Intravenous Magnesium Sulphate for Analgesia after Caesarean Section: A Systematic Review

Objective To summarise the evidence for use of intravenous magnesium for analgesic effect in caesarean section patients. Background Postcaesarean pain requires effective analgesia. Magnesium, an N-methyl-D-aspartate receptor antagonist and calcium-channel blocker, has previously been investigated for its analgesic properties. Methods A systematic search was conducted of PubMed, Scopus, MEDLINE, Cochrane Library, and Google Scholar databases for randomised-control trials comparing intravenous magnesium to placebo with analgesic outcomes in caesarean patients. Results Ten trials met inclusion criteria. Seven were qualitatively compared after exclusion of three for unclear bias risk. Four trials were conducted with general anaesthesia, while three utilised neuraxial anaesthesia. Five of seven trials resulted in decreased analgesic requirement postoperatively and four of seven resulted in lower serial visual analogue scale scores. Conclusions Adjunct analgesic agents are utilised to improve analgesic outcomes and minimise opioid side effects. Preoperative intravenous magnesium may decrease total postcaesarean rescue analgesia consumption with few side effects; however, small sample size and heterogeneity of methodology in included trials restricts the ability to draw strong conclusions. Therefore, given the apparent safety and efficacy of magnesium, its role as an adjunct analgesic in caesarean section patients should be further investigated with the most current anaesthetic techniques.


Introduction
Magnesium, the fourth most common cation in the body, has been used in medicine and anaesthesia for a range of applications including seizure prevention in preeclampsia, tocolysis, asthma management, and dysrhythmias [1][2][3].
Magnesium sulphate also appears e ective at improving postoperative analgesia [4]. e postoperative analgesic e ect of magnesium appears to extend to nonobstetric patients undergoing general or neuraxial anaesthesia [1]. Magnesium has an antagonistic e ect at the N-methyl-D-aspartate (NMDA) receptor [5], as well as calcium-channel blocker properties [6]. Antagonism at the NMDA receptor is thought to alter the mechanism of central hypersensivity and to subsequently decrease analgesic requirements including opioid consumption [7]. Postoperative shivering also appears to be decreased in patients given preoperative magnesium [8].
In patients who undergo caesarean section, adequate postoperative pain management allows early and su cient ambulation and breastfeeding [9] and reduces complications including thromboembolic events, pneumonia, and poor bonding with the newborn while also improving patient satisfaction [9,10]. E ective postoperative analgesia may enable patients to be discharged from hospital earlier following caesarean deliveries [11]. Consequently, several analgesia modalities and adjuncts, of which magnesium may be included, have been studied and reviewed for postcaesarean analgesia [9].
To our knowledge, no systematic review exists summarising the evidence for use of intravenous magnesium as an adjunct for postcaesarean section analgesia. erefore, the aim of our review and our research question was to identify, evaluate, and summarise the evidence for the use of intravenous magnesium sulphate, compared to placebo control, for analgesic e ect in parturients undergoing caesarean section and qualitatively review its e ectiveness.

Methods
A systematic literature search on PubMed, Scopus, MEDLINE, Cochrane Library, and Google Scholar databases was performed using the search terms "(caesarean OR cesarean) AND magnesium", all in title, with no language or other lters. All reference lists of articles found were also manually screened. e last search was conducted in May 2017. is review included only published studies and did not explicitly search for non-English trials, though no such trials were retrieved with the search strategy outlined.
Requirements for inclusion in this review were (1) intervention randomisation, (2) control group arm, (3) use of intravenous magnesium pre-or intraoperatively in parturients undergoing caesarean section, and (4) reporting on postoperative analgesia outcomes. Exclusion criteria included any trials that were retrospective or observational or studies on epidural or intrathecal use of magnesium sulphate.
All papers produced by the search methodology were screened. Papers not meeting the inclusion criteria were excluded. Each trial was assessed for risk of bias using the methodology described by the Cochrane Handbook [12] through examination of selection bias, performance bias, and detection bias. Postoperative pain and analgesic requirement outcomes, as reported in the original trials, were then compared qualitatively and are reported.

Results
Initial search methodology produced 51 studies which were then assessed for compliance with inclusion and exclusion criteria and were sorted as in Figure 1.
Twenty-one papers were excluded due to magnesium administration route other than intravenous. Sixteen were excluded due to absence of analgesic outcomes. One was excluded for reporting of intraoperative outcomes only with no postoperative analgesia outcomes reported [13]. Two were excluded due to retrospective or audit methodologies [14,15]. Screening of references yielded one extra study which was subsequently excluded due to lack of placebo control in an observational study [16].

