Investigators reported adjusted prevalence estimates of chemsex across four survey years collected by the RiiSH study. The adjusted estimates were 6.1% in 2022, 4.2% in 2023, 5.1% in 2024 and 6.8% in 2025. The authors described prevalence as low and relatively stable across this period. Trend estimation used log-binomial regression models, according to the abstract, and the sample comprised 5,728 respondents.
The year-to-year variation was small. Still, the 2025 estimate returned to a level similar to 2022. The study team did not present event-level drug or dosing details in the abstract; if such granularity exists, the preprint points readers to supplementary materials or underlying data access notes.
Data came from the RiiSH study, an annual online cross-sectional survey of gay, bisexual and other men who have sex with men conducted between 2022 and 2025. Participants were recruited through social networking and dating apps and via community partners, the authors noted. This sampling strategy targets community-attached men rather than clinic-only attendees.
For trend analyses the investigators applied log-binomial regression models to estimate crude and adjusted chemsex prevalence over time. For the assessment of factors associated with chemsex, the team used logistic regression and prespecified seven candidate variables: age, sexual orientation, education, employment, HIV status, recent HIV PrEP use, and recent sexual health service (SHS) attendance.
The cohort included 5,728 respondents; the preprint abstract does not give a full breakdown of age bands, education levels or employment categories beyond the associations reported. The authors reported comparing those aged 35–59 to a reference group aged 16–34 when evaluating age-related differences in chemsex likelihood. Recent engagement with SHS and recent HIV PrEP use were recorded and analysed as candidate predictors.
Because this is a community sample recruited through apps and partners, responders likely reflect a mixture of men who access formal sexual health services and those who do not, the authors suggested by the inclusion of SHS use as an explanatory variable.
Logistic regression results, as presented in the abstract, identified several variables associated with higher odds of reporting chemsex. Participants who identified as gay had higher odds of chemsex compared with other sexual orientations (adjusted odds ratio [aOR]=1.57; 95% CI 1.11–2.29; p=0.01). Unemployment was strongly associated with chemsex (aOR=3.08 versus full-time employed; 95% CI 2.19–4.26; p<0.001).
Age was associated with chemsex: those aged 35–59 had greater likelihood compared with the 16–34 age band (p<0.001). Recent HIV PrEP use carried a markedly elevated adjusted odds (aOR=3.79 versus no recent use; 95% CI 2.63–5.56; p<0.001). Recent sexual health service (SHS) attendance was also associated (aOR=2.23; 95% CI 1.76–2.83; p<0.001). Finally, living with HIV was associated with higher odds of chemsex (aOR=3.16 versus negative or unknown status; 95% CI 2.43–4.09; p<0.001).
The authors framed these associations as markers for where clinical and support services might intersect with populations reporting chemsex. They did not report additional model diagnostics or alternative variable codings in the abstract; detailed model outputs may be in the full preprint or supplementary files.
Data indicated a consistent relationship between service engagement and chemsex: men reporting recent HIV PrEP use and recent SHS attendance were more likely to report chemsex. Investigators reported this pattern as supportive of integrating specialist support services for chemsex within existing sexual health service settings.
Integration could mean targeted screening, referral pathways, or co-located psychosocial and substance-use support within SHS, the authors suggested by implication in their conclusions. The abstract does not specify particular models of care or evaluate any intervention; it reports associations that point to where services might focus resources.
The authors declared this work as a preprint that has not undergone peer review; they cautioned it should not yet guide clinical practice. Ethical approval was obtained from the United Kingdom Health Security Agency Research Ethics and Governance Group (reference R&D 524), and participant consent procedures were confirmed.
The preprint states the underlying dataset cannot be made publicly available for ethical and legal reasons. The abstract does not provide event-level drug type, dosing, or other granular chemsex behaviour data; readers should consult the supplementary materials or contact the authors for clarifications where permitted. The absence of public data and the preprint status are concrete limitations reported by the study team and should temper immediate translational use of these findings.