We reviewed the results in 43 knees in 39 patients (8 male and 31 female) who had recurrent patellar dislocation and subluxation and were treated by a modified Elmslie-Trillat method (lateral retinacular release, medial transfer of the patellar tendon without advancement, plication of the medial retinaculm, and advancement of the vastus medialis) . Four of the female patients had bilateral dislocations. Four patients were habitual and 35 recurrent. The mean age at the time of operation was 18.4 years (range, 14 to 36) . The results were analyzed after follow-ups ranging from a minimum of 6 months to 12 years and 3 months (average; 3 years, 8 months) .
Using MacNab's (1952) score, the patients were evaluated for both subjective and objective functions. There were good or excellent results in 32 patients (75%) and in 1 patient (2%) the result was rated a failure. There were no significant complications.
The study confirmed that a tangential radiograph of the patellofemoral joint, made with the knee in 30° flexion, is reliable in determining patellar displacement. But no significant relationship was noted between the roentgenographic patellofemoral relationships and the results. In 10 knees we measured the torsional alignment of the lower leg by computed tomography. Two of the 8 patients who had torsional malalignment were graded poor or fair. We concluded that a patient who has a torsional malalignment must be considered carefully to determine whether or not the Elmslie -Trillat method is indicated.
