Effect of tramadol as an adjuvant to local anesthetics for brachial plexus block: A systematic review and meta-analysis

Background Tramadol, a 4-phenyl-piperidine analog of codeine, has a unique action in that it has a central opioidergic, noradrenergic, serotonergic analgesic, and peripheral local anesthetic (LA) effect. Many studies have reported contradictory findings regarding the peripheral analgesic effect of tramadol as an adjuvant to LA in brachial plexus block (BPB). This meta-analysis aimed to evaluate the effects of tramadol as an adjunct to LA in BPB during shoulder or upper extremity surgery. Methods We searched the PubMed, EMBASE, Cochrane, KoreaMed databases, and Google Scholar for eligible randomized controlled trials (RCTs) that compared BPB with LA alone and BPB with LA and tramadol. Primary outcomes were the effects of tramadol as an adjuvant on duration of sensory block, motor block, and analgesia. Secondary outcomes were the effects of tramadol as an adjuvant on time to onset of sensory block and motor block and on adverse effects. We performed the meta-analysis using Review Manager 5.3 software. Results We identified 16 RCTs with 751 patients. BPB with tramadol prolonged the duration of sensory block (mean difference [MD], -61.5 min; 95% CI, -95.5 to -27.6; P = 0.0004), motor block (MD, -65.6 min; 95% CI, -101.5 to -29.7; P = 0.0003), and analgesia (MD, -125.5 min; 95% CI, -175.8 to -75.3; P < 0.0001) compared with BPB without tramadol. Tramadol also shortened the time to onset of sensory block (MD, 2.1 min; 95% CI, 1.1 to 3.1; P < 0.0001) and motor block (MD, 1.2 min; 95% CI, 0.2 to 2.1; P = 0.010). In subgroup analysis, the duration of sensory block, motor block, and analgesia was prolonged for BPB with tramadol 100 mg (P < 0.05) but not for BPB with tramadol 50 mg. The quality of evidence was high for duration of analgesia according to the GRADE system. Adverse effects were comparable between the studies. Conclusions In upper extremity surgery performed under BPB, use of tramadol 100 mg as an adjuvant to LA appears to prolong the duration of sensory block, motor block, and analgesia, and shorten the time to onset of sensory and motor blocks without altering adverse effects.


Introduction
A brachial plexus block (BPB) provides anesthesia and analgesia during surgery involving the upper limb and for acutely painful conditions, and is the most frequent plexus block performed by anesthesiologists. It is worthwhile to explore the options for extending pain relief while minimizing the adverse effects of local anesthesia. Local anesthetics (LAs) have been used with various perineural adjuvants, including dexamethasone [1,2], clonidine [3], dexmedetomidine [4], opioids [5], and magnesium [6], to enhance the quality and duration of anesthesia and postoperative analgesia.
Systemic opioids have been used to relieve pain during surgery for many years, but the effects of perineural opioid adjuvants on BPB are controversial. Some studies have reported that addition of opioids such as fentanyl, alfentanil, morphine, buprenorphine, and meperidine to BPB improves sensory block, motor block, and analgesia, but other studies have found no such effect [7][8][9].
Tramadol administered parenterally or orally has proven effective in managing acute postoperative pain in adults [10]. Tramadol is a unique opioid with two modes of action for inhibition of pain, i.e., an opioid action mediated by the μ receptor and a non-opioid action mediated by α 2 -adrenergic and serotoninergic activity [11,12]. The monoaminergic activity of tramadol inhibits the descending pain pathways, resulting in suppression of nociceptive transmission at the spinal level [13]. Tramadol also exhibits LA properties by blocking K + channels [14]. Clinically, intradermal administration of tramadol provides local anesthesia for minor skin procedures [15]. Many studies have characterized the effects of tramadol as an adjuvant to LA in BPB [16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31]. However, these studies have yielded variable results regarding the analgesia-enhancing effects of tramadol when used in BPB; while some studies showed a beneficial effect, others showed no benefit.
The purpose of this meta-analysis and systematic review was to evaluate the effects of tramadol as an adjunct to LA in BPB on the onset and duration of sensory block, motor block, and analgesia, as well as the adverse effects associated with BPB when used for shoulder and upper extremity surgery.
KoreaMed databases as well as Google Scholar were systematically searched for RCTs performed in adults (aged older than 18 y) up to November 2015 without language restrictions. The search strategy comprised the following key words: ("tramadol") and ("local anesthetic") and ("axillary block" or "brachial plexus block" or "infraclavicular block" or "interscalene block" or "supraclavicular block") as outlined in Supporting Information (S1 File).

