Child Sexual Abuse and Human Papillomavirus Infection

Abstract

To the Editor. Stevens-Simon et al are to be commended for their study in what remains probably among the most persistent controversies in child sexual abuse research: the significance of genital human papillomavirus (HPV) infection.1 Their study showed that HPV DNA was detected in 5 (16%) of the 31 girls with confirmed or suspected sexual abuse and none of the 9 girls found to not have been abused. However, the statement in the results of this article (“P < .05, Fisher's exact test”) is in error; this difference did not approach statistical significance, possibly because of the small number of nonabused children (uncorrected χ2 = 1.66, P = .1198; Fisher's exact test, 1-tailed P= .258; 2-tailed P = .570; see Fig 1).2 Given the difference in rates of detection of HPV DNA, at least 25 nonabused girls with no HPV DNA detectable would have to have been enrolled to achieve a statistically significant difference. This underscores one of the challenges of this type of analysis: enrolling a sufficient number of nonabused children in a study that includes vaginal lavage and perineal swabs.In Reply. We appreciate the comments by Dr Beck-Sague and regret that the statistical test as noted in the abstract and results of our paper are in fact incorrect. Thus, the difference we noted in detection of genital HPV in abused versus nonabused girls (5 of 16, 16%, vs 0 of 9) can only be described as a “trend” rather than a statistically significant difference. We agree with her conclusion that acceptable approaches to obtaining genital samples from control children remains a problematic methodologic barrier to performing more definitive studies of genital HPV infection in sexually abused children. Thank you for your letter regarding this issue.