The development of new-onset atrial fibrillation (AF) frequently complicates the postoperative phase of coronary artery bypass grafting (CABG) surgery. If not prevented by medication, the incidence of AF has ranged from 5 to 100%, depending on the technique of the arrhythmia detection (1,2). The AF mostly arises within the first 5 postoperative days, and recurrence is common (3). A variety of factors are associated with the generation of AF, but the exact pathophysiologic mechanisms are still incompletely understood.
Cardiac surgery induces sympathetic nervous overactivity, further exaggerated by withdrawal of β-adrenergic antagonist therapy (4,5). In fact, an independent role of catecholamines in provocation of postoperative AF has been demonstrated (6). Concordant with this, β-blocker treatment reduces the incidence of postoperative AF (2). Successful response also has been accomplished by digoxin and by drugs belonging to the Vaughan Williams class III (7,8).
Sotalol, a β-blocker with an ancillary class III effect, has effectively prevented postoperative AF compared with a pure β-blocker or placebo (9-12). However, it is not established whether the antiarrhythmic advantage truly is a manifestation of prolonged repolarization, considering the rather low dosing regimens used.
Our study compared the efficacies of low-dose sotalol and metoprolol in the prevention of clinical sustained AF emerging shortly after CABG and examined whether prolongation of repolarization by sotalol can be demonstrated in this setting.
MATERIALS AND METHODS
Patients who were scheduled for CABG were screened. Ineligible were patients with concomitant valve replacement, antiarrhythmia surgery, or contraindications to β-blockers, such as asthma, sick sinus syndrome, second or third grade atrioventricular conduction disturbance, or poor left ventricular function with ejection fraction < 20%. Also excluded were patients who had long QT syndrome or needed class I or III antiarrhythmic medication. The final inclusion occurred early on the first postoperative morning, provided that the patient was in sinus rhythm > 50 beats/min (beats/min) and needed no inotropic treatment for heart failure. Altogether 191 consecutive patients were included and randomized to two treatment groups.
On the first postoperative morning, all eligible patients were assigned consecutive numbers. Because of the randomization process, equalization of the treatments occurred after every sixth or eighth patient, also at random.
The patients gave informed consent before the operation. The protocol was approved by the ethical committee of the Department of Cardiothoracic Surgery of the Helsinki University Central Hospital.
All the operations were performed by standard techniques used at the institution and were constant throughout the study period. They consisted of mild systemic hypothermia, topical cooling, and cardiac arrest. The protection of the heart was maintained by cold crystalloid antegrade cardioplegia containing sodium, 120 mM; potassium, 8 mM; magnesium, 16 mM; calcium, 1.2 mM; chloride, 160 mM; and bicarbonate, 10 mM. Pump prime consisted of Ringer's solution. Revascularization by using both arterial and vein grafts was performed as completely as surgically possible.
The patients were randomized to receive metoprolol, 75 mg/day (n = 93), or sotalol, 120 mg/day (n = 98), both in three divided daily doses, in a single-blinded manner in which the study subjects were unaware of the treatment being administered. If the sinus heart rate (HR) remained > 90 beats/min and no heart failure was present, the daily metoprolol dose was increased to 100-150 mg and sotalol dose to 160-240 mg. If bradycardia (< 50 beats/min) appeared, the starting dose was reinstituted. Digitalis, diuretics, angiotensin-converting enzyme inhibitors, dihydropyridine calcium antagonists, and antithrombotic medication were allowed. Potassium supplementation was used as long as needed. Treatment of acute AF episodes was according to the attending surgeon's or cardiologist's practice and consisted mostly of intravenous digoxin, β-adrenergic blocker, oral quinidine, or a combination of these. Neither quinidine nor any other medication was started as preventive therapy.
Study end points
The study period consisted of the first 6 postoperative days. Within that time, the appearance of any sustained AF was recorded. Other sustained atrial arrhythmias, ventricular tachycardia, and ventricular fibrillation also were recorded until the end of the study period. The study medication was terminated if symptomatic bradycardia (<40 beats/min), symptomatic hypotension (systolic blood pressure <90 mm Hg), torsade de pointes ventricular tachycardia, pulmonary edema, or a severe surgical complication occurred.
