Magnesium in brachial plexus blocks: A meta-analysis, meta regression, and trial sequential analysis

Abstract

Background and Aims:

Evidence is inconclusive regarding the role of magnesium sulphate (MgSO₄) on the onset of sensory block in brachial plexus blocks (BPBs). A meta-analysis was conducted to assess the impact of adding MgSO₄ to local anaesthetics (LA) used in BPBs on the onset time of sensory block.

Methods:

Medline, Embase, Google Scholar, and bibliography searches were consulted to identify randomised clinical trials using MgSO₄ versus control in BPBs. The primary outcome was the onset of sensory block. The secondary outcomes measured were duration of sensory and motor block, as well as the duration of analgesia. Continuous data were expressed as standard mean difference (SMD) with 95% confidence intervals (CI) and analysed using a random-effects model. Subgroup analysis was also performed based on magnesium dose. Study quality was assessed using a JADAD scoring system.

Results:

Thirty-two studies were included in this study. MgSO₄ led to a quicker onset of sensory block (n = 1768; N = 28; SMD = −1.00; 95% CI: −1.46, −0.55; P < 0.0001). MgSO₄ also led to a quicker onset of motor block (n = 1768; N = 28; SMD = −1.15; 95% CI: −1.62, −0.68; P < 0.00001), and prolonged duration of analgesic effect (n = 1368; N = 21; SMD = 4.22; 95% CI: 3.36, 5.08; P < 0.00001). Subgroup analysis showed that MgSO₄ doses of >200 mg compared to ≤200 mg prolonged the duration of sensory block (MD = 179.00 min vs 144.31 min) and motor block (MD = 145.26 min vs 116.52 min). It also showed that MgSO₄ >200 mg led to a longer duration of analgesia compared to MgSO₄ ≤200 mg (MD = 231.65 min vs 173.58 min).

Conclusion:

Using MgSO₄ as an adjunct to LA in BPBs leads to a faster onset of sensory block. MgSO₄ also prolongs the duration of analgesia, with optimal doses likely being ≤200 mg.

INTRODUCTION

Peripheral nerve blocks are a routine part of current anaesthetic practice, with evidence demonstrating a reduced stress response, earlier mobilisation, and high patient satisfaction.[1,2] Regional anaesthesia for the upper limb can be provided safely and effectively using brachial plexus blocks (BPBs).[1] To increase the speed of onset and prolong the duration of action, adjuvants have been used with local anaesthetics (LA) to varying degrees of success.[2]

A commonly used adjuvant in peripheral nerve blocks is magnesium sulphate (MgSO₄), which elicits an anti-nociceptive effect primarily through the antagonism of N-methyl-D-aspartate receptor ion channels, modulation of calcium influx, and activation of nitric oxide pathways.[3] Adjuvants can speed the onset of sensory block, thereby improving theatre efficiency and operating output.[4] For MgSO₄, however, robust safety data, the optimal dose, and its impact on the onset of sensory block are unclear.[5] Li et al.[6] and Zeng et al.[7] performed meta-analyses showing no statistical difference (P = 0.83 and 0.48, respectively) between the two groups (saline vs MgSO₄) when assessing the speed of onset of sensory block. However, subgroup analysis revealed a mixed picture: no difference in the speed of onset of sensory block for BPBs, but a faster onset with MgSO₄ for femoral and sciatic nerve blocks. Small sample sizes and significant heterogeneity hamper further analysis. The dose of magnesium used may influence this block parameter, with 250 mg of MgSO₄ resulting in a faster onset of sensory block after brachial plexus injection, but 125 mg resulting in no difference.[8]

We, therefore, performed a meta-analysis to evaluate and assess the impact of MgSO₄ supplementation on the rapidity of sensory block onset when added to LA agents during peripheral nerve blocks in patients undergoing regional anaesthesia for elective surgical procedures.

METHODS

The Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines were adhered to throughout this study.[9] The study protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) at the National Institute for Health Research (ID: CRD42023457706).

Information sources and search strategy

The bibliographic databases, namely MEDLINE, EMBASE, and Google Scholar, were searched. All studies reported between February 2006 and June 2025 were included. The keywords used as part of the search terms included ‘magnesium sulphate’, ‘magnesium sulfate’, ‘magnesium’, ‘brachial plexus’, ‘axillary’, ‘supraclavicular’, ‘infraclavicular’, ‘interscalene’, ‘block’, ‘motor’, and ‘sensory’. Where possible, the following filters were applied: ‘human’ and ‘adults’. The searches were limited to papers published in English only and available in full [Supplementary Figure 1 ^{(2.2MB, tif)} and Supplementary Tables 1 and 2]. Bibliography searches were conducted on all included studies to identify additional suitable studies. All studies compared the analgesic/anaesthetic effects of LA alone versus LA and MgSO₄. R.K. and S.P. separately screened the studies identified during the process. If there were any disagreements, B.A., M.R., and B.R. were consulted.

Supplementary Table 1.

Electronic search strategy for Google Scholar

Search Fields	Outcome
Search terms	(“magnesium sulphate” OR “magnesium sulfate” OR magnesium) AND (“brachial plexus” OR axillary OR supraclavicular OR infraclavicular OR interscalene) AND block AND (motor OR sensory)
Limitations	Year range: 2006-2025, sort by relevance
Where any words occurred	All in the title
Results	3

Supplementary Table 2.

