Facts, fancies and follies of drug-induced QT/QTc interval shortening

Abstract

Parallels exist between drug-induced QT/QTc prolongation and shortening. However, these parallels are largely superficial and the experience with drug-induced QTc prolongation and its potential proarrhythmic link cannot be directly applied to drug-related QTc shortening. The congenital short QT syndrome (SQTS) is clearly much less prevalent than congenital, long QT syndrome, possibly some 1000 times. If the same discrepancy exists between arrhythmic susceptibility to drug-induced QTc prolongation and shortening, it is questionable whether regulatory burden should be imposed on drugs that might cause serious arrhythmia, once in many millions of exposures. Further, majority of torsadegenic drugs block the I_Kr current which is susceptible to the drug blockade because of the corresponding channel geometry. There is no parallel known for drug-induced QTc shortening. Also, all drugs that prolong QTc interval massively cause torsade de pointes tachycardia in more than exceptional isolated instances. On the contrary, digitalis that causes substantial QTc shortening is not known to trigger frequently ventricular arrhythmias. Moreover, most available population QTc data were obtained with Bazett's correction which produces erroneous QTc shortening at slow heart rates. Safety limits derived from such data are inappropriate. Because practically all new drugs undergo the so-called thorough QT study, drug-induced QTc shortening will not go unnoticed for any new pharmaceutical. Describing drug-related QTc shortening in the label seems sufficient to avoid treatment of the rare SQTS subjects. Intensive investigations of QTc-shortening drugs (similar to those of drugs with positive thorough QT studies) do not seem to be warranted.

This article is a commentary on Shah, pp. 58–69 of this issue and is part of a themed section on QT safety. To view this issue visit http://www3.interscience.wiley.com/journal/121548564/issueyear?year=2010

Figure 1

Example of a QT/heart rate and QTc/heart rate distribution in a drug-free recording of a healthy subject. The blue squares show the uncorrected QT interval. The red circles show the Bazett-corrected QTc interval. Note that the range of the QTc (B) values is larger than the range of QT values. While the accurate individualized heart rate correction (green diamonds) leads to a practically horizontal line of QTc ∼384 ms irrespective of heart rate, Bazett's correction leads to artificial QTc shortening at slow and QTc prolongation at fast heart rates. In this data set, the standard deviation of uncorrected QT intervals, Bazett-corrected QTc intervals and individually corrected QTc intervals is 17.2, 24.2 and 3.5 ms respectively.

Figure 2

Examples of QTc shortening on placebo observed in two different unpublished thorough QTc studies. Each graph shows a section of changes in individually corrected QTc intervals versus baseline. These placebo-induced changes are likely to be caused by autonomic conditioning due to the stress of the closed and restricted environment of clinical units.