Twelve-lead rest electrocardiograms (ECGs) were recorded on the preoperative and each postoperative morning. HR, QRS, and QT durations were analyzed from the preoperative and day 6 recordings. QRS duration was measured from the beginning of the Q or R wave to the end of the S wave in lead V2. QT duration was measured from the initiation of the Q or R wave to the point where a line drawn tangentially along the descending part of the T wave crossed the TP baseline. In cases in which a biphasic T wave emerged, the terminal descending or ascending part of it was used for drawing. A separate U wave was disregarded. To study the difference between the groups, we attempted to overcome the confounding effect of HR on QT interval with three methods:
- Correcting by the formula proposed by Bazett, QTC = QT/RR1/2(13);
- Plotting the uncorrected QT intervals against the corresponding HRs; and
- Averaging the uncorrected QT durations of the rest ECGs by 10 beat/min groups of sinus HRs between 40 and 100 beats/min.
Cardiac rhythm was observed by central and bedside monitors for the first 2 days or longer if indicated, and every arrhythmia was recorded on paper. Thereafter, a 12-lead ECG was recorded when a patient complained of palpitations or irregular pulse was observed. AF was defined as irregular QRS complex intervals without detectable regular atrial activity and considered as clinical sustained AF if it continued for ≥ 15 min on bedside monitor. AF noticed on morning routine ECG or lasting long enough to be confirmed by 12-lead ECG were recorded as well. The tracings were analyzed blindly by a cardiologist. The time of onset, duration, and mean ventricular rate of AF were recorded.
In addition to clinically relevant hemodynamic and biochemical follow-up measurements, plasma potassium, plasma magnesium, and plasma creatinine concentrations were measured on postoperative days 1, 2, and 6. Perioperative myocardial infarction was defined as a new pathologic Q wave appearing on serial ECGs, or creatine kinase MB isoenzyme at a concentration of ≥ 100 IU, measured on postoperative day 1. Relative heart volume was determined from plain chest radiographs by measuring the three orthogonal axes of heart and using ellipsoid assumption for volume calculation (14). Mean heart-rate levels from the postoperative days 2-5 were calculated as the mean of three HR measurements made by counting the supine resting pulse rate over 1 min. The preoperative data concerning medical history, details of coronary angiography, and biplane left ventricular cineangiography were collected from patients' files.
Based on the previous study conducted at the institution, the incidence of postoperative AF was expected to be 25-30% (3). To identify a 50% reduction in the incidence of AF at an α level of 5% with 80% power, the study groups should include ∼100 patients in each (15). Continuous variables were analyzed by the two-tailed Student's t test and by the Mann-Whitney rank sum test for nonnormal distributions. The paired t test or analysis of variance with repeated measures was used to analyze the serial changes in continuous variables. Univariate analyses of the categoric variables were performed by the two-tailed Fisher's Exact test or the χ2 test. The stepwise logistic-regression analysis was used to test which variables, raised by the univariate analyses, independently predicted the occurrence of AF. A p value of <0.1 for entering the variables into the model was expected. The actuarial curves for the first episode of AF in the two study groups were plotted by using the Kaplan-Meier estimation, and the difference was tested with the log-rank test. The difference in uncorrected QT intervals between the treatment groups was tested by two methods:
- Least-squares linear-regression equations were calculated for preoperative and postoperative QT duration-HR plots. Equality of the regression lines across the groups was tested by computing the 95% confidence intervals (CIs) for the difference between the slopes of the lines and for the vertical distance between the lines (16); and
- The averaged QT durations in the specified subgroups based on HR were compared with analysis of variance.
All data were analyzed on the intention-to-treat basis and are expressed as mean ± SD. A p value of < 0.05 was regarded as statistically significant.
The groups were well balanced for all the registered pre- and intraoperative characteristics (Table 1). The mean daily starting dose of metoprolol was 74 ± 11 mg, and of sotalol, 116 ± 14 mg. Sixty-nine (74%) of the metoprolol patients and 85 (87%) of the sotalol patients completed the study with the starting dose (p < 0.05). The dose had to be increased in 17 (18%) of the metoprolol group and in 10 (10%) of the sotalol group. The mean daily doses of these patients were 101 ± 7 and 149 ± 16 mg, respectively. Six (7%) of the metoprolol and two (2%) of the sotalol patients did not tolerate the designed dosage scheme for the entire study period. The final daily doses of these patients were 54 ± 14 and 40 ± 0 mg, respectively. The medication had to be terminated prematurely in one patient in each group. Symptomatic bradycardia, hypotension, torsade de pointes, or sustained monomorphic ventricular tachycardia or ventricular fibrillation were not observed.
