Concern has been raised that pre-exposure prophylaxis (PrEP) could lead to, or contribute to, a new epidemic of sexually transmitted infections (STIs) other than HIV.
No other STI combines incurability with near-100% lethality in the absence of treatment as HIV does. But other STIs can have serious health consequences. The bacterial STIs – gonorrhoea, syphilis, chlamydia and chlamydia’s ‘superstrain’, lymphogranuloma venereum (LGV), as well as other infections like Mycoplasma genitalium and trichomoniasis – are generally easily curable but may have serious health consequences in some people if left untreated. Syphilis may also cause serious symptoms in early disease, and drug resistance in gonorrhoea is of considerable concern.
Viral STIs include herpes, genital warts and hepatitis B and hepatitis C; the latter two can eventually cause fatal liver disease if left untreated and human papillomavirus (HPV) can cause cervical, anal and oral cancer. The rest of this briefing, however, will concentrate mainly on gonorrhoea, chlamydia and syphilis, as these are the infections that have seen large increases in recent years.
The questions we will aim to answer in this briefing are:
- To what extent have STI infections actually increased in recent years?
- Are these increases especially concentrated in gay men, especially since PrEP became available?
- Are increases in STIs correlated with other changes in behaviour such as reduced condom use, ‘seroadaptive’ behaviours such as serosorting (preference for sexual partners of the same HIV status) or seropositioning (decisions on sex roles dependent on HIV status)?
- Does PrEP directly lead to STI increases? Or are they just happening at the same time?
- Are the observed increases in STIs caused by people taking more tests for them?
- Could increased testing and treatment rates for STIs actually help reduce STIs? Is there any evidence for this already happening and if not, why not?
This briefing will not discuss messaging strategies to deal with STIs in the era of PrEP, or the psychological and social benefits of PrEP.
STI epidemiology – history
There is no doubt that bacterial STI diagnoses, especially in men who have sex with men, have increased since their historic low in the 1990s when condoms were the only effective HIV prevention measure and fear of HIV led to their widespread use. However, it’s important to put recent increases into a historical context. Are STIs now at a historic high?
Gonorrhoea and syphilis data have been collected in the UK for a century, ever since the Venereal Disease Act of 1917 set up the country’s network of specialist, anonymous STI clinics.
In 2017 and 2018, there were around 45,000 and 56,000 cases of gonorrhoea respectively. A recent review by Public Health England found that that these were the highest annual numbers of cases since 1986, but gonorrhoea diagnoses exceeded this figure between 1928 and 1930. They remained at about 40,000 a year throughout the 1930s, fell by about 20% during World War II, then peaked abruptly in 1946 to 54,000 as soldiers came home from the front.
The advent of antibiotics brought that figure down to about 20,000 during the 1950s but gonorrhoea cases rose to over 60,000 cases during the 1970s, with an all-time high of over 64,000 a year between 1973 and 1977. Cases then began to fall rapidly and reached a historic low of 10,000 a year between 1993 and 1995.
However, diagnoses started to rise again as soon as antiretroviral combination therapy for HIV became available. There was a period between 2004 and 2010 when diagnoses were contained below 20,000 a year, but they have more than doubled since then. Diagnoses rose to nearly 44,700 in 2017, representing a 22% increase over the previous year and an 8% increase relative to 2015.
In the USA, gonorrhoea cases have been recorded since 1941, when prevalence was 140 cases per 100,000 population (0.14%). As in the UK, there was a post-war peak in cases so that national prevalence was 0.3% in 1946, with a longer and higher peak in the 1970s reaching 0.5%. Cases fell to a historic low of less than 0.1% per head of population in the 1990s and early 2000s but since 2011 have more than doubled, to 0.2%.
Gonorrhoea is an adaptable organism that retains the capacity to outwit medications. (Resistance is discussed later in this briefing.) In addition, diagnoses are not only sensitive to how many tests are done but also the anatomical sites that are tested; the UK has a better record than most European countries for rectal smear tests but does not always test for pharyngeal (throat) gonorrhoea, which is an important site of infection.
