Safety and efficacy of different administration order on paradoxical reactions following midazolam in children: a randomised, controlled trial


 Study Objective Midazolam is commonly used for Anesthesia induction. However, the paradoxical reactions increase risk of self-injury and bring pressure on anesthesiologist. Although various drugs have been showed effective, but the potential adverse effects limit the application. We compared safety and efficacy of different administration order on the occurrence of paradoxical reactions in children.Design Prospective, randomized, double-blinded study.Setting Operating room of a university-affiliated hospital.Patients 1306 ASA physical status I pediatric patients, aged 6 months to 12 years, scheduled for general anesthesia.Interventions Patients were divided into three groups. In the fentanyl group, patients were given fentanyl followed by midazolam at induction; In the midazolam group patients received midazolam followed by fentanyl at induction; In the blank group, patients were given saline followed by fentanyl at induction. Propofol was administrated 2.5 mg/kg intravenously (IV).Measurements Onset of paradoxical reactions and BIS values were observed and recorded.Main Results Preferentially and slowly injecting fentanyl could decrease the incidence of paradoxical reaction (p<0.05). Children displaying paradoxical reactions were mostly aged 8 months-2 years old (p = 0.001). Bispectralindex values increased significantly when paradoxical reactions burst out (p< 0.05). There were no significant differences in terms of side-effects.Conclusion Changing the administration order was safe and effective to decrease the incidence of paradoxical reactions.

panic, rage, violent behavior, and self-mutilating behavior, 1 which often appears after a brief period of calm state. 2 Several studies have demonstrated that the occurrence of paradoxical reactions following midazolam was associated with injection rate, 3 dose, age, 4 personality and mental factor, 5 genetic background 6 and alcoholism