- Question: Do the adjuvant pharmaceutics prolong the duration of sensory and motor blockade for peripheral nerve block?
- Findings: These findings provide evidence for the consideration of dexmedetomidine, clonidine, especially dexamethasone, as adjuvants to prolong the duration of peripheral nerve block and improve analgesia.
- Meaning: The adjuvant pharmaceutics especially dexamethasone may be used to prolong the duration of local anesthetic for peripheral nerve block; however, more high-quality randomized controlled trials (RCTs) of dexamethasone should be performed in the future.
Peripheral nerve blocks (PNB) with local anesthetics (LAs) are the most commonly used for limb surgery, with the advantages of increasing analgesic effect and reducing postanesthesia care unit stay, opioid use and cost, while still being easy to administer uterus.1,2 However, these advantages may be offset by the limited duration of current LAs, and, especially, be weakened during postoperative pain control.3 While the analgesic duration can be prolonged by increasing the LA dose, the side effects and risk of potential neurotoxicity can also be increased.4 Anesthetists have sought strategies to prolong the benefits of single-shot PNBs beyond the pharmacological duration of commonly used LAs.5 Adjuvant analgesic strategies are a technically simple method that can be used to extend the analgesic duration and decrease the potential risk of side effects by reducing the LA dose.6 As ultrasound techniques have developed, indwelling perineural catheters can effectively provide analgesia for several days via prolonged LA infusion. However, their utility is limited by difficulties with catheter removal, rarely infection, or inherent secondary failure rate.7,8
Many LA adjuvants, including opioids, epinephrine, corticosteroids, α2-adrenergic receptor (α2AR) agonists, and magnesium sulfate, have demonstrated the synergistic analgesic effects of LAs to improve the quality of PNB.1,4 LA adjuvants perform this effect by several mechanisms, such as local vasoconstriction limiting systemic uptake, directly acting on peripheral nerves, or having systemically anti-inflammatory effects.1 Some randomized controlled trials (RCTs), systematic reviews, and meta-analyses have indicated the facilitatory effects of perineural administration of LA adjuvants to prolong the analgesia duration.1,9,10 Despite this evidence, the debate regarding which LA adjuvant provides superior, prolonged effects on PNBs with the fewest side effects remain.
Based on the previous studies to date, we selected several commonly used LA adjuvants that were directly or indirectly compared to detect the facilitatory effects of adjuvant pharmaceutics. Therefore, the purpose of this network meta-analysis (NMA) was to compare the efficacy of dexmedetomidine, fentanyl, morphine, sufentanil, dexamethasone, clonidine, epinephrine, tramadol, magnesium sulfate (MgSO4), and buprenorphine in prolonging the analgesia of LA during PNBs.
The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines were followed during the preparation of this NMA including registration of the study methods before study conduct. The study protocol was prospectively registered with PROSPERO, the international prospective register of systematic reviews (CRD42020187866).
Two of the authors (C.L. and W.Y.) independently retrieved relevant studies published before May 18, 2020 from electronic databases including PubMed, EMBASE, Web of Science, and the Cochrane Library. ClinicalTrials.gov was also evaluated for ongoing and unpublished trials. The medical subject heading (MeSH) and text words relating to dexmedetomidine, fentanyl, morphine, sufentanil, dexamethasone, clonidine, epinephrine, tramadol, MgSO4, and buprenorphine were sought as intervention. Results were combined using the Boolean operator “AND” with the search terms with Mesh and text words including PNB, local anesthesia, infiltration anesthesia, “anesthetics, local” [MeSH], “nerve block” [MeSH], “adjuvants, pharmaceutic” [MeSH], adjuvant, and regional anesthesia. Finally, the search was limited to RCTs. No manual searching was performed in this NMA, and only English publications were included.
Inclusion and Exclusion Criteria
Parallel group RCTs were included to detect the effects of adding perineural agents as an adjuvant to LA on sensory block (SB) time or motor block (MB) time in patients undergoing limb surgery with PNBs. Randomized, and single- or double-blinded trials including adults (age ≥18 years) published in full manuscript were considered. RCTs with 2 LA adjuncts combined using were also included. We did not restrict different dosages of adjuvants or LAs. To reduce imbalance of studies with different LAs, the studies involving short acting LAs were excluded. We also excluded studies with irrelevant methods including general anesthesia, intrathecal anesthesia, continuous catheter-based nerve blocks, and adjuvant analgesia. Studies were also excluded if they did not include at least one of the designated primary outcomes of the NMA or if both the mean values and standardized differences could not be obtained. Since the inclusion of comparisons with only a single study direct comparison can increase the absolute error in meta-analyses,11 interventions with only a single direct comparison study were removed.
