Author + information
- Received December 24, 1996
- Revision received April 7, 1997
- Accepted April 21, 1997
- Published online August 1, 1997.
- Kishore J Harjai, MDA,
- Sameh K Mobarek, MDA,
- Jorge Cheirif, MD, FACCA,1,1,
- Louis-Marie Boulos, MDA,
- Joseph P Murgo, MD, FACCA and
- Freddy Abi-Samra, MDA,*
- ↵*Dr. Freddy Abi-Samra, Ochsner Clinic, 1514 Jefferson Highway, New Orleans, Louisiana 70121.
Objectives. We sought to evaluate the effect of clinical factors on recovery of atrial function after cardioversion for atrial fibrillation.
Background. Lack of effective mechanical atrial function (EMAF) after cardioversion of atrial fibrillation predisposes to thromboembolic complications and delays improvement in functional capacity.
Methods. Fifty-two patients underwent cardioversion (group I, electrical cardioversion, n = 40; group II, pharmacologic or spontaneous cardioversion, n = 12) for atrial fibrillation. Serial transmitral inflow Doppler variables were recorded after cardioversion until EMAF (atrial filling velocity >0.50 m/s) was seen. Clinical variables (age, duration of atrial fibrillation, left ventricular ejection fraction, left atrial diameter, underlying cardiovascular disease, antiarrhythmic drug therapy and mode of cardioversion) were tested for an association with the outcomes of recovery of atrial function by day 3 and day 7.
Results. Effective mechanical atrial function recovered in 68% of patients by day 3 and in 76% by day 7 after cardioversion. The mode of cardioversion was significantly associated with recovery of atrial function by day 3 in bivariate and multivariate analyses (odds ratio 0.12, 95% confidence interval 0.01 to 1.0, for electrical cardioversion). None of the variables had an association with recovery of atrial function by day 7. Group I patients took a longer time to recover atrial function than group II patients (p = 0.012). In addition, group I patients had a significantly lower peak atrial filling velocity (mean [±SD] 0.39 ± 0.19 m/s vs. 0.56 ± 0.16 m/s) and a higher early filling to atrial filling velocity ratio (2.5 ± 1.2 vs. 1.5 ± 0.5) after cardioversion.
Conclusions. A high proportion of patients recover EMAF within 1 week after cardioversion. Patients who undergo electrical cardioversion display a greater degree and a longer duration of mechanical atrial dysfunction than those who convert pharmacologically or spontaneously.
Atrial mechanical dysfunction after cardioversion for atrial fibrillation has received considerable attention in the last few years . Its significance is corroborated by the occurrence of thromboembolic events after cardioversion, despite exclusion of left atrial appendage thrombus by transesophageal echocardiography before cardioversion [2–4]. These thromboembolic events, which occur in the absence of preexisting left atrial appendage clot, are believed to be causally related to atrial dysfunction after cardioversion, which predisposes to thrombus formation . Because normal atrial contraction is believed to lag behind conversion to sinus rhythm by up to a few weeks , patients undergoing cardioversion are often anticoagulated for a prolonged period.
Restoration of normal atrial contraction may be related to several clinical factors, such as patient age, mode of cardioversion, left atrial size, left ventricular function, associated cardiovascular disease and myocardial depressant effect of prophylactic antiarrhythmic therapy . The effect of these clinical variables on recovery of effective mechanical atrial function (EMAF) has not been adequately assessed. This study was conducted to examine the role of clinical variables in recovery of EMAF. We hypothesized that electrical cardioversion would be associated with a longer duration and greater degree of left atrial mechanical dysfunction than pharmacologic or spontaneous cardioversion.
