Abstract
PURPOSE: The incidence of cardiac autonomic neuropathy (CAN) in patients with type 1 diabetes (T1D) is frequently underestimated. Individuals with T1D and CAN have an increased mortality risk, mainly from cardiovascular causes. The objectives of the present study were to assess the clinical and laboratory characteristics associated with CAN in patients with T1D and verify the ability of multiple clinical factors to help identify patients with this condition. METHODS: 102 patients with T1D were evaluated for CAN using standardized cardiovascular reflex testing. Clinical characteristics were used to compute a numerical score for CAN diagnosis and a ROC curve elaborated for assessment of the best cutoff to predict CAN. This score was then applied to the second sample of 120 patients. The sensitivity, specificity, and positive and negative predictive values were calculated. RESULTS: Prevalence of CAN was around 35% in the first sample of patients and just below 20% in the second sample. Hypertension, total cholesterol, triglycerides, postprandial sweating, diastolic blood pressure, abnormal right and left 10 g monofilament, retinopathy, and nephropathy were considered independent predictors of CAN. The CAN-score cut-off was 16.88. This yielded a sensitivity of 50%, specificity 73.8%, positive predictive value 22.9%, and negative predictive value 90.5%. CONCLUSION: The use of a subset of clinical and laboratory characteristics can be more accessible than the cardiac reflex tests and more accurate than a single isolated characteristic.