Electronic search strategy for EMBASE

Search Fields	Outcome
Search terms	(“magnesium sulphate” OR “magnesium sulfate” OR magnesium) AND (“brachial plexus” OR axillary OR supraclavicular OR infraclavicular OR interscalene) AND block AND (motor OR sensory)
Limitations	Year range: 2006-2025, publication type: articles
Where any words occurred	Title, original title, and abstract
Results	33

Eligibility criteria

Studies that met the following criteria were included: complete articles that were published in English, RCTs that compared LA alone to LA and MgSO₄, compared BPBs only, and compared the analgesic effect of LA alone versus LA and MgSO₄.

Studies were excluded if patients received a general anaesthetic before or after a nerve block, intravenous magnesium was used, or other adjuncts were included within the block mix, to reduce confounding factors. Review articles, case reports, and animal studies were also excluded.

Selection of studies and data extraction

Participants of interest were adults over the age of 18 years who received BPBs with LA, undergoing a variety of surgical procedures. Interventions referred to the administration of MgSO₄ as an adjuvant to LA, with the dose used determined by the authors of the studies. The comparison was between LA and MgSO₄ versus LA alone. The primary outcome was the speed of onset of sensory block; secondary outcomes were efficacy parameters such as sensory and motor effects, duration or degree of analgesia, and adverse effects.

Two authors (R.K. and S.P.) reviewed the article titles and abstracts retrieved from the initial searches and excluded irrelevant studies. The remaining studies were then retrieved in full and assessed for suitability. If studies were not readily accessible for download, authors were not contacted, and the study would be excluded from analysis.

Data were independently collected by two authors, R.K. and S.P., using a standardised data collection form created in Microsoft Excel.[10] Discrepancies were resolved through review and discussion. If any discrepancies remained, B.A, R.M., and R.B. were consulted. The data extracted included the following items: name of the author, year published, number of patients in the study/relevant arms of the study, type of block performed, dose of magnesium, dose of LA, final volume of LA mixture, time of sensory and motor block onset, duration of sensory and motor block, and duration of analgesia. The results were recorded as mean ± standard deviation for continuous variables or dichotomous outcomes as events. If any results were recorded in hours, they were converted into minutes.

Risk-of-bias assessment

The JADAD score (1–5) provides a validated method of describing trial quality and was used to assess the included studies.[11] R.K. and S.P. scored each trial independently using the 5-point scoring system. The final scores were assigned by consensus in cases where discrepancies arose with the initial scores. A risk-of-bias assessment was conducted using the Cochrane Risk of Bias tool including, reviewing for sequence generation (selection bias), allocation sequence concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective outcome reporting (reporting bias), and other potential sources of bias.[12]

Outcomes

The endpoints in this analysis were as follows: (1) onset of sensory block, defined as the time interval between the administration of LA mix to the loss of sensation, as part of a pin prick test, as defined by authors; (2) duration of sensory block, defined as the time interval between the return to normal sensation from loss of sensation, as part of a pin prink test, as defined by authors; (3) onset of motor block defined as the time interval between the administration of LA mix to the loss of motor power as part of a modified Bromage score, as defined by authors; (4) duration of motor block, defined as the time interval between the return to normal motor power from the loss of motor power, as part of a modified Bromage score, as defined by authors; (5) duration of analgesia defined as the time interval between the onset of sensory block to the first request for analgesia, as defined by authors; (6) incidence of rescue analgesia defined as number of patients requesting for postoperative analgesia in the first 24 h. The corresponding effect on outcomes based on MgSO₄ doses of ≤200 mg and >200 mg was also assessed. Results were presented as the standard mean difference and the mean difference.

Statistical analysis

A pairwise meta-analysis was conducted to assess the (1) efficacy and (2) safety of MgSO₄. All data analyses were performed using Review Manager (Cochrane RevMan) (Version 9.6.0) software.[13] Continuous data were expressed as standard mean difference (SMD) with 95% confidence intervals (CI), due to inter-study variation in outcome measurement, and analysed using a random-effects model and an inverse variance method. The mean difference (MD) was also used, where relevant, to express continuous data in a clinically meaningful manner. Statistical significance was demonstrated if the 95% CI did not include 0. The odds ratio (OR) and 95% CI were calculated using a random-effects model and the Mantel–Haenszel (M–H) method for dichotomous variables. The results were deemed statistically significant if the 95% CI did not include 1.0. For all statistical tests, significance was set at P < 0.05. Clinical variation between studies resulted in variations in the results. To overcome this, heterogeneity was measured using the I² statistic. This value describes the level of variation in study findings that is due to differences in methodology, rather than to chance.

A dose of ≤200 mg or >200 mg of MgSO₄ was chosen for the subgroup analysis, as almost an equal number of studies fit into each subgroup. It was thus decided that this was the most suitable midpoint for analysis.

Secondary analysis, in the form of meta-regression, was performed using the Comprehensive Meta-analytic software.[14] to determine if heterogeneity was due to the amount of LA, the amount of MgSO₄, the JADAD score, and the method of block performance (peripheral nerve stimulator vs ultrasound vs peripheral nerve stimulator plus ultrasound). A cumulative analysis for the above variable was performed, including the ‘one study removed’ analysis, where necessary.

Publication bias was represented using a funnel plot. This graph illustrates the effect size on the x-axis and sample size on the y-axis. In this case, the sample size was reported as the log of the standard error of the effect size, where n represents the number of participants and N represents the number of trials. If small trials are inappropriately represented, the plot will appear asymmetrical. Further secondary analyses, such as Egger’s regression test, Begg and Mazumdar’s rank correlation tests, and precision modelling, including Duval and Tweedie’s ‘trim and fill effect’, were performed to find missing studies that might impart asymmetry to the funnel plot. These were done for the primary outcome, and where relevant, we created a scatter plot.

We performed a trial sequential analysis (TSA) by using TSA software (Version 0.9.5.10 Beta).[15] The Sidik-Jonkman random effects model was chosen to calculate the Z-statistic, as it is less likely to underestimate the heterogeneity between the included trials. The strength of the existing evidence can be derived from the risk of type-1 and type-2 statistical errors; the percentages were set at 5%, 20%, and 10%, respectively, to result in a power of 80% and 90%. To regulate the risk of a type-2 error, an extension of the Lan and DeMets alpha-spending function was used to adjust the threshold, which represents what is denoted as futility boundaries. The strength of the available evidence can be considered by determining the required information size (IS) for a conclusive and reliable meta-analysis. It can be derived from the risk of type-1 and type-2 statistical errors, which we set at 5% and 20%, respectively, resulting in a power of 80%. This was done solely for the primary outcome.

RESULTS

Database searches, bibliographic searches, and full-text screening identified 77 articles, of which 32 were included[8,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46] in this study for analysis. Six articles were excluded after being retrieved for review [Figure 1].[47,48,49,50,51,52]

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) flowchart

Not all studies defined their measured outcomes, such as the onset of motor block or the duration of motor block. Where definitions were provided, we used the respective authors’ definition for the various outcomes described in the included RCTs.

Appraisal of study quality

There were 32 included RCTs; 13 RCTs had a JADAD score of ≤2, and 19 RCTs had a JADAD score of ≥3 [Table 1]. Further robust statistical analysis was performed to seek confounders, given the JADAD score of <2 in 13 RCTs.

Table 1.

