Diagnostic Tool Predicting Results of Electrodiagnostic Tests in Patients Presenting With Musculoskeletal Symptoms

Patients presenting with musculoskeletal symptoms are fairly common in any neurologist practice, their presentation raises concern of underlying radiculopathy as a cause of their symptoms. The role of electrodiagnostic study has been variable in such patients. Creating a diagnostic tool will be of most help to provide physicians of predictive value of such tests before ordering and help guide them in finding the most appropriate test.

Retrospective chart review study of 444 patients older than 18 years. The patients were referred to the electrodiagnostic laboratory with musculoskeletal symptoms such as neck, back, or limb pain; extremity numbness; or to rule out radiculopathy.

A total of 308 patients met predefined inclusion criteria. Variables such as age, extremity numbness, back pain, and other symptoms were identified as significant predictors for abnormal results. A diagnostic tool was developed using these variables, scoring each as normal (0) or abnormal (1), with age as the only numerical input. For nerve conduction study (NCS) prediction, significant factors included age [P < 0.01, odds ratio (OR) 0.95], extremity numbness (P = 0.03, OR 0.42), other symptoms (P = 0.04, OR 0.29), and normal magnetic resonance imaging (P = 0.02, OR 2.99). For electromyography (EMG) prediction, age (P < 0.01, OR 0.95), back pain (P = 0.02, OR 0.43), and other symptoms (P = 0.03, OR 0.29) were significant. The logistic regression models for predicting abnormal NCS and EMG exhibited areas under the receiver operating curves of 0.807 and 0.777, respectively. With 95% specificity, the NCS model had a sensitivity of 24.10%, positive predictive value of 68.97%, and negative predictive value of 72.84%, while the EMG model had a sensitivity of 32.14%, positive predictive value of 75%, and negative predictive value of 74.67%.

This study highlights the potential for using a scoring system to predict NCS and EMG outcomes based on key clinical variables.