The history of syphilis is a little simpler as the syphilis organism rarely develops antibiotic resistance and is diagnosed via blood tests, which are more consistently performed. Between 1930 and 1945, UK cases ranged between 8000 and 16,000 a year, with the same 1946 peak as gonorrhoea, to over 20,000 cases. Cases then fell with the introduction of penicillin to a low of 700-800 cases in the late 1950s, before rising to a peak of about 3000 cases a year in the late 1970s. The ratio of male-to-female cases rose throughout the 1960s and 1970s, suggesting that male-to-male transmission was important in the pre-HIV era. Infections in the USA followed the same trends.
Cases then started to fall and reached a historic low of 280 in 1995, just before the advent of HIV combination therapy, with about a third of male cases being in gay men. There was a sudden doubling of cases in 2001, which was also the first year that cases in gay men formed more than 50% of male cases. Since then cases have risen sixfold overall, to over 6000, and tenfold in gay men, but only by 75% in women. Syphilis is now 18 times more common in men than women.
Epidemiology of other STIs
Increases in other STIs have been less marked. Chlamydia is more prevalent in women than men and, as an STI that is asymptomatic in 50% of cases in men and 75% in women, diagnoses may be as much influenced by the number of people testing as actual incidence. Male chlamydia cases have risen by 15% in the last ten years but these pale into insignificance compared with a 224% increase in gonorrhoea and a 183% increase in syphilis, both concentrated in gay men.
LGV is caused by a variant of chlamydia which has evolved considerably greater virulence and can cause unpleasant long-term complications. The UK has had over 40% of the LGV cases in Europe, and two-thirds of them are in London. LGV was extremely rare until 2003 when there was an outbreak in the Netherlands, with the first UK cases appearing in 2004. There were 919 UK cases in 2016, somewhat lower than in previous years.
Diagnoses of the most common viral STIs have not increased, unlike the bacterial ones. First diagnoses of genital herpes have increased by 11% in the last ten years but have stayed more or less constant for the last six years at about 32,000 cases a year.
Diagnoses of genital warts, caused by HPV, decreased by 28% between 2008 and 2017, from 78,000 to 59,000. This decrease, seen in young people and in women and men alike, has been ascribed to widespread vaccination of adolescent girls since 2008 and ‘herd immunity’ among male partners. The decline has not been seen in gay men, who are recommended to ask for HPV vaccination. HPV vaccination was extended to adolescent boys in the UK in July 2018.
While 180,000 people in the UK or one in 366 members of the general population may have chronic hepatitis B infection (HBV), most cases are among people who acquired it abroad or many years ago. New infections are now quite uncommon due to vaccination. There were only 445 acute or suspected acute cases of HBV in 2017, representing one per 125,000 people in the general population. Although declining, incidence is about twice as high in London. Fifty-five per cent of cases were ascribed to heterosexual contact and 15% to male/male sex, with the route of most other infections unknown.
The estimated prevalence of chronic hepatitis C infection (HCV) is the UK is roughly the same as chronic HBV at 210,000 cases or one in 314 people. In contrast to HBV it is spread largely by sharing needles. It is extremely common in current or former injecting drug users in the UK, with a consistent prevalence rate of 50% in this group, of whom 70% are now aware of their infection.
However HCV is also spread sexually, with a 2013 survey finding that 1.2% of HIV-negative and 7.7% of HIV-positive men in London were HCV positive. A global meta-analysis of studies of gay men between 1984 and 2012 found that the pooled hepatitis C annual incidence rate in men with HIV was 0.53% a year – or, in other words, that HCV incidence in gay men with HIV is of the same order of magnitude as HCV incidence is in HIV-negative gay men (which a 2012 UK study estimated at 0.43% a year).
More recently, studies from several European cities have found incidence rates of hepatitis C that are the same or higher in HIV-negative men than in HIV-positive men. In Lyon and London, 45% and 40% respectively of gay men with new hepatitis infections were HIV negative. Of these, 67% and 81% respectively were taking PrEP or had done during the previous year.
Syphilis and gonorrhoea specifically in gay men
Of the common STIs, it is only syphilis and gonorrhoea (aside from HIV) that have become especially concentrated in gay men. In 2001, NAM aidsmap was already raising concern that hundreds of cases in syphilis in gay men had been reported in the previous two years, compared with a handful in the mid-90s.
In the UK the proportion of male syphilis cases that were in gay men increased from 58% in 2006 to 85% in 2014. Similarly the proportion of male cases of gonorrhoea that were in gay men increased from 27% in 2008 to 70% in 2015.
