• BACKGROUND
    • We reviewed all revision operations performed by a single surgeon on previous hemi- and total shoulder arthroplasties from November 1987 to March 2005.
  • METHOD
    • Thirty-five patients' charts were reviewed to determine the causes of failures. In addition, their outcomes after the revision surgery were analyzed.
  • RESULTS
    • Overall, results were satisfactory in 71% by Neer criteria at a mean follow-up of 27.6 months. Outcomes were related to reason for failure. When failure was because of glenoid erosion, loosening, or humeral loosening, the mean improvement in forward elevation (FE) (28.9 degrees ) and external rotation (ER) (16.1 degrees ) was significantly better (P=.024 FE; P=.000 ER) than when the failure was because of infection, soft-tissue problems, or pain of undetermined origin (FE=-5.6 degrees ; ER=-6.8 degrees ).Likewise, UCLA scores in the first group were significantly better than in the second group (P=.003). In the first group, 16/18 patients were satisfied, while in the second group only 4/17 were satisfied.
  • CONCLUSION
    • Our data suggest that patients whose revisions are because of glenoid erosion or component loosening can expect to have better outcomes than those whose revisions are performed for infection, instability, or other soft-tissue problems.