Study Design Matters: Using a Rigorous Analytic Design to Examine the Association Between Rectal Douching and...

Prior studies have identified an association between rectal douching and prevalent rectal Neisseria gonorrhoeae (GC) and Chlamydia trachomatis (CT), but few studies have examined associations with incident infections.

We used data from ExGen, a 48-week cohort study conducted in Seattle, Washington, 2016 to 2018. Each week, participants completed a survey and self-collected rectal specimens, which were tested for CT/GC at the end of the study. Our primary analysis was a case-crossover analysis. We identified individuals with incident rectal GC or CT and included 1 case-week (the first week of the incident rectal infection) and 2 randomly selected control-weeks (weeks where the participant tested negative for rectal GC/CT). We used multivariate conditional logistic regression to estimate the association between douching and rectal GC/CT. In an exploratory analysis, we created a "synthetic" cross-sectional study of the longitudinal data to mirror the design of other studies.

There were 140 individuals in ExGen. Our case-crossover sample comprised 17 GC case-weeks and 34 matched GC control-weeks, and 22 CT case-weeks and 44 matched CT control-weeks. Participants reported douching during 64.7% of GC case-weeks and 70.6% control-weeks (adjusted odds ratio, 0.47; 95% confidence interval, 0.07-3.10). Douching was reported at a similar frequency among CT case-weeks (45.5%) versus control weeks (47.7%; adjusted odds ratio, 1.16; 95% confidence interval, 0.26-5.07). In our exploratory synthetic cross-sectional study, rectal douching was significantly associated with rectal GC but not CT.

Douching was not associated with rectal GC/CT in this methodologically rigorous analysis. A better understanding of the mechanisms of douching products and impact of long-term use may be warranted.