Clinical Relevance and Management of Drug-Related QT Interval Prolongation

Michael A. Crouch, PharmD, Lynn Limon, PharmD, Angela T. Cassano, PharmD


Pharmacotherapy. 2003;23(7) 

In This Article

Drugs Associated with QT Prolongation

Drugs associated with QT prolongation, torsade de pointes, or both are classified by therapeutic class and under each grouping are subdivided into three categories: definite, probable, and proposed association. In a definite association, a positive temporal concordance of drug administration and the event is documented (per published data retrieved from MEDLINE searches), the event disappears or lessens with drug discontinuation, and the event recurs with rechallenge. Additional evidence may include in vitro or in vivo data showing the drug's effects on cardiac ion channels. The second category, probable association, consists of drugs that most likely represent a true risk, but for which there is less complete evidence. The third category contains drugs that have been proposed to cause risk. Drugs may be assigned to this category because of inadequate or contradictory published case reports. In addition, some drugs have been labeled as causing QT prolongation because other drugs in their pharmacologic class have a definite risk of increasing the QT interval. The labeling content derives, of course, from United States Food and Drug Administration (FDA) or manufacturer suggestions, but for these particular drugs, there is no published evidence to support the purported risk. Finally, there are some drugs that have been associated with QT prolongation for which neither published evidence nor product labeling exists, but they are included for the sake of completeness. Table 1 summarizes all these data and includes pertinent information from the product labeling.[11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84]

Reports associating anesthetic agents and QT prolongation have yet to yield a definitive association. Three agents have been purported to cause QT prolongation: enflurane, isoflurane, and halothane. Of these agents, the risk of QT prolongation appears highest with isoflurane.

Probable Association. Halothane, isoflurane, and enflurane significantly prolonged the QTc in one study, but all prolongation was less than 10 msec.[85] Results of reports with these agents vary and at times are discordant. Prolongation of the QT interval appears to be more associated with isoflurane.[86,87] Results of studies with enflurane have been contradictory; some investigations report lower risk as compared with that of halothane,[88,89] with another suggesting the QT interval was prolonged significantly with enflurane but not halothane.[90] Further data suggest that halothane shows similar and comparable effects on ventricular activation to those of enflurane.[91] The clinical relevance is questionable, however, as one study showed enflurane increased the QTc interval significantly after start of anesthesia but with no marked influence on inducibility of clinical ventricular tachycardias.[92]

The proarrhythmic effect of antiarrhythmic drugs is well known and reviewed elsewhere.[93,94,95,96,97] Briefly, effects on the QT interval have been noted with types IA, IC, and III agents (Vaughan Williams classification system).[2] According to the classification, type I agents have their primary effect on the sodium channels, whereas the type III agents are specific to the potassium channels. Types IA and IC agents also alter the movement of potassium, with somewhat greater effect with the type IA agents compared with type IC agents. The rate, magnitude, and clinical effect of QT prolongation differ greatly among these agents.

Definite Association. Type IA agents include quinidine, disopyramide, and procainamide. These drugs primarily affect the rate of fast sodium entry into cardiac cells but also affect outward potassium flow, acting to slow conduction velocity and prolong refractoriness.[93,96] Quinidine is implicated as causing syncope in 2-6% of patients, which is hypothesized to be the result of QT prolongation and associated torsade de pointes. The syncope is associated with therapeutic or subtherapeutic serum concentrations and usually occurs early in therapy.[5] Risk factors for quinidine proarrhythmia (up to 28% of these patients have torsade de pointes) include preexistent QT interval abnormalities, concomitant digoxin, other drugs known to prolong the QT interval, impaired left ventricular function, and hypokalemia.[96] Starting quinidine therapy in the hospital with 2-4 days of close observation enables identification of 55-77% of torsade de pointes cases. Potassium and magnesium levels should be high normal before starting therapy. Of importance, the reaction may occur after low doses and at low serum concentrations.[98] A prolonged QTc is one indicator of impending torsade de pointes, although monitoring for prominent U waves, extrasystoles, and U-wave augmentation after extrasystole may be more useful.[5] The frequency of torsade de pointes with disopyramide and procainamide appears less than that with quinidine. For procainamide, its active metabolite N-acetylprocainamide has been implicated; torsade de pointes has occurred when N-acetylprocainamide levels, rather than parent drug levels, are high.[96]

Type IC antiarrhythmics include flecainide and propafenone. Like the type IA drugs, these agents also decrease the rate of sodium entry (fast channels) into cardiac cells and have some effect, albeit less than that of type IA agents, on the potassium channels. Unlike the type IA drugs, type IC agents profoundly slow conduction velocity while leaving refractoriness relatively unaltered. Flecainide is well known for its proarrhythmic effects in patients with established heart disease, based on the results of the Cardiac Arrhythmia Suppression Trial.[99] Of importance, proarrhythmic effects in this trial appeared to be due to excessive slowing of conduction. In normal doses, flecainide usually has little effect on repolarization.[18,100] In a series of 152 patients receiving flecainide, 4.6% demonstrated proarrhythmic effects. Many of the patients had increased flecainide concentrations or the presence of other antiarrhythmic agents, which likely confounded the situation.[100]

One group of authors[101] retrospectively compared the electrocardiographic effects of flecainide and propafenone in patients recruited for a study to treat atrioventricular node reentry tachycardia. Flecainide significantly increased QTc dispersion (a measure of heterogeneous ventricular repolarization on the 12-lead ECG that may predict an increased risk of proarrhythmia), adjusted QTc dispersion, mean QT, and QTc.[102] These results are supported by another investigation in which QT prolongation was believed to be reflective of the universal extended QRS interval, a use-dependent property of flecainide.[103] For propafenone, also a type IC agent, no effect on the markers of repolarization occurs, although it appears to increase the QRS interval, showing effect on depolarization.[101] Although type IC antiarrhythmics may be proarrhythmic, their primary risk is not associated with repolarization, and QT prolongation is not a prominent risk.

