Efficacy of intravenous magnesium for the management of non‐post operative atrial fibrillation with rapid ventricular response: A systematic review and meta‐analysis

Intravenous magnesium (IV Mg), a commonly utilized therapeutic agent in the management of atrial fibrillation (AF) with rapid ventricular response, is thought to exert its influence via its effect on cellular automaticity and prolongation of atrial and atrioventricular nodal refractoriness thus reducing ventricular rate. We sought to undertake a systematic review and meta‐analysis of the effectiveness of IV Mg versus placebo in addition to standard pharmacotherapy in the rate and rhythm control of AF in the nonpostoperative patient cohort given that randomized control trials (RCTs) have shown conflicting results.


| INTRODUCTION
Atrial fibrillation (AF) is the most common cardiac arrhythmia encountered in clinical practice and contributes substantially to patient morbidity and mortality. 1,2 AF with rapid ventricular response (RVR), defined as a heart rate (HR) of >110 beats per min without signs of hemodynamic compromise, 3 left untreated can substantially worsen patient outcomes leading to increased rates of morbidity with stroke, rate-related ischemia, cardiomyopathy, and may potential precipitate cardiogenic shock. 4,5 Despite guidelines, significant variation in clinical practice exists regarding rate and rhythm control strategies often utilizing betablockers, nondihydropyridine calcium channel antagonists, digoxin, amiodarone, intravenous magnesium (IV Mg), cardioversion, or a combination of these therapies. 3 Due to its low cost and favorable side-effect profile with minimal negative inotropy, IV Mg has gained widespread use. 6,7 It is believed to exert its effect on ventricular rate via the reduction of cellular automaticity and prolongation of atrial and atrioventricular nodal refractory period via its effect on potassium channels and intracellular calcium accumulation. 8 Nevertheless, a limited number of previous clinical trials have shown conflicting or inconclusive results. 9,10 Recently, further randomized control trials (RCTs) have been published 11,12 increasing the patient cohort by 29% and utilizing more contemporary medical therapy.
In a recently published RCT by our group, the use of IV Mg in a contemporary cohort of patients found no additional benefit in achieving rate or rhythm control compared to the standard of care. 12 In contrast, other studies have found IV Mg effective in rate control early in the care of patients with AF RVR. 11, 13 We, therefore, sought to perform a meta-analysis on placebocontrolled, randomized clinical trials conducted in patients with nonpostoperative AF with RVR examining the efficacy of IV Mg in addition to study-defined standard of care for both rate and rhythm control.

| Data sources and search strategy
An electronic literature search was undertaken until the end of August 2022. PubMed, MEDLINE, EMBASE databases, and Google Scholar were searched with no restriction on publication date or language. Searches were performed using Medical Subject Heading (MeSH) and keywords that included but were not limited to; "magnesium," "magnesium sulfate," "intravenous magnesium," "atrial fibrillation," "atrial fibrillation with rapid ventricular response" and "rapid atrial fibrillation." Example search strategies are provided in the Supporting Information: Material Table S1. References of reviewed articles were also screened to identify further relevant studies.
This review was registered in the PROSPERO database (ID: CRD42033414179).

| Study selection and data extraction
Study inclusion criteria were as follows; 1. RCT; 2. Studies comparing IV Mg to placebo in addition to standard of care 3. The primary endpoint of HR reduction and/or reversion to sinus rhythm. Studies were excluded if magnesium was not compared with placebo (i.e., compared directly to antiarrhythmic agents or electrical cardioversion), the setting for AF was postoperative, rate and rhythm control outcomes were not reported, or studies were not performed with human subjects. Two authors (A. E. and J. G.) independently used the search and selection criteria to identify trials for inclusion. All citations were first screened at the title and abstract level to determine suitability for inclusion with full-text articles retrieved and reviewed with suitable trials identified for inclusion. Disagreement was resolved through discussion and consensus with the senior author (J. N.). The risk of bias for each trial included in the analysis was assessed by two authors (A. E. and J. G.) using the Cochrane Collaboration Assessment Tool (Supporting Information: Figure S2).

| Outcome assessment
The primary efficacy outcome was successful rate control and/or reversion to sinus rhythm/rhythm control as defined by the individual trial. The secondary safety endpoint was patient experienced side effects. Outcomes were independently reviewed by 2 authors (A. E. and J. G.), and any discrepancies were resolved by consensus with the senior author (J. N.).

| Statistical analysis
Descriptive statistics are presented as mean with standard deviation for continuous variables and absolute percentages for categorical variables. Data were analyzed utilizing a random-effects model for the primary and secondary outcomes. Summary statistics are reported as pooled odds ratios with 95% confidence intervals (CI).
Statistical heterogeneity was quantified using the I 2 statistic, where I 2 ≥ 80% was considered significant inter-study heterogeneity.