This means that even in terms of STI clinic attendees, gay and bisexual men are now 17 times more likely to be diagnosed with syphilis than heterosexual men and nearly six times more likely to be diagnosed with gonorrhoea.
In contrast, while there have been increases in the proportion of other STIs in gay and bisexual men, they still form a minority of cases and increases have been relatively slight. Between 2006 and 2015, the proportion of genital herpes cases that were seen in gay men increased from 5% to 10%, as did genital warts cases, and the proportion of chlamydia cases from 4% to 20%.
In the US the estimated annual incidence of gonorrhoea in gay men was 5% in 2017 – roughly 25 times the general-population incidence, though this varies geographically. On the west coast and in New York there is much more concentration among gay men than in the south. In terms of absolute numbers diagnosed, there were actually slightly fewer cases of gonorrhoea in men than women until 2012, when male cases first outstripped female. In gay men, a slight upward trend in gonorrhoea cases accelerated from 2013 onwards and still maintains an upwards trend.
Surveillance in the rest of western Europe is more haphazard, less complete and less easy to compare between countries. However, a similar pattern to the UK in diagnoses of syphilis in gay men has been seen in the largest countries. For instance there were virtually no cases of syphilis seen in surveys in either the UK, Germany or France in 1998: in 2001 there were 370 in the UK, 420 in Germany and 200 in France; and in 2008, the last year in which directly comparable surveys in gay men were made, cases had increased to 1600 in the UK, 1500 in Germany, and to 500 in France (where reporting is less complete). By 2016 the European Centre for Disease Prevention and Control (ECDC) documented 6470 cases in total (not just in gay men) in the UK, 7175 in Germany and 1742 in France. The ECDC estimated that 66% of these cases were in gay men. There was an upward inflection in cases in gay men documented in western Europe between 2012 and 2015, with a slight slackening off in the rate of increase in 2016, the last year documented.
A note on gonorrhoea resistance
Gonorrhoea resistance causes significant concern and may be one of the most significant medical risks of increased STI prevalence. The Neisseria gonorrhoeae bacterium easily becomes resistant to antibiotics. Originally treated with sulphonamides and then penicillin in the 1940s, gonorrhoea became progressively resistant to those drugs and then to tetracycline and ciprofloxacin, which replaced them.
As the proportion of gonorrhoea resistant to ciprofloxacin in the UK climbed to 50% in gay men and 20% in heterosexuals, a switch was made to a whole new class of antibiotics – the cephalosporins. A single drug from this class, cefixime, was used for first-line treatment of gonorrhoea starting in 2006.
However, the proportion of gonorrhoea with resistance to cefixime rapidly increased in gay men from about 5% in 2008 to 31% in 2010. As a result, in 2011 the recommended therapy changed, to dual combination therapy. This combined another cephalosporin, ceftriaxone, with the macrolide drug azithromycin.
Emphasising this as the only approved regimen for gonorrhoea, and guarding against over-treatment, appeared to have positive results both in the US, where rates of resistance to ceftriaxone fell tenfold between 2011 and 2014, and in Europe, where in the UK and Belgium resistance to both drugs halved between 2011 and 2014. No cases of ceftriaxone-resistant gonorrhoea were detected in Europe in 2016, compared with seven in 2013.
Doctors, however, were faced with the dilemma that up to 3% of patients are allergic to ceftriaxone. To preserve the action of azithromycin, the 2019 British Association for Sexual Health and HIV guidelines reverted back to solo ceftriaxone. At the moment cases of gonorrhoea resistant to ceftriaxone are rare, but higher levels of resistance in south and south-east Asia have led to predictions that increases over here are likely; it is notable that the first cases of multi-drug-resistant gonorrhoea seen in the UK were in heterosexual men who had acquired it in south-east Asia.
Conclusion: trends in STI epidemiology
- The truly exceptional period in the epidemiology of STIs both in the UK and globally was the 1990s, when the fear of HIV and the adoption of safer sex led to historic lows in STI diagnoses.
- Diagnoses of most STIs have been increasing since then, beginning well before the advent of PrEP.
- Diagnoses have yet to reach the historic highs of the late 1970s to early 1980s, though there is no reason to suppose they will not.
- In recent years, diagnoses of three STIs – HPV/genital warts (due to vaccination), hepatitis B (due to vaccination and treatment), and HIV itself (for many reasons) have declined, but all others continue to rise.