Class III agents (sotalol, amiodarone, ibutilide, and dofetilide) cause torsade de pointes by a combination of calcium-dependent early afterdepolarizations and enhanced dispersion of repolarization. Sotalol-associated torsade de pointes has been observed in 1-4% of patients, with the risk increasing with increasing dosages and with accumulation of drug due to renal impairment or older age. In a safety analysis of sotalol, 4.3% of patients had a proarrhythmic event.[104] The rate increases with total daily dose (1.8% in dosages up to 320 mg/day, 4.5% in dosages up to 480 mg/day, and 6.8% in dosages up to 640 mg/day).[105] Risk factors for torsade de pointes include baseline arrhythmia, heart failure or cardiomegaly, baseline QTc, elevated serum creatinine level, and patient's sex. Women show a 3-fold greater chance of developing torsade de pointes.[106] Sotalol slows heart rate through its -adrenergic antagonism, and both bradycardia and electrolyte disturbances favor early afterdepolarizations.[5,94] Sotalol prolongs repolarization by blocking the IKr channel. It has no effect on the slow component (IKs) and minimal effect on the inward rectifier current.

For amiodarone, the frequency of proarrhythmic events has been reported as 2%.[107] Overall, the rate of occurrence of proarrhythmia appears low[108]; the package insert comments that the rate is infrequent (< 2%).[14] Amiodarone's mechanism for prolonging repolarization includes blockade of IKr channel, the fast sodium channel, and the slow inward calcium channel by means of the L-calcium channel. The lower frequency of amiodarone-associated torsade de pointes may be related to its reduced calcium influx.[94] Amiodarone still has a high risk of causing torsade de pointes, but its occurrence rate is much lower than that of other type III antiarrhythmic agents.

Ibutilide increases the refractory period and prolongs the action potential duration by means of slow inward currents and potassium currents. Among patients in the ibutilide clinical program, QT prolongation was reported, and 4.3% of patients developed torsade de pointes before the drug received FDA approval status.[109] Overall, the rate of torsade de pointes is reported at 1.7% with ibutilide.[19] Female sex and non-Caucasian race, along with low heart rate and heart failure, appear to increase the risk of ibutilide-induced torsade de pointes.[109]

Dofetilide is the newest class III antiarrhythmic to be marketed. Because of its later approval and known risk of QT prolongation, it has been held to a higher standard than previously approved antiarrhythmics. Its approval was contingent on developing a program with a treatment algorithm to decrease the risk of QT prolongation. Overall, the frequency of torsade de pointes was 0.8% across all dosages, in patients receiving the agent for supraventricular arrhythmias, if appropriately dosed for renal function. The range was 0.3-10.5%, with the rate increasing with increasing dosages. Dofetilide blocks the IKr channel, with no effect on the other repolarizing potassium currents.[17] Most torsade de pointes cases that occurred with dofetilide in clinical trials happened within 3 days of starting the drug; therefore, all patients who are now prescribed the agent must be monitored in a setting where they can be continuously monitored for 3 days. A baseline QTc interval and calculated creatinine clearance must be obtained to choose the appropriate dosing regimen. All prescribers must complete a training program to prescribe the drug.[17]

Proposed Association. Although adenosine is not an agent classically associated with QT prolongation, there are case reports implicating this drug. In one case report, the patient had long QT syndrome,[110] whereas two other cases of torsade de pointes occurred in patients with a normal heart and normal baseline QT intervals.[111]

Felbamate and fosphenytoin are both considered to increase the risk of QT prolongation and torsade de pointes, although their effects have not been well characterized.

Proposed Association. For felbamate, no cases could be retrieved through a MEDLINE search. In the product labeling, torsade de pointes is listed based on postmarketing reports.[24] No specific numbers, concomitant drugs, or risk factors are noted. Likewise for fosphenytoin, no cases could be retrieved through a MEDLINE search. In the product labeling, QT prolongation is reported as an infrequent adverse event (1/100-1/1000).[25]

Tricyclic and tetracyclic antidepressants affect the QT interval; however, cardiotoxicity primarily occurs in the overdose setting. Selective serotonin reuptake inhibitors (SSRIs) affect the QT markedly less than do the antidepressants released before the SSRIs came to market, although a probable association appears to exist.

Definite Association. In a review article on antidepressant overdoses published in 1982, the author evaluated reports dating back to the 1960s on conduction disturbances associated with overdoses of amitriptyline, desipramine, doxepin, imipramine, and nortriptyline.[112] Electrocardiographic findings included prolongation of the PR and QRS intervals, as well as ST- and T-wave changes. In critical care patients with tricyclic overdoses, QRS prolongation and QT prolongation were reported in 42% and 42% of patients, respectively.[113] Despite documented abnormal ECG readings, serious cardiac complications (e.g., ventricular arrhythmias) are low, affecting 0.02-2.6% of patients who had an overdose.[112,114,115] Prolongation of the QTc interval has been observed with regular dosing of these agents, as well as with the agent maprotiline, in adults and children.[116,117,118,119,120] In a recent epidemiologic study in which 495 patients receiving psychotropic therapy were evaluated,[121] those taking tricyclic antidepressants were at highest risk for QTc interval lengthening compared with other antidepressant agents. Cardiotoxicity of the tricyclics is attributed to their quinidine-like effects. Imipramine and amitriptyline inhibited the current mediated by HERG at concentrations achieved with normal dosages, with the amount of blockade increasing with increasing concentrations.[122] Overall, the frequency of QT prolongation is not well characterized in this class of agents, but it does appear to be uncommon.

