Re: Jennifer R. Rider, Kathryn M. Wilson, Jennifer M. Sinnott, Rachel S. Kelly, Lorelei A. Mucci, Edward L. Giovannucci. Ejaculation Frequency and Risk of Prostate Cancer: Updated Results with an Additional Decade of Follow-up. Eur Urol 2016;70:974–82

We would like to comment on the methodology used by Rider et al [1] in their study. We first consider the reference group. The investigators selected an intermediate group (4–7 ejaculations permonth [EPM]) since the sample sizewas low for the groupwith 0–3 EPM, but this is not appropriate. In general, researchers must choose the lower or upper group as the reference category; in this case, perhaps the authors could have recategorized ejaculation frequency to choose the lower group. Second, the study included health professionals aged between 40 and 75 yr in 1986, but the exposure was measured in 1992 in a survey asking about average EPM when respondents were aged 20–29 yr and 40–49 yr and in the year before the survey. This could have led to information bias, specifically recall bias, so the exposure may have been measured inappropriately [2]. In addition, the researchers excluded men with T1a disease, but they do not describe the reasons why they decided this. This exclusion could have led to selection bias. Perhaps the frequencymight not be that high; however, it is important to include all prostate cancer cases. Finally, we identified a different statistical analysis that led to a statistically significant positive association, but statistical significance is the least interesting thing about the results. It is well known that the more you analyze, the