Purpose  

This study was designed to determine whether biofeedback is more effective than diazepam or placebo in a randomized, controlled trial for patients with pelvic floor dyssynergia-type constipation, and whether instrumented biofeedback is necessary for successful training.

Methods  

A total of 117 patients participated in a four-week run-in (education and medical management). The 84 who remained constipated were randomized to biofeedback (n=30), diazepam (n=30), or placebo (n=24). All patients were trained to do pelvic floor muscle exercises to correct pelvic floor dyssynergia during six biweekly one-hour sessions, but only biofeedback patients received electromyography feedback. All other patients received pills one to two hours before attempting defecation. Diary data on cathartic use, straining, incomplete bowel movements, Bristol stool scores, and compliance with homework were reviewed biweekly.

Results  

Before treatment, the groups did not differ on demographic (average age, 50 years; 85 percent females), physiologic or psychologic characteristics, severity of constipation, or expectation of benefit. Biofeedback was superior to diazepam by intention-to-treat analysis (70 vs. 23 percent reported adequate relief of constipation 3 months after treatment, chi-squared=13.1, P<0.001), and also superior to placebo (38 percent successful, chi-squared=5.7, P=0.017). Biofeedback patients had significantly more unassisted bowel movements at follow-up compared with placebo (P=0.005), with a trend favoring biofeedback over diazepam (P=0.067). Biofeedback patients reduced pelvic floor electromyography during straining significantly more than diazepam patients (P<0.001).

Conclusions  

This investigation provides definitive support for the efficacy of biofeedback for pelvic floor dyssynergia and shows that instrumented biofeedback is essential to successful treatment.