Mirels Scores in Patients Undergoing Prophylactic Stabilization for Femoral Metastatic Bone Disease in the Veterans Administration Healthcare System

Introduction: A retrospective review was performed for patients in the Veterans Administration Healthcare System who underwent prophylactic stabilization of the femur for metastatic disease. The goal was to evaluate indications for prophylactic stabilization through Mirels criteria. Methods: All veterans who underwent inpatient prophylactic femoral stabilization between October 2010 and September 2015 were identified. Procedure and demographic variables were collected by using chart review. Provider notes and radiographs were reviewed to calculate Mirels scores. Results: Ninety-two patients underwent confirmed prophylactic stabilization for metastatic bone disease. Lung cancer and multiple myeloma accounted for most lesions. The mean Mirels score was 10.3 (range 7 to 12). 3.2% of patients had a score of 7, and 6.5% had a score of 8. Most lesions were peritrochanteric (66%) and lytic (85%). There was more variability in size (mean 2.3), with 15% being under one third of bony width and 38% between one and two-thirds. The mean pain score was 2.5; 73% reported functional pain. Of lytic and peritrochanteric lesions (53% of all cases), 55% were less than two-thirds the cortical width and 31% lacked functional pain. Conclusion: This retrospective study of prophylactically stabilized metastatic lesions revealed that more than 90% of patients had Mirels scores greater than 8, suggesting a substantial risk of pathologic fracture. Over half of all stabilized lesions were peritrochanteric and lytic. These criteria alone achieve a minimum Mirels score of 8; however, one-third of these lacked functional pain. Notably, Mirels' original paper found location and type criteria to be the least predictive of impending fracture. Contrariwise, functional pain was the most accurate predictor. Multiple studies have found poor specificity of the Mirels criteria. The high scores achievable by the location and type criteria may represent an overrepresentation of their contribution to fracture risk. Reconsideration of the relative weights of each criterion warrants further examination.