Accuracy of QT c and QTI for Detection of Autonomic Dysfunction

Background: Correlates of QT interval duration have been described although their effects on its ability to identify autonomic neuropathy have not. Methods: We examined the ability of QT c and QTI to detect pharmacologically simulated autonomic dysfunction (PSAD) in persons without bundle branch block, U waves, or long QT syndrome by reviewing 249 articles published through 1996 describing the influence of adrenergic beta antagonists or atropine on QT duration. Six of the articles described effects of intravenous drug administration on the ECG among 94 individuals in sinus rhythm. Autonomic dysfunction was pharmacologically simulated in a subset of 30 men and women via coadministration of both drugs. We used logistic regression to estimate accuracy of QT c and QTI for PSAD, reported as area under summary receiver operating characteristic curves (AUC [95% Cl]) and sensitivity (95% Cl) of test thresholds with specificity of 0.80. Results: Sensitivity of QT C > 436 ms 1/2 was 0.20 (0.09–0.38) and AUC QTC , 0.54 (0.41–0.66). A QTI < 95% was similarly insensitive, 0.30 (0.16–0.48), and AUC QTI equally low. However, stratum‐specific AUC QTI ' was higher than overall AUC QTI , 0.69 (0.64–0.74) when age mean (37.7 years), 0.77 (0.73–0.81) in males, and 0.87 (0.77–0.97) in participants without history of arrhythmia. Sensitivity of QTI thresholds in these strata, range 0.41 to 0.62, was 2.2 to 7.8 times greater than sensitivity, range 0.08 to 0.19, of equally specific QT c thresholds. In a model combining age (y), gender, and arrhythmia, AUC QTI was 0.88 (0.86–0.90). Conclusions: In isolation, QT c and QTI inaccurately detect PSAD. When confounding is taken into consideration, low QTI identifies PSAD with greater sensitivity and accuracy than high QT c . A.N.E. 1999;4(3):257–266