Article Text

Download PDFPDF

Original article
Does HIV pre-exposure prophylaxis use lead to a higher incidence of sexually transmitted infections? A case-crossover study of men who have sex with men in Los Angeles, California
  1. Matthew R Beymer1,2,
  2. Michelle A DeVost2,
  3. Robert E Weiss3,
  4. Rhodri Dierst-Davies4,
  5. Chelsea L Shover2,
  6. Raphael J Landovitz1,5,
  7. Corinne Beniasians2,
  8. Ali J Talan2,
  9. Risa P Flynn2,
  10. Robyn Krysiak2,
  11. Kayla McLaughlin2,
  12. Robert K Bolan2
  1. 1 David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, CA, USA
  2. 2 Health and Mental Health Services, Los Angeles LGBT Center, Los Angeles, California, USA
  3. 3 Department of Biostatistics, Jonathan and Karin Fielding School of Public Health, University of California, Los Angeles, California, USA
  4. 4 Deloitte Financial Advisory Services LLP, New York, New York, USA
  5. 5 Center for Clinical AIDS Research and Education (CARE), Los Angeles, California, USA
  1. Correspondence to Matthew R Beymer, Los Angeles LGBT Center, Los Angeles, CA 90028–6213, USA; mbeymer{at}


Background Pre-exposure prophylaxis (PrEP) is an effective method for reducing HIV incidence among at-risk populations. However, concerns exist over the potential for an increase in STIs following PrEP initiation. The objective of this study is to compare the STI incidence before and after PrEP initiation within subjects among a cohort of men who have sex with men in Los Angeles, California.

Methods The present study used data from patients who initiated PrEP services at the Los Angeles LGBT Center between October 2015 and October 2016 (n=275). A generalised linear mixed model was used with a case-crossover design to determine if there was a significant difference in STIs within subjects 365 days before (before-PrEP period) and 365 days after PrEP initiation (after-PrEP period).

Results In a generalised linear mixed model, there were no significant differences in urethral gonorrhoea (P=0.95), rectal gonorrhoea (P=0.33), pharyngeal gonorrhoea (P=0.65) or urethral chlamydia (P=0.71) between periods. There were modest increases in rectal chlamydia (rate ratio (RR) 1.83; 95% CI 1.13 to 2.98; P=0.01) and syphilis diagnoses (RR 2.97; 95% CI 1.23 to 7.18; P=0.02).

Conclusions There were significant increases in rectal chlamydia and syphilis diagnoses when comparing the periods directly before and after PrEP initiation. However, only 28% of individuals had an increase in STIs between periods. Although risk compensation appears to be present for a segment of PrEP users, the majority of individuals either maintain or decrease their sexual risk following PrEP initiation.

  • sexual health
  • gay men
  • prophylaxis
  • syphilis
  • chlamydia infection
View Full Text

Statistics from


Pre-exposure prophylaxis (PrEP) is a biomedical prevention strategy where HIV-negative individuals take a once-daily antiretroviral pill to prevent contracting HIV. Multiple randomised controlled trials have demonstrated that PrEP can effectively reduce the incidence of new HIV infections between 44% and 75% among different populations of high-risk uninfected individuals.1–6 The first large cohort study of PrEP efficacy found a 44% reduction in the incidence of HIV between the treatment and placebo groups among persons engaging in high-risk sexual behaviour.1 A second study found that the incidence of HIV was 1.2 infections per 100 person-years in the treatment group versus 9 infections per 100 person-years in the group that received deferred PrEP.7

The success of PrEP in these large-scale trials prompted the US Centers for Disease Control and Prevention (CDC) to recommend PrEP initiation for individuals at increased risk for HIV exposure, including sexually active, HIV-negative gay, bisexual and other men who have sex with men (MSM). The CDC indicates PrEP use for MSM meeting at least one of the following criteria in the past 6 months: (1) engaging in any condomless anal intercourse; (2) having had any diagnosis of an STI; and/or (3) being in an ongoing sexual relationship with an HIV-positive partner.8

An analysis by the CDC found that 25% of HIV-negative MSM in the USA between the ages of 18 and 59 met the CDC guidelines for PrEP use.9 However, the latest published data show that only between 4% and 10% of MSM report using PrEP.10 11