Study selection
The studies included in this analysis were peer-reviewed RCTs that compared BPB with LA alone and BPB with LA and tramadol for shoulder or upper extremity surgery in adult patients. Review articles, case reports, letters to the editor, commentaries, proceedings, laboratory studies, and other non-relevant studies were excluded. Two authors (JBJ and YKC) independently assessed the articles for compliance with the inclusion/exclusion criteria. Any disagreement was resolved by discussion or consultation with a third independent investigator (HWS).

Data extraction and assessment of outcomes
The primary outcomes were the effects of tramadol as an adjuvant to LA on duration of sensory block, motor block, and analgesia. The secondary outcomes were the effects of tramadol as an adjuvant to LA on time to onset of motor block and sensory block and on the adverse effects of BPB for shoulder and upper extremity surgery.
Using standardized forms, two authors (JBJ and JYP) independently extracted the following data: the name of the first author, year of publication, type of surgery, type and dose of LA, volume of LA, dose of tramadol, number of patients, technique used for nerve guidance (landmark, nerve stimulator, or ultrasound guidance), type of BPB approach (axillary, infraclavicular, interscalene, or supraclavicular), definitions of sensory and motor block (duration of sensory block, duration of motor block, duration of analgesia, onset of sensory block, and onset of motor block), and adverse effects (nausea, vomiting, pruritus, and sedation). In our analysis, there were two studies that contained more than two groups for tramadol as an adjuvant to LA (one by Kabachi et al. [24] that included arms receiving tramadol 100 mg and 200 mg and the other by Robaux et al. [29] that included arms receiving tramadol 40 mg, 100 mg, and 200 mg). Data from RCTs with more than two intervention groups need to be combined into a single group according to the formula for combining groups in the Cochrane Handbook [32]. However, we used only the data for the 50 mg and 100 mg doses in the metaanalysis for comparison of the effects of tramadol according to dose strength. We attempted to contact the authors of studies that had insufficient or missing data. If contact was not possible, we extrapolated data from the study text or tables to obtain the relevant information. Values for the duration and time to onset of sensory or motor block were converted into minutes and the adverse effects of BPB were reported as the number of patients. The control group included patients who received LA alone in BPB and the intervention group included those who received LA with tramadol in BPB during surgery.

Assessment of risk of bias
Two authors (JBJ and YKJ) independently evaluated the quality of the RCTs by using the risk of bias tool in Review Manager (RevMan 5.3, The Cochrane Collaboration, Oxford, UK). Quality was evaluated using the following seven potential sources of bias: random sequence generation, allocation concealment, blinding of the participants, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and other sources of bias. The methodology for each RCT was graded as "high," "low," or "unclear" to reflect either a high, low, or uncertain risk of bias, respectively.

