Introduction

Atrial fibrillation (AF) is the most common cardiac arrhythmia in the clinical setting and is characterized with fast and irregular abnormal atrial electrophysiological activities. The prevalence of AF increases substantially with age, ranging from about 1% in young adults to nearly 10% in those >80-year-old (Kannel et al. 1998). In the mainland China, approximately 10 million people are affected with AF (Hu and Sun 2008; Zhou and Hu 2008). AF is an independent risk factor for ischemic stroke (Wolf et al. 1991). In the mainland China, the prevalence of stroke in AF patients was much higher than that in non-AF population (Hu and Sun 2008; Zhou and Hu 2008). Common AF is associated with multiple risk factors, including age, valvular heart disease, hypertension, and left ventricular dysfunction (Go et al. 2001). Lone AF is referred to as AF which is not associated with other cardiovascular diseases. Lone AF accounts for nearly 30% of all AF cases (Brand et al. 1985; Lévy et al. 1999). Some AF patients have a positive family history, and several epidemiological studies suggest that genetic factors may play an important role in the development of AF, especially in patients with lone AF (Wang 2008).

Recently, major advances have been made in identifying the genetic factors underlying AF. To date, several genetic loci have been mapped for monogenetic Mendelian AF by different research groups, including our group (Wang 2008; Lai et al. 2003; Tsai et al. 2008). Several AF genes have been identified; some genes encode ion channels (e.g. SCN5A, KCNQ1, KCNE2, KCNJ2, KCNA5, and KCNH2) (Wang 2008; Chen et al. 2003; Sébillon et al. 2003; Yang et al. 2004; Hong et al. 2005; Olson et al. 2005, 2006; Xia et al. 2005). Most interestingly two non-ion channel genes have been recently identified for AF, including NUP155 reported by our group (Zhang et al. 2008) and NPPA (Hodgson-Zingman et al. 2008).

Although identification of genes for familial AF has provided enormous insights into the molecular pathogenic mechanisms of AF, they account for only a small proportion of AF cases. The most common form remains sporadic AF. The common AF is a complex disease which is caused by multiple genes, environmental factors, and gene–environment interactions. One large-scale case–control study using a genome-wide association strategy (GWAS) has been reported for AF in 2007. One single nucleotide polymorphisms (SNP) on chromosome 4q25, rs2200733, showed strong association with AF, and the finding was replicated in two other populations of European descent and a Chinese population from Hong Kong (Gudbjartsson et al. 2007). Follow-up studies replicated the 4q25 association with AF in additional five independent European populations (Anselmi et al. 2008; Kääb et al. 2009). However, the association in the Hong Kong Chinese population in the original GWAS report has not been replicated yet. Thus, it is unknown whether the 4q25 AF association remains significant in the most typical Chinese population from the mainland China because the Hong Kong population is derived mostly from a couple of provinces from Southern China. In this study, we tested the association between SNP rs2200733 and AF in a more representative Chinese Han population (GeneID). We further assessed the association of rs2200733 with lone AF in contrast to other types of AF.

It is interesting that SNP rs2200773 was found to be significantly associated with ischemic stroke (IS) in a separate GWAS from Iceland, and the finding was replicated in one German cohort and one UK cohort (Gretarsdottir et al. 2008). In this study, we also assessed the association between SNP rs2200733 and ischemic stroke in a Chinese population (GeneID).

Materials and methods

Ascertainment of patients

The Chinese GeneID population is a large cohort from the mainland China that aims to identify susceptibility genes for various cardiovascular diseases in the Chinese Han population. The study subjects have been enrolled from multiple hospitals in Wuhan, Dalian, and Beijing. The studies were approved by appropriate local institutional review boards on human subject research, and conformed to the guidelines set forth by the Declaration of Helsinki. Written informed consent was obtained from the participants. The participants completed a health questionnaire, and had fasted blood samples drawn. Genomic DNA was isolated from blood samples using standard protocols with the Wizard® Genomic DNA Purification Kit (Promega Corporation, Madison, WI, USA).

Diagnosis criteria

The AF cohort in GeneID consists of patients undergoing AF ablation as well as AF patients seen at the outpatient clinics over a period of last 5 years. Diagnosis of AF was based on standard diagnostic criteria (Fuster et al. 2006), and performed by a panel of cardiologists on the basis of electrocardiograms (ECG) and/or Holter ECG. Patients with other types of cardiac arrhythmias were excluded. Patients with cardiomyopathies and valvulopathies were also excluded based on data from echocardiography (Echo). AF patients were classified as “Paroxysmal AF”, “Persistent AF”, and “Permanent AF” according to ACC/AHA/ESC AF guidelines (Fuster et al. 2006). We limited the recruitment of AF patients to those affected men and women aged ≤75 years but with prior or current evidence of AF. An AF patient was diagnosed as “lone AF” if he/she had no hypertension, coronary artery disease (CAD), congenital heart disease, congestive heart failure, or diabetes. An individual was considered as an “AF control” if no AF, other types of arrhythmias, ischemic stroke, valvulopathies, and cardiomyopathies were detected by ECG, Echo, or magnetic resonance imaging (MRI)/computed tomography (CT).

