We present the results of literature search in a PRISMA flow chart (Fig. 1). 3789 records were identified through database searching. After duplicating, screening the initial title and abastract,81 full texts were retrieved for comprehensive evaluation. Finally, 18 RCTs were included which contained 1175 patients met the inclusion criteria in the present meta-analysis included in the final analyses.
The patients of these 18 RCTs whose average age older than 60. These articles contained 595 patients in lidocaine groups and 580 patients who received normal saline or nothing served as control. Studies were carried through in patients undergoing abdominal, cardiac, orthopedic, urinary, endoscopic surgery. Nine of the studies reported surgery time, and all studies mentioned duration of infusion. A lidocaine bolus (1-2mg/kg) was given before surgical incision in 17 of the 18 RCTs whereas all eligible RCTs followed by a continuous infusion during and after operation. The characteristics of the included trials were summarized in Table 1.
Risk of bias within studies
We evaluated the quality of the included RCTs on the basis of the Cochrane Handbook of Systematic Reviews of Interventions, version 5.1.0 (updated March 2011).The entire risk of bias concerning selection bias, performance bias, attrition bias, detection bias and other bias exposed low risk of bias in more than 65% across all included all studies (Fig. 2 and Fig. 3).
1. Postoperative pain
Postoperative pain was evaluated by different pain scores. The pain scores were provided in totally 11 studies of 18 RCTs which had sufficient quantitative data to be combined through meta-analysis [19-29].Five studies assessed postoperative pain on a visual analogue scale (VAS) from 0 to 100mm and five studies used the VAS from 0 to 10cm measurement, while one trial evaluated postoperative pain with NRS.
1.1 Pain score at rest
As shown in Fig. 4, postoperative pain scores at rest were evaluated in 10 studies[19-28].Meta-analysis of pain data at rest was statistically significant decrease VAS scores in lidocaine group compared with the control group at 2h (7 studies, n=386, SMD -1.30 95%CI -1.90 to -0.70), at 4h (7 studies, n=364, SMD -1.20 95%CI -1.91 to -0.49), at 6h (3 studies, n=208, SMD -0.87 95%CI -1.72 to -0.02), at 8h (5 studies, n=284, SMD -0.84 95%CI -1.40 to -0.27), at 12h (4 studies, n=267, SMD -0.73 95%CI -1.14 to -0.32), at 18h (3 studies, n=208, SMD -0.79 95%CI -1.78 to 0.20), at 24h (10 studies, n=569, SMD -0.39 95%CI -0.66 to -0.11), at 48h (8 studies, n=458, SMD -0.25 95%CI -0.59 to 0.10). The outcome adequately revealed that perioperative intravenous lidocaine infusion can reduced postoperative pain scores in elderly patients of abdominal, cardiac, orthopedic, urinary, endoscopic surgery at 2h, 4h, 6h,8h, 12h, 24h after surgery, but did not decreased the pain scores at 18h and 48h after surgery.
1.2 Pain score at movement
There were 4 trials reported the pain scores postoperatively at movement[22-24, 29].The pain score at movement after surgery was decreased with intravenous lidocaine at 2h (2 studies, n=151, SMD -1.02 95%CI -2.05 to -0.00), 12h (2 studies, n=151, SMD -0.84 95%CI -1.18 to -0.51), one study of Moeen demonstrated yield similar effect at 6h (n=111, SMD -1.42 95%CI -1.84 to -1.01) and at 18h (n=111, SMD -0.70 95%CI -1.09 to -0.32); although there was no statistical differences at 4h (1 studies, n=40, SMD -0.58 95%CI -1.21 to 0.05), at 24h (4 studies, n=277, SMD -0.10 95%CI -0.51 to 0.30), at 48h (4 studies, n=277, SMD -0.15 95%CI -0.56 to 0.26).The detailed result are showed in Fig. 6.
We performed the robust of the evidence by sensitivity analysis, The difference of the combined effect was detected by different effect model. The estimated effect(95%CI) for the sensitivity analysis remained robust about the pain scores at different point times from 2h to 48h after surgery when selecting random effect model or fixed effect model, which showed that the results were reliable.
2.Length of hospital stay
17 trials[19-24, 26-36] of 18 RCTs reported the length of hospital stay between groups. The overall pooled results from the meta-analysis showed that lidocaine markedly shortened the time of hospital stay compared with the control group. Intervention substantial heterogeneity was noted (MD -0.30 95%CI -0.50 to -0.09); forest plot is shown in Fig. 7.
