Analgesic effect and safety of single-dose intra-articular magnesium after arthroscopic surgery: a systematic review and meta-analysis

To examine the analgesic effect and safety of single-dose intra-articular (IA) magnesium (Mg) after arthroscopic surgery. Pubmed, Embase and Cochrane library were searched through in January 2016. Eight RCTs and eight experimental studies were included. The IA Mg exhibited a significantly lower pain score when compared with placebo (MD, −0.41, 95% CI, −0.78 to −0.05, p = 0.03). There was no significant difference between Mg and bupivacaine in terms of pain relief and the time to first analgesic request. Furthermore, statistically significant differences both in pain score (MD, −0.62, 95% CI, −0.81 to −0.42, p < 0.00001) and time to first analgesic request (MD, 6.25, 95% CI, 5.22 to 7.29, p < 0.00001) were observed between Mg plus bupivacaine and bupivacaine alone. There was no statistically significant difference among the various groups with respect to adverse reactions. Most of the included in vitro studies reported the chondrocyte protective effect of Mg supplementation. There were also two in vivo studies showing the cartilage protective effect of IA Mg. The single-dose IA Mg following arthroscopic surgery was effective in pain relief without increasing adverse reactions, and it could also enhance the analgesic effect of bupivacaine. In addition, Mg seemed to possess the cartilage or chondrocyte protective effect based on experimental studies.

Scientific RepoRts | 6:38024 | DOI: 10.1038/srep38024 and reduce the chondrotoxicity of local anesthetics. Furthermore, some broader issues related to the analgesic ability and chondrocyte protective effect of IA Mg in other settings, such as osteoarthritis, are to be addressed. Therefore, the objectives of this systematic review and meta-analysis were to investigate the analgesic effect of single-dose IA based on the following comparisons: (1) Mg versus placebo, (2) Mg versus bupivacaine and (3) Mg plus bupivacaine versus bupivacaine alone, after arthroscopic surgery. Another objective was to assess the protective effect of Mg on chondrocyte or cartilage based on randomized controlled trials (RCTs), in vitro and in vivo studies.

Methods
Literature search. This meta-analysis was in accord with the Preferred Reporting Items for Systematic review and Meta-analyses statement 31 . The electronic databases of Pubmed, Embase and Cochrane library were searched through in January 2016 using a series of logic combination of keywords and text words related to arthroscopic knee surgery, magnesium and randomized controlled trials (RCTs) (Appendix 1). The Pubmed and Embase databases were also searched through in January 2016 to retrieve in vitro and in vivo experimental studies of Mg supplementation (Appendix 1). No restriction was imposed, and the references of the retrieved articles and reviews were evaluated. Study selection. Two researchers reviewed all the retrieved titles, abstracts and full texts independently.
Disagreements were resolved through discussions and/or consultations to a thrid researcher. All of the eligible trials must meet the following criteria: (1) patients undergone arthroscopic surgery; (2) including a treatment group of single-dose IA Mg or Mg plus bupivacaine for postoperative pain relief; (3) including a controlled group of IA placebo or bupivacaine alone, (4) other interventions should be balanced between comparative groups, (5) RCTs. The exclusion criteria for this study were: (1) case series, reviews, protocols or vitro studies; (2) non-RCTs; (3) abstract or full text was not available; (4) no IA injection; (5) meeting abstract.
Data extraction. The available information and outcomes of each included study were extracted by two independent researchers. The retained data included the first author, year of publication, size of each group, doses of intervention, follow-up time points, type of operation, type of anesthesia and injection time. The Cochrane risk of bias table was used to assess the methodological quality of the included RCTs. A total of seven potential risks of bias were evaluated: random sequence generation, allocation concealment, blinding of participants, blinding of outcome assessment, incomplete outcome data, selective reporting and other bias (mainly including the conflict of interests) 32 . Each risk item was evaluated based on a three-level rating system: low risk, unclear risk and high risk. Studies with three or more unclear or high risk of bias were considered as poor methodological quality.
The primary outcome of interests for this study was the effects of IA Mg and Mg plus bupivacaine on postoperative pain control. If a study reported multiple pain scales, the highest one on the hierarchy of pain scale related to the outcomes was adopted, as described by Jüni and colleagues 33 . The secondary outcomes of interest were the time interval to the first request of analgesia and the side effects. If data was presented in figures, the GetData software would be used (http://getdata-graph-digitizer.com/index.php) to extract data from the figures. If the outcomes were reported in terms of the median and the range or interquartile range, the median would be considered as mean and the standard deviation would be estimated by the range or interquartile range 32 .

