Europace (2006) 8, 297–301
doi:10.1093/europace/eul010
A simple point score system for predicting the efficacy
of external rectilinear biphasic cardioversion for
persistent atrial fibrillation
Sebastian Stec*, Aleksander Gorecki, Beata Zaborska, and Piotr Kulakowski
Department of Cardiology, Postgraduate Medical School, Grochowski Hospital, Grenadierow 51/59 Street,
04-073 Warsaw, Poland
Received 6 May 2005; accepted after revision 15 January 2006; online publish-ahead-of-print 16 March 2006
KEYWORDS
Persistent atrial fibrillation;
Cardioversion;
Rectilinear biphasic
waveform;
Initial energy;
Point score system
Aims To develop a simple point score system that can accurately predict the optimal energy of initial
rectilinear biphasic (RLB) waveform shock for cardioversion (DC) of persistent atrial fibrillation (AF).
Methods and results Data from 302 consecutive patients with AF who underwent a step-up protocol of
sequential shocks of 50 J—from 1 up to 2 J/kg—200 J of RLB waveform DC were prospectively examined.
Using a logistic regression model, three variables independently predicted the need for 2 J/kg shocks:
AF duration .7 months, previous DC, and increased left atrial (LA) diameter .4.5 cm. A simplified point
score system (REBICAF score) that spans from 0 to 4 was developed. The score gives two points for AF
duration .7 months and one point for previous DC or LA diameter .4.5 cm. The area under the receiver
operator curve (ROC) of the proposed score for predicting the need for 2 J/kg shock was 0.84. There was
a progressive increase in the need for 1 J/kg, 2 J/kg, and 200 J as the point score increased (P , 0.001,
x 2 test for trend). More than 90% cumulative success rate was achieved in the low- (0–1), intermediate(2), and high-REBICAF (3–4) score subgroups with 1 J/kg, 2 J/kg, and 200 J RLB shocks, respectively.
Conclusion A simple point score system is useful in prediction of successful initial RLB energy for DC
of AF.
Introduction
Direct-current cardioversion remains the gold standard for
restoration of sinus rhythm in patients with persistent
atrial fibrillation (AF).1,2 It is desirable to reduce the total
number of shocks and use the lowest energy shock likely
to terminate AF to avoid deleterious myocardial and haemodynamic effects.2 Recent studies have shown that the use of
the rectilinear biphasic (RLB) defibrillator in unselected
populations of AF allowed a 99% success rate of cardioversion, the highest ever percentage reported for cardioversion2–11 using mono- or biphasic waveforms.
The Joint Committee of the European Society of
Cardiology/American College of Cardiology/American
Heart Association guidelines recommended 200 J as an
initial energy setting for monophasic cardioversion.2 There
is, however, limited information on factors influencing
effectiveness of RLB shocks and initial energy needed for
RLB cardioversion of persistent AF.2–10
* Corresponding author. Tel/fax: þ48 22 8109802.
E-mail address: [email protected]
We have developed a simple scoring system for assessing
the need for high-energy of RLB cardioversion and the likelihood of an initial RLB shock to terminate persistent AF. Our
purpose in designing the REBICAF (REctiliner BIphasic
Cardioversion of persistent AF) point score system was to
provide clinicians with a tool to achieve a high success
rate of the first RLB shock with the lowest energy and
a small cumulative energy in unselected patients with
persistent AF.
Methods
The study group was composed of 302 consecutive patients at a
single centre who underwent elective external cardioversion for
persistent AF between August 2002 and August 2004. Standard
clinical criteria were used to select patients for cardioversion.
The patients were effectively anticoagulated for at least 4 weeks
before cardioversion. According to the physicians’ decision,
pharmacological anti-arrhythmic treatment was begun or continued
in some patients before cardioversion. In all cases, two metal handheld paddle electrodes (diameter of 8 cm) were positioned in an
antero-lateral (right infraclavicular area—ventricular apex) configuration. The RLB waveform was delivered by the RLB defibrillator
(Zoll M-Series Biphasic, Zoll Medical Corp., USA) with energy setting
& The European Society of Cardiology 2006. All rights reserved. For Permissions, please e-mail: [email protected]
298
1–10, 25, 50, 75, 100, 120, 150, and 200 J. A pre-defined step-up
protocol was prospectively used, with sequential shocks in a synchronized fashion of 50 J, followed by 1 J/kg, 2 J/kg (maximum
200 J), and 200 J. The end point of the protocol was defined as
the restoration of sinus, atrial paced or junctional rhythm, or delivery of four shocks. Success was defined as interruption of AF for
.1 min (immediate recurrence of AF within 1 min of the shock
was treated as an unsuccessful procedure and such patients underwent the next step of protocol). The study protocol complied with
the Declaration of Helsinki and was approved by the Ethics
Committee of our institution. All patients signed informed consent
before cardioversion.
