Dislocations and the femoral head size in primary total hip arthroplasty

The dislocation rate of 3197 Charnley prostheses with 22 mm head in which the surgery was done between 1979 and 1991 in 2 orthopaedic centers was compared with that of 2875 Lubinus prostheses with 32 mm head in which the surgery was done between 1980 to 1991 in 3 other centers. A 1-year followup showed an equal rate of dislocation (2.4%-2.5%) in the 2 groups and included 75% of the 201 dislocated hips. Almost all of the late dislocations occurred with the Charnley prosthesis, resulting in a total dislocation rate of 3.7% compared with 2.9% with the Lubinus prosthesis. Regardless of the type of prosthesis used, there was a higher risk of dislocation in patients with nonhealed hip fractures and in arthroplasties performed by less experienced surgeons. When these 2 variables were removed, the small femoral head was not associated with an increased risk of dislocation. However, there were 77 of 118 (65%) recurrent dislocations in the Charnley group, compared with 37 of 83 (45%) in the Lubinus group, and the relative risk of a dislocated hip arthroplasty becoming recurrent increased by 2.3 times if the small femoral head was used. The number of reoperations also were doubled in this group. Almost 4 times as many dislocations were documented within 2 weeks after surgery after any type of prosthesis inserted through a posterior approach compared with the transtrochanteric approach, but there was no increase in rate of recurrence or revision.