Regression equation analysis enhances detection of conduction slowing beyond axonal loss in diabetic neuropathy

Abstract

Objectives: To evaluate the utility of regression analysis in the assessment of conduction slowing in diabetic distal symmetrical polyneuropathy (DSP) and in identifying superimposed demyelination beyond fiber loss. Background: Causes of conduction slowing beyond pure axonal loss has been attributed to an additional demyelinating component. We therefore evaluated the utility of regression analysis in the assessment of conduction slowing in diabetic DSP and in identifying superimposed demyelination beyond fiber loss. Methods: We previously established regression analysis to develop confidence intervals that assess the range of conduction slowing from primary demyelination in patients with chronic inflammatory demyelinating polyneuropathy (CIDP). In this study, by using the regression equations, we analyzed conduction slowing in patients with diabetic DSP. Results: Mean conduction velocity (CV) was significantly slower in diabetic DSP than in the non-diabetic DSP for all tested nerves. More patients were found to fulfill the regression equation criteria in the diabetic group compared to the non-diabetic group (47.0 % vs. 23.3 %). The estimated likelihood of having more than two motor nerves with CV slowing in the demyelination range by American Academy of Neurology or regression equations criteria was significantly higher in the diabetic DSP (0.73) compared to non-diabetic DSP (0.52). Conclusions: Conduction slowing in diabetic DSP beyond what is expected exclusively from axonal loss could be identified by regression analysis.

Keywords

• Regression analysis
• Diabetic neuropathy
• Amyotrophic lateral sclerosis
• Conduction velocity slowing
• Demyelination