Journal of the New Zealand Medical Association, 23-September-2011, Vol 124 No 1343
Prediction of cardiac rhythm 1 year following cardioversion for atrial fibrillation
Amjad K Hamid, A Mark Richards, Ian G Crozier, John G Lainchbury, Iain Melton, Paul G Bridgman, Suetonia C Palmer, Chris M Frampton, M Gary Nicholls
Atrial fibrillation (AF) affects 0.4-1% of the general population. Prevalence increases with age to 8% in those over 80 years.1–3 AF is associated with increased risk of stroke, heart failure and all-cause mortality. The significant costs of management are driven largely by hospitalisation.1,4 Treatment consists of either ventricular rate control combined with anticoagulation, or attempts at restoring sinus rhythm (SR).
Robust randomised controlled trials suggest rate control is not inferior to attempted rhythm control regarding survival and morbidity, and is appropriate for many patients with recurrent or persistent AF.1,5,6 Similar findings were observed in patients with congestive heart failure.7 Accurate case selection for cardioversion (CV) is required to offer the best chance of sustained SR.
Whilst CV for AF has been used for over 40 years, follow-up periods have often been brief with reported rates of sustained SR of 44 to 69% at 4 weeks and 23 to 79% at 1 year.8-14 Factors predicting atrial fibrillation recurrence include duration of arrhythmia,9,10,13,14,16-19 type of arrhythmia (AF or flutter),14 echocardiographic indices,9,10,18,20 age,1,13,14,17 gender,1,12 underlying cardiac disease,5,11,12,17 frequency of paroxysmal AF and previous CV attempts,1,16,17 the use of rate-controlling or anti-arrhythmic drugs,11,17,18 functional status (NYHA Class),13,14,19 hypertension or pulmonary disease,5,11,12 the initial energy used to achieve CV21 and restoration of SR with drug therapy.9
We report findings from a large consecutive cohort of patients with AF/flutter presenting over a 2-year period regarding maintenance of SR over 1 year, predictors of rhythm outcomes, and mortality and cardiovascular morbidity.
The study, approved by the Regional Ethics Committee, was carried out in Burwood and Christchurch Hospitals which serve a population of 481,431 (2001 population census). All patients with AF/flutter undergoing CV between December 2000 and December 2002 were included. Patients were categorised into those who underwent immediate CV within 72 hours from AF/flutter onset (IC) and elective CV after receiving warfarin for at least 4 weeks with an INR at 2 to 4 for 2 weeks prior to CV (EC).
A transoesophageal echocardiogram was first performed on patients with AF duration of more than 48 hours, IC group, or they had not been on Warfarin long enough; EC group. Warfarin was continued for at least 6 weeks post CV. Demographics, known duration of arrhythmia, presenting symptoms, comorbid conditions, medications and the number of DC shocks and Joules delivered were recorded.
In preparation for CV, all patients had a 12 lead electrocardiograph (ECG), routine blood tests and echocardiography according to American Society of Echocardiography Guidelines22. Using a monophasic defibrillator, CV was performed following the intravenous administration of propofol and remifentanil or midazolam with the defibrillator paddles placed on the chest in the anterior and posterior positions. The standard energy selection was 200 Joules for the first shock followed, if required, by 2 shocks each at 360 Joules. CV was considered successful if the patient maintained SR for more than 20 minutes. Cardiac rhythm and medical status were recorded at 6 weeks and 1 year following CV.
Apart from 38 patients, all CV in the EC group were performed by one of the authors (AH). The primary end point of the study was cardiac rhythm at 6 weeks and 1 year. Secondary outcomes included total and cardiovascular mortality and cardiovascular morbidity as determined from the hospital electronic Patient Management System records.
Statistics—Data are shown as mean (± SD or SEM), median (interquartile range) when non-parametric, or frequency (%). Comparisons of categorical and continuous variables between the IC and EC groups were conducted using Chi-square and independent t-tests or Mann-Whitney U tests respectively. Univariable regression analysis was used to determine predictors of SR at 6 weeks and 1 year after cardioversion. Within the EC group only, in view of adequate sample size for robust results, the independent predictive power of atrial flutter, hypertension, medications, requirement for >1 shock, duration of AF/atrial flutter and LA diameter>40 mm to predict SR at 6 weeks and 1 year after EC was tested using multivariable logistic regression. P<0.05 indicated statistical significance. Analyses were conducted using SPSS version 13.
Over 2 years, 53 patients underwent IC and 508 patients EC. One IC patient and 12 in the EC group were lost to follow up leaving 52 and 496 respectively. Patients in the IC group underwent CV while receiving low molecular weight heparin. None of the IC group had undergone prior CV. Seventy EC patients (15%) had one and 17 (4%) had two or more previous cardioversions.
Patient characteristics are shown in Table 1. Males presented with AF/atrial flutter at a younger age than females (statistically significant within EC group, P<0.001).
Table 1. Clinical, treatment & echocardiography data (mean ± SEM, median [interquartile range] or number [%])
P<0.001, † P<0.01, ‡ P<0.05 for comparisons among groups using independent t-test, Mann-Whitney U tests and Chi-square tests as appropriate. All percentages are given to the nearest whole percent.
Initial success rates were similar in both groups regardless of the presenting rhythm. In the EC group, 86.3% were successfully cardioverted (87.4% in AF and 86.1% in atrial flutter). In the IC group 94.2% were cardioverted (94.4% in AF and 93.8% in atrial flutter patients). A significantly higher percentage of IC patients were in SR at both 6 weeks and 1 year than in the EC group (65.4% versus 43.3 % and 57.7% versus 30% respectively—Figure 1).
