Lidocaine vs. Other Local Anesthetics in the Development of Transient Neurologic Symptoms (TNS) Following Spinal Anesthesia: A Meta-Analysis of Randomized Controlled Trials

The use of lidocaine in spinal anesthesia may increase the risk of transient neurological symptoms (TNS) according to previous meta-analyses. However, the previous meta-analyses lacked data on some other local anesthetics and thus, more evaluations are still needed to compare the effect of lidocaine on the development of TNS. The objective of this study was to compare the risk of TNS according to lidocaine versus other local anesthetics in patients undergoing spinal anesthesia. A total of 39 randomized controlled trials with 4733 patients were analyzed. The incidence of TNS was 10.8% in the lidocaine group and was 2.2% in the control groups (risk ratio (RR) 4.12, 95% confidence interval (CI) 3.13 to 5.43, p < 0.001). In subgroup analysis, lidocaine increased the incidence of TNS compared with other local anesthetics except mepivacaine, ropivacaine or sameridine. The risk of TNS was higher in the hyperbaric (p < 0.001) or isobaric lidocaine (p < 0.001) group compared with the control group, but there were no differences found between the two groups when hypobaric lidocaine was administered (p = 1.00). This study confirmed that lidocaine for spinal anesthesia still causes TNS more frequently than most other local anesthetics, especially when hyperbaric or isobaric lidocaine was used.


Introduction
Lidocaine is an attractive regional anesthetic for ambulatory surgery. It offers a rapid onset and fast recovery of both motor and sensory block [1]. However, when compared with other local anesthetics, the use of lidocaine in spinal anesthesia has been known to be associated with increased risk of transient neurological symptoms (TNS) [2,3], hindering its application in ambulatory spinal anesthesia. Other local anesthetics including mepivacaine, low-dose bupivacaine, procaine, articaine, levobupivacaine, ropivacaine and 2-chloroprocaine have been suggested as replacement drugs.
TNS generally occurs in patients with single injection spinal anesthesia within the first 24 hours [2]. TNS consists of pain in the lower extremities without abnormalities in neurologic and radiologic examination [2]. In previous systematic reviews and meta-analyses [2,3], lidocaine has a significantly higher relative risk of developing TNS compared with most other local anesthetics, but the previous meta-analyses lacked data on ropivacaine, levobupivacaine, and chloropocaine; thus, more evaluations

Data Synthesis and Statistical Analysis
Data synthesis and meta-analysis were performed using Revman 5.3 software (Cochrane Collaboration, Oxford, UK) and R version 3.6.1 (R Foundation for Statistical Computing, Vienna, Austria). Since the incidence of TNS was dichotomous variable, the authors calculated the risk ratio (RR) as a pooled estimate. The inverse variance method and random effect models were employed. A continuity correction of 0.5 was applied to zero total events trials [21]. The findings were presented as a forest plot with 95% confidence intervals. A subgroup analysis was conducted according to the local anesthetics which were used in the control group. Additional subgroup analyses were carried out to investigate any relationship between the incidence of TNS and baricity/concentration. The heterogeneity among the studies was evaluated by I 2 statistic. I 2 could be interpreted in the following manner: 0% to 40% might not be important; 30% to 60% may represent moderate heterogeneity; 50% to 90% may represent substantial heterogeneity; and 75% to 100% considerable heterogeneity [22]. A publication bias was assessed by construction of a funnel plot and the linear regression test. Sensitivity analysis (leave one study out) was conducted to confirm the robustness of the results.

Methodology Quality and Risk of Bias
The methodology quality and risk of bias are summarized in Figure 2. In each study, all patients were randomized to receive intrathecal lidocaine or other local anesthetics; however, the randomization method was unclear in eight studies. Twenty one studies maintained allocation concealment, but the other studies failed to describe it clearly. The risk of performance bias was high in 9 studies and unclear in 20 studies. It might be important for anesthesiologists to be aware of drugs for patient safety and sufficient block when performing spinal anesthesia. Unlike performance bias, the risk of detection bias was low overall. The risk of attrition bias, reporting bias, and other biases were low in more than 75% of the studies evaluated. Reasons for each risk of bias are shown in Table S2 used for spinal anesthesia and the results of three sameridine groups were combined. Patients who received general anesthesia due to insufficient spinal block were excluded from the final analysis because it was unclear whether local anesthetics were administered in the cerebrospinal fluid. Details of each trial are summarized in Table 1. The 5% hyperbaric lidocaine was most used, followed by 2% isobaric. The dose of lidocaine used in each trial varied from 10 to 100mg. Specific details, including concentration, baricity, doses and adjuvants of study drugs, are summarized in Table 2.