Probable Association. Fluoxetine, paroxetine, and sertraline are other antidepressants that have been implicated as causing QT prolongation. This class of drugs has been associated with even fewer negative cardiac events than those of the tricyclic and tetracyclic agents. In animal cardiac tissue, fluoxetine and citalopram inhibited sodium and calcium channels, similar to results observed with tricyclic agents.[123] Further investigation has found that, compared with tricyclic antidepressants and doxepin, no significant changes in the QT interval were noted with fluoxetine.[120,124] Cases of fluoxetine-associated toxicity have been reported (torsade de pointes and QT prolongation).[125,126,127] Ventricular arrhythmias, not specified, are reported as rare (1 in 10,000) in the product labeling.[30] Insufficient evidence exists to adequately predict QT prolongation or prevent torsade de pointes during fluoxetine therapy. A few cases of QT prolongation implicating paroxetine and sertraline also have been published.[128,129] The data with citalopram and QT prolongation are in overdose situations.[130]

Proposed Association. Another antidepressant that has been implicated is venlafaxine. The only reference noting QT prolongation indicates it occurred when the drug was given with other agents known to affect the QT interval.[131] The product labeling notes QT prolongation.[36]

The QT interval-prolonging effects of the antihistamines terfenadine and astemizole are well characterized, particularly when drug-drug interactions are present. Both of these agents have been removed from the United States market because of these effects. Other antihistamines have been implicated but lack clear associations.

Definite Association. Initial reports of terfenadine's effect on the QT interval appeared in the literature in the early 1990s, demonstrating the parent compound, not the metabolite, was implicated.[132] The combination of terfenadine with ketoconazole,[133,134] erythromycin,[135] and grapefruit juice[136] resulted in elevated terfenadine concentrations and alterations of the QT interval. Astemizole also results in QT prolongation, typically when drug-drug interactions are present. Cases of QT prolongation with astemizole overdose have been reported.[137,138] The mechanism of QT prolongation with these agents appears to be secondary to potassium channel blockade leading to delayed repolarization.[139]

Proposed Association. One investigation with diphenhydramine overdose demonstrated a significant, although moderate, change in the QTc interval.[140] Clemastine also has been purported to cause QT prolongation, but the data to support this effect are based on one animal study.[141] Loratadine appears to lack significant effects on the QT interval, even at a high plasma concentration,[142] although a nested case-control analysis identified marginally increased risk with loratadine, as well as with astemizole and terfenadine.[143] QT prolongation does not appear to occur with certirizine, fexofenadine, or desloratadine.[144,145,146]

As a class, these agents recently have been under scrutiny for their effects on the ECG. Macrolides, fluoroquinolones, imidazole antifungals, and antimalarials are best known for prolonging the QT interval.

Definite Association. Erythromycin prolongs the QT interval by blockade of the IKr channel, with effects increasing as drug concentrations increase.[147,148,149,150] Erythromycin exhibits electrophysiologic effects similar to those of class III antiarrhythmic drugs.[98] Although oral erythromycin-induced prolonged QTc interval and frequent premature ventricular contractions have been reported (confounded by quinidine administration),[151] the occurrence of QTc prolongation with possible torsade de pointes predominantly has been reported with intravenous administration.[150,152,153,154,155,156] In a case series,[157] the ECGs of seven patients were analyzed. Twelve of 13 separate intravenous erythromycin administrations demonstrated a statistically significant increase in both the QT and QTc durations compared with baseline values. Ventricular arrhythmias occurred in three patients who received erythromycin 1000 mg faster than 15 mg/minute. In addition, two reports include patients who had liver impairment; presumedly they were unable to metabolize erythromycin normally and drug accumulated.[158,159] One group investigated whether female sex was a risk factor in developing erythromycin-induced cardiac adverse reactions, and the data suggested a female predominance of cardiac arrhythmias while taking erythromycin compared with men (58% vs 32%, respectively).[4] A more recent review of macrolide adverse reactions supports the finding of female sex, increasing age, concomitant drug therapy, and comorbid disease as risk factors.[160]

Besides erythromycin's direct effect on the QT interval, pharmacokinetic interactions can occur. Drug-drug interactions with this agent are well known to lead to electrocardiographic changes with agents that are metabolized through cytochrome P450 (CYP) 3A4. For example, metabolism of terfenadine, astemizole, and cisapride will be inhibited, potentially leading to accumulation of those agents and QT prolongation (see Drug Interactions).

The antimalarial agent halofantrine was widely used internationally before it received FDA approval status in the United States. Inhibition of the HERG channel by halofantrine was recently documented as the cellular mechanism responsible for repolarization problems with this agent.[161] In a prospective study,[162] 20 patients (11 with malaria, 9 without) received halofantrine 1500 mg (a normal dose). In both groups, QTc prolongation occurred, although no clinical consequences were noted. The mean QTc for both groups did not exceed 460 msec. The highest mean value for QTc interval occurred as serum levels of halofantrine and its major metabolite peaked. Additional case reports of QT prolongation and/or torsade de pointes have been published in predisposed patients (female, long QT syndrome).[163,164,165] Given the wide use of halofantrine outside the United States and its known ability to inhibit HERG channels, it is surprising that more repolarization problems have not been reported. Clinically, QTc prolongation occurs at normal doses. It seems prudent to monitor patients closely, at least for the first dose of the two-dose regimen (the second dose given on day 7).

There are many case reports of QTc prolongation and torsade de pointes associated with pentamidine administration in patients with human immunodeficiency virus (HIV) infection.[166,167,168,169,170,171,172] Many of these cases are difficult to interpret because of confounding variables such as concurrent administration of other drugs known to prolong the QTc interval, electrolyte abnormalities, or renal dysfunction. Although a specific mechanism has not been identified, presumedly pentamidine's similarity in chemical structure to that of procainamide may explain pentamidine's effects on repolarization. Additional case reports of QT prolongation with accompanying ventricular tachycardia have been associated with pentamidine infusions and intramuscular injections.[173,174] In a comparative study, no significant differences were noted in QTc or the frequency or complexity of ventricular arrhythmias between 16 patients treated with pentamidine and 11 treated with trimethoprim-sulfamethoxazole.[175] In another small study (10 patients),[176] the investigators also found no difference between pre- and post-pentamidine ECGs compared with the ECGs of patients not receiving pentamidine when evaluated prospectively.

