To the Editor: In their review of the effects of antipsychotics on the cardiac QTc interval, Alexander H. Glassman, M.D., and J. Thomas Bigger, Jr., M.D. (1), gave a somewhat misleading view. They contended that the QT interval is "modestly associated with torsade de pointes" (p. 1775) and that "the greater the duration, the more likely torsade de pointes becomes" (p. 1776) but cite no references in support of these assertions. An alternative view is that the degree of prolongation of the QT interval is an inadequate and imprecise predictor of the likelihood of arrhythmia. Indeed, this latter view is supportable on theoretical grounds; given the number of processes taking place during the QT interval (depolarization, plateau, repolarization) and the range of ion channels potentially involved (2), it would be remarkable if the extent of lengthening of the QT interval were directly and linearly related to the probability of arrhythmia.

The authors also contended that the QT interval is the "best predictor [of arrhythmia] available" (p. 1775). There is an alternative view here, too. Greater QT dispersion, a measure of repolarization heterogeneity (3), is linked to greater cardiac mortality in certain populations (4, 5) and is held by some to be a more useful indicator of proarrhythmia than are QTc changes (6).

Their conclusions about the relative risks presented by different antipsychotics also seem to rely more on conjecture than on scientific examination. To express certainty about the link between the use of thioridazine and haloperidol and the development of torsade de pointes and sudden death is unfounded. Case reports make this association likely, but in the absence of case-control studies, the link remains no more than probable. Similarly, the frequency of adverse case reports on QTc intervals and different antipsychotics is no basis for unequivocal statements about the relative risk of arrhythmia with different drugs.

The overall impression given by the authors is that clinicians should judge the likely cardiac toxicity of antipsychotics largely by reference to their relative effects on the QT interval. That is, the greater the prolongation of the QT interval, the more dangerous the drug. I believe this suggestion to be based on inappropriate extrapolation and interpretation of available data. The precise truth, I suspect, will turn out to be much more complex than Drs. Glassman and Bigger suggested.