Incidence of atrial fibrillation
New-onset AF was detected in 30 (32%) of the metoprolol patients and in 16 (16%) of the sotalol patients (p < 0.01). The actuarial curves for the appearance of the first AF started to separate on day 2 (Fig. 1; p < 0.01). The first AF episode occurred on 2 ± 1 days in both groups (p = NS). The ventricular rate of AF did not differ between the groups (127 ± 32 beats/min in the metoprolol patients and 129 ± 29 beats/min in the sotalol patients; p = NS). In patients with recurrent AF, 2 ± 1 episodes per patient were noticed in both the metoprolol and the sotalol patients (p = NS).
To convert AF to sinus rhythm, quinidine was given in 19 (20%) of 93 metoprolol patients and in nine (9%) of 98 sotalol patients (p = 0.06). The administration of other antiarrhythmic drugs did not diverge between the groups. Electrical cardioversion was used in one patient in each group.
The sinus HRs did not differ between the groups preoperatively. There was a significant postoperative increase in sinus HRs in both groups (p < 0.001). The sinus HR with sotalol therapy was lower than that with metoprolol therapy during the whole treatment period (p < 0.001; Fig. 2).
Preoperative QRS, QT, and QTc intervals were comparable between the groups. Sotalol or metoprolol treatment did not alter QRS duration. QT interval shortened significantly in the metoprolol group (p < 0.001) but was unchanged in the sotalol group (p < 0.001 between the groups). QTc increased equally (p < 0.001) in the metoprolol and sotalol patients (p = NS between the groups; Table 2).
After the treatment, the uncorrected QT-heart-rate lines revealed prolongation of QT interval in the sotalol patients as a group compared with the metoprolol patients (p < 0.05; Fig. 3). However, among the specified heart-rate groups, the QT durations tended to differ in the slowest group only (p = 0.1). There were no differences preoperatively.
Regression analysis of the QT-heart-rate plots showed a significant prolongation of the uncorrected QT duration by sotalol treatment: QT = 512 − 2.01 × HR before the treatment and QT = 559 − 2.26 × HR after the treatment, with the mean difference of the parallel regression lines being 31 ms; 95% CI, 20-42 ms; p < 0.01 (Fig. 4A). The treatment did not change the slopes of the lines. Metoprolol did not affect the QT duration: QT = 501 − 1.78 × HR before the treatment and QT = 490 − 1.48 × HR on the treatment (mean difference, 10 ms; 95% CI, -1-20 ms; p = NS; Fig. 4B). The slopes did not differ.
Creatine kinase MB isoenzyme levels on postoperative day 1 and plasma magnesium and potassium levels on the first 2 days did not differ between the groups. However, plasma creatinine concentrations on the first (92 ± 20 vs. 85 ± 14 mM; p < 0.05), second (100 ± 28 vs. 90 ± 19 mM; p < 0.05), and sixth day (104 ± 27 vs. 96 ± 18 mM; p < 0.05) were higher in the sotalol patients.
Predictors of the occurrence of atrial fibrillation
In univariate analysis, older age (62 ± 8 vs. 57 ± 8 years; p < 0.001), larger preoperative radiologic heart volume (480 ± 90 vs. 450 ± 80 ml/m2; p < 0.05), greater plasma magnesium concentration on the first day (0.91 ± 0.10 vs. 0.87 ± 0.10 mM; p < 0.05), lower plasma potassium concentration on the third day (4.41 ± 0.36 vs. 4.62 ± 0.55 mM; p < 0.05), and increased number of new-onset left bundle branch block (7 vs. 1% of the patients; p < 0.05) were noted in patients with postoperative AF. Of these, multivariate analysis revealed age (p < 0.001), randomization to metoprolol treatment (p < 0.01), plasma magnesium concentration (p < 0.05), and heart volume (p = 0.07) as independent predictors of the occurrence of AF.
In patients with or without AF, QT (387 ± 47 vs. 381 ± 37 ms; p = NS) and QTc (429 ± 37 vs. 429 ± 35 ms; p = NS) durations, their changes or QT-HR plots (QT = 549 − 2.15 × HR and QT = 525 − 1.87 × HR; mean difference, 3 ms; 95% CI, −8−14 ms; p = NS) did not differ. Furthermore, QRS duration or mean sinus HRs within the whole study population or within the groups did not differentiate the patients with AF.
This study indicates that sotalol is superior to metoprolol in preventing the occurrence of postoperative AF. Prolongation in repolarization by sotalol could be demonstrated even with modest doses, suggesting that the specific class III effect of sotalol is present. The advantage of sotalol over metoprolol was not compromised by the appearance of proarrhythmic events.