Confounding factors aside, it appears patients with HIV infection who have Pneumocystis carinii pneumonia and who receive parenteral pentamidine therapy are at increased risk of QTc prolongation and ventricular arrhythmias. Usually patients with this pneumonia are acutely ill, are compromised in one or more ways, and are taking multiple drug regimens. Corticosteroids, often given in this setting, can cause hypokalemia. Although authors disagree about the increase in risk associated with pentamidine, the general recommendation is to monitor at least serial ECGs, and if QTc prolongation is found, monitor more closely or choose an alternative therapy.

Almost all reported cases of QTc prolongation or torsade de pointes were in patients prescribed parenteral pentamidine for at least 1 week. A case of inhaled pentamidine-associated torsade de pointes has been published.[177] However, in a study of 100 patients taking inhaled pentamidine twice/week, the authors found no QT prolongation for the duration of the patients' therapy.[178] Of interest, other authors describe a case of torsade de pointes in a patient who received one dose of inhaled pentamidine (hypokalemia and low ionized calcium present).[177]

Various fluoroquinolones have been implicated as causing QT prolongation, with the greatest concern with sparfloxacin and grepafloxacin.[180,181,182] Both of these agents, now removed from the U.S. market, blocked the IKr channel to a much greater degree than do the other agents in this class. As of November 1999, 145 cases of QT-related cardiac events/49,000 patients were reported with sparfloxacin. With grepafloxacin, seven cardiac-related deaths and three cases of torsade de pointes were reported.[183,184]

Probable Association. The inhibition rates of seven fluoroquinolones for their ability to interact with the IKr channel have been investigated.[185] Sparfloxacin was the most potent compound, displaying a value of 18 µM for the concentration that inhibits 50% of the HERG channel current, followed by grepafloxacin 50 µM, moxifloxacin 129 µM, gatifloxacin 130 µM, levofloxacin 915 µM, ciprofloxacin 966 µM, and ofloxacin 1420 µM. As previously mentioned, sparfloxacin and grepafloxacin have been removed from the U.S. market.

Moxifloxacin is a new agent that received FDA approval status in 2000 in the oral form and in 2001 in the parenteral form. Given its more recent evaluation by the FDA, this drug is an example of the FDA's focus on the cardiotoxicity risks of new drugs.[53] The tendency to prolong QT interval was found early in moxifloxacin's development[186]; thus, many patients at risk were excluded from clinical trials. More than 4000 patients were evaluated during the clinical development of moxifloxacin, none of whom had an attributable cardiovascular morbidity or mortality.[53] Among those who received moxifloxacin, the mean QTc prolongation was 6 msec, unlikely to be of clinical relevance.[187] In another investigation, 228 moxifloxacin-treated patients received concurrent drug therapy with high-risk agents.[188] Four cases of arrhythmia were reported in 3780 patients: three in patients who received moxifloxacin alone and one in a patient who also received another drug with purported QT prolongation risk. One case of torsade de pointes was reported in a patient who received moxifloxacin; this patient had risk factors for QT prolongation, including hypokalemia, coronary heart disease, and a cardiac pacemaker suggesting underlying cardiac disease.[189] The risk of QTc prolongation with oral moxifloxacin appears low, given the in vitro data and the clinical data available to date, although a single case has been report in an elderly female patient with multiple risk factors.[184] The risk may be increased with the intravenous formulation; a report of a mean QTc interval prolongation of 12.1 msec was observed.[190]

Gatifloxacin (oral and parenteral) is another newer fluoroquinolone. Because studies in healthy volunteers showed a mean QTc prolongation of just 2.9 ± 16.5 msec in trials, no exclusions were made related to this drug once it was marketed. Patients with different risk factors (hypokalemia, concurrent therapy with high-risk drugs like cisapride) were thus included in the trials, and none had adverse reactions related to QTc prolongation.[45] A safety study evaluated almost 16,000 patients receiving gatifloxacin, of whom 4906 had an underlying cardiac condition and/or were taking cardiac drugs. None of the three cardiac adverse events were believed to be associated with gatifloxacin.[191] Other cases of torsade de pointes have been reported; however, these cases appear to have confounding variables (sotalol therapy, fluconazole therapy, bradycardia, hypomagnesemia).[192,193] One investigator obtained the cases of torsade de pointes associated with fluoroquinolones sent to the FDA between January 1996 and May 2001 by voluntary reporting, adjusted to cases/10 million prescriptions.[194] The number of cases/10 million prescriptions was gatifloxacin 27, levofloxacin 5.4, ofloxacin 2.1, ciprofloxacin 0.3, and moxifloxacin 0. Gatifloxacin's product labeling warns against giving this agent to patients with long QT syndrome and with uncorrected hypokalemia. It should not be given to those receiving type IA or III antiarrhythmics and should be used with caution if the patient is receiving other drugs that prolong the QT interval.[45]

Like erythromycin, clarithromycin inhibits the CYP3A4 isoenzyme system.[195] Therefore, a patient taking this agent is theoretically at risk for the same drug interactions that may occur with erythromycin.[40,196] As monotherapy, published evidence to support that clarithromycin adversely affects the QT interval is limited, perhaps because, unlike erythromycin, it is only available orally and thus has lower serum concentrations. Case reports with clarithromycin and QT prolongation, torsade de pointes, or both have been published.[197,198] Clarithromycin's mechanism of prolonging the QT interval is likely the blockade of the IKr channel.[199] During a 2001 FDA advisory committee reviewing the application of telithromycin, data were presented from national prescription reports documenting more torsade de pointes occurrences with clarithromycin than with erythromycin.[200] The authors, in a later publication, suggested potentially biased reporting due to the scrutiny of cisapride use in the time period reviewed; many of the reports were from patients who were taking both clarithromycin and cisapride.[160]