1.1 Study patients.
Hemodynamically stable patients referred for cardioversion for atrial fibrillation between May 1995 and March 1996 were considered for inclusion in this study. No exclusions were made based on age, gender, concomitant disease states, ventricular function, cause or duration of atrial fibrillation or medical therapy. Patients who had any documented evidence of atrial flutter on the electrocardiogram (ECG) and those who were unable to follow up for logistic reasons were not included. Of 54 patients considered for participation in the study, two were excluded because of failure of cardioversion (n = 1) or predominant atrial flutter on the ECG (n = 1). Thus, the study group consisted of 52 patients successfully cardioverted to sinus rhythm. Patients were classified as group I if they underwent electrical cardioversion (n = 40) and group II if they cardioverted with medications (n = 10) or spontaneously (n = 2).
The baseline characteristics of the study group are shown in Table 1. Underlying cardiovascular diseases were present in 85% of the patients and included one or more of the following: hypertension (58%), ischemic or nonischemic dilated cardiomyopathy (21%), coronary artery disease (17%), significant valvular disease (8%), hypertrophic cardiomyopathy (4%) or pericardial disease (4%). In eight patients (15%), no structural cardiovascular disease was identified. A specific precipitating factor was identified in only one patient (2%) who had atrial fibrillation after cardiac surgery. At the time of enrollment, patients were on the following medications for ventricular rate control or antiarrhythmic effect: digoxin (n = 17, 33%), beta-adrenergic blocking agents (n = 31, 60%), calcium channel blocking agents (verapamil or diltiazem, n = 10, 19%), amiodarone (n = 10, 19%), flecainide (n = 4, 8%) or propafenone (n = 1, 2%).
The duration of atrial fibrillation was <28 days in 14 patients; all of these patients presented with their first episode of atrial fibrillation. In the other 38 patients, the duration of atrial fibrillation was ≥28 days (28 days to 3 months in 12 patients [32%], >3 to 6 months in 11 [29%], >6 to 12 months in 9 [24%], >12 to 24 months in 3 [8%], >24 months in 1 [3%] and unknown in 2 [5%]).
The protocol was approved by the Institutional Review Board of the Alton Ochsner Medical Foundation. All patients provided written informed consent. All procedures were conducted in accordance with institutional guidelines.
The decision to perform electrical or pharmacologic cardioversion was made on the basis of the clinical situation, in consultation with the referring cardiologist, and was independent of enrollment into the study protocol. For electrical cardioversion, fasting patients received intravenous sedation with methohexital sodium (total dose 0.5 to 1.5 mg/kg body weight). A Physio-Control (LIFEPAK 8) direct current cardioverter was used in all patients. Thirty-two patients (80%) cardioverted with a single electrical shock; multiple shocks were required in the other eight patients. The total amount of electrical energy used for cardioversion was 379 ± 285 J per patient.
For pharmacologic cardioversion, patients were hospitalized and monitored in a telemetry setting. Telemetry monitoring technicians notified the investigators as soon as cardioversion to normal sinus rhythm occurred, and they obtained ECG rhythm strips for confirmation of normal sinus rhythm. Cardioversion occurred during treatment with the following medications: oral sotalol (n = 3), oral quinidine (n = 1), intravenous procainamide (n = 2), intravenous metoprolol (n = 3) or intravenous diltiazem (n = 1). In two patients, cardioversion to sinus rhythm occurred before any medications were instituted.
1.3 Echocardiographic studies.
Transthoracic two-dimensional imaging and pulsed wave Doppler echocardiographic studies were obtained in all patients using a Hewlett-Packard Sonos 1000 or 1500 ultrasound machine equipped with 2.5- and 3.5-MHz phased array transducers. Transmitral inflow velocities were recorded from the apical four-chamber view with the sample volume positioned between the tips of the mitral valve leaflets. To ensure standardization, all studies were performed in the morning hours with the patient in the supine position after 10 min of rest. To minimize technical sampling errors, we attempted to have the same sonographer perform studies whenever possible. Studies were recorded on S-VHS videotape for later review by one of the investigators who had no knowledge of the patient’s clinical background. An average of 5 to 7 beats was studied with pulsed wave recording of the transmitral flow velocities. Peak velocities and time velocity integrals of early filling (E) and atrial filling (A) waves were determined on each study.
