The lateral subvastus approach combined with an osteotomy of the tibial tubercle is a recognised, but rarely used approach for total knee replacement (TKR). A total of 32 patients undergoing primary TKR was randomised into two groups, in one of which the lateral subvastus approach combined with a tibial tubercle osteotomy and in the other the medial parapatellar approach were used. The patients were assessed radiologically and clinically using measurement of the range of movement, a visual analogue patient satisfaction score, the Western Ontario McMasters University Osteoarthritis Index and the American Knee Society score. Four patients were lost to the complete follow-up at two years. At two years there were no significant differences between the groups in any of the parameters for clinical outcome. In the lateral approach group there was one complication due to displacement of the tibial tubercle osteotomy and two osteotomies took more than six months to unite. In the medial approach group, one patient had a partial tear of the quadriceps. There was a significantly greater incidence of lateral patellar subluxation in the medial approach group (3 of 12) compared with the lateral approach group (0 of 16) (p = 0.034), but without any apparent clinical detriment. We conclude that the lateral approach with tibial tubercle osteotomy is a safe technique with an outcome comparable with that of the medial parapatellar approach for TKR, but the increased surgical time and its specific complications do not support its routine use. It would seem to be more appropriate to reserve this technique for patients in whom problems with patellar tracking are anticipated.