Prolongation of the QT interval has been reported with the azole antifungal agents, including ketoconazole, itraconazole, and fluconazole. Compared with most agents discussed in this article, the azole antifungal agents are probably more important in their role of inhibiting metabolism and increasing serum concentrations of other agents known to increase the risk of QT prolongation and possible torsade de pointes rather than direct QT prolongation. Azoles inhibit the CYP3A4 isoenzyme,[47,48,49,50,195] although fluconazole has a much lower inhibition than that of the other two agents. Ketoconazole also blocks HERG and may have inherent effects on the potassium currents.[201] Ketoconazole alone has been evaluated, and after 4 days of standard therapy, it produced a mean increase in the QTc interval of 5.5 msec, as compared with placebo.[202]

Two reports discuss the risk of fluconazole and QT prolongation.[203,204] In one report, the authors noted the lack of confounding variables.[203] In the second report, however, intravenous fluconazole and intraperitoneal dosing were used and resulted in markedly elevated plasma levels of fluconazole (216 µg/ml compared with an expected 18-28 µg/ml).[204] Fluconazole was stopped, and the patient's cardiac abnormalities corrected in 3 days. Ketoconazole and itraconazole have been available only for oral administration (although itraconazole recently became available for parenteral administration), but fluconazole as an intravenous agent allows rapid accumulation of drug with potentially high plasma levels.

Quinine is the isomer to quinidine, and it has been shown to have an effect on the QT interval, albeit much less than that of quinidine. Nonetheless, it has been demonstrated to have effects similar to those of quinidine.[205] In addition, quinine, on occasion, has been reported to cause torsade de pointes, although confounders were present.[206]

Proposed Association. Other antiinfective agents for which risk of QT prolongation has been proposed are foscarnet, ganciclovir, levofloxacin, trimethoprim-sulfamethoxazole, and mefloquine. For foscarnet and ganciclovir, the data are quite limited. Foscarnet can cause electrolyte abnormalities, such as hypokalemia, a known risk factor for prolonging QT intervals.[43] Two patients developed torsade de pointes during ganciclovir infusion. Both had reoccurrence of the arrhythmia on rechallenge.[207] At this time, not enough evidence has accumulated to help predict ganciclovir's effects on repolarization.

Levofloxacin, mentioned in the previous section, also carries a proposed association. The manufacturer had to revise the labeling of levofloxacin to reflect 15 cases of QT interval-related events and torsade de pointes reported to the FDA's MedWatch program through mid-1999.[184] Details of causality were not available for review. The product labeling states that patients had confounding factors, and a causal relationship was not established.[51] A case of lengthened QTc interval and polymorphic ventricular tachycardia with levofloxacin has been reported.[208] In another study investigating the inhibition rates of seven fluoroquinolones for their ability to interact with the IKr channel, levofloxacin was less potent than sparfloxacin, grepafloxacin, moxifloxacin, and gatifloxacin.[185] In an article mentioned earlier,[194] 5.4 cases of torsade de pointes were reported to the FDA per 10 million prescriptions of levofloxacin between January 1996 and May 2001. The risk of QT prolongation with levofloxacin is low. Per the product labeling, patients should not have hypokalemia or significant bradycardia, or be prescribed a class IA or III antiarrhythmic before starting the drug.[51]

Through the many years of trimethoprim-sulfamethoxazole use, including the more recent surge as a treatment for P. carinii pneumonia associated with HIV infection, only two cases of QT prolongation were identified through a MEDLINE search.[209,210] Certainly if an association exists with trimethoprim-sulfamethoxazole prolonging the QT interval, the risk is low. Trimethoprim-sulfamethoxazole probably does inhibit a rare type of IKr channel, which was implicated in one published case of prolonged QT interval after receiving this drug.[211]

Data to support the antimalarial mefloquine's proposed prolongation of the QT interval are sparse. Most studies document no significant changes in the QTc interval during mefloquine administration.[212,213] Other investigations, however, comment that the QTc interval was longer.[214] Results of one animal study suggest that mefloquine does not alter the QTc itself, but it enhances the effects of halofantrine.[215] Apparently a synergistic prolongation of the QT interval occurs when mefloquine is coadministered with halofantrine.[216]

Various antipsychotic agents have been implicated as causing QT prolongation. A clear relationship has been established with the butyrophenones (haloperidol, droperidol, and pimozide), as well as thioridazine, mesoridazine, and ziprasidone.

Definite Association. The butyrophenone class of neuroleptics includes pimozide (Tourette's syndrome), haloperidol (antipsychotic), and droperidol (antiemetic occasionally used in psychiatric emergencies). Pimozide is a potent inhibitor of HERG channels.[217] Because pimozide is a well-established substrate of CYP3A4 isoenzyme activity, patients are at risk for significant drug interactions if they are taking any of the CYP3A4 inhibitor drugs.[59,195] Haloperidol also blocks HERG channels, but has little effect on other I channels.[218] One group[219] showed that droperidol exhibits class III antiarrhythmic drug properties and blocks the IKr channel even at low plasma concentrations.[73] Given that haloperidol, droperidol, and pimozide are all documented inhibitors of ion channels, patients taking any of these drugs are at increased risk of developing repolarization problems. For haloperidol and droperidol, the risk increases with increasing doses. For droperidol, patients also appear to be at risk with standard doses. Droperidol is discussed in greater detail in the gastrointestinal section, since its primary indication is as an antiemetic.

