Unicompartmental knee replacement: a comparison of constrained and unconstrained designs.

Seventy-six patients who had eighty-seven unicompartmental knee replacements were followed for an average of fifty-three months (range, two to twelve years). The operation was on the medial side in eighty-two knees and on the lateral side in five. Fifty replacements were unconstrained and thirty-seven were constrained. Of the fifty knees that had an unconstrained replacement, forty-nine (98 per cent) had a good or excellent result, compared with only twenty-six (70 per cent) of the knees that had a constrained replacement; the difference is significant (p = 0.0007). No knee that had an unconstrained replacement had a poor result, compared with nine (24 per cent) of the knees that had a constrained replacement (p = 0.0009). Four (8 per cent) of the fifty knees that had an unconstrained replacement later had a revision total knee arthroplasty, compared with ten (27 per cent) of the thirty-seven knees that had a constrained replacement; the difference is significant (p = 0.04). Noteworthy degenerative changes in the opposite compartment occurred in only one of the eighty-seven knees (a knee in which an unconstrained prosthesis had been inserted).