• 
• 

Three Part Question

In a patient with [recent onset atrial fibrillation], is [magnesium sulphate effective] in achieving [control of ventricular rate]?

Clinical Scenario

A 72 year old woman attends the Emergency Department with newly diagnosed atrial fibrillation of 48 hours duration. You decide to treat her by controlling her ventricular rate and wonder whether magnesium sulphate is a suitable agent.

Search Strategy

Medline 1966- 10/2004 using Ovid interface.
[(exp atrial fibrillation OR atrial fibrillation.mp] AND (exp magnesium OR magnesium.mp)]

Search Outcome

The search produced 133 papers of which 8 were relevant and of sufficient quality.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
DeCarli 1986 USA	45 consecutive patients with symptomatic atrial fibrillation of varying duration, receiving intravenous digoxin to control VR.	Prospective, non-randomised blinded study.	Dose of digoxin required to achieve control of VR in magnesium deficient and non-deficient patients.	Magnesium deficient patients required twice the dose of intravenous digoxin to achieve control of VR.	Small study Not acute af but symptomatic patients requiring VR control.
Gullestad 1993 Norway	35 patients with stable AF of <1/52 duration (rate >100/min) given MgSO4 (2x 5mmol boluses followed by 0.4mmol/min infusion) or verapamil (2x 5mg boluses followed by 0.1mg/min infusion). The infusions continued for 1 hour after cardioversion to a maximum of 24 hours.	Unblinded randomised trial.	Conversion rate @ 4 & 24 hours.	No significant difference in cardioversion rate @ 24 hours (53% Mg vs. 40% verapamil) but Mg was quicker (47% @ 4 hours vs. 5% for verapamil).	Small study Short follow up (4 hours). Did not study long-term morbidity and mortality. Not an ED population.
			Proportion with rate <100/ min @ 4 hours.	Verapamil significantly better than Mg at reducing HR @ 4 hours (65% vs. 27%).
			Side-effects reported.	Significantly more side-effects were noted with verapamil (25% hypotension; 15% CCF) vs. no side-effects for Mg.
Hays 1994 USA	15 patients presenting to an ED with newly diagnosed stable AF (unknown duration) with HR >99/min., given either Mg (2g bolus followed by 4g infusion over 4 hours) or placebo. At 30 mins. All patients were given 0.5mg digoxin over 30 mins.	Double-blinded trial.	Heart rate @ 30 mins & 4 hours.	In the 1st 30 mins. Mg significantly better than placebo at reducing HR (16 +/- 7% vs. 2 +/- 5%) (p<0.02). No significant difference in HR reduction @ 4 hours between digoxin alone (18 +/- 10%) or digoxin-Mg (26 +/- 7%) (p=0.08).	Small study. Very brief follow up. Not clear if randomised.
			Cardioversion rate @ 4 hours.	There was a non-statistically significant difference in cardioversion rate between digoxin alone (37.5%) and digoxin-Mg (14.2%).
Brodsky 1994 USA	18 outpatients with stable AF <7/7 duration (100-200 /min), all on standardised digoxin regime, given either Mg (2g bolus followed by 8g infusion over 6 hours) or placebo.	Randomised controlled trial.	Proportion with HR <100/min. within 24 hours	100% Mg-digoxin had HR <90/min. after 24 hours vs. 50% with digoxin alone (p<0.05)	Small study. Short follow up. Not an ED population.
			Cardioversion rate within 24 hours.	No significant difference in cardioversion rate between groups (60% Mg-digoxin group vs. 38% digoxin alone)
Moran et al 1995 Australia	42 ICU patients with an atrial tachyarrhythmia with VR >120bpm of > 1 hour duration and K+>4mmol/l. Patients treated with either an initial bolus of magnesium (0.037g/kg) fb an infusion (0.025g/kg/hr) adjusted to maintain serum magnesium levels between 1.5-2.0mmol/l or a bolus of amiodarone (5mg/kg) fb infusion of 10mg/kg/24hrs Excluded if systolic bp <80mmHg or acute renal failure	Prospective block randomised clinical trial of non-consecutive patients	Conversion of tachyarrhythmia to SR	21 patients randomised to magnesium group, 71% of whom had af/21 patients randomised to amiodarone group, 52% of whom had af. 67% patients in Mg group, and 33% in amiodarone group converted to SR. This was significantly different from 2 hours. For those patients who did not convert, both magnesium and amiodarone significantly reduced VR by mean 19 bpm (p=0.001) within 0.5hrs.	Non-consecutive trial with no power calculation and small numbers. Mixture of ICU patients with tachyarrhythmias other than af therefore limited applicability to ED. No control group for spontaneous conversion to SR.
			Control of VR	There was no significant difference between the groups in control of VR
			Effect on systolic bp	There was no significant reduction in bp in either group
Eray et al 2000 Turkey	19 patients with af > 3 days duration and uncontrolled VR >120 bpm were given 2g MgSO4 iv bolus fb 1g/hr for 6 hours. Magnesium deficiency assessed by measurement of 24 hour magnesium excretion in urine. Iv diltiazem given at end of 1st hour if VR still uncontrolled Excluded if low sats/ bp orthopnoea/tachypnoea and renal failure	Prospective, unblinded, non-randomised	Control of VR in magnesium deficient and non-deficient patients.	VR decreased significantly compared to baseline at 15,30 and 60 minutes. No difference between magnesium deficient and non-deficient patients.	Small numbers Effect of magnesium only studied over 1 hour Chronic af, >3 days duration but did require control of VR
Chiladakis 2001 Greece	46 patients admitted to a cardiology department with stable paroxysmal AF <12 hours duration and HR >100/min. They were given either Mg (2.5g bolus followed by 7.5g over 6 hours) or diltiazem (25mg bolus, 12.5mg/hour over 6 hours)	Randomised trial	HR hourly	Nearly identical significant reduction in HR during both treatment regimes: pre-treatment to 1 hour (p<0.05) and 3 hours (p<0.001).	Small study. Not ED population but not dissimilar in characteristics. Not placebo controlled (it would have been interesting to know the spontaneous cardioversion rate in the group though not particularly ethical ). iv diltiazem not licensed in the UK.
			BP every 15mins.	No changes in BP with either treatment
			Cardioversion rate.	Significantly more patients cardioverted with Mg than with diltiazem (57% vs. 22% p=0.03).
			LV ejection fraction post cardioversion.	Same LVEF for both groups (Mg 59.6 +/- 8.8%; diltiazem 59.2 +/- 10%)
Cybulski 2004 Poland	22 patients with af <24 hours duration with either low serum K+ or Mg++ levels compared to 31 patients with normal serum electrolytes, both groups receiving infusion of 1000ml of 10% dextrose, 10U of short acting insulin, 80mmol K+ and 8g of MgSO4	Unblinded non-randomised study	Conversion to SR	19% of 115 consecutive patients with acute onset af had deficiency of K+, Mg++ or both.	Small non-randomised trial where there is no information about the likelihood of spontaneous conversion to SR Patients selected as part of alternative trial Unclear how rapidly infusion given
			Number of patients presenting with af<24 hours duration with K+ or Mg++ deficiency	Those patients with K+/Mg++ deficiency were more likely to revert to SR (86% vs. 39% p<0.001) than those without any deficiency