Slow adoption of PrEP may be due to numerous factors including concerns held by both potential PrEP users as well as by prescribing physicians. Reservations about taking PrEP at the patient level include both short-term and long-term drug side effects,12–15 out-of-pocket costs,12 13 15 16 the possibility for drug resistance given PrEP continuation following HIV infection13 and fear of being stigmatised as promiscuous.17

A recent survey of physicians that asked about their willingness to prescribe PrEP reported that 96% of respondents noted concerns about patient adherence to the regimen.18 Other studies of physician attitudes have cited similar adherence concerns19 20 as well as concerns that PrEP users may be more likely to forego using condoms once on PrEP, leading to higher incidence of other STIs.19 21

The concept of exhibiting greater sexual risk after adoption of a safety measure like PrEP is known as risk compensation or risk homeostasis. A study by Brooks et al in 2012 found that 60% of HIV-negative MSM surveyed (n=25) said they would either decrease or abandon condom use with the adoption of PrEP.22 However, other studies have found contrary results, with potential PrEP users stating that they would not be likely to reduce condom use.23 24 Additionally, a large cohort study of PrEP users over the first 9 months of use found that users reported the same amount or fewer sex partners over follow-up when compared with baseline.25 While self-reported intentions following PrEP use and risk behaviour assessments are informative, measuring biomarkers like STIs may be a more objective indicator of risk compensation among PrEP users.

Three studies have examined STIs among PrEP users. McCormack et al 7 compared the STI incidence between PrEP users and non-PrEP users and found no differences between the groups. A second study found that after 6 and 12 months of PrEP use, 30% and 50% of PrEP users had been diagnosed with any STI, respectively.26 A third study found that 35% of PrEP users were diagnosed with an STI during follow-up.27

These studies were limited in that they only examined STI rates after PrEP use and therefore did not control for STI rates among the same individuals prior to initiating PrEP. By comparing STIs within subjects, between-person confounders are greatly minimised and a more rigorous assessment of STI changes can be performed. The primary objective of this study is to use a case-crossover design to determine if there is a difference in STIs in the periods before and after PrEP for each subject analysed. The secondary objective is to determine if there are any behavioural differences between individuals who exhibit an increase in STIs versus individuals who demonstrate either no change or a decrease in STIs in the periods before and after PrEP initiation.

Materials and methods

The Los Angeles LGBT Center (the Center) is a Federally Qualified Health Center located in the Hollywood neighbourhood of Los Angeles, California. The Center offers HIV/STI testing, transgender healthcare services, primary healthcare, HIV care and biomedical prevention services (eg, PrEP). The Center began prescribing PrEP through its West Hollywood (Center WeHo) satellite location in October 2015.

All patients who receive HIV/STI testing through the Center WeHo are administered a baseline risk assessment to determine demographics, recent sexual behaviour, substance use behaviours and knowledge of non-occupational postexposure prophylaxis (PEP) and PrEP. For those who do not report already taking PrEP, HIV/STI testing counsellors provide education about PrEP and ask patients about both their self-perceived candidacy for PrEP and their likelihood to initiate PrEP.

Patients who indicate a desire to initiate PrEP are referred to a PrEP linkage coordinator to set up the initial appointment with a PrEP provider. Each patient is evaluated to rule out any medical contraindications to PrEP, including prevalent or incident HIV infection. Labs are subsequently taken for STIs (gonorrhoea, chlamydia and syphilis), hepatitis B, hepatitis C and renal function. Patients are educated about medication adherence and side effects, and a 30-day prescription is written for PrEP, following provider approval. To continue receiving PrEP prescriptions, patients are required to return 30 days after the baseline visit, and every 3 months thereafter for HIV and STI testing, lab review to measure renal function and review of medication adherence.

At baseline and follow-up visits, patients are asked to self-collect urine and rectal specimens for gonorrhoea and chlamydia testing. A lab technician then swabs the throat to test for gonorrhoea and takes a blood sample to test for syphilis, HIV and renal function. The blood sample is also used to test for HIV antibodies through an HIV rapid test. Individuals who test HIV antibody negative have a nucleic acid amplification test which is sent for acute HIV screening. Individuals who are positive for any STI except HIV are called back for treatment typically within 2–7 days. Individuals who test positive for HIV are taken off PrEP and immediately linked to HIV care by an HIV linkage-to-care specialist.