Assessment of mitral inflow variables was performed on the day of cardioversion (within 6 h after cardioversion) and on days 1, 3 and 7. In each patient, echocardiographic follow-up was continued until such time as one of the following end points was reached: recovery of EMAF (i.e., atrial filling wave velocity ≥0.5 m/s [n = 37, 71%]) , recurrence of atrial fibrillation (n = 5, 10%) or completion of the study protocol 7 days after cardioversion (n = 8, 15%). Two patients (4%) were lost to follow-up.
Left ventricular ejection fraction was estimated from the parasternal long- and short-axis views and apical four- and two-chamber views. Left atrial dimension was measured at end-systole in the parasternal long-axis view.
1.4 Effective mechanical atrial function.
This was assessed by transthoracic Doppler echocardiography in accordance with Manning et al. and defined as the presence of atrial filling waves with a peak velocity ≥0.5 m/s. Thus, by definition, the end point of recovery of EMAF is a composite of persistence of normal sinus rhythm and achievement of adequate atrial filling wave velocity after cardioversion. In contrast, failure to recover EMAF represents either a lack of adequate A wave velocity with normal sinus rhythm, or recurrence of atrial fibrillation, which precludes satisfactory atrial filling waves.
1.5 Independent variables and outcomes.
Independent clinical variables tested for an association with recovery of EMAF are age, duration of atrial fibrillation (<28 days or ≥28 days), left ventricular ejection fraction (<50% or ≥50%), mode of cardioversion (group I, electrical cardioversion; or group II, pharmacologic or spontaneous cardioversion), left atrial diameter, use of antiarrhythmic medications (including beta-blockers and calcium channel blockers) and presence of associated cardiovascular disease, including hypertension. Outcomes were prospectively defined and included recovery of EMAF by day 3 and recovery of EMAF by day 7.
1.6 Statistical analysis.
Comparisons between continuous and categoric variables were performed using the unpaired ttest. For comparisons of categoric variables, frequency tables and chi-square analyses were used. To evaluate the effect of clinically relevant variables on the outcomes of recovery of EMAF by day 3 and recovery of EMAF by day 7 after cardioversion, bivariate analysis was initially performed using logistic regression. Odds ratios and confidence intervals were calculated for categoric variables. To assess the independent effect of variables on the outcomes, backward stepwise regression was used. Life-table analysis was used to assess the difference between group I and group II with respect to time to recovery of effective atrial mechanical contraction.
All analyses were performed using the BMDP version 7.0 software. Data are expressed as mean value ± SD for continuous variables, and as percentages for categoric variables. Statistical significance was established at p < 0.05.
2.1 Clinical characteristics.
Patients in groups I and II were similar with respect to age, duration of atrial fibrillation, presence of coexisting cardiovascular diseases, left ventricular ejection fraction, left atrial diameter and use of antiarrhythmic medications at the time of enrollment (Table 1).
2.2 Recovery of EMAF after cardioversion.
Using life-table analysis, earlier recovery of EMAF was seen in patients who cardioverted pharmacologically or spontaneously compared with those who underwent electrical cardioversion (p = 0.01) (Fig. 1). Between the two groups, the difference in the proportion of patients with recovery of EMAF is more pronounced up to day 3 and begins to taper off on day 7. Fifty-two percent of all patients recover EMAF by day 1, 68% by day 3 and 76% by day 7.