A total of 16 baseline characteristics arranged in a dichotomous
fashion were screened as candidate predictor variables of the
need for 2 J/kg RLB shock. They were tested in a multivariate, stepwise (backward elimination) logistic regression model.12 Variables
associated with P , 0.05 were retained in the final model. The fit
of the model was calculated by the Hosmer–Lemeshow statistic.12
After development of the multivariate model, the REBICAF point
score was calculated for those variables that had been found to
be independent predictors of the need for 2 J/kg shock in the multivariate analysis. The score was then constructed by an arithmetic
sum of the points assigned for variables present before cardioversion. Differences in the cumulative success rates for increasing
REBICAF score values were assessed using the x 2 test. The predictive
performance of the model was evaluated by using the C-statistic.
Data of continuous variables are expressed as median, and
mean + SD. Value of median was chosen to dichotomize some
continuous variables (age, body weight, and AF duration) for
multivariate analysis. A P-value , 0.05 was considered statistically
significant. All analyses were performed using STATA 7.0 statistical
software (Statcorp, College Station, TX, USA).
Results
The mean age of the study population was 67.4 + 10.4 years
(range: 38–90) with AF that had been present for mean
8.7 + 5.2 months (range: 1–24). Clinical characteristics of
the study population are shown in Table 1.
Overall, RLB cardioversion was effective in interrupting
AF in 295 of 302 patients (97.7%), with mean 1.98 + 0.94
shocks. The mean cumulative energy delivered during
repeated shocks of this protocol was 165 + 140 J.
The cumulative efficacy of RLB cardioversion was 35% for
the 50 J shocks, 75% for the 1 J/kg shocks, 91% for 2 J/kg
shocks, and 98% for the 200 J shocks.
Within 24 h of observation period, there were only
four patients with early recurrence of persistent AF after
successful RLB shocks (1.4%). Early recurrences were
observed in patients after successful cardioversion by 2 J/kg
RLB shock (two patients) and by 200 J RLB shock (two
patients).
We did not observe any adverse effects related to
procedure including transient ischaemic attack, stroke,
pacemaker dysfunction, symptomatic bradycardia requiring
temporary pacing, or exacerbation of heart failure.
Although skin burn or skin discomfort scale after RLB
cardioversion was not closely monitored, no significant
skin injury or persistent chest pain was reported.
Independent risk factors associated with the need for
2 J/kg RLB shock were AF duration .7 months, previous cardioversion, and increased LA diameter .4.5 cm (Table 2).
Of all the variables studied, duration of AF was the best
predictor of failed low-energy RLB waveform cardioversion.
The need for 2 J/kg shocks increased by 39% (6.6% vs. 45.9%,
S. Stec et al.
Table 1 Patient population
Characteristic
Study group (n ¼ 302)
Age (years) [mean + SD] (range)
Age (years) [median]a
Male
Body weight (kg) [mean + SD] (range)
Body weight (kg) [median]a
No heart disease found
Systemic hypertensiona
Without LVH
With LVH
Valvular heart disease
Coronary artery disease
Previous myocardial infarctiona
Dilated cardiomyopathy
Permanent pacemakera
Diabetesa
Previous stroke/TIA
AF duration (months) [mean + SD]
(range)
AF duration (months) [median]a
Previous AF episodesa
Previous cardioversiona
LA diameter .4.5 cma
LVEF , 40%a
Drug treatment (any)
Sotalola
Amiodaronea
Propafenonea
Verapamil or Diltiazema
b-Blockersa
67.4 + 10.4 (38–90)
70
169 (56%)
82.4 + 12.4 (60–126)
78
7 (2%)
244 (81%)
183 (80%)
61 (20%)
56 (19%)
114 (38%)
65 (22%)
15 (5%)
34 (11%)
69 (23%)
13 (4%)
8.7 + 5.2 (1–24)
7
221 (73%)
95 (31%)
116 (38%)
36 (12%)
280 (97%)
51 (17%)
45 (15%)
25 (8%)
5 (2%)
199 (66%)
Values are expressed as median, mean + SD, or n (%). LVH, left ventricular hypertrophy with interventricular septum 1.4 cm; TIA, transient
ischaemic attack; LVEF, left ventricular ejection fraction.
a
Factors included in the multivariate analysis.
