Introduction

Sevoflurane has been introduced into clinical practices as an inhaled volatile anesthetic since 1992. This medication is particularly effective for inhalation induction and maintaining the effects of general anesthesia on pediatric patients due to its inherent stability, minimal respiratory pungency and minimal blood-gas partition coefficient1. Another advantage of sevoflurane is its ability to rapidly induce anesthetic effects in a controllable manner once injected.

Unfortunately, postoperative behavioral disturbance was predominantly observed in patients who received pediatric surgeries accompanied by sevoflurane as anesthetic. Another major issue caused by sevoflurane is the significant increase in the incidence of emergence agitation (EA). For instance, the incidence of emergence agitation was increased from 12–13% to 56% when sevoflurane was introduced as the main agent2,3.

Emergence agitation resulted from general anesthesia is usually characterized by either disorientation or abnormal excitation during the early stage of patient recovery. However, more severe symptoms such as sympathetic activation and arrhythmia are likely to be observed, which may further impede the recovery of patients. Some researches argued that the toxicity of sevoflurane may affect the central nervous system and trigger EA, while others suggested that other factors including age may contribute to EA4. Since sevoflurane is likely to induce EA in certain circumstances, prophylactic medicine has been introduced into sevoflurane in order to enhance the recovery of patients and reduce the risk of postoperative behavioral disturbance. Conventional prophylactic medicine includes sedative-hypnotic, opioid receptor agonist and narcotic analgesic and they have been introduced into sevoflurance in clinical practices. On the other hand, treatments for preventing EA include midazolam, dexmedetomidine, clonidine, ketamine, propofol and fentanyl and they appear to have significant difference in pharmacological characteristics. As a result, the effectiveness and safety of these treatments should be verified in clinical practices.

This study enabled us to compare the effectiveness and safety of placebo, midazolam, dexmedetomidine, clonidine, ketamine, propofol and fentanyl which are commonly introduced as prophylactic treatments. We incorporated various endpoints in our study so that both direct and indirect comparison can be comprehensively achieved.

Materials and Methods

Two phases were involved in this study. Phase one was collecting all the articles about the efficacy and safety of seven auxiliary medications that are introduced into pediatric sevoflurane anesthesia. Phase two was meta-analysis on a select group of these techniques.

Search strategy

Articles complied with the selection criteria were thoroughly searched, including PubMed, Embase and other databases. The following keywords and searching terms including their corresponding synonyms were used to retrieve the corresponding articles according to standard PICOS (population, intervention, comparison, outcome, study design) criteria: pediatric anesthesia (population), clonidine, dexmedetomidine, fentanyl, ketamine, midazolam, propofol (intervention and comparison) and randomized controlled trial (study design), emergence agitation (primary outcome).

Inclusion and exclusion criteria

Literature inclusion criteria: (1) researching type: randomized controlled trials; (2) researching objects: children between the age of six months and fourteen years who received sevoflurane as anesthetic (3) interventions: single or mixed clonidine, dexmedetomidine, fentanyl, ketamine, midazolam, propofol; (4) outcomes contain at least one of the followings: EA, postoperative nausea and vomiting (PONV), the number of patients who required analgesic, pediatric anesthesia emergence delirium (PAED), the extubation time, the emergency time and the duration of PACU stay. Literature exclusion criteria: (1) non-randomized controlled trials; (2) research objects were not complied with the inclusion criteria; (3) literatures which were not written in English; (4) duplicated literatures which were published by the same author; (5) literatures in which data integrity cannot be guaranteed. A Jadad Scale table concerning randomization, blinding and withdraw was used as an appendix to qualify the included papers (Table S1).

Outcome measures and data extraction

Data extraction was performed using a standard approach: two researchers (W. C. Wang and P. Huang) extract the corresponding data from literatures independently including the sample size and data integration was also carried out for each study. The number of paper included varied between researchers, and difference in data extraction was used for correction. Any disagreement or different opinions with respect to data extraction and integration was resolved by a third researcher (X. L. Guo).

Statistical analysis

First, we accomplished a conventional meta-analysis on the selected data. Odds ratios (OR) were selected as the appropriate statistics for comparing binary outcomes whereas standardized mean difference (SMD) were selected for comparing continuous outcomes. Apart from that, the 95% CI were also obtained in order to assess the precision of the corresponding statistics. Heterogeneity across studies was assessed by the statistic of I2 and significant heterogeneity was presented if I2 > 50%. The fixed-effect model was implemented if studies are homogeneous in nature (P-value of heterogeneity >0.05). By contrast, the random-effects were chosen in the case of significant heterogeneity (P-value for heterogeneity <0.05).