C.L. and D.W. independently viewed titles, abstracts, and full texts according to the selection criteria, and then extracted relevant data using a standardized data sheet. Data extracted included primary author, year of publication, comparative groups, sample size, type and dose of LA, type and dose of perineural of LA adjuvants, analgesic effects, pain score, opioid consumptions, and primary and secondary outcomes. Any disagreement was discussed with a third reviewer (C.X.) until a consensus was reached. The trial authors were contacted about any missing or unclear data. The mean was estimated for continuous outcomes using the reported value or median and the standard deviation was derived from the reported confidence interval (CI), interquartile range, or P value, if necessary.12
We designated SB and MB duration (in hours), defined as time from completion of LA injection to full recovery from SB and MB, as primary outcomes. Secondary outcomes included duration of analgesia (in hours), defined as time to first analgesic request (FAR). We also planned to evaluate cumulative analgesic consumption during the first 24 hours postoperatively and pain severity at 12 and 24 hours postoperatively. However, given the lack of available data of opioid consumption and pain scores after data extraction, this evaluation could not be assessed in this NMA.
Assessment of Bias
Using the Cochrane Risk of Bias instrument, 2 reviewers independently assessed the risk of bias in each study including random sequence generation, allocation concealment, performance bias, detection bias, and attribution bias. Each bias was subsequently graded as a “high,” “uncertain,” or “low” risk of bias. Confidence evidence in NMA (CINeMA, version 0.6.1) was used to evaluate the certainty of confidence contribution to the NMA based on the 5 essential elements of the GRADEpro, which included instruction, study limitations, imprecision, inconsistency (heterogeneity and incoherence), indirectness, and across-studies bias.13
Contributions of direct comparisons were summarized for the risk of bias assessment for each outcome by blinding rank. Additionally, to evaluate the presence of small study effects, the comparison-adjusted funnel plots for each outcome were visually inspected. To ensure appropriate comparison in the funnel plots, the order of treatments in the data set were set from the oldest to newest.
To estimate effect sizes for continuous variables of primary and secondary outcomes, the results were calculated as mean differences (MDs) with 95% CIs for the primary outcomes and secondary outcome. To visualize network geometry and node connectivity, a network plot was produced for all outcomes.14 NMAs were fit within a frequentist framework using a multivariate random effects meta-analysis model that accounts for correlations between effect sizes in studies with more than 2 groups.15,16
Network consistency and a common heterogeneity parameter across all treatment contrasts were assumed for all analyses. For treatment comparisons, the MDs and 95% CIs are presented that account for uncertainty in variance estimates in league tables. Summary treatment effects are also presented with 95% CIs and their corresponding prediction intervals (PrIs) for all placebo comparisons in forest plots.16 To obtain treatment hierarchies, we used a parametric bootstrap procedure with 10,000 resamples to compute ranking probabilities for all ranks and outcomes.15 Surface under the cumulative ranking curve (SUCRA) was computed to give a ranking probability of each treatment under different outcomes.14 NMAs were conducted using the “mvmeta” and “network” packages in STATA 14.0.
Statistical heterogeneity was assessed in each pairwise comparison with the I2 and τ2 statistic and its associated P values. I2 > 50% was considered to indicate statistical heterogeneity. Consistency was evaluated using a global inconsistency, and side-splitting inconsistency model and loop inconsistency for local inconsistency. Since different LAs with various analgesic times may cause heterogeneity, a subgroup analysis was performed for each LA. For the subgroup analysis, the CIs and PrIs for all comparisons were summarized and presented in interval plots.