| RESULTS
Of the 76 studies returned in the search results, 9 studies were included in the final analysis as shown in Figure 1. [11][12][13][14][15][16][17][18][19] All studies except one were published in the English language. 11 In brief, these 9 one study was triple-blinded, seven were double-blinded and in one study the details of blinding were unknown. The dose of IV Mg varied between studies from 2.5 to 10 g with 5 g being the most common dose administered. One trial included two treatment arms of a lower and higher dose administering 4.5 and 9 g, respectively. 13 The definition of rate control varied between studies and ranged from <90 to <110 and some studies included relative reductions in HR from a baseline of ≥20%. Four studies examined both rate and rhythm control and 5 rate control alone with time to the measurement of outcome ranging between 2 and 24 h. An analysis of publication bias (Supporting Information: Figure S1) showed a symmetrical distribution of most studies (with the exception of 1 trial) in the funnel plots for rate and rhythm control, suggesting there is no significant publication bias.
Baseline patient characteristics are summarised in Table 2. Serum magnesium levels were reported in five studies with all studies having mean levels within the normal range. Only three studies reported serum potassium levels, with mean levels within normal limits.
Cardiovascular risk factors were infrequently reported with hypertension being the most frequently reported risk factor. The standard of care, timing, and therapy utilization are summarised in Table 3.
Rate and rhythm control agents used varied between studies and were infrequently reported. Most studies utilized digoxin, beta-blockers, calcium channel blockers, and amiodarone with digoxin being the most commonly used agent. Of the nine trials included in the analysis, five studies enforced an initial lockout period where medication administration was not allowed. This varied between studies ranging from initial medical contact to a maximum of 2 h.

| Efficacy outcomes
The primary efficacy outcome of ventricular rate control was significantly higher in patients receiving IV Mg compared to placebo (odd ratio [OR] 1.87, 95% CI 1.13−3.11, p = .02). Similarly, reversion to sinus rhythm/rhythm control was more common in patients receiving IV Mg compared to placebo (OR 1.45, 95% CI 1.04−2.03, p = .03, I 2 = 24%) as shown in Figure 2.
A sensitivity analysis of studies with an initial lockout period revealed no difference in achieving rate (OR 0.97, 95% CI 0.59−1.59, p = .89, I 2 = 0%) or rhythm (OR 1.07, 95% CI 0.68−1.70, p = .77, I 2 = 37%) control with the use of IV Mg in addition to standard care as seen in Figure 3. For studies without a lockout period, there was statistically significant greater odds of rate (OR 2.56, 95% CI 1.86−3.51, p < .01, I 2 = 20%) and rhythm (OR 2.03, 95% CI 1.23−3.35, p = .005, I 2 = 0%) control with IV Mg use in addition to standard of care, as seen in Figure 4.

| Safety outcomes
Magnesium was very well tolerated among study participants with safety outcomes and side effects reported in five of the nine trials. In this analysis, IV Mg appears to have a greater effect on the likelihood of rate control compared to rhythm control (OR 1.87 vs. 1.45) albeit with greater study heterogeneity in the former (I 2 = 59% vs. I 2 = 24%). It is recognized that spontaneous reversion is common in the first 48 h of new-onset paroxysmal AF. 21 Whilst it is unclear in our study population, an uneven distribution toward chronic AF, which is associated with impaired rhythm control, 22  This is of particular relevance given the relative decline in digoxin utilization for AF management in the contemporary era. 25  Early rhythm control may be more effective in the long term with lower risks of adverse cardiovascular outcomes. 32 Similar to the mechanism of action of amiodarone, magnesium has inhibitory effects on the potassium rectifier channels and may aid in cardioversion 33,34 and this should be considered when administered.
Although the rate of IV infusion influences the side effects experienced, 35  not observe a significant difference between those receiving magnesium versus placebo, and therefore prescribers should feel confident that this will not be a significant clinical concern.

| LIMITATIONS
The paucity of patient baseline characteristics reported between studies, including cardiovascular risk factors must be considered. It is well established that a history of congestive cardiac failure, atrial size on echocardiography, and whether AF is chronic, acute, permanent or paroxysmal contributes to the likelihood of reversion or rate control.
Only two studies described whether patients had a history of chronic AF 12,13 and one study excluded patients with known or chronic AF. 14 Studies involving more selected patient cohorts are required to evaluate whether more specific subgroups of patients may be more or less likely to benefit from IV Mg therapy.
It is known the magnesium is a cofactor for the sodiumpotassium ATPase pump and hypomagnesemia may result in disturbances of intracellular potassium and calcium. Unfortunately, baseline magnesium and potassium levels were not adequately reported. Out of nine studies, four did not report magnesium concentrations and only two trials reported potassium concentrations. Due to limited data, it was not possible to perform an appropriately powered sub-group analysis to evaluate differences in outcomes stratified by serum magnesium and potassium levels.
There was considerable variability in the standard of care between RCTs. Digoxin was administered in six of the nine RCTs as the predominant form of rate control. Contemporary practice often utilizes digoxin as a second line following the administration of betaor calcium channel blockers, or in those with decompensated heart failure or left ventricular dysfunction given the positive inotropic properties. 3 The high rate of digoxin use likely reflects clinician preference or standard of care at the time of the RCT. In an attempt to define the isolated effects of Mg, many studies utilized a lockout period. Whether IV Mg improves rate and rhythm control with specific antiarrhythmics in the management of AF with RVR is beyond the scope of this study.

| CONCLUSION
IV Mg increases the likelihood of reversion to sinus rhythm and showed an increased likelihood of achieving study-defined rate control. IV Mg is a safe and well-tolerated treatment and represents a reasonable adjunct to standard medical therapy of patients presenting with AF with RVR in a nonpostoperative setting.