- The rises in the bacterial STIs gonorrhoea and syphilis are especially marked, and they are also the two bacterial STIs most concentrated among gay and bisexual men.
- In Europe, syphilis cases, over 80% of them in gay and bisexual men, have risen 50% since 2020 and gonorrhoea cases, two-thirds of them in gay and bisexual men, have more than doubled, from 35,000 to 80,000 a year.
- We will look below at the reasons for this, and whether PrEP contributes to the increases in the most recent years.
Behaviour change and condom use
It is hard to deny that a gradual, long-term decline in condom use from the high rates of the AIDS era was the first and one of the most important reasons for the rise in STIs in recent years. Surveys have shown increases in condomless sex that started almost as soon as HIV combination therapy started reducing mortality from AIDS in the late 1990s.
Although direct comparisons cannot be made from one survey to another (because the characteristics of respondents changed), the Gay Men’s Sex Surveys (GMSS) in the UK are one of the longest longitudinal data sets investigating gay and bisexual men’s risk behaviour in the world.
Starting in 1993, most of the surveys have either asked all respondents whether they had had condomless anal sex during the year or asked that question solely of men who said they had had anal sex. (GMSS has used quite a strict definition of condomless sex; just one instance during the previous year counts as answering ‘yes’, regardless of the actual or perceived HIV status of partners, their viral load, or their relationship status.)
From 1993-1997, roughly 33% of all respondents answered that they had had at least one episode of condomless anal sex during the year. Between 1998 – when antiretroviral therapy became widely available – and 2007 this proportion increased to around 42%. Since then, surveys have been less frequent but in 2008 the figure was 54% and in 2014 64%.
To put it another way, up till 1997 a third of gay men were not using condoms consistently; by 2014 only a third were using them consistently.
In some years the question has been whether gay men who have anal sex have had condomless sex during the year, which excludes the men who had no anal sex and is therefore a higher figure. This question was first asked in 1998, when the proportion was 47%. The proportion was 59% by 2001 and in surveys up to 2010 the figure hovered around the mid-50s. By the next survey, published in 2014, the figure had increased to 76%.
When discussing condoms and whether the decline in their use has been the root cause of increases seen in STIs other than HIV, it is worth noting that condoms are not as effective against other STIs as they are against HIV.
For instance, a prospective study of self-reported condom use in 929 attendees at five US STI clinics in 2012 found that consistent and correct condom use (i.e. no breaking, slipping, leakage, putting on after first insertion, or taking off before last withdrawal) only reduced bacterial STI diagnoses by 60% – lower than the 70-90% reduction rates seen for HIV in similar analyses.
This is partly because most STIs are not spread exclusively via genital contact. As far back as 2002, studies were already showing that a particularly high proportion of gonorrhoea infections in gay men were in the throat (pharyngeal), often asymptomatic, and spread through oral sex. Subsequent research has reaffirmed this. Similarly, estimates of the proportion of syphilis in gay men spread by oral sex or kissing range from 33% to 50%.
In other words, even if it were possible, making condom use routine again might not have the expected impact against other STIs that it might have in HIV.
Using condoms has never been the only measure taken to avoid HIV infection, especially by gay men. As condom use became less frequent, but before biomedical prevention measures including PrEP and treatment as prevention ('U=U') were adopted or had a significant effect, people with HIV and at risk of it were adopting what were called 'seroadaptive' behaviours that attempted to minimise the risk of sex if it was condomless.
Serosorting (restricting condomless sex to partners of one’s own HIV status) was a major behavioural phenomenon in the 2000s. From 2001 onwards, for instance, in San Francisco, while the amount of condomless sex increased in HIV-positive gay men, and stayed level in HIV-negative men, there was a sharp drop in reported condomless sex between partners of different HIV status. This trend was especially notable in men with HIV, in whom condomless sex 'across the serodivide' fell by two-thirds. Similar trends were seen in other surveys, including in HIV-positive women.
The reason serosorting was led by HIV-positive people was because they could be certain of their status. As a number of studies found, for HIV-negative people it was a much less secure way of ascertaining safety. Many people relied on guessing partners’ HIV status – a process that led to increased stigma against the type of person seen as likely to have HIV.
