Cemented total hip arthroplasties (THAs) were performed through a Charnley transtrochanteric approach in 1162 patients from 1970 through 1986. Trochanteric separations numbered 58 (5%). Rate of nonunion was related to gender, preoperative diagnosis, and prior THA or endoprosthesis. Nonunion patients had lower mean Charnley pain and function mean scores than union patients at the 45- and 49-month examinations, respectively. A Trendelenberg gait was noted in 17% of nonunions compared with 6% of united patients. Revision rates were nearly three times higher in nonunion patients. These results indicate there is a quantifiable risk for a surgeon who chooses to use a trochanteric osteotomy.