Methodology Quality and Risk of Bias
The methodology quality and risk of bias are summarized in Figure 2. In each study, all patients were randomized to receive intrathecal lidocaine or other local anesthetics; however, the randomization method was unclear in eight studies. Twenty one studies maintained allocation concealment, but the other studies failed to describe it clearly. The risk of performance bias was high in 9 studies and unclear in 20 studies. It might be important for anesthesiologists to be aware of drugs for patient safety and sufficient block when performing spinal anesthesia. Unlike performance bias, the risk of detection bias was low overall. The risk of attrition bias, reporting bias, and other biases were low in more than 75% of the studies evaluated. Reasons for each risk of bias are shown in Table  S2.  [20]. The randomization and allocation concealment were well performed in most studies, but the method was not described in several studies. The performance bias was unclear or high in most studies, whereas the detection bias was low overall. The risk of attrition bias, reporting bias and other biases were low in most studies. Abbreviations: + = low risk of bias, ? = unclear risk of bias, -= high risk of bias.

Outcome Synthesis
A total of 39 studies included 4733 patients; 2209 patients were allocated to the lidocaine group and 2524 patients to the control group. The incidence of TNS was 10.8% (238/2209) in the lidocaine group and was 2.2% (56/2524) in the control group. The risk of TNS after spinal anesthesia was significantly higher in the lidocaine group than in the control group (Risk ratio (RR) = 4.12, 95% confidence interval (CI) = 3.13 to 5.43, p < 0.001), with a low level of heterogeneity (I 2 = 0%, p = 0.61) (Figure 3). A Figure 2. Risk of bias summary and graph using the Cochrane Risk of Bias tool [20]. The randomization and allocation concealment were well performed in most studies, but the method was not described in several studies. The performance bias was unclear or high in most studies, whereas the detection bias was low overall. The risk of attrition bias, reporting bias and other biases were low in most studies. Abbreviations: + = low risk of bias, ? = unclear risk of bias, -= high risk of bias.

Outcome Synthesis
A total of 39 studies included 4733 patients; 2209 patients were allocated to the lidocaine group and 2524 patients to the control group. The incidence of TNS was 10.8% (238/2209) in the lidocaine group and was 2.2% (56/2524) in the control group. The risk of TNS after spinal anesthesia was significantly higher in the lidocaine group than in the control group (Risk ratio (RR) = 4.12, 95% confidence interval (CI) = 3.13 to 5.43, p < 0.001), with a low level of heterogeneity (I 2 = 0%, p = 0.61) (Figure 3). A symmetrical funnel plot and linear regression test showed insignificant results for publication bias (p = 0.206) ( Figure S1). Sensitivity analysis revealed the robustness of the results (Table S3). Omitting one study [28] decreased the RR to 3.51, but still maintained the significance.

Figure 4.
Forest plot for subgroup analysis. A subgroup analysis was conducted according to the local anesthetics which were used in the control group. The incidence of TNS was significantly higher in the lidocaine group than in the bupivacaine (p < 0.001), levobupivacaine (p < 0.001), prilocaine (p = 0.001), chloroprocaine (p = 0.037), procaine (p = 0.003) and articaine group (p < 0.001). However, there were no differences in the risk of TNS between the lidocaine group and mepivacaine (p =0.728), ropivacaine (p = 0.052) and sameridine group (p = 0.545). Abbreviations: LDC = lidocaine, LA = local anesthetics, CI = confidence interval.