Table 2 Independent predictors of the need for high-energy
(2 J/kg) RLB shock in step-up protocol of cardioversion for
persistent AF
Factors
OR (95% CI)
P-value
Body weight .78 kg
AF duration .7 months
Previous cardioversion
LA diameter .4.5 cm
1.90 (0.96–3.83)
9.70 (4.40–21.50)
3.12 (1.50–6.30)
2.60 (1.30–5.30)
0.063
,0.001
0.002
0.006
Hosmer–Lemeshow test P . 0.05, ROC ¼ 0.84.
P , 0.001) if AF persisted for more than 7 months before
cardioversion.
AF duration .7 months had at least a two-fold higher
magnitude than the two other risk factors, so the point
score system assigned a value of 2 when this variable was
present. Previous cardioversion and increased LA diameter
.4.5 cm had a similar magnitude, so they were each
assigned a value of 1. The REBICAF score categorized
patients by the number of points (range: 0–4) as shown in
Figure 1.
There was a consistent, progressive, significant pattern of
increased need for the 1 J/kg, 2 J/kg, and 200 J RLB shocks
Scoring system for cardioversion of AF
as the REBICAF score increased in the test cohort
(P , 0.001, x 2 test for the trend).
As shown in Figure 2, there was a 15-fold difference in the
need for 2 J/kg RLB shock between patients with 0 or 1
REBICAF score compared with the highest REBICAF score
value. The C-statistic of the model for the need for 2 J/kg
RLB shock to achieve termination of AF was 0.84 with the
best-fit of the model.
Because 50 J RLB shock had a low success rate in this
step-up protocol, the next energy settings were tested for
the evaluation of the REBICAF score system. A similar need
for 1 J/kg RLB shock was observed in patients who scored
0 and 1 points as well as 200 J for those scoring 3 and 4
points. Therefore, those groups were combined for the
final conclusion and the cohort was divided into low-,
intermediate-, and high-REBICAF score subgroups (0–1
points, 2 points, and 3–4 points, respectively).
Compared with low- and intermediate-REBICAF score
groups, patients in the high-REBICAF score group had significantly higher incidence of LA diameter .4.5 cm (20 and 20
vs. 81%, P , 0.01), previous cardioversion (27 and 20 vs.
48%, P , 0.01), LVEF , 40% (4 and 7 vs. 27%, P , 0.01),
tendency for higher weight (77.8 + 9.3 and 84.8 + 13.3
vs. 86.6 + 13.1 kg), and significantly longer duration of AF
(5.1 + 1.5 and 9.1 + 3.9 vs. 13.4 + 5.7 months, P , 0.001).
299
The use of propafenone, sotalol, b-blockers, verapamil,
and diltiazem was similar across subgroups. Amiodarone
was administered more often in high-REBICAF score group
compared with low- and intermediate-REBICAF score
groups (25 vs. 10 and 11%, P , 0.01).
In patients without parameters included in the REBICAF
score system (REBICAF score ¼ 0), no patient required the
fourth step of the protocol. Acute failure of RLB cardioversion (n ¼ 7) and early recurrences of AF within 24 h (n ¼ 4)
were observed only in the intermediate- and high-REBICAF
score groups (1%/3% and 6%/2%, respectively).
We found that more than 90% cumulative success rate was
achieved by 1 J/kg RLB shock in the low-REBICAF score
group, 2 J/kg RLB shock in the intermediate-REBICAF score
group, and 200 J RLB shock in the high-REBICAF score group
(Figure 1). Therefore, REBICAF score-based protocol (initial
energy: 1 J/kg for the low-REBICAF score group, 2 J/kg for
the intermediate-REBICAF score group, and 200 J for the
high-REBICAF score group, then 200 J for every subgroup)
was proposed. Calculation of parameters revealed that
REBICAF score-based protocol would have achieved termination of persistent AF with significantly lower number of
shocks and mean cumulative energy compared with the
evaluated protocol (Table 3).
