Cardiac autonomic neuropathy in diabetes mellitus: prevalence, risk factors and utility of corrected QT interval in the ECG for its diagnosis.

OBJECTIVES

To study the prevalence and risk factors for cardiac autonomic neuropathy (CAN) and the utility of prolongation of corrected QT interval (QTc) in the ECG to diagnose CAN in patients with diabetes mellitus.

DESIGN AND SETTING

Cross-sectional study conducted among patients attending the diabetic clinic of a teaching hospital.

METHODS

The prevalence of CAN among 100 patients with type 1 and type 2 diabetes mellitus was assessed by the five autonomic function tests by Eving's methodology. The CAN score in each patient and its relationship to the QTc interval were analysed. Possible influences of age, duration of diabetes and coexistent peripheral neuropathy on the occurrence of CAN also were studied.

RESULTS

The prevalence of CAN was 60%. Univariate analysis showed a significant association between CAN and higher age (odds ratio (OR) 15.75), prolongation of QTc (OR 5.55), duration of disease over 10 years (OR 2) and peripheral neuropathy (p<0.001) in patients with type 1 diabetes. Significant risks for CAN among patients with type 2 diabetes were coexistent peripheral neuropathy (OR 14), prolonged QTc (OR 9.75), higher age (OR 7.2) and disease duration over 10 years (OR 1.92) in univariate analysis, but none of them showed independent risk in multivariate analysis. Disease duration over 10 years resulted in QTc prolongation in a significant numbers of cases with type 1 (p<0.001) and type 2 (p = 0.006) diabetes. The sensitivity, specificity and positive predictive value of QTc prolongation for the diagnosis of CAN were 77%, 62.5% and 77% in type 1 and 76.5%, 75% and 81.3% in type 2, respectively. Higher CAN scores correlated with longer QTc intervals (coefficient of correlation 0.73; p<0.001).

CONCLUSIONS

The prevalence of CAN in diabetes mellitus is high. Higher age, longer duration of diabetes and peripheral neuropathy are significant risk factors. QTc interval in the ECG can be used to diagnose CAN with reasonable sensitivity, specificity and positive predictive value.