Development and Validation of an Outpatient Clinical Predictive Score for the Diagnosis of Duchenne Muscular Dystrophy/Becker Muscular Dystrophy in Children Aged 2–18 Years

Abstract

Abstract Introduction: There is no bedside clinical examination-based prediction score for Duchenne muscular dystrophy/Becker muscular dystrophy (DMD/BMD) in children with neuromuscular diseases (NMDs) presenting with proximal limb-girdle weakness. Methods: We compared the details of 200 cases of lower motor neuron type of weakness and had some proximal limb-girdle muscle weakness and divided them into 2 groups: with/without a confirmed diagnosis of DMD/BMD. We determined the predictive factors associated with a diagnosis of DMD/BMD using multivariate binary logistic regression. We assessed our proposed prognostic model using both discrimination and calibration and subsequently used the bootstrap method to successfully validate the model internally. Results: A total of 121 patients had DMD/BMD and the rest of the patients had other diagnoses. Male gender, presence of Gower’s sign, valley sign, toe walking, calf pseudohypertrophy, and tongue hypertrophy were independent predictors for a confirmed diagnosis of DMD/BMD and included in the final CVT2MG score (Calf pseudohypertrophy, Valley sign, Toe walking, Tongue hypertrophy, Male gender, and Gower’s sign). The final model showed good discrimination (AUC = 87.4% [95% CI: 80.5–92.3%, P < 0.001]) and calibration (P = 0.57). A score of 6 or above appeared to be the best cutoff for discriminating between the DMD/BMD group and the rest of the group with both sensitivity and specificity of 98%. The interrater reliability was almost perfect between two pediatric neurologists and strong between a pediatric neurologist and a pediatric neurology trainee resident (k = 0.91 and 0.87). Conclusion: The CVT2MG score has good sensitivity and specificity in predicting a confirmed diagnosis of DMD/BMD in subsequent tests.