Atrial fibrillation (AF) is the most common clinically significant cardiac arrhythmia1–3 and is associated with high mortality4 and morbidity and reduced quality of life.5 The prevalence and incidence of AF are high and increase with age.6–8 The costs and hospitalizations attributable to AF have risen markedly and a continued increase is expected, due to population aging.9,10

Clinical cardiovascular risk factors such as obesity, hypercholesterolemia, diabetes mellitus, and hypertension are associated with higher risk of AF.11–13 Some electrocardiographic characteristics have also been related to AF. Among them, P-wave duration and interatrial blocks (IAB) have been associated with AF risk in some studies.14–20 Interatrial blocks are recognized as an electrocardiographic abnormality based on the presence of a P-wave duration ≥ 120 milliseconds, called partial IAB, or ≥ 120 milliseconds with a biphasic P-wave morphology in the inferior leads, called advanced IAB.21,22 Interatrial blocks have also been associated with a higher risk of supraventricular tachyarrhythmias,23–26 and the confluence of IAB and this type of arrhythmias has been recently named the Bayés syndrome.27

The evidence of a relationship between AF and P-wave characteristics, both duration and morphology, in the general population is scanty. We hypothesized that long P-waves and the presence of advanced IAB are independently associated with the risk of AF. The aim of this study was to determine the individual and combined associations between 2 P-wave characteristics, duration and morphology, and risk of developing AF in a population-based study.

The median follow-up of the cohort was 7.12 years. The flowchart of participant inclusion is shown in Figure 2. Those attending the follow-up were younger, a high proportion were women, and had lower levels of systolic blood pressure, glycaemia, and triglycerides, as well as a lower proportion of treatment for hypertension, diabetes, and hypercholesterolemia at the baseline visit (Table 1 of the supplementary material). The analysis included 80 individuals with AF and 160 controls. Participant characteristics are shown in Table 1. The mean age was 74.3 years, and most of the participants were men. No differences in the prevalence of hypertension, hypertension treatment, diabetes, diabetes treatment, or in the mean values of systolic blood pressure, diastolic blood pressure, or heart rate were observed between cases and controls. However, the group of AF cases presented with higher body mass index than the group of controls.

Figure 2.

Flowchart of the participants included in this study. AF, atrial fibrillation; ECG, electrocardiogram.

(0.39MB).

P-wave characteristics and differences between cases and controls are shown in Table 2. P-wave duration was longer in the group of AF cases than in the control group. The presence of IAB, partial or advanced, was also more frequent in AF cases than in controls. Atrial premature beats were only observed in cases, not in controls. P-wave duration was associated with heart rate (Pearson correlation coefficient = −0.167; P value = .010), body mass index (Pearson correlation coefficient = 0.121; P value = .062) (Table 2 of the supplementary material), and hypertension and hypertension treatment (Table 3 of the supplementary material).

In the multivariate conditional logistic regression models including body mass index as a covariate (Table 3), we observed that P-wave duration (as a continuous or categorical variable) and the presence of partial or advanced IAB were directly associated with AF. Body mass index was also related to AF with odds ratios (OR) ranging between 1.065 and 1.071 in the logistic regression models. Additional adjustment for baseline heart rate and hypertension did not affect the magnitude of the observed associations (Table 4 of the supplementary material). When considering P-wave duration, we assessed the linearity of the association with AF and observed that the nonlinear component was statistically significant (P = .019) (Figure 3). We observed increased OR for AF when the duration exceeded 110 milliseconds; longer P-wave durations did not additionally increase the OR for AF.

Figure 3.

Smoothing spline of the association between P-wave duration and atrial fibrillation (P-wave duration limited to a range of 80 to 180 milliseconds). OR, odds ratio.

(0.08MB).

The association of abnormal P-wave morphology, defined by the presence of advanced IAB, with AF tended to be similar to that of partial IAB (Table 3). When considering P-wave duration and morphology simultaneously, we observed that the association with AF was present when P-wave duration was longer than 110 milliseconds, independently of the coexistence of advanced IAB. The OR for AF in the group with advanced IAB was very similar to that of the group with partial IAB (Table 3).

As a post hoc statistical power estimation and according to the sample size of our study and assuming an alpha risk of 0.05 and a beta risk of 0.20 in a 2-sided test, the ORs that could be detected as statistically significant was ≥ 4.5 on comparison of the group with advanced IAB vs the group with normal P-wave, and was ≥ 3.3 on comparison of the group with advanced IAB vs the group with partial IAB.

The results of this study confirm the association between P-wave duration and AF. We observed that AF was associated with P-wave duration > 110 milliseconds, and that longer P-wave durations did not further increase the magnitude of the association. Partial and advanced IAB morphology were also associated with AF; however, the presence of advanced IAB did not increase the magnitude of the association with AF beyond that of partial IAB or P-wave duration between 110 and 119 milliseconds.