Description of the Studies Included
The research identified 16,364 potentially relevant studies (PubMed 5269, Embase 2225, Cochrane Library 6906, Web of Science 1963, Clinical Trials.gov 1 completed trial with unpublished data), including 10,499 unique records. Overall 284 full-text articles were retrieved after the exclusion of 10,153 reports on the basis of titles and abstracts and 61 records without full-text. Of these 285 full-text articles, 53 trails including 3469 participants were included (Figure 1). Although we intended to include sufentanil, epinephrine, and morphine as LA adjuvants during PNB, the involved studies were excluded due to single study comparisons, thus, adjuvants under this NMA study included dexamethasone, dexmedetomidine, buprenorphine, MgSO4, clonidine, fentanyl, and tramadol. Therefore, 7 interventions were involved in the duration of SB, 6 of them for MB time, and 6 of them with time of FAR evaluation.
The basic characteristics and information of the 53 enrolled articles are described in Supplemental Digital Content, Tables 1 and 2, and Supplementary References, https://links.lww.com/AA/D586, including 49 articles with 59 study arms for SB, 49 articles with 57 study arms for MB, and 38 articles with 43 study arms for FAR. Overall 3469 participants were randomized to different interventions, including 1588 patients randomized to a placebo group. The network plot of SB time, MB time, and FAR time are shown in Figure 2A–C, respectively.
Risk of Bias, Heterogeneity, and Consistency
The risk of bias summary and graph are presented in Supplemental Digital Content, Figures 1 and 2, https://links.lww.com/AA/D586. The comparison-adjusted funnel plots for SB time, MB time, and FAR time were asymmetric, suggesting that a small study or publication bias was observed in this NMA (Supplemental Digital Content, Figure 3, https://links.lww.com/AA/D586). Based on contributions of each risk of bias (Supplemental Digital Content, Figure 4, https://links.lww.com/AA/D586) and indirectness contributions of each study to the network estimate (Supplemental Digital Content, Figure 5, https://links.lww.com/AA/D586), the confidence rating of comparisons for primary and secondary outcomes using CINeMA are presented in Supplemental Digital Content, Figure 6, https://links.lww.com/AA/D586. The confidence rating of primary and secondary outcomes had 4 high evidence level and 1 very low evidence level for mixed comparisons, and 2 high evidence level and 4 very low evidence level for indirect comparisons.
The NMA of all indirectly and directly compared treatments were performed in our study, which showed that SB time (I2 = 99.01%, τ2 = 6.04), MB time (I2 = 99.26%, τ2 = 6.62), and time of FAR (I2 = 98.48%, τ2 = 6.58) had high risks of heterogeneity. However, for global inconsistency testing, no inconsistency was detected for SB time (χ2  = 4.13; P = .13), MB time (χ2  = 5.06; P = .08), and time of FAR (χ2  = 0.65; P = .42). Local inconsistency was assessed using a side-splitting inconsistency model and loop inconsistency, but no inconsistency was found (data not shown). Due to no detectable inconsistency, there was no disagreement between direct and indirect comparisons.
SB Time, MB Time, and Time of FAR in NMA
The results of NMA for SB time and MB time are presented as MD and 95% CIs in Figure 3A. SB time was available for 59 intervention arms (49 included studies, 3269 patients) including 7 active treatments and 1 placebo therapy group (Figure 2A). As presented in Figure 4A, with respect to SB time, the results of the NMA demonstrated that dexamethasone (5.73 hours, 95% CI, 4.16–7.30), dexmedetomidine (4.51 hours, 95% CI, 3.52–5.50), fentanyl (3.59 hours, 95% CI, 0.11–7.06), clonidine (2.75 hours, 95% CI, 1.46–4.04), and MgSO4 (2.81 hours, 95% CI, 0.01–5.60) could significantly prolong the SB time of LA as adjuvants.
MB time was available for 57 intervention arms (49 included studies, 3021 patients) including 6 active treatments and sham therapy group (Figure 2B). The results of NMA suggested that dexamethasone (4.20 hours, 95% CI, 2.51–5.89), dexmedetomidine (4.04 hours, 95% CI, 2.98–5.11), fentanyl (4.42 hours, 95% CI, 0.78–8.06), and clonidine (2.93 hours, 95% CI, 1.69–4.16) were associated with a significantly prolonged MB time of the LA (Figures 3A and 4B).