The inclusion criteria for this analysis are: (1) MSM; (2) initiation of a PrEP regimen at the Center’s Sexual Health and Education Program between October 2015 and October 2016; (3) at least one STI testing visit within 365 days prior to PrEP initiation; (4) at least one STI testing visit within 365 days after PrEP initiation; and (5) no previous reported use of PrEP or PEP before initiating PrEP at the Center. An MSM is defined as an individual who was assigned a male sex at birth, reported that their gender identity is male, and either reported a sexual orientation of gay or bisexual or reported sex with a man at the last sexual encounter. Follow-up STI testing data were available up to the end of May 2017.

Statistical methods

This study uses a case-crossover design with an intent-to-treat analysis to assess changes in STI incidence between the period prior to PrEP initiation (before-PrEP) and the period after PrEP initiation (after-PrEP). The case-crossover design has each individual serving as their own control (also known as an interrupted time series design). The before-PrEP period consists of any STI testing visits in the 365 days prior to PrEP initiation as well as the STI testing visit during the PrEP enrolment visit. The after-PrEP period consists of the STI testing visits up to 365 days after PrEP initiation. However, not all individuals have 365 days of data as only those who started PrEP in May 2016 or earlier could have complete data. An intent-to-treat design was used since reliable biological measures of PrEP adherence (intraerythrocytic tenofovir diphosphate levels in dried blood spots) were not taken for individuals receiving PrEP.

We analyse all bodily sites/STIs in a single model. A generalised linear mixed model (GLMM) with a person-random effect is used to determine the rate ratio of STI incidence in the period after PrEP to the period before PrEP. Predictors in the model include the period variable (before-PrEP or after-PrEP) and the bodily site/STI tested (eg, urethral gonorrhoea, rectal chlamydia) and the interaction between these two variables. The effects of interest are the STI rate ratio of the STI rate after PrEP initiation to the STI rate before PrEP initiation. The outcomes for each person contained results of STI testing for each possible bodily site/infection combination: urethral gonorrhoea, rectal gonorrhoea, pharyngeal gonorrhoea, urethral chlamydia, rectal chlamydia and syphilis. The GLMM was a log link random intercept Poisson model with log of time since last visit as an offset (subsequently referred to as the Poisson GLMM). We included age group, race/ethnicity and sexual orientation in the base model to estimate STI rate differences across demographic groups.

We fit a second GLMM model with indicators of the before-PrEP and after-PrEP periods as covariates, and the number of visits in each period as the outcome, to determine if detection bias was potentially present. Lastly, we compared individuals who had an increase in STIs with individuals who had no change or a decrease in STIs to determine if there was a difference in either self-reported substance use or the number of sexual partners in the past 3 months. All analyses were performed using SAS V.9.4 (SAS Institute).


Two hundred and seventy-five individuals met the inclusion criteria for analysis. The majority of the sample was between the ages of 30 and 39 at baseline, White or Latino, gay and reported having a college degree or higher (table 1).

Table 1

Demographics of men who have sex with men administered PrEP at the Los Angeles LGBT Center (n=275), October 2014 to May 2017

The before-PrEP period had a lower number of STI testing visits (n=755) when compared with the after-PrEP period (n=908) (table 2). Overall gonorrhoea prevalence decreased from 11.7% in the before-PrEP period to 10.3% in the after-PrEP period. In contrast, syphilis prevalence increased from 1.5% to 3.5%. Chlamydia prevalence remained unchanged at 9.6% and prevalence for gonorrhoea, chlamydia and/or syphilis (any STI) remained the same at approximately 20%. Similar results were seen for gonorrhoea, chlamydia and syphilis incidence (table 3).

