|Home | About | Journals | Submit | Contact Us | Français|
Torsade de pointes (TdP) is a stereotyped polymorphic ventricular tachycardia characterized by a cyclical shifting of the QRS axis (“twisting” around the points of the isoelectric line) preceded by a prolonged QT interval (Figure 1). It is the quintessential arrhythmia of the long-QT syndrome, whether congenital or acquired, and results from a complex interplay among structural, metabolic, genetic and pharmacologic determinants. Epidemiologic risk factors such as gender, electrolyte imbalance, ischemia and QT-prolonging drugs are well established. In this review, we explore the epidemiology, proposed mechanisms, electrocardiographic risk factors and genetic architecture of TdP—particularly highlighting models for studying TdP and updated with the potential clinical implications from recent genetic association studies of QT interval and sudden cardiac death (SCD).
The incidence of TdP is difficult to estimate as it can manifest as syncope or SCD without electrocardiographic documentation of the rhythm. United States vital statistics recorded from 1989 to 1998 suggest that more than 300,000 lives are lost to SCD annually and the proportion of SCD among cardiac causes of death has increased.1 As many as 80% of individuals dying suddenly are found to have coronary artery disease with acute or prior myocardial infarction, while 10-20% have no evidence of structural heart disease.2 Presumptive cardiac arrhythmia has been suggested by autopsy studies as the most likely cause of sudden death in many persons less than 35 years of age who are free of structural heart disease.3
Results from post-mortem analyses have added to the greater appreciation for the potential role of primary electrical disorders such as TdP as an underlying cause of SCD. Chugh et al performed an autopsy investigation of 270 hearts from victims of SCD and found that 14 hearts (5%) were structurally normal.4 Postmortem molecular screening in 12 cases of unexplained SCD found that 2 out of 12 cases had the same missense mutation in exon 7 of the KCNH2 potassium channel gene.5 Tester et al performed a similar post-mortem mutation analysis of 49 cases of unexplained SCD, without structural heart disease on autopsy, and reported that over one third of cases involved an identifiable putative long-QT syndrome (LQTS) gene mutation with a significantly increased proportion of mutation carriers in females compared to males (44% vs. 7%, p < 0.008).6 These findings suggest that electrical disorders, whether genetic or not, may represent an important contribution to SCD in general and to SCD in individuals with structurally normal hearts in particular.
The QT interval is an easily measured and non-invasive risk factor, prolongation of which is associated with SCD in the general population. The heart rate-corrected QT interval (QTc) has a graded influence on increasing risk of SCD even among patients without recognized LQTS gene mutations.7,8 But in observational community-based studies it is unclear whether a) the QT association with SCD is explained by TdP or b) the QTc is a non-functional marker of other independent SCD risk factors such as coronary disease.9-11 Additional electrocardiographic risk predictors are discussed below.12,13
Female compared to male sex is associated with a two- to three-fold increased frequency of TdP, particularly after puberty.14,15 The sex differences may involve the influence of gonadal steroids on repolarization. Testosterone has been shown to shorten action potential duration in guinea pig ventricular myocytes, likely due to suppression of the L-type calcium current and enhancement of the delayed-rectifier IKs current.16 Testosterone also diminishes dofetilide-induced proarrhythmia in female rabbits.17 In guinea pigs, progesterone shortens action potential duration mostly through inhibition of inward calcium current and enhancement of the IKs delayed rectifier current.18 The role of estrogen as a contributor to disproportionate risk of TdP in women remains unclear.19
Electrolyte imbalances are a common risk factor for TdP, most importantly hypokalemia. Thiazide diuretics are associated with SCD, which may be mediated by their potassium-wasting effects.20 Repletion of potassium has been shown to improve repolarization in the setting of congenital or acquired QT prolongation.21,22 Aldosterone antagonists, which reduce potassium loss, were shown in the EPHESUS trial to reduce rates of SCD by 37% in patients with reduced left ventricular ejection fraction following acute myocardial infarction, although direct myocardial effects could be an alternate explanation.23
