Global Epidemiologic Characteristics of Sexually Transmitted Infections Among Individuals Using Preexposure Prophylaxis for the Prevention of HIV Infection

1Preexposure prophylaxis (PrEP; emtricitabine and tenofovir disoproxil fumarate) for the prevention of HIV infection is safe and effective when there is a high level of adherence.,,, The World Health Organization recommends the use of PrEP in subpopulations at substantial risk of HIV (ie, incidence >3 per 100 person-years). Operationally, this means that PrEP services are prioritized for men who have sex with men (MSM) in all world regions. Preexposure prophylaxis is also offered to the HIV-negative partner in HIV-serodiscordant partnerships as a bridge to viral suppression in several countries. In countries in East and Southern Africa with a high burden of HIV, PrEP services are provided for sex workers or for young women when the epidemiologic characteristics warrant. There is increasing interest and investment in implementing PrEP in low- and middle-income countries (LMICs) by large donors, such as the US President’s Emergency Plan for AIDS Relief and the Global Fund to Fight AIDS, Tuberculosis and Malaria. The Bill and Melinda Gates Foundation and Unitaid have also made substantial investments in PrEP in LMICs. However, recent estimates of the global burden of sexually transmitted infections (STIs) stress the need to consider programs that could address the synergistic epidemic of HIV and STIs.

2Global guidelines dictate that PrEP programs focus on people at substantial risk for HIV, who are the same population at risk for other STIs. With growing interest in PrEP, more members of key populations are motivated to engage with health care systems than ever before. This change provides a unique opportunity to package PrEP services with more comprehensive sexual and reproductive health services at a moment of peak receptivity, particularly in LMICs where such services are currently limited. This plan is consistent with the World Health Organization Sustainable Development Goals to end the HIV epidemic and other communicable diseases, to improve sexual and reproductive health, and to achieve universal health coverage.

3In recent years, access to PrEP has shifted from provision in the context of demonstration projects to wider implementation through national health systems. To synthesize the latest available data to inform policies and practice around the provision of STI services within PrEP programs, we conducted a systematic review to estimate the prevalence and incidence of STIs among PrEP users. We supplemented data from the systematic review with data from key PrEP implementers who provided unpublished STI data. Previous systematic reviews have aimed to compare STI rates among PrEP users and nonusers, focused only on MSM, used data almost exclusively from high-income countries (HICs), and had limited search strategies.,, Since those reviews, an expanding body of PrEP studies from LMICs provides additional data. Unlike previous reviews, we aimed to describe the STI burden among PrEP users to highlight the potential lost opportunities if STI services are not provided for individuals initiating PrEP as well as persistent PrEP users. In particular, we contribute to the literature by providing pooled estimates according to anatomical site (ie, pharyngeal, genital, or anal site) that are valuable for informing STI testing recommendations and cost-effectiveness analyses.

4This review was conducted in 2 stages. First, a systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) checklist (PROSPERO registration: CRD42018116721). Second, a contact list of 82 PrEP implementers and/or researchers provided by the World Health Organization and some of us (J.J.O., J.D.T., F.T.-P., I.H.-M., and P.M.) was used. An email invitation to contribute unpublished STI data was sent to individuals on the contact list with a follow-up email 1 week later if there was no response. No financial incentives were offered for contributing the data.

5We followed the guidelines in the Cochrane Handbook for Systematic Reviews of Interventions, version 5.1. The following 9 databases were searched from inception to November 20, 2018, without language restriction: Ovid MEDLINE (and In-Process and Other Nonindexed Citations and Daily), Ovid Embase, Ovid Global Health, Ovid EconLit, EBSCO CINAHL Plus, EBSCO Africa-Wide Information, Web of Science Core Collection, VHL LILACS, and Ovid Northern Light Life Sciences Conference Abstracts. The 2 key concepts anchoring our search strategy were STIs and PrEP (full details in eAppendix 1 in the Supplement). We included data from routine implementation programs (PrEP, prospective cohorts, randomized clinical trials, or demonstration projects of oral PrEP) that reported at least 1 of the following: frequency of STI testing and laboratory-confirmed STI positivity (incidence or prevalence). We included data from key STIs: Chlamydia trachomatis; Neisseria gonorrhoeae; Treponema pallidum; Trichomonas vaginalis; Mycoplasma genitalium; hepatitis A, B, and C; and herpes simplex virus. We excluded systematic reviews, letters, editorials, studies using only qualitative research methods, duplicated results from the same study, laboratory studies about testing STI diagnostic performance, and studies restricting study populations by clinical outcomes (eg, men with urethritis or women with cervicitis). We manually searched the references of existing systematic reviews,, to ensure our search strategy included all relevant articles. Once duplicates were removed, the titles and abstracts of articles were independently screened by at least 2 reviewers (M.K.S. and V.A.) according to a list of eligibility criteria; disagreements were discussed with 1 of us (J.J.O.). Data were reviewed by 1 of us (J.J.O.) for consistency and accuracy. Variables used for the data extraction are summarized in eAppendix 2 in the Supplement. We obtained missing data from articles of interest by contacting the corresponding authors. We emailed PrEP implementers to request data related to STI prevalence and/or incidence. Unpublished data were included if they fulfilled the same inclusion criteria, and at the time of request, these data have not yet been published or incorporated into existing publications.

6Baseline prevalence was defined as STI diagnoses within 3 months of starting PrEP and confirmed by laboratory test results. Incidence was defined as STI diagnoses while the individual was taking PrEP and calculated as the number of new laboratory-confirmed STI cases divided by the total duration of exposure to PrEP, calculated as cases per 100 person-years. We extracted reported incidence rates and their 95% CIs when provided. If unavailable, we calculated the incidence by dividing the reported numbers of STI cases and time at risk, and we manually calculated the 95% CIs using the delta method to derive log rates and SEs. When time at risk was not available, we contacted authors for these data and excluded articles when we could not confidently measure STI prevalence or incidence.

7Random-effects meta-analysis was used to calculate across-study pooled estimates of STI prevalence and STI incidence to account for sampling error and heterogeneity. Pooled estimates and 95% CIs were generated using a Freeman-Tukey–type double arcsine transformation to adjust for variance instability. Statistical heterogeneity between studies was assessed with the I2 statistic. Predefined subgroup meta-analyses were based on the following covariates: anatomical site (oropharyngeal, anorectal, or genital), study populations (MSM only or mixed [MSM and non-MSM]), type of study (observational or experimental), and country income level (HIC or LMIC). Observational studies include settings in which there may be additional user costs for STI testing (but could also be paid through a private insurance company, national health insurance, or from philanthropic groups) and thus may result in less systematic STI screening. Experimental studies follow a predefined study protocol for STI testing and thus may have more systematic STI screening. High-income country was defined as any country with a gross national income per capita of US $12 056 or more in 2017. Random-effects metaregression models were conducted to examine the association of these variables with the effect size. Funnel plots were generated to assess for the possibility of small-study effects that may be associated with publication bias. The Egger test was performed to confirm the presence of this bias. All analyses were conducted using Stata, version 13.1 (StataCorp LLC). We evaluated the methodological quality using the Joanna Briggs Institute critical assessment tool for prevalence and incidence studies. A score of 5 (out of 10) or above was deemed to be of sufficient quality to be included in the review.

8Of 3325 articles identified, 88 (71 published and 17 unpublished) met the inclusion criteria for prevalence and incidence data (Figure 1). Table 1 summarizes the characteristics of these studies: data came from 26 countries, mostly from HICs (62 [70%]) and from MSM-only programs (58 [66%]). Table 2 provides more data on included studies, all of which were deemed to be of sufficient methodological quality as determined by the Joanna Briggs Institute tool (ie, score of ≥5).,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, A summary of the countries that provided data is shown in Figure 2.

9Table 3 shows that, among studies reporting a composite outcome of any chlamydia, gonorrhea, and early syphilis, the pooled prevalence was 23.9% (95% CI, 18.6%-29.6%). The prevalence of chlamydia or gonorrhea by anatomical site was highest in the anorectum (chlamydia, 8.5% [95% CI, 6.3%-11.0%]; gonorrhea, 9.3% [95% CI, 4.7%-15.2%]) compared with genital sites (chlamydia, 4.0% [95% CI, 2.0%-6.6%]; gonorrhea, 2.1% [95% CI, 0.9%-3.7%]) and oropharyngeal sites (chlamydia, 2.4% [95% CI, 0.9%-4.5%]; gonorrhea, 4.9% [95% CI, 1.9%-9.1%]). The forest plots for the pooled prevalence by subgroups are provided in eAppendix 3 in the Supplement. For example, the prevalence of chlamydia differed by study population (MSM, 6.9% [95% CI, 5.4%-8.6%]; mixed, 10.7% [95% CI, 0%-38.0%]), study type (observational, 7.9% [95% CI, 5.6%-10.4%]; experimental, 3.1% [95% CI, 1.1%-6.1%]), and country income level (HIC, 7.5% [95% CI, 5.7%-9.6%]; LMIC, 6.6% [95% CI, 2.2%-12.8%]).

10In studies that reported a composite outcome of any chlamydia, gonorrhea, and early syphilis, the pooled incidence was 72.2 per 100 person-years (95% CI, 60.5-86.2 per 100 person-years). The incidence of chlamydia or gonorrhea by anatomical site was highest in the anorectum (chlamydia, 29.9 per 100 person-years [95% CI, 24.1-37.1 per 100 person-years]; gonorrhea, 21.6 per 100 person-years [95% CI, 16.4-28.4 per 100 person-years]) compared with genital sites (chlamydia, 10.4 per 100 person-years [95% CI, 9.2-11.8 per 100 person-years]; gonorrhea, 9.9 per 100 person-years [95% CI, 8.3-11.8 per 100 person-years]) and oropharyngeal sites (chlamydia, 4.6 per 100 person-years [95% CI, 3.3-6.3 per 100 person-years]; gonorrhea, 19.7 per 100 person-years [95% CI, 16.0-24.3 per 100 person-years]). Compared with oropharyngeal chlamydia, the reported incidence of oropharyngeal gonorrhea was significantly higher. The forest plots for the pooled incidence by subgroup are provided in eFigures 1 to 11 in the Supplement (eAppendix 3 in the Supplement). The incidence of chlamydia differed by study type (observational, 22.4 per 100 person-years [95% CI, 18.6-27.0 per 100 person-years]; experimental, 17.0 per 100 person-years [95% CI, 8.7-33.3 per 100 person-years]) and country income level (HIC, 22.1 per 100 person-years [95% CI, 18.5-26.5 per 100 person-years]; LMIC, 8 per 100 person-years [95% CI, 5.6-11.5 per 100 person-years]).

11A few observations from the metaregression results are notable (eTables 1-7 in the Supplement). The prevalence of gonorrhea was higher in studies that enrolled MSM only (adjusted odds ratio [AOR], 1.11 [95% CI, 1.00-1.22]) compared with studies also containing non-MSM populations (eTable 2 in the Supplement). The incidence of chlamydia was higher in the anorectum (AOR, 7.25 [95% CI, 4.83-10.86]) and genital sites (AOR, 2.20 [95% CI, 4.83-10.86]) than in oropharyngeal sites, and it was higher in HICs (AOR, 4.92 [95% CI, 2.35-10.32]) than in LMICs (eTable 5 in the Supplement). Visual inspection of the funnel plots and the Egger test found an indication of small-study effects, with underestimation of the true chlamydia incidence rate (eFigure 7 in the Supplement). The incidence of gonorrhea was lower in genital sites than in oropharyngeal sites (AOR, 0.50 [95% CI, 0.32-0.77]), and it was higher in HICs than in LMICs (AOR, 7.03 [95% CI, 2.62-18.88]; eTable 6 in the Supplement). Visual inspection of the funnel plots and the Egger test found an indication of small-study effects, with underestimation of the true gonorrhea incidence rate (eFigure 6 in the Supplement). The incidence of early syphilis was higher in HICs (AOR, 3.93 [95% CI, 1.36-11.41]) than in LMICs (eTable 7 in the Supplement). Visual inspection of the funnel plots and the Egger test found an indication of small-study effects, with underestimation of the true early hepatitis C incidence rate (eFigure 11 in the Supplement).

12This systematic review and meta-analysis consolidates the published and unpublished evidence of the high STI burden among individuals initiating PrEP as well as among persistent PrEP users. Our findings underscore the lost opportunities if STI services are not provided for individuals initiating PrEP and highlights the opportunity to harness the growing interest in providing PrEP programs globally to be a gateway to provide more comprehensive sexual and reproductive health services for PrEP users. There are opportunities for economies of scope and scale to control STIs by leveraging the growing infrastructure of PrEP delivery and access to higher-risk individuals. Synergistically, the identification of high-risk individuals with STIs can be a gateway for the provision of PrEP. Implementing more frequent STI screening and testing and partner services among high-risk individuals may potentially lessen the effect of STI epidemics., As we strengthen the delivery of sexual and reproductive health services for PrEP users globally, there may also be a positive flow-on effect for nonusers living with HIV who also are at high risk for STIs, and other nonusers may also be able to access these services.

