Introduction
AF is the most important risk factor for stroke; it has a deleterious effect on longevity with a doubling of all cause mortality [1]. The estimation of the probability of recurrent AF by using a simple parameter might guide the clinician in the management of these patients. P wave dispersion constitutes a recent contribution to non-invasive electro cardiology. This electrocardiographic measurement reflects a disparity in atrial conduction [2]. The objective of this study is to assess the value of P wave dispersion as a predictor of AF recurrence in patients with and without structural heart disease to detect and try to trace which group of patients is susceptible to recurrence of AF (the high risk group), in an attempt to follow and try to prevent the occurrence of serious complications among them.
Methods
Sixty two patients who had AF and successfully converted to sinus rhythm were included. They were followed up for recurrence of AF for six months. Any patient with failed cardioversion, received antiarrhythmic agents 4 weeks before and after cardio-version, severe hypertension, ACS, heart failure, severe pulmonary disease, pulmonary embolism, sick sinus syndrome or patients with open heart surgery within 3 months were excluded from the study. All patients were subjected to the following: thorough history, standard 12 lead ECG for PWD calculation, Echo-doppler to measure LA diameter, Left ventricular systolic (EF%) and diastolic (E, A, E/A ratio) functions.
Results
According to AF recurrence, the patients were classified into: Group I included 36 patients with recurred AF and Group II included 26 patients with preserved sinus rhythm (PSR). Maximum P wave duration was statistically significant longer in group I than that in group II (P<0.04) and at a cutoff point of >110ms sensitivity was 88.9% and specificity of 73.1% for patients with AF recurrence. PWD was of highly statistically significant values in group I (71±21ms) than those in group II(40±15ms) (P<0.000) with sensitivity of 75% and specificity of 88.5% at cutoff point >80.5ms. Statistically significant left ventricular diastolic dysfunction (in the form of impaired and pseudonormal relaxation) and increased left atrial diameter were more obvious in group I than those in group II(P<0.000&P<0.007 respectively). On the other hand, there was insignificant decrease in EF% in group I than in group II. Logistic regression analysis for P max, PWD, LAD and EF% revealed that PWD is independent predictor for AF recurrence (r=0.585, P<0.000). Receiver operator curve for AF recurrence group showed that PWD had the largest area under the curve (AUC=0.975) than those of P max (AUC=0.885) or LAD (AUC=0.955).
Discussion
Recurrence of atrial fibrillation after electrical cardioversion of chronic AF is not uncommon, where up to 57% of AF recurrence after cardioversion take place during the first 30 days after direct current shock therapy. However, it remains unclear which parameter (s) predict clinical recurrence [3].There is no universally accepted value that defines a prolonged PWD as calculated by 12- lead electrocardiography [4]; however, our results would point toward employing a higher cutoff to more accurately identify patients at risk for recurrent AF. We conclude that PWD, which is a non-invasive and simple parameter, could be applicable to predict the high risk group for AF recurrence after cardioversion, but a larger study is needed to consolidate these results.
Figure (1):
Interactive plot after ROC curve for The PWD as regarding AF recurrence
Figure (2):
Comparison ROC curves for PWD, LAD and P max
Table (1):
Patients’ characteristics in both groups
| Item no. (%) | AF recurrence | P value | ||
|---|---|---|---|---|
| Recurred N=36 | PSR N=26 | |||
| 1-Age (mean±SD) | 55±13 | 50±13 | NS | |
| 2-Sex | Male | 21 (58.3%) | 14 (53.8%) | NS |
| Female | 15 (41.7%) | 12 (46.2%) | ||
| 3-Hypertension | Hypertensive | 21 (58.3%) | 13 (50.0%) | NS |
| Non-hypertensive | 15 (41.7%) | 13 (50.0%) | ||
| 4-DM | Diabetic | 15 (41.7%) | 10 (38.5%) | NS |
| Non-diabetic | 21 (58.3%) | 16 (61.5%) | ||
| 5-Heart diseases | CAD | 30 (61.6%) | 22 (84.6%) | NS |
| VHD | 3 (19.2%) | 1 (3.8%) | ||
| Lone AF | 3 (19.2%) | 3(11.5%) | ||
PSR =Preserved sinus rhythm, CAD =Coronary artery disease, VHD = Valvular heart diseases, NS = Non-significant
Table (2).
The ECG and echocardiographic findings in both groups
| Item Mean±SD Or no. (%) |
AF recurrence | P value | |
|---|---|---|---|
| Recurred N=36 | PSR N=26 | ||
| 1-Max P duration (ms) | 116.33±10.00 | 105.46±8.45 | 0.04* |
| 2-PWD (ms) | 71±21 | 40±15 | 0.000*** |
| 3-LAD (cm) | 4.40±.91 | 4.00±.87 | 0.007** |
| 4-EF (%) | 44.83±10.89 | 46.58±8.64 | NS |
| 5-Diastolic function | |||
| NF | 5 (13.9%) | 16 (61.5%) | 0.000*** |
| IR | 15 (41.7%) | 3 (11.5%) | 0.000*** |
| PNR | 16 (44.4%) | 7 (26.9%) | 0.000*** |
PSR= Preserves sinus rhythm, PWD=P wave dispersion, LAD= Left atrial diameter, NF= Normal function, IR= Impaired relaxation, PNR= Pseudonorn relaxation
* = significant,
** = highly significant,
*** = Very highly significant
Table (3):
The logistic regression analysis for predictors of AF recurrence.
| Item | R- square | P - value |
|---|---|---|
| PWD | .585 | .000 |
| PMAX | .056 | .755 |
| LAD | .570 | .487 |
| EF | .046 | .479 |
References
- 1.Poli D. Age, CHADS2 define bleeding risk in anticoagulated AF patients. Thromb Res. 2007;121:347–52. doi: 10.1016/j.thromres.2007.05.014. [DOI] [PubMed] [Google Scholar]
- 2.Turhan H, Kose H, Celik T, et al. Atrial fibrillation recurrence after cardioversion: is there is a simple electrocardiographic parameter to predict it? Int J of cardiology. 2006;113(3):435–436. doi: 10.1016/j.ijcard.2005.10.023. [DOI] [PubMed] [Google Scholar]
- 3.Boriani G, Diemberger I, Biffi M, et al. P wave dispersion and short-term vs. late atrial fibrillation recurrences after cardioversion. Int J Cardiol. 2005;101:355–361. doi: 10.1016/j.ijcard.2004.03.039. [DOI] [PubMed] [Google Scholar]
- 4.Boriani G, Diemberger I, Biffi M, et al. PWD&short-term vs. late AF recurrences after DC. Int J Cardiol. 2005;101:355–361. doi: 10.1016/j.ijcard.2004.03.039. [DOI] [PubMed] [Google Scholar]