. An example comes from the PrEPX study in the state of Victoria, Australia. While the risk of STIs increased by 71% in men who were new to PrEP in their first year in the study, during that year 52% of men did not have an STI, while 25% had had two or more, accounting for 76% of infections, and 13% had had three or more, accounting for 53%.

Similarly, an American study published last September (Beymer) showed that although urethral chlamydia increased by 83% in men after they started PrEP and syphilis rose threefold, only 28% of men actually had an increase in STI risk as individuals.

Another American study (Rendina) presented at last year's IAS conference was able to measure STI infections in a subgroup of men who stopped PrEP before, during and after their time on PrEP. The study found that while the number of condomless sex acts men reported tripled during their time on PrEP, the proportion diagnosed with a rectal STI only increased from 7% to 10%, and fell back to 2% when they stopped PrEP. The author commented that his study showed association, but not causation. He said: “[Men] are taking PrEP when they are engaging in high risks and they are stopping PrEP at times when they are no longer at high risk.”

Conclusion

The meta-analysis and accompanying evidence shows that although STI rates are generally high in gay and bisexual men taking PrEP, there is a large variation between studies and the populations they involve, and some of the studies referenced in the discussion section find that STI infections are concentrated amongst a particularly at-risk minority.

The authors of the meta-analysis comment: “Despite the heterogeneity of some of our results, our overall findings suggest that the incidence rates of various STIs among gay and bisexual men who engage in high-risk sexual behaviour is high.” In other words, the kind of sex that exposes people to HIV unless they take PrEP also exposes them to the other STIs – and so, as the authors add, “This subgroup of men can benefit from access to STI testing and treatment at close intervals.”

They add: “The use of PrEP is a highly effective means of preventing HIV and should be embedded in a comprehensive programme targeting primary and secondary prevention and treatment of other STIs.”