Haloperidol is administered orally, intramuscularly, and/or intravenously, occasionally in high doses to control psychomotor agitation due to delirium. Cases of QTc prolongation and torsade de pointes have been reported, although primarily in high-dose situations. In a case series, three patients were noted to develop QTc prolongation without torsade de pointes with haloperidol.[220] In another report, a patient who underwent coronary bypass grafting surgery experienced torsade de pointes without evidence of QTc prolongation.[221] Various other cases of prolongation have been reported, and a clear association exists between haloperidol and the QT interval, torsade de points, or both.[57,222,223,224] Two authors[225] reviewed 11 case reports of haloperidol and droperidol conduction disturbances in critical care patients. Problems occurred when patients received more than 50 mg/24 hours of either agent intravenously. Thirteen of 18 patients reviewed had a history of cardiovascular disease. The authors recommend a baseline ECG with measurement of potassium and magnesium concentrations. Patients should be monitored, and those receiving higher doses should receive particular attention.

A group of authors[226] who obtained the FDA MedWatch reports for pimozide monotherapy indicated that 11 patients reportedly had cardiac arrhythmias. Furthermore, pimozide prolonged the QT interval in isolated perfused rabbit hearts. Pimozide in overdose has been associated with torsade de pointes.[227]

Numerous cases of electrocardiographic changes and ventricular arrhythmias with thioridazine have been reported.[228,229,230,231,232,233,234,235,236,237,238,239] This adverse reaction has been reported with both standard dosing as well as in overdose situations. A recent epidemiologic study[121] found that in patients taking psychotropic therapy, those receiving thioridazine were at higher risk for QTc interval lengthening compared with other phenothiazines (only droperidol put patients at higher risk).

Thioridazine has quinidine-like effects, slowing fast sodium entry and cardiac conduction and prolonging ventricular repolarization.[229,239] One group[240] demonstrated thioridazine's blockade of IKr. In 1996, other investigators[241] showed that thioridazine's effects on the QT interval increased with increasing dosages and that the parent drug is primarily responsible. Thioridazine's metabolism will be slowed by any drug that inhibits the CYP2D6 isoenzyme.[63]

Although the risk of QT prolongation increases with higher dosages, patients taking any dosage of thiorizadine are at risk. In a July 2000 letter from the manufacturer, a warning appeared that the product labeling now recommends thioridazine only as second-line therapy.[242] Wording is specific regarding its contraindication with drugs that inhibit the CYP2D6 isoenzyme (e.g., fluoxetine, paroxetine), in patients with long QT syndrome, in those with a baseline QTc greater than 450 msec, and/or in those with a history of cardiac arrhythmias. An ECG and potassium level at baseline are recommended and should be measured periodically.

Receiving almost no attention was another letter from the same manufacturer for mesoridazine in September 2000, with a black box warning added to the product labeling.[243] A dose-dependent prolongation of the QT interval has been noted. It too is now relegated to second-line therapy.[243]

Ziprasidone is an antipsychotic agent that received FDA approval status in its oral formulation in February 2001 and in its intramuscular formulation in mid-2002. It is a good example of the attention the FDA and pharmaceutical manufacturers are focusing on cardiotoxicity. Ziprasidone had been in the FDA pipeline since 1997. In 1998, a review by the Cardio-Renal Advisory Committee prompted a "non-approvable" letter because of QTc prolongation findings. As a result, the manufacturer performed a study to compare ziprasidone's QTc interval rates with those of risperidone, olanzapine, quetiapine, haloperidol, and thioridazine.[244] Ziprasidone's mean increase from baseline values of QTc was 9-14 msec greater than that of the other agents. The exception was that the effect of ziprasidone on the QTc interval was 14 msec less than that of thioridazine. These results were partially responsible for label changes required for thioridazine, as mentioned above. The manufacturer, as part of a postmarketing requirement for ziprasidone, will study sudden unexpected death in any patient who took the drug.[245] In placebo-controlled studies, ziprasidone increased the QTc interval 10 msec more than did placebo at the highest recommended daily dose of 160 mg.[64] Increases to more than 500 msec were observed in 2 of 2988 ziprasidone-treated patients and in 1 of 440 placebo-controlled patients. In clinical trials completed to support the Swedish application of ziprasidone, an average QT prolongation of 2.86 msec was observed. Prolongation of 30-60 msec was found in 11.1% of ECGs of those taking ziprasidone and in 6.2% of those taking placebo.[246] Product labeling includes warnings against administering ziprasidone with other drugs that prolong the QT interval (a list is provided in the package insert[64]), in patients with long QT syndrome, in patients with recent myocardial infarction, and in patients with uncompensated heart failure. Patients should undergo baseline measurements of potassium and magnesium concentrations. Torsade de pointes has not been reported with this agent.[64]

Probable Association. Chlorpromazine, developed in 1950, was the first widely used neuroleptic. Although there are a few reports associating chlorpromazine with QT prolongation and torsade de pointes,[247,248,249,250] these are overshadowed by the proliferation of cases associated with thioridazine. In the epidemiologic study mentioned earlier[121] in patients receiving psychotropic therapy, those taking thioridazine were at higher risk for QTc interval lengthening compared with other phenothiazines, including chlorpromazine. Product labeling for chlorpromazine mentions "nonspecific, usually reversible, Q and T wave distortions" as a class effect with phenothiazines.[56]

Proposed Association. Two reports of prolonged QT interval implicated quetiapine in the overdose setting.[251,252] In the manufacturer's study comparing ziprasidone with other antipsychotics, quetiapine prolonged the QT interval a mean of 14.5 msec from baseline.[244] The mean changes from QTc interval baseline of the tested antipsychotics were thioridazine 35.6 msec, ziprasidone 20.3 msec, quetiapine 14.5 msec, risperidone 11.6 msec, olanzapine 6.8 msec, and haloperidol 4.7 msec. Risperidone may have an association with changes on the QT interval, but not enough evidence has accumulated to help predict risperidone's effects on repolarization. In a study comparing QT prolongation potential in feline hearts,[253] risperidone prolonged the QT interval by 19.4 ± 2.2%, compared with haloperidol's 26.2 ± 0.7% increase. Lesser prolongation was observed with sertindole, clozapine, and olanzapine.[251] Sertindole was removed from the market in 1998 because of safety concerns, as some data suggested a potential for effects on the QT interval.[254,255]

Although these agents are not commonly noted for their effects on the QT interval, one chemotherapeutic agent, arsenic trioxide, clearly leads to this adverse effect.