2.3 Clinical variables affecting recovery of atrial function by day 3.
Thirty-four (68%) of the 50 patients who completed follow-up recovered EMAF by day 3. This included 23 (61%) of 38 patients from group I and 11 (92%) of 12 patients from group II. On bivariate analyses, age, duration of atrial fibrillation, left ventricular ejection fraction, presence of cardiovascular disease, use of antiarrhythmic therapy or left atrial diameter did not show any significant association with recovery of EMAF by day 3. The mode of cardioversion was significantly associated with this outcome; thus, patients who underwent electrical cardioversion had an odds ratio of 0.12 for recovery of EMAF by day 3, indicating that electrical cardioversion, compared with pharmacologic or spontaneous cardioversion, was associated with a lower probability of achievement of EMAF by day 3. Odds ratios and confidence intervals were not significant for any of the other categoric variables (Table 2). On multivariate analysis, the mode of cardioversion remained significantly associated with recovery of EMAF by day 3, when adjusted for age, presence of cardiovascular disease or use of antiarrhythmic therapy. However, when adjusted for duration of atrial fibrillation, left ventricular ejection fraction or left atrial diameter, the association between mode of cardioversion and recovery of EMAF by day 3 was no longer statistically significant.
2.4 Clinical variables affecting recovery of EMAF by day 7.
By day 7, 76% of all patients recovered EMAF. This group included 27 (71%) of the 38 electrically cardioverted patients who completed the study and 11 (92%) of 12 pharmacologically or spontaneously cardioverted patients. None of the independent variables used in this analysis had an association with the outcome of recovery of EMAF by day 7 on bivariate or multivariate analysis. The odds ratios and 95% confidence intervals for categoric variables are shown in Table 2.
2.5 Comparison of pulsed wave Doppler transmitral recordings after cardioversion.
A comparison between group I and group II with respect to mitral inflow variables obtained within 6 h after cardioversion is shown in Table 3. Group I patients who underwent electrical cardioversion showed a significantly lower peak atrial filling wave velocity (A), a lower time velocity integral of atrial filling wave and a higher ratio of peak early filling velocity to peak atrial filling velocity (E/A ratio) compared with group II. No significant difference was found in the peak early filling wave velocity (E) or the time velocity integral of the early filling wave between the groups.
Four main findings emerge from our study.First, a large proportion of patients who undergo cardioversion for atrial fibrillation recovered EMAF early after cardioversion. Approximately one in five patients recovered EMAF within 6 h after cardioversion. More than half of all patients recovered EMAF by the first day after cardioversion and more than three-fourths recovered EMAF by 1 week after cardioversion. Furthermore, 11 (92%) of 12 patients who underwent pharmacologic or spontaneous cardioversion recovered EMAF by day 3. Second, of all the clinical variables tested on bivariate or multivariate analysis, only the mode of cardioversion was seen to have any impact on recovery of atrial function. Thus, patients who underwent electrical cardioversion were significantly less likely to recover EMAF by the third day after cardioversion (odds ratio 0.12) compared with those who cardioverted pharmacologically or spontaneously. This effect remains significant when adjusted for patient age, presence of underlying cardiovascular disease or use of antiarrhythmic drug therapy; however, it is no longer significant when adjusted for left ventricular ejection fraction, left atrial diameter or duration of atrial fibrillation. Third, the effect of the mode of cardioversion on recovery of EMAF is not statistically significant at 1 week after cardioversion, implying that the possible detrimental effect of electrical current on atrial function is an early phenomenon and wears off between 3 and 7 days after cardioversion. Accordingly, among the two groups of patients, the greatest difference in the proportion of patients recovering EMAF is seen in the first 3 days after cardioversion (Fig. 1). Fourth, mitral inflow variables recorded within 6 h after cardioversion show a significantly lower peak atrial filling velocity, a lower time velocity integral of the atrial filling wave and a higher early filling to atrial filling peak velocity ratio (E/A ratio) in the electrically cardioverted patients, implying a greater degree of atrial dysfunction in these patients. This is consistent with the delayed recovery of atrial function in electrically cardioverted patients.
