Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Circ Arrhythm Electrophysiol. Author manuscript; available in PMC 2010 August 1.
Published in final edited form as:
PMCID: PMC2760022

Familial aggregation of Atrial Fibrillation – a study in Danish Twins



Heritability may play a role in non-familial atrial fibrillation (AF). We hypothesized that a monozygotic (MZ) twin whose co-twin was diagnosed with AF would have an increased risk of the disease compared to a dizygotic (DZ) twin in the same situation.

Methods and Results

A sample of 1137 same-sex twin pairs (356 MZ and 781 DZ pairs) where one or both members were diagnosed with AF were identified in The Danish Twin Registry. Concordance rates were twice as high for MZ pairs than for DZ pairs regardless of gender, 22.0% vs. 11.6% (p<0.0001). In a Cox regression of event free survival times, we compared the time span between occurrences of disease in MZ and DZ twins. The unaffected twin was included, when his or her twin-sibling (the index twin) was diagnosed with AF. After adjustment for age at entry, MZ twins had a significantly shorter event free survival time (hazard ratio: 2.0 (95% confidence interval (CI): 1.3 – 3.0)) thereby indicating a genetic component. Using biometric models, we estimated the heritability of AF to be 62 % (55 % – 68 %), due to additive genetics. There were no significant differences across genders.


All the analyses of twin similarities in the present study indicate that genetic factors play a substantial role in the risk of AF for both genders. The recurrence risk for co-twins (12–22%) is clinically relevant and suggests that co-twins of AF-affected twins belong to a high-risk group for AF.

Keywords: MesH, Atrial Fibrillation, Arrhythmia, Risk Factors, Twin Study, Genetics


Atrial fibrillation (AF) is a common type of arrhythmia, currently affecting more than 5% of the Western population over 65 years1. AF is associated with risk of thromboembolic complications, heart failure and death2. It is commoner in men than women and in those with heart failure, valvular heart disease, hypertension, diabetes, and with increasing age3, 4. In approximately 12%(−30%) of AF patients no concomitant heart disease is present, a condition known as ”lone AF”5, 6. Although AF is a common disease and constitutes a major problem in public health7, current treatment methods for AF are far from satisfactory. Various treatment strategies are currently deployed but still recurrence rate is high and many patients finally develop permanent AF refractory to any attempt to obtain sinus rhythm (SR) including electric cardioversion. This lack of true success in treatment of AF is partly explained by the scarce knowledge of the pathogenesis of AF.

It is not known why the majority of individuals with hypertension or valvular or structural heart disease remain in SR even in an advanced age while others develop AF in the absence of these or any other known risk factors. These facts make it reasonable to consider AF a multifactorial disease in which genetic factors may play an important role in defining the risk for the individual of developing the disease.

In 1947 Wolff described three brothers who all were diagnosed with AF at a young age8, and since then there has been a debate about the heritability of this arrhythmia. From kindreds with familial AF we know that certain mutations can cause the disease, but until now only private mutations not found outside the actual family have been described911. On the other hand common polymorphisms have been shown to associate to risk of AF in patients with non-familial AF and in the general population at large1216. These findings indicate that AF to some degree is a polygenetic disease. An attempt to estimate the heritability of AF in the general population has been done with data from the Framingham Heart Study, showing that parental AF is a risk factor for AF in offspring12. Likewise a study of the general population in Iceland demonstrate that Icelandic patients with AF are more closely related to each other than the rest of the population suggesting a significant degree of heritability in non-familial AF17.

No relatives are genetically closer to each other than monozygotic twins. When twin populations are sampled and followed, it is possible to collect epidemiological data concerning more common diseases, their distribution, incidence, and prevalence, and their relation to sex, age, zygosity etc. The classical comparisons between monozygotic (MZ) twins (with completely identical genes) and dizygotic (DZ) twins (with no more genes in common than ordinary sib pairs), can provide estimates of heritability, i.e. the extent to which the variation in a given population with respect to a certain trait is gene dependent18. In the present study we investigated the incidence of AF among Danish twins whose twin-sibling (i.e. index twin) already had developed the disease and from calculations of concordance rates and hazard ratios, we aimed at estimating the co-twin risk of being diagnosed with AF during follow-up and the heritability of non-familial AF.

