Safety of Patients Reason for FDA Black Box Warning on Droperidol

To the Editor: I appreciate the brief report by Drs. Habib and Gan on the case reports examining perioperative use of droperidol (1). Their letter adds useful insight into the cases that prompted the FDA to issue a black box warning about droperidol. I wish to raise several issues in defense of the FDA decision about the black box warning on droperidol: Let us assume briefly that droperidol was responsible for the arrhythmias in the 10 case reports. Since we know that droperidol is safe in the overwhelming majority of patients, we can reasonably assume that patients in whom droperidol precipitates an arrhythmia must be susceptible to ventricular arrhythmias. In fact, this has been clearly documented (2). It is therefore entirely to be expected that in reviewing case reports for droperidol associated arrhythmias, one would identify multiple risk factors, exactly as done by Drs. Habib and Gan. Thus, we don’t really learn anything about whether droperidol caused these patients to experience arrhythmias by the identification of additional risk factors for ventricular arrhythmias, since such factors would be expected covariates of droperidol-induced arrhythmias. Having explained away several of the cases by identifying other risk factors, and noting the lack of detailed information in the other cases, Drs. Habib and Gan conclude that in no case “was there evidence of a cause and effect relationship.” This is expected. This was not a double blind, randomized trial. Lacking appropriate controls, case reports of rare events can almost always be explained away (3). Individuals who published case reports appeared more convinced of the causative role of droperidol (4–6) or its chemical cousin haloperidol (7–9). The important question is whether there is any mechanistic basis for concern. Here the data are unequivocal. First, droperidol causes dose-dependent QT prolongation in patients (10,11) and the effect is readily demonstrated in vitro (12). As nicely summarized by Prielipp and Balser (13), droperidol blocks the I(k)r potassium channel encoded by the HERG gene at concentrations of 10 nmol/L (14). Given a volume of distribution of 2.04 L/kg (15) and a molecular weight of 369 daltons, we can readily calculate that a droperidol dose of 0.625 mg would yield a plasma concentration of 12 nmol/L. The genetics of long QT syndrome are complex, but clearly place a small number of patients at risk of drug-induced arrhythmias (16,17). Rare mutations of the MiRP1 gene predispose patients to arrhythmias from drugs that affect HERG-encoded potassium channels (18). Genetic predisposition towards long QT has low penetrance, but those individuals with the genetic trait and without clinically evident QT prolongation are still at increased risk of arrhythmias from drugs that block potassium channels (19). The genetic, cellular, and animal data paint a convincing story of genuine risk from droperidol for a very small numbers of patients. When combined with suggestive case reports, this is a legitimate concern. The package insert is the appropriate instrument for communicating legitimate concerns to physicians. Habib and Gan suggest an incidence of torsade of about 1 in 1 million doses. Individual physicians will have a different perspective on such risk from a regulatory agency. If I’m throwing up, please give me droperidol—the risk sounds vanishingly small. However, if I am a regulatory agency, that same small risk translates into a handful of preventable deaths. Unfortunately, the true causal relationship between low-dose droperidol and torsade will probably never be established in humans. The incidence is too low for any feasible study to establish the safety of droperidol. Lacking the definitive study, we will be forced to continue examining inconclusive case reports, looking carefully at animal and cellular data, and drawing the best conclusions we can for our patients. Since the FDA has to think about the safety of many millions of patients, their position is likely to be more conservative than the conclusions of individual physicians. Steven L. Shafer, MD