Abnormal QT responses to adenosine in subjects with long-QT syndrome

In the interesting article by Viskin et al.,1 adenosine-induced bradycardia was found to produce a higher increase in QT and a lower decrease in QTc (calculated using the Bazett formula) in subjects with long-QT syndrome (LQTS) than in controls. However, a major issue in this study concerns the reliability of QT correction when large ranges of heart rates are analysed. The Bazett formula overcorrects at slow rates and undercorrects at high rates.2 More generally, the accuracy of correction formulas is limited by the high inter-individual variability of the QT/RR relation,3 so that they should only be used for an approximate adjustment of QT over a narrow range of heart rates.4 In the study by Viskin et al.,1 mean RR interval ranged from 580 to 1670 ms in the LQTS group, and from 550 to 2240 ms in the controls. These wide ranges suggest that caution is required in interpreting the observed between-group differences in QTc. Moreover, the appropriateness of QT correction during abrupt changes in heart rate should be considered critically. Most correction formulas utilize models obtained by rest electrocardiograms in subject cohorts, but do not describe the electrophysiological process of delayed adaptation of repolarization to rapid changes in heart rate, i.e. QT hysteresis. Although QT correction has been used in studies on LQTS patients under dynamic conditions such as epinephrine administration, measurements were generally taken at steady state during infusion.5 The effect of QT hysteresis throughout the sudden adenosine-induced deceleration–acceleration sequence may be reasonably relevant. Therefore, when QT is adjusted to the preceding RR interval in these conditions, the resulting value does not represent an effectively rate-corrected QT (i.e. it does not reflect the QT expected at 60 bpm). Also, this value has controversial biological meaning, because it is derived by applying formulas calculated in steady-state conditions to data recorded in non-steady-state conditions. An alternative analysis could be performed using the QT/RR plot obtained over a few minutes of recording during the test. This may allow (i) avoidance of bias due to QT correction over large RR ranges; (ii) quantification of QT hysteresis by evaluation of the QT/RR loop; and (iii) estimation of QT/RR slope, obtained either from raw data or after QT lag compensation by resynchronization of QT and RR changes.6 This analysis might also be helpful in discriminating different LQTS genotypes. QT hysteresis could differ between genotypes, as shown by different dynamic responses to sympathetic stimulation.7 Also, the higher increase in QT at maximal bradycardia in the LQTS group than in the controls is in accordance with studies showing steeper QT/RR slope in LQTS patients than in healthy subjects.8 However, current evidence suggests that an increased QT/RR slope exists in LQT2 and particularly in LQT3 patients, but not in those with LQT1 genotype, as in these subjects, a paradoxical prolongation in QT occurs at fast rates.9,10 On the basis of these considerations, dynamic assessment of QT rate dependence and hysteresis could be clinically intriguing and may improve the diagnostic accuracy of the adenosine challenge test.