Characteristics of studies

Reference	Country of Origin	JADAD score	Number of patients		BPB type	Block method	Dose of LA	Local Anaesthetic Mix
			Control Group	MgSO4 Group				Control Group	MgSO4 Group
Aggarwal et al.[16]	India	2	40	40	Supraclavicular	LM	150 mg	30 mL 0.5% ropivacaine + 1 mL NaCl 0.9%	30 mL 0.5% ropivacaine + 150 mg MgSO4
Ashry et al.[17]	Egypt	3	30	30	Supraclavicular	USG	500 mg	25 mL 2% lignocaine + 5 mL NaCl 0.9%	25 mL 2% lignocaine + 250 mg MgSO4
Attri et al.[18]	India	4	30	30	Supraclavicular	PNS	225 mg	30 mL 0.75% ropivacaine	30 mL 0.75% ropivacaine + 150 mg MgSO4
Beiranvand et al.[19]	Iran	2	30	30	Infraclavicular	USG	360 mg	18 mL 2% lignocaine + 12 mL NaCl 0.9%	18 mL 2% lignocaine + 1000 mg MgSO4 + 10 mL NaCl 0.9%
Deshpande and Patil.[20]	India	5	30	30	Axillary	USG	150 mg	30 mL 0.5% ropivacaine + 1 mL NaCl 0.9%	30 mL 0.5% ropivacaine + 150 mg MgSO4
Dogru et al.[21] (1)	Turkey	3	20	20	Axillary	PNS	200 mg	40 mL 0.5% levobupivacaine + 1 mL 0.9% NaCl 0.9%	40 mL 0.5% levobupivacaine + 150 mg MgSO4
Dogru et al.[21] (2)	Turkey	3	20	20	Axillary	PNS	100 mg	40 mL 0.25% levobupivacaine + 1 mL NaCl 0.9%	40 mL 0.25% levobupivacaine + 150 mg MgSO4
Elhusein et al.[22]	Egypt	3	20	20	Supraclavicular	USG	140 mg	28 mL 0.5% bupivacaine + 2 mL NaCl 0.9%	28 mL 0.5% bupivacaine + 200 mg MgSO4
Elyazed and Mogahed.[23]	Egypt	5	35	35	Infraclavicular	USG + PNS	175 mg	35 mL 0.5% ropivacaine + 4 mL NaCl 0.9%	35 mL 0.5% ropivacaine + 150 mg MgSO4 + 2.5 mL NaCl 0.9%
Fahmy et al.[24]	Egypt	3	21	21	Interscalene	USG	100 mg	20 mL 0.5% bupivacaine + 5 mL NaCl 0.9%	20 mL 0.5% bupivacaine + 500 mg MgSO4
Ghali et al.[25]	Egypt	5	30	30	Supraclavicular	USG	115 mg	23 mL 0.5% bupivacaine + 2 mL NaCl 0.9%	23 mL 0.5% bupivacaine + 100 mg MgSO4
Gunduz et al.[26] (1)	Turkey	3	15	15	Axillary	LM + PNS	NA	5 mg/kg 2% prilocaine + NaCl 0.9% to 35 mL total volume	5 mg/kg 2% prilocaine + 100 mg MgSO4 + NaCl 0.9% to 35 mL total volume
Gunduz et al.[26] (2)	Turkey	3	15	15	Axillary	LM + PNS	NA	5 mg/kg 2% prilocaine + NaCl 0.9% to 35 mL total volume	5 mg/kg 2% prilocaine + 150 mg MgSO4 + NaCl 0.9% to 35 mL total volume
Güneş and Güneş.[45]	Turkey	5	30	30	Infraclavicular	USG + PNS	100 mg	20 mL 0.5% bupivacaine + 5 mL NaCL 0.9%	20 mL 0.5% bupivacaine + 1 mg MgSO4 + 4 mL NaCL 0.9%
Gupta et al.[27]	India	2	62	62	Supraclavicular	LM	120 mg	24 mL 0.5% ropivacaine + 6 mL NaCl 0.9%	24 mL 0.5% ropivacaine + 150 mg MgSO4+ 5.5 mL NaCl 0.9%
Haghighi et al.[28]	Iran	5	30	30	Axillary	LM	NA	5 mg/kg 2% lignocaine + 43 mL NaCl 0.9%	5 mg/kg 2% lignocaine + 600 mg MgSO4+ 40 mL NaCl 0.9%
Hassan and Abdelkareem.[29]	Egypt	1	20	20	Supraclavicular	USG	93.75 mg	25 mL 0.375% levobupivacaine	25 mL 0.375% levobupivacaine + 200 mg MgSO4
Hassan et al.[30]	Egypt	2	15	15	Supraclavicular	USG	300 mg	12 mL 2% lignocaine + 12 mL 0.5% bupivacaine + 1 mL NaCl 0.9%	12 mL 2% lignocaine + 12 mL 0.5% bupivacaine + 100 mg MgSO4
Karthik et al.[31]	India	4	25	25	Supraclavicular	USG	187.5 mg	25 mL 0.75% ropivacaine + 1 mL NaCl 0.9%	25 mL 0.75% ropivacaine + 250 mg MgSO4
Kaur et al.[32]	India	5	35	35	Supraclavicular	PNS	135 mg	27 mL 0.5% ropivacaine + 3 ml NaCl 0.9%	27 mL 0.5% ropivacaine + 250 mg MgSO4 + NaCl 0.9% to 30 mL total volume
Mandloi et al.[33]	India	2	30	30	Infraclavicular	PNS	NA	0.5% ropivacaine	0.5% ropivacaine + 2 mg/kg MgSO4
Mathur and Singh.[34]	India	1	30	30	Supraclavicular	N/A	150 mg	30 mL 0.5% ropivacaine + 0.5 mL NaCl 0.9%	30 mL 0.5% ropivacaine + 250 mg MgSO4
Metias et al.[35]	Egypt	2	12	12	Supraclavicular	USG	140 mg	28 mL 0.5% bupivacaine + 2 mL NaCl 0.9%	28 mL 0.5% bupivacaine + 200 mg MgSO4
Mukherjee et al.[36]	India	5	50	50	Supraclavicular	LM + PNS	150 mg	30 mL 0.5% ropivacaine	30 mL 0.5% ropivacaine + 150 mg MgSO4
Nahida et al.[37]	India	1	30	30	Supraclavicular	USG	176.25 mg	23.5 mL 0.75% ropivacaine + 1.5 mL NaCl 0.9%	23.5 mL 0.75% ropivacaine + 150 mg MgSO4
Patel et al.[38]	India	2	30	30	Supraclavicular	USG	150 mg	30 mL 0.5% ropivacaine + 1.5 mL NaCl 0.9%	30 mL 0.5% ropivacaine + 150 mg MgSO4
Piplai and Roy.[39]	India	2	40	40	Supraclavicular	PNS	112.5 mg	15 mL 0.75% ropivacaine + 25 mL of NaCl 0.9%	15 mL 0.75% ropivacaine + 1000 mg MgSO4 + 23 mL NaCl 0.9%
Rao et al.[40]	India	5	30	30	Supraclavicular	PNS	NA	1.5 mg/kg 0.5% bupivacaine + 43 mL NaCl 0.9%	1.5 mg/kg 0.5% bupivacaine + 600 mg MgSO4 + 40 mL NaCl 0.9%
Sadafule et al.[41]	India	4	30	30	Supraclavicular	USG	140 mg	28 mL 0.5% bupivacaine + 2 mL NaCl 0.9%	28 mL 0.5% bupivacaine + 250 mg MgSO4
Shridevi and Asokan.[42]	India	2	30	30	Supraclavicular	USG	176.25 mg	23.5 mL 0.75% ropivacaine + 1.5 mL NaCl 0.9%	23.5 mL 0.75% ropivacaine + 150 mg MgSO4
Shukla et al.[43]	India	4	19	19	Supraclavicular	USG	145 mg	29 mL 0.5% ropivacaine 1 mL NaCl 0.9%	29 mL 0.5% ropivacaine + 250 mg MgSO4
Taneja et al.[44]	India	2	30	30	Supraclavicular	LM	150 mg	30 mL 0.5% ropivacaine + 1 mL NaCl 0.9%	30 mL 0.5% ropivacaine + 150 mg MgSO4
Verma et al.[8] (1)	India	5	30	30	Supraclavicular	USG	100 mg	20 mL 0.5% bupivacaine + 5 mL NaCl 0.9%	20 mL 0.5% bupivacaine + 125 mg MgSO4+ 3.75 mL NaCl 0.9%
Verma et al.[8] (2)	India	5	30	30	Supraclavicular	USG	100 mg	20 mL 0.5% bupivacaine + 5 mL NaCl 0.9%	20 mL 0.5% bupivacaine + 250 mg MgSO4 + 2.5 mL NaCl 0.9%
Youssef et al.[46]	UAE	3	50	50	Supraclavicular	USG	145 mg	29 mL 0.5% ropivacaine + 1 mL NaCL 0.9%	29 mL 0.5% ropivacaine + 250 mg MgSO4

MgSO₄=Magnesium Sulphate; NaCl – Sodium Chloride; BPB=Brachial Plexus Block; LM=Landmark; USG=Ultrasound-Guided; PNS=Peripheral Nerve Stimulator; LA=Local Anaesthetic; (1)=study arm 1; (2)=study arm 2

Characteristics of studies

Thirty studies included patients with ASA status I–III. The remaining two did not mention the ASA status of their included patients.[37,42] There were no statistically significant differences in ASA status among the included RCTs. Additional characteristics of the studies are documented [Table 1].

