topeeng.gif (8383 bytes)

[ Scientific Activity - Actividad Científica ] [ Brief Communications - Temas Libres ]

Major Role of P-wave Signal Averaged Electrocardiography in
Predicting the Occurrence of Atrial Fibrillation in
Patients Undergoing Open Heart Surgery

Caravelli Paolo, Musumeci Giuseppe, Gherarducci Gherardo,
Tartarini Giuseppe, Mariotti Rita, Bortolotti Umberto, Mariani Mario.

CardioToracic Department. University of Pisa
Via Paradisa, 2. Pisa, ITALY.

Abstract
Introduction
Material and Methods
Results
Discussion

Abstract
Aims: to identify the risk factors for the development of atrial fibrillation (AF) in patients (pts) after open heart surgery (OHS) by means of P-wave signal averaged electrocardiography (SAECG).
Methods and Results: We studied 60 pts (39 M, 21 F, mean age 65± 10 years), submitted either to coronary artery bypass grafting (CABG) or aortic valve replacement (AoVR). We performed a preoperative time domain analysis of the P-wave SAECG recording in order to measure filtered P wave duration (fPWD) and root mean square voltage of the last 10 and 20 ms of the atrial depolarization (RMS 10,20). Twentysix pts (44%), Group A (Gr.A), developed an AF episode 2,75±1,9 days after OHS. The remaining 34 pts composed the control group: Group B (Gr.B). No differences between the two groups were found respect to age, sex, echocardiographic parameters, total P wave duration on standard ECG, percentage of AoVR or CABG procedures, mean value of aortic gradient, severity of CAD, mean number/pt of CABG, aortic clamp time, cardiopulmonary by-pass time. Results showed that fPWD was significantly longer (139±11vs 112±12 ms: p<0,05) while RMS 10 and 20 were significantly smaller ( 2.9± 1.0 mV vs. 4.2± 1.1 mV, P =0.02; 4.3± 2.2 mV vs. 6.1± 2.0 mV, P < 0.05) in Gr.A than Gr.B. Assuming 135 ms as threshold value for fPWD , the onset of postoperative AF could be predicted with a sensitivity of 84.6%, a specificity of 73.5 %, a negative predictive value of 86.2%, and a positive predictive value of 70.9% (P < 0.01).
Conclusion: a fPWD shorter than 135 ms at SAECG is a powerful, accurate, and independent predictor of the AF occurrence after OHS.

Top

Introduction:

Atrial fibrillation (AF) is the commonest arrhythmia following open heart surgery, with an incidence between 10% and 40%, which puts patients (pts) in danger of complications. Though AF seldom represents a life-threatening condition, under certain circumstances, such as an impairment of left ventricular function or incomplete myocardial revascularization, it may cause haemodynamic instability, with consequent prolongation of hospital stay , or, even worse, embolic episodes. Therefore, it would be useful to define pre-, intra-, or post-operative parameters that can identify patients at risk of developing AF after open heart surgery (OHS). Recently, the attention has been focused on P wave high resolution electrocardiography (SAECG) for the study of atrial conduction and for risk-stratification concerning the development of episodes of AF. It has been shown that a prolonged P wave filtered with the Signal Averaging technique is associated with a history of paroxysmal AF in patients matched for age and heart disease. Previous reports also indicated that the study of the Signal Averaged P wave can predict the development of AF in patients undergoing cardiac surgery.
The aim of the present study is to further evaluate the role of signal averaged P-wave duration versus P-wave duration in the standard ECG, and clinical, demographic and echocardiographic variables in the prediction of AF after open heart surgery.

