IN 2001, the US Food and Drug Administration (FDA) posted a safety alert regarding droperidol use indicating that “cases of QT prolongation and/or torsade de pointes (TdP)” had been reported with this drug.*A warning was issued that contraindicated the use of droperidol in patients with known or suspected QT prolongation, imposed the recording of a 12-lead electrocardiogram before administration in all patients to determine whether a prolonged QT interval was present, and recommended that electrocardiogram monitoring be continued for 2–3 h after treatment to monitor arrhythmias. Moreover, administration of droperidol to patients at risk of developing prolonged QT interval (e.g. , patients older than 65 yr or receiving volatile anesthetics or intravenous opiates) was recommended with “extreme caution.” The impact of this warning was reinforced by a consensus guideline published in 2003, which recommended, as a consequence of the warning regarding droperidol, the use of 5-hydroxytryptamine type 3 receptors antagonists as first-line antiemetics.1Although droperidol is still available, in some institutions its use was discontinued, as shown by Nuttall et al. ,2who recorded a dramatic decrease in droperidol use from approximately 12% between 1998 and 2001 to 0% between 2002 and 2005.
Since the FDA warning, controversy has increased, and there has been extensive debate in the anesthesiology literature. Many believe that this warning was unjustified given the efficacy of droperidol as an antiemetic, the lack of published evidence of droperidol-induced arrhythmias during decades of use, and the absence of overt toxicity when administered at low doses. On the other hand, a “precaution principle approach” was justified by the known dose-dependent QT interval prolongation and risks of torsades de pointes at the high doses of droperidol used in psychiatry.
In this issue of Anesthesiology, Nuttall et al. contribute to the assessment of droperidol’s toxicity.2These authors report a retrospective study on QT prolongation/TdP and sudden death in a large anesthesia survey divided into two 3-yr time periods before and after the FDA warning. This study showed a significant decrease in the overall proportion of patients with QT prolongation, TdP, or death within 2 days following surgery after the eviction of droperidol. During these 6 yr of observation, almost 300,000 surgical procedures were performed, and only three cases of potential or certain TdP were observed. Except for two cases occurring after cardiac surgery, which is known to be a risk factor, one case of sudden death was observed in a 48-yr-old woman after orthopedic surgery. Although not proven, the responsibility of droperidol could not be ruled out in this patient because she received prophylactic droperidol 10 h before she was found to be in cardiac arrest. Interestingly, another QT-prolonging agent, ondansetron, could have played a role in the occurrence of this adverse event. Even if the droperidol exposure was assessed from the records of only 1 per 1,000 of the entire population, the authors calculated the maximum risk of arrhythmia induced by droperidol to be 3.6 per 10,000. However, if taking into account this case of sudden death where droperidol had a possible role, this maximal risk would increase by 50% to 5.4 per 10,000. Finally, the authors conclude that the FDA warning is “excessive and unnecessary.”2
Does this study provide sufficient evidence to question the FDA warning on droperidol and call for its withdrawal?
Several drugs with limited effects on QT interval duration have recently been removed from the market because they had the potential to cause QT interval prolongation and TdP.3Although the potential for TdP is extremely low, it can lead to death in some subjects. This, in itself, is sufficient to mandate some form of warning. The International Conference on Harmonisation and Medicine Agencies in the United States, Europe, and Japan have issued guidelines on preclinical and clinical drug development targeting the effects of new chemical entities on ventricular repolarization.†‡§Old drugs are also concerned when they are being developed for new indications or populations or when administered at a new dose or route of administration that results in significantly higher exposure than those previously approved. Briefly, a drug is considered to bear a potential risk of proarrhythmia when the maximal increase of QTc interval, compared with placebo, has an upper bound of the 95% confidence interval over 10 ms.§
Droperidol has several electrophysiologic characteristics that the guidelines view as potentially harmful. Droperidol blocks HERG, one of the main ionic currents that underlies QT interval duration.4,5Results of clinical evaluations are not consistent. White et al. 6failed to demonstrate statistically significant QT interval prolongation with 1.25 mg droperidol. However, they found a 22-ms QT prolongation with droperidol compared with 12 ms with placebo, and their study was only powered to detect QTc change of 15% (i.e. , approximately 60 ms). In a study that was not placebo controlled, we found a 17-ms QT interval prolongation with 0.75 mg droperidol. Therefore, although not definitively proven or studied according to the guidelines, droperidol can prolong the QT interval even at a low dose and belongs to the increasing list of noncardiac drugs for which some form of warning is justified.
Although the level of QT prolongation indicated as problematic in the International Conference on Harmonisation guideline may seem very low and is not necessarily associated with proarrhythmia, it emphasizes the pharmacodynamic response rather than the perceived epidemiologic risk. From a safety point of view, such a conservative approach is appropriate. For example, despite limited QT interval prolongation,8isolated cases of TdP were reported to the FDA during terfenadine use,9whereas its risk in the population was shown to be similar to that of other antihistamines in epidemiologic studies.10That is, it is clear that the risks of TdP with QT-prolonging drugs can be underestimated in epidemiologic studies and still exist in some individuals. This justifies the regulatory policy of putting warning boxes on all QT-prolonging drugs. Therefore, from a regulatory point of view, the study by Nuttall et al. does not in itself justify removal of the warning on droperidol use. Nevertheless, these authors should be commended for performing the first epidemiologic study addressing the issue of sudden death and torsades associated with the use of droperidol. Their results should prompt the FDA to reconsider and lessen the warning on droperidol.
Setrons also have the capacity to block HERG at high concentrations,11and although their influence on QT prolongation has not been extensively studied clinically in the perioperative period, there are indications that at least ondansetron can induce QT prolongation in this setting.7Given that this drug may have contributed in part to one case of sudden death,2the study of Nuttall et al. emphasizes the need to reinforce the observation of setrons’ effects on ventricular repolarization.
Finally, if one considers the estimated maximal risk of droperidol-induced proarrhythmia (3.6 per 10,000), this would still represent a risk 60 times greater than that of epidural hematoma after epidural anesthesia, whose risk is approximately 1 in 168,000 in the United States.12Even if it is not fatal, no anesthesiologist worldwide would consider the risk of epidural hematoma negligible and accept to perform everyday epidural anesthesia without any caution. Therefore, although the precise format of the warning certainly remains a matter for debate, the warning itself is still justified because one has to be more stringent on safety issues than on efficacy issues.
Université Pierre et Marie Curie-Paris 6, Department of Pharmacology; Assistance Publique–Hôpitaux de Paris, Saint-Antoine Hospital, Division of Clinical Pharmacology; Institut National de la Santé et de la Recherche Médicale (INSERM) CIC9304, Paris, France. firstname.lastname@example.org