External validation of a clinical risk score for the presence of cardiovascular autonomic neuropathy in type 1 diabetes

Aims: To validate in an independent external population a CAN Risk Score previously developed in type 1 diabetes (T1D) and validated for cardiovascular autonomic neuropathy (CAN) with a good diagnostic accuracy. Methods: Forty-seven participants with T1D (age 47.7 ± 13.2 years, duration of diabetes 30.0 (19.0–40.5) years, 24 males) underwent 4 cardiovascular reflex tests (CARTs) to diagnose early and confirmed CAN (according to 1 or 2 abnormal results). CAN Risk Score was calculated from resting heart rate, HbA1c, retinopathy and/or nephropathy, cardiovascular disease, HDL cholesterol, systolic blood pressure and smoking (range 0–10). Results: Eleven participants (23.4 %) had CAN. The CAN Risk Score was higher in subjects with overall CAN (early and confirmed) (P = 0.0498) and with confirmed CAN (P = 0.0142) compared to those without, and correlated with CARTs severity (rho = 0.32, P = 0.026), Expiration/Inspiration ratio (r = −0.33, P = 0.0258) and Valsalva ratio (r = −0.47, P = 0.0015). A CAN Risk Score ≥ 4 was found in 19 participants and was associated with the presence of confirmed CAN (P = 0.0129). The CAN Risk Score showed an area under the ROC curve (AUC) of 0.802 ± 0.080 for confirmed CAN, and at the cut-off of 4, sensitivity, specificity and negative predictive values of 85.71 %, 67.50 % and 96.43 %. Conclusions: This study confirmed the diagnostic value of the CAN Risk Score and supports its inclusion in a diagnostic algorithm to identify candidates for CARTs, thereby reducing universal screening. Using routinely available clinical data as categorical variables, the score is easy to calculate and implement in clinical settings.