Increased risk of TdP is associated with structural heart disease,24,25 including ischemic heart disease.26 In a study of patients receiving dofetilide, 32 cases of TdP were observed; NYHA class III or IV dilated cardiomyopathy was associated with TdP with an odds ratio of 3.9 (CI 1.1-13.1) after controlling for QTc and gender.27 A recent trial of dronedarone in patients with severe heart failure was stopped prematurely due to increased mortality predominantly from heart failure death but also with a trend toward increased rates of SCD.28
Proarrhythmic effects of antiarrhythmic agents with strong IKr blocking effects are well recognized, particularly in the setting of risk factors as mentioned above, and have been a dominant factor in the lack of their more widespread use in arrhythmia prevention.29 But non-cardiac medications with comparatively modest effects on repolarization have been associated with TdP.30-32 When the risk has outweighed the benefit or non-arrhythmic alternatives in class exist, several of these compounds have been pulled from the market by regulatory agencies (Table 1). QT prolongation or other markers of TdP potential (e.g. potassium channel binding assays) are one of the greatest factors leading to the loss of promising therapies from the drug development pipeline.33,34 There has been an increased call for awareness of medications that prolong the QTc and/or predispose to TdP which has led to development of consensus-based lists of implicated drugs including an online registry located at www.qtdrugs.org.35,36
However, acquired QT prolongation is an imperfect surrogate for TdP risk. Some agents clearly prolong the QT interval but only infrequently lead to TdP. For example, amiodarone is known to display characteristics of all four Vaughan-Williams class antiarrhythmic effects including the blockade of both IKr and IKs delayed rectifier potassium channels (class III), sodium channels (class I), L-type calcium channels (class IV) and adrenergic receptors (class II).37 In canine models chronic amiodarone administration has been shown to prolong action potential duration and QTc but provide antiarrhythmic effects with a reduction in TdP rates.38,39 Recent rabbit models suggest a dose-related effect of amiodarone with lower doses inhibiting IKr and promoting arrhythmic activity while escalating doses preferentially block late sodium channel currents thereby suppressing proarrhythmia.40
Ranolazine, a novel anti-anginal agent known to inhibit late sodium channel current and IKr potassium current (thus prolonging the QT interval), was demonstrated in the large MERLIN-TIMI 36 randomized-controlled trial to have no significant increased risk of mortality or symptomatic documented arrhythmia.41 Secondary analyses found ranolazine use associated with reduced non-sustained ventricular tachycardia despite QTc prolongation,42 an observation supported by dog models.43 The recently designed RAID (Ranolazine Implantable Cardioverter-Defibrillator Trial) clinical trial is actively enrolling patients with increased risk for SCD to investigate the potential anti-arrhythmic properties of ranolazine (clinicaltrials.gov NCT01215253). The examples of amiodarone and ranolazine support a model whereby class III IKr - and IKs -blocking effects can be counteracted by sodium channel blocking or other anti-arrhythmic effects, leading to the corollary that not all drug-induced QT prolongation is created equal.
It has been proposed that the risk of TdP results from the culmination of multiple risk factors existing in combination with some degree of genetic predisposition. Dr. Dan Roden originally coined the term “reduced repolarization reserve” to refer to a model in which modifiable risk factors interact with an underlying fixed substrate, including genetic factors, to lead to proarrhythmia.44 Since cardiac tissue has multiple redundant repolarization mechanisms it would be unlikely for a single lesion to result in TdP. Rather, akin to contemporary carcinogenesis theories, the phenotype of TdP is the result of “multiple hits” which could include genetic predisposition—congenital LQTS, subclinical rare gene mutations, or predisposing common allelic variants as discussed—later as well as environmental or other non-genetic factors such as hypokalemia, IKr -blocking drugs, ischemia, etc.45
The mechanisms of proarrhythmia leading to TdP are not fully understood. However, experimental evidence suggests that the occurrence of TdP is a downstream endpoint in a chain of events involving a complex array of ion channel interactions that set up a substrate for arrhythmia in the setting of an appropriate trigger. Well described theories include dispersion of repolarization, triangulation, reverse-use dependence, short-term variability (also known as instability) and early-afterdepolarizations (EADs). We consider in turn these proposed mechanisms as well as a selection of relevant models for understanding TdP pathogenesis (Table 2).46,47