Several studies have identified variables associated with a higher risk of AF. Most of these studies have focused on clinical variables, such as age, hypertension, diabetes, cardiac morbidity, obesity, or obstructive sleep apnea.12,13,32 Few studies have also considered electrocardiographic characteristics, such as signs of left ventricular hypertrophy or P-R interval duration.33 In the last decade, P-wave characteristics have attracted the attention of several groups.34–36

The Framingham Heart Study (FHS),17,18 the Copenhagen ECG study19 and the Atherosclerosis Risk in Communities study (ARIC)18,20 have analyzed the association between several P-wave indices and the risk of AF. In all 3 studies, a statistically significant relationship between P-wave duration and the risk of AF was observed. The ARIC and the FHS studies compared the risk of AF in those individuals with a P-wave duration ≥ 120 milliseconds with those below this cutoff and reported a significant association (meta-analyzed hazard ratio [HR], 1.55; 95%CI, 1.29-1.85).18 In the Copenhagen ECG Study, P-wave duration was stratified in 5 groups, considering P-wave between 90 and 111 milliseconds as the reference group. The authors reported a HR for AF of 1.22 in the group with a P-wave of 112 to 119 milliseconds, a HR for AF of 1.50 in the group with a P-wave of 120 to 129 milliseconds and a HR for AF of 2.06 in the group with a P-wave ≥ 130 milliseconds (all of them statistically significant).19

Our data suggest that the magnitude of the association is higher and, in concordance with Nielsen et al.,19 appears at shorter P-wave duration, when the P-wave duration is greater than 110 milliseconds. The inclusion of individuals with P-wave duration between 110 and 119 milliseconds in the reference group could decrease the magnitude of the association reported in the FHS and the ARIC studies. There is an open discussion about the P-wave duration threshold to define interatrial block (100, 110 or 120 milliseconds). We believe that this definition should be based on the clinical implications of the P-wave duration rather than on its population distribution as has been proposed.37

The assessment of the pathological mechanisms explaining this association was beyond the scope of our study. However, previous reports have shown that P-wave duration is highly correlated with atrial fibrosis and fatty infiltration.38 Platonov et al.,39 reported that this fibrosis and fatty infiltration was more manifest in patients with permanent AF than in those with paroxysmal AF, in which the presence of abnormal electric activity within the pulmonary veins plays a major role. However, the presence of fibrosis in paroxysmal AF, at least in elderly patients, has also been documented.40,41 Interestingly, in our study we mainly identified permanent/persistent AF and missed most of the paroxysmal AF episodes. Therefore, the association between P-wave duration and AF in our study is based primarily on permanent/persistent AF and could be related to this fibro-fatty atrial infiltration.

Until recently, the diagnosis of IAB was not generally accepted, and both atrial enlargement and atrial blocks were considered nonspecific atrial abnormalities. Recent results from the ARIC study confirm that the presence of advanced IAB is associated with an increased AF risk (HR, 3.09; 95%CI, 2.51-3.79).20 In our study, the presence of A-IAB was also associated with AF, although this association was not statistically significant (OR, 2.09; 95%CI, 0.78-5.64). We also analyzed whether the combination of a long P-wave duration and the presence of advanced IAB increased the risk of AF compared with a P-wave duration ≥ 110 milliseconds. We did not observe an increased risk when both factors were present; therefore, the risk would be directly related to P-wave duration. However our study is hampered by low statistical power and these results should be interpreted with caution. De Bacquer et al.,15 have reported that the combination of long P-wave and morphological P-wave alterations (notching or deflection) significantly increased the risk of AF. Additional studies should analyze the combined effect of long P-wave and A-IAB on AF risk.

Recent studies have shown that the inclusion of some electrocardiographic markers (P-wave duration, or PR interval or left ventricular hypertrophy) did not improve predictive capacity, as measured by discrimination or reclassification, of the CHARGE-AF risk function.18,42 Further studies are warranted to assess whether the inclusion of P-wave duration or the presence of advanced IAB could improve the predictive capacity of this type of risk functions.

P-wave duration is associated with a higher risk of developing AF, beginning at 110 milliseconds. Advanced IAB morphology did not seem to provide an additional AF risk beyond that of P-wave duration, although the reduced number of participants with advanced IAB limits the conclusion about this P-wave morphology pattern. Further studies are required to validate this P-wave duration threshold and to analyze the added predictive value of this electrocardiographic marker.

This work was supported by the Spanish Ministry of Economy and Innovation through the Carlos III Health Institute [Red de Investigación Cardiovascular RD12/0042, FIS99/0655, FIS99/0013-01, FIS 99/9342, PI081327, MS12/03287], European Funds for Development, and by the Catalan Research and Technology Innovation Interdepartmental Commission [SGR 2014-1195]. I.R. Dégano was funded by the RECERCAIXA Program, Obra Social “La Caixa” [RE087465].

None declared.

.