Incidence of atrial fibrillation
The incidence of AF in the metoprolol group, 32%, is within the range of the results in previous trials among patients undergoing CABG without or with low-dose β-blocker prophylaxis. The incidence slightly exceeds the 26% rate previously reported from our institution when a similar metoprolol treatment regimen was used (3). Although Holter recording is a more sensitive technique than clinical detection used here, it is not necessary in revealing AF in postoperative patients because they are identified by clinical symptoms (3).
β-Adrenoceptor blockade prophylaxis
The capability of several β-blockers to attenuate the sympathetic contribution on AF occurrence in patients undergoing CABG is established (9,10,17-21). In a meta-analysis of randomized controlled trials, Andrews et al. (2) showed a markedly reduced likelihood of developing postoperative supraventricular arrhythmias with β-blocker treatment (OR, 0.28; CI, 0.21-0.36). The effect seems to be applicable not only to clinically observed arrhythmias but also to arrhythmias detected by Holter monitoring.
Sotalol prevents supraventricular and ventricular arrhythmias (12), and its efficacy in postoperative state has been tested in three previous trials. In the first published study by Jansen et al. (9), oral sotalol, 240 mg daily, was compared with oral metoprolol, 150 mg daily, and placebo. Sotalol significantly reduced the occurrence of supraventricular arrhythmias: 2.4, 15.3, and 36.0% in the sotalol, metoprolol, and placebo groups, respectively. The statistical power, however, suffers from the small number of study subjects. Suttorp et al. (11) found that sotalol, 160 mg daily, reduced the incidence from 33% in the placebo-treated controls to 16%. In another study by the same group (10), the slightly increased success rate of high dose (240 mg daily) over low dose (120 mg daily) sotalol was shadowed by increased numbers of adverse effects, necessitating discontinuation of the drug (10.5 vs. 2.8%). With adequate statistical power, our data demonstrate suppression of the incidence of postoperative AF from 32 to 16% with sotalol, agreeing with the prior studies and affirming proper tolerability of the low dose.
Mechanisms behind the preventive efficacy
The superiority of sotalol in preventing the development of AF could be the result of a better blocking effect on β-adrenoceptors alone or together with its additional class III properties. Supporting the former, the mean sinus HR, the only clinically available determinant of β-blockade, was higher in the metoprolol group (Fig. 2). From pharmacologic point of view, the β-adrenoceptor blocking potency ratio of metoprolol to sotalol is ∼2.7-3.3, making the dosages comparable (22). However, the longer plasma elimination half-time and almost complete bioavailability of sotalol might provide more sustained β-blockade (12,22). Although lack of a placebo control group limits us in quantifying the sinus-rate reduction by metoprolol treatment, sinus rates were comparable to the rates previously reported from our institution of metoprolol-treated postoperative patients and lower than in placebo patients in other studies (3,11). Sinus-rate slowing by sotalol is not exclusively attributed to β-blockade, because D-sotalol, the dextroisomer of sotalol with a pure repolarization-prolonging effect, also decreases HR by lengthening the sinoatrial node action-potential duration (23).
Sotalol prolongs repolarization and lengthens the atrial refractory period (12). This would augment the pure β-blocker effect to prevent postoperative AF. The class III influence on ventricular level is demonstrated by QT prolongation. The interpretation, however, is confounded by changes in HR. A widely used method to overcome this is correcting QT intervals for HRs according to certain formulas. The outcome depends on the predominant HRs of the groups and is under the influence of all factors that can alter HR. A method free of correction bias is to compare QT intervals in samples obtained at specified HRs (24,25). We applied such a method to analyze group differences and showed that sotalol prolonged QT intervals significantly more than metoprolol at each corresponding sinus HR level. This difference was not evident by using the traditional Bazett's formula. Application of Bazett's equation with β-blocker therapy has led to incorrect results (24). The tendency toward a larger QT prolongation at lower rates found in our study (Fig. 3) is in agreement with the reverse use-dependent effect of sotalol, as demonstrated with monophasic action-potential recording technique (26). The experimental evidence that class III antiarrhythmic properties are achieved only by relatively high doses of sotalol (27) is challenged by these clinical data.