Whole organism models have the disadvantage of competing effects of anesthesia required for in vivo rabbit and dog experiments,48,49 as well as that of the translation and torsion of the left ventricle in intact animals. Yan, Shimizu and Antzelevitch have described an experimental model involving canine left ventricular wedge preparations to better understand the action potential in an electrically coupled environment50 while removing the confounding variable of anesthesia. In this model a transmural wedge is resected from the anterior wall of the left ventricle along with native branches of the left anterior descending coronary artery. Microelectrodes can then be used to record action potentials in epicardium, midmyocardium, and endocardium as well as subendocardial Purkinje fibers. This model has established that M cells, comprising the midmyocardium, have disproportionate prolongation of action potential duration (APD) in response to hypokalemia or class-III antiarrhythmics when compared with other layers of cells within the left ventricle myocardium. IKr -blocking agents such as d-sotalol were also found to preferentially prolong the APD of the M cell (Figure 2).51
Heterogeneity of action potential duration across the layers of conductive tissue within the ventricular wall locally, termed transmural dispersion of repolarization (TDR), predisposes to TdP52,53 in in vivo models. TDR has been demonstrated by Kozhevnikov et al in anesthetized dog models of the intact left ventricle with three-dimensional mapping of activation and repolarization during dofetilide infusion.54 Medina-Ravell et al demonstrated that increased TDR with biventricular pacing promoted TdP in a rabbit wedge model and correlated in humans with a similar biventricular pacing protocol with prolonged QT and JT intervals as well as polymorphic VT on surface ECGs.55
Isolated left ventricular wedge preparation models have limited resolution to detect the impact of between-region heterogeneity of repolarization. In an anesthetized canine model Opthof et al demonstrated greater apical-basal dispersion of repolarization arguing that some heterogeneity in repolarization is spatial (whole heart) as opposed to transmural.56 This hypothesis has been further supported in anesthetized swine models by Xia et al in which intact swine left ventricular endocardial and epicardial mapping was performed and spatial dispersion of repolarization was found to account for regional differences in peak to end of the T-wave (discussed below).57,58
Hondeghem et al developed an automated system called SCREENIT utilizing an isolated intact rabbit heart (Langendorff model) to try to predict compounds that could potentiate proarrhythmia.59 They then described the characteristics of “TRIaD” (Triangulation, Reverse-use dependence, Instability, and Dispersion) as a conceptual framework for understanding mechanisms of TdP.60-66
Using the Langendorff rabbit heart model, action potential duration (APD) is quantified by inserting septal recording electrodes into the subendocardial Purkinje fibers and measuring from the midpoint of the upstroke of the monophasic action potential to the end of the repolarization phase in subendocardial left ventricular septum.60 With particular class-III antiarrhythmics prolongation of APD is observed and the monophasic action potential develops a more triangular morphology losing the stereotypical plateau appearance of phase 2 in concert with blunted phase-3 repolarization (Figure 3). This is known as triangulation and is associated with proarrhythmia in animal models.
Reverse-use dependence in TRIaD is described as lengthened action potential duration in the setting of depolarization following a pause.60,61 This phenomenon is thought to set up short-term-variability, also known as instability, that may precede TdP. Due to the stochastic variation of slowly activating sodium current, a long RR interval or pause obligates a prolonged APD. If sinus cycle length is regular, then an inevitably shortened TR interval (RR minus APD) leads to a shortened APD which is then followed by a longer TR and a longer APD. This can become an oscillating short-long-short cycle and lead to the stereotyped pattern that can herald an episode of TdP. Instability, defined as short-term variation in APD, is thought to contribute to proarrhythmia as demonstrated in the Langendorff model. Dispersion of repolarization, the last component of the TRIaD conceptual framework, has been described above.
Early afterdepolarizations, in the setting of triangulation, instability and dispersion of repolarization, are thought to be a common trigger initiating polymorphic VT. Initial evidence for EADs generated in the intact electrically-coupled left ventricle was reported by Yan et al in which both canine wedge prep and rabbit isolated left ventricle preparations were challenged with exposure to dl-sotalol and azimilide (both IKr blockers) and monitored for phase 2 EADs.67 They found both dl-sotalol and azimilide produced frequent phase 2 EADs but progression to TdP-like polymorphic VT required evidence of marked increase in transmural dispersion of repolarization.
