Clinical trials have demonstrated high efficacy of both daily PrEP and event-driven PrEP (2-1-1/on-demand) for HIV prevention. Event-driven PrEP involves taking two tablets before sex and one tablet twice after sex. Both regimens are used in Dutch clinical practice, but evaluating their real-world effectiveness requires routine-care surveillance data. The study compared HIV incidence between daily and event-driven PrEP use within the Dutch national PrEP program to assess differences in effectiveness in a programmatic setting.
This analysis used surveillance data from the Dutch national PrEP program covering August 1, 2019 through December 31, 2025. Data originated from the national registration of Centers for Sexual Health consultations (SOAP). Inclusion criteria for the analysis were individuals aged 16 years or older who initiated PrEP and had at least one follow-up consultation after initiation. At each consultation the PrEP regimen used since the last visit was recorded, and person-time was modeled as time-varying according to the regimen reported at each visit.
The investigators calculated HIV incidence rates per 100 person-years and used Cox proportional hazards models to estimate hazard ratios for HIV acquisition comparing regimens. Multivariable models adjusted for sociodemographic factors, sexual behavior, and history of sexually transmissible infections. The study used pseudonymized SOAP data and relied on the existing legal framework in the Netherlands that does not require additional ethical review for retrospective analyses of routine surveillance data.
A total of 16,469 individuals initiated PrEP and had at least one follow-up visit and were included in the analysis. The median follow-up duration was 2.0 years (IQR 0.8–4.0). Median age at baseline was 33 years (IQR 27–44). The study population was primarily composed of men who have sex with men (15,843 persons). Additional participants included 579 transgender and gender diverse persons, 45 women, and two men who have sex with women.
Across the cohort there were 49 new HIV diagnoses during 41,092 person-years of follow-up, yielding an overall incidence of 0.12 per 100 person-years (95% CI 0.09–0.16).
When examined by reported regimen during person-time, 42 HIV infections occurred during periods when participants reported event-driven PrEP use, corresponding to an incidence of 0.20 per 100 person-years (95% CI 0.15–0.27). Seven infections occurred during periods of daily PrEP use, corresponding to an incidence of 0.04 per 100 person-years (95% CI 0.02–0.07).
Cox proportional hazards models adjusted for sociodemographic characteristics, reported sexual behaviour, and prior sexually transmissible infections. In the multivariable analysis, time-varying use of event-driven PrEP was associated with a higher hazard of HIV acquisition compared with daily PrEP. The adjusted hazard ratio reported was 7.0 (95% CI 3.0–16.4). The authors note that overall HIV incidence remained low in the program despite this difference between regimens.
The study found a seven-fold higher adjusted hazard of HIV acquisition during periods of event-driven PrEP use compared with daily PrEP in the Dutch national PrEP program. The authors interpret this finding as potentially driven by lower adherence to the event-driven (2-1-1) dosing schedule in routine practice, which may reduce real-world effectiveness compared with clinical trial settings.
Based on these results, the authors recommend enhanced, person-centered counseling for individuals interested in, or using, event-driven PrEP in real-world settings. They call for research to identify the specific domains where users need support and the preferred methods of delivering that support. Details on the exact adherence measures, behavioral patterns, or which counseling strategies might be most effective were not reported in the abstract and would require consultation of the full preprint or supplementary materials.
The analysis used pseudonymized surveillance data from SOAP. Under Dutch law, retrospective analyses of routinely collected surveillance data do not require additional ethical review or informed consent; individuals can opt out of data sharing when visiting Centers for Sexual Health. The authors stated that there was no funding for this study. One author reported institutional unrestricted research grants from Gilead Sciences and participation in advisory boards paid to his institute; all other authors declared no competing interests.
Pseudonymized individual participant data from SOAP can be requested for scientific use through submission of a methodologically sound proposal to the SOAP registration committee. Proposal forms and further information are available via the provided contact (soap{at}rivm.nl); data requestors will be required to sign a data access agreement. The authors indicate that further research should investigate which domains of support and which delivery methods are most acceptable and effective for people using event-driven PrEP. Specifics on adherence measurement, intervention content, or implementation strategies were not provided in the abstract and would need to be sought in the full preprint or supplementary files.