The ischemic stroke cohort in GeneID consists of patients enrolled from the Stroke Specialty Clinic at three hospitals in Beijing. Diagnosis of ischemic stroke was based on the World Health Organization (WHO) criteria (Goldstein et al. 1989) by a panel of clinical neurologists. All patients had MRI or CT performed. ECG data were missing for the majority of ischemic stroke patients; thus, it was impossible to classify the patients with ischemic stroke into other sub-categories. Controls for ischemic stroke were defined as individuals without a history of stroke, and had normal MRI/CT imaging at the time of evaluation.

Genotyping of SNP rs2200733

SNP rs2200733 was genotyped using a Rotor-Gene™ 6000 High Resolution Melt system (Corbett Life Science, Concorde, NSW, Australia). Genotyping was performed in a total of 25 μL polymerase chain reaction (PCR) volume containing 1 μL of LC Green dye, 5 pmol of each primer, 25 ng of genomic DNA, 2.5 μL of 10 × PCR buffer with 1.5 mmol/L MgCl2, 5 mmol deoxynucleotide triphosphates, and 1 U of Taq polymerase. Two positive controls for each genotype (T/T, T/C, and C/C) were included in each run. A total of 24 cases and controls were randomly selected for verification of genotyping results using direct DNA sequence analysis. DNA sequence analysis was performed with forward and/or reverse primers using the BigDye® Terminator v3.1 Cycle Sequencing Kits on an ABI PRISM® 3100 Genetic Analyzer (Applied Biosystems, Foster City, CA, USA).

Statistical analysis

SNP rs2200733 genotypes were tested for Hardy–Weinberg equilibrium among controls using PLINK v1.05. Allelic and genotypic association of rs2200733 with AF was assessed using Pearson’s 2 × 2 and 2 × 3 contingency table χ2 test (SPSS, version 13.0). Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated using the χ2 test (SPSS, version 13.0). When the case–control samples were stratified, Breslow–Day tests were performed to analyze the homogeneity between ORs from each sub-group (SPSS, version 13.0). Multivariate analysis was performed by incorporating age, sex, hypertension (HT), and CAD as covariates by using multivariate logistic regression (SPSS, version 13.0). Empirical P values were determined using the PLINK v1.05 program with 100,000 Monte–Carlo simulations.

Power analysis

Power analysis was carried out using nQuery Advisor version 6.0.

Results

Power analysis

We studied 383 AF patients and 851 non-AF controls from the GeneID population consisting of the typical, mainland Chinese Han individuals covering 10 provinces from Central and Northern China. The average age at diagnosis was 58 ± 14 years for cases and 59 ± 11 years for controls (Table 1). Among the AF cases, 62.9% was classified as paroxysmal AF, 33.2% as persistent AF, and only 3.9% as permanent AF; 32.6% of patients had lone AF; 13.5% of AF patients had ischemic stroke and 6.0% had positive family history (Table 1). The male/female ratio and the rate of CAD and hypertension are shown in Table 1. For power analysis, we utilized the known population parameters of an OR of 1.42 and T allele frequency of 0.528 in the Hong Kong cohort from the previously published report (Gudbjartsson et al. 2007). Our study with 383 AF cases and 851 controls can provide statistical power of 97.9% at a Type I error rate of 0.05, which suggests that our sample size provides sufficient power to identify the association between SNP rs2200733 and AF.

Table 1 Clinical characteristics of the GeneID study population
Full size table

Our ischemic stroke study involved 811 patients with ischemic stroke and 688 non-stroke controls from the GeneID population. The average age at diagnosis was 61 ± 11 years for cases, and 61 ± 9 years for controls (Table 1). When using the setting of an OR of 1.4 and C allele frequency of 0.472 as reported previously (Gudbjartsson et al. 2007; Gretarsdottir et al. 2008), our study cohort can provide a statistical power of 99.6% at 0.05 Type I error, which strongly suggests that our GeneID cohort has sufficient power to identify the association between rs2200733 and ischemic stroke.

