ObjectiveTo explore the treatment methed of recurrent patellar dislocation associated with old osteochondral fracture and to evaluate its effectiveness. MethodsBetween August 2010 and August 2014, 12 cases of recurrent patellar dislocation with old osteochondral fracture were treated. There were 4 males and 8 females with an average age of 18.3 years (range, 15-24 years). The left knee was involved in 7 cases and the right knee in 5 cases. All the patients had a history of patellar dislocation, the average interval from injury to first hospitalization was 7.6 months (range, 6-13 months). At preoperation, the range of motion (ROM) of the injured knee was (89.17±13.11)°; the Lysholm score was 56.67±18.91; the Q-angle was (17.50±5.28)°; and tibial tuberosity-trochlear groove (TT-TG) distance was (18.33±4.03) mm. The Q-angle was more than 20° and TT-TG distance was more than 20 mm in 6 of 12 cases. There were 6 cases of patellar osteochondral fracture, 5 cases of lateral femoral condylar osteochondral fracture, and 1 case of patellar osteochondral fracture combined with lateral femoral condylar osteochondral fracture. After osteochondral fracture fragments were removed under arthroscope, lateral patellar retinaculum releasing and medial patellar retinaculum reefing was performed in 2 cases, medial patellofemoral ligament (MPFL) reconstruction combined with both lateral patellar retinaculum releasing and medial patellar retinaculum reefing in 4 cases, and MPFL reconstruction, lateral patellar retinaculum releasing, medial patellar retinaculum reefing, and tibial tubercle transfer in 6 cases. ResultsAll wounds healed by first intention with no complication of infection, haematoma, skin necrosis, or bone nonunion. All patients were followed up 12-60 months with an average of 24.2 months. At 3 months after operation, all patellar dislocations were corrected; the Q-angle was (13.33±1.37)° and the TT-TG distance was (12.17±1.17) mm in 6 patients undergoing tibial tubercle transfer, showing significant differences when compared with preoperative values[(22.50±2.17)° and (21.33±2.34) mm] (t=15.25, P=0.00; t=8.27, P=0.00). All patients achieved relief of knee pain and knee locking; the knee ROM and the Lysholm score at last follow-up were (120.42±11.57)° and 89.25±9.71, showing significant differences when compared with preoperative ones (t=-11.61, P=0.00; t=-8.66, P=0.00). ConclusionIt has satisfactory short-term effectiveness to remove old osteochondral fragments that can not be reset and to correct patellar dislocation for recurrent patellar dislocation with old osteochondral fracture.