Milan et al recently introduced another potential model for study of myocardial repolarization and arrhythmogenesis using zebrafish. This in vivo model poses unique practical advantages in that the zebrafish is transparent during embryogenesis, can be paralyzed and orally perfused without anesthesia, and can be easily genetically manipulated and reproduced. They observed common QT-prolonging compounds such as haloperidol and astemizole can significantly prolong the QT interval on the zebrafish surface ECG compared to non-QT-prolonging controls.68
Another interesting new model in the study of repolarization was recently reported as a cellular in vitro model using induced pluripotent stem cells (iPS) derived from dermal fibroblasts of an individual with known LQTS type 1.69 In this model Moretti et al used a retrovirus vector to deliver human transcription factors into dermal fibroblasts to generate iPS cells with a known KCNQ1 gene mutation. These iPS cells were then directed to differentiate into contracting embryoid bodies expressing myocardial cell markers. Microelectrode recordings were used to infer atrial, nodal and ventricular type embryoid bodies. EADs in the LQT1 embryoid bodies were induced with exposure to catecholamines but was not observed in controls. This model represents a potential opportunity for in vitro study of human genotype-phenotype relationships and recent investigations have examined KCNH2 iPS cells and response to HERG-binding drugs, but several hurdles must be overcome including identification of appropriate controls and more robust phenotyping.70,71
Electrocardiographic QT interval prolongation has been the most commonly identified ECG manifestation associated with the risk of TdP. Indeed, a prolonged QT interval is a part of the diagnostic criteria for congenital long-QT syndrome;72 a QTc > 500 ms has been consistently associated with risk of syncope or sudden cardiac death in patients with congenital LQTS.73-75 Several studies have reported association of QT interval prolongation in unselected individuals with risk of sudden cardiac death. For example, a Dutch population-based study of approximately 8000 elderly men and women over the age of 55 showed a nearly three-fold risk of sudden cardiac death associated with a more modest QTc prolongation (> 450 ms found in 6% of males, > 470 ms in 5% of females).76 Chugh et al also demonstrated in a population-based case-control study of patients with known CAD, that after controlling for diabetes and QT-prolonging drugs, a prolonged baseline QTc was independently predictive of SCD (OR 5.53; 95% CI, 3.20 to 9.57).26 Increased QTc dispersion has been identified in population studies as a predictor of cardiac death77-79 but a direct association with TdP has not been well-validated.80
T-wave morphology and parameters have been increasingly studied to find additional ECG predictors of TdP. Based on compelling data from animal models demonstrating the cellular basis of the electrocardiographic T wave (Figure 2), perhaps the most studied ECG parameter is the interval between the peak of the T wave and the end of the T wave (Tpeak-end). It has been suggested that since the T wave itself more specifically reflects phase 2 and phase 3 of the monophasic action potential repolarization it more accurately represents elements predisposing to TdP.51,52,66,81 Microvolt T-wave alternans (beat-to-beat alternation in the morphology and amplitude of the T wave) has been extensively studied as a marker for stratifying risk for ventricular arrhythmias but has not been validated as a predictor of TdP.82
Yamaguchi et al tested this hypothesis in vivo studying 27 patients with documented drug-induced QT prolongation, of whom 12 developed TdP.83 Logistic regression analysis in this small sample found that Tpeak-end was a more reliable predictor of TdP than both QTc and QTc dispersion. This observation was further supported by Shimizu et al84 in patients with ventricular hypertrophy, by Topilski et al85 in patients with bradyarrhythmias and notched T waves, and by Watanabe et al86 in patients with documented non-sustained ventricular tachycardia who demonstrated increased risk of both induced and non-induced VT in the setting of prolonged Tpeak-end.
In vivo associations between T-wave morphology and TdP have been further supported in humans by Topilski et al who noted that patients without known LQTS gene mutations who had notched T waves had a strong correlation of longer Tpeak-end with risk for TdP.85 This has also been supported by Bozkaya et al who demonstrated predisposition to ventricular arrhythmias in patients with prolonged Tpeak-end, notched T waves or U waves.87
The oscillating short-long-short sequence of variability in QTc interval is known as short-term variability (STV) and has also been shown to predict TdP. Thomsen et al performed some of the first in vivo studies in anesthetized dogs with chronic AV block and demonstrated a dose-dependent increase in d-sotalol-induced TdP that was not explained by incremental prolongation of the QTc but rather by an increase in STV.88,89 A case-control pilot study by Hinterseer et al compared 20 patients with documented TdP with 20 matched controls and demonstrated significantly increased baseline STV assessed by surface ECG in cases, in the absence of QTc prolongation.90
It has become increasingly appreciated that the allelic architecture—a function of the frequency and the effect size of all genetic variants contributing to a trait—underlying the complex phenotype of myocardial repolarization is one involving a spectrum of genetic variants. At one end of this spectrum are rare variants such as those underlying congenital long-QT syndromes which have strong effects on myocardial repolarization leading to phenotypes such as QTc prolongation and a strong predisposition to syncope or SCD. These mutations are typically rare, private to individual families, and not found upon screening the general population at large. They arise spontaneously in a single individual and may be inherited although strong negative selective pressures due to lethality before reproduction often prevent such variants from rising in frequency within the general population.91 Less clinically apparent are rare variants with more moderate effects on repolarization representing variable or incomplete penetrance of heritable LQTS mutations which may have a more ambiguous inheritance pattern.92 Finally, common allelic variants discovered as single nucleotide polymorphisms (SNPs) associated with modest effects on myocardial repolarization have been identified in unselected populations, but their relationship to TdP remains yet to be determined. We consider in turn these three classes of alleles (genetic variants).