3.1 Clinical variables affecting atrial function after cardioversion.
Several investigators previously attempted to identify factors that affect the recovery of atrial function after cardioversion for atrial fibrillation. Manning et al. demonstrated that immediately after cardioversion and at 24 h and 1 week after cardioversion, atrial mechanical function is better in patients with atrial fibrillation <2 weeks in duration than in those with atrial fibrillation >6 weeks in duration. Although there was no difference in the left atrial diameter between the groups with atrial fibrillation of “brief” (<2 weeks), “moderate” (2 to 6 weeks) or “prolonged” (>6 weeks) duration, other patient characteristics such as age, left ventricular ejection fraction, underlying cardiac diseases, concomitant antiarrhythmic therapy and mode of cardioversion were not separately reported for the three groups. Any one or combination of these factors could conceivably influence recovery of EMAF independent of the duration of atrial fibrillation. We included all of these variables in our models using recovery of EMAF by day 3 and recovery of EMAF by day 7 after cardioversion as the outcome measures. Interestingly, duration of atrial fibrillation before cardioversion did not seem to affect the outcome significantly.
In 78 patients with atrial fibrillation who underwent cardioversion (electrical cardioversion in 60 patients, pharmacologic in 18), Abascal et al. found no significant difference in the peak early filling and peak atrial filling velocities at 24 h after cardioversion between the electrically and pharmacologically cardioverted patients. They concluded that left atrial dysfunction after cardioversion is independent of the mode of cardioversion. Their conclusion is in contrast to our findings, which suggest that patients who require electrical cardioversion seem to display greater atrial mechanical stunning. The difference between their and our findings may be related to the duration of atrial fibrillation in the two studies. All the patients in the study reported by Abascal et al. had atrial fibrillation >4 weeks, whereas 27% of the patients in our study had atrial fibrillation <4 weeks.
Mattioli et al. addressed the effect of underlying cardiovascular disease on recovery of effective atrial function in 60 patients with atrial fibrillation. Restoration of sinus rhythm occurred spontaneously in 20 patients and was accomplished pharmacologically in 40 patients. It was found that EMAF recovered earlier in patients without cardiovascular disease and in those with underlying hypertension only, compared with those patients who had ischemic cardiomyopathy. Duration of atrial fibrillation <24 h and normal left atrial size were also associated with early recovery of atrial function .
3.2 Relation between left atrial dysfunction and thromboembolic complications.
Thromboemboli are known to occur after cardioversion in nonanticoagulated patients despite “negative” transesophageal echocardiograms immediately before cardioversion [8, 9]. It is most likely that persistent or increased atrial stasis after cardioversion results in the formation of fresh, loosely adherent thrombi and subsequent embolism. Lack of mechanical activity contributes to stasis in the left atrium . In view of the excellent sensitivity and specificity of transesophageal echocardiography for detection of left atrial appendage clots , it is unlikely that these events are related to small, preexistent clots missed by the precardioversion transesophageal echocardiogram. The occurrence of embolism in up to 7% of patients undergoing cardioversion [8, 12, 13], the close temporal relation between the embolic event and the procedure and, oftentimes, the paucity of other potential sources of embolism [8, 9]greatly undermine the role of other sources of embolism. Thus, an increase in the thrombogenic milieu seems to occur immediately after cardioversion, and patients are predisposed to atrial thrombus formation, presumably until such time as “normal” atrial and atrial appendage function resumes.
3.3 Current recommendations for anticoagulation and clinical implications of our findings.
Anticoagulation for 2 to 4 weeks before cardioversion is currently recommended to allow for adherence and endothelialization or resolution of existing thrombi, as well as to prevent new clot formation. In addition, anticoagulation for up to 4 weeks after cardioversion is recommended to allow for late resumption of atrial activity [1, 8]. Recent studies have proposed the use of transesophageal echocardiography to screen patients for atrial thrombi before cardioversion [14, 15]to reduce the requirement for prolonged anticoagulation before cardioversion. It is believed that in patients who have no demonstrable left atrial thrombi on transesophageal echocardiography before cardioversion, long-term anticoagulation beforecardioversion may be deferred . Existing reports, however, clearly support the use of anticoagulation for a period aftercardioversion [8, 9].