Table 2 illustrates the significant predictors of SR maintenance at 6 weeks and 1 year. Combining known duration of AF/flutter with the requirement for >1 shock to achieve SR provided a graded prediction regarding outcome at 6 weeks and 1 year (Figure 2) which was superior to the prediction using either factor alone. Neither previous CV nor CHADS2 score predicted rhythm outcome at 6 weeks or 1 year on either univariable or multivariable analysis.
Figure 1. The percentage of patients in SR at 6 weeks and 1 year according to the number of shocks required to achieve SR in both Immediate and Elective Cardioversion groups combined, n=477
Morbidity and mortality—At 1 year, there was no difference in mortality between groups (IC one death; 1.9%, EC 11 patients; 2.2%) or new-onset heart failure (2 patients; 3.8% compared with 14 patients; 2.8%). Combined cardiovascular events including fatal and non-fatal acute coronary syndromes, stroke (1 patient in the IC; 1.9%, 7 patients in the EC; 1.4%), hospitalisation due to recurrence of AF/flutter and heart failure affected a higher percentage of patients in the IC group (24.7%) than in the EC group (14.8%; P=0.025).
Figure 2. Maintenance of SR at 6 weeks and 1 year according to a combination of known duration of AF/atrial flutter and number of shocks required for successful cardioversion
Only 43.3% and 30% of the EC group remained in SR at 6 weeks and 1 year respectively. The success rate was significantly higher in the IC cohort (65.4% and 57.7% respectively). We have documented, for the first time, that initial resistance to electrical CV, reflected in the number of shocks required to restore SR, is one indicator of the likelihood of reverting to AF.
Along with others,9,13,14,16,18-20 we identified arrhythmia duration as an independent predictor of rhythm outcome. With long-established AF/flutter, electrical remodelling becomes established rendering the atria resistant to resynchronisation and vulnerable to early return of AF or flutter. Our results, along with earlier reports,9,10,13,14,16-20 suggest that every effort should be made to ensure patients are referred early for CV.
The ability of echocardiographic indices to predict outcome following CV is controversial. In our patients, as reported elsewhere,9,20 left atrial dilatation was predictive of return to AF/flutter – but on univariate analysis only.
Higher rates of sustained SR in IC compared with EC patients presumably reflect younger age, a higher proportion in atrial flutter, shorter duration of arrhythmia, less prior cardiovascular disease and more frequently normal cardiac structure.
The role of amiodarone in preventing atrial fibrillation post cardiac surgery and post CV is well established.23-26 Accordingly, we noted pre-treatment with amiodarone appeared protective against recurrence of AF/atrial flutter. IC and EC patients on beta-blockers were more likely to remain in SR, although statistically not significant, similar to earlier findings suggesting that at least some beta-blockers improve rates of sustained SR after CV25-27 and for patients with heart failure beta-blockers reduce the incidence of new AF onset.28
Despite evidence that the renin-angiotensin system promotes arrhythmogenesis15 whereas blockade of the system (with ACE inhibitors or angiotensin II receptor blockers) can be inhibitory, we found ACE inhibitors had no impact on post CV outcome. Those on ACE inhibitors, however, were more likely to have impaired cardiac function. Accordingly, it is likely that any protective effect of ACE inhibitors against return of AF was masked by the underlying cardiac disorders for which they were prescribed. Similarly in our group of patients, being on statins did not seem to influence SR maintenance.
Hypertensive cardiovascular disease is the most common antecedent of AF29,30. Accordingly, one might have anticipated hypertension to be a negative, rather than positive predictor of outcome. However, hypertension did not predict outcome at 1 year, and did not retain independent predictive power on multivariate analysis. Similar to earlier findings,14 the presence of atrial flutter rather than AF was a positive predictor of sustained SR.
Limitations—The effect of medications on cardiac rhythm cannot be ascertained with confidence from our data. Beyond the factors we have evaluated, a variety of medications and chromosomal variants may have played a role in determining success or failure in maintaining SR.
A monophasic defibrillator was utilised in all patients undergoing cardioversion in this study. Current guidelines and recent robust papers recommend the use of biphasic shocks and the extrapolation of our findings to biphasic defibrillators may not be appropriate.
In summary, in a large consecutive cohort from a single referral centre, 57.7% of patients requiring immediate cardioversion for AF/atrial flutter remained in SR after CV at 1 year whereas the figure was only 30% in patients undergoing elective cardioversion. The number of shocks required to achieve SR, and the known duration of AF/atrial flutter are independent additive predictors of rhythm at 6 weeks and 1 year after elective CV.
Competing interests: None.
Author information: Amjad K Hamid, SMO, PhD Student*; A Mark Richards, Professor*; Ian G Crozier, Cardiologist†; John G Lainchbury, Cardiologist†; Iain Melton, Cardiologist†; Paul G Bridgman, Cardiologist†; Suetonia C Palmer, Nephrologist*; Chris M Frampton, Biostatistician*; M Gary Nicholls, Professor*
*Department of Medicine, Otago University – Christchurch
†Department of Cardiology, Christchurch Hospital, Christchurch
Acknowledgement: This research was funded by a grant from the National Heart Foundation of New Zealand.
Correspondence: Amjad K Hamid, Cardiology Department, Christchurch Hospital, Private Bag 4710, Riccarton Avenue, Christchurch, New Zealand. Fax: +64 (0)3 3648303; email: firstname.lastname@example.org
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads)
Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals