Predicting the deltoid tuberosity index in proximal humerus fractures using fracture characteristics and patient age: development of the LBQ-PHF score

Abstract

Background: The aim of this study was to investigate (1) whether fracture pattern and age are associated with local bone quality (LBQ), and (2) whether a scoring system based on these variables is able to predict LBQ in proximal humerus fractures (PHF). Materials and methods: A retrospective study was performed of all acute PHF at a Level 2 trauma center with plain radiographs and CT between June 2009 and March 2022. Local bone quality was measured by using the deltoid tuberosity index (DTI). In addition to age and gender, fracture morphology was categorized using the following classification systems: Neer, Resch, AO Foundation/Orthopaedic Trauma Association (AO/OTA), and Hertel/LEGO. Additionally, coronal head alignment was calculated by measuring the head-shaft angle. Results: Only the Resch classification system revealed a significant relationship between fracture type and bone quality, as there was a significant association between coronal head alignment and DTI (p = 0.001). Valgus head alignment was observed significantly more frequent in patients with low bone quality (p = 0.002). Multinomial logistic regression analysis revealed a significant relative risk ratio for age (RRR = 0.97, [95% CI, 0.94-1], p = 0.039) and a non-significant trend for DTI (RRR = 1.26, [95% CI, 0.96–1.64], p = 0.092) for occurrence of anatomic relative to valgus head alignment. Using a DTI cut-off value of 1.3 instead of 1.4, age and also varus head alignment were identified as significant predictors of LBQ (OR = 1.12, [95% CI, 1.1–1.15], p < 0.001; OR = 0.54, [95% CI, 0.3–0.96], p = 0.037). A scoring system called the LBQ-PHF score (local bone quality in proximal humerus fractures), developed based on these two variables was able to predict LBQ with a sensitivity of 79.2% and a specificity of 86.7%. Conclusion: Age and coronal humeral head alignment are independent predictors of LBQ in PHF. A simple scoring system developed based on these variables is able to assess BQ with solid predictive characteristics.