Pitfalls in studies of children with monosymptomatic nocturnal enuresis.

In this issue Ferrara et al. present an important placebo-controlled, three-arm, double-blind, double-dummy, study on the treatment of nocturnal enuresis, demonstrating that homotoxicology is superior to placebo but less effective than desmopressin. Nocturnal enuresis is a disease with a heterogeneous aetiology and complex pathophysiology. The fact that different therapies may result in a wide range of responses is, therefore, not surprising. Differences in success rate can, therefore, be largely attributed to selection bias in the sub-populations. This consideration must be taken into account for every study design, to avoid premature interpretation of the results. Positive results in a paper are only not subject for discussion if both methodology and study population fulfil the highest standards, because negative results are not likely to be reported. Several points of weakness are present in the majority of studies, such as (a) inappropriate subtyping of the patients (terminology) or (b) epidemiological data, (c) insufficient documentation of patients' characteristics, (d) lack of plausible explanation as to why the placebo effect might be absent, (e) the heterogeneity of the severity of bedwetting. All these may lead to false positive and/or false negative results. In this commentary we try to tackle these different issues which might be relevant for the interpretation even of placebo-controlled studies like that of Ferrara et al.