In addition to HR, QT duration is profoundly influenced by autonomic nervous alterations, diurnal variations, catecholamines, plasma electrolyte levels, drugs, and some diseases. Of the relevant variables, the study groups differed during the treatment only in mean daily sinus rates, plasma creatinine levels (within reference range; <115 mM in our laboratory), and proportions of quinidine-treated patients to convert AF. Plasma creatinine levels do not directly affect repolarization, and plasma potassium or magnesium concentrations did not differ between the groups. Thus the class III effect of sotalol remains a likely explanation to the observed QT prolongation. Although this class III involvement in the outcomes of the previous studies has been proposed, the authors have not reported QT and QTc durations or their changes.
Our study demonstrated the class III effect as a prolongation of ventricular repolarization even in the lowdose sotalol treatment. However, because QT and QTc durations, their changes, or QT-HR plots did not differ between the patients with or without AF, the role of the class III effect in the prevention of postoperative arrhythmias remains unclear. This study cannot provide an estimate of atrial refractoriness, the actual determinant of AF prevention. Similarly, the HRs did not differentiate the patients with AF. Based on these facts, the contribution of the class III effect or HR on the more efficient AF prevention by sotalol is suggested only indirectly.
Predictors of atrial fibrillation
Older age, randomization to metoprolol treatment, high postoperative plasma magnesium concentration, and large preoperative heart volume were shown independently to increase the susceptibility to AF. Age as a risk factor was the most uniform finding in a number of previous studies (3,28,29). Enlarged cardiac volume to predict postoperative AF probably reflects more advanced coronary artery disease and denotes that structural factors in addition to catecholamine overload are related to the pathogenesis of postoperative AF.
Postoperative decline in plasma magnesium concentration is well recognized and even considered to provoke AF (30,31). In contrast, we reported that in magnesium-treated patients, supraphysiologic concentrations after CABG were related with increased AF incidence (3). Systematic use of β-blockade, attenuating the sympathetic stimulation-induced hypomagnesemia (32), and lack of deep hypomagnesemia (values below reference range) could account for the unexpected observation. The effect of magnesium on the development of postoperative AF was beyond the scope of this study, but it may be hypothesized that its role is delineated by the depth of hypomagnesemia and by the intensity of sympathetic drive.
The open design of the study makes possible a bias in the collection and identification of arrhythmic events. Another concern arises from the arrhythmia documentation that was performed only in patients whose arrhythmia lasted long enough to be recorded. Short periods of AFs might have been missed. Consequently some caution is warranted in the interpretation of the results.
This study shows that sotalol in low doses significantly reduces the incidence of AF compared with a standard β-blocker shortly after CABG surgery. An obvious class III effect, prolongation in ventricular repolarization, together with a more efficient HR slowing were demonstrated. Whether these mechanisms underlie the enhanced antifibrillatory effect could not be explicitly verified.
Acknowledgment: We thank Dr. Matti Viitasalo, M.D., for valuable criticism during the preparation of the manuscript.
1. Lauer MS, Eagle KA, Buckley MJ, DeSanctis W. Atrial fibrillation following coronary artery bypass surgery. Prog Cardiovasc Dis
2. Andrews TC, Reimold SC, Berlin JA, Antman EM. Prevention of supraventricular arrhythmias after coronary artery bypass surgery: a meta-analysis of randomized control trials. Circulation
3. Parikka H, Toivonen L, Pellinen T, Verkkala K, Järvinen A, Nieminen MS. The influence of intravenous magnesium sulphate on the occurrence of atrial fibrillation after coronary artery bypass operation. Eur Heart J
4. Reed HL, Chernow B, Lake CR, et al. Alterations in sympathetic nervous system activity with intraoperative hypothermia during coronary artery bypass surgery. Chest
5. Salazar C, Frishman W, Friedman S, et al. Beta-blockade therapy for supraventricular tachyarrhythmias after coronary surgery: a propranolol withdrawal syndrome? Angiology
6. Kalman JM, Munawar M, Yapanis A, et al. Atrial tachyarrhythmias following coronary bypass surgery: sympathetic mechanisms. In: Kingma JH, van Hemel NM, Lie KI, eds. Atrial fibrillation: a treatable disease?
Dordrecht: Kluwer Academic Publishers, 1992:211-25.