Materials and methods

The Danish Twin Registry is a nationwide and population-based registry, established in 1954. The registry contains information about twins born between 1870 and 2004 and identified through church records or through the Central Office of Civil Registration. The registry currently holds data of more than 75,000 Danish twin pairs. All twins in the registry are ascertained independently of any disease. Zygosity is self-reported and established through a questionnaire with questions on the degree of similarity between twins in a pair19, 20. Validity of this kind of classification has been evaluated through blood samples and misclassification found to be less than 5% of cases21.

Information on diagnosis (World Health Organization; Inte rnational Classification of Diseases, 8th edition: code 427.4; 10th edition: code I48.9) was collected from The Danish National Patient Registry. This registry contains information about all patient contacts with all Danish Hospitals since 1977, including the specific hospital and department, admission- and discharge dates, and diagnoses (link:

Through merging information from the two registries, all twin pairs where at least one of the twins was diagnosed with AF were identified. For all twins identified the following information was sampled: zygosity, sex, time of birth, vital status, time of death or immigration, and time of first diagnosis of AF. From the initial dataset twins born earlier than January 1st, 1912 were excluded in order to avoid misclassification due to diagnosis not registered in The Danish National Patient Registry. Twins with unknown zygosity and twin pairs with information only on one twin were also excluded from the dataset. Opposite sex twin pairs were excluded because data on opposite twin pairs are only available for a fraction of the cohorts under study22. Furthermore, before entering the Cox proportional hazards model, twin pairs where the co-twin was lost to follow up before diagnosis of index twin, were excluded. Finally one pair of twins was excluded due to doubt of diagnosis as the index twin apparently was diagnosed at the age of seven.

Statistical analysis

Analysis of proband-wise concordance rates

The proband-wise concordance rate is the probability that a twin gets the disease given that his or her twin partner already has developed the disease. Concordant twin-pairs are defined as pairs where both twins have been diagnosed with AF and discordant twin-pairs as pairs where only one twin has the diagnosis. The proband-wise concordance rate is preferred to other types of concordance rates because it does not vary with the ascertainment probability, and is found to equal on average the population case-wise rate23.

As MZ-twins are completely identical for all genetic factors and DZ-twins share on average 50% of their genes, higher concordance rates amongst MZ-twins are interpreted to be caused by genetic factors. Differences between concordance rates for MZ- and DZ-twins are tested with the Chi square test with one degree of freedom. The concordance rates was calculated manually, using the formula for proband-wise concordance rates.

Cox proportional hazards model

The genetic effect was explored using survival analysis in the Cox proportional hazards model. The diagnosis free time after diagnosis of the index twin was compared between MZ and DZ twins. The index twin, is defined as the twin first diagnosed with AF within a twin pair. The co-twin is defined as the twin who is the twin-sibling of the index-twin.

A shorter time span between occurrences of disease in MZ twins would be indicative of a genetic effect. Thus, for each twin pair we calculated the time span from diagnosis of the index twin, to diagnosis of the co-twin. For some pairs this time was right censored due to end of follow-up or death. In a Cox model the relation between time from diagnoses and zygosity, sex, and age was assessed. In further analysis, we included an interaction term between sex and zygosity, to assess whether the genetic effect is dependent on sex. In order to investigate the genetic effect in younger twins, we analysed the twins divided into two stratae, defined as ≥ 65 years and < 65 years. Results are given as hazard ratios with 95% confidence limits. Cumulative incidence curve was constructed from Kaplan-Meier coordinates and differences between strata (zygosity) evaluated by log-rank test. We used SAS version 9.1. and the procedure ”proc phreg”, calculating the Cox proportional hazards model.

Biometric models

Biometric models were used to estimate the heritability. The relative importance of genetic factors is here assessed by the liability approach. Liability is based on threshold models, reflecting prevalence on a latent distribution of liability. Individuals above this threshold are assumed to have the trait of interest while individuals under the threshold are assumed not to have the trait24.

The correlation between relatives – in the present study MZ- and DZ-twins – is used to partition the correlation into components attributable to shared genes and environments, in order to estimate heritability. The correlation coefficients rmz and rdz calculated for MZ- and DZ-twins respectively, provides information of genetic as well as environmental influences and assuming equal environments, differences in the two correlation coefficients will represent the influence of genetic factors. The following relations are the standard assumptions about the quantitative genetics:


where a2 corresponds to the proportion of the total variance associated with additive genetic effects (A), d2 with dominant genetic effects (D), c2 with shared environmental effects (C), and e2 with non-shared environmental effects (E). All components are assumed to be independent. Different genetic models can be tested combining these elements but no more than three components can be simultaneously represented in a model for MZ- and DZ-twins reared together. An ADE model refers to decomposition of frailty Z = A + D + E, an AE model refers to Z = A + E, etc. Dominant and shared environmental factors cannot be represented simultaneously because of confounding of the two in a study of twins reared together. The model assumes no epistasis, no gene-environment interaction, and no assortative mating. Selection of the best fitted model is based on Akaike information criterion ((AIC=χ2−2·df). We used the Mx-software25 for the estimation procedures, and all models were fitted to contingency tables separately for males and females.