Allelic association of rs2200733 and AF in a Chinese Han population

There was no deviation from Hardy–Weinberg equilibrium for SNP rs2200733 in the control groups (P > 0.05; supplementary Table 1). We found highly significant association between SNP rs2200733 and AF in the Chinese GeneID population (Table 2). The T allele confers risk of AF. The P value for allelic association reached 3.7 × 10−11 with an OR of 1.81 and 95% CI from 1.51 to 2.15 (Table 2). The homozygous OR for rs2200733 was 2.83 (Table 2).

Table 2 Analysis of allelic association of rs2200733 T allele with AF and ischemic stroke in the GeneID population
Full size table

The association between SNP rs2200733 and AF was more significant in the lone AF group (OR = 2.40, P = 1.3 × 10−9) than in the other AF group (OR = 1.59, P = 6.2 × 10−7).

When the cases and controls were divided into a male group and a female group, the association remained significant in both groups with P values of 8.5 × 10−8 (OR = 1.83) and 8.6 × 10−5 (OR = 1.77), respectively (Table 2; Fig. 1). Then, the cohort was divided into different age groups, and the P values for association were 7.7 × 10−9 (OR = 2.07) for individuals with the age of <60 years and 3.6 × 10−4 (OR = 1.57) for older individuals (≥60 years). Association for AF without CAD was significant (P = 5.9 × 10−11, OR = 3.0), whereas that for AF with CAD became non-significant (P = 0.155). Association for AF without hypertension was significant (P = 3.3 × 10−8, OR = 1.96), whereas that for AF with hypertension was less significant (P = 4.5 × 10−4, OR = 1.61) (Table 2).

Fig. 1
figure 1

Comparison of odds ratios (ORs). Vertical axis OR; horizontal axis different groups of patients with AF or ischemic stroke sub-divided based on various criteria. Each filled rhombus with a line segment refers to the OR and 95% CI for each subgroup. The asterisk denotes the statistical significant difference between ORs of subgroups by homogeneity test (P < 0.05)

Full size image

Genotypic association of rs2200733 and AF in a Chinese population

We found statistically significant genotypic association between rs2200733 and AF. The association was more significant with the assumption of a dominant or additive model (P = 4.1 × 10−12 and 9.3 × 10−11, respectively) than a recessive model (P = 4.2 × 10−4) (Table 3).

Table 3 Analysis of genotypic association of rs2200733 with atrial fibrillation and ischemic stroke under different models of inheritance in the GeneID population
Full size table

To adjust for potential confounding factors, we performed a multivariate logistic regression analysis. Both allelic and genotypic association between SNP rs2200733 and AF remained highly significant after adjusting for possible risk factors, including age, gender, CAD, and hypertension (Tables 2, 3). Empirical P values were estimated by performing 100,000 Monte–Carlo simulations and found to be significant (Tables 2, 3).

Lack of association between rs2200733 and ischemic stroke in a Chinese Han population

We carried out a large case–control association study with 811 ischemic stroke patients and 688 controls to assess a previously reported association (Gretarsdottir et al. 2008) between SNP rs2200733 and ischemic stroke. No significant P values were obtained for either allelic or genotypic association between rs2200733 and ischemic stroke (P = 0.42 and 0.45, respectively; Tables 2, 3). The association remained non-significant in either male or female cohorts.

We also analyzed the genotyping data by dividing the cases into early onset (<60 years, 39.8%) and late onset (≥60 years, 60.2%) ischemic stroke patients. The association between rs2200773 and late onset ischemic stroke was not significant (P > 0.05; Tables 2, 3). A significant P value was obtained with the early onset group only for both allelic (P = 0.01, risk allele C OR = 1.32, 95% CI = 1.07–1.64) and genotypic frequencies (P = 0.005 with a recessive model) (Tables 2, 3). The association remained significant after adjusting for covariates of sex and age and after 100,000 Monte–Carlo simulations (Tables 2, 3). It is notable that the risk allele for early onset ischemic stroke is C in the GeneID population (Table 2), instead of the T allele for risk of ischemic stroke in the European populations.

Discussion

A 2007 GWAS report identified significant association between SNP rs2200733 on chromosome 4q25 and AF in an Icelandic population, and further studies replicated the association in a Sweden cohort, a US cohort, and a Chinese population from Hong Kong (Gudbjartsson et al. 2007). Later, the association in Caucasian populations was further replicated in German, Italian, Netherland, and US Framingham/Vanderbilt cohorts (Anselmi et al. 2008; Kääb et al. 2009). On the other hand, the association in the Hong Kong population has not been replicated yet. In the present study, we have carried out a large-scale case–control association study with 383 Chinese AF patients and 851 matched controls for SNP rs2200733. The results showed highly significant association between SNP rs2200733 and AF in the GeneID Chinese Han population. Both allelic and genotypic associations were highly significant even after adjustment for covariates (Tables 2, 3). The Chinese population in Hong Kong was derived mainly from Guangdong and Guangxi Provinces in Southern China. This study has expanded the association between rs2200733 and AF to the more representative mainland Chinese Han population with nearly 10 million AF patients (Hu and Sun 2008; Zhou and Hu 2008).

