Seventy-five women with rheumatoid arthritis according to the 1987 criteria of the American Rheumatism Association were examined. Mean age was 61.9 +/- 12.5 years, mean disease duration 14 years. Sixty-three were or had been on steroids (median cumulative prednisolone dose 2.5 g). Maximal voluntary knee extensor and flexor strength (Nm) was assessed at 30 degrees/s by an isokinetic dynamometer. Walking ability was expressed as walking and stair-climbing time (s). Markers of disease activity included number of swollen and tender joints, pain as recorded by the patients on a visual analogue scale (VAS), and disability as scored by the Stanford Health Assessment Questionnaire (HAQ). Muscle strength, walking time (50 m) and stair-climbing time were reduced on average by 30%, and increased by 28% and 54% (p<0.0001), respectively, compared to 67 age-, weight- and height-matched healthy women. Associations between muscle strength and cumulative or current steroid dose were not found after correction for age and disease duration. Significant linear correlations were found between knee extensor strength and walking time (r =-0.78, p<0.0001) and stair-climbing time (r = -0.76, p<0.0001). Similar correlations were found for flexor strength. The correlations remained significant (Rpartial ranging from -0.64 to -0.69, p<0.0001) in multiple regression analyses adjusting for age, height, weight, disease duration, number of swollen and tender joints, and VAS and HAQ scores. In conclusion, negative effects of steroids on muscle strength were not demonstrated. Leg muscle strength is an important and independent determinator of walking ability in RA.