Statistical analyses.
The outcome measures investigated in this meta-analysis were postoperative pain intensity and the time to first request of analgesia. Due to the low incidence of adverse reactions, only qualitative analysis were conducted on relevant adverse effects. Firstly, we calculated the overall mean difference (MD) and its corresponding 95% confidence interval (CI) between comparative groups for postoperative pain intensity at last follow up. Then a subgroup analysis were conducted by calculating the MD at different time points of follow up. If all included trials reported the pain intensity at a same postoperative time point, an overall MD between comparative groups for pain intensity was calculated at this time point of follow up. We also calculated the overall MD and its related 95% CI for the time to first request of analgesia. The heterogeneity of effect size across trials was tested by Q statistics (p < 0.05 was considered heterogeneous). If there was a significant heterogeneity among the studies, the random-effects model was used; otherwise, the fixed effects model was acceptable. We also examined the I 2 statistic, which measures the percentage of the total variation across studies due to heterogeneity (I 2 > 50 was considered heterogeneous). We further conducted sensitivity analysis to explore possible explanations for heterogeneity and to examine the influence of various exclusion criteria on the overall MD.
Begg's tests 34 and funnel plots were performed to assess the publication bias. Statistical analyses were performed using Review Manager 5 software (RevMan 5, The Cochrane Collaboration, Oxford, UK) and STATA version 11.0 (StataCorp LP, College Station, Texas). A p value less than 0.05 was considered to be statistically significant.

