Electrophysiological diagnosis using the coefficient of variation of R‐R intervals and a point‐of‐care nerve conduction device is highly correlated to the diagnosis of diabetic polyneuropathy using conventional electromyographs

ABSTRACT Aims/Introduction While treatment for diabetic polyneuropathy (DPN) is still developing, progress has stagnated. The alignment between pathological neurodegeneration in DPN and patients' subjective symptoms is often low, yet these symptoms are frequently used for diagnosis. This reliance has hindered the development of effective drugs to prevent neurodegeneration. This study aims to establish an objective electrophysiological diagnostic method that could support future treatment development and validate its accuracy. Materials and Methods This retrospective multicenter cohort study involved hospitalized diabetic patients. A total of 314 patients underwent nerve conduction studies using standard electromyography and a simplified nerve conduction testing device (NC‐stat/DPNCheck™), along with electrocardiogram‐based coefficient of variation of R‐R intervals (CV R‐R ). Patients with a severity classification of Stage 2 or higher based on electromyography were defined as having DPN. Logistic regression analysis was used to identify significant factors explaining DPN presence, followed by ROC analysis to determine optimal cutoff values for diagnosis. Results Significant factors included resting CV R‐R , sural nerve conduction velocity (SNCV), and amplitude of sensory nerve action potential (SNAP). SNCV had the highest area under the curve (AUC = 0.823). The optimal cutoff values were 1.62% for CV R‐R , 46.5 m/s for SNCV, and 10.5 μV for SNAP. Diagnosing DPN based on abnormalities in two or more of these three conditions yielded an accuracy of 79.3%. Conclusions The established diagnostic criteria of DPN demonstrate high performance and are expected to be applicable in clinical settings.