The results were retrospectively analyzed of 47 subtrochanteric fractures of the femur treated with a 95 degrees condylar blade-plate to establish whether two different surgical techniques yielded different results. Before 1981, treatment consisted of extensive visualization of the fracture lines, permitting anatomic reduction of all fragments, stable internal fixation with the blade-plate, and optional autologous bone grafting as recommended by the AO group. Twenty-four fractures were treated accordingly and constituted Group I of this study. In 1981, visualization of the fracture lines was abandoned, especially at the medial cortex; an indirect reduction technique was used to gain optimal alignment and stability without aiming at anatomic reduction, and bone grafting was discontinued. Twenty-three patients were treated accordingly and constituted Group II. The use of prophylactic antibiotics as a routine for all major trauma was instituted at the time the surgical technique was changed. Thus, only two of the 24 patients in Group I received antibiotics as opposed to 20 of the 23 patients in Group II. Average time to bony union for those fractures that healed primarily was 5.4 months in Group I and 4.2 months in Group II. Delayed or nonunion was 16.6% in Group I and 0% in Group II, and the infection rate was 20.8% versus 0% in the two groups. The four cases with a delayed union were aseptic, but three of the four nonunions were infected. The functional end result was comparable for both groups.(ABSTRACT TRUNCATED AT 250 WORDS)