Performance of the Fresno-Quebec Rule in identifying patients with concomitant fractures not requiring a radiograph before shoulder dislocation reduction: a multicenter retrospective cohort study

Abstract

Background and importance Although shoulder dislocation diagnosis is often solely based on clinical examination, physicians may order a radiograph to rule out a concomitant shoulder fracture before performing reduction. The Fresno-Québec decision rule aims to identify patients requiring a radiograph before reduction to avoid unnecessary systematic imaging. However, this novel approach needs further validation. Objective To evaluate the performance of the Fresno-Québec rule in identifying patients who do not require a prereduction radiograph and assess the variables associated with a clinically significant fracture. Design, settings, and participants A multicenter, retrospective cohort study from 2015 to 2021. Data were extracted from three ED university-affiliated tertiary-care centers. Patients aged ≥18 years with a final diagnosis of anterior glenohumeral dislocation were included. Outcomes measure and analysis Accuracy metrics [sensitivity (Se), specificity (Sp), positive (PPV), negative predictive value (NPV), positive likelihood ratio (PLR) and negative likelihood ratio (NLR)] of the Fresno-Québec rule were measured. Multivariable logistic regression model was used to identify variables associated with the presence of a concomitant clinically significant fracture. Main results A total of 2129 patients were included, among whom 9.7% had a concomitant fracture. The performance metrics of the Fresno-Québec rule were as follows: Se 0.96 95% confidence interval (0.92–0.98), Sp 0.36 (0.34–0.38), PPV 0.14 (0.12–0.16), NPV 0.99 (0.98–0.99), PLR 1.49 (1.42–1.55) and NLR 0.12 (0.06–0.23). A total of 678 radiographs could have been avoided, corresponding to a reduction of 35.2%. Age ≥40 years, first dislocation episode [odds ratio (OR) = 3.18 (1.95–5.38); P < 0.001], the following mechanisms: road collision [OR = 6.26 (2.65–16.1)], low-level fall [OR = 3.49 (1.66–8.28)], high-level fall [OR = 3.95 (1.62–10.4)], and seizure/electric shock [OR = 10.6 (4.09–29.2)] were associated with the presence of a concomitant fracture. Conclusion In this study, the Fresno-Québec rule has excellent Se in identifying concomitant clinically significant fractures in patients with an anterior glenohumeral dislocation. The use of this clinical decision rule may be associated with a reduction of approximately a third of unnecessary prereduction radiographs.