Association of QT-Prolonging Medications With Causes of Sudden Death

The research reported in the article by Simpson et al examined the association of QT-prolonging medications (QTdrugs list at www.crediblemeds.org) with autopsy-defined causes of sudden death and used trauma victims as a comparison group. The authors confirmed the long-standing association of QT-prolonging drugs and a presumed diagnosis of sudden cardiac death, but they also went a step further to perform what is essentially a subgroup analysis to compare the strength of the association of exposure to these drugs (measured as a novel QTPM score) with either 1) cases adjudicated to be “autopsy-defined SAD” or 2) cases with “autopsy-defined non-SAD.”

Autopsy-defined SAD was considered to be any “death for which there was no identifiable alternative cause (eg, acute cerebrovascular event, tamponade, vascular rupture, pulmonary embolism, occult overdose, acute heart failure with pulmonary edema).” This assumes that the anatomic finding at each autopsy is the only contributor to the patient’s death and that death was not also due to an arrhythmia. The subgroup analysis found that the odds ratio was only significant for an association of QTPM score in the cases with “autopsy-defined non-SAD” and not with “autopsy-defined SAD.” The authors concluded that the autopsy demonstrated that the increased risk (SAD associated with QT-prolonging drugs) was specific for the group with a “non-arrhythmic” but not “arrhythmic” cause of death, and cautioned that prior estimates of SAD with QT-prolonging drugs may be overestimates.

An alternative explanation for these findings could be that drug-induced SAD is more likely to occur when the patient is taking QT-prolonging drugs and also has a clinical event that causes ischemia, hypoxia, heart failure, or stroke, which are known risk factors for drug-induced TdP. Each of these factors would be expected to enhance the risk of TdP with QT-prolonging drugs. The lower odds ratios (1.59–3.03; P=.22) for the SAD subset may well be caused by design weaknesses acknowledged by the authors (incomplete medication exposure, unproven score for QTPM exposure) and the inclusion of the “Conditional Risk of TdP” category of drugs in the QTPM score (22 of the 51 drugs do not prolong QT).

We know from prior experience that conducting subset analyses can be misleading, especially when the subgroups are imbalanced by comorbidities, as are these, by definition. A plausible alternative explanation for the authors’ findings is that patients who suffer coronary ischemia, stroke, heart failure, or a hypoxic event while taking QT-prolonging drugs are at even greater risk of SCD than those only taking QT-prolonging drugs. Based on these data, it seems unwarranted to conclude that the risk of SCD with QT-prolonging drugs may have been overestimated. The results of this study could equally support the conclusion that drug-induced SAD is underestimated because it may often be attributed to a clinical event that produces anatomic findings and an autopsy diagnosis.

Disclosure: This research used this author’s (RLW) web-based QTdrugs lists as a resource.