Congenital LQTS is a rare familial disorder with a prevalence estimated at 1:2000 to 1:5000; genotype-phenotype relationships and clinical course have been well-described.75,93 A survey of 262 unrelated LQTS patients found that the vast majority (78%) of mutations among LQTS patients were present in a single family or individual.91 The majority of identified mutations involve the LQT1, LQT2, or LQT3 genes. KCNQ1 (LQT1) and KCNH2 or HERG (LQT2) encode voltage-gated inward rectifying potassium channels and reduced repolarization current (loss of function of IKs and IKr respectively) can result from mutations causing intrinsic channel or protein trafficking defects.75 SCN5A (LQT3) encodes the cardiac sodium channel and the failure of these channels to close properly due to gain-of-function mutations can lead to action potential prolongation.94
In fact, some populations with unique demographic histories may appear to violate the rule that LQTS-causing mutations are private to individual families. This has been uniquely demonstrated in Finland in which four distinct founder mutations comprise as much as 73% of cases of LQTS.95 A recent screening of 6,263 individuals from a population-based survey in Finland identified 27 individuals who carried one of the four founder mutations demonstrating an increased prevalence of ≈0.4% or one in 250.96
LQTS mutations with incomplete penetrance have been shown to lead to the variable phenotype seen in many patients with congenital LQTS. This particular subset of patients is an important group for further study because they may help us to better understand why some patients carrying an identifiable congenital LQTS gene mutation may have a normal QTc at baseline but suffer sudden cardiac death in the setting of adrenergic challenge, hypokalemia or exposure to a QT-prolonging drug.
Some of the earliest evidence of incomplete penetrance was reported by Priori et al who studied LQTS probands and found that 33% of 46 genotyped family members thought to be clinically unaffected carried the same familial LQTS gene as the proband.92 A subsequent study by the same investigators genotyped 430 probands and 1115 family members demonstrating an average penetrance of 60%.97 Thus, while LQTS is a familial disorder with autosomal dominant inheritance leading to an average 50% gene transmission, phenotype penetrance was reduced suggesting that several so-called unaffected patients might actually be susceptible to TdP in the setting of an appropriate exposure. Lehtonen et al evaluated 16 cases of TdP and found that 3 individuals (19%) carried one of the four Finnish founder LQTS mutations despite a normal baseline QTc, demonstrating both incomplete penetrance of mutation carriers and the inconsistent resting QTc correlation with TdP risk.98
Several lines of evidence support the hypothesis that environmental modifiers can be an important second hit to individuals with latent mutations. Shimuzu et al have shown that epinephrine significantly increased QT and Tpeak-end intervals in asymptomatic LQT1 mutation-carriers with QTc < 460 ms.99 Vyas et al demonstrated that epinephrine challenge can unmask subclinical LQT1 gene-carriers with a median increase in QT interval of 78 ms.100 Jeyaraj et al evaluated largely unaffected LQT2 mutation carriers with baseline QTc < 450 ms and demonstrated an increased Tpeak-end interval after exposure to erythromycin, a known hERG channel inhibitor.101 Yang et al studied patients with presumed drug-induced QTc prolongation and found that 10-15% of cases had an identifiable mutation in a known congenital LQTS coding region.102
Genome-wide association studies (GWAS) involve testing the majority of common variation, which can be efficiently assayed in large sample sizes. GWAS has recently provided more insight into the heritability of complex traits such as myocardial repolarization by identifying common allelic variants that individually have a modest effect on the quantitative continuous QTc.103 These common variants could be associated with TdP and SCD by incrementally reducing repolarization reserve—an effect that could be augmented by environmental factors such as gender, ischemia, structural heart disease, hypokalemia, or QT-prolonging drug exposure. An additional advantage of GWAS is the ability to use large population-based cohorts without ascertainment on disease.
