Antidiabetic Treatment Intensity and Diabetic Peripheral Neuropathy in a County-Level Metabolic Management Center: A Real-World Cross-Sectional Study

INTRODUCTION: Diabetic peripheral neuropathy (DPN) is a common disabling complication of type 2 diabetes (T2DM), but real-world data from Chinese county-hospital populations are limited. This study estimated the prevalence of DPN and its associated risk factors, including antidiabetic treatment intensity, in a county-level Metabolic Management Center (MMC) cohort. METHODS: This cross-sectional study enrolled 2420 consecutive outpatients and inpatients with T2DM. DPN was defined as large-fiber neuropathy confirmed by nerve conduction study/electromyography (NCS/EMG) using a standardized MMC protocol; diagnosis did not require neuropathic symptoms or signs. Multivariable logistic and ordinal regression were used to examine associations, with treatment intensity modeled as both a three-tier ordinal and categorical variable. Prespecified sensitivity analyses with progressively extensive covariate adjustment assessed confounding by indication. RESULTS: The prevalence of NCS/EMG-defined DPN was 12.5%. Higher antidiabetic treatment intensity was associated with increased odds of DPN (per-tier ordinal odds ratio (OR) 1.86, 95% CI 1.50-2.30, P < 0.001), with similar effect sizes across sensitivity analyses. History of hypothyroidism (OR 1.796, 95% CI 1.183-2.727, P = 0.006), greater neck circumference (OR 1.022 per cm, 95% CI 1.004-1.041, P = 0.020) and lower free thyroxine (fT4, OR 0.967 per pmol/l, 95% CI 0.940-0.994, P = 0.018) were also associated with DPN. Recorded diabetes duration (adjusted OR (aOR) 1.000, 95% CI 0.998-1.001, P = 0.818) and single baseline (HbA1c, aOR 0.979, 95% CI 0.926-1.036) showed no clear independent association, which we interpret cautiously given uniformly poor glycemic control and the limitations of single-timepoint measurement. CONCLUSIONS: In this county-hospital T2DM cohort, higher treatment intensity, hypothyroidism, and greater neck circumference were independently associated with NCS/EMG-defined large-fiber DPN and may inform risk stratification in resource-limited settings. The 12.5% prevalence reflects electrophysiological large-fiber abnormality rather than clinically defined DPN per Toronto criteria, and all observed associations are non-causal and hypothesis-generating.