Time of FAR as a secondary outcome was involved in 43 intervention arms (38 included studies, 2567 patients) including 6 active treatments and 1 sham therapy group (Figure 2C). The results of time of FAR are presented as MD and 95% CI in Figure 3B. The results of NMA indicate that dexamethasone (8.71 hours, 95% CI, 6.63–10.79), fentanyl (6.70 hours, 95% CI, 3.11–10.29), dexmedetomidine (5.25 hours, 95% CI, 4.08–6.43), tramadol (6.03 hours, 95% CI, 2.21–9.84), and clonidine (3.35 hours, 95% CI, 1.82–4.87) were associated with significant longer time of FAR (Figure 4C).
Supplemental Digital Content, Figure 7 and Table 3, https://links.lww.com/AA/D586, present the ranking probabilities and SUCRA values for SB time, MB time, and time of FAR. Based on the positive results of NMA, the treatment with the highest probabilities of being the most efficacious in prolonging SB time was dexamethasone (93.7%), whereas tramadol (35.6%) was least efficacious. For MB time, dexamethasone (79.1%) had the highest probability and MgSO4 (31.9%) had lowest probability of prolonging the LA. For the time of FAR, dexamethasone (95.0%) had the highest probability of being the longest time and MgSO4 (21.5%) had lowest probability. Dexamethasone as an adjuvant for LAs presented the highest probability to prolong the duration of LAs based on the ranking probabilities assessment.
Subgroup Analysis According to Different LAs
Due to the high risks of heterogeneity and publication bias observed in this study, a subgroup analysis was performed according to different LAs, with the hope of reducing this bias, since different LAs have different analgesia times. In our study, there were 3 LAs included in the NMA, including bupivacaine, levobupivacaine, and ropivacaine. However, only ropivacaine had enough included studies to perform a subgroup NMA, including 28 studies for SB time, 29 studies for MB time, and 20 studies for time of FAR (Supplemental Digital Content, Table 2 and Figure 8, https://links.lww.com/AA/D586). In this subgroup NMA study, high risks of heterogeneity for SB time (I2 = 97.64%, τ2 = 4.88), MB time (I2 = 98.97%, τ2 = 5.88), and time of FAR (I2 = 98.19%, τ2 = 2.66) were observed, and the publication bias could also be found. Although the high risks of heterogeneity still existed, we got similar results during this subgroup analysis. Compared with the placebo group, dexamethasone, dexmedetomidine, and clonidine significantly prolonged SB time (5.59 hours, 95% CI, 3.67–7.51; 4.51 hours, 95% CI, 3.32–5.70; 2.95 hours, 95% CI, 1.64–4.26, respectively) (Figure 5A) and MB time (4.17 hours, 95% CI, 1.85–6.48; 4.03 hours, 95% CI, 2.70–5.35; 3.10 hours, 95% CI, 1.76–4.43, respectively, Figure 5B). Dexamethasone (10.01 hours, 95% CI, 7.95–12.08), dexmedetomidine (4.78 hours, 95% CI, 3.73–5.83), and clonidine (3.41 hours, 95% CI, 2.21–4.60) also demonstrated a significantly longer time of FAR in the subgroup analysis (Figure 5B). In addition, dexamethasone as LAs adjuvant significantly prolonged time of FAR compared with dexmedetomidine (5.23 hours, 95% CI, 2.92–7.54) or clonidine (6.61 hours, 95% CI, 4.29–8.92).
This systematic review and NMA of adjuvants to LAs for PNBs included data from 53 clinical trials including 3469 patients who were randomized to 7 interventions or a sham therapy. Our findings provide further evidence about the analgesic efficacy of different adjuvants. Based on primary and secondary outcomes, the duration of SB and FAR is prolonged by 2 to 4 hours when dexmedetomidine, fentanyl, dexamethasone, or clonidine are added to LAs as adjuvants. However, dexmedetomidine, dexamethasone, and clonidine can prolong the duration of MB about as long. The subgroup analysis also supports the facilitatory effects of dexamethasone, dexmedetomidine, and clonidine, and indicates that dexamethasone has a longer time of FAR than clonidine or dexmedetomidine.
