# Predictors of Peripheral Neuropathy in Metabolic Disease: A Multivariable Analysis Incorporating the Toronto Clinical Scoring System and Sudomotor Assessment

## TL;DR

This study identifies age as a key predictor of peripheral neuropathy in diabetes patients and suggests sudomotor assessment can help diagnose the condition.

## Contribution

The study introduces a multivariable analysis combining the Toronto Clinical Scoring System and sudomotor assessment to predict peripheral neuropathy in diabetes.

## Key findings

- Age was the strongest independent predictor of peripheral neuropathy, with each additional year increasing the odds by 6%.
- Sudomotor assessment showed strong diagnostic ability for diabetic neuropathy with an AUC of 0.816.
- Adding sudomotor assessment improved the predictive model's performance.

## Abstract

Background and Objectives: Peripheral neuropathy (PNP) is a frequent and debilitating complication among patients with diabetes mellitus (DM) and other metabolic conditions, substantially affecting morbidity, functional status, and quality of life. Identifying predictors of PNP is essential for optimizing early diagnostic strategies and improving long-term management outcomes. The aim of this study was to determine the predictive factors of PNP in a cohort of patients with DM. Materials and Methods: A cross-sectional study including 117 patients diagnosed with DM assessed for PNP was conducted. All patients were evaluated clinically and biologically. PNP was clinically assessed using the Toronto Clinical Scoring System (TCSS) score and sudomotor function by Sudoscan. Results: The patients included were mostly males with type 2 DM and metabolic syndrome phenotypes. Moreover, the patients with PNP were much older than those without PNP (65 [57–69] vs. 59.50 [46–68] years, p = 0.008), with a longer duration of DM (10 [6–15.50] vs. 5.5 (2–14] years, p = 0.019), and associated autonomic diabetic neuropathy (χ2 = 24.382, p < 0.001). Furthermore, TCSS and Sudoscan were correlated with a history of PNP, especially Sudoscan, which showed a very good discriminative ability for diabetic neuropathy diagnosis (AUC = 0.816). In a multivariable logistic regression including age, DM duration, and HbA1c, age was independently associated with PNP, with each additional year increasing the odds of neuropathy by approximately 6% (OR = 1.06, 95% CI 0.02–1.09, p = 0.002). When age was excluded, DM duration showed a borderline association with PNP (OR = 1.055, CI95% 0.997–1.117), suggesting potential overlap between these variables. Adding sudomotor assessment to the initial model improved the model performance (AUC 0.70–0.72). Conclusions: Age emerged as the main independent predictor of diabetic neuropathy, highlighting the role of cumulative metabolic exposure in the development of neural damage. Moreover, sudomotor assessment may have a complementary role in PNP assessment.