We studied the incidence, indications, and changing patterns of total hip arthroplasty (THA) revision by comparing all consecutive THA revisions performed at a single institution between July 1989 and June 1991 with all those performed at the same institution between January 2000 and December 2000. Aseptic loosening was the main reason for THA failure during both periods. Between the periods, the number of revisions for polyethylene wear, instability, and leg-length discrepancy increased, and the number of revisions for conversion of failed hemiarthoplasty to THA and for deep infection decreased significantly. We evaluate the reasons for these trends in an attempt to identify preventable causes of failure and to stimulate further research on prevention.