Sub-analysis of early onset AF

In order to study the effect of zygosity and sex in the younger age groups, we used a cut-off age of 65 year at the time of diagnosis of proband. This value was chosen as a compromise in order to define a population as young as possible without loosing the possibility to perform the analysis at all due to too low number of events. The estimates was calculated as odds ratio with 95% confidence limits from a 2×2 table using the Chi square and Fisher’s exact test.


A total of 1,137 twin pairs were identified, 356 MZ and 781 DZ pairs. Of these 1,137 pairs, 92 were concordant and 1,045 were discordant pairs. Mean age of the concordant pairs at inclusion (time of first diagnosis) was 68.2 years and for the discordant pairs 67.7 years, (Table 1).

Table 1
Baseline characteristics of participants by zygosity and diagnostic status at inclusion.

Concordance rates

Proband-wise concordance rates are shown in Table 2. Analysing all 1,137 twins in one group regardless of gender showed concordance rate for MZ twins approximately twice the rate for DZ twins: 22.0% vs. 11.6% (p<0,0001). Stratifying for gender did not change these figures significantly: 23.3% vs. 11.4% (p<0,001) for women and 21.1% vs. 11.7% (p=0,001) for men.

Table 2
The number of concordant and discordant twin pairs and proband-wise concordance rates for atrial fibrillation, by sex and zygosity.

Cox proportional hazards model

From the original total number of 1137 twin pairs we identified 806 pairs, 255 MZ and 551 DZ pairs, where the co-twins were alive after the diagnosis of the index twin and thus fitted into a Cox proportional hazard model. The risk of AF was significantly associated with monozygosity (hazard ratio (HR): 2.0; 95% confidence interval (CI): 1.3 – 3.0; p=0.0009).

The disease rate is twice as high in MZ twins compared to DZ twins (Error! Reference source not found.Figure 1) and this difference is clearly statistically significant, indicating that AF has a genetic component. As expected entry-age had a strong effect on AF risk (HR: 1.5; 95 % CI: 1.2 – 1.9) for each 10 year increase; p<0,0001). Disease rate was 30 % higher in men, but not statistically different from the female rate (HR: 1.3; 95 % CI: 0.84 – 2.0; p = 0.24). When analysing two age groups separately, we find a stronger genetic effect of zygosity in twins <65 years than twins ≥ 65 years. (HR: 2,9, p = 0,0036 vs HR: 1,65, p = 0,0514), the difference is not statistically significant (p = 0,21). This effect was expected, as we see a tendency for the time to diagnosis of the co-twin to decrease as a function of the age of the index-twin at time of his or her diagnosis. The effects of age as an independent risk factor plays a role.

Figure 1
Plot of cumulative incidence of atrial fibrillation as a function of zygosity for both genders analyzed together. CI=confidence interval.

In the analysis allowing for a sex dependent effect of zygosity, this appeared to be stronger for women (HR: 2.1; 95 % CI 1.2 – 4.0; p=0.017) than for men (HR: 1.9; 95 % CI 1.1 – 3.3; p=0.021), but the difference between the two genders was not statistically significant (p=0.78). This difference was even more pronounced though still not statistically significant, when analysed for the younger age groups. Analysing only patients younger than 65 years at time of diagnosis of proband showed odds ratio for women at 5.2 ( 95% CI: 1.2 – 23.1; p=0.034) and for men 2.7 ( 95% CI: 1.1 – 6.7; p=0.020) (Table 3).

Table 3
Distribution of events stratified for gender and zygosity and with age at diagnosis of index-twin less than 65 years (n = 391).

