Cardiovascular autonomic neuropathy (CAN) is a major microvascular complication of type 2 diabetes mellitus (T2DM) that affects cardiovascular regulation and is associated with increased morbidity. This study aimed to identify clinical and laboratory factors associated with CAN in hospitalized T2DM patients and to construct a practical predictive model for individual risk estimation.
The analysis was retrospective and included 831 consecutive hospitalized patients with T2DM treated at Suqian Hospital Affiliated to Xuzhou Medical University between September 2024 and December 2025. The cohort was randomly split 7:3 into a training set (n = 581) and a validation set (n = 250). Within the training set patients were classified by presence or absence of CAN into non-CAN (n = 307) and CAN (n = 274) groups for comparative analyses and model development.
The investigators compared demographic, anthropometric, biochemical, hematologic, comorbidity and medication variables between CAN and non-CAN patients in the training set. Key differences observed (all P < 0.05) included higher values in the CAN group for age, diabetes duration, HbA1c, neutrophil count, platelet count, neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), monocyte-to-lymphocyte ratio, systemic immune-inflammation index (SII) and systemic inflammatory response index. Conversely, the CAN group had lower height, body weight, total and direct bilirubin, alanine aminotransferase, aspartate aminotransferase, fasting C‑peptide and fasting insulin.
Comorbid conditions and prior therapies were more prevalent among CAN patients: diabetic nephropathy, diabetic retinopathy (DR), diabetic peripheral neuropathy, histories of hypertension, cardiovascular disease and stroke, and higher proportions treated with antiplatelet agents and insulin (all P < 0.05).
Candidate predictors were screened using least absolute shrinkage and selection operator (LASSO) regression, which selected seven variables: DR, age, diabetes duration, HbA1c, NLR, PLR, and SII. These candidates were entered into multivariable logistic regression. The final independent influencing factors for CAN identified were DR (OR = 2.29, 95% CI: 1.48–3.56), longer diabetes duration (OR = 1.04 per unit increase, 95% CI: 1.01–1.08), advancing age (OR = 1.06 per year, 95% CI: 1.04–1.09), elevated HbA1c (OR = 1.23, 95% CI: 1.13–1.36), increased NLR (OR = 2.23, 95% CI: 1.36–3.69) and elevated PLR (OR = 1.01, 95% CI: 1.00–1.02). SII was selected by LASSO but was not retained as an independent predictor in the final multivariable model reported in the abstract.
A nomogram prediction model was constructed using the six independent variables (DR, age, diabetes duration, HbA1c, NLR, PLR) to provide an intuitive, individualized estimate of CAN risk for patients with T2DM. The nomogram visualizes the combined contribution of each predictor to estimated risk.
Discrimination: The nomogram demonstrated good discriminative performance with area under the receiver operating characteristic curve (AUC) values of 0.839 (95% CI: 0.807–0.871) in the training set and 0.787 (95% CI: 0.732–0.843) in the validation set. Sensitivity and specificity were 81.4% and 71.7% in the training set, and 71.3% and 73.3% in the validation set, respectively.
Calibration: Calibration curves showed close agreement between predicted and observed CAN outcomes in both datasets. The Hosmer-Lemeshow goodness-of-fit test indicated satisfactory calibration for the training set (χ2 = 6.701, P = 0.461) and the validation set (χ2 = 12.139, P = 0.096).
Clinical utility: Decision curve analysis indicated the model provides net clinical benefit over a wide range of threshold probabilities — 1% to 88% in the training set and 3% to 80% in the validation set — supporting potential applicability for risk-guided decision making.
The results highlight routinely available clinical features and simple inflammatory indices (NLR, PLR) as components of an accessible risk prediction tool for CAN in T2DM. The nomogram could be used in inpatient settings to identify patients at higher risk for CAN and to target monitoring, further autonomic testing, or interventions addressing glycemic control and comorbid vascular complications.
This report is based on a single-center retrospective cohort of hospitalized patients; the abstract does not report prospective validation or external, multicenter testing. The diagnostic criteria or testing methods used to define CAN are not detailed in the abstract. Other potential confounders and the timing or thresholds for laboratory measures are not specified here. The study period, cohort selection criteria beyond hospitalization, and management protocols were described at a high level; further methodological detail requires consultation of the full text.
All authors declared no conflicts of interest in the source report.