A meta-analysis of serosorting found that while HIV-negative people who serosorted were half as likely to acquire HIV as people who did not, it was the least effective 'seroadaptive' behaviour. One hundred per cent condom use was more effective, and 'seropositioning' was even more effective than that. Seropositioning involves HIV-negative men adopting the insertive role in anal sex and HIV-positive men the receptive role.
What is clear is that since PrEP came along, there has been a complete reversal in gay men’s ideas of what the safest HIV prevention strategy would be. A study from New York, for instance, found that PrEP was now regarded as by far the safest HIV prevention strategy, ahead of U=U and considerably ahead of seropositioning (they were not asked about serosorting).
This change in perception of the safest way to avoid HIV has been accompanied by a considerable reduction in sexual anxiety among gay men using PrEP and a relaxation of serosorting behaviour.
The importance of networks
The fact that gay men are once more willing to 'cross the serodivide' may significantly contribute to the increases seen in STIs in the last few years.
One of the most convincing explanations of why HIV rates are exceptionally high in certain communities – such as black gay men in the USA or African migrants in Europe – despite similar or lower rates of risk behaviour than in their compatriots, is that racism and cultural isolation means that the population largely has sex only with other members of their minority group. This creates a tightly-bound network in which most of the sexually active population is connected closely to other people, including people with HIV – or STIs.
Another example of the power of network effects is the fact that Iceland, of all countries, has the highest rate of chlamydia in the world, due to the often-asymptomatic infection recirculating among the country’s small and relatively isolated population.
HIV serosorting had a similar effect on STIs in gay men. The HIV-positive population was only 10% the size of the negative population but had exceptionally high rates of condomless sex between each other. This led to such a concentration of STIs among HIV-positive gay men that some – such as LGV and sexually transmitted hepatitis C – were almost exclusively seen in HIV-positive men.
However, as soon as serosorting was relaxed, and more HIV-positive and HIV-negative men started having sex with each other, a greater number of infections such as hepatitis C started to occur in HIV-negative men, as has been noted above.
Conclusion: STIs and behaviour change
A rise in STIs following the historic low of the HIV years was therefore probably inevitable, and the reasons for it unsurprising. The rise started just after the millennium, at the time when the availability of antiretroviral therapy first started to reduce people’s fears of the consequences. It has accelerated since 2013-14, when awareness of the efficacy of both PrEP and of treatment as prevention as measures against HIV started to seep into the consciousness of at-risk populations.
The questions to ask at this point are:
- Are these increases all part of an inevitable historic rebound from the days of AIDS?
- Or do PrEP and treatment as prevention have causative roles to play in sustaining, and possibly accelerating, the observed increases in STIs?
- If they do, is this due to phenomena such as the fall in condom use and increased willingness to have serodiscordant sex? Or are other factors at play?
- In particular, do increases in the number and frequency of HIV tests and, alongside them, STI tests, have a part to play? Are we seeing more STIs mainly because we are looking for them more?
- And if so, and if PrEP and awareness of the benefits of immediate HIV treatment lie behind increases in the regularity of STI testing, could they in the longer term have a beneficial effect on STI incidence? Is there any evidence of this happening?
Does the use of PrEP lead to more STIs?
Whether PrEP actually leads to further increases in STIs, or is instead being adopted by people who would already be likely to get them, is unclear on an epidemiological basis.
Some randomised and cohort studies of PrEP have found rises in STIs following the initiation of PrEP. Others have found no increase. And in some, high rates of STIs preceded the adoption of PrEP, suggesting the diagnoses may have been motivators for people to start PrEP.
People taking part in PrEP studies have certainly had very high rates of STIs. The proportion of subjects diagnosed with a bacterial STI during follow-up was 43% during the Ipergay study and 57% during the PROUD study in people taking PrEP. In the DISCOVER study, the annual incidence of bacterial STIs was almost 100% – this does not mean everyone in the study caught an STI but that the annual number of STI diagnoses nearly equalled the number of participants in the study.
However, only one of these studies, PROUD, was designed to detect whether PrEP influenced sexual risk behaviour, by comparing participants who knew they were taking PrEP with ones who knew they were not. The proportion diagnosed with STIs during follow-up was 57% in participants allocated to start PrEP immediately and 50% in those allocated to start it a year later. This difference was not statistically significant, and there was no significant difference between arms for individual STIs either.