Statistical analysis
The statistical analysis was performed using RevMan 5.3. We computed the mean difference (MD) with its 95% confidence intervals (CIs) for continuous variables and the relative risk (RR) with corresponding 95% CIs for dichotomous outcome data. The overall data were determined using a Z-test. All reported P-values are two-sided. A two-sided P-value < 0.05 was considered to be statistically significant. Statistical heterogeneity was estimated using the I 2 statistic, which was deemed to be significant when I 2 values were above 50%. The Mantel-Haenszel or inverse variance fixed-effects model was used for the study without significant heterogeneity, while the Mantel-Haenszel or inverse variance random-effects model was used for the study with significant heterogeneity. Sensitivity analyses were performed by excluding studies with a high risk of bias.
We performed subgroup analyses for primary outcomes on the basis of type of BPB approach (axillary, infraclavicular, interscalene, or supraclavicular), dose of tramadol (50 mg or 100 mg), type of LA (intermediate-acting LA [lidocaine, mepivacaine, or prilocaine] or long-acting LA [ropivacaine, bupivacaine, or levobupivacaine]), and volume of LA used for BPB ( 30 mL or >30 mL).
If the funnel plot was visually asymmetric or if the P-values were < 0.1 on Egger's linear regression test, the presence of a possible publication bias was suspected. In such cases, a trimand-fill analysis was performed to confirm publication bias.

Predefined sources of heterogeneity and GRADE guidelines
There was heterogeneity with regard to the definitions of times to onset and duration of sensory block and motor block. Therefore, we assessed the strength of evidence from the RCTs using the GRADE (Grades of Recommendation, Assessment, Development, and Evaluation) guidelines. The GRADE tool evaluates the quality across RCTs for each outcome. Based on key elements, including study quality, consistency, directness, precision, and publication bias, the GRADE tool classifies the strength of the synthesized evidence into four categories: high quality (further research is very unlikely to change the confidence in the estimate of effect); moderate quality (further research is likely to have an important impact on the confidence in the estimate of effect and may change the estimate); low quality (further research is very likely to have an important impact on the confidence in the estimate of effect and is likely to change the estimate); and very low quality (there is a high degree of uncertainty about the estimate).

Publication bias
We evaluated a funnel plot for every comparison and estimated the publication bias using Egger's linear regression method. Egger's linear regression method indicated the publication bias for the following comparisons (>10 studies for comparison): duration of sensory block (P = 0.00985), duration of motor block (P = 0.01386), duration of analgesia (P = 0.00995), and  Results of the meta-analysis
7. GRADE guidelines. In subgroup analysis according to tramadol dose, the duration of sensory block, motor block, and analgesia was prolonged in the studies with tramadol 100 mg for BPB but not in the studies with tramadol 50 mg. When the strength of the evidence was evaluated using the GRADE guidelines, there was high evidence that tramadol 100 mg with LA for BPB prolonged the duration of analgesia when compared with LA alone for BPB in patients undergoing upper extremity surgery ( Table 4). The overall quality assessment was downgraded by inconsistency of effect, heterogeneity, and publication bias, but upgraded by the larger treatment effect and the presence of a dose-response relationship.

