The increased risk of drug-related torsades de pointes (TdP) and death in women taking ion-channel-active antiarrhythmic drugs was first described 20 years ago. In 1993, Makkar1 identified all case reports or case series of drug-related TdP, finding 332 patients described in 83 articles. The proportion of women versus men in these reported series was compared with the proportion of women versus men having received prescriptions for these drugs, determined using pharmacy-based databases. There were significantly more observed cases of TdP in women for multiple QT-prolonging drugs, including the class 1a agents quinidine, procainamide and disopyramide, the class III agent sotalol and the multiple-class agent amiodarone, than would have been predicted from the percentage of women prescribed the drug (see Figure 1).
While this study could have been limited by reporting bias, as only cases reported in the literature were analysed, clinical trials have since provided more robust data demonstrating that gender does influence the risk of TdP with QT-prolonging drugs. In a database derived from 22 clinical trials of oral d,l-sotalol,2 TdP developed in 2% of the 2,336 men and in 4% of the 799 women. After adjustment for other risk factors for TdP, including a history of ventricular arrhythmia or congestive heart failure (CHF), sotalol dose and renal function, women had three-fold greater odds of developing TdP than men. While women are more likely than men to require pacemakers with the use of other negatively chronotropic anti-arrhythmics,3 the sex difference in TdP risk with sotolol was not explained by differential dose-related bradycardic responses in women versus men.2 Similarly, in two recent studies of dofetilide, Diamond CHF and Diamond Myocardial infarction (MI), female gender was associated with a doubling of the odds of developing TdP, even after controlling for baseline QT.4 Intravenous ibutilide also is more likely to cause TdP in women.5
The recent Rate Control versus Electrical Cardioversion (RACE) trial of therapeutic strategies for atrial fibrillation further supported the increased danger of anti-arrhythmic drugs for women.6 While there was no difference in adverse events between those in the rate-control versus rhythm-control arms in men, women in the rhythm-control arm were three times more likely than women in the rate-control arm to develop adverse events. The similar Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) trial did not show a difference between men and women in arrhythmic events, but the overall pro-arrhythmia rate in that study was low.7
Underlying Pharmacokinetic Differences Between Men and Women
As recently researched and reviewed,8 there are some differences in pharmacokinetics between men and women that may contribute to these differences in pro-arrhythmia. Men on average weigh more, most drug-dosing studies have been conducted in men and for most drugs dosing is not weight-dependent, which could potentially contribute to higher levels in women. Metabolic enzymes exhibit varying gender effects. There are a group of enzymes called Cytochrome p450 – some are more active in men, some in women. Furthermore, renal excretion is greater in men, who show 10% greater glomerular filtration rate (GFR) after correction for bodyweight.8 However, while these differences are documented, they seem unlikely to explain the increased antiarrhythmic drug toxicity observed in women. In many renally excreted drugs, such as dofetilide or sotalol, dosing is based on GFR, which includes gender correction in the formula. For other drugs, such as quinidine, TdP is not dose-related, but occurs often with lower serum levels. Ibutilide’s differential impact on QTc in women has been demonstrated to be independent of level.9 Thus, pharmacokinetic differences may contribute, but are unlikely to be a primary cause of gender differences in drug-induced QT prolongation.
Differences in Repolarisation
A more likely explanation for the increased propensity of women to develop TdP in response to QT-prolonging drugs lies in well-documented differences in repolarisation in men versus women, which were first noted by Bazett in 192010 and later confirmed in the Framingham Heart Study.11 Rautaharju12 looked at evolution of the QT interval with age in a group of over 17,000 individuals in the National Health and Nutrition examination. In childhood, QT interval (corrected for HR) does not differ between boys and girls. However, the QT interval begins to shorten in males at puberty, decreasing by a mean of 20msec, and increases back to values similar to those in women in middle age. These QT differences appear due to changes in velocity of early repolarisation, as J-point and ST-angle are steeper in men, a difference also noted to decrease with age.13 The longer QT at baseline may predispose women to TdP when confronted with further lengthening of the QT by drugs. The physiological explanation for the difference in QT interval length between men and women may lie in differences in electrophysiological properties of the myocytes, in circulating hormones, in autonomic function or in interactions among these factors.
The Role of Circulating Female Hormones
The impact of circulating hormones on QT interval in women can be evaluated through studies of QT during the menstrual cycle, defined by varying hormone levels, and also through studies of post-menopausal hormone replacement therapy. Measurement of resting QTc at different phases of the menstrual cycle has not shown differences.14 However, interventions that may amplify repolarisation differences have suggested a role for circulating hormones. Burke et al15 measured QT and QTc in 23 female and 20 male volunteers, mean age 27 years, during three phases of the menstrual cycle in women, and three similarly spaced days for men, both at baseline and after autonomic blockade. At baseline, there were again no differences in QTc in women during the different phases of the menstrual cycle. However, after autonomic blockade, QTc was longer during the follicular phase than luteal, suggesting a QT-prolonging influence of circulating oestrogen and/or a protective effect of progesterone on repolarisation in women (see Figure 2). In a later study, Rodriguez et al.9 infused healthy volunteers with ibutilide at varying phases of the menstrual cycle, finding ibutilide-induced QT prolongation to be greater during the follicular and ovulatory phases than the luteal. In this study, QT prolongation correlated inversely with progesterone levels, but not with oestradiol, suggesting a protective effect of progesterone.
