Prospective evaluation of magnetic resonance imaging and physical examination findings in patients with greater trochanteric pain syndrome

Objective. This study examined patients with greater trochanteric pain synd rome (GTPS) to determine the prevalence of gluteus medius pathology by util izing magnetic resonance imaging (MRI), and to evaluate the presence of Tre ndelenburg's sign, pain on resisted hip abduction, and pain on resisted hip internal rotation as predictors of a gluteus medius tear in this group of patients. Methods. Twenty-four subjects with clinical features consistent with GTPS w ere recruited. A standard physical assessment was performed at study entry, including assessment of the 3 specific physical signs. Following this init ial assessment, MRI of the affected hip was performed. A 1.5T whole body MR I system was utilized, with T1 and T2 fast spin-echo sequences performed in the coronal and axial planes. All MR images were reviewed in random order by a single radiologist. In 12 patients, the 3 physical signs were assessed at study entry and at 2 months by the same observer and the intraobserver reliability for each of the signs was calculated. Results. All subjects were women (median age 58 years, range 36-75 years). The median duration of symptoms was 12 months (range 12-60 months). MRI fin dings were as follows: 11 patients (45.8%) had a gluteus medius tear, 15 pa tients (62.5%) had gluteus medius tendinitis (pure tendinitis in 9 patients and tendinitis with a tear in 6 patients), 2 patients had trochanteric bur sal distension, and 1 patient had avascular necrosis of the femoral head. T rendelenburg's sign was the most accurate of the 3 physical signs in predic ting a tendon tear, with a sensitivity of 72.7% and a specificity of 76.9%. Moreover, Trendelenburg's sign was the most reliable measure, with a calcu lated intraobserver kappa of 0.676 (95% confidence interval 0.270-1.08). Conclusion. The results support the hypothesis that gluteus medius tendon p athology is important in defining GTPS. In this series, trochanteric bursal distension was uncommon and did not occur in the absence of gluteus medius pathology. The physical findings suggest that Trendelenburg's sign is the most sensitive and specific physical sign for the detection of gluteus medi us tears, with an acceptable intraobserver reliability. Further delineation with MRI, especially in patients with a positive Trendelenburg's sign, is recommended prior to any consideration of surgery in this group of patients . Finally, with the pathology of this condition defined, the challenge will be to devise and assess, by randomized controlled trial, an appropriate tr eatment strategy for this group of patients.