Definite Association. Arsenic trioxide recently received FDA approval status for the treatment of relapsed and resistant cases of acute promyelocytic leukemia. It has been used for many years, both as a therapeutic agent and as a poison. Cardiotoxicities of arsenic poisoning include QRS complex broadening, prolonged QT intervals, ST segment depression, T-wave flattening, and multifocal ventricular tachycardia.[256,257] One group of authors[258] reported QT prolongation that developed into torsade de pointes in 3 of 19 patients being treated with arsenic trioxide. All 3 patients had normal QTc intervals before arsenic trioxide treatment. Risk factors such as electrolyte imbalance were corrected. Torsade de pointes occurred at 12, 16, and 42 days after the start of treatment. In another case series,[259] eight patients treated with arsenic trioxide, all prospectively monitored, demonstrated QT prolongation during induction therapy. Four patients developed nonsustained ventricular tachycardia and had to be treated with antiarrhythmic agents.

The manufacturer recommends a baseline 12-lead ECG and measurement of electrolytes and serum creatinine before commencing therapy. Once therapy is under way, any QTc finding greater than 500 msec needs to be corrected to 460 msec before therapy can continue.[65] Others have recommended that all patients receiving arsenic trioxide receive regular electrocardiographic monitoring. Patients who had arrhythmias in this case series developed them between 5 and 117 days of therapy. This accumulated toxicity was found in the three reported cases of torsade de pointes mentioned previously,[258] suggesting tissue accumulation.

Probable Association. For tamoxifen, one supportive study in rabbit ventricular myocytes showed that, at clinically relevant concentrations, tamoxifen blocks the I channel as well as the inward calcium current.[260] The authors suggested that effects on the calcium current act as a protective device against effects on repolarization, especially since in their study tamoxifen blocked the I channel to a greater extent than did quinidine. The product labeling reports in the overdose section that prolonged QT intervals occur with dosages 6 times higher than normal.[66] There does appear to be a risk of prolonged QT interval with tamoxifen, but only at greater than normal dosages.[261]

In addition to antiarrhythmic agents, other cardiovascular drugs have been implicated in QT prolongation. Clear associations have been demonstrated for bepridil and probucol.

Definite Association. Bepridil is a calcium channel blocker indicated for second-line therapy of angina. It has class I antiarrhythmic activity, affecting fast sodium inward currents. Various case reports of bepridil-induced torsade de pointes have been reported in the literature.[262,263] Its effects on I channels were investigated in a study comparing bepridil, nitrendipine, diltiazem, verapamil, and mibefradil (no longer on the U.S. market). The HERG channel is blocked in a concentration-dependent manner most by bepridil, followed by verapamil and mibefradil. Nitrendipine and diltiazem have negligible effects on HERG channels.[264] Bepridil has a dose-dependent effect on the QT interval. Mean prolongation of the QTc interval appears to be 8-10%, with some outliers having a QTc interval 25% over baseline.[264] Torsade de pointes has been reported in patients with long QT intervals and hypokalemia[67,265,266,267]; the rate was 0.01% in one French series.[265]

The prescribing information for bepridil includes strong language about using this agent as second-line antianginal therapy. Dosing titration should be slow, allowing 8 days for a steady-state concentration and remeasurement of the QT interval. Bepridil is contraindicated in those with a history of arrhythmia, with hypotension, with uncompensated cardiac insufficiency, with long QT syndrome, or in those taking drugs known to prolong the QT interval.[57]

Probucol, no longer marketed, was associated with QT prolongation since initial reports from the early 1980s.[268,269,270] Subsequently, the effects of probucol on the QT interval were evaluated in a controlled investigation. The mean QTc interval of this small population was lengthened after probucol administration, although not statistically significantly compared with placebo. In those patients showing an increase, the increase in the QT interval was 11-70 msec over baseline.[271] Other authors noted that the prolongation of the QT interval is directly related to plasma probucol concentrations.[272] Additional studies verified that the QT interval is prolonged with probucol therapy.[273] The clinical relevance, however, was questioned.[274] Yet, case reports of QT prolongation and development of torsade de pointes or syncope were published.[275,276] Results of additional investigations suggest that QTc prolongation is likely to appear in female patients and in patients with long baseline QTc or with low serum albumin levels.[277,278]

Proposed Association. Mibefradil is another calcium channel blocker (T channels) proposed to cause QT prolongation. It was withdrawn from the U.S. market because of multiple drug-drug interactions. However, QT interval alterations were reported, as was torsade de pointes in patients receiving the agent.[97,279] It remains uncertain whether proarrhythmic effects with mibefradil were due to the drug itself or to drug-drug interaction with other agents affecting the QT interval.

Although data are lacking, indapamide may lead to QT prolongation. Diuretic-induced hypokalemia is believed to be the primary cause with this agent. Indapamide was implicated in two reports of torsade de pointes; hypokalemia was present in both, concomitant disopyramide in one.[280,281] In one investigation,[282] the effect of the diuretic agent indapamide on the I channel in guinea pig myocytes was evaluated. They found that indapamide blocks the IKs channel; however, proarrhythmic effects with indapamide appear most likely due to hypokalemia induced by this diuretic agent.[282]

Other agents implicated for QT prolongation, but for which no published data in humans are available to support such a claim, include isradipine, moexipril-hydrochlorothiazide, and nicardipine. In fact, published data for isradipine and nicardipine suggest these drugs have no significant effect on the QT interval.[283,284,285] It is important to note that QT prolongation is listed in the package inserts for all three agents.[69,70,71]

This class of agents has received a great deal of attention related to QT prolongation. Cisapride is well known for its effects on the ECG, most notably when drug-drug interactions are present. Distribution of cisapride in the United States currently is restricted. Droperidol, mentioned briefly in the antipsychotic section, also has a definite association.

