The commentary by Morris et al. [1] builds a model for the impact of circumcision on prostate cancer based on three axioms for which there is little or no support. They begin with the notion that prostate cancer is caused by underlying inflammation or infection (possibly sexually transmitted diseases (STDs)), although a link to a specific pathogen remains elusive. This fairly new idea needs further study before basing treatment models on it, especially considering the lack of a correlation between the incidence of prostate cancer and cancer of the cervix, an increase in prostate cancer mortality among Roman Catholic priests, and being inconsistent with other studies [2].

Second, they postulate that uncircumcised men are at greater risk of STDs based on one study, while ignoring the eleven studies that show either no difference or that circumcised men have an overall greater risk of acquiring an STD.

Finally, they postulate that uncircumcised men have a greater risk of prostate cancer based on four studies. Two of these studies are >40 years old and compare the risks of Jews to non-Jews. Prostate cancer incidence varies with race. In the USA, African Americans have nearly twice the risk of Whites, who have twice the risk of Asian Americans [3]. In a study that compared cases of prostate cancer to cases of BPH, circumcised non-Jews had three times the risk of prostate cancer than Jews (Odds ratio (OR) 3.23, 95% CI 1.56–6.69), indicating that Jews, as a racial category, were at lower risk [4]. The other two studies relied on patient report, an unreliable practice, to determine circumcision status.[5,6] The British study failed to consider ethnicity or race, so circumcision status might have been a marker of socioeconomic status [5]. The American study did consider race and found that uncircumcised Blacks had 2.6 times the risk of prostate cancer compared with Blacks circumcised as newborns, while the risk was 1.7 times higher for uncircumcised Whites compared with Whites circumcised as newborns [6].

Their third premise also ignores contrary evidence. In one study, when only non-Jews were considered, uncircumcised men trended towards lower risk of prostate cancer (OR 0.86, 95% CI 0.46–1.58) [7], while in another study no difference was found (OR 0.93, 95% CI 0.55–1.58) [8]. Similarly, no association has been found between PSA levels and circumcision status [9]. If circumcision reduced the risk of prostate cancer, it would be expected that the age-adjusted incidence of prostate cancer would decrease as the circumcision rate increased. In fact, the opposite has been documented [10]. Likewise, if Morris et al. [1] are correct, one would expect the incidence of prostate cancer in Europe to be 36–60% greater than the incidence in the White USA population. For the period 1983–1987, before PSA screening, the age-adjusted incidences of prostate cancer in Denmark, Finland, Iceland, Norway, and Sweden were 48.9, 61.8, 85.0, 71.8, and 81.6 per 100,000, respectively [11]. In 1986 the age-adjusted incidence among USA Whites was 86 per 100,000 [10].

Morris et al. failed to consider race. If Blacks constitute 11% of the USA population and are at twice the risk of prostate cancer, using the racially specific risks the number of cases of prostate cancer by circumcising everyone is 174,997 instead of 167,471. This decreases the estimated cost savings by US$ 125 million.

Morris et al. conveniently chose to not estimate the lost opportunity costs of circumcising 562 704 infants. Using a cost of $195 per circumcision and the standard discount rate of 5% over 70 years, the lost opportunity costs are $3.3 billion, which overwhelms the $1.1 billion in estimated prostate cancer costs.

The concept of using circumcision to reduce prostate cancer risk has no biological or epidemiological foundation. Even if one is willing to depart from reality and use the most extreme assumptions as put forth by Morris et al., circumcision of infants is not cost-effective.