Biometric models

Contingency tables along with prevalence for atrial fibrillation divided into sex and zygosity groups are shown in Table 4. Tetrachoric correlations for males and females are given in Table 5 together with p-values for test of homogeneity of thresholds across zygosity groups. Females show significantly different thresholds (p<0.01) caused by differing prevalence in the zygosity groups, while there is no difference for males (p=0.08). Assuming the same thresholds across zygosity groups within the sexes there is a significant difference in thresholds (and thereby prevalence) between males and females (p<0.001). Model fit statistics for the biometric models are shown in Table 6, the AE-model provided the best fitting model (lowest AIC) for both males and females. The heritability of atrial fibrillation was estimated to be 67% (95% CI: 57%–76%) among females and 59% (95% CI: 49%–67%) among males due to additive genetic effects (Table 7). The difference in heritability of males and females was not statistically significant and the common estimate is 62% (95% CI: 55%–68%).

Table 4
Contingency tables and prevalence of atrial fibrillation among Danish Twins subdivided into sex and zygosity groups.
Table 5
Tetrachoric correlation
Table 6
Model fit statistics for biometric model
Table 7
Variance components for AE-model


In the present study of Danish twins with AF we demonstrate that once a twin is diagnosed with AF, the probability of his or her co-twin getting the disease is associated to zygosity. Two studies have previously addressed the heritability of non-familial AF or AF in the general population, concluding that risk of AF to some extent is a heritable condition12, 26, but these studies have not been able to disentangle the effect of common familial environment and genetic factors. As MZ twins have completely identical genes, no relatives are genetically closer to each other than they, and we assume that any variation in concordance rates of AF between MZ and DZ twins will be gene dependent.

Comparisons of concordance rates as well as analysis of twin similarities indicate that genetic effects play a significant role for the risk of AF for both genders. Concordance rates for MZ twins are more than twice the rates for DZ twins when both genders are analysed in one group, 22.0% vs. 11.6%, and this difference is highly significant (p<0.0001) Analysing both genders separately did not change these figures (Table 2). The analysis of twin similarities further supports this, by estimating the heritance of AF to be 62 % due to additive genetic factors (i.e. the effect of the genetic factors are acting additively and not primarily as gene-gene interaction which would have been reflected in a more than factor 2 difference in the correlation for MZ and DZ twins) (Table 7).

Cox proportional hazard regression models with age and gender as covariates showed a hazard ratio of 2.0, for developing AF for a MZ twin once his or her twin-sibling had developed the disease. Analysing twins younger than, and older than 65 years separately, showed that this effect is even stronger in the younger twins (HR: 2,9, p = 0,0036 vs HR: 1,65, p = 0,0514), although the difference is not stastistically significant. We had expected that AF in the younger twins would be more genetic, because they do not have the same degree of concomitant diseases that increases the risk of AF. Age and gender are known to be strongly associated to risk of AF with male sex and increasing age as predictors of increasing risk, but in our model only age and zygosity was significantly associated with risk for AF. Furthermore the increase in risk associated to zygosity was higher than that associated to 10 year increase in age. Testing the distribution of events stratified for gender and zygosity in the age groups younger than 65 years, showed OR for AF of 5.2 for women and 2.7 for men (Table 3), suggesting a stronger association between zygosity and AF for women than for men in the younger age groups although the difference between the two groups were non-significant. This finding is in accordance with data from the Framingham Heart Study12 where female AF had a stronger association to risk of AF in off-spring than male AF. In the present analysis this trend is increasing with decreasing age and we speculate if the explanation may be that female AF in the younger age groups is more likely to be “lone AF” and therefore of a more genetic nature. This would imply that AF in the older age groups to a higher degree is associated to concomitant conditions as hypertension, ischemic heart disease, diabetes, and obesity - all conditions with a male predominance. Disorders with a genetic predisposition often occur at a younger age. As mentioned AF primarily is a disease of the elderly but occasionally patients as young as in their early twenties are seen. The younger the patients are, the more likely they will present with AF in the absence of major predisposing conditions and therefore presumably with AF of a more genetic nature. Although we see a trend towards this conclusion in claiming that young female participants in our study present the strongest association between zygosity and AF, present data suggest that AF in all age groups has a considerable genetic component.

This study provides estimates of the sources of the variability in AF occurrence. Currently, a number of loci that confer increased vulnerability to AF and variance in candidate genes have been found both through linkage studies and genome-wide association studies. Association between loci on chromosome 1027 and 628 have been reported and the list of candidate genes is currently expanding: KCNQ19, KCNJ210, KCNE514, KCNE211 and the genes behind the renin angiotensin system15 and PITX229. The heritability estimates gives a very helpful overall estimate of the influence of genetic factors and can give us information about the prospect of identifying further genetic variants of importance for AF.


Monozygosity associates with increased risk of AF for twins whose twin-sibling has been diagnosed with AF. Assuming that the major difference between MZ and DZ twins is the degree of genetic similarity, we interpret this difference in risk of AF to be caused by genetic factors, and biometric models suggest a degree of heritability in AF as high as 62%.

Study limitations

The quality of the diagnoses in The Danish National Patient Registry is a topic for discussion. In a former validation of the registry, review of 116 medical records by a cardiologist confirmed the diagnoses in 112 cases.30 Twins diagnosed with AF before 1977 and not on any later occasion will not be registered in The Danish National Patient Registry. This potential misclassification was minimised by only including twins older than 65 years at the time of start of the registry.

Misclassification regarding the diagnosis is most likely independent of zygosity, and will make the concordance rates smaller in both groups of twins, thus making the groups more similar, and increasing the probability of the null hypothesis. Despite this fact, we have found a strong heritability estimate.

We make the assumption that any differences in concordance rates between MZ and DZ twins is due to the difference in their degree of genetic relatedness. This is standard twin methodology and all twin heritability estimates are made under the ”equal environment assumption” i.e. the degree of intrapair similarity due to the common environment is equal in monozygotic and dizygotic twins. This assumption has been much debated and in vestigated and it has generally been shown to be a valid and robust assumption31.

The majority of excluded twins, were excluded due to lack of follow-up information about the co twin. This was most frequent among the older twin pairs thus reflecting differences in data handling and collection over time and not differential reporting depending on zygosity or disease status. Regarding the group of twins born before 1912, 27 twin pairs that otherwise qualified for inclusion were excluded: 17 DZ (same sex) and 10 MZ pairs. None of these pairs were concordant with respect to disease. From the population in the analysis, the expected distribution of zygosity and concordance in a group of 27 twin pairs would be 19 DZ (same sex) and 8 MZ pairs and of these 2 pairs would be concordant. It is therefore not likely that any significant bias is introduced by this exclusion.

General Practioners (GP) in Denmark do not register ICD diagnoses in the NPR. Therefore, affected twins who have not been hospitalized, but only diagnosed by their GP, will not be included in this study. This means that the actual number of both MZ and DZ twins with AF may be higher. Certain clinical conditions, such as hypertension, ischaemic heart disease, diabetes and hyperthyroidism, are strong risk factors for AF. Our study design unfortunately did not give us the opportunity to assess the distribution of these risk factors among our twin population. As these conditions also are suspected to be at least to some degree heritable, this may be a potential bias.

Monozygotic twins know that they share most of their phenotypic traits, as they are “identical twins”, as opposed to dizygotic twins, who may not resemble each other neither physically nor psychologically. A monozygotic twin with a twin-sibling diagnosed with AF, may therefore be more eligible to seek a physician, in order to be examined for this same disease, than a dizygotic twin with a costwin diagnosed with AF. This will overestimate monozygotic concordance rates.

Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia. Familial clustering of AF has been known for more than 70 years. In population studies, children of parents with AF have a nearly double or even greater, risk of AF. Mutations and polymorphisms in genes coding for ion channels involved in cardiac repolarization and connexions may be involved in promoting AF. The present twin study adds confirms the basic notion that inherited factors play a role in the development of AF, and suggests that the importance of heritable factors may be greater than previously suggested. We find that monozygotic (MZ) twins have twice the concordance rate for AF, as dizygotic (DZ) twins. A MZ twin has twice the risk of developing AF, if his/her co-twin is diagnosed with AF, as compared to a DZ twin in the same situation. We estimate the heritability of AF to be as high as 67%. Family history is an important factor to take into consideration in the evaluation of patients at risk for developing AF. Understanding the genetic components and how they contribute to the pathophysiology leading to AF may lead to new approaches in diagnosis, prevention and treatment of this common cardiac arrhythmia.


Funding sources

Supported by The Danish National Research Foundation Centre for Cardiac Arrhythmia (DARC), The John and Birthe Meyer Foundation, and The Danish Cardiovascular Research Academy (DaCRA).

Lasse Steen Ravn was the recipient of a research fellowship from DaCRA.