Results
Search results and selected articles. The literature search retrieved a total of 49 citations for potential RCTs examining the analgesic effect of single-dose IA Mg after arthroscopic surgery and 491 citations for potential in vitro and in vivo experimental studies of Mg supplementation, respectively. Two evaluators screened the titles/abstracts and full-texts of all eligible articles. Eventually, 8 RCTs 23-28 (513 patients) and 8 experimental studies 13,14,29,30,[35][36][37][38][39][40] (6 in vitro and 2 in vivo studies) were qualified for final analysis (Fig. 1). The characteristics of the included RCTs are presented in Table 1. The characteristics and results of the included in vitro and in vivo experimental studies are listed in Table 2. According to the Cochrane risk of bias table (Appendix 2), three trials 25,27,29 did not describe their random sequence generation (high risk of selection bias); six trials 24-26,28-30 did not design a clear allocation concealment plan (unclear or high risk of selection bias); all trials adopted the double-blind method except one 27 which did not describe the blinding method of participants (high risk of performance bias); one study 30 was rated as unclear risk for both incomplete data and selective reporting; six studies 23-28 did not present a competing interest statement (unclear risk of other bias). In addition, three trials 25,27,30 were rated as low methodological quality for having three or more unclear or high risk of bias.
Effects of pain relief. Mg versus placebo. Five trials 23,24,27,29,30 were eligible for assessing the postoperative pain score of IA Mg versus placebo after arthroscopic surgery at the last follow-up time point (24 or 48 hours). The combined results showed a significantly lower pain score of Mg (MD, − 0.41, 95% CI, − 0.78 to − 0.05, p = 0.03). Substantial heterogeneity was observed (p = 0.0006, I 2 = 80%) (Fig. 2). Sensitivity analyses after excluding studies with poor methodological quality and studies involving femoral nerve block in each group or involving other types of surgery rather than single knee meniscectomy all showed positive results with reduced heterogeneity (p > 0.05) (Appendix 3). One of the studies reported data with median and range was further excluded in sensitivity analysis, and the result was similar. Meanwhile, the positive results of subgroup analyses remained unchanged at different follow-up time points (2, 12 or 24 hours) (Appendix 4). Begg's rank correlation test suggested no evidence of publication bias among included studies (p = 0.806).
Mg versus bupivacaine. Three trials were 24,27,28 eligible for assessing the postoperative pain score of IA Mg versus bupivacaine (including levobupivacaine) after arthroscopic surgery at the last follow-up time point (24 or 48 hours). The combined results showed no significant difference between Mg and bupivacaine (MD, 0.17, 95% CI, − 0.92 to 1.26, p = 0.76). Substantial heterogeneity was observed (p = 0.0002, I 2 = 88%) (Fig. 2). Sensitivity analyses after excluding studies with poor methodological quality, or studies involving levobupivacaine rather than bupivacaine, or data reported by median and range showed the same results (Appendix 3). Meanwhile, the results of subgroup analyses remained to be insignificant at different follow-up time points (1, 2, 4, 12 or 24 hours) (Appendix 4). Begg's rank correlation test suggested no evidence of publication bias among included studies (p = 0.296).
Mg plus bupivacaine versus bupivacaine alone. Three trials [24][25][26] were eligible for assessing the postoperative pain score of IA Mg plus bupivacaine versus bupivacaine alone after arthroscopic surgery at the last follow-up time point (18 or 24 hours). The combined results showed that the effect of pain relief of Mg plus bupivacaine approached Time to first analgesic request. Mg versus placebo. Four trials 23,24,27,30 were eligible for assessing the time interval before the first request for analgesic medication of IA Mg versus placebo after arthroscopic surgery. The combined results showed a significantly longer time interval of Mg (MD, 3.59, 95% CI, 0.26 to 6.93, p = 0.03). Substantial heterogeneity was observed (p < 0.00001, I 2 = 99%) (Fig. 3). However, sensitivity analyses after excluding studies with poor methodological quality and studies involving femoral nerve block in each group or involving other types of surgery rather than single knee meniscectomy all denied such positive results (p > 0.05) (Appendix 3). Begg's rank correlation test suggested no evidence of publication bias among included studies (p = 0.308).
Mg versus bupivacaine. Three trials 24,27,28 were eligible for assessing the time interval before the first request for analgesic medication of IA Mg versus bupivaciane after arthroscopic surgery. The combined results showed no significant difference between Mg and bupivaciane (MD, − 0.82, 95% CI, − 5.83 to 4.20, p = 0.75). Substantial heterogeneity was observed (p < 0.00001, I 2 = 99%) (Fig. 3). Sensitivity analyses after excluding studies with poor methodological quality or studies involving levobupivacaine rather than bupivacaine reached the same results (p > 0.05) (Appendix 3). Begg's rank correlation test suggested no evidence of publication bias among included studies (p = 1.000).  Mg plus bupivacaine versus bupivacaine alone. Four trials [24][25][26] were eligible for assessing the time interval before the first request for analgesic medication of IA Mg plus bupivacaine versus bupivacaine alone after arthroscopic surgery. The combined results showed a significantly longer time interval of Mg plus bupivacaine (MD, 6.25, 95% CI, 5.22 to 7.29, p < 0.00001). Substantial heterogeneity was observed (p = 0.04, I 2 = 69%) (Fig. 3). Sensitivity analyses after excluding studies with poor methodological quality reached the same positive results with reduced heterogeneity (p = 0.22, I 2 = 34%) (Appendix 3). Begg's rank correlation test suggested no evidence of publication bias among included studies (p = 1.000).

Safety. Adverse reactions reported in RCTs.
Only four included RCTs 27-30 reported adverse reactions (including knee effusion, nausea, vomiting, flushing, shivering, hypotension, bradycardia and drowsiness), as illustrated in Table 2. There was no statistically significant difference between comparable groups (including IA Mg versus placebo) in each trial.

