Seventy-nine knees (45 patients) with a diagnosis of avascular necrosis of the distal femur treated between 1978 and 1989 were evaluated. All patients had a corticosteroid association (had been treated with >30 mg of prednisone for >2 weeks predating by at least 6 months the onset of avascular necrosis). Thirty-two knees were managed with protected weightbearing and rest. Core decompression was performed at a minimum of 3 months after the onset of symptoms in another 47 knees. The knees treated with protected weightbearing had an average asymptomatic period of only 11 months and all but 6 (18%) proceeded to total knee replacement within 6 years. Core decompression yielded good or excellent results in 73% of the knees at an average followup of 11 years (range, 4-16 years). Of the 13 knees with failed core decompression, 7 were asymptomatic for greater than 5 years. A subset of 26 knees from each group was matched for age, gender, diagnosis, Ficat and Arlet Stage, and length of followup. The matched noncore group had 23% survival as compared with 74% survival in the core group. This long term followup suggests that core decompression may slow the rate of symptomatic progression of avascular necrosis of the knee. In addition, core decompression may extend the symptom free interval in certain patients and may delay the need for more extensive procedures such as total knee arthroplasty.