This study reviews 356 consecutive operations for the insertion of femoral-head prostheses. Early and late results are analyzed, and in their light, tentative conclusions are offered as to indications, choice of prosthesis, and technique.

Of the 356 prostheses used, the types were Eicher, four; Judet, sixty-two; Frederick Thompson, 127; and Austin Moore, 163. Of all these, 286 were inserted for fracture of the femoral neck or fracture complications: ninety-five for recent fractures, 111 for non-union, and eighty for avascular necrosis. The remaining seventy were inserted for miscellaneous reasons—arthritis of various types, idiopathic avascular necrosis, or revision of previous operations.

We followed 232 of these hips for more than one year after operation; the average follow-up was 40.5 months. The results for these 232 were classified as to type of prosthesis and the indication for operation. The Eicher and Judet types are discussed only briefly, the Eicher because of the small number used and the Judet because were discontinued its use in 1954. The three major complications with the Judet prosthesis were breakage of the prosthesis, loosening of its stem in the femur, and foreign-body reaction to the acrylic.

The 290 operations using the Thompson or Moore type prostheses are discussed in detail. In increasing order of frequency the incidence of good or excellent results was for recent fractures of the femoral neck, 84.8 per cent; for avascular necrosis, 79.5 per cent; and for non-union after fracture, 71.4 per cent. The results obtained in the miscellaneous group were much poorer (46.5 per cent).

Early complications encountered with the Moore and Thompson types of prosthesis included infection (2.1 per cent), dislocation (1.0 per cent), and fracture of the femur during operation (4.5 per cent). The mortality rate for the first six weeks was 1.4 per cent. Late complications found among the 183 patients with Moore and Thompson prostheses, whom we were able to follow for a year or more, were loosening of the stem in the femoral shaft, distal migration of the prosthesis, and intrusion of the head of the prosthesis into the pelvis. Because loosening and migration distally were encountered more commonly with the Thompson type of prosthesis, we now prefer the Moore type provided that the neck remnant is long enough to allow-seating at the proper level. When the neck is too short, we use the Thompson. Intrusion of the prosthesis into the pelvis was seen with about equal frequency with both types and was usually associated with pain and a poor result. We think that the incidence of this serious complication (roughly 15 per cent in this series) would be higher in a series followed longer: this complication is a major objection to the indiscriminate use of prostheses.

The choice of surgical approach should vary with the condition of the patient. A flexion contracture of the hip of more than 20 degrees is one such condition of special importance.

This study leads us to think that the principal indications for use of a femoral-head prosthesis are non-union of the femoral neck and avascular necrosis of the femoral head after fracture. We also think that there is a limited place for use of a primary prosthesis in acute displaced femoral-neck fractures, but that for this condition the choice of this method must be highly individualized, not based on age alone. It seems clear from the analysis of our results that although the femoral-head prosthesis has a valuable place in our resources, it is (as Stinchfield has said) no panacea, for even when chosen with caution and inserted with skill, serious complications result.