A meta-analysis of PrEP studies and rollout programmes in 2019 found that the average annual bacterial STI diagnosis rate among gay and bisexual men taking part in the highest-quality PrEP studies was 84%, with specific rates of 42% for chlamydia, 40% for gonorrhoea and 9.5% for syphilis.
Looking back at previous studies of the relationship between PrEP and STI rates, the first meta-analysis to compare STI rates in studies of gay men taking PrEP and studies of gay men not on PrEP caused considerable concern 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.
A study based on data collected in Seattle from 2014 to 2016 found that while chlamydia rates did increase after men started on PrEP, rates of syphilis and gonorrhoea were elevated 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, a meta-analysis of eight PrEP studies 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.
Meta-analyses average out individual features of studies, and there was actually a wide variation between individual studies. Out of the eight studies, 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 saw a sustained statistically significant increase in STIs between baseline and follow-up. One was the Kaiser Permanente rollout of PrEP in northern California, which saw a 48% increase in STIs following the start of PrEP. 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 tripled (198% increase) after people started PrEP.
It is interesting that this increase was observed in Australia. The researchers themselves drew attention to the “unprecedented rate of enrolment” into PrEP there. 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.
A US study was able to measure STI infections in men 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% before to 10% during PrEP use, and fell back to 2% after stopping it. 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.”
Rises in STIs vary strongly between individuals who start PrEP. An example comes from the PrEPX study in the state of Victoria, Australia, the successor to Vic-PrEP. In people who started PrEP for the first time, the annual incidence of STIs in this study increased by 71%, from 69.5 diagnoses per 100 people a year in the year before starting PrEP to 98.4 diagnoses per 100 people a year in the year after. However, 52% of men did not have an STI during that year: the rise in STIs took place largely in the 25% of men who had two or more, accounting for 76% of infections, including 13% who had three or more infections, accounting for 53%.
The mechanism for the increases in STIs that have been seen does not appear to be a reduction in condom use, at least in people using PrEP, but rather an increase in partner numbers. In the PrEPX study, there was not a statistically significant difference in STI rates according to the frequency of condom use, either with casual or regular partners. The relative lack of significance of condom-use frequency may be due to the fact that condom use was already very low in study participants even before starting PrEP.
The number of partners with whom participants had receptive anal sex was much more significant than condom use. Compared to people who had had one to five such partners in the last six months, people with between eleven and 20 partners were 91% more likely to have an STI, and people with 21 to 50 partners, 117% more likely.
Are more STIs due to more STI testing?
One significant confounder of the observed rise in sexually transmitted infections is that there has also been a significant increase in testing for them. This applies to people receiving PrEP as well as the wider population. It is an increase not only in the overall number of tests but in the appropriate testing of all anatomical sites and in screening of people who do not report symptoms.
More testing, especially of asymptomatic STIs, will at least temporarily inflate the number of diagnoses. So it is important to establish the extent to which the rises in STI diagnoses are rises in incidence: are we finding more STIs primarily because we are testing more often for them?
Gonorrhoea and chlamydia are often both asymptomatic and self-limiting; the immune system can eventually get rid of the infection (though not always, which is what can give rise to serious consequences). Syphilis is not self-limiting and can have serious acute symptoms as well as chronic ones, but its early stages can often be missed.
In the UK, the number of STI tests performed at sexual health clinics increased by 17.5% between 2013 and 2017, with the largest increases in women (20%) and gay men (44%).
In gay men especially, part of this increase may have been due to the demand for testing brought on by PrEP: central London clinics had been offering medical support for people buying PrEP online since spring 2016 and saw the first significant decline in HIV incidence that year. The IMPACT demonstration study, which began in October 2017, would in itself be expected to increase the number of HIV and STI tests performed by gay men by an additional 20 to 25% if all its participants test quarterly and had not done with the same frequency before.
Furthermore, the UK is one of the few countries in Europe which has tested appropriately in the past for rectal, urethral and pharyngeal (throat) infections. For instance, the first EMIS – a large pan-European survey of gay men – asked its participants if they had had an STI test in the last year, and only 30% had. However, many “STI tests” consisted solely of a blood test, which cannot detect chlamydia or gonorrhoea. Fewer than 50% of EMIS’s gay and bisexual male respondents had had a urethral swab and only 16% a rectal swab, indicating that about one in 20 of EMIS’s 180,000 respondents had had a test in the previous year that could detect two of the most common STIs in gay men in every site.