Top

Material and Methods: 

Sixthy consecutive pts (39 M and 21 F, mean age 65± 10 years) who had undergone open heart surgery at our Department were included in the present study. Thirty-two pts (53%) (28 M and 4 F, mean age 65± 6 years), had ischemic heart disease and underwent coronary artery bypass grafting. Twenty-eight pts (47%) (11 M and 17 F, mean age 64± 13 years), with aortic valve disease and normal coronary arteries, underwent valve replacement with a mechanical or biological prosthesis. Pts with history of paroxysmal AF, with arrhythmias at rest, and with thyroid dysfunction were excluded. Pts who were on antiarrhythmic drugs at time of operation or who had suspended antiarrhythmic drugs since less than 5 drug half-lives were also excluded. P wave duration was measured in ms on the bipolar lead I, II and III of the surface ECG, with a paper speed of 50 mm/s (sPWD). The total P-wave duration was defined as the time interval between earliest onset of the P-wave in any of the I, II and III leads on the latest offset in any of the I,II and III leads. All patients underwent coronary catheterization and M-mode, 2D, and Doppler transthoracic and transesophageal echocardiographic evaluation. Pts with mitral stenosis or mitral regurgitation greater than +/+++ were excluded as well. The day before surgery, all patients underwent Signal Averaged P wave analysis with a commecrcially avaiable instrument provided with specific software. Recordings were taken with the patient supine, in a silent well-heated environment, in order to minimize the artifacts due to muscular contraction. The 3 orthogonal bipolar leads X,Y, Z were used, with the QRS complex as trigger. The averaging procedure, with a bi-directional 30-250 Hz filter, was protracted until at least 200 beats were collected and the background noise was less than 0.1 mV. In all pts the filtered P wave duration (fPWD, ms) and the mean squared voltage of the last 10, 20 and 30 ms of the P wave (RMS10, RMS20, RMS30, mV) were assessed. The operative variables examined were: aortic cross-clamp time (Ao-CT, min), cardiopulmonary bypass time (CPB-T, min), and the number and type (arterial or venous) of conduits used for myocardial revascularization. Duration of hospital stay was also recorded. All pts were under continuous ECG monitoring during hospital stay. The episodes of AF lasting more than 15 minutes, both spontaneously and pharmacologically ceased, were considered relevant. Plasma electrolyte concentrations were repeatedly assessed during hospitalization. Data are reported as mean ± standard deviation. Statistical analysis was performed with Student t test and Pearson c2 test. P values less than 0.05 were considered significant.

Top

Results:

There were neither intraoperative nor hospital deaths. Episodes of AF, as previously defined, were observed in 26 pts (43%). The prevalence of AF in the two groups of pts suffering from ischemic heart disease and aortic valve disease was 44% and 43%, respectively. The first episode of AF appeared after a mean of 2.8±1.9 days after operation, with a mean of 1.8±0.4 episodes per patient and a mean duration of 108±69 minutes. Sinus rhythm spontaneously recovered only in 2 pts (6.7%), while AF persisted in spite of antiarrhythmic therapy in other 2 pts who had undergone aortic valve replacement and were discharged from hospital after 20 days with good haemodynamic performance and normofrequent ventricular response. In the remaining 22 pts, sinus rhythm was restored with drug therapy. The study population was grouped according to the occurrence of AF: Group A: 26 pts (19 M, 7 F; mean age 66±12 years) who experienced episodes of AF after operation and Group B: 34 pts (23 M, 11 F; mean age 64±7 years) who remained in sinus rhythm. The 2 groups did not significantly differ as for as sex, age, body weight, systolic and diastolic blood pressure, heart rate at rest and history of cardiovascular disease (Tab. 1). No differences between the 2 groups were found relative to total P wave duration on standard ECG (Gr. A 126±2 vs Gr. B 123±3 ms). The echocardiographic parameters considered were similar as well; ejection fraction (Gr. A 50±4% vs Gr. B 53±4%), the left atrial diameter from the parasternal long axis view was almost the same in the 2 groups (Gr. A 39±5 mm vs. Gr. B 37±7 mm). Aortic transvalvular gradient was not significantly different between pts suffering from aortic valve disease in the 2 Groups (Fig. 1). No significant differences between the 2 groups were detected concerning severity of coronary artery disease and number and type of conduits employed for myocardial revascularization. In addition, all pts showed similar values of plasma electrolytes on the day of onset of AF. Aortic cross-clamp time and cardiopulmonary bypass time were longer in Gr. A, but the difference was not statistically significant (Ao-CT: 68± 36 vs. 56± 36 minutes, P = n.s.; CPB-T: 102± 52 vs. 76± 21 min., P = n.s.). Signal Averaged P wave analysis showed significant differences between the 2 groups: filtered P wave duration (fPWD) was significantly longer in Gr. A than in Gr. B (144± 11 ms vs. 117± 12 ms, P < 0.05), while the mean squared voltage of the last 10 and 20 ms of the P wave was significantly lower in Gr. A than in Gr. B (RMS10: 2.9± 1.0 mV vs. 4.2± 1.1 mV, P =0.02; RMS20: 4.3± 2.2 mV vs. 6.1± 2.0 mV, P < 0.05) (Fig. 2). Assuming 135 ms as threshold value for fPWD, the onset of postoperative AF could be predicted with a sensitivity of 84.6%, a specificity of 73.5 %, a negative predictive value of 86.2%, and a positive predictive value of 70.9% (P < 0.01) (Tab. 2). No differences were detected considering P wave duration on the standard ECG. The occurrence of AF was well tolerated hemodynamically, but determined a longer hospital stay: 11± 4 days in Gr. A vs. 8± 1 days in Gr. B (P<0.05).

