To the Editor:

Borowitz et al (J Am Board Fam Pract 2003;16:213–8) reported that painful defecation was the primary precipitant of childhood constipation and recommended several interventions to reduce the risk of chronic constipation. We believe, however, that their study results do not strongly support these conclusions and should be interpreted more cautiously.

The retrospective case-control design of the study is the source of its limitations. The authors attempted to measure the association between painful bowel movements and subsequent constipation by asking parents of currently constipated children and control subjects whether they remembered their children having painful or difficult bowel movement during the prior 3 months. We believe parents of constipated children who are presenting for medical care would be much more likely than parents of control subjects to perceive their child’s recent bowel movements as painful or difficult. Such parents would also be more likely to recall episodes of painful or difficult defecation. Consequently, the observed association between parental perception of previous painful defecation and current constipation can be explained (at least in part) by recall bias.

Borowitz et al nevertheless assert that painful defecation is the “primary precipitant” of constipation in early childhood, implying that painful defecation is causally linked to constipation. Even if painful defecation were strongly associated with later constipation, it is plausible that a true causal factor (eg, low fluid intake, genetic susceptibility) could lead both to painful defecation and subsequent constipation. In other words, painful defecation could be associated constipation without being its primary precipitant.

The authors advocate several interventions to reduce the risk of chronic constipation, including counseling parents of asymptomatic children and dietary modification or laxative medication for constipated children. Their study, however, was not designed to evaluate the effectiveness of these interventions in reducing the risk of chronic constipation, and each of these interventions has plausible harms. Although the effectiveness of these interventions is plausible, family physicians should regard these recommendations as expert opinions until they have been substantiated by well-designed studies.