Table 2

Prevalence of STIs in the 365 days before and after PrEP initiation (n=275), October 2014 to May 2017

Table 3

Incidence of STIs in the 365 days before and after PrEP initiation (n=275), October 2014 to May 2017

In the Poisson GLMM, there were no significant differences in urethral gonorrhoea (P=0.95), rectal gonorrhoea (P=0.33), pharyngeal gonorrhoea (P=0.65) or urethral chlamydia (P=0.71) between periods (table 4). There was a 29% increase in rectal chlamydia (P=0.01) and 164% increase in syphilis diagnoses (P=0.02). There were no significant differences in STIs by age group (P=0.39), race/ethnicity (P=0.14) or sexual orientation (P=0.92). In a second model, there was a significant increase in any STI from the before-PrEP to the after-PrEP periods (rate ratio=1.36; 95% CI 1.06 to 1.74; P=0.02).

Table 4

Poisson generalised linear mixed model comparing each STI in the 365 days before PrEP and the 365 days after PrEP (n=275), October 2014 to May 2017*

To determine if detection bias was present in this analysis (ie, detection of more STIs as an artefact of more STI testing), we used a second GLMM to determine if there was a difference in the number of visits per unit time between periods. There was a modest 7% decrease in the average number of STI testing visits from the before-PrEP period (mean=2.72; median=2; SD=1.7) to the after-PrEP period (mean=2.52; median=2; SD=1.7) (P<0.0001).

Of the 275 unique individuals in the analysis, the plurality was not diagnosed with STIs in either the before or after period (38%) followed by individuals who had an increase in STIs (28%), a decrease in STIs (24%) and the same number of STIs in the before and after periods (10%). When comparing individuals who had an increase in STIs and individuals with no STIs, a decrease, or no change, there were no differences by self-reported methamphetamine use (P=0.59), nitrates use (P=0.42), ecstasy use (P=0.15) and number of sexual partners in the past 3 months (P=0.34).


Our study showed that there were no significant differences in urethral gonorrhoea, rectal gonorrhoea, pharyngeal gonorrhoea or urethral chlamydia between the before-PrEP and after-PrEP periods. In contrast, there was a 29% and 164% increase in rectal chlamydia and syphilis, respectively, between periods. When analysing any STI as an outcome, there was a significant overall increase in STIs between periods. To the best of our knowledge, this is the first study to compare each patient’s STI testing results before and after PrEP initiation to determine if there was a true change in STIs.

The increase in syphilis between periods may be partly explained by two factors. First, syphilis chancres may be in areas not covered by a condom during intercourse, and therefore transmission may occur despite a condom being used.28 Second, this finding is consistent with an overall upward trend in syphilis seen in the population of patients testing for STIs at the Center who are not using PrEP. In the year before PrEP was prescribed at the Center (November 2014 to October 2015), syphilis positivity among non-PrEP users was 2.9%. In the next year (November 2015 to October 2016), syphilis positivity increased to 3.4% among non-PrEP users. Therefore, there appears to be an overall increase in syphilis diagnoses, regardless of PrEP use.

The increasing overall trend was not observed in the general population for rectal chlamydia in the years before and after PrEP initiation at the Center. Rectal chlamydia in the periods before and after PrEP availability at the Center in the non-PrEP population remained consistent at 9.8%. Further studies in other jurisdictions should replicate these analyses to determine if our finding of an increase in rectal chlamydia is an aberration or an actual trend.

Our study has notable limitations. Although we included all STI testing visits up to 365 days after PrEP initiation, a full 365-day period was not available for all patients included in the analysis since the analysis period ended in May 2017. For example, an individual who began PrEP in October 2016 would only have contributed 8 months to the after-PrEP period. However, our protocols require patients to return at 3-month intervals to refill their PrEP prescription, and a majority of individuals included had at least two STI testing visits in the after-PrEP period. Furthermore, we adjusted for time under observation by fitting incidence models. A second limitation is that the results may have been different had we included individuals who had previously used PEP or PrEP. We hypothesised that individuals who have taken PEP or PrEP in the past are fundamentally different because they have prior experience with the medication. A third limitation is that we were unable to account for individuals who initiated PrEP and failed to take their medication yet still had STI testing. While this scenario is possible, our PrEP providers indicated that this was rare in the population included. A fourth limitation is that the case-crossover study design is observational and non-randomised. Furthermore, it is possible that changes in STI incidence may be attributable to factors beyond PrEP use. Lastly, the Center tests all PrEP patients for pharyngeal chlamydia, but this testing is not standard testing for non-PrEP patients. Therefore, we were unable to analyse changes in pharyngeal chlamydia since no data in the before-PrEP period were available.