Studies of hormone-replacement therapy in post-menopausal women may provide evidence of effects of female hormones on repolarisation. In a small study of short-term oestrogen replacement (four weeks), there was no difference in QT or QTc with rest or exercise after hormone therapy.16 However, larger studies of longer-term hormone replacement do suggest a small but measurable role for circulating hormones. Kadish et al.17 analysed 34,378 women in the observational arm of the Women’s Health Initiative, and found that those taking unopposed oestrogen had a longer QTc than those on no hormone-replacement therapy (426 versus 423 months), but that those on oestrogen plus progesterone were similar to those on no replacement, suggesting a small QT-prolonging effect of oestrogen, with progesterone exerting a protective effect. Similar findings were seen in the population-based Atherosclerotic Risk in Communities (ARIC) study.18
Experimental studies support the hypothesis that circulating oestrogen influences repolarisation. In guinea pig ventricular myocytes, oestrogen at physiological concentration suppresses potassium current density (IKr) but not the slow component of the delayed rectifier current (IKs) or calcium currents. Pre-application of oestrogen-receptor blockers does not block the effect of oestrogen on IKr.19 However, other studies have shown IKs blockers but not IKr blockers to lead to greater action potential duration prolongation in female than male guinea pigs.20
Role of Circulating Male Hormones
The influence of circulating androgens on repolarisation has been examined through studies of men undergoing castration due to medical therapy for prostate cancer or trauma. Biadoggia et al. found that castrated men exhibited significantly longer QT intervals than age-matched normal men, and, in the small group receiving testosterone replacement, QT intervals shortened back to normal. Women with virilisation syndromes due to endocrine abnormalities displayed shorter QT than normal women, further emphasising the role of androgens.21
In basic studies, circulating testosterone also influences repolarisation. In guinea pig myocytes, testosterone decreases action potential duration (APD) in a dose-dependent manner. Testosterone enhances IKs while suppressing L-type Ca2 currents (ICa,L).22 In another study, long-term pre-treatment of female guinea pigs with dihydrotesterone decreased dofetilide-induced APD prolongation and decreased early after-depolarisations, despite identical oestradiol levels, again supporting a protective role of testosterone.23
Different autonomic responses between men and women may also influence the differing rates of pro-arrhythmia with QT-prolonging drugs. In the study by Burke,15 double autonomic blockade with propranolol and atropine led to QTc lengthening in both men and women, but the gender difference was unchanged, implying that resting autonomic differences do not explain the gender difference in QTc, which is related to factors other than resting autonomic differences. However, it is possible that differences in acute autonomic responses between men and women may have differential influences on repolarisation, affecting impact susceptibility to drug-induced torsades.
Effects of autonomic arousal on QT interval are well-documented. In a study by Toivonen et al.,24 participants were asked to wear holters while on call, and to record the times of pages that woke them from sleep. The QT interval was relatively longer for a given HR during stress, implying modulation of the QT–RR relationship by the autonomic nervous system. In a similar study, Nakagawa et al.25 asked healthy volunteers to wear holter monitors for 24 hours and then set an alarm for 5:00am. In the overall group, 20 seconds after the alarm RR shortened and QTc lengthened; however, while the RR interval shortened to a greater extent in men, the QTc lengthened to a greater extent in women. (see Figure 3). The greater RR shortening in men is consistent with prior studies demonstrating that cardiovascular reactivity to stress is greater in men.26,27 The greater increase in QTc in women suggests that repolarisation in women may be more sensitive to sympathetic stimulation, another potential mechanism of the increased toxicity seen with QT-prolonging drugs.
Gender Differences in Electrophysiological Properties of the Myocardium
Experimental studies have examined the electrophysiological differences in repolarisation between men and women. In isolated rabbit purkinje fibres, APD is longer in those isolated from female animals, and these female-derived myocytes are more prone to developing early after-depolarisations in response to provocative drugs.28 Liu et al.29 looked at cycle-length dependence of the QT interval in a Langendorff-perfused rabbit model. At relatively fast cycle lengths, female rabbit hearts displayed a slightly longer QT. However, as the cycle length was lengthened, the female hearts displayed a significantly larger lengthening of the QT than the male hearts, i.e. a steeper QT–RR relationship. Current density of IKr (one of the major repolarising outward potassium channels implicated in TdP) was less in the female myocytes than the male. There was no difference in the voltage dependence or activation or deactivation kinetics of IKr in the female myocytes. This reduced outward IKr may be one explanation of the slower repolarisation process and longer QT seen in females.
Gender and Beta-blockade
While beta-blockers are not ion channel-active ‘antiarrhythmic agents’, they are the one class of agents shown to be protective against sudden cardiac death. In the Metoprolol CR/XL Randomised Intervention Trial in Congestive Heart Failure (MERIT-HF) trial, patients with New York Health Authority (NYHA) class II–IV CHF and decreased ejection fraction were randomised to receive metoprolol or placebo. Beta-blockers improved not only total mortality, but also sudden cardiac death. In separate analyses of men and women (one-quarter of the overall group), the relative risk reduction for mortality with beta-blockade was similar: 21% in women and 18% in men.30 In a combined analysis of the three largest studies of beta-blockers in CHF – MERIT, the Cardiac Insufficiency Bisoprolol Study II (CIBIS-II) and Carvedilol Prospective Randomized Cumulative Survival (COPERNICUS) – the point-estimate for risk reduction for sudden cardiac death by beta-blockers was nearly identical between men and women (see Figure 4).30 Similarly, in a combined analysis of five large trials of the use of metoprolol in patients post-MI, the survival curves for men and women show a close to identical benefit for men and women.31
Women are more susceptible than men to pro-arrhythmia from QT- interval-prolonging agents. This increased vulnerability stems from longer repolarisation in women. The specific mechanisms of this difference include effects of circulating male and female hormones, as well as differences in underlying electrophysiological properties of the myocardium, and possibly differences in acute autonomic responses, between men and women. Care should be taken in the use of QT- prolonging drugs in women.