To minimize the cost and potential morbidity associated with anticoagulation, future research should focus on identifying subsets of patients who do not need prolonged anticoagulation after cardioversion. Studies reporting on the duration of atrial mechanical dysfunction after cardioversion will be the basis for such trials. However, because most left atrial thrombi reside in the left atrial appendage , before any strategy of short-term postcardioversion anticoagulation relying on the recovery of atrial mechanical function is evaluated, it needs to be demonstrated that left atrial mechanical function is an accurate surrogate marker of left atrial appendage mechanical function. Although this may intuitively seem to be so, it has not yet been documented. If EMAF is indeed a surrogate marker of atrial appendage function, clinical rationale would indicate that the duration of anticoagulation be shortened in patients who show early recovery of EMAF. Clearly, these implications apply only to those patients who remain in sinus rhythm after cardioversion.
The findings of our study also have a potential impact on predicting improvement in exercise capacity after cardioversion in symptomatic patients with atrial fibrillation. It is well described that improvement in symptom-limited maximal oxygen consumption and anaerobic threshold is delayed and does not occur immediately after cardioversion . This has been ascribed to the delay in improvement in stroke volume that follows an increased atrial contribution to left ventricular filling. Our findings imply that patients undergoing electrical cardioversion may have delayed improvement in exercise capacity after cardioversion compared with those patients who cardiovert spontaneously or pharmacologically.
3.4 Study limitations.
Several factors other than atrial function can influence the peak velocity of the atrial filling wave. These include the loading conditions of the heart, ventricular compliance, heart rate and appropriate placement of the sample volume. Therefore, the peak velocity of the atrial filling wave is, at best, an indirect and less than ideal way of assessing atrial function. Precautions were taken, however, to minimize technical sampling errors and to ensure standardization of measurements. We also reviewed the medical records of enrolled patients to look for any major events during the study period that could possibly have influenced the peak velocity of the atrial filling wave, independent of atrial function. No major hemodynamic changes that could potentially affect recovery of EMAF were seen in our group during the study period.
We sincerely acknowledge the technical and nursing assistance provided by Susan Revall, Nancy Davison, RN, Kim Drake, RN, and Melanie Bradley-Brown during the course of this study. We are thankful to Lauren Camardelle for editorial assistance with the manuscript.
↵1 This study was performed during Dr. Cheirif’s tenure of Clinician-Scientist Award 92004390 from the American Heart Association, Dallas, Texas.
This study was presented in part at the annual meeting of the American Society of Echocardiography, Orlando, Florida, June 1997 by Dr. Harjai as part of the Young Investigator Competition.
- atrial wave filling velocity
- early wave filling velocity
- ratio of early transmitral flow velocity to atrial flow velocity
- electrocardiogram, electrocardiographic
- effective mechanical atrial function
- Received December 24, 1996.
- Revision received April 7, 1997.
- Accepted April 21, 1997.
- The American College of Cardiology
- Manning WJ,
- Leeman DE,
- Gotch PJ,
- Come PC
- Fatkin D,
- Kuchar DL,
- Thorburn CW,
- Feneley MP
- Missault L,
- Jordaens L,
- Gheeraert P,
- Adang L,
- Clements D
- Black IW,
- Fatkin D,
- Sagar KB,
- et al.
- Manning WJ,
- Silverman DI,
- Katz SE,
- et al.
- Abascal VM,
- Dubrey S,
- Ochoa MR,
- Davidoff R,
- Falk RH
- Grimm RA,
- Stewart WJ,
- Maloney JD,
- et al.
- Resnekov L,
- McDonald L
- Black IW,
- Hopkins AP,
- Lee LCL,
- Jacobson BM,
- Walsh WF
- Manning WJ,
- Reis GJ,
- Douglas PS
- Lipkin DP,
- Frenneaux M,
- Stewart R,
- Joshi J,
- Lowe T,
- McKenna WJ