The accuracy of intra-operative fluoroscopy in evaluating the reduction quality of intertrochanteric hip fractures

Purpose To investigate the capability of intra-operative fluoroscopy to determine the reduction quality of intertrochanteric fractures and to determine which view (anteroposterior [AP] or lateral) can better predict the reduction quality. Methods A retrospective analysis of 128 patients with intertrochanteric fractures. Two observers were asked to independently evaluate the quality of reduction (positive or non-positive support) based on intra-operative fluoroscopy (AP and lateral view). Results based on CT scans were considered as the gold standard. Sensitivity, specificity, percentage of correct interpretations, and agreement were calculated. Results At the first reading, sensitivity, false-negative rate, specificity, false-positive rate, and percentage of correct interpretations were 86.1%, 13.9%, 69.4%, 30.6%, and 79.7% for junior resident and 81.0%, 19.0%, 67.3%, 32.7%, and 75.8% for senior resident (all p > 0.05). It was highly predictive of a reliable cortical support when cortical position in AP view was consistent with that in lateral view (85/85, 100% for junior, and 86/86, 100% for senior). Lateral view was generally predictive of a final cortical support when the position between AP and lateral view was inconsistent (90.7% [39/43] vs 9.3% [4/43] for junior, p < 0.001; 92.9% [39/42] vs 7.1% [3/42] for senior). Comparable results were reported at the second reading. The average inter-observer and intra-observer agreement was 0.670 and 0.654, respectively. Conclusions We should ensure the cortical support for both AP and lateral view are positive intra-operatively as far as possible. Even if this is not possible, we need to make sure the lateral position is positive support.