Some of the earliest studies of SNPs suggested that polymorphisms may be common in one population and rare in another yet produce a consistent repolarization phenotype. Splawski et al identified an amino acid-altering polymorphism of the cardiac sodium channel gene SNC5A S1102Y, of which the minor Y allele was identified in 13/23 African-American cases of arrhythmia and/or QT-prolongation compared to only 13/100 African-American controls reflecting a significantly increased risk of arrhythmia associated with this variant.104 Burke et al further validated the association of Y1102 with SCD in blacks105 while Chen et al reported this same allele Y1102 in a Caucasian family (the allele is rare in individuals of European ancestry, <1%) and noted a strong association with syncope, ventricular fibrillation and SCD.106
Common QT variants have recently been identified in several genes, some known to be involved in myocardial repolarization and some novel. Early GWAS investigating QTc prolongation identified common variants in the nitric oxide synthase 1 adaptor protein, NOS1AP, a gene involved in regulating neuronal nitric oxide synthase now known to modulate cardiac repolarization.107-110 Two much larger GWAS meta-analyses were recently published.111,112 The analysis of the QTGEN consortium involved 13,685 individuals of European ancestry and identified variants in five loci (NOS1AP, KCNQ1, KCNE1, KCNH2, SCN5A) known to be involved in myocardial repolarization as well as variants in five newly identified loci collectively explaining approximately 6% of QT interval variability. The analysis of the QTSCD consortium involved 15,842 and also validated the identification of common variants in known LQTS genes (KCNQ1, KCNH2, SCN5A, KCNJ2) as well as variants in novel loci without known roles in cardiac electrophysiology. Table 3 lists common genetic variants associated with the QT interval.
Recent studies provide evidence that common variants identified as single nucleotide polymorphisms (SNPs) can also unmask latent rare familial mutations with incomplete penetrance. For example, Crotti et al genotyped an older female proband who presented with ventricular fibrillation as well as acquired QTc prolongation and diagnosed LQT2 involving a rare KCNH2 mutation that was shared by multiple relatives in her family.113 However, the proband also carried a common polymorphism K897T on the non-mutant allele that involved KCNH2. Relatives with LQT2 who did not also carry the polymorphism K897T were asymptomatic suggesting that common variants can be modifiers of the expression of latent congenital LQTS mutations, although limited numbers of individuals with and without the polymorphism and the mutation prevent definitive conclusions. Similarly, a South African LQT1 founder population harboring a mutation in KCNQ1-A341V was studied to examine the modifying effect of common variation in NOS1AP, which encodes a nitric oxide synthase adaptor protein and is known to have quantitative QT effects.114 NOS1AP alleles associated with QT interval lengthening were significantly associated among LQT1 carriers with QTc prolongation as well as cardiac arrest and sudden cardiac death. Subsequent studies have further validated the complex interplay of common variants found in individuals with congenital LQTS.115-117
There has been limited investigation into common variants and their association with other electrocardiographic measures of repolarization such as Tpeak-end, short-term variability, etc.—all potential ECG manifestations of disordered repolarization. A study by Porthan et al examined SNPs (known to influence QTc) and their association to Tpeak-end as well as specific T-wave morphologies. 118 The results of this study demonstrated very modest influence of polymorphisms on Tpeak-end but suggested little contribution of common variants to the prognostic value of T-wave morphology.
Prolongation of the QT interval is associated with sudden cardiac death. It is therefore plausible that common genetic variation associated with longer QT interval will be associated with increased risk of SCD. Common variation at NOS1AP has been studied to understand its association with SCD since it has been the most strongly associated locus in GWAS. A study from the Cardiovascular Health Study and Atherosclerosis Risk in Communities studies119 found that a common allele at NOS1AP associated with longer QT interval was in fact associated with SCD in white adults. Meta-analysis of these findings with an independent study from the Netherlands further strengthened the association.120
More recently, a candidate gene study of torsade de pointes cases and controls121 demonstrated association of the KCNE1 missense D85N SNP that has been associated with QT interval duration112 and congenital Long QT Syndrome.117 These studies provide proof of principle that identifying common variants related to an SCD risk factor can also identify SCD and TdP variants. Ongoing work is testing other common QT interval variants for a relationship to SCD and drug-induced QT prolongation.
TdP remains a clinically important problem in the context of a large burden of SCD. The QT interval continues to play a powerful role in predicting risk of SCD although it is a crude and imperfect predictor of TdP. We have learned a great deal from genetic studies, most recently in the form of GWAS, in better appreciating the complex allelic architecture underlying channelopathies predisposing to TdP. The potential role of newer models including zebrafish68 and induced pluripotent stem cells69 provide exciting opportunities for improving our ability to understand the complexities of this disease.
C.N.-C. was supported by the National Institutes of Health (HL080025, HL098283), the Doris Duke Charitable Foundation and the Burroughs Wellcome Fund.