Since the identification of opioid receptors in the peripheral nervous system, the interest in opioids alone or a combination with LAs for PNB has been aroused for several decades.17 The results of our NMA also support that the addition of fentanyl to LAs for PNB prolongs the duration of anesthesia and analgesia. However, from the results of indirect comparisons, fentanyl was not superior to dexamethasone, clonidine, or dexmedetomidine for extending the active time of LAs. Due to an insufficient number of studies involved, the efficacy of fentanyl could not be further assessed by subgroup analysis. Therefore, more RCTs are needed to study the improvement of analgesia induced by administrated fentanyl to LAs.
Clonidine, an α2 adrenergic agonist, was initially used as an antihypertensive, but 1 study reported that when given through an interscalene catheter without an LA, it could provide better analgesia compared with the systemic administration of the same dose.18 One meta-analysis provided a high level of evidence suggesting that clonidine added to a LA for peripheral nerve or plexus block prolonged the duration of postoperative analgesia by about 2 hours.19 Similar results were observed in this NMA study. Dexmedetomidine, another α2 adrenergic agonist, has been proposed as a safe and effective LAs adjuvant to extend the duration of single-shot block.20,21 One recent meta-analysis further demonstrated that perineural dexmedetomidine improved brachial plexus block onset, quality, and analgesia.22 Correspondingly, the results of this NMA also indicate that dexmedetomidine could significantly prolong duration of SB and FAR.
Previously studies reported that dexmedetomidine has more favorable pharmacokinetic and pharmacodynamic properties than clonidine and might be an interesting option for neuraxial anesthesia and analgesia.23,24 Our NMA shows that dexmedetomidine has longer duration of SB than clonidine, but does not significantly prolong MB time. This suggests that dexmedetomidine might be a better choice than clonidine as an adjuvant to LA for PNB. The ranking probabilities also support this finding. However, as an α2 adrenergic agonist, the benefits of dexmedetomidine or clonidine should still be carefully weighed against the possible risks of sedation, fainting, bradycardia, and hypotension.19,20
Dexamethasone is a corticosteroid drug that has been added to regional anesthetics such as lidocaine, bupivacaine, and other sodium channel blocking anesthetics to prolong the duration of analgesia.25,26 The underlying mechanisms of analgesia prolongation are poorly understood and may include the inhibition of nociceptive C-fibers and/or anti-inflammatory effects.25,27 Choi et al1 have suggested that the addition of dexamethasone to regional anesthetics prolonged the duration of analgesia by approximately 6 hours with no observed adverse events. In this NMA, we found dexamethasone significantly prolonged the duration of analgesia 4 to 7 hours. In addition, during the subgroup analysis, dexamethasone prolonged the duration of FAR about 10 hours, and had longer time of FAR than dexmedetomidine or clonidine. Based on this finding, dexamethasone may be the best potential choice to act as adjuvant for extending analgesia of LAs. However, the prolonged MB duration of dexamethasone should not be ignored, which may delay ambulation in patients undergoing surgery of a lower extremity.
Our review and NMA has several limitations. We did not evaluate the side effects induced by adjuvants administration, including peripheral nerve injury, depth of sedation, postoperative nausea/vomiting, and hemodynamic side effects, because of the limited data available from the included studies. Thus, we are unable to comment on the safety profile in light of our findings. Additionally, the high level of heterogeneity was not successfully resolved with subgroup analyses, which may reduce the external validity of our finding. This suggests that some unaccounted factors may have affected these results. First, all patients underwent surgery without general anesthesia. Previous reports have suggested some general anesthesia drugs may have effects that last into the postoperative period until FAR.19 Second, we did not consider the impact induced by some narcotics, since intravenous sedation could prolong the duration of analgesia effect of LAs during regional block. Third, some intrinsic factors—smaller sample sizes of individual studies, the potential variation in the study population, variations of surgery procedure, not distinguishing between different dosages of adjuvants or LAs—may have led to high level of heterogeneity by the large variation in the magnitude of effect across all included studies. The wide array of different PNBs over a wide time period and different nerve blocks without differentiation may be the source of the heterogeneity. In recent years, development of nerve stimulation and ultrasound guidance techniques has increased the accuracy and efficacy of regional blocks. In addition, only English publications were evaluated. It is possible that some trials may have been missed in the search, thereby introducing a potential bias. Further, adjuvants effects are, at least in part, mediated systemically. Although we are able to evaluate the effect of these adjuvants, this analysis is unable to distinguish between systemic and local effects. Chong et al28 indicated that perineural dexamethasone prolonged the duration of analgesia by 3.77 hours compared to intravenous dexamethasone, but Pehora et al29 suggested that both perineural and intravenous could similarly prolong duration of SB in upper limb surgery. Zhao et al30 show that only dexamethasone combined with epinephrine could prolong the effects of analgesic duration compared to intravenous route, otherwise perineural and intravenous had equivalent effect as adjuvants on regional anesthesia.