Discussion
Figure 1 Cumulative success rate 50 J, 1 J/kg, then 2 J/kg RLB
shock, and finally 200 J RLB shock according to the REBICAF point
score system and the overall results of the step-up protocol
(numbers 0–4 represent values of REBICAF point score).
This study demonstrates that three variables (AF duration
.7 months, LA diameter .4.5 cm, and previous cardioversion) routinely obtained before external cardioversion can
be used to construct a simple point score system that is predictive of the need for higher energy settings of the RLB
defibrillator to achieve termination of AF. The predictors
were derived from the logistic regression model that confirmed their independent predictive power for a 2 J/kg
RLB shock. The scoring system categorized patients into
low-, intermediate-, and high-energy requirement strata
and allowed prediction of about 95% success rate of initial
RLB shock with low cumulative energy and number of
shocks. The scoring system is simple to calculate and easy
to apply in clinical practice.
The cumulative success rate of RLB cardioversion
achieved in our study is in agreement with results previously
published by the groups of Niebauer and Siaplaouras.6,9
Niebauer et al.6 reported that the antero-posterior paddle
position was effective in 99% of patients who underwent
RLB waveform cardioversion of AF.
In a small randomized study, Siaplaouras et al.9 showed that
the electrode position (antero-posterior vs. antero-lateral)
Table 3 Comparison of cardioversion calculated according to
step-up protocol and REBICAF score-based protocol
Parameters of
cardioversion
50 J–1 J/kg–
2 J/kg–200 J
protocol
REBICAF
score-based
protocol
Overall success rate (%)
Mean cumulative energy (J)
Number of shocks
97.7
165 + 140
1.98 + 0.94
97.7
147 + 83
1.08 + 0.31
Figure 2 Need for 2 J/kg RLB shock according to the REBICAF point
score system for persistent AF.
P , 0.05 vs. the REBICAF score-based protocol.
P , 0.008 vs. the REBICAF score-based protocol.
300
did not influence the efficacy of external RLB cardioversion
with regard to acute success (94.9 vs. 95.2%, P ¼ ns) and
early recurrence of AF (11.6 vs. 4.8, P ¼ ns). We demonstrated that with antero-lateral (right infraclavicular area—
ventricular apex) position, we were able to achieve
similar success rate. Moreover, the investigators in previous
studies used adhesive pads in antero-posterior and/or anterolateral configurations.6,9 Therefore, metal hand-held paddle
electrodes and adhesive pads delivering RLB waveform in
antero-lateral position seem to be equally effective for
cardioversion of persistent AF.
There are limited data on initial energy required for
successful RLB waveform cardioversion of persistent AF.
Siaplaouras et al.9,10 reported that an initial shock of 120 J
is successful in about 75–80% of patients with persistent
AF, whereas this figure rises to about 95% with 200 J
(including patients with early recurrence of AF after RLB
cardioversion). Comparison of our data with results of
120–150–200–200 J RLB cardioversion protocol used by
Siaplaouras et al.10 revealed that REBICAF score basedprotocol would have achieved termination of persistent
AF with significantly lower number of shocks and mean
cumulative energy (Table 3).
Neal et al.11 reported no release of cardiac troponin after
RLB waveform shocks; however, they found a positive correlation between skin erythema and skin discomfort after
shocks with increasing cumulative energies. Therefore, it
seems valuable to set RLB energy based on pre-existing
data in order to limit the number of shocks, cumulative
energy, and employ a simplified protocol. In addition, in
some specific groups of patients (e.g. patients with pacemakers) reduction of cumulative energy is particularly
beneficial.
The approach undertaken in developing the REBICAF score
is similar to that taken by Antman et al.13 for risk stratification and choice of treatment in patients with non-ST
elevation acute coronary syndromes. We used a similar statistical methodology to develop a scoring system for prediction of the need for high RLB energy to terminate AF and to
guide the initial energy setting of the RLB shock in patients
admitted for elective cardioversion of persistent AF.