Reference List

1. Kannel WB, Wolf PA, Benjamin EJ, Levy D. Prevalence, incidence, prognosis, and predisposing conditions for atrial fibrillation: population-based estimates. Am J Cardiol. 1998;82:2N–9N. [PubMed]
2. Frost L, Engholm G, Johnsen S, Moller H, Husted S. Incident stroke after discharge from the hospital with a diagnosis of atrial fibrillation. Am J Med. 2000;108:36–40. [PubMed]
3. Frost L, Engholm G, Moller H, Husted Decrease in mortality in patients with a hospital diagnosis of atrial fibrillation in Denmark during the period 1980–1993. Eur Heart J. 1999;20:1592–1599. [PubMed]
4. Friberg J, Scharling H, Gadsboll N, Jensen GB. Sex-specific increase in the prevalence of atrial fibrillation (The Copenhagen City Heart Study) Am J Cardiol. 2003;92:1419–1423. [PubMed]
5. Brand FN, Abbott RD, Kannel WB, Wolf PA. Characteristics and prognosis of lone atrial fibrillation. 30-year follow-up in the Framingham Study. JAMA. 1985;254:3449–3453. [PubMed]
6. Levy S, Maarek M, Coumel P, Guize L, Lekieffre J, Medvedowsky JL, Sebaoun A. Characterization of different subsets of atrial fibrillation in general practice in France: the ALFA study The College of French Cardiologists. Circulation. 1999;99:3028–3035. [PubMed]
7. Alpert JS, Petersen P, Godtfredsen J. Atrial fibrillation: natural history, complications, and management. Annu Rev Med. 1988;39:41–52. [PubMed]
8. Wolff L. Familial Auricular Fibrillation. N Engl J Med. 1947
9. Chen YH, Xu SJ, Bendahhou S, Wang XL, Wang Y, Xu WY, Jin HW, Sun H, Su XY, Zhuang QN, Yang YQ, Li YB, Liu Y, Xu HJ, Li XF, Ma N, Mou CP, Chen Z, Barhanin J, Huang W. KCNQ1 gain-of-function mutation in familial atrial fibrillation. Science. 2003;299:251–254. [PubMed]
10. Xia M, Jin Q, Bendahhou S, He Y, Larroque MM, Chen Y, Zhou Q, Yang Y, Liu Y, Liu B, Zhu Q, Zhou Y, Lin J, Liang B, Li L, Dong X, Pan Z, Wang R, Wan H, Qiu W, Xu W, Eurlings P, Barhanin J, Chen Y. A Kir2.1 gain-of-function mutation underlies familial atrial fibrillation. Biochem Biophys Res Commun. 2005;332:1012–1019. [PubMed]
11. Yang Y, Xia M, Jin Q, Bendahhou S, Shi J, Chen Y, Liang B, Lin J, Liu Y, Liu B, Zhou Q, Zhang D, Wang R, Ma N, Su X, Niu K, Pei Y, Xu W, Chen Z, Wan H, Cui J, Barhanin J, Chen Y. Identification of a KCNE2 gain-of-function mutation in patients with familial atrial fibrillation. Am J Hum Genet. 2004;75:899–905. [PubMed]
12. Fox CS, Parise H, D'Agostino RB, Sr, Lloyd-Jones DM, Vasan RS, Wang TJ, Levy D, Wolf PA, Benjamin EJ. Parental atrial fibrillation as a risk factor for atrial fibrillation in offspring. JAMA. 2004;291:2851–2855. [PubMed]
13. Lai LP, Su MJ, Yeh HM, Lin JL, Chiang FT, Hwang JJ, Hsu KL, Tseng CD, Lien WP, Tseng YZ, Huang SK. Association of the human minK gene 38G allele with atrial fibrillation: evidence of possible genetic control on the pathogenesis of atrial fibrillation. Am Heart J. 2002;144:485–490. [PubMed]
14. Ravn LS, Hofman-Bang J, Dixen U, Larsen SO, Jensen G, Haunso S, Svendsen JH, Christiansen M. Relation of 97T polymorphism in KCNE5 to risk of atrial fibrillation. Am J Cardiol. 2005;96:405–407. [PubMed]
15. Tsai CT, Lai LP, Lin JL, Chiang FT, Hwang JJ, Ritchie MD, Moore JH, Hsu KL, Tseng CD, Liau CS, Tseng YZ. Renin-angiotensin system gene polymorphisms and atrial fibrillation. Circulation. 2004;109:1640–1646. [PubMed]
