BJUI published a survey of ≈1000 men in Belgium, concluding that male circumcision reduces penile sensitivity . This conclusion is at odds with an extensive literature review by the American Academy of Pediatrics , two large randomised controlled trials (RCTs) [3, 4], physiological measurements  and a recent survey of 10 000 men in Germany .
Several problems with the Belgian survey may explain the discrepancy. First, we question the data presented and its interpretation. A 5-point Likert scale was used for seven parameters in three locations (dorsal, lateral and ventral) of the glans and shaft (total 42 sites), but the actual differences were numerically small (1–11% for all but one measurement) for the 21 measurements yielding P < 0.05. If the statistical analyses had corrected for multiple testing few, if any, significant P-values would have been obtained. Moreover, we would question whether differences this small are biologically significant. Furthermore, we question whether it is realistic for a respondent to rate their orgasm intensity as a result of stimulating, for example, their lateral shaft alone. In contrast to the subjective measures used in this study, empirical testing has found no difference in sensation between circumcised and uncircumcised men .
The study design poses substantial problems. It involved a self-selected convenience sample in contrast to RCTs representing ‘level 1 evidence’ [3, 4]. Self-selection is highly likely to over-enrol opponents of male circumcision or men with sexual problems. This possibility is supported by the observation that 22.6% of the respondents were circumcised – a circumcision rate that greatly exceeds the proportion of men in European countries who are circumcised; the 6.7% rate in the German study  being more typical.
Rather than focusing on data of others, the authors criticise any contrary findings by others as arising from personal biases. For example in stating that the two large African RCTs were biased, they ignore the obvious fact that each trial took place in a region in which few local men were circumcised, as it would be inefficient to try to recruit from within populations having few eligible participants. The Kisumu trial in Kenya was done among Luo men in part for this reason. The RCTs offered a unique ability to assess sexual satisfaction after male circumcision, as they recruited and randomised 4456 men in one trial and 2784 in the other, all of whom had been sexually experienced before the procedure, and thus had the basis for a personal comparison. This is not possible in men who receive circumcision in infancy or childhood, as was the case for over half of the Belgian respondents.
Bronselaer et al.  say men circumcised aged > 10 years reported less sexual pleasure and more discomfort. Men in the RCTs were well matched for age, sexual behaviours and other relevant variables. Each trial found no decrease in sexual function, sensitivity or satisfaction [3, 4]. In the Kenyan trial over half of the men reported increased penile sensitivity and greater ease of reaching orgasm after undergoing circumcision . Similarly, most other studies of men evaluated before and after circumcision have shown few, if any, differences in sexual function. It is likely that imbalances between the groups may account for the findings of Bronselaer et al. Because participants were not randomised or ‘blinded’ in their convenience sample, many men may have been circumcised for medical reasons. Pre-existing medical problems (e.g. recurrent balanitis, posthitis or phimosis) might have lasting effects. Thus, surveys such as this one should either exclude circumcisions for medical reasons or should be conducted in populations in which non-medical circumcisions are commonplace.
They say ‘Sorrels et al. included sensitivity of the foreskin in their assessment and found a clear link between tactile penile sensitivity and sexual pleasure’, whereas Sorrells et al.  did not assess sexual pleasure. Furthermore, statistically rigorous analysis of the data of Sorrells et al. has shown no significant differences between the circumcised and uncircumcised groups.
In conclusion, substantial flaws in the design, statistical analysis and interpretation of data in this Belgian study reduce its interpretability.