Two automatic QT algorithms compared with manual measurement in identification of long QT syndrome
Received 6 February 2009
Abstract
Background
Long QT syndrome (LQTS) is an inherited disorder that increases the risk of syncope and malignant ventricular arrhythmias, which may result in sudden death.
Methods
We compared manual measurement by 4 observers (QTmanual) and 3 computerized measurements for QT interval accuracy in the diagnosis of LQTS:
1.QT measured from the vector magnitude calculated from the 3 averaged orthogonal leads X, Y, and Z (QTVCG) and classified using the same predefined QTc cut-points for classification of QT prolongation as in manual measurements;
2.QT measured by a 12-lead electrocardiogram (ECG) program (QTECG) and subsequently classified using the same cut-points as in (1) above;
3.The same QT value as in (2) above, automatically classified by a 12-lead ECG program with thresholds for QT prolongation adjusted for age and sex (QTinterpret).
The population consisted of 94 genetically confirmed carriers of KCNQ1 (LQT1) and KCNH2 (LQT2) mutations and a combined control group of 28 genetically confirmed noncarriers and 66 unrelated healthy volunteers.
Results
QTVCG provided the best combination of sensitivity (89%) and specificity (90%) in diagnosing LQTS, with 0.948 as the area under the receiver operating characteristic curve. The evaluation of QT measurement by the 4 observers revealed a high interreader variability, and only 1 of 4 observers showed acceptable level of agreement in LQTS mutation carrier identification (κ coefficient >0.75).
Conclusion
Automatic QT measurement by the Mida1000/CoroNet system (Ortivus AB, Danderyd, Sweden) is an accurate, efficient, and easily applied method for initial screening for LQTS.
ABSTRACT