Trials Meeting Inclusion
Criteria. Ten randomised trials meeting inclusion criteria were returned by the search methodology described. e assessment of risk of bias for each is outlined in Table 1. Seven of the ten trials were deemed to have low risk of bias and a high-quality rating per the GRADE approach [12]. ree trials, Agrawal et al. [17], Davoudi et al. [18], and Safavi et al. [19], were deemed to possess an unacceptably unclear risk of bias across all domains due to inability to discern whether the study protocol was insu cient or omitted. Subsequently, in line with the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions, these studies were downgraded to a moderate quality rating per the GRADE approach [12] and excluded. e seven subsequently included trials (Table 2) encompassed a total of 530 patients, 260 of whom received intravenous magnesium sulphate preoperatively. Four trials administered magnesium concurrently with general anaesthesia [20][21][22][23] (240 patients: 110 receiving magnesium sulphate) while the other three were with neuraxial anaesthesia [11,24,25] (290 patients: 150 receiving magnesium sulphate).

Magnesium Administration Protocols.
All seven trials administered placebo control and intravenous magnesium interventions preoperatively as a bolus. However, three trials also administered an ongoing magnesium infusion for 24 hours postoperatively following the preoperative bolus [11,24,25]. e trial by Paech et al. [11] included two intervention arms (high dose and low dose), both with bolusinfusion regimens.
Dosage regimens used by each trial are outlined in Table 3. Doses ranged from 25 mg/kg to 50 mg/kg. One study used a standard dose of 6 grams as a bolus [24] which would likely be the highest dose-to-weight studied. 50 mg/kg was the most commonly studied bolus dose, with four trials assessing a group with this dose [11,[20][21][22]. Of those with postoperative infusion protocols, two groups 16 excluded due to absence of analgesic outcomes 18 excluded due to intrathecal and/or epidural route 3 excluded due to local incisional wound in ltration route 2 excluded due to retrospective methodologies 1 excluded due to absence of control group in an observational study 1 excluded due to intraoperative only outcomes; no post-op analgesia outcomes 51 reports yielded 35 reports studying magnesium therapy in caesarean section with pain outcomes 10 randomised-control trials of intravenous magnesium with analgesia outcomes   Anesthesiology Research and Practice received a standard 2 g/hr infusion [11,24], one group received 1 g/hr [11] and one group received a weight-based infusion of 10 mg/kg/hr for 24 hours [25].

Analgesic
Requirements. Five trials measured time to rst analgesia requirement (Table 4). e e ect of magnesium on prolonging time to rst rescue analgesia varied in signi cance between trials. ree of the ve found signi cant prolongation compared with control, p < 0.01 [20,24,25]. e remaining two found IV magnesium comparable with IV placebo [11,23].
Total analgesic requirement in the immediate 24 hours post caesarean was measured in six of the seven trials. Intravenous magnesium resulted in signi cantly less analgesic consumption than placebo in four of the six, p < 0.01 [20][21][22]25]. Likewise, Elgebaly et al. [24] did not measure total analgesic consumption, though did measure and reveal signi cantly lower frequency of analgesic intake in their magnesium group (2.5) compared to control (3.0; p < 0.01) [24] in the rst 24 hours post caesarean, which may be a comparable outcome. e two trials which did not reach signi cance included limitations; Helmy et al. [23] had the smallest intervention sample size (n � 20) of any trial, and Paech et al. [11] was the only trial to provide patientcontrolled epidural analgesia (PCEA) postoperatively. It is unclear if these discrepancies may have resulted in failure to reach signi cance.

Ranked Pain
Outcomes. All included trials used a visual analogue scale (VAS) to measure patient-ranked pain scores (Table 5). However, time points and frequency of measurements varied widely from trial to trial. Elgebaly et al. [24] measured VAS scores only at the time of rst analgesia requirement and did not reveal a signi cant di erence between intervention and control groups. Similarly, Paech et al. [11] only measured a VAS score at one time point for patient analgesia satisfaction at 48 hours postoperatively which did not reveal bene t in the intervention group compared to control (p � 0.449).
e remaining ve trials serially measured VAS scores. Magnesium intervention resulted in signi cantly lower VAS scores at the 24-hour time point in two trials [22,25] (p < 0.05) and consistently lower VAS scores at multiple time points in the same trials (p < 0.05).
However, Mireskanadari et al. [21] reported lower VAS scores in the intervention group at one, six, and 12 hours (p < 0.05), but not at 24 hours [21]. Similarly, Elrahman and Youssry [20] reported lower VAS scores at six and twelve hours in magnesium patients (p < 0.05), but not at two or at 24 hours. Two trials, Paech et al. [11] and Helmy et al. [23], measured VAS scores at comparable values to the placebocontrol group at several time points. However, these intervention arms do show discrepancies as outlined in Section 3.3.