In 3 studies[26, 32, 34]of the 18 RCTs, lidocaine in cardiac surgery as the intervention was effective in shorting duration of hospital stay (MD -0.71 95%CI -1.37 to -0.05); Nevertheless, results from 8 studies[19-22, 28, 29, 35, 36] in gastrointestinal surgery were not shown to reducing length of hospital stay (MD -0.20 95%CI -0.54 to 0.14) ;6 studies[23, 24, 27, 30, 31, 33] in other surgery revealed yielded lidocaine infusion can decreased the length of hospital stay (MD -0.30 95%CI -0.50 to -0.09).
1. Postoperative opioid consumption
Eleven articles with 552 patients reported the outcome of postoperative opioid consumption. Morphine was applied for postoperative pain relief in 6 trials[20, 23, 25, 27, 30, 33], 4 trials[19, 21, 26, 28] applied fentanyl, one trial offered meperidine. Pooled meta-analysis revealed the lidocaine group had fewer opioid requirements (MEQ, mg) after surgery in comparison to the control group (SMD -0.31 95%CI -0.61 to -0.01).
2. Gastrointestinal recovery
Six studies[20, 25, 29, 30, 33, 35] with 285 patients reported data on the time to first flatus and the time to first defaecation/bowel movement.
2.1 Time to first flatus
The combined analysis did not demonstrate a significant reduction of time to flatus in lidocaine group compared with the control group (MD -3.24 95%CI -9.90 to 3.41).
2.2 Time to first defaecation/bowel movement
There was enough evidence of effect for lidocaine infusion in contrast with the control group to shorten the time to first defaecation/bowel movement (MD -8.64 95%CI -16.00 to -1.28).
3. Postoperative neuropsychological function status scale
Only one studies proved lidocaine administration during and after surgery does not reduce the high rate of postoperative cognitive dysfunction.
4. Adverse events
The complication of nausea was reported in 9 trials[19, 21, 22, 27-29, 33, 35, 36],whereas three articles[19, 21, 28] showed the complication of vomiting. It illustrated respectively that there was significant reduction in nausea in lidocaine group contrasted with control group (n=479 OR 0.56 95%CI 0.37 to 0.84). There was no statistical difference about vomiting between lidocaine group and control group (n=120 OR 0.47 95%CI 0.19 to 1.1) but there was a slightly lower incidence of vomiting in lidocaine group when compared with the control group.
Risk of bias across studies
The funnel plot was used to investigated the risk of bias across studies (publication bias). For outcome of postoperative pain, the funnel plot at 24h after surgery suggested funnel plot asymmetry on visual inspection, publication bias was likely to be present (Fig. 5). The funnel plot for the duration of hospital stay suggested funnel plot was symmetrical on visual inspection, indicating that publication bias was not likely to be present (Fig. 8).
Additional analyses such as subgroup analysis, sensitivity analysis and publication bias analysis were performed to evaluate which surgery may benefit from IV lidocaine administration (subgroup analysis), the robust of result in terms of risk of bias from single study (sensitivity analysis) and the appearance of risk of bias across analysis (publication bias). By this means we researched the primary outcome, such as subgroup analysis for length of hospital stay, sensitivity analysis for postoperative pain at 24h after surgery and publication bias analysis for length of hospital stay and postoperative pain at 24h after surgery.
- Subgroup analysis. The preplanned subgroups analysis on type of surgery accounted for heterogeneity about all subgroups (Fig. 7). Subgroup analysis was conducted and it was shown that IV lidocaine short the length of hospital stay only in cardiac surgery subgroup and other surgery subgroup compared to control, however, in gastrointestinal surgery subgroup, it was not statistical difference, which may be resulted by the small effect of IV lidocaine in shortening length of hospital stay for aged patients and a weaker physiological state in old folks , however, heterogeneity of three group was accepted (I2<50%). This result indicated that aged patients undergoing cardiac surgery may more likely benefit from perioperative IV lidocaine and the debate that whether perioperative IV lidocaine shorten the length of hospital stay in aged patients undergoing gastrointestinal surgery may resolved by more high qualities and large simple size RCTs in the future.
- Sensitivity analysis. We tested the robust of the evidence by sensitivity analysis in the outcome of postoperative pain score due to the heterogenicity was unacceptable (I2>50%). The difference of the combined effect was detected by comparing fixed effect model and random effect model. The estimated effect(95%CI) for the sensitivity analysis remained robust about the pain scores at 2h, 4h, 6h, 8h, 12h, 24h, 48h after surgery when selecting different effect model, which showed that the results were reliable. However, the estimated effect(95%CI) for the sensitivity analysis did not remained robust at 18h (RE model: SMD -0.79 95%CI -1.78 to 0.20 VS FE model: SMD -0.69 95%CI -0.98 to -0.41). The outcome adequately revealed the result of postoperative pain score at 18h was unreliable.