Discussion
The present meta-analysis confirmed that lidocaine used for spinal anesthesia still causes TNS more frequently than most other local anesthetics (bupivacaine, levobupivacaine, prilocaine, chloroprocaine, procaine, and articaine) except for mepivacaine, ropivacaine or sameridine. This is the first study that analyzed the role of baricity and concentration of lidocaine as potential risk factors for TNS, and the subgroup analysis showed that hyperbaric and isobaric lidocaine showed higher TNS rates than the others, and higher rates of TNS have been observed in all concentration categories of lidocaine.
The incidence of TNS of lidocaine was about 4 times higher than that of other local anesthetics in this study, which supported the result of the previous meta-analysis by Zaric et al. [2] which included 16 RCTs with a total of 1467 patients. Since then, many RCTs have been still conducted to compare the incidence of TNS between lidocaine and other local anesthetics. A recent meta-analysis including 24 studies with 2226 patients compared the risk of TNS by using direct and indirect comparison [3]. However, in this meta-analysis, more RCTs, including 39 studies with 4733 patients (more than twice) were analyzed and lidocaine was compared with more various local anesthetics. However, the risk ratio and prevalence of developing TNS of the current study was slightly lower than those of the previous studies; Zaric [3]. These difference can be explained in part by that many RCTs of the present meta-analysis reported no case of TNS in the lidocaine group [4][5][6][7][8][9][10][11][12][13][14][15][16][17][18], and a continuity correction of 0.5 was applied to these zero total events trials to prevent the overestimation of the risk of TNS [21]. Among other local anesthetics, chloroprocaine and mepivacaine have similar characteristics with lidocaine in terms of rapid onset time and short duration [47,48]. Subgroup analysis suggested that the incidence of TNS of chloroprocaine was lower than that of lidocaine. This finding significantly differs from the previous results [2,3]. However, the previous meta-analyses included only one or two RCTs comparing the effect of lidocaine and chloroprocaine. This study included four RCTs and no case of TNS in the chloroprocaine group was reported [17,24,26,46]. This would appear to indicate that chloroprocaine may be an attractive alternative to lidocaine for the short ambulatory surgery with fast onset and quick recovery time [48]. On the other hand, there was no difference in the incidence of TNS between lidocaine and mepivacaine, which was consistent with the result of the previous studies [2,3]. The idea that ropivacaine could decrease the development of TNS is still controversial. Although Zaric et al. [2] found that there was no difference in the risk of TNS between lidocaine and ropivacaine, Forget et al. [3] found that ropivacaine could decrease the risk of TNS than lidocaine. However, as expected, two studies included smaller number of studies and the latter study estimated pooled effect size by mostly indirect comparison. Surprisingly, the present study found more studies comparing the effect of lidocaine and ropivacaine, and estimated the pooled effect size by using a direct comparison.
Subgroup analysis suggested that no cases of TNS were found in the hypobaric lidocaine group whereas previous studies showed that the TNS of the lidocaine group occurred regardless of the baricity and concentration [49][50][51]. This result can be explained by low doses (10-20 mg) of hypobaric lidocaine group administered. Ben-David et al. [52] also reported that small doses of hypobaric lidocaine reduced the risk of TNS more than large doses of hypobaric lidocaine. However, this needs to be interpreted with caution since low doses of local anesthetics may be insufficient for adequate regional block [53]. Regarding the concentration of lidocaine, higher rates of TNS have been observed in all categories of concentration, which confirms the previous finding that altering the lidocaine concentration had no influence on the prevention of TNS [54].
This meta-analysis has a few limitations. First, various definitions of TNS were used in each study. Generally, TNS is defined as pain originating in the gluteal region and radiating to both lower extremities [2]. Some studies included considered TNS as only pain [4], while several other studies regarded TNS as pain and abnormal sensation (hypoesthesia or dysesthesia) [26,30]. Moreover, the anatomical regions (back, thigh, buttock or lower extremity), involving TNS, varied in each study. Furthermore, some studies did not specify details and/or a definition of TNS [35,47]. This variance with respect to TNS may have created bias, influencing the exact frequency of TNS. Second, specific types of surgery and position may be considered as risk factors of TNS, such as knee arthroscopy and lithotomy position. In the present study, various types of surgery and surgical position were included, and this may induce a bias in the results. Subgroup analysis according to the surgical position may provide a better overview. However, surgical position is heterogeneous and is not defined in some trials and the subgroup analysis according to the position, which may induce inaccurate results with bias. Third, only one study compared lidocaine to articaine with 134 patients, which may not be enough to conclude that the frequency of TNS with articaine is less than with lidocaine. Similarly, one RCT compared sameridine with lidocaine with 140 patients and there were no cases of TNS in the sameridine group.

Conclusions
In conclusion, the risk of developing TNS after spinal anesthesia with lidocaine was significantly higher than with bupivacaine, levobupivacaine, prilocaine, chloroprocaine, procaine or articaine. In addition, hyperbaric and isobaric lidocaine showed higher TNS rates than other lidocaines.