This study assessed the effects of postoperative exercise and intra-articular polysulfated glycosaminoglycan (PSGAG) on the repair of osteochondral defects in the carpal joints of ponies. Eighteen ponies with normal carpi had osteochondral defects (mean dimensions 2.4 cm x 0.9 cm) created arthroscopically on the dorsal aspect of the distal articular surface of the radial carpal bone. The ponies were randomized (while balancing for age [range, 2 to 15 years; median, 5.0 years]) to two groups--nine ponies were exercised and nine were stall confined. Beginning at surgery, six ponies in each group received five weekly intra-articular injections of PSGAG (250 mg) in one joint and lactated Ringer's solution in the contralateral joint; the remaining three ponies in each group received lactated Ringer's solution in both joints. The incremental exercise schedule on a circular, rotating walker was begun six days after surgery and occurred twice daily, reaching a maximum of 0.7 miles of walking and 2.7 miles of trotting by the third postoperative month. The effects of treatment on the joint tissues were determined by weekly lameness examinations and measurement of the range of carpal joint motion, carpal radiographs at six and 17 weeks after surgery, synovial fluid analysis, and cytologic evaluation of alcohol-fixed synovial fluid specimens at weeks 1 through 4 and week 17, and histology of the synovial membrane. Ultrasound images of the carpi were acquired before operation and at weeks 1, 2, 4, 8, 10, 13, and 17. Ponies were euthanatized 17 weeks after surgery. Exercise, without medication, caused more lameness throughout the study compared with no exercise. Exercised, nonmedicated ponies had the greatest limitation to carpal flexion (more painful joints), and nonexercised, nonmedicated (control) ponies had the least limitation to flexion. Radiographic scores indicated that the exercised, nonmedicated ponies had significantly (p < .05) more signs of osteoarthritis than exercised, medicated and control ponies. Ultrasonographic measurements indicated that exercise, without medication, caused the greatest increase in combined measurement of the joint capsule thickness and synovial fluid accumulation at all postoperative times. Synovial lining cell numbers in the synovial fluid from exercised ponies were significantly (p < .05) higher than in nonexercised ponies at week 1, and this trend continued at weeks 4 and 17 (p < .1). There were significantly (p < .05) more morphologic abnormalities in the synovial lining cells from exercised than from nonexercised ponies at week 17. Medication with PSGAG enabled exercised carpal joints to be flexed significantly further from weeks 2 through 6 compared with nonmedicated joints.(ABSTRACT TRUNCATED AT 400 WORDS)