Primary outcome: onset of sensory block

The pooled data showed that the use of MgSO₄ with LA quickens the onset of sensory block compared to LA alone, with a mean difference (MD) of 2.16 min faster vs LA alone (SMD = 1.00; 95% CI: −1.46, −0.55; P < 0.0001 (I² = 95%)) [Figure 2 and Table 2].

Figure 2.

Forest plot – onset of sensory block. The forest plot shows that the onset of sensory block was slower in the control group than when magnesium sulphate was used as an adjunct in BPBs. Std mean difference IV = standard mean difference–inverse variance; SD = standard deviation; CI = confidence interval

Table 2.

Primary outcome and subgroup analysis of the primary outcome

	Number of studies (participants)	Overall SMD [95% CI]	Overall MD [95% CI]	Subgroup analysis
				Number of studies (participants)	SMD [95% CI]		MD [95% CI]
					MgSO4 dose ≤200 mg	MgSO4 dose >200 mg	MgSO4 dose ≤200 mg	MgSO4 dose >200 mg
Onset of sensory block	28 (1768) [8,16–18, 20–25,27–29,31–35,37–46]	−1.00 [-1.46, -0.55] (P<0.0001)	-2.16 [-3.02, -1.29] (P<0.00001)	27 (1708) [8,16–18,20–25, 27–29,31,32,34,35,37–46]	-1.05 [-1.50, -0.59] (P<0.00001)	-1.06 [-2.04, -0.08] (P=0.03)	-1.68 [-2.33, -1.03] (P<0.00001)	-2.93 [-5.50, -0.37] (P=0.03)

SMD=standard mean difference; MD=mean difference; CI=confidence interval

The MgSO₄ dosage (≤200 mg and >200 mg) subgroup analysis showed that MgSO₄ quickens the onset of sensory block, compared to LA alone (SMD = −1.05; 95% CI: −1.52, −0.58; P < 0.0001) [Figure 3]. There were no subgroup differences observed (P = 0.98, I² = 0%). Sensitivity analysis, excluding outliers based on effect size and wide confidence intervals, did not alter the P value or the direction of the overall outcome. Doses of MgSO₄ ≤200 mg resulted in a quicker onset to sensory block versus LA alone (SMD = −1.05; 95% CI: −1.50, −0.59; P < 0.00001), with an even faster onset for MgSO₄ doses of >200 mg vs LA alone (SMD = −1.06; 95% CI: −2.04, −0.08; P = 0.03). Subgroup analysis was also done for the type of LA used, where ropivacaine (SMD = −1.20; 95% CI: −1.81, −0.59; P = 0.0001) and levobupivacaine (SMD = −1.58; 95% CI: −2.89, −0.27; P = 0.02) were significant, in contrast to bupivacaine (SMD = −0.70; 95% CI: −1.59, 0.19; P = 0.12) and lignocaine (SMD = 0.07; 95% CI: −2.31, 2.45; P = 0.95) [Figure 4]. However, the sample sizes for each LA were not large enough; thus, the results should be interpreted cautiously. Sensitivity analysis performed by excluding studies with a JADAD score of <3 did not alter the overall results or the direction or P value.

Figure 3.

Forest plot: subgroup analysis – onset of sensory block. This forest plot shows that a magnesium sulphate dose of >200 mg led to a faster onset of sensory block when compared to a dose of ≤200 mg. Std mean difference IV = standard mean difference–inverse variance; SD = standard deviation; CI = confidence interval

Figure 4.

Forest plot: subgroup analysis based on LA type – onset of sensory block. This forest plot shows that magnesium sulphate quickened the onset of sensory block with ropivacaine and levobupivacaine, but there was no difference with bupivacaine and lignocaine, when compared to the control group. Std mean difference IV = standard mean difference–inverse variance; SD = standard deviation; CI = confidence interval

Secondary outcomes

The mean duration of sensory and motor block was prolonged with the addition of MgSO₄ by 161.31 min and 130.27 min, respectively. The mean duration of both sensory and motor blocks was prolonged to a greater degree in MgSO₄ doses >200 mg compared to doses of ≤200 mg [Table 3]. There were no subgroup differences observed for the duration of sensory block (P = 0.66, I² = 0%) or for the duration of motor block (P = 0.11, I² = 59.8%). The onset of motor block was faster with the addition of MgSO₄ (SMD = −1.15; 95% CI: −1.62, −0.68) compared to LA alone (MD of 3.50 min faster than with LA alone) [Figure 5]. Subgroup analysis also showed that a MgSO₄ dose of >200 mg led to a faster onset of motor block (SMD = −1.49; 95% CI: −2.34, −0.63; MD of 5.30 min faster than LA alone) when compared to a dose of ≤200 mg (SMD = −0.97; 95% CI: −1.54, −0.40; MD of 2.25 min faster than LA alone), with P < 0.00001 and I² = 95% [Table 3 and Figure 6]. No subgroup differences were noted (P = 0.32, I² = 0%).

Table 3.