Comparison with Other
Interventions. Two of the included trials compared intravenous magnesium sulphate with other active intervention arms. Helmy et al. [23] included a group which received a bolus of IV ketamine 0.3 mg/kg (n � 20) which was compared with an IV magnesium sulphate 30 mg/kg bolus (n � 20) group and IV normal saline control group (n � 20). In regard to postoperative pain and analgesia, the performance of magnesium was inferior to that of ketamine in prolonging the time to rst analgesia requirement, lowering VAS scores (at two and six hours), and decreasing total 24-hour analgesic requirements. Ketamine produced signi cantly more favourable results compared to control while magnesium failed to do so [23].
Elgebaly et al. [24] compared the e cacy of an IV magnesium bolus and infusion as an adjunct to 0.5% bupivacaine spinal with a combination of spinal bupivacaine and an intrathecal bolus of 25 μg spinal fentanyl for postoperative analgesia. Magnesium performed comparably to fentanyl; both interventions prolonged time to rst analgesic request and reduced total 24-hour analgesic requirements. Magnesium, however, produced fewer side e ects than intrathecal fentanyl including pruritus and nausea and vomiting, while also causing less perioperative sedation in comparison to fentanyl (p < 0.01) [24].

Other Outcomes.
In those trials measuring neonatal outcomes, magnesium performed comparably to placebo controls. Apgar scores were comparable in the ve trials that reported them [11,20,21,23,25]. One trial routinely measured umbilical artery and vein pH, which was comparable between magnesium and control groups [11]. One trial routinely measured time to ambulate unassisted postoperatively, which found patients in the magnesium group ambulate sooner (4.2 hours) than control (6.3 hours, p < 0.01) [20].

Side E ects.
Magnesium intervention did not result in toxicity in any of the trials. e incidence of postoperative nausea and vomiting or shivering was comparable to control in all trials that reported these outcomes. However, IV magnesium resulted in signi cantly greater surgeonreported intraoperative blood loss than control in the trial by Paech et al. (500 mL median in high-dose magnesium group, 475 mL in low-dose group, and 400 mL in control group; p < 0.01) [11].
is increased loss is likely due to the tocolytic properties of magnesium sulphate [11].
is increased blood loss did not result in blood transfusion or additional uterotonic use in any case. Mireskandari et al. [21] and Elrahman and Youssry [20], however, recorded no signi cant di erence in blood loss between patients receiving magnesium than control (p > 0.05).  Maulik et al. [25] reported an increased incidence of intraoperative hypotension in magnesium intervention patients than control (p < 0.001), which was readily corrected with vasopressors [25].

