A randomized controlled trial of anorectal biofeedback for constipation

Abstract

The purposes of this study were: (1) to examine the efficacy of anorectal biofeedback (AB) for constipation compared to a biofeedback control (BC) treatment and (2) to examine the extent to which self-reported childhood sexual/physical abuse predicted biofeedback outcome. Twenty-one patients with pelvic floor dyssynergia were randomized to either (1) an AB arm, where patients learned to isolate the anal sphincter using an electromyography probe, or (2) a BC arm that controlled for the nonspecific effects of biofeedback, where patients learned to relax trapezius or temporalis muscles with EMG feedback. Both treatments were delivered by registered nurses for six sessions. Prior to randomization and post-treatment, patients completed the validated Constipation Severity Instrument and two measures of quality of life (QOL), the Irritable Bowel Syndrome-QOL, and the SF-36. Generalized estimating equations examined the within-group and between-group differences over time. Pre- and post-treatment data were obtained for six AB and nine BC patients. AB patients' overall constipation severity scores decreased by 35.5% (vs. 15.3%), and their obstructive defecation symptom scores decreased by 37.9% (vs. 19.7%) compared to BC. A similar pattern was shown on the IBS-QOL. On the SF-36 Mental Health Composite (MCS), AB scores improved 28.0% compared to BC scores, which worsened 12.7%. Those without (vs. with) a childhood sexual/physical abuse history showed improvement on the MCS post-biofeedback. While our sample was statistically underpowered, AB produced clinical improvements in constipation severity and QOL.