We observed similar factors associated with the need for
the high-energy RLB shock as reported for failure of monophasic waveform cardioversion. In contrast to previous
studies, body weight, age, and ejection fraction ,40% did
not reach statistical significance for the necessity of 2 J/kg
shocks (Table 3). In a previous study, multivariate analysis
documented that short duration of AF, presence of atrial
flutter, and younger age were independent predictors
of success, whereas LA enlargement, underlying heart
disease, and cardiomegaly predicted failure of monophasic
cardioversion.14
Of all the variables studied, duration of AF was the best
predictor of failed low-energy RLB waveform cardioversion,
which is in agreement with current guidelines, and studies
on monophasic and biphasic cardioversions.2,5,8,13 Current
guidelines emphasize anatomical and electrophysiological
mechanisms, the so-called atrial remodelling, that
promote the progressive persistence of this arrhythmia and
make the interruption of AF more difficult.2
Moreover, it is important for the future to develop a
system expressing not only the likelihood of success of
initial energy setting of RLB cardioversion, but also to
S. Stec et al.
predict long-term maintenance of sinus or atrial-paced
rhythm. There are some clinical, echocardiographic, and
biochemical factors that have been proved to be associated
with the maintenance of sinus rhythm after successful
monophasic cardioversion.14–16 Because of the higher efficacy of the biphasic than the monophasic waveform cardioversion, further studies of factors determining short- and
long-term maintenance of sinus rhythm after RLB cardioversion are needed.6,9,11,15 It will facilitate clinical decisionmaking and improve cost-effectiveness by avoiding RLB
cardioversion in patients unlikely to benefit from a rhythmcontrol strategy.
Limitations
This study has some limitations. Although it included a relatively large number of patients, the anti-arrhythmic drug
treatment was not randomly assigned, but prescribed individually according to the patient’s history. This might have
accounted for the lack of any difference in cardioversion
success between patients who were on anti-arrhythmic
drugs at the time of cardioversion and those who were
not. In contrast, higher rate of amiodarone treatment in
high-point score group might have improved the rate of successful cardioversion and recurrence of AF within the observation period. Moreover, we did not record immediate
recurrence of AF; as this appeared during general anaesthesia, physicians went to the next step of the protocol to
achieve prolonged AF termination. From the academic
point of view, failure cardioversion and immediate recurrence of AF are different entities. However, from a clinical
point of view both these outcomes mean failure of cardioversion to restore sinus rhythm; therefore, a clinical definition was used in our paper. Moreover, we used simple
median values of body weight and AF duration instead of
using them as continuous variables. It is possible that the
performance of the model could be improved by using continuous variables; however, the point score system might
become more complex and not easily calculated in routine
practice.
There is also the question of point score used in the wide
spectrum of patients presenting with paroxysmal, shortlasting AF, or persistent AF with duration shorter than 4
weeks (accepted for cardioversion on the basis of effective
anticoagulation or after transoesophageal echocardiography). Because of significant differences in efficacy and electrical parameters among biphasic defibrillators, the results
of this study can be applied only to RLB waveform cardioversion. Therefore, the REBICAF point score system needs
further prospective validation in a variety of clinical settings
and populations with AF.
We did not evaluate skin burn scale, pain scale, or troponin release after RLB waveform cardioversion. However, no
significant pain or skin injury leading to additional intervention or delay in discharge were observed. A previous study
reported no troponin increase after RLB waveform cardioversion, regardless of the total energy used.11 The low
rate of early recurrences of AF (within 24 h after cardioversion) limited statistical analysis of their relationship
with the level of energy delivered.
Results are defibrillator-specific and may not be generalized to devices from other manufacturers.
Scoring system for cardioversion of AF
Conclusions
The antero-lateral hand-held metal paddle position of RLB
cardioversion is effective in about 98% of patients with persistent AF. The duration of AF .7 months, left atrial enlargement .4.5 cm, and previous attempt at cardioversion are
independent predictors of the need for higher energy of
RLB shock for termination of AF. On the basis of those
three factors, the REBICAF score can be calculated (range:
0–4 points) with a very good predictive performance for
the need of high-energy requirements to terminate the
arrhythmia. The REBICAF score system enables physicians
to categorize patient’s requirements for the initial RLB
energy with a high likelihood of successful cardioversion.
Initial energy setting according to REBICAF score-based protocol seems to achieve cardioversion of persistent AF in
about 95% of patients, and moreover, with a low cumulative
energy and number of shocks. The data from this study could
be used when considering cardioversion in patients with
long-standing AF.