Secondary outcomes and subgroup analysis of secondary outcomes

	Number of studies (participants)	Overall SMD [95% CI]	Overall MD [95% CI]	Subgroup analysis
				Number of studies (participants)	SMD [95% CI]		MD [95% CI]
					MgSO4 dose ≤200 mg	MgSO4 dose >200 mg	MgSO4 dose ≤200 mg	MgSO4 dose >200 mg
Duration of sensory block	29 (1777) [8,16–18,20–23, 25,26,28–46]	3.90 [3.21, 4.58] (P<0.00001)	161.31 [143.98, 178.64] (P<0.00001)	27 (1657) [8,16,17,20–23, 25,26,28–32,34–46]	3.68 [2.71, 4.65] (P<0.00001)	4.01 [2.95, 5.07] (P<0.00001)	144.31 [123.87, 164.75] (P<0.00001)	179.00 [144.13, 213.88] (P<0.00001)
Onset of motor block	28 (1768) [8,16–18, 20–25,27–29, 31–35,37–46]	-1.15 [-1.62, -0.68] (P<0.00001)	-3.50 [-4.91, -2.09] (P<0.00001)	27 (1708) [8,16–18, 20–25,27–29, 31,32,34,35,37–46]	-0.97 [-1.54, -0.40] (P=0.0009)	-1.49 [-2.34, -0.63] (P=0.0006)	-2.25 [-3.62, -0.87] (P=0.001)	-5.30 [-8.67, -1.94] (P=0.002)
Duration of motor block	29 (1833) [8,16–18, 20–29,31–38, 40–46]	3.53 [2.87, 4.19] (P<0.00001)	130.27 [114.94,145.59] (P<0.00001)	28 (1773) [8,16–18, 20–29,31,32,34–38, 40–46]	3.83 [2.91, 4.74] (P<0.00001)	2.78 [1.85, 3.71] (P<0.00001)	116.52 [104.03, 129.02] (P<0.00001)	145.26 [95.36, 195.16] (P<0.00001)
Duration of analgesia	21 (1368) [8,16–18, 20,23,24,27,30–36, 38,41,43-46]	4.22 [3.36, 5.08] (P<0.00001)	197.07 [166.16, 227.97] (P<0.00001)	20 (1308) [8,16–18, 20,23,24,27,30–32, 34–36,38,41,43-46]	4.26 [3.04, 5.47] (P<0.00001)	3.90 [2.71, 5.09] (P<0.00001)	173.58 [131.08, 216.08] (P<0.00001)	231.65 [169.72, 293.59] (P<0.00001)

SMD=standard mean difference; MD=mean difference; CI=confidence interval

Figure 5.

Forest plot – onset of motor block. The forest plot shows that the onset of motor block was slower in the control group than when magnesium sulphate was used as an adjunct in BPBs. Std mean difference IV = standard mean difference–inverse variance; SD = standard deviation; CI = confidence interval

Figure 6.

Forest plot: subgroup analysis – onset of motor block. This forest plot shows that magnesium sulphate dose of >200 mg led to a faster onset of motor block when compared to a dose of ≤200 mg. Std mean difference IV = standard mean difference–inverse variance; SD = standard deviation; CI = confidence interval

The duration of analgesia was prolonged with MgSO₄ compared to LA alone (MD = 197.07 min) [Table 3 and Figure 7]. However, the duration of analgesia was longer for MgSO₄ doses of >200 mg than for ≤200 mg (231.65 min vs 173.58 min, respectively) [Figure 8]. Upon sensitivity analysis, excluding the outliers, there was no difference between the groups regarding subgroup differences (P = 0.44, I² = 0%). Furthermore, a subgroup analysis based on different types of LA used was significant for all types; however, the results for individual LA should be treated with caution due to the small sample sizes for each LA considered (ropivacaine, bupivacaine, levobupivacaine, and lignocaine). Sensitivity analysis performed by excluding studies with a JADAD score of <3 did not alter the overall results or the direction or P value for any of the secondary outcomes, except for the incidence of rescue analgesia.

Figure 7.

Forest plot – duration of analgesia. This forest plot shows that magnesium sulphate prolonged the duration of analgesia when compared to the control group. Std mean difference IV = standard mean difference–inverse variance; SD = standard deviation; CI = confidence interval

Figure 8.

Forest plot: subgroup analysis – duration of analgesia. This forest plot shows that the duration of analgesia was prolonged more for magnesium sulphate doses of >200 mg compared to ≤200 mg. Std mean difference IV = standard mean difference–inverse variance; SD = standard deviation; CI = confidence interval

The incidence of rescue analgesia was higher in the LA only group, with an odds ratio (OR) of 0.32; 95% CI: 0.15, 0.70; P = 0.004. The number needed to treat (NNT) for rescue analgesia is 8 in favour of the MgSO₄ group. The relative risk reduction was 24% if MgSO₄ were to be used in the LA mixture. However, performing sensitivity analysis based on JADAD score ≥3 revealed OR = 0.36; 95% CI: 0.11, 1.18; P = 0.09, indicating the influence of a low JADAD score on the overall results. Hence, the overall results should be treated with utmost caution and warrant a more focused standardised approach in future trials to ascertain the impact of adding MgSO₄ when assessing the incidence of rescue analgesia.

Analysis of bias and internal consistency

Meta-regression of the data for LA, amount and volume of MgSO₄, JADAD score, and the method of block administration showed no significant influence [Table 4]. The risk of bias table indicated an overall ‘high’ or ‘some concern’, primarily due to potential bias in outcome measurement (D4) and the selection of reported results (D5) [Figure 9]. This was largely due to a lack of available information regarding the assessor’s awareness of the study intervention and the availability of pre-specified analysis plans. However, the JADAD score for most of the studies demonstrated high study design quality. Meta-regression for the primary outcome, with an overall test of the model where P = 0.93 and R² = 0.00, indicates no variance in the true effects.

