Long-term comparison of the Charnley, Muller, and Trapezoidal-28 total hip prostheses. A survival analysis.

Seven hundred ninety-one total hip arthroplasties (242 Charnley prostheses, 146 Muller prostheses, and 403 Trapezoidal-28) performed between 1969 and 1980 by one surgeon were evaluated using survival analysis to predict success and/or failure over time. Only 6.4% of the population was excluded because of failure to return for follow-up evaluation. All other patients were included in the analysis. It is expected that at 10 years after operation, 91% of the Charnley, 88% of the T-28, and 80% of the Muller prostheses will survive (P less than .05). The acetabular cup of the Muller prosthesis showed significant premature loosening, compared with the Charnley and T-28 prostheses (P less than .0001 and P less than .05, respectively). The Muller femoral stem was inferior only to the Charnley (P less than .025). The T-28 femoral stem showed no difference in success compared with the other two prostheses, despite 10 T-28 femoral stem fractures. Multivariate survivorship regression analysis revealed significant factors that may have predisposed the above failures. Women (P less than .0005) and older patients (P less than .0005) had significantly higher success rates. Significant intrinsic factors included radiolucency about the entire acetabular bone-cement interface (P less than .0005), fracture in the acetabular cement (P less than .005), and radiolucency about the entire femoral bone-cement interface (P less than .0005).