16. Gudbjartsson DF, Arnar DO, Helgadottir A, Gretarsdottir S, Holm H, Sigurdsson A, Jonasdottir A, Baker A, Thorleifsson G, Kristjansson K, Palsson A, Blondal T, Sulem P, Backman VM, Hardarson GA, Palsdottir E, Helgason A, Sigurjonsdottir R, Sverrisson JT, Kostulas K, Ng MC, Baum L, So WY, Wong KS, Chan JC, Furie KL, Greenberg SM, Sale M, Kelly P, MacRae CA, Smith EE, Rosand J, Hillert J, Ma RC, Ellinor PT, Thorgeirsson G, Gulcher JR, Kong A, Thorsteinsdottir U, Stefansson K. Variants conferring risk of atrial fibrillation on chromosome 4q25. Nature. 2007;448:353–357. [PubMed]
17. Arnar DO, Thorvaldsson S, Manolio TA, Thorgeirsson G, Kristjansson K, Hakonarson H, Stefansson K. Familial aggregation of atrial fibrillation in Iceland. Eur Heart J. 2006;27:708–712. [PubMed]
18. Hauge M. The Danish Twin Register. Oxford University Press. 1981
19. Skytthe A, Kyvik K, Holm NV, Vaupel JW, Christensen K. The Danish Twin Registry: 127 birth cohorts of twins. Twin Res. 2002;5:352–357. [PubMed]
20. Skytthe A, Kyvik K, Bathum L, Holm N, Vaupel JW, Christensen K. The Danish Twin Registry in the new millennium. Twin Res Hum Genet. 2006;9:763–771. [PubMed]
21. Christiansen L, Frederiksen H, Schousboe K, Skytthe A, von Wurmb-Schwark N, Christensen K, Kyvik K. Age- and sex-differences in the validity of questionnaire-based zygosity in twins. Twin Res. 2003;6:275–278. [PubMed]
22. Skytthe A, Kyvik K, Bathum L, Holm N, Vaupel JW, Christensen K. The Danish Twin Registry in the new millennium. Twin Res Hum Genet. 2006;9:763–771. [PubMed]
23. McGue M. When assessing twin concordance, use the probandwise not the pairwise rate. Schizophr Bull. 1992:171–176. [PubMed]
24. Neale MC. Methodology for Genetic Studies of Twins and Families. Dordrecht: Kluwer Academic Publisher; 1992.
25. Mx: Statistical Modeling. 4th ed. Richmond VA: Department of Psychiatry, Medical College of Virginia; 1999.
26. Arnar DO, Thorvaldsson S, Manolio TA, Thorgeirsson G, Kristjansson K, Hakonarson H, Stefansson K. Familial aggregation of atrial fibrillation in Iceland. Eur Heart J. 2006;27:708–712. [PubMed]
27. Brugada R, Tapscott T, Czernuszewicz GZ, Marian AJ, Iglesias A, Mont L, Brugada J, Girona J, Domingo A, Bachinski LL, Roberts R. Identification of a genetic locus for familial atrial fibrillation. N Engl J Med. 1997;336:905–911. [PubMed]
28. Ellinor PT, Shin JT, Moore RK, Yoerger DM, MacRae CA. Locus for atrial fibrillation maps to chromosome 6q14–16. Circulation. 2003;107:2880–2883. [PubMed]
29. Gudbjartsson DF, Arnar DO, Helgadottir A, Gretarsdottir S, Holm H, Sigurdsson A, Jonasdottir A, Baker A, Thorleifsson G, Kristjansson K, Palsson A, Blondal T, Sulem P, Backman VM, Hardarson GA, Palsdottir E, Helgason A, Sigurjonsdottir R, Sverrisson JT, Kostulas K, Ng MC, Baum L, So WY, Wong KS, Chan JC, Furie KL, Greenberg SM, Sale M, Kelly P, MacRae CA, Smith EE, Rosand J, Hillert J, Ma RC, Ellinor PT, Thorgeirsson G, Gulcher JR, Kong A, Thorsteinsdottir U, Stefansson K. Variants conferring risk of atrial fibrillation on chromosome 4q25. Nature. 2007;448:353–357. [PubMed]
30. Frost L, Vestergaard P. Alcohol and risk of atrial fibrillation or flutter: a cohort study. Arch Intern Med. 2004;164:1993–1998. [PubMed]
31. Derks EM, Dolan CV, Boomsma DI. A test of the equal environment assumption (EEA) in multivariate twin studies. Twin Res Hum Genet. 2006;9:403–411. [PubMed]