Polymorphisms and atrial fibrillation: sorting the wheat from the chaff

In the last 5 years, we have gained an increasing appreciation for the genetic contribution to atrial fibrillation (AF). While Mendelian families with AF have been reported, they have typically been considered rare. Studies from Framingham1 and Iceland2 have demonstrated a genetic basis for AF in the general population. Family members of those with AF have an odds ratios of ∼1.8 for developing the arrhythmia. This risk is considerably greater for younger patients1,2 and those with lone AF.3 One approach to identify common genetic determinants of a disease in a population is to perform an association study between a single nucleotide polymorphism (SNP) and the condition of interest. Such a study can be performed using one to many hundreds of thousands of SNPs across the genome. In the case of AF, numerous prior associations have been reported between AF and variants in the cardiac sodium and potassium channels, gap junction proteins, and inflammatory markers, among others. However, the studies to date have generally been underpowered, have not been replicated, and have a low pre-test probability of the SNP actually causing AF. Sinner et al. have described the relationship between the K897T SNP in KCNH2/HERG and AF in >1200 affected individuals and 2400 controls.4 Based on the well-known relationship between variation in KCNH2/HERG and ventricular repolarization, the authors hypothesized that this gene may also be associated with alterations in atrial repolarization, and thus AF. Starting with 40 SNPs across the KCNH2 gene, they determined the frequency of each SNP in 671 cases and 694 controls and found five associated with AF. These five SNPs were then analysed in an additional 531 cases with AF and 1781 control subjects, and one, the well-known K897T polymorphism, emerged. This SNP had an odds ratio of 1.25 [95% confidence interval (CI) 1.11–1.41, P = 3.3 × 10−4] with the 897K allele being associated with an increased risk of AF, while the 897T allele had a reduced risk. This association remained after correction for the age and sex of the subjects. With such a large, well-powered cohort, these investigators have raised the standard for future association studies on AF. However, this study was subject to some limitations. It would have been interesting to know the relationship between AF and ventricular repolarization in this population; unfortunately electrocardiograms were unavailable from the control subjects. This study is also limited by the lack of information on other covariates associated with AF, such as congestive heart failure and hypertension. Further, all of the subjects in the study were of European descent so the results many not be applicable to other races and ethnicities. Finally, with an odds ratio of 1.25, this variant in KCNH2 has a relatively modest overall effect in the population, implying that additional genetic variants for AF remain to be identified. What do we know of the K897T polymorphism? This SNP has a minor allele frequency of ∼23% in Caucasians and is therefore common in the general population. As summarized in Table 1, it has been extensively studied for its role in ventricular repolarization. In 2002, Pietila and colleagues described the association between the QT interval and K897T in 413 individuals. They found that the minor allele, 897T, was associated with QT prolongation but only in the subset of 187 women. The following year, Bezzina et al. reported the opposite, that 897T was associated with QT shortening in 1382 subjects and was more predictive in females. This result was replicated by investigators from the KORA study who found that 897T is associated with a 1.9 ms reduction in the QT interval per allele and with a greater decrease in women than men. Recently, a similar result was noted in the Framingham Heart Study, with a reduction in the corrected QT interval by 1.6 ms per 897T allele. Thus, while there is some inconsistency, on the whole the preponderance of epidemiological data supports the association between the QT shortening effect and the 897T allele. Summary of studies examining the relationship between the SNP K897T in KCNH2/HERG and the QT interval For reference, the K897T SNP (A2703C) is rs1805123 and has a minor allele frequency of ∼23% in Caucasians. n = number of subjects; *noted in women only. Summary of studies examining the relationship between the SNP K897T in KCNH2/HERG and the QT interval For reference, the K897T SNP (A2703C) is rs1805123 and has a minor allele frequency of ∼23% in Caucasians. n = number of subjects; *noted in women only. The results from these association studies create an apparent paradox. It has been well described in humans and in animal models of AF that electrical remodelling leads to shortening of the atrial action potential duration and that AF begets AF. There are also several reports of AF associated with short QT syndrome.5 Therefore, we would anticipate that AF would be associated with the 897T allele; however, it is the more common 897K allele that is associated with AF in the current study. How then do we reconcile the current findings with the expectation that AF is usually associated with shortening of atrial refractory periods? Unfortunately, a clear answer does not emerge from functional studies of this polymorphism. Characterization of the 897T and K alleles in cell lines has been inconsistent, with multiple different effects observed from several experienced research groups.6–9 In most cases, investigators have seen a reduction in channel current density as well as subtle effects on channel activation, inactivation, and deactivation. The net effect of a reduction in channel activity would be expected to result in a prolongation of the action potential duration and QT interval. The observed differences between the clinical findings of a shortening of the QT interval and the expression studies might be due to subtle differences in experimental conditions, or the inherent limitations in heterologous expression systems. It is interesting to note that there are a small number of cases where AF has been associated with prolonged repolarization. Two distinct gain-of-function mutations in KCNQ1 have been reported in a family and an individual with AF. While such an alteration would be expected to result in QT shortening, QT prolongation was noted in both cases.10,11 This discrepancy may reflect the failure of in vitro systems to recapitulate fully the intracardiac milieu of accessory subunits, signalling molecules, and other ion channels that interact to regulate repolarization. There is also the report of AF in an individual with a mutation of KCNA5 that encodes IKur. This loss-of-function mutation leads to prolongation of repolarization in atrial myocytes.12 Thus, we cannot discount the possibility that prolongation of atrial repolarization can under certain circumstances lead to AF, possibly through triggered activity. In sum, these disparate clinical, electrophysiological, and genetic findings highlight our limited understanding of the determinants of atrial repolarization. The study of Sinner et al.4 provides a foundation for further investigation of the relationship between KCNH2 and AF, and will ultimately require replication at other centres. It also emphasizes that large cohorts with adequate power are necessary when examining genetic associations, a particularly important point given the emergence of genome-wide studies for AF and other cardiovascular traits. Conflict of interest: none declared.