Definite Association. When initially marketed, cisapride was considered a relatively safe drug, although reports of cisapride-related arrhythmias in adults were available. In those case reports, all drug-related arrhythmias occurred with concomitant administration of a drug known to increase cisapride levels, a higher than normal dosage of cisapride, or concomitant administration with another agent known to prolong the QT interval. By the end of 1999, there were 341 reports of cisapride-associated arrhythmias, including 80 deaths.[286,287] Cisapride is primarily metabolized by CYP3A4, the isoenzyme responsible for metabolism of most drugs handled by the CYP enzyme system. As a substrate, its metabolism is inhibited, which results in higher concentrations of the active parent, causing the cardiotoxicity, probably by blocking I channels.[288] Well-known CYP3A4 inhibitors include clarithromycin, diltiazem, erythromycin, grapefruit juice, itraconazole, ketoconazole, and protease inhibitors.[289]

As mentioned earlier in the antipsychotic section, droperidol has solid data supporting QT prolongation. A recent epidemiologic study found that in patients receiving psychotropic therapy, those taking droperidol were at highest risk for QTc interval lengthening compared with those taking other psychotropic agents.[121] Dose-dependent QTc interval prolongation after droperidol therapy was found in 20 surgical patients who received one of three dosage regimens, and it is listed in the product labeling.[290] Droperidol exhibits class III antiarrhythmic drug properties and even at low plasma concentrations, blocks IKr.[219] In late 2001, droperidol's labeling was changed to reflect new indications, new dosage recommendations, and a new black box warning. The FDA received reports of 20 cases of torsade de pointes with droperidol. Eighteen persons died; six of these deaths were attributable to torsade de pointes or QT prolongation.[291]

Proposed Association. Octreotide has been suggested to cause QT prolongation, although our search revealed no published reports on octreotide. In the precautions section of the product labeling, an association is noted of QT prolongation in patients with acromegaly who receive this drug.[72] Of note, electrocardiographic changes are common in patients with acromegaly.

Dolasetron, an agent for nausea, has been purported to cause effects on the QT interval.[292] The data are sparse and the association is weak; yet, the reaction is mentioned in the package insert.[74]

As a class, serotonin receptor agonists have been implicated in QT prolongation. Most data supporting this adverse effect are from sumatriptan, the agent that has been marketed the longest in the United States.

Probable Association. When initially marketed, the serotonin receptor agonists for the treatment of migraines were touted for their high safety profile. In fact, studies demonstrated no significant effects on the ECG.[293,294] The most concerning aspect of its use was in patients with established ischemic heart disease. In 3-5% of patients, symptoms of chest pressure occurred, but electrocardiographic monitoring indicated these symptoms are not typically associated with cardiac dysfunction.[295] Concerns with sumatriptan, however, have been raised, and a fatal cardiac arrhythmia has occurred during sumatriptan administration.[296] Prolongation of the QT interval is mentioned in its labeling.[76] Although our search revealed no published articles to support this adverse effect, other structurally similar migraine treatments have been implicated as causing QT prolongation, including naratriptan and zolmitriptan. Each has specific comments regarding QT prolongation within their respective package inserts.[75,78] Rizatriptan has been associated with nonspecific arrhythmias.[77] Almotriptan's product labeling does not mention QT prolongation or arrhythmias.[297]

Various other agents have been linked with QT prolongation and in some cases with torsade de pointes. In most situations, a paucity of data exists, but based on FDA-approved labeling, some agents are categorized as having a probable association.

Probable Association. Levomethadyl, along with methadone, was evaluated for its ability to block the HERG channel. Both agents were more potent inhibitors (at clinically relevant concentrations), when compared with fentanyl, meperidine, codeine, morphine, and buprenorphine.[298] A group of authors[299] reported on 17 methadone-treated patients who developed torsade de pointes; all of these patients received high-dose therapy (mean daily methadone dose 397 mg). Fourteen of the 17 patients had at least one potential risk factor for prolonged QT interval, including hypokalemia and concomitant administration of a drug associated with prolonged QT interval.

The muscle relaxant tizanidine is associated with QT prolongation. Initial data in animals revealed this adverse reaction. Our search revealed no published data available in humans; yet, QT prolongation is mentioned in the product labeling.[84]

Proposed Association. Epinephrine has been linked to QT prolongation, but whether this is a direct effect of increased catecholamine concentrations is not certain. Epinephrine-induced hypokalemia, mediated through the -adrenergic receptors, could be the mechanism of this effect. Discordant data have been published. Epinephrine infusions in one investigation were found to have no effect on the QT interval.[300] Certainly, catecholamines can result in afterdepolarizations. Epinephrine-induced torsade de pointes has been reported, although cases appear to be in predisposed individuals.[301,302,303] Epinephrine may prolong the action potential duration and result in increased QT dispersion.[303,304]

Two cases of tacrolimus-associated QTc prolongation followed by torsade de pointes have been published.[305,306] Contributing factors included concurrent trimethoprim-sulfamethoxazole, hypokalemia, concurrent ganciclovir, and a family history suggestive of congenital long QT syndrome. Tacrolimus is chemically a macrolide, related to erythromycin, that has a strong association with QTc prolongation and torsade de pointes. In addition, there is evidence supporting tacrolimus prolongation of the action potential by its effects on the potassium channel.[307]

A final agent for which there are scant data is salmeterol. Prolongation is mentioned in the salmeterol package insert,[82] although our search revealed no published data supporting this effect. In fact, some foreign data suggest salmeterol has no effect on the QT interval.[308,309]