Comment(s)

Digoxin: Hayes. Only paper comparing Mg with placebo (& only for 30 mins) Mg (2g) better than placebo at reducing HR in the 1st 30 mins of treatment of stable AF. Mg (1g/hour) no better than Mg-digoxin at reducing HR @ 4 hours and both poor at cardioverting. Brodsky. Mg (2g bolus followed by 8g over 6 hours) + digoxin significantly better than digoxin alone at reducing HR at 24 hours (100% vs. 50%) but no better at cardioverting AF <7/7 duration. Calcium Channel Blockers: Gullestad Mg (10mmol bolus + 0.4mmol/min) as good as verapamil (10 mg bolus + 0.1mg/min) at cardioversion but verapamil better at reducing HR in patients with AF <1/52, though with higher incidence of symptomatic hypotension and worsening of CCF. Chiladakis Mg (10g over 6 hours) better than diltiazem (100mg over 6 hours) at cardioverting acute paroxysmal AF, but both equally effective at HR reduction with minimal side-effects. Magnesium and Amiodarone: Moran In an ITU population, with only a limited percentage of patients having atrial fibrillation, magnesium superior to amiodarone at achieving SR and as effective as amiodarone for controlling ventricular rate in those patients who didn't convert to SR. Electrolyte Disorders and AF: Cybulski Magnesium and potassium deficiency is common in new onset af and supplementation is suggested to convert these patients to SR though importantly there was no control group. De Carli Magnesium deficiency may be common in patients with symptomatic patients with af and that those patients who were deficient required larger doses of digoxin to control VR. Eray Magnesium infusion significantly decreased VR over 1st hour but no control group and no significant difference between Mg deficient and Mg replete patients as measured by urinary excretion of Mg.

Clinical Bottom Line

All these studies were small and have moderate quality. There is little good evidence to suggest that magnesium will cardiovert patients with atrial fibrillation. Placebo controlled trials show no effect of magnesium on cardioversion rates, and those trials that suggest improved cardioversion rates with magnesium do not report spontaneous cardioversion rates. Magnesium, alone or in combination with other drugs, may have a limited role in achieving control of ventricular rate especially in magnesium deficient patients.

References

1. DeCarli C, Sprouse G, LaRosa JC. Serum magnesium levels in symptomatic atrial fibrillation and their relation to rhythm control by intravenous digoxin. Am J Cardiol 1986; 57(11):956-9.
2. Gullestad L, Birkeland K, Molstad P, Hoyer MM, Vanberg P, and Kiekshus J. The effect of magnesium versus verapamil on supraventricular arrhythmias. Clin Cardiol. 1993; 16; 429-434.
3. Hays JV, Gilman JK and Rubal BJ. Effect of magnesium sulfate on ventricular rate control in atrial fibrillation. Ann Emerg Med. 1994; 24; 61-64.
4. Brodsky MA, Orlov MV, Capparelli EV, Allen BJ, Iseri LT, Ginkel M, and Orlov YSK. Magnesium therapy in new-onset atrial fibrillation. Am J Cardiol. 1994; 73; 1227-1229.
5. Moran JL, Gallagher J, Peake SL, Cunningham DN, Salagaras M, Leppard P. Parenteral magnesium sulfate versus amiodarone in the therapy of atrial tachyarrhythmias: a prospective randomized study. Crit Care Med 1995;23(11):1816-1824.
6. Eray O, Akca S, Pekdemir M, Eray E, Cete Y and Oktay C. Magnesium efficacy in magnesium deficient and nondeficient patients with rapid ventricular response atrial fibrillation. Eur J Emerg Med 2000;7:287-90.
7. Chiladakis JA, Stathopoulos C, Davlouros P, and Manolis AS. Intravenous magnesium sulfate versus diltiazem in paroxysmal atrial fibrillation. Intern J Cardiol. 2001; 79; 287-291.
8. Cybulski J, Budaj A, Danielewicz H, Maciejewicz J, Ceremuzynski L. A new-onset atrial fibrillation: the incidence of potassium and magnesium deficiency. The efficacy of intravenous potassium/magnesium supplementation in cardioversion to SR. Polish Heart Journal 2004;60(6): 578-81.