In a large population-based case-control study, Wright et al1 reported a 15% reduction (95% confidence interval [95% CI], 1%-27%) in the risk of prostate cancer in men circumcised before their first sexual intercourse. They hypothesize that this is due to protection against a sexually transmitted infectious cause of prostate cancer. We believe that this result is unlikely to reflect a causal association between circumcision and the risk of prostate cancer, for 2 unrelated reasons.
First, we believe the underlying premise, that data support a potential sexually transmissible cause of prostate cancer, is not correct. The association between prostate cancer risk and a history of sexually transmitted infections (STIs) is tenuous at best. Large-scale prospective studies of the risk of prostate cancer in men with human immunodeficiency virus (HIV) with more than 1 million person-years of follow-up have found that rates of prostate cancer are significantly decreased compared with the general population (meta- standardized incidence ratio, 0·70; 95% CI, 0.55-0.89).2 Because HIV is sexually transmissible, these men are also at an increased risk of other STIs. In addition, the rates of most confirmed infection-related cancers are increased in men with HIV, including cancers due to Epstein-Barr virus (non-Hodgkin and Hodgkin lymphoma), Kaposi sarcoma herpes virus (Kaposi sarcoma), human papillomavirus (cervical, anal, penile, and oral cavity cancers), and hepatitis B and C viruses (liver cancer). It is striking that prostate cancer rates are not found to be increased in these men.
Second, strong selection biases related to socioeconomic status (SES) operating in case-control studies may have caused an artefactual association between circumcision and a reduced risk of prostate cancer in the study by Wright et al.1 In the United States, circumcision rates are substantially lower in men of lower SES; a population-based study found that circumcision rates increased from 60% in men with less than a high school education to 84% in men with more than a high school education.3 Response rates have also been found to be universally lower in controls of low SES in population-based studies.4 In this case, this would have led to an underrepresentation of uncircumcised (low SES) men among the controls, which could potentially account for the weak protective association with circumcision that Wright et al report.1 Differential response rates (78% for cases and 67% for controls) enhance the plausibility of selection bias as an explanation for the observed association.
There is abundant evidence that male circumcision results in a reduction in some types of STIs.5 However, the balance of evidence suggests that prostate cancer risk is related to neither an infectious cause nor circumcision.