7. Csicsko JF, Schatzlein MH, King RD. Immediate post-operative digitalization in the prophylaxis of supraventricular arrhythmias following coronary artery bypass. J Thorac Cardiovasc Surg
8. Butler J, Harriss DR, Sinclair M, Westaby S. Amiodarone prophylaxis for tachyarrhythmias after coronary artery surgery: a randomized, double blind, placebo controlled trial. Br Heart J
9. Janssen J, Loomans L, Harink J, et al. Prevention and treatment of supraventricular tachycardia shortly after coronary bypass grafting: a randomized open trial. Angiology
10. Suttorp MJ, Kingma JH, Tjon Joe Gin RM, et al. Low and high dose sotalol versus propranolol in prevention of supraventricular tachyarrhythmias early after coronary artery bypass surgery. J Thorac Cardiovasc Surg
11. Suttorp MJ, Kingma JH, Peels HOJ, et al. Effectiveness of sotalol in preventing supraventricular tachyarrhythmias shortly after coronary artery bypass grafting
. Am J Cardiol
12. Singh BN, Phil D. Sotalol: a beta blocker with unique antiarrhythmic properties. Am Heart J
13. Bazett HC. An analysis of time relations of electrocardiogram. Heart
14. Raphael MJ, Donaldson RM. The normal heart: methods of examination. In: Sutton D, ed. A textbook of radiology and imaging.
4th ed. Edinburgh: Churchill Livingstone, 1987:538-57.
15. Machin D, Campbell MJ. Statistical tables for design of clinical trials.
Oxford: Blackwell Scientific Publications, 1987:10-34.
16. Altman DG, Gardner MJ. Calculating confidence intervals for regression and correlation. Br Med J
17. Daudon P, Corcos T, Gandjbakhch I, Levasseur J-P, Cabrol A, Cabrol C. Prevention of atrial fibrillation or flutter by acebutolol after coronary bypass grafting. Am J Cardiol
18. Lamb RK, Prabhakar G, Thorpe JAC, Smith S, Norton R, Dyde JA. The use of atenolol in the prevention of supraventricular arrhythmias following coronary artery surgery. Eur Heart J
19. Khuri SF, Okike OS, Josa M, et al. Efficacy of nadolol in preventing supraventricular tachycardia after coronary artery bypass grafting
. Am J Cardiol
20. Silverman NA, Wright R, Levitsky S. Efficacy of low-dose propranolol in preventing postoperative supraventricular tachyarrhythmias: a prospective, randomized study. Ann Surg
21. Vecht RJ, Nicolaides EP, Ikweuke JK, Liassides C, Cleary J, Cooper WB. Incidence and prevention of supraventricular tachyarrhythmias after coronary bypass surgery. Int J Cardiol
22. Cruickshank JM, Prichard BNC, eds. Pharmacodynamics. Beta-blockers in clinical practice.
Edinburgh: Churchill Livingstone, 1987:9-176.
23. Kato R, Ikeda N, Yabek SM, Kannan R, Singh BN. Electrophysiologic effects of the levo- and dextrorotatory isomers of sotalol in isolated cardiac muscle and their in vivo pharmacokinetics. J Am Coll Cardiol
24. Viitasalo M, Karjalainen J. QT intervals at heart rates from 50 to 120 beats per minute during 24-hour electrocardiographic recordings in 100 healthy men: effect of atenolol. Circulation
25. Karjalainen J, Viitasalo M, Mänttäri M, Manninen V. Relation between QT intervals and heart rates from 40 to 120 beats/min in rest electrocardiograms of men and a simple method to adjust QT interval values. J Am Coll Cardiol
26. Schmitt C, Brachmann J, Karch M, et al. Reverse use-dependent effect of sotalol demonstrated by recording monophasic action potentials of the right ventricle. Am J Cardiol
27. Nattel S, Feder-Elituv R, Matthews C, et al. Concentration dependence of class III and beta-adrenergic blocking effects of sotalol in anesthetized dogs. J Am Coll Cardiol
28. Fuller JA, Adams GG, Buxton B. Atrial fibrillation after coronary artery bypass grafting
: is it a disorder of the elderly? J Thorac Cardiovasc Surg
29. Leitch JW, Thomson D, Baird DK, Harris PJ. The importance of age as a predictor of atrial fibrillation and flutter after coronary artery bypass grafting
. J Thorac Cardiovasc Surg
30. Fanning WJ, Thomas CS Jr, Roach A, Tomichek R, Alford WC, Stoney WS Jr. Prophylaxis of atrial fibrillation with magnesium sulfate after coronary artery bypass grafting
. Ann Thorac Surg
31. England MR, Gordon G, Salem M, Chernow B. Magnesium administration and dysrhythmias after cardiac surgery: a placebo-controlled, double-blind, randomized trial. JAMA
32. Flink EB, Brick JE, Shane SR. Alterations of long-chain free fatty acid and magnesium concentrations in acute myocardial infarction. Arch Intern Med