Acetabular component orientation as an indicator of implant luxation in cemented total hip arthroplasty

OBJECTIVE: To determine the sensitivity and specificity with which acetabular component angles of inclination and version could be used, alone or in combination, to predict luxation of cemented total hip arthroplasties (THA). STUDY DESIGN: Comparison of retrospectively selected cases and controls SAMPLE POPULATION: All THA performed at the University of Florida between 1991 and 1998 with the BioMedtrix system and for which at least 2 months of radiographic follow-up were available. All THA performed at the University of Georgia with the BioMedtrix system which subsequently luxated. METHODS: Acetabular component inclination angle (IA) and acetabular version angle (VA) were determined for each THA. Data were grouped according to outcome - luxation or no luxation - with the luxated cases from the 2 institutions pooled. Receiver operator characteristic (ROC) analysis was used to evaluate decision rules for using IA and VA as tests for detecting postoperative luxation. Sensitivity and specificity for luxation and 95% confidence bounds were computed with selected values of IA and VA as cut-points. RESULTS: The nonluxation group consisted of 68 THA with a median follow-up time of 5 months (range, 2-60 months). The luxation group consisted of 12 THA with a mean time to luxation of 36 days. The nonluxation group had a mean +/- standard deviation (SD) IA and VA of 40.3 degrees +/- 8.9 degrees and 71.1 +/- 13.6 degrees, respectively, whereas the luxation group had a mean +/- SD IA and VA of 34.7 degrees +/- 12.6 degrees and 72.9 degrees +/- 16.6 degrees, respectively. An IA cut-point of 37.8 degrees achieved 58.3% sensitivity and 57.4% specificity. A VA cut-point of 73 degrees achieved 75.0% sensitivity and 51.5% specificity. IA and VA considered simultaneously achieved a 50.0% sensitivity and 88.2% specificity. CONCLUSIONS AND CLINICAL RELEVANCE: ROC analysis indicated that both IA and VA considered individually or simultaneously were poor indicators of luxation. Although extreme values of IA may predict luxation with high specificity, the potential for luxation cannot be excluded based on apparently appropriate values of IA and VA. The results of this study also indicate that a successful outcome is possible with a wide range of acetabular component positions.