It is worth noting that the frequency of the risk allele T of SNP rs2200733 in the Caucasian populations deviates greatly from that in the Chinese population, <30 to >50%, respectively (Hapmap data, http://www.hapmap.org). These results suggest that SNP rs2200733 is a more common genetic risk factor in the Chinese population than in the Caucasian populations.

We compared the frequency of rs2200733 T allele in the previously published Hong Kong Chinese AF cases (60.5% of 333 cases) (Gudbjartsson et al. 2007) and that in our GeneID cases (65.6% of 383 cases), and found a significant difference (P = 0.003). No significant difference was found between two control groups (P = 0.09). It appears that SNP rs2200733 has a stronger effect on AF in our GeneID population (OR = 1.81, 95% CI = 1.51–2.15) than on Hong Kong population (OR = 1.42, 95% CI = 1.16–1.73). The P value for homogeneity of ORs was statistically significant (P = 0.024) (Fig. 1), suggesting that the effect of SNP rs2200733 on AF was significantly greater in the mainland Chinese population than in Hong Kong population. Further studies with large sample sizes are needed to determine the varying effects of SNP rs2200733 in different ethnic background and whether the difference in this study may be partly due to the small sample size of the Han Chinese in the Hong Kong study by Gudbjartsson et al. (2007) and the GeneID study.

Lone AF is AF without detectable cardiovascular disease such as congestive heart failure, high blood pressure, CAD, etc. Approximately 30% of AF cases belong to lone AF. Genetic factors were assumed to play a more important role in the development of lone AF than of AF associated with other cardiovascular diseases (Ellinor et al. 2005; Arnar et al. 2006). We tested this hypothesis using the GeneID population. When we divided the total AF cases into “lone-AF” and “other AF” groups, the frequency of the risk T allele was higher in the lone AF group (70.8%) than in the other AF group (61.6%) (P = 0.013). The OR for the “lone AF” group was 2.40, whereas that for the “other AF” group was 1.59. A significant difference was identified between these two ORs (P = 0.022) (Fig. 1). These results suggest that the influence of genetic factors is stronger for lone AF than for other types of AF and that rs2200733 confers a stronger risk in the lone AF population than in other types of AF populations.

The OR in AF cases diagnosed under 60 years of age was slightly higher than AF cases over 60 years (2.07 vs. 1.57), but the difference was not significant (P = 0.12) (Table 2; Fig. 1). We noted that the proportion of lone AF cases in the two different age groups differed greatly: 48.2 versus 17%, respectively (P value was 1.8 × 10−10). Hypertension is known to be a risk factor for AF. The association between rs2200733 and AF was significant in both AF with hypertension and AF without hypertension (Table 2). The results suggest that hypertension may not influence the contribution of rs2200733 to risk of AF. CAD is also a known risk factor for AF. The association between rs2200733 and AF was significant in AF group without CAD but not in the AF group with CAD (Table 2). The results suggest that SNP rs2200733 may not confer risk of AF in those patients also affected with CAD.

SNP rs2200733 was also identified in a 2008 GWAS for ischemic stroke in Iceland. The association was replicated in a German cohort and a UK cohort (Gretarsdottir et al. 2008). Our results suggest, however, that rs2200733 is not a risk factor for ischemic stroke in the Chinese population (P = 0.43). The association analysis was further carried out with the cases divided into two groups: one group with early onset ischemic stroke (<60 years of age) and the other group with late onset disease (≥60 years of age). A significant P value was obtained for early onset ischemic stroke but not with late onset disease. Despite the significant P value for the association between rs2200733 and early onset ischemic stroke, the risk allele in the GeneID early onset ischemic stroke cohort is the C allele, instead of the T allele in the European populations. The results suggest that the association between rs2200733 and early onset ischemic stroke is likely a false positive finding, or that rs2200733 is not a causative so that two different alleles of this SNP in different ethical populations are in linkage disequilibrium with the risk allele of an unidentified causative SNP. Thus, future studies in both Caucasian and Chinese populations are needed to further assess the association of SNP rs2200733 with ischemic stroke.

In conclusion, we found highly significant association between SNP rs2200733 on chromosome 4q25 and AF in the mainland Chinese Han population. The results expanded the association between rs2200733 and AF from a Hong Kong population to a more representative Chinese Han population.