The results of this NMA suggest that the LA adjuvants dexamethasone, dexmedetomidine, and clonidine could significantly prolong the duration of sensory, MB, and FAR. Further, if prolonged MB is not a problem and a long time to first analgesic request is intended, dexamethasone may be preferable. However, regardless of which adjuvant is selected, physicians should continuously monitor patients for the duration of MB.
Name: Chengluan Xuan, MD, PhD.
Contribution: This author helped substantially to conception and design, acquisition of data, analysis and interpretation of data, writing the abstract, introduction, data acquisition, analysis and interpretation, discussion, and revising the final manuscript.
Name: Wen Yan, MD, PhD.
Contribution: This author helped with acquisition of data, analysis and interpretation of data.
Name: Dan Wang, MD, PhD.
Contribution: This author helped with acquisition of data, analysis and interpretation of data.
Name: Cong Li, MD.
Contribution: This author helped with acquisition of data, analysis and interpretation of data.
Name: Haichun Ma, MD, PhD.
Contribution: This author helped design the study and revised it critically for important intellectual content.
Name: Ariel Mueller, MA.
Contribution: This author validated and revised it critically for important intellectual content.
Name: Jingping Wang, MD, PhD.
Contribution: This author helped with the conception and design, analysis and interpretation, revising the final manuscript, and approved the final submitted manuscript.
This manuscript was handled by: Markus W. Hollmann, MD, PhD.
1. Choi S, Rodseth R, McCartney CJ. Effects of dexamethasone as a local anaesthetic adjuvant for brachial plexus block: a systematic review and meta-analysis of randomized trials. Br J Anaesth. 2014;112:427–439.
2. Fredrickson MJ, Krishnan S, Chen CY. Postoperative analgesia for shoulder surgery: a critical appraisal and review of current techniques. Anaesthesia. 2010;65:608–624.
3. Becker DE, Reed KL. Local anesthetics: review of pharmacological considerations. Anesth Prog. 2012;59:90–101.
4. Wu HH, Wang HT, Jin JJ, et al. Does dexmedetomidine as a neuraxial adjuvant facilitate better anesthesia and analgesia? A systematic review and meta-analysis. PLoS One. 2014;9:e93114.
5. Boezaart AP, Davis G, Le-Wendling L. Recovery after orthopedic surgery: techniques to increase duration of pain control. Curr Opin Anaesthesiol. 2012;25:665–672.
6. Opperer M, Gerner P, Memtsoudis SG. Additives to local anesthetics for peripheral nerve blocks or local anesthesia: a review of the literature. Pain Manag. 2015;5:117–128.
7. Aveline C, Le Hetet H, Le Roux A, et al. Perineural ultrasound-guided catheter bacterial colonization: a prospective evaluation in 747 cases. Reg Anesth Pain Med. 2011;36:579–584.
8. Bowens C Jr, Briggs ER, Malchow RJ. Brachial plexus entrapment of interscalene nerve catheter after uncomplicated ultrasound-guided placement. Pain Med. 2011;12:1117–1120.
9. Yaghoobi S, Shahamat H, Alizadeh A, Khezri MB. Comparing postoperative analgesic effect of dexmedetomidine or dexamethasone added to lidocaine through infraclavicular block in forearm surgery. Clin J Pain. 2019;35:766–771.
10. Aliste J, Layera S, Bravo D, et al. Randomized comparison between perineural dexamethasone and dexmedetomidine for ultrasound-guided infraclavicular block. Reg Anesth Pain Med. 2019;44:911–916.
11. Dias S, Ades AE, Welton NJ, et al. Chapter 12: Validity of network meta-analysis. In: Network Meta-Analysis for Decision-Making. 2018:1st ed. Wiley Press; 351–374.