Table 1

table1.gif (8546 bytes)
table1a.gif (9547 bytes)

 

fig1.gif (5076 bytes)

Fig.1: The echocardiographic parameters considered were similar as well; ejection fraction and the left atrial diameter from the parasternal long axis view was almost the same in the 2 groups. Aortic transvalvular gradient was not significantly different between patients suffering from aortic valve disease in the 2 Groups

 

fig2.gif (5960 bytes)

Fig. 2: Filtered P wave duration (fPWD) was significantly higher in Group A than in Group B, while the mean squared voltage of the last 10 and 20 ms of the P wave was significantly lower in Group A than in Group B. No differences between the two groups were found respect to total P wave duration on standard ECG (sPWD).

Table 2

table2.gif (5338 bytes)

Top

Discussion: 

The incidence of episodes of AF in pts undergoing coronary artery bypass grafting or aortic valve replacement has been reported in the range of 5% to 40%. In the present study, we observed an incidence of 43%, with most first episodes occurring within 72 hours from operation and lasting less than 48 hours, in agreement with the most recent reports. The analysis of clinical and instrumental data taken either preoperatively, or intra-operatively, or postoperatively, showed that only Signal Averaged P wave analysis (fPWD, RMS10, and RMS20) could predict the occurrence of episodes of AF after OHS. The assessment of P wave duration on the standard ECG did not show any significant difference between ischemic and aortic pts, thus confirming the higher predictive value of SAECG. Though the occurrence of AF did not significantly impair the haemodynamic performance in our pts, it determined a longer hospital stay (P<0.05), as observed by others. Our results are consistent with those of other studies, such as Steinberg’s et al., Klein’s et al., Stafford’s et al. These studies differ from our study because the authors chose a cut-point for fPWD different from the one identified in our population (Table 3). The threshold value of 135 ms allowed the highest sensitivity, specificity, negative and positive predictive value in the present population. The choice of a lower cut-point compared with other studies minimized the number of false negatives, thus providing a very high sensitivity (84.6%) and negative predictive value (86.2%).
In conclusion, the present study documenting that a fPWD duration shorter than 135 ms at SAECG is a powerful, accurate, and independent predictor of the occurrence of AF after cardiac surgery, indicates that fPWD is the only predictor of AF after OHS. Therefore, it is suggested to preoperatively perform SAECG in patients undergoing heart surgery in order to identify pts at risk and monitor them more carefully, both clinically and electrocardiographically, justifying to take into consideration a possible prophylactic drug therapy, as put forward by previous papers.

Table 3

table3.gif (7783 bytes)

 

Questions, contributions and commentaries to the Authors: send an e-mail message (up to 15 lines, without attachments) to arritmias@listserv.rediris.es , written either in English, Spanish, or Portuguese.

Top


© CETIFAC
Bioengineering
UNER

Update
Nov/12/1999