Background:

PrEP follow-up in primary care involves repeated contacts and regular STI screening. Higher chlamydia (CT) and gonorrhoea (NG) diagnoses in PrEP users may reflect surveillance bias (“testing paradox”), and frequent screening increases antibiotic exposure with potential antimicrobial-resistance pressure.

Research questions:

Among men who have sex with men (MSM) in a real-world PrEP programme, how do PrEP status and STI screening intensity jointly relate to observed CT/NG incidence and prescribing of STI-related antibiotics?

Method:

Retrospective cohort using records from Maccabi Healthcare Services (Israel). Men aged ≥21 years with ≥1 rectal CT/NG PCR test (Jan 2019–Jun 2024) contributed consecutive 3-month intervals. Lagged PrEP exposure: never, current (≥70% days covered), recent (stopped ≤30 days), or stopped (>30 days). Screening intensity was log(1+tests/interval); repeat positives within 60 days were excluded. Outcomes were CT/NG diagnoses and prescriptions for doxycycline and ceftriaxone (cephalexin negative control). Mixed-effects negative binomial models with person-time offset and patient random intercept estimated aIRRs, including PrEP×intensity interaction and adjustment for age, SES, and prior STIs.

Results:

6,943 individuals contributed 174,516 person-months. Screening was more frequent during current PrEP (74% of intervals tested) than never PrEP (~23%). At annual screening, current PrEP versus never PrEP showed large aIRRs for NG (4.2, 95% CI 3.8–5.3) and CT (5.2, 4.0–6.6), attenuating with more frequent testing (quarterly: NG 2.2, 1.9–2.5; CT 2.7, 2.3–3.1; monthly: NG 0.8, 0.7–1.0; CT 1.0, 0.8–1.2). We recorded 10,248 doxycycline and 7,476 ceftriaxone prescriptions; at annual screening, current PrEP had higher prescribing (doxycycline aIRR 3.1, 2.7–3.6; ceftriaxone 3.6, 3.0–4.4), which diminished at monthly screening (0.9, 0.8–1.1; 0.8, 0.7–1.0).

Conclusions:

Apparent PrEP-associated increases in STI diagnoses and antibiotic use depended strongly on screening intensity, consistent with surveillance bias. The large cohort and sensitivity analyses support robustness. Risk-stratified screening in general practice may preserve timely STI care while reducing avoidable antibiotic exposure.

Points for discussion:

How should primary care/PrEP programmes operationalise risk-stratified screening intervals while maintaining continuity and equity?

What is the acceptable trade-off between earlier asymptomatic detection and antibiotic stewardship/AMR risk in routine PrEP follow-up?

How should health systems interpret trends in STI incidence when screening intensity changes over time (programme evaluation and policy)?