Table 4.

Meta regression Table

Meta Regression	MgSO4 amount	Volume of MgSO4	LA amount	Volume of LA	Type of LA	JADAD Score	Method of block
2-sided P	0.67	0.24	0.76	0.64	(Composite) 0.51	0.62	(Composite) 0.49

MgSO4 = magnesium sulphate, LA = local anaesthetic

Figure 9.

Risk-of-bias summary

The precision funnel plot does not demonstrate any differences [Figure 10]. Duval and Tweedie’s ‘trim and fill’ did not show any unaccounted-for studies on the left or the right of the mean using a random- or fixed-effects model, although Egger’s regression showed P = 0.0005. This suggests that inherent flaws may be embedded within the trials, as the meta-regression did not significantly impact the overall outcome (P = 0.93) when all covariates were considered. The R² analogue 0 explained that there was no between-study variance. The scatter plot equation was y = −0.83 − 0.0002 × dose of MgSO₄, showing that for every unit increase in dose of MgSO₄, there was a corresponding decrease in effect size.

Figure 10.

Precision funnel plot

Trial sequential analysis

According to the TSA [Figure 11], the sample size achieved exceeded the required IS of n = 919. We found that the IS was reached for both doses of magnesium used (n = 772 and 822, respectively) for MgSO₄ doses of ≤200 mg and >200 mg. The required IS for ultrasound-guided blocks (n = 477) was easily reached compared to the IS for peripheral nerve stimulator blocks (n = 1753), demonstrating that more RCTs are needed to confirm the benefit of using peripheral nerve stimulators.

Figure 11.

Trial sequential analysis for the onset of sensory block. The blue z curve represents the cumulative sum of participants; the vertical red line represents the required information size (IS); the outer inward sloping red coloured lines represent the trial sequential monitoring boundaries; the inner sloping lines signify the futility boundary; the brown upper and lower horizontal lines represent the conventional boundary

DISCUSSION

This meta-analysis showed that MgSO₄ quickens the onset of sensory block when used as an adjunct in BPBs compared to LA alone.

Utilising more studies and patients, and excluding confounders such as general anaesthetics, we showed that adding MgSO₄ shortened the onset of sensory block with statistical significance (MD of 2.16 min vs 0.25 min and 1.29 min in the Li et al.[6] and Zeng et al.[7] studies, respectively). However, the clinical significance of this slightly quicker onset time would likely be minimal at an organisational level in terms of improved efficiency, but may benefit the patient.[53] We further established that a MgSO₄ dose of ≤200 mg shortened the onset of sensory block, with a slightly quicker onset observed for a dose of >200 mg compared to LA alone.

Secondary outcome data showed that MgSO₄ resulted in a longer duration of sensory and motor block. Furthermore, this study established that MgSO₄ shortened the onset of motor block. The TSA performed for the onset of motor block far exceeded the IS (n = 828), but the IS (n = 1781) for MgSO₄ dose ≤200 mg fell short but did not reach the futility boundary, indicating that more RCTs are required in this area for a definitive conclusion in comparison to >200 mg (n = 704). The duration of sensory and motor blocks was prolonged for different types of LA used. However, the sample sizes for bupivacaine, prilocaine, and lignocaine were small compared to ropivacaine; therefore, this result should be interpreted cautiously.

Li et al.[6] and Zeng et al.[7] demonstrated that MgSO₄ prolonged sensory block duration (MD = 106.69 min vs 114.59 min), albeit to a lesser extent than this study (MD = 161.31 min). A longer sensory block duration would be beneficial in shorter procedures performed under peripheral nerve blocks alone, thereby avoiding general anaesthesia and its associated complications, such as postoperative nausea and vomiting (PONV) and pulmonary complications.[54] Like this study, Li et al.[6] similarly determined that the duration of motor block and onset of motor block were prolonged and quicker, respectively, with MgSO₄ (MD = 89.95 min and 1.17 min, respectively).

This meta-analysis demonstrated that MgSO₄ as an adjunct prolonged the analgesic effect compared to LA alone (MD = 197.07 min). A possible explanation is that reducing calcium influx through the antagonism of voltage-gated N-methyl-D-aspartate receptors reduces central sensitisation of peripheral nociceptor stimulation following a painful stimulus.[55] Li et al.[6] also demonstrated a prolonged analgesia time with the addition of magnesium (MD = 124.66 min vs LA alone); however, only seven studies were analysed compared to 21 in this study.

A subgroup analysis showed that a MgSO₄ dose of >200 mg extended the duration of analgesia more than when compared to a dose of ≤200 mg (by ~60 min). Additionally, the TSA showed that the sample sizes are adequate to conclude that MgSO₄ prolongs the duration of analgesia. This conclusion is of clinical importance if the primary aim of the block is for postoperative analgesia, where, for example, nerve catheters need to be avoided. Peng et al.[56] demonstrated a prolonged analgesic time through improved pain scores with 400 mg of MgSO₄ used as a perineural adjunct. Doses as high as 1 g have been safely used in transverse abdominal plane blocks[57] and as low as 100 mg in axillary nerve blocks.[26] However, further research would be required to demonstrate the optimal MgSO₄ dose for use in BPBs. As mentioned above, using a dose of ≤200 mg is adequate to provide analgesia, and any higher doses may not be necessary to achieve the same outcome. Given the potential risks associated with high doses of MgSO₄, this outcome highlights the clinical significance of considering the MgSO₄ dose when used as part of a perineural block.