A PrEP demonstration project in New York City found evidence that increased testing rates were contributing to at least some of the rise in STI diagnoses in gay men on PrEP. In this study, from 68 to 83% of the STIs diagnosed were asymptomatic, and the researchers estimated that 24% of STIs would have been missed even if participants had screened for STIs every six months, as US guidelines recommend, instead of every three months as in the study.
In other words, an increase in testing, and especially frequency of testing, often diagnose STIs that would simply have remained undiagnosed before, especially as chlamydia and gonorrhoea infections are often self-limiting.
Going back to the PrEPX study in Australia, the observed, large, increases in STI diagnoses became much smaller when testing frequency was controlled for. Among those starting PrEP, the number of clinic visits increased from 3.2 in the year before PrEP to 4.7 in the year after, and tests for individual STIs increased from 8.5 to 12.9. In multivariate analysis, controlling for testing frequency turned the 71% increase in STI diagnoses to a 21% absolute increase in incidence.
Similarly, an observed increase of 84% in chlamydia turned into an absolute increase in incidence of 38%. After controlling for testing, increases in gonorrhoea (at any site) were no longer significant, and there was no increase in syphilis even before controlling for testing.
While confirming, therefore, that some of the rises in STI incidence seen in Victoria state’s PrEP users were real, it also confirms that most of the rises in incidence were caused by more testing and more diagnosis.
Might frequent testing and treatment reduce STIs?
It is clear that high rates of testing and treatment for HIV are starting to produce substantial falls in HIV incidence in some countries and communities, as the proportion of people with HIV who are infectious shrinks. Might the same apply to STIs?
A couple of mathematical models suggest that, if large numbers of the gay and bisexual men and trans people most at risk of both HIV and STIs are enrolled in PrEP programmes, then the frequent STI testing that should result might lead to substantial drops in STI incidence.
A modelling study presented in 2017 found that if PrEP became widespread among gay men in the US, diagnoses of STIs would rise in the first year, but would fall thereafter.
If the testing interval was once every six months, for instance, the annual incidence of all STIs in all gay men would fall from about 5.4% a year after starting a PrEP programme to 4% after three years, and to less than 2% after ten years.
This was based on an assumption that 40% of all sexually active gay men take PrEP, but also that condom use in gay men on PrEP would fall by 40%. Incidence would decrease further if condom use did not fall.
There was no support for the hypothesis that PrEP and reduced condom use would lead to rises in STIs. In this model, even zero condom use under PrEP could not transform the STI rates seen in the general gay population into the rates seen in PrEP seekers.
A second modelling study based on Dutch data was published in 2019. This found that if 75% of the gay and bisexual men in the Netherlands who met the criteria for PrEP started taking it, and therefore testing for HIV and STIs every three months, this would not only cause a 70% drop in HIV incidence – in both PrEP users and other gay men – but would lead to even greater reductions in gonorrhoea.
In the ‘base case’ of the model, gonorrhoea would almost be eliminated, with a 97% fall in gonorrhoea cases relative to 2018, the year the model assumed the PrEP programme was started.
That would be if condom use did not change. But even if the PrEP programme resulted in a 75% rise in the likelihood of condomless anal sex in gay men, it would nonetheless lead to a 73% drop in gonorrhoea cases. PrEP would not lead to a rise in gonorrhoea unless condom use fell almost to zero.
There is already evidence that higher testing rates can achieve falls in some STIs, notably hepatitis C. One study found that new cases of hepatitis C among HIV-positive gay and bisexual men seen at three clinics in London had declined by nearly 70% since 2015, which was attributed to regular hepatitis C screening and a treatment-as-prevention effect resulting from wider use of direct-acting antiviral (DAA) therapy. A previous study in the Netherlands found the same thing. In Australia, a study among a different population, people who inject drugs, found that testing and DAA treatment had reduced the community prevalence of hepatitis C viremia (a detectable viral load) from 43% in 2015 to 25% in 2017. The British HIV Association has declared that hepatitis C could be eliminated in the UK by 2021 if these trends continue.