Discussion
Magnesium, an NMDA receptor antagonist [5] and calcium channel blocker [6], has been utilised intravenously as an adjunct analgesic in various surgical procedures with varying results [26]. However, evidence for its analgesic role post caesarean section has not been systematically reviewed. is review consolidates the evidence for postcaesarean analgesia outcomes following intravenous magnesium which revealed con icting results. e trend amongst available evidence suggests that analgesic requirements in the immediate 24-hour postoperative period may be reduced with IV magnesium. Beyond this, there is currently insu cient evidence for the role of preoperative intravenous magnesium for postcaesarean analgesia.
Unfortunately, a meta-analysis of the included studies was neither appropriate nor possible due to the small sample size of intervention groups and heterogeneity of methodology, inconsistent outcome points, and di ering rescue analgesic practices and types. Assessment of publication bias, such as inclusion of a funnel plot, was not performed due to too few included studies. Subsequently, the aim of our review was to qualitatively assess the evidence for intravenous magnesium sulphate as a postoperative analgesic adjunct.
In regard to administration protocols, no argument can be made for using a bolus-infusion protocol over a bolusonly protocol on the evidence available. Likewise, neither anaesthetic type (spinal or general) revealed apparent bene t, although a study with direct comparison between these patient groups has not been conducted. A bolus of 50 mg/kg was the most commonly used dose; however, beyond citing similarity to dosing used for preeclampsia, no trials justi ed their dosing. As discussed by Helmy et al., a lower-dose bolus of 30 mg/kg may be insu cient [23]. e  usual dosing of magnesium for preeclampsia is 4 grams as a loading dose and an infusion up to 24 hours at 1 g/hr [27].
Six of the seven included studies reported on total 24-hour postoperative analgesic requirement, and ve of seven serially reported on postoperative VAS scores. With the exception of the study by Paech et al. [11] which was the only study to utilise PCEA and the study by Helmy et al. [23] in which an outcome di erence may have been obscured by the low patient number in the intervention group, the remaining studies supported an analgesic e ect in regard to improvement in 24-hour analgesic consumption and serial VAS scores. Our review found that there is insu cient evidence to support the role of intravenous magnesium in increasing time to rst postoperative rescue analgesia requirement.
Other authors [22] have postulated that the superior analgesic e cacy of PCEA may have masked the analgesia e ect of magnesium in the study by Paech et al. [11]. In addition, Helmy et al. [23] conceded in their study that failure to reach signi cant outcome di erences from control may have been due to their low magnesium bolus dose (30 mg/kg compared to the most common 50 mg/kg) and lack of an ongoing infusion protocol [23]. e safety of IV magnesium for use as an analgesic adjunct can be translated from use elsewhere in obstetrics, or for analgesia in other surgical procedures, with reviews concluding no serious side e ects at doses as high as 28 grams over 24 hours with no di erence in morbidity compared with placebo [26,28]. A Cochrane review of magnesium use in preeclamptic parturients supported its safety at doses up to 1 g/hr [27]. Further, magnesium for a range of indications has an established safety pro le in parturients [3]. Increased surgeon-reported intraoperative blood loss was the single side e ect identi ed in this review [11]. However, the e ect of magnesium on uterine tone and subsequent requirement of additional oxytocin has been described elsewhere [29] and should be considered.
Strengths of our review include the exclusive criteria of intravenous magnesium for analgesic e ect in parturients postcaesarean section. All trials were randomised and blinded, though not without some remaining risk of bias due to limited speci cs on allocation concealment. e majority of trials reported on VAS scores, time to rst analgesic request, and total 24-hour analgesic consumption allowing comparison.
ere were several limitations in this review. Firstly, the majority of trials had small intervention groups (n � 20-42) perhaps limiting the ability to reach signi cance, for example, in the smallest group by Helmy et al. [23], or also introducing risk of bene t detection by chance [30]. e possibility that potential side e ects may similarly have been undetected secondary to small sample sizes should also be considered. Secondly, there was inconsistency of end points and their reporting as well as di ering modes of rescue analgesia. Further, there was variation in the dose and administration protocols used.
ere was also infrequent measurement and reporting of side e ects including haemodynamics, nausea, vomiting, sedation, and shivering. Finally, the low control group pain scores in some trials [11] secondary to highly e ective analgesia (PCEA) may also have confounded the ability to identify intervention bene t, given low baseline pain [31]-90% median pain control satisfaction in control group [11]. Speci c measurements such as length of hospital stay and long-term (>48 hours) analgesia requirements failed to be reported in any trial with the exception of six-week followup data by Paech et al. [11]. e most signi cant limitation of our review to provide relevance to current practice is the failure of any study to utilise a modern approach to obstetric anaesthesia and postoperative pain management. Only three of the seven trials [11,24,25] were conducted with spinal anaesthesia, the current standard in modern obstetric anaesthesia [32], and only one trial [11] used intrathecal opioids (fentanyl) with their neuraxial technique. Longer-acting intrathecal opioids such as spinal morphine combined with local anaesthetic would be the most current and relevant mode of anaesthesia for testing IV magnesium as an analgesic adjunct [32,33]. Likewise, no trial adopted a postoperative analgesia regimen in line with the current standard; a multimodal regimen including regular paracetamol and nonsteroidal anti-in ammatory drugs (NSAIDs) with rescue opioids [34]. It is di cult to comment whether the modes of anaesthesia utilised in the included trials impacted on their ability to identify intervention bene t; however, it undoubtedly limits the applicability of the results to current practice.
Future research agenda following our review and its limitations could comprise a randomised-control trial of intravenous magnesium as an adjunct to spinal anaesthesia using local anaesthetic and long-acting opioid followed by a multimodal analgesia regimen. Such a trial would ideally include two intervention arms: one employing a bolus-only regimen and the other employing a bolus-infusion regimen. A scheduled protocol to monitor for adverse e ects is essential. Further, there remains a paucity of data on long-term outcomes including length of hospital stay and wound pain at six weeks which should also be addressed.
On balance, we conclude that preoperative intravenous magnesium has the potential to improve postcaesarean outcomes by decreasing total postoperative analgesic requirements. In the case of opioid rescue analgesia, this has the potential to decrease adverse e ects including nausea, vomiting, sedation, and constipation while maintaining satisfactory analgesia [35]. However, this potential bene t needs to be further evaluated in forthcoming trials with adequate sample size and more current anaesthetic techniques. By ensuring satisfactory postcaesarean analgesia, patients may be able to ambulate and breastfeed earlier with greater patient satisfaction [9].

Conclusion
Preoperative intravenous magnesium sulphate may decrease 24-hour rescue analgesia requirements in patients following caesarean section with no serious adverse e ects. Further trials utilising a current anaesthetic approach are required.
Anesthesiology Research and Practice 7