12. Higgins JPT, Green S. Cochrane Handbook for Systematic Reviews of Interventions. Version 5.1.0 [updated March 2011]. The Cochrane Collaboration, 3/2011. Available at: www.cochrane-handbook.org
13. Nikolakopoulou A, Higgins JPT, Papakonstantinou T, et al. CINeMA: an approach for assessing confidence in the results of a network meta-analysis. PLoS Med. 2020;17:e1003082.
14. Chaimani A, Higgins JP, Mavridis D, Spyridonos P, Salanti G. Graphical tools for network meta-analysis in STATA. PLoS One. 2013;8:e76654.
15. White IR, Barrett JK, Jackson D, Higgins JP. Consistency and inconsistency in network meta-analysis: model estimation using multivariate meta-regression. Res Synth Methods. 2012;3:111–125.
16. Mutz J, Vipulananthan V, Carter B, Hurlemann R, Fu CHY, Young AH. Comparative efficacy and acceptability of non-surgical brain stimulation for the acute treatment of major depressive episodes in adults: systematic review and network meta-analysis. BMJ. 2019;364:l1079.
17. Stein C. Peripheral mechanisms of opioid analgesia. Anesth Analg. 1993;76:182–191.
18. Iskandar H, Benard A, Ruel-Raymond J, Cochard G, Manaud B. The analgesic effect of interscalene block using clonidine as an analgesic for shoulder arthroscopy. Anesth Analg. 2003;96:260–262.
19. Pöpping DM, Elia N, Marret E, Wenk M, Tramèr MR. Clonidine as an adjuvant to local anesthetics for peripheral nerve and plexus blocks: a meta-analysis of randomized trials. Anesthesiology. 2009;111:406–415.
20. Williams BA, Hough KA, Tsui BY, Ibinson JW, Gold MS, Gebhart GF. Neurotoxicity of adjuvants used in perineural anesthesia and analgesia in comparison with ropivacaine. Reg Anesth Pain Med. 2011;36:225–230.
21. Agarwal S, Aggarwal R, Gupta P. Dexmedetomidine prolongs the effect of bupivacaine in supraclavicular brachial plexus block. J Anaesthesiol Clin Pharmacol. 2014;30:36–40.
22. Vorobeichik L, Brull R, Abdallah FW. Evidence basis for using perineural dexmedetomidine to enhance the quality of brachial plexus nerve blocks: a systematic review and meta-analysis of randomized controlled trials. Br J Anaesth. 2017;118:167–181.
23. Bagatini A, Gomes CR, Masella MZ, et al. Dexmedetomidina: farmacologia e uso clínico [Dexmedetomidine: pharmacology and clinical application.]. Rev Bras Anestesiol. 2002;52:606–617.
24. Ramadhyani U, Park JL, Carollo DS, Waterman RS, Nossaman BD. Dexmedetomidine: clinical application as an adjunct for intravenous regional anesthesia. Anesthesiol Clin. 2010;28:709–722.
25. Movafegh A, Razazian M, Hajimaohamadi F, Meysamie A. Dexamethasone added to lidocaine prolongs axillary brachial plexus blockade. Anesth Analg. 2006;102:263–267.
26. Martinez V, Fletcher D. Dexamethasone and peripheral nerve blocks: on the nerve or intravenous? Br J Anaesth. 2014;113:338–340.
27. Dawson RL, McLeod DH, Koerber JP, Plummer JL, Dracopoulos GC. A randomised controlled trial of perineural vs intravenous dexamethasone for foot surgery. Anaesthesia. 2016;71:285–290.
28. Chong MA, Berbenetz NM, Lin C, Singh S. Perineural versus intravenous dexamethasone as an adjuvant for peripheral nerve blocks: a systematic review and meta-analysis. Reg Anesth Pain Med. 2017;42:319–326.
29. Pehora C, Pearson AM, Kaushal A, Crawford MW, Johnston B. Dexamethasone as an adjuvant to peripheral nerve block. Cochrane Database Syst Rev. 2017;11:CD011770.
30. Zhao WL, Ou XF, Liu J, Zhang WS. Perineural versus intravenous dexamethasone as an adjuvant in regional anesthesia: a systematic review and meta-analysis. J Pain Res. 2017;10:1529–1543.