Our study also had important strengths. The first strength is that this is the first study, to our knowledge, that had STI testing data both before and after PrEP initiation on the same individuals. Thus, we were able to use an individual’s STI testing data before PrEP as a control to determine if STIs increased after starting PrEP.

The second strength of our study was the design of the PrEP protocol, which required patients to return at 3-month intervals to renew their PrEP prescription. Regularly testing at 3-month intervals provides more complete STI data collection, and it assures a better standard of care for the patient. Data from PrEP users receiving services at Callen-Lorde Community Health Center showed that 77% and 68% of STIs at 3 and 9 months, respectively, after starting PrEP would have been missed if providers relied solely on symptom assessment.29 Although the CDC recommends testing for STIs at 6-month intervals, the data from the current study and the study conducted at the Callen-Lorde Community Health Center show that 3-month intervals will help ensure timely treatment of STIs. Provided other clinics decide to implement similar PrEP protocols, mandatory testing as a condition for continuation of PrEP use will help protect both the patient and their partners from preventable STI morbidity.

Importantly, we found a non-monotonic trend in STI incidence between periods. A small segment of PrEP users had a higher number of STIs in the year following PrEP initiation (28%), a finding suggestive of risk compensation, while the remainder either had the same number of STIs as before or had a reduction in STIs, although this may be partly explained by the somewhat limited follow-up after PrEP initiation as compared with the longer run-up prior to PrEP initiation. While we were not able to detect differences by demographics, substance use or number of sexual partners, there may be unmeasured behavioural predictors (eg, percentage of sexual experiences where a condom was used) or psychological factors (eg, sexual compulsivity, depression) that distinguish PrEP users who contract a greater number of STIs after PrEP initiation from PrEP users who maintain or lower their sexual risk profile. Future studies should analyse the roles of substance use, sexual compulsivity and partner networks among PrEP users to more closely determine what factors lead to a higher incidence of STIs following PrEP initiation.

PrEP is an effective method for reducing the incidence of HIV, but vigilance is needed to keep the threat of other STIs at bay. As PrEP use becomes more widespread, prompt identification and treatment of STIs is necessary to prevent forward transmission of STIs. Expansion of PrEP access, identification of those at highest risk for STI acquisition, STI testing requirements at 3-month intervals and continuing to identify early HIV infections to quickly initiate treatment could comprise an effective HIV and STI strategy in this new era of biomedical prevention.

Key messages

  • Among men who have sex with men (MSM) pre-exposure prophylaxis (PrEP) users in Los Angeles, there was no change in gonorrhoea or urethral chlamydia from the year before to the year after PrEP initiation.

  • There was a modest increase in syphilis and rectal chlamydia from the year before to the year after PrEP initiation.

  • There was a non-monotonic trend in STIs in the periods before and after PrEP initiation among MSM in Los Angeles.


View Abstract


  • Handling editor Catherine H Mercer

  • Contributors MRB: study concept, data cleaning, design, analysis, writing and revisions. MADV: study concept, data cleaning, design, analysis and manuscript review. REW: study concept, design, analysis and manuscript review. RDD, CLS, RJL, CB: study concept and manuscript review. AJT, RPF, RK, KM: manuscript review. RKB: study concept, design, writing and manuscript review.

  • Funding MRB was supported by the UCLA Postdoctoral Fellowship Training Program in Global HIV Prevention Research (Currier and Gorbach, PIs); T32MH080634. REW was supported by the Center for HIV Identification, Prevention, and Treatment (CHIPTS) NIMH grant P30MH058107; the UCLA Center for AIDS Research (CFAR) grant 5P30AI028697, Core H.

  • Competing interests None declared.

  • Ethics approval The study received approval from the University of California, Los Angeles South General Institutional Review Board (SGIRB) (IRB number: 00004474; Project number: 16-000452) .

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles