Abstract
HIV pre-exposure prophylaxis (PrEP) effectively reduces new HIV diagnoses. High rates of incident bacterial sexually transmitted infections (STIs) have been observed in patients eligible for and adherent to PrEP. Observational studies generally report low long-term retention in PrEP care. Limited data exist on the rates of bacterial STI diagnosis upon re-engagement with PrEP services. We conducted a retrospective chart review within the HIV prevention program of an urban academic medical center in New York City. Eligible patients started PrEP from 2015 to 2019, then resumed PrEP services after a gap in care of at least 180 days. Demographic, clinical, and laboratory data were used to characterize the patient population and rates of bacterial STI diagnosis at re-engagement. In total, 286 patients were identified, with 316 qualifying re-engagement visits. Twenty-nine percent of patients had continued PrEP during the care gap, and 30% reported discontinuing medication due to a perceived change in risk. A new STI was diagnosed at 19% of re-engagement visits. There was no statistically significant difference in rates of new STI between individuals returning on or off PrEP, nor between those with perceived lower risk and those without. Individuals who fall out of PrEP services and subsequently re-engage remain at high risk of bacterial STI during the gap in care, regardless of whether PrEP medication is continued or the patient perceives themselves to be at lower HIV acquisition risk. Providers should strongly encourage patients discontinuing PrEP to remain engaged in sexual health services. Alternatives to clinic-based PrEP care must still include regular bacterial STI screening.
Introduction
Oral HIV pre-exposure prophylaxis (PrEP) using emtricitabine with tenofovir disoproxil fumarate or tenofovir alafenamide is a highly effective intervention for the prevention of HIV infection, and an integral part of the United States end the HIV epidemic campaign. 1 One advantage of oral PrEP is that it can easily be initiated, or used “on demand,” during periods of elevated HIV acquisition risk and then discontinued if risk factors change.
High rates of PrEP discontinuation are well documented, with 27–72% of individuals lost to follow-up in recent reports, 2 –4 and a substantial but variable proportion citing change in risk as the reason for leaving PrEP care. 3,5 –9 High rates of resumption, between 27% and 60%, are also reported, 6,10,11 and there is likely an additional subset of individuals who do not attend follow-up appointments but continue taking medication. 12
Multiple investigators have noted high rates of bacterial sexually transmitted infection (STI) diagnosis among PrEP users. 13 –15 However, little is known about the ongoing risk of people who discontinue PrEP in the context of falling out of PrEP care. One study in New York City reported fewer condomless sex acts among people who discontinued PrEP, as well as lower rates of rectal gonorrhea/chlamydia diagnosis using self-collected testing. 16 Our study aims to further evaluate the ongoing HIV risk among patients who return to PrEP care after a prolonged gap by measuring the incidence of new bacterial STI diagnosis in this population. In addition, we wished to characterize patient demographics, reasons for falling out of care, and PrEP medication use during care interruption.
In this brief report, we present data on rates of bacterial STI diagnosis, PrEP medication continuation, and reasons for loss to follow-up among patients re-engaging in PrEP care.
Methods
Participants and procedures
Data were taken from a previously described cohort of patients assessed for comprehensive HIV prevention services at three outpatient clinics comprising the HIV prevention program within a large academic medical center in New York City. 17 Demographic and clinical data were abstracted from the clinical record by the study authors. This study was approved by the Columbia University Institutional Review Board with a waiver of written informed consent. PrEP start dates ranged from February 1, 2015, to July 25, 2019.
Patients were selected for inclusion if they had a documented PrEP start date and at least one gap of ≥180 days between documented HIV tests, indicating a gap in appropriate PrEP care during this time. Exclusion criteria included an incorrect medical record number, loss to follow-up after the initial PrEP visit, documented visits during the “gap” in care, or evidence that the patient never initiated PrEP.
Of the 696 individuals initiating PrEP in our database, we identified 286 patients and 316 distinct clinic visits meeting inclusion criteria. The mean patient age was 31 years (range 17–63), with 96.1% identifying as male at birth. In total, 136 patients (47.6%) identified as Hispanic of any race, 63 (22.0%) as non-Hispanic Black, 58 (20.3%) as non-Hispanic White, 9 (3.1%) as non-Hispanic Asian, and 8 (2.8%) as other non-Hispanic, with data missing for 12 participants.
In total, 243 patients (86.5%) reported a sexual preference for men, 30 (10.7%) for both men and women, and 8 (2.8%) for women only, with data missing for 5 patients. A majority reported penetrative or receptive anal (96.8%) or oral (97.5%) sex.
Measures and data analysis
Clinical records from each re-engagement visit were reviewed to abstract the patient's reported reason for the gap in PrEP care, as well as whether they had continued taking PrEP and from what source. All test results for gonorrhea, chlamydia, syphilis, and HIV were included for analysis and reviewed by study authors. Gonorrhea and chlamydia testing was performed on urine specimens as well as rectal and throat swabs at the discretion of the treating clinician. Positive tests at re-engagement in care were considered to represent new infection.
A new syphilis diagnosis was defined as a newly positive rapid plasma reagin (RPR) with confirmatory TPA-ABS (traditional testing algorithm), or fourfold rise in RPR titer compared with the most recent result. If previous RPR results were unavailable, evidence of a new infection was determined by a clinician's decision to treat. Odds ratios (ORs), confidence intervals (CIs), and p values were calculated for rates of new STI diagnosis between clients reporting continued PrEP use versus not taking PrEP upon the re-engagement visit. Differences in STI rates between patients returning to care on or off PrEP were assessed using Fisher's exact test with a two-tailed p value. Analyses were carried out using Epi Info. 18
Results
At the time of re-engagement, 83 patients (29.1%) reported still taking PrEP across 87 instances, with 4 of these individuals having 2 gaps in care. In total, 203 individuals reported being off PrEP across 229 instances, with 24 having 2 gaps in care, and 1 patient with 3 gaps. The mean duration of care interruption including all qualifying instances was 289 days.
Of the individuals who reported continued PrEP use during their gap in care, 40% obtained prescriptions from our program, 28.7% from an outside provider, and 30% had provider documentation of medication adherence without noting a specific source.
We present rates of bacterial STI diagnosis at re-engagement visits, with results stratified according to reported ongoing PrEP medication use at the time of the visit (Table 1). Overall, a new STI was diagnosed at 19.9% of all re-engagement visits. New STI diagnoses were more common in those returning on PrEP versus off PrEP (36.5% vs. 25.3%), but this difference was not statistically significant. In terms of specific STI diagnoses, rates of both rectal gonorrhea (OR 3.22, CI 1.04–9.92, p = 0.04) and syphilis (OR 3.25, CI 1.06–9.97, p = 0.04) were more common in those returning still on PrEP versus off PrEP. There were no new diagnoses of HIV over the course of the study.
Bacterial Sexually Transmitted Infection Diagnosis at Pre-Exposure Prophylaxis Re-Engagement Visit
CI, confidence interval; CT, chlamydia trachomatis; GC, Neisseria gonorrhea; PrEP, pre-exposure prophylaxis; STI, sexually transmitted infection.
We also examined patients' stated reason for loss to PrEP follow-up at each re-engagement visit (Table 2). The most commonly reported reason for leaving PrEP care was a lower perceived risk of HIV acquisition, followed by insurance issues, loss of access due to travel or work, and difficulty making appointments. Only one instance of discontinuation due to side effects (mild nausea) was documented. Of the instances after PrEP discontinuation due to lower perceived risk, a new STI diagnosis was made at 9 of 55 (16.3%) compared with 31 of 174 (17.9%) visits after discontinuation for any other reason (OR 0.9, CI 0.40–2.04, p = 0.49).
Reported Reasons for Pre-Exposure Prophylaxis Medication Discontinuation
PrEP, pre-exposure prophylaxis.
Discussion
This study builds on a previously reported analysis of individuals falling out of PrEP care at an urban HIV prevention program 7 by describing those who went on to re-engage in care. Rates of bacterial STI diagnosis were used as a proxy indicator for risk of HIV acquisition during this time.
Of those who stopped PrEP during the gap in care, the most common reason for medication discontinuation was a perceived lower risk of HIV acquisition, presumably corresponding to a change in risk behavior. However, rates of new STI diagnosis at the time of re-engagement did not differ between those who returned on PrEP versus off PrEP, indicating that this perception did not correspond to differences in risk. This adds to previous reports of discrepancy between individuals' perceived risk of HIV acquisition and objective measures of risk. Similarly, there was no statistically significant difference in overall rates of bacterial STI diagnosis between those who reported continued use of PrEP versus those who stopped taking PrEP.
Among those who discontinued PrEP medication, a high proportion cited structural barriers such as difficulty making appointments and loss of access due to travel or work as reasons for falling out of care. (It should be noted that despite the high proportion of patients citing cost as a barrier, uninsured patients in New York state can have most PrEP costs covered by the PrEP assistance program.) This represents a compelling argument for the use of alternative care delivery methods such as televisits.
However, in light of the high rate of STI diagnoses noted in our population, any transition away from in-person care must be balanced with increased opportunities for appropriate STI screening (e.g., at-home testing). Future expansion of PrEP care beyond the traditional clinic model will need to integrate access to the recommended bacterial STI screening in addition to HIV testing.
Our study has some limitations. The generalizability of our results is limited by a cohort comprised primarily of MSM, although this generally reflects the demographics of PrEP use in the United States. In addition, records from outside health care providers were not available, which may have led to underestimation of new STI diagnoses for those individuals who obtained care outside of our institution. Future PrEP research should be inclusive of women, and should also address opportunities to improve STI screening outside the clinic setting for those who have difficulty attending appointments.
Finally, data from this study preceded the onset of the COVID-19 pandemic, after which resources were directed away from STI testing and treatment. In 2020, PrEP use declined, 19 and rates of bacterial STI (particularly syphilis and gonorrhea) increased. 20 Data from this series likely underestimate rates of PrEP discontinuation and STI diagnosis over this period.
Together, our findings suggest that bacterial STI and, by proxy, HIV acquisition risk remain high during lapses in PrEP care. Though some individuals may discontinue and resume care as their HIV risk changes, our data do not support a decreased risk during this time. Every effort should be made to counsel patients around the importance of continuing visits for sexual health care and continued discussion of PrEP appropriateness even when there is a perceived reduction in risk.
Footnotes
Authors' Contributions
J.M. contributed to writing—original draft, and formal analysis. L.B. was involved in conceptualization, methods, and formal analysis. J.Z. was in charge of supervision. Others were involved in writing—review and editing.
Author Disclosure Statement
No competing financial interests exist.
Funding Information
Research reported in this publication was supported by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health under award number 5UM1AI069470-14 (M.S. and J.Z.), K23AI150378 (J.Z.), L30AI133789 (J.Z.), and 5T32AI100852-10 (J.M.).