The purpose of this study was to classify and characterize responders and non-responders to pelvic floor muscle (PFM) exercise in treatment of female stress urinary incontinence (SUI). Fifty-two women with SUI, mean age 45.9 years (24-a), participated in a randomized P N exercise trial for 6 months. Treatment was either intensive exercises (IE) or moderate exercises canied out at home (HE).Each of five parameters used to measure improvement of SUI (pad test, conversion of a closure pressure [cough] from negative to positive, the patient's own assessment, leakage index, and social activity index) was given one point if present. Responders to treatment were defined by cumulative score 4 and 5, borderline responders by 2 and 3, and nonresponders by 0 and 1. Classified as responders were 65% in the IE group and 20.6% in the HE group. There were no non-responders but 35% borderline responders in the IE group.The responders in the IE group were significantly older, had a longer history of SUI, a higher body mass index (BMI), stronger PFM, more severe SUI measured by leakage and social activity index before treatment, negative closure pressure, lower resting maximum urethral pressure (MUP), and in addition were more motivated for treatment than borderline responders.The discriminant function classified the responders to treatment by a combination of four initially observed variables: maximum flow rate, PFM strength, duration of SUI symptoms, and score on the leakage index. A11 patients classified as responders by use of the cumulative score in the HE group were correctly classified by the discriminant fUnCtiOn. 0 1992 Wiley-Liss. IncAbbreviations: BMI, body mass index; FUPL, functional urethral profile length; HE, home exercises; IE, intensive exercise; MUP, maximum urethral pressure; PFM, pelvic floor muscle; SUCM, simultaneous urethrocystometry; SUI, stress urinary incontinence; VAS, visual analogue scale.