Moreover, the network meta-analysis was conducted in the same manner and the surface under the cumulative ranking curve (SUCRA) in order to rank the corresponding interventions. SUCRA, a transformation of the mean rank, provides a hierarchy of treatments and accounts for the location and variance of clinical outcomes. Higher accumulative SUCRA values indicate better treatment ranks, which is equal to 1 when the treatment is certain to be the best.

Results

Literature search results

We identified a total of 1,598 publications and 537 of them were removed since they are either duplicated literatures, comments, letters and case reports. Another 605 publications were removed since they were not related to the research topic and 411 of the remaining articles contain incomplete data. As a result of this, 45 articles published from 1999 to 2015 were complied with the selection criteria (Fig. 1)5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49. A total of 4,032 cases were included and the detailed baseline characteristics of the included studies were displayed in Table 1. A Jadad Scale table concerning randomization, blinding and withdraw was used as an appendix to qualify the included papers (Table S1).

Table 1 Main characteristics of included studies.
Full size table
Figure 1
figure 1

The flow chart of literature selection.

Full size image

Conventional meta-analysis

We carried out conventional meta-analysis to compare the efficacy and safety of seven auxiliary medications that are introduced into pediatric sevoflurane anesthesia (Table 2). Clonidine, dexmedetomidine, fentanyl and ketamine and propofol significantly reduced the risk of EA (Fig. 2). The same approach was adopted to evaluate the relative safety of these auxiliary medications compared to placebo. Both clonidine and dexmedetomidine were associated with a decrease in the risk of PONV. Furthermore, patients with dexmedetomidine experienced a reduced risk of sedative. Fentanyl exhibited less favorable results than the placebo with respect to PONO, the emergency time and the duration of PACU stay. However, Fentanyl showed compelling results with respect to RA and PAED. Ketamine exhibited convincing results in both PAED and the emergency time. We also observed that patients treated with clonidine, dexmedetomidine, fentanyl and midazolam and propofol exhibited significantly longer emergency response time compared to placebo. Patients treated with propofol were associated with a downward trend of RA and PAED.

Table 2 Pair-wise meta-analyses of direct comparisons between the six drugs and placebo.
Full size table
Figure 2
figure 2

The forest plot of different treatment on emergence agitation from network meta-analysis.

Full size image

Network meta-analysis

We also carried out pair wise comparisons among these medications through network meta-analysis Table 3: patients treated with clonidine, dexmedetomidine, fentanyl, ketamine and propofol were less likely to have EA. Fentanyl exhibited the least favorable results with respect to PONV compared to the other six auxiliary medications whereas clonidine and dexmedetomidine exhibited more compelling results than placebo. Additionally, dexmedetomidine, fentanyl, ketamine and midazolam were less likely to result in sedatives use compared to placebo. Our study also demonstrated that dexmedetomidine, fentanyl and ketamine significantly reduced the average PAED in comparison to placebo and dexmedetomidine appeared to be more effective than clonidine with respect to PAED.

Table 3 The efficacy (emergence agitation) and tolerability (PONV, RA, PAED, extubation time, emergency time, duration of PACU stay) of six treatments according to the network meta-analysis using odds ratios (ORs) or standard mean differences (SMDs) and corresponding 95% credible interval (CrI).
Full size table

Besides, we compared the average extubation and emergency time for determine the overall safety of these medications. Patients treated with dexmedetomidine exhibited significantly longer extubation time compared to those who were given placebo. On the other hand, patients treated with clonidine, dexmedetomidine, fentanyl and ketamine and midazolam exhibited significantly shorter emergency time compared to those treated with propofol, and ketamine group had significantly shorter average emergency time compared to the midazolam group. The comparison of duration for the corresponding treatments revealed that both clonidine and fentanyl demonstrated relatively longer duration of PACU stay compared to placebo whereas such a figure in the propofol group is significantly shorter than that in the clonidine group.