Achieving this with the more contagious and often asymptomatic bacterial STIs is more difficult. England has had a national chlamydia screening programme for young people under 25 since 2003. Modelling showed that between 26% and 43% of 16 to 24 year olds would have to be tested and treated for chlamydia yearly to have an effect in infections. A 2009 report by the National Audit Office found that only 50% of primary care trusts were achieving 26% coverage, and since then coverage has declined and the proportion of young people testing positive for chlamydia has increased, from 7.5% in 2013 to 9% now.
However, it is arguable that high rates of testing and treatment might have a more positive effect in smaller, more targeted populations with very high background incidence. At London’s 56 Dean St, the largest sexual health clinic in Europe, there was a huge increase in gonorrhoea testing from 3000 in the fourth quarter of 2013 to 11,000 in the third quarter of 2015. This was followed by a decline from 1600 gonorrhoea diagnoses during that quarter to 1100 per quarter for the next seven quarters, up until June 2016, despite testing rates staying at the same level. This apparent decline in gonorrhoea infections was to some extent reflected in a national decline in gonorrhoea diagnoses from 22,000 to 17,000 in 2016.
However, this decrease was not sustained. Gonorrhoea diagnoses started increasing again, both at Dean St and nationally, in the second half of 2017. Although resistant infections may have had a part to play, this was most likely because clinics found it hard to cope with a surge in demand for appointments coupled with NHS cuts that reduced the number available.
If 'test and treat' against STIs is not a complete answer, how about adding in 'STI PrEP'? So far there have only been a couple of studies of the idea of giving gay men at very high risk of STIs antibiotic PrEP in the form of the cheap and safe antibiotic doxycycline, already used widely for malaria prophylaxis.
In a pilot study with 30 gay men in the US, there were 70% fewer STIs in the men given doxycycline. In a later study in France, 233 gay men took doxycycline for an average of 8.7 months. There was a fall of 73% in cases of syphilis and 70% in chlamydia – but no fall in gonorrhoea, possibly due to the presence of circulating doxycycline-resistant gonorrhoea.
- There is no doubt that there has been a considerable increase in STIs since the historic lows of the 1990s. The exceptions are in HIV itself and in the two STIs for which we have vaccines – hepatitis B and HPV.
- Gonorrhoea and syphilis, as well as some less common conditions such as LGV, have become proportionally more common in gay and bisexual men, in the UK, the US and in western Europe.
- A number of influences have contributed to the rises in STIs and may have done so sequentially. Fewer gay men employed 100% condom use as their primary HIV prevention method. Also, ‘seroadaptive’ methods such as serosorting and seropositioning became more popular. These may have had some efficacy against HIV but less against STIs.
- While there has been an upward inflection in the already steady increase in STI diagnoses in gay men since 2010-2013, it is impossible to ascribe this to PrEP. Any PrEP-specific effect would be unlikely to be seen until use became widespread, as it has done only even in the USA only in the last two to four years.
- A combination of factors ranging from falls in condom use to a relaxation of sexual segregation between HIV-positive and HIV-negative gay men is more likely to have been the cause, as these trends have been around for longer.
- There is some evidence that a proportion of gay men starting PrEP do experience an additional rise in STI infections, especially of syphilis and gonorrhoea. One study has found that this was associated with a rise in the number of partners with whom they have receptive anal sex.
- However, only a minority of gay men experience this increased incidence of STIs and it is difficult to disentangle cause and effect; does PrEP have a disinhibitory effect on risk behaviour, or is it sought because the person sees that their risk behaviour and HIV risk is rising, or is likely to?
- Another complicating factor is that STI testing rates have increased in all populations and especially in gay men. The PrEPX study found that most – though not all – of the increase in STI diagnoses after men started PrEP was due to more STI testing.
- The potential exists for more testing and treatment of STIs to lead to falls in new STIs, as more people get treated sooner after infection. This has already been observed with hepatitis C.
- As PrEP use requires engagement with sexual health services, PrEP might facilitate this happening.
- Although condoms are not as fully effective against most STIs as they are for HIV, they remain the most effective preventative strategy, at least for rectal and vaginal infections.
A final thought. PrEP’s role in reducing sexual anxiety has been cited as a significant positive effect by PrEP users, especially gay men, in a number of studies. Thought needs to be given therefore to how to develop messages that lead to behaviours protective against HIV without re-instating sexual anxiety among people at risk of HIV.