By prolonging analgesic effects, patient recovery and satisfaction can be improved by delaying the use of postoperative analgesia such as opiates, from which side effects can often be problematic for patients.[58] This meta-analysis showed that the incidence of rescue analgesia was higher in the control group in the first 24 h. Therefore, using MgSO₄ as an adjunct in BPBs demonstrates a clinical advantage in delaying the potential need for such medications. Conversely, delaying the requirement for postoperative analgesia does not necessarily equate to an overall reduction in postoperative analgesic requirements. This study did not examine that; however, Zeng et al.[7] demonstrated that magnesium reduced total analgesic consumption in the 24-h postoperative period for patients who received a BPB. However, their TSA showed that sample sizes were not adequately powered.

The NNT for rescue analgesia was 8 in favour of the MgSO₄ group. However, the required IS computed using TSA was inadequate for this outcome (n = 480 vs a required size of n = 579). Therefore, the results should be interpreted cautiously, and more focused RCTs should address this issue.

From the papers included in the study, perioperative adverse effects of adding MgSO₄ were either not mentioned or insignificant when compared to LA alone. These side effects included nausea (incidence of 3.7% in control and MgSO₄ groups from 10 studies), vomiting (incidence of 1.48% in control and MgSO₄ groups from 5 studies), pruritis (incidence of 0% in control and MgSO₄ groups from 1 study), dyspnoea (incidence of 0% in control and MgSO₄ groups from 1 study), Horner’s syndrome (incidence of 0% in control and MgSO₄ groups from 4 studies), hypotension (incidence of 1.5% vs 1.3% in control and MgSO₄ groups, respectively from 12 studies) and bradycardia (incidence of 0.3% vs 0% in control and MgSO₄ groups, respectively from 12 studies). However, Li et al.[6] found no significant difference between the two groups. Zeng et al.[7] found a reduction in the incidence of PONV in those who received MgSO₄, across different types of blocks; however, there was no significant difference in those who received BPBs (n = 220; OR = 0.75; 95% CI: 0.30, 1.88). The current available evidence is inconsistent regarding the long-term impact of perineural MgSO₄, but interestingly, our study demonstrates more benefit in using perineural MgSO₄ compared to Desai et al.[59]

Limitations

This study has several limitations that should be acknowledged. This study focused on BPBs; therefore, we can only conclude that using MgSO₄ as an adjunct in such cases is beneficial. It would be sensible to extrapolate these results to other types of nerve blocks, but this meta-analysis cannot confirm this. There was also a degree of variation between the included papers in their definitions of the primary and secondary outcomes. Additionally, some papers did not define the parameters of their outcomes, which made interpretation difficult. A standardised definition for the measured outcomes would be required in future RCTs to enable better evidence synthesis.

Visual analogue scores (VAS) were assessed in 15 of the included studies; however, analysing the data in any meaningful way was difficult due to inconsistency across the studies in how they were analysed. This study, therefore, was unable to assess this outcome. More focused and consistent reporting in VAS scores would be recommended in future RCTs exploring this. Similarly, more standardised and focused RCTs are needed to ascertain the impact of adding MgSO₄ to LA when considering the incidence of rescue analgesia, given the current evidence.

There was also significant heterogeneity for all outcomes, including the primary outcome. However, meta-regression, sensitivity analysis, and subgroup analysis were performed to examine heterogeneity among the confounders, but we could not detect any; thus, it is more likely that there is intra-study variation than inter-study variation. Egger’s regression showed P = 0.0005, demonstrating that inherent flaws might be embedded within the trials themselves, as neither meta-regression nor trim and fill, precision modelling, sensitivity analysis, or R² analogue could identify any reasons for variance in the observed effect size.

CONCLUSION

Using magnesium sulphate as an adjunct to local anaesthetics in brachial plexus blocks quickens the onset of sensory block. It has been shown that a magnesium sulphate dose of ≤200 mg is optimal for the onset of sensory and motor block, as well as the duration of sensory and motor block and analgesia. However, further randomised controlled trials are required to establish definitive conclusions about the onset of motor block when using magnesium sulphate ≤200 mg, peripheral nerve stimulators, the incidence of rescue analgesia usage, and the optimal LA type, due to inadequate sample sizes.

Presentation at conferences/CMEs and abstract publication

EBPOM World Congress 2024 – London. 04/07/2024. Poster presentation.

Study data availability

The dataset generated and/or analysed for this study is available from the corresponding author on reasonable request.

Disclosure of use of artificial intelligence (AI)-assistive or generative tools

The authors confirm that no AI tools or language models (LLMs) were used in the writing or editing of the manuscript, and no images were manipulated using AI, except that software has been used for grammar corrections and references.

Declaration of use of permitted tools

The scales, scores, figures, and tables are freely available and not copyrighted.

Authors contributions

SP: Data collection, manuscript writing and reviewing, data analysis, and project administration. RKK: Data collection, manuscript writing and reviewing. RM: Conceptualisation, Manuscript writing and reviewing, supervision. AB: Conceptualisation, Manuscript writing and reviewing, data analysis, supervision. RB: Manuscript writing and reviewing, data analysis, and conceptualisation.

Supplementary material

This article has supplementary material and can be accessed at this link. Supplementary Material at http://links.lww.com/IJOA/A49.

Conflicts of interest

There are no conflicts of interest.

Supplementary Figure 1

Electronic search strategy for PUBMED

Funding Statement

Nil.