31. Lifetime prevalence of anal intercourse among sexually active female youth and young female sex workers: a comparative systematic review and meta-analysis
Owen BN., Brock PM., Shubber Z., Baggaley RF., Butler AR., Pickles M., Brisson M., Drolet M., Boily M-C.
Background Anal cancer incidence has increased over the past twenty years. Heterosexual anal intercourse (AI) is a risk factor for HPV and HIV infection but is under-researched and ill-understood. We compare AI practices among young, sexually active general population females and female sex workers (FSW). Methods: We searched PubMed for heterosexual AI studies among young females, including FSWs (mean age <25 years), to December 2012. Study estimates were pooled using a random-effects model. Results: Fifty-four studies (42 average-risk, 6 higher-risk, 6 FSWs) were included. Most studies among general population youth were conducted in North America using self-administered questionnaires. All FSW studies were conducted in Africa and Asia using face-to-face interviews (FTFI). Pooled estimates of lifetime AI prevalence were larger among average-risk (23.6%, 95% CI 20.4–26.7) and higher-risk youth (25.5%, 95% CI 11.7–39.2) than FSWs (12.8%, 95% CI 8.4–17.3), but highly heterogeneous (I2 >90%). However, pooled AI estimates by continent of average-risk youth (Asia = 13.9%, 95% CI = 1.7–29.5; Africa = 18.4%, 95% CI 0.9–35.8) were more similar to those of FSWs (Asia = 16.0%, 95% CI 10.3–21.6; Africa = 9.7%, 95% CI 3.8–15.7). Estimates of average-risk youth reporting via FTFI (12.1%, 95% CI 0.7–23.5) were likewise similar to those among FSWs (12.8%, 95% CI 8.4–17.3). Pooled AI prevalence estimates among FSWs were higher in studies conducted after 2001 than in earlier studies. Conclusions: AI is common among sexually active females and may be increasing; it could therefore be an important determinant of HPV transmission and anal cancers. AI is as or more common among the general population youth than young FSWs but this may be confounded by continent, interview method and other unmeasured variables.