Current issues of ACP Journal Club are published in Annals of Internal Medicine


Review: Behavioral interventions plus laxatives are effective for defecation disorders, but biofeedback does not add benefit


ACP J Club. 2002 Jul-Aug;137:28. doi:10.7326/ACPJC-2002-137-1-028

Source Citation

Brazzelli M, Griffiths P. Behavioural and cognitive interventions with or without other treatments for defaecation disorders in children. Cochrane Database Syst Rev. 2002;(1):CD002240 (latest version 1 Jul 2001). [PubMed ID: 11687154] (All 2002 articles were reviewed for relevancy, and abstracts were last revised in 2008.)



In children with defecation disorders, are behavioral modification (B-mod) techniques and cognitive therapies more effective than conventional treatments?

Data sources

Studies were identified by searching 12 databases (including the Cochrane Incontinence Group Trials Register and the Enuresis Resource and Information Centre Register) in March 2001, scanning reference lists, and contacting experts.

Study selection

Studies were selected if they were randomized or quasi-randomized controlled trials of cognitive, behavioral, or cognitive and behavioral interventions in children with a history of fecal soiling with or without constipation.

Data extraction

The quality of studies was assessed. Data were extracted on participant characteristics, interventions, and outcomes.

Main results

14 studies (12 randomized controlled trials, 843 children) met the selection criteria. Children had encopresis (i.e., inappropriate passage of stool in children > 4 y of age) in 12 studies and fecal incontinence resulting from congenital abnormalities in 2 studies. Study duration ranged from 2 weeks to 12 months. 8 studies compared conventional treatment (laxatives, dietary advice, and toilet training) plus biofeedback with conventional treatment alone. The groups did not differ for the number of children who were not cured or improved at 12 or 18 months (Table). 1 study compared biofeedback training plus laxatives with biofeedback alone. More children in the biofeedback-plus-laxatives group than in the biofeedback-alone group were not cured or improved at 12 weeks and 12 months (Table). 2 studies compared B-mod plus laxatives with B-mod alone. The laxative group had fewer unsuccessful children than did the B-mod–alone group at 6 and 12 months (Table). 1 study compared B-mod (diet modifications and scheduled toileting) with mineral oil (laxative); groups did not differ for any outcomes. 1 study compared B-mod plus psychotherapy with B-mod alone; treatment results were similar in both groups. 1 study compared 3 groups: B-mod (incentive programs and toilet training) plus laxatives, laxatives alone, and biofeedback. B-mod plus laxatives were better than laxatives alone for reducing the number of children who were not cured or improved (Table).


In children with defecation disorders, biofeedback does not add benefit to conventional treatment. The combination of behavioral modification techniques and laxatives is more effective than either intervention alone.

Source of funding: National Health Service.

For correspondence: Ms. M. Brazzelli, University of Aberdeen, Aberdeen, Scotland, UK. E-mail

Table. Unsuccessful rates for cognitive and behavioral interventions for child defecation disorders*

Comparisons Follow-up Weighted event rates RRI (95% CI) NNH (CI)
Conv + biofeed vs conv 12 mo 52% vs 45% 16% (−4 to 40) Not significant
18 mo 52% vs 46% 15% (−11 to 48) Not significant
Biofeed + lax vs biofeed 12 wk 91% vs 56%† 63% (16 to 149) 3 (2 to 10)
12 mo 91% vs 64%† 43% (5.3 to 108) 4 (3 to 28)
B-mod + lax vs B-mod 6 mo 43% vs 60% 28% (5 to 46) 6 (4 to 28)
12 mo 36% vs 51% 30% (3 to 49) 7 (4 to 48)
B-mod + lax vs lax Not reported 15% vs 55%† 73% (36 to 90) 3 (2 to 7)

*B-mod = behavioral modifications; biofeed = biofeedback; conv = conventional treatment (laxatives, dietary advice, and toilet training); lax = laxatives. Other abbreviations defined in Glossary; RRI, RRR, NNH, NNT, and CI calculated from data in article.
†Event rates not weighted.


Encopresis interferes with normal social and psychological development and strains family relationships. Brazzelli and Griffiths have done a great service with this systematic review of one important aspect of the management of this condition. This review provides some evidence that conventional treatment works, although that was not its primary purpose. The combination of B-mod plus laxatives is more effective than B-mod alone or laxatives alone. This finding should be of some comfort to those who treat children with this condition, but the authors point out the appalling lack of scientific evidence about each component of conventional therapy, including the relative efficacy of various types and doses of laxative.

The main new piece of information from this review is the conclusion that biofeedback adds nothing to conventional treatment. Biofeedback trains children to tighten and relax their perineal muscles and increase the efficiency of defecation. This review suggests that even when that objective is achieved, it does not correlate with successful resolution of encopresis. In fact, there seems to be a trend toward worse outcomes with biofeedback, and when the one study that was strongly positive and accounted for the heterogeneity was removed, the outcomes were worse. This finding needs further explanation and study. Is our understanding of the mechanism of this condition incorrect? Or is there a great deal of variability in the skill and manner in which biofeedback was used? The clinical bottom line is that widespread adoption of biofeedback has no role in the care of children with this condition.

The balancing of the specific components of conventional treatment (e.g., laxatives, diet, education, and stress reduction) in specific families remains very much an art rather than a science. A review directed at conventional therapy is indicated but will almost certainly leave many unanswered questions. Since encopresis is not a rare condition, these questions can be addressed with carefully designed studies.

Michael E. Moffatt, MD
University of Manitoba
Winnepeg, Manitoba, Canada