Discussion
Our systemic review and meta-analysis indicates that use of tramadol as an adjuvant to LA in BPB prolongs the duration of sensory block, motor block, and analgesia and that it shortens the time to onset of sensory block and motor block without any change in adverse effects. There was some heterogeneity between the studies with regard to definitions of analgesia, sensory block, and motor block. There was high evidence according to GRADE guidelines that tramadol 100 mg with LA for BPB prolonged the duration of analgesia when compared with LA alone for BPB. To our knowledge, this is the first systematic review to evaluate the effect of tramadol as an adjuvant to LA in BPB for shoulder and upper extremity surgery. In the past, there have been contradictory results regarding the effect of opioids as an adjuvant to LA in BPB. Saryazdi et al. [7] reported that addition of different opioids (meperidine, buprenorphine, morphine, and fentanyl) to lidocaine in axillary BPB achieved no statistically significant difference in duration of sensory block or motor block between the study groups.
Tramadol has unique modes of action, including weak opioid activity via the μ receptor, α 2adrenergic and serotonergic agonistic activity, and LA properties via blockade of K + channels [33][34][35].
Our study included 16 studies that examined the effect of tramadol as an adjuvant to LA for BPB and also included quality control. However, the studies included in the review showed high heterogeneity. Generally, the type of surgery performed often determines the selection of BPB approach (interscalene, supraclavicular, infraclavicular, or axillary). This can affect the duration of analgesia at the surgical site. As an example, interscalene approaches are used for shoulder surgery, whereas axillary approaches tend to be used more for surgery on the forearm and hand. This difference in approach contributes to different results and clinical heterogeneity. We performed the meta-analysis using RevMan statistical software and performed subgroup analysis for various items (type of BPB approach, dose of tramadol, type of LA, volume of LA used for BPB) to identify the source of the heterogeneity (S1 Table). We could not find any difference in the duration of sensory, motor block, or analgesia according to type of BPB approach, but we did identify a dose-response effect of tramadol (50 mg, 100 mg) on the duration of sensory block, motor block, and analgesia.
Tramadol as an adjuvant for BPB in our review shortened the time to onset of sensory block and motor block. These findings are attributed to the potentiating effect of opioids and the peripheral LA-like effect of tramadol. The mechanism underlying the LA effect of tramadol is different from that of LA; the action of LA is generated by blocking Na + channels, but tramadol exerts its effect by blocking K + channels, as does meperidine [34]. A previous study showed that tramadol was as effective as lidocaine when injected subcutaneously in patients undergoing minor superficial procedures [36]. For the variable route of tramadol during BPB with LA, sensory and motor blocks enhanced by a perineural adjuvant to LA, but not by systemic administration (31).
Typical adverse effects of tramadol are headache, nausea, vomiting, dizziness, and sedation when it is used for analgesia (10,31). We could not detect any differences in adverse effects between studies in our meta-analysis, which could reflect low plasma concentrations of tramadol. Use of tramadol as an adjuvant in BPB causes fewer symptoms than does intravenous administration of tramadol (36). There have been no reports of nerve damage attributed to tramadol in animal or human studies. The US Food and Drug Administration has not approved perineural administration of tramadol as it has for dexamethasone.
A recent systematic review of various adjuvants for peripheral nerve block [36] reported results for tramadol that contradict the findings of our systematic review. The authors of that review reported that perineural tramadol had no effect on sensory or motor block, and recommended not using tramadol as an adjuvant in peripheral nerve block. However, their review included only 5 RCTs of tramadol as an adjuvant to LA in BPB [22,24,26,29,31], and omitted many other relevant RCTs [16-20, 23, 25, 28]. Furthermore, they also included RCTs for other types of nerve block, such as psoas block [37] and paravertebral block [38]. Unlike that review of tramadol, we systematically searched for and identified the 16 studies on tramadol used as an adjuvant alone in BPB [16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31], and analyzed the effects of tramadol on sensory block, motor block, and analgesia using systemic meta-analysis software. Generally, the degree of nerve block is determined by the type of nerve, the anatomic site of the nerve, and the type of nerve block.
Our review has several limitations. First, the studies included in the review contained considerably clinical heterogeneity with regard to type of BPB approach, dose and volume of drug, and type of guidance used for BPB. Based on the clinical assumption that different types of BPB may lead to different sensory or motor block characteristics and analgesia. Second, the definitions of outcomes of interest such as time to onset and duration of sensory block, motor block, and analgesia varied widely between the studies. Third, this review pertains to the duration of sensory block, motor block, and analgesia, and highlighted publication bias as ascertained by the trim-and-fill analysis. As a result, the findings of our meta-analysis were influenced by publication bias among the included studies.
However, our review also has several strengths. The main strength is that we tried to include all relevant databases and RCTs in our search. The methodology used was strong, with registration of the protocol for the review on PROSPERO and use of RevMan software.

Conclusions
Our study provides evidence that tramadol 100 mg is a potential adjuvant for use with LA in BPB. Adjuvant tramadol prolonged the duration of sensory block, motor block, and analgesia and shortened the time to onset of sensory block and motor block without altering the incidence of adverse effects.