Discussion:
The duration of the P wave is a result of the conduction velocity in the atrial working myocardium and the distance the electrical activation has to travel. Functional and structural changes in the atrial muscle affect both of these parameters - they slow down conduction velocity and extend the way to travel due to the enlargement of the atria. This results in an extension of the P wave duration and a change in its morphology. The same changes in structure and function are responsible for generating atrial arrhythmias including atrial fibrillation [8]. Additionally, the ongoing arrhythmia leads to progression of the above mentioned changes. In particular the enlargement of left atrium is clearly associated with an increase in filling pressure of the left ventricle and probably dependent on the ventricular rate [12]. Data on the effect of AF on muscle conduction are more scarce and equivocal [13].
The main result of our study is to show a longer duration of the P wave in patients with average long-term persistent atrial fibrillation compared to patients with paroxysmal arrhythmia and sinus rhythm at the time of the study. The degree of this elongation (about 10%) of the P wave duration is not only statistically but also clinically significant. It should be emphasized that the P waves measured by us in both groups markedly exceed the normal values of 120 ms (144.6 vs 159.9 ms, respectively). Similar comparisons are not numerous in the available literature [14, 15]. As our subgroups of patients suffering from paroxysmal and persistent AF are comparable according to age, gender distribution, comorbidities and anti-arrhythmic medication, it indicates that this additional prolongation is caused only by the presence of the prolonged episodes of arrhythmia. Our measurements were not performed immediately after the cardioversion shock, so the influence of direct current can be skipped. Additionally the direct current flow during cardioversion had little or no effect on the P wave duration in one small study immediately after the procedure and on the next day [16].
Even if diabetic patients were not frequently presented in our studied group there was a noticeable difference between diabetics and non-diabetics in terms of P wave duration. Those with DM were observed to have a longer P wave than participants without it. This is in line with the other clinical observations even the direct evidence lacks in human. Diabetes is presumed to be a risk factor for atrial fibrillation and the topic has been reviewed quite often. A meta-analysis of different cohort and case control studies investigating the correlation of DM and AF, showed that individuals with DM had a 40% greater risk of AF compared to unaffected individuals [17]. There is only sparse literature to be found about DM leading to electrical changes of the atrial substrate [18]. In an experimental setting of DM it was associated with increased atrial fibrosis, interatrial conduction delay and greater inducibility of AF [19]. Another animal study confirmed those results with additional interesting observations of either P wave prolongation in diabetic rats without left atrial enlargement, for which the authors accounted diabetic changes in the gap junction protein Cx [20]. Similar outcomes were obtained in patients with impaired fasting glucose leading to significantly prolonged interatrial conduction times and consecutive decrease left atrial emptying volume and fraction [21].
In another subgroup of our patients the chronic kidney disease was found to be a predictor of longer P wave duration. In the literature some researches made already the association between maximum P wave duration and exacerbation of the renal condition until the defined end points of hemodialysis, death or a specified decline in estimated glomerular filtration rate [22. 23]. Based on our results it could be assumed, that a vice versa influence of CKD on the P wave duration is occurring as well, possibly because of a simple fluid overload. Referring to the patients included in our study, only a small number of 9 (7.6%) patients presented with CKD as a comorbidity but it was discovered to be a statistically significantly related to the P wave duration. This needs further investigation in other studies, not distorted by the small number of CKD patients. Atrial fibrillation is frequently described together with a renal dysfunction but mainly as a preceding comorbidity but no relation was found for CKD being the reason of AF. Nevertheless our results indicate such possibility making the subject worth to be studied.
The anti-arrhythmic medication influences the electrophysiological properties on the working myocardium, in particular the conduction speed and refractory period which could influence the P wave duration. The results of our study do not support such concept. Amiodarone is a class III antiarrhythmic agent acting mainly as potassium channel blocker, characterized by prolongation of the refractory period and atrial repolarization. It has been shown to be effective in maintaining sinus rhythm and preventing arrhythmia episodes in patients with paroxysmal atrial fibrillations. Even if in one small study researchers described the amiodarone-related increase in P wave duration, this was a small experimental animal study and the conditions were not comparable in sinus rhythm in human, present in our study [24]. The relationship between P wave duration and amiodarone administration was similarly negated in a study conducted by Sasaki et al. [25].
In contrast to amiodarone the treatment with sodium channel blocker could theoretically influence the P wave duration. Propafenone is an IC class agent which blocs the fast sodium channels, slowing down the conduction velocity in the working myocardium. According to literature data there is no direct relationship between the dose of propafenone and the duration of the P wave, however the same study confirms a weak correlation between the treatment with propafenone and the elongation of the P wave duration [26]. Our data do not confirm this finding. One should however emphasize that our propafenone treated patients’ group was not large.
Based on our results the theoretical model resulting from ROC curves indicated the estimated P wave duration differentiating patients between sinus rhythm and persistent atrial fibrillation groups. In the literature, this approach has not been presented so far, so our value of P wave duration - 148 ms can only be referred to studies indicating the importance of this parameter in the prediction of sinus rhythm maintenance after electrical cardioversion. In 1999 Aytemir and co-authors investigated the P wave signal-averaged ECG in 73 patients after successful cardioversion. During 6 months follow-up period the recurrence of AF was observed in 31 patients and in 42 patients sinus rhythm was maintained. The researchers found no difference between the groups according to gender, age, presence of organic heart disease, left atrial diameter, left ventricular ejection fraction, use of antiarrhythmic drug, and duration of atrial fibrillation. The filtered P wave duration was statistically longer in patients with recurrence of atrial fibrillation 138.4 vs. 112.5 ms. A filtered P wave duration of 128 ms was had a sensitivity of 70% and specificity of 76% for the detection of recurrence of atrial fibrillation [27]. On the other hand in the study of Perzanowski et al. the maximum duration of the P wave did not differentiate patients who remained in sinus rhythm or experienced a recurrence of arrhythmia (142 vs 145 ms; p=n.s.) [28]. As the authors did not mention the methodology of P wave duration measurements it should be assumed that they used simple standard 12 lead ECG without any more precise equipment. This lack of precision could be the cause of their results. In the study of Gonna and co-workers a 12-lead ECG was recorded after electrical cardioversion for persistent AF in 77 patients and repeated after 1 month. Compared with the sinus rhythm group, the one with recurrent AF had more patients with P wave duration exceeding 142 ms. Using a cutoff of <142 ms for P wave duration the authors showed a sensitivity of 64.6% and specificity of 62.1% for sinus rhythm maintenance. In multiple regression analysis the P wave duration longer than 142 ms was the only independent predictor of AF recurrence [29]. The above-mentioned considerations indicate unequivocally that the prolongation of the P wave is clearly a risk factor for paroxysm of atrial fibrillation and more advanced stages of the arrhythmia, which is in line in our results. Moreover in different settings we produced the evidence which supports the previous findings. According to the higher values of the P wave duration obtained by us, it should be remembered that the precise methodology used in our study is qualitatively different from that of other researchers [28, 29]. This is a reason that already a few years ago we confirmed the lack of P wave dispersion, assessed in some of the above papers, which is related to the inaccuracy of the measurement [30].
In summary the ongoing atrial fibrillation in form of moderately long persistent arrhythmia influences negatively the structural and functional atrial remodeling. This occurs independently from age and gender, sort of anti-arrhythmic treatment but can be slightly related to some comorbidities.