Radiographic interpretation of carpometacarpal arthroplasty: correlation between radiographic loosening and clinical outcome.

OBJECTIVE: To assess the radiographic appearance and interpretation of loosening in patients following first carpometacarpal joint pyrohemisphere arthroplasty.

MATERIALS AND METHODS: A retrospective review over a 6-year period was performed identifying patients with first carpometacarpal pyrohemisphere implants. All postoperative radiographs were reviewed and compared with clinical information. Loosening was defined as greater than 1 mm of periprosthetic lucency or increasing lucency on serial studies. Adverse clinical outcome was defined as infection, complex regional pain syndrome, subluxation or an inability to return to activities of daily living. Statistical analysis included Cohen's Kappa coefficient to measure inter-reader agreement for radiographic interpretation and the agreement between radiographic loosening and an undesired or adverse clinical outcome.

RESULTS: Thirty-one implants in 26 patients were reviewed. A total of 73 postoperative radiographs were reviewed. The mean age of the study subjects was 58 years. The mean radiographic follow-up was 13 months (range: 1-56). All arthroplasties were performed for osteoarthritis. Only one repeat surgery was performed. There was good agreement amongst readers with regard to radiographic interpretation, but the strength of agreement between radiographic loosening and adverse clinical outcome was poor. The sensitivity of radiography in predicting an adverse clinical outcome was 63%, specificity 65%, positive predictive value 39%, and negative predictive value 83%.

CONCLUSION: Although radiography may be useful in the correct clinical context, it should not be utilized as the sole predictor of adverse clinical outcomes following carpometacarpal arthroplasty.