The corresponding SUCRA values of seven pediatric sevoflurane anesthesia auxiliary medications with respect to each efficacy and safety endpoint were illustrated in Table 4, Fig. 3 and Figs S1a–S6a. Fentanyl had the highest cumulative ranking probability with respect to EA and PAED (EA, 88.8%; PAED, 83.9%) whereas both ketamine and dexmedetomidine demonstrated robust results with respect to EA (70.5% and 66.7%, respectively); clonidine exhibited the most compelling SUCRA values with respect to PONV and RA (PONV, 91.6%, RA, 75.0%) and ketamine ranked the best with respect to the emergency time (96.0%). More importantly, placebo exhibited the highest cumulative ranking probability with respect to the extubation time and PACU, therefore other medications may trigger several adverse effects which are reflected by longer extubation time and PACU (Extubation Time, 80.7%; the PACU, 92.2%).

Table 4 Relative ranking of six drugs assessed by estimated and predictive probabilities using SUCRA values.
Full size table
Figure 3
figure 3

The cumulative ranking probabilities of different treatment on emergence agitation.

Full size image

Discussion

In current study, we conducted a network meta-analysis to compare the relative efficacy and safety of six prophylactic treatments including clonidine, dexmedetomidine, fentanyl, ketamine, midazolam and propofol. Our results showed that fentanyl, ketamine, and dexmedetomidine are significantly associated with a lower risk of EA and PAED together with enhanced effectiveness compared to the placebo. It appears that dexmedetomidine is more appropriate than others and such a conclusion is supported by Fang et al. reporting that dexmedetomidine was the most appropriate medication with respect to EA prevention50.

One potential explanation for above conclusion is that dexmedetomidine is an α(2)-adrenoceptor agonist with several analgesic, anxiolytic and sedative properties. It is suspected that these properties may enhance the hemodynamic stability, hence contributing to risk reduction of EA51,52. It is acknowledged that pain relief medicine is able to reduce anesthesia-related EA effectively23,29,53. However, some researchers argued that the use of general analgesic is not effective in reducing the risk of EA54. Dahmani et al. demonstrated that the sedation triggered by dexmedetomidine played a key role in reducing the risk of EA during the recovery period55. Therefore, we suspect that the reduction in the risk of EA is likely to be triggered by the analgesic and anxiolytic roles of dexmedetomidine. Apart from that, dexmedetomidine has somehow neuroprotective effects which are able to reduce neurocognitive impairment resulted from anesthetics56. Meanwhile, Robert et al. reported that the neuroprotective effect of dexmedetomidine resulted from the increase of expression levels of Mdm2 and Bcl-2, up-regulating the neurotrophic factor-Cyclic AMP response element-binding protein (BDNF-CREB) and activating the ERK signaling pathways57,58,59.

This study demonstrates that fentanyl is particularly more effective than dexmedetomidine in reducing the risk of EA and PAED. As suggested by Fenmei et al., fentanyl is able to reduce the risk of EA in a non-specific way regardless of its undiscovered relationship with postoperative pain and EA60,61,62. This may be explained by the fact that fentanyl has a durable analgesic and sedative effect. However, fentanyl has excitatory effects on the gastrointestinal smooth muscle and both patients in the fentanyl group are more likely to experience PONV and RA compared to those in the dexmedetomidine group. Furthermore, the effect of ketamine on risk reduction of PAED and EA is almost equal to that contributed by dexmedetomidine which is consistent with a study conducted by Dahmani et al.55 ketamine is an aspartate receptor antagonist which not only exhibits similar sedative and hypnotic effects to those of dexmedetomidine but also contain strong analgesic effects63,64,65,66.

This study is a network meta-analysis which compares different types of prophylactic treatments including clonidine, dexmedetomidine, fentanyl, ketamine, midazolam, and propofol. However, some limitations should be further addressed by future researchers due to the nature of network meta-analysis. For instance, there may be significant variations with respect to design, sample size and patient selection which cannot be incorporated by our network meta-analysis. Apart from that, the unequal number of interventions for each endpoint did not enable us to carry out a cluster analysis. In summary, our findings suggested that dexmedetomidine should be considered as the most appropriate prophylactic treatment that can be introduced into sevoflurane anesthesia. We recommend researchers to carry out specific following-up studies so that the long-term effects of these interventions can be discovered.

Additional Information

How to cite this article: Wang, W. et al. Efficacy and Acceptability of Different Auxiliary Drugs in Pediatric Sevoflurane Anesthesia: A Network Meta-analysis of Mixed Treatment Comparisons. Sci. Rep. 6, 36553; doi: 10.1038/srep36553 (2016).

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.