We compared the effects of treatment with bipolar and monopolar radiofrequency energy on 30 osteochondral sections harvested from 22 patients with spontaneously occurring chondromalacia who were undergoing knee arthroplasty. Specimens with chondromalacia grades 2 or 3 were randomly assigned to one of two bipolar or one monopolar treatment groups. All samples were marked and mounted on a jig to allow simulation of an arthroscopic surgical procedure with a flow rate of 100 ml/min of a balanced electrolyte solution at 22 degrees C. Under arthroscopic visualization, the designated area was treated until smooth, and the total treatment time was recorded. There was no difference in patients' ages, chondromalacia grade, or cartilage thickness among groups. Significant chondrocyte death, as determined by cell viability staining with confocal laser microscopy, was observed with each group. The bipolar devices produced significantly greater depths of chondrocyte death (2228 +/- 1003 microm and 2810 +/- 517 microm) than did the monopolar device (737 +/- 391 microm). The bipolar devices caused cell death to subchondral bone significantly more often (13 of 20 specimens) than did the monopolar device (0 of 10 specimens). Caution should be used in treating fibrillated cartilage with radiofrequency energy, particularly with the bipolar devices tested.