13The high pooled prevalence of STIs among those starting PrEP reinforces the belief that we are reaching groups at high risk for HIV and STIs, and the high pooled incidence emphasizes the need for ongoing STI testing and treatment services because PrEP users remain at high risk for STIs. Our study complements other meta-analyses of STI incidence among MSM only,,; however, we extend their findings by examining sources of heterogeneity according to anatomical site of detection, study population composition, country income level, and study type. We noted a high level of heterogeneity in our pooled estimates, which may be due to additional factors, including differences in background HIV prevalence in country or setting, case mix of populations (ie, sampling different underlying populations: different distributions of socioeconomic status, race/ethnicity, age, or sexual mixing networks), study designs (variable inclusion criteria for PrEP, different frequency of testing), and STI diagnostic protocols (eg, the Pre-exposure Prophylaxis Initiative [iPrEx] trial, analyzed urethral samples for chlamydia or gonorrhea only if leucocytes were present in urine, whereas Australian demonstration projects did not impose such reliance on urine leucocytes). Nevertheless, despite this high level of heterogeneity between studies, the consistently high STI prevalence and incidence reported in individual studies cannot be ignored.

14This systematic review uncovered several important gaps in evidence. First, we found only 1 article that reported antimicrobial-resistant M genitalium among PrEP users. With expected high yields of positive samples from PrEP users, PrEP programs may be useful as sentinel surveillance sites for STI–antimicrobial resistance monitoring for N gonorrhoeae and M genitalium. Second, there are inconsistencies in how STI prevalence and STI incidence are reported, precluding their inclusion in meta-analyses. For future meta-analyses, reporting the number of cases with person-years at risk or incidence rates with 95% CIs would be a minimum requirement. We recommend disaggregating STI prevalence and STI incidence by pathogen and subpopulations (eg, age, sex, or transgender identity).

15Our study is useful to advocate for improved access to STI services for PrEP users and to inform program design and cost-effectiveness analyses. There is a clear need to facilitate the development of affordable, accurate, and easy-to-use point-of-care tests for STIs and developing models for STI case management in resource-constrained settings. A reevaluation is needed of how diagnostic costs can be reduced and how economies of scope and scale may be gained from using the existing infrastructure of cartridge-based molecular diagnostic machines that are used for other diseases (such as tuberculosis). The current interest, demand, and support for PrEP services in LMICs is predicated on a need to provide PrEP as simply and cheaply as possible. Therefore, a tension exists between the increasing costs and complexity of PrEP implementation and the opportunity and need to provide effective STI services. A market and technology landscape report for STI diagnostics (similar to HIV self-testing) would be a helpful resource for PrEP programs. Furthermore, guidance from international authorities, such as the World Health Organization, will be needed to define what may be considered as essential sexual health services compared with enhanced services, particularly in resource-constrained settings.

16There are ongoing challenges in implementing integrated STI services within PrEP programs. The key challenges are related to STI diagnostics, program logistics of combined STI and PrEP delivery, and lack of STI capacity building. Particularly for LMICs, there is a lack of access to triple–anatomical site sampling (ie, testing from oropharyngeal, urogenital, and anorectal sites), which is critical for detecting STIs in MSM. This situation is usually related to lack of funding, so considerations should be given to the burgeoning evidence for pooled samples testing. A robust economic case is pertinent because cost has been raised as a major barrier, even in HICs where direct user costs may be incurred by those with no health insurance.

17The strength of our review is the inclusion of data from 26 countries including non-MSM populations, LMIC settings, and previously unreported STI data. Our findings should be considered in light of several limitations. First, there is a potential for selection and detection bias. The high STI prevalence for individuals starting PrEP may reflect the inclusion criteria for some PrEP programs (ie, some clinicians may encourage same-day referral for PrEP when a rectal STI is diagnosed). The pooled incidence may be overestimated owing to more frequent testing and from more anatomical sites. Second, not all PrEP-related publications focused on reporting STI data. We mitigated this factor by approaching PrEP programs for unpublished STI data. Third, we included only laboratory-confirmed STIs. Therefore, most estimates came from HICs where diagnostics were available, whereas estimates obtained in LMICs are representative of externally funded research programs.

18Given the high STI burden among individuals initiating PrEP and among persistent PrEP users, there are opportunities to leverage the global interest in PrEP policy and the development of programs to actively promote the integration of STI services, which includes appropriate asymptomatic testing, treatment, and targeted vaccination. Currently, fewer STI data are available from programs offering PrEP to women, young people, serodiscordant couples, and transgender individuals outside HICs. More data would help guide recommendations on the frequency and optimal STI testing approaches for all population groups accessing PrEP.