Acknowledgements
This study was supported by the grant 501-1-2-10-29/02 from the
Postgraduate Medical School, Warsaw, Poland. S.S. was supported
by a grant from the Foundation of Polish Science.
References
1. Lown B, Perloth MG, Kaibey S, Abe T, Harken DW. “Cardioversion” of
atrial fibrillation: a report on the treatment of 65 episodes in 50 patients.
N Engl J Med 1963;269:325–31.
2. Fuster V, Rydén LE, Asinger RW et al. ACC/AHA/ESC Guidelines for the
Management of Patients with Atrial Fibrillation: Executive Summary. A
Report of the American College of Cardiology/American Heart
Association Task Force on Practice Guidelines and the European Society
of Cardiology Committee for Practice Guidelines and Policy
Conferences (Committee to Develop Guidelines for the Management of
Patients With Atrial Fibrillation) Developed in Collaboration With the
North American Society of Pacing and Electrophysiology. Circulation
2001;104:2118–50.
301
3. Mittal S, Ayati S, Stein KM et al. Transthoracic cardioversion of atrial
fibrillation: comparison of rectilinear biphasic versus damped sine wave
monophasic shocks. Circulation 2000;101:1282–7.
4. Page RL, Kerber RE, Russell JK et al. Biphasic versus monophasic shock
waveform for conversion of atrial fibrillation: the results of an international randomized, double-blind multicenter trial. J Am Coll Cardiol
2002;39:1956–63.
5. Ricard Ph, Levy S, Trigano J et al. Prospective assessment of the minimum
energy needed for external electrical cardioversion of atrial fibrillation.
Am J Cardiol 1997;79:815–6.
6. Niebauer MJ, Brewer JE, Chung MK, Tchou PJ. Comparison of the rectilinear biphasic waveform with the monophasic damped sine waveform
for external cardioversion of atrial fibrillation and flutter. Am J Cardiol
2004;93:1495–9.
7. Koster RW, Dorian P, Chapman FW, Schmitt PW, O’Grady SG, Walker RG. A
randomized trial comparing monophasic and biphasic waveform shocks
for external cardioversion of atrial fibrillation. Am Heart J 2004;147:e20.
8. Wozakowska-Kaplon B, Janion M, Sielski J, Radomska E, Bakowski D,
Bartkowiak R. Efficacy of biphasic shock for transthoracic cardioversion
of persistent atrial fibrillation: can we predict energy requirements?
Pacing Clin Electrophysiol 2004;27:764–8.
9. Siaplaouras S, Buob A, Rotter C, Böhm M, Jung J. Biphasic transthoracic
cardioversion of atrial fibrillation: impact of different electrode positions. [Abstract]. Eur Heart J 2004;66:2942.
10. Siaplaouras S, Buob A, Rotter C, Bohm M, Jung J. Impact of biphasic electrical cardioversion of atrial fibrillation on early recurrent atrial fibrillation and shock efficacy. J Cardiovasc Electrophysiol 2004;15:895–7.
11. Neal S, Ngarmukos T, Lessard D, Rosenthal L. Comparison of the efficacy
and safety of two biphasic defibrillators waveforms for the conversion of
atrial fibrillation to sinus rhythm. Am J Cardiol 2003;92:810–4.
12. Hosmer DW, Lemeshow S. Applied Logistic Regression. New York: Wiley;
1989.
13. Antman EM, Cohen M, Bernink PJ et al. The TIMI risk score for unstable
angina/non-ST elevation MI: a method for prognostication and therapeutic decision making. J Am Med Assoc 2000;284:835–42.
14. Van Gelder IC, Crijns HJ, van Gilst WH, Verwer R, Lie KI. Prediction of
uneventful cardioversion and maintenance of sinus rhythm from directcurrent electrical cardioversion of chronic atrial fibrillation and flutter.
Am J Cardiol 1991;68:41–6.
15. Kosior D, Szulc M, Piatkowski R, Rabczenko D, Torbicki A, Opolski G.
Factors determining long-term maintenance of sinus rhythm after cardioversion of persistent atrial fibrillation. Kardiol Pol 2003;59:128–41.
16. Wozakowska-Kaplon B, Opolski G. Effects of sinus rhythm restoration in
patients with persistent atrial fibrillation: a clinical, echocardiographic
and hormonal study. Int J Cardiol 2004;96:171–6.