OBJECTIVES: To determine whether gut focused behavioural treatment (biofeedback) is a useful therapy in multiple sclerosis patients referred for constipation, incontinence, or a combination of these symptoms. Most patients with multiple sclerosis complain of constipation, faecal incontinence, or a combination of the two. Patients rate these bowel symptoms as having a major impact on their life. Until now the management of these problems has been empirical, with a lack of evaluated therapeutic regimes. METHODS: Thirteen patients (eight women, median age 38 years, median duration of multiple sclerosis 10 years) complaining of constipation, with or without faecal incontinence underwent a median of four sessions of behavioural treatment. Anorectal physiological tests were performed before therapy. Impairment and disability were rated with the Kurtzke score and the Cambridge multiple sclerosis basic score (CAMBS). Patients were contacted a median of 14 months after completion of treatment. RESULTS: A beneficial effect was attributed to biofeedback in five patients. Mild to moderate disability, quiescent and non-relapsing disease, and absence of progression of multiple sclerosis over the year before biofeedback were predictive of symptom improvement. No physiological test predicted the response to therapy. CONCLUSION: Biofeedback retraining is an effective treatment in some patients with multiple sclerosis complaining of constipation or faecal incontinence. A response is more likely in patients with limited disability and a non-progressive disease course.