Discussion
This systematic review and meta-analysis was performed on a total of 8 RCTs (published 2006 to 2015) and 8 in vitro and in vivo experimental studies. The most important finding of the present study is that the administration of single-dose IA Mg at the end of arthroscopic surgery was effective in pain relief without increasing adverse reactions when compared with placebo, and exhibited a comparable analgesic effect in comparison with bupivacaine. In addition, IA Mg could enhance the analgesic effect of bupivacaine. Another important finding is that Mg seemed to possess cartilage or chondrocyte protective effects according to the included experimental studies. Thus, IA Mg should perhaps be considered as an alternative to local anesthetics for pain relief after arthroscopic surgery. However, the optimal concentration and dosage of IA Mg still needs to be further explored.
Mg is a physiological antagonist of N-methyl-D-aspartate (NMDA) receptor, which is essential to the development and functionality of the nervous system; it serves as a target for the treatment of cognitive impairment, depression, schizophrenia and pain [41][42][43][44][45] . Mg could be effective in pain relief by exerting the antinociceptive effect 46 , inhibiting TNF-α 47 , and modulating hypesthesia and hyperalgesia 48,49 , by blocking the NMDA receptor. The pain relief effect of Mg has been illustrated by several studies in many cases, such as postoperative sore throat 50 , tourniquet pain 51 , major non-laparoscopic gastrointestinal surgery 52 , diabetic neuropathic pain 53 , cardiac surgery 54 , and cancer-related neuropathic pain 55 . The present study also demonstrated that the single-dose IA Mg was effective in pain relief after arthroscopic surgery.
Several studies, including our previous surveys, have revealed that dietary and serum Mg were negatively associated with the prevalence of knee osteoarthritis [56][57][58][59] , indicating that Mg could possess the cartilage or chondrocyte protective effects. In addition to the included experimental studies (Table 3), Mg was reported to be able to regulate the level of sex determining region Y-box 9, which plays a vital role of cartilage growth plates and is required in the successive steps of chondrogenesis 60 . Furthermore, calcium is critical to the regulation of many cellular physiological functions; the overload of calcium is detrimental to the mitochondrial function 61 . Thus, the inhibition of NMDA receptor by Mg could also decrease the entry of extra-cellular calcium into cells, and thereby exhibits the chondrocyte protective effects. In combination with the findings of the present study, it seems that IA Mg not only can be regarded as an alternative to local analgesics relying on its potential chondrotoxicity, but also shows the cartilage and chondrocyte protective effects. Previous in vitro studies have indicated that local anesthetics such as bupivacaine, lidocaine and ropivacaine exert chondrotoxicity by reducing the chondrocyte viability 13,14 . The cell death or IA crystal formation caused by mitochondrial DNA damage in chondrocytes could possibly explain the potential toxicity of local anesthetics 62 . However, the exact mechanisms still needs to be further explored.

Strengths.
The present study has several strengths. Firstly, this is the first meta-analysis that examined the analgesic effects of single-dose IA Mg after arthroscopic surgery and demonstrated its effectiveness and safety. Secondly, this study included both in vitro and in vivo experimental studies to elucidate the potential cartilage and chondrocyte protective effects of Mg supplementation. Thirdly, a comprehensive literature search was performed in several major databases to cover as many eligible RCTs or experimental studies as possible. Thus, the chance of missing any relevant study was fairly low. Lastly, the sensitivity and subgroup analyses supported the robustness of most findings.

Limitations.
Limitations of the present study should also be acknowledged. Firstly, the number of retrieved RCTs was limited and the sample size of each trial was relatively small. This may bias the results. Secondly, a variety of factors may contribute to the heterogeneity of some indexes and affect the results, including the differences in the types of surgery, the dosage of Mg, the setting of follow-up time point, the ages of patients, the severity of pain and the unequal indications for arthroscopy. Especially, the selected dosages of IA Mg were ranged from 500 to 1000 mg, which required to be further investigated to determine the optimal dosage and to demonstrate the safety as the dosage increases. After all, two studies 37,38 suggested that high level of Mg (50 and 100 mM) or extra-high concentration of Mg may not always be beneficial. Thirdly, because of the limited number of included RCTs, this study was unable to compare IA Mg with other local analgesics such as morphine and ropivacaine, which had been demonstrated effective after arthroscopic surgery by our previous meta-analyses 3,8 . However, other relevant studies suggested that Mg exhibited the similar analgesic effects when compared with morphine and amplified the effect of morphine or ropivacaine 22,[63][64][65][66][67] . Finally, the limited retrieval of in vivo studies (one was about the collagen induced OA model, which is a severe model; the other one was related to stem cells) in the related area may limit the conclusion that Mg supplementation possess the chondroprotective effect.

Conclusion
Single-dose IA Mg at the end of arthroscopic surgery was effective in pain relief without increasing adverse reactions, and it could also enhance the analgesic effect of bupivacaine. In addition, Mg seemed to exhibit the cartilage or chondrocyte protective effect according to the experimental studies. Perhaps IA Mg should be considered as an alternative to local anesthetics after arthroscopic surgery. However, the optimal concentration and dosage of IA Mg still needs to be explored in further fully powered RCTs, to fully address this point.