Among 70 patients with arthritis who were receiving satisfactory maintenance therapy with sulindac (300 to 400 mg daily), 64% had no detectable sulindac sulfide (active metabolite) in one to four random urine specimens. However, 36% had 1.0 to 7.8 (mean, 2.2 +/- 1.4) micrograms/ml sulindac sulfide in urine, similar to the therapeutically effective concentrations found in 24 concurrent plasma specimens (1.4 to 9.0 micrograms/ml). Ten patients had sulindac sulfide in only one or two of two to four urine specimens. Thus, 36% of the patients had pharmacodynamically significant concentrations of sulindac sulfide in urine, presumably capable of suppressing the cyclooxygenase pathway responsible for prostaglandin synthesis in the kidney and elsewhere. The findings suggest individual variability in the capacity for renal oxidation of sulindac sulfide to inactive metabolites, perhaps related to genetic or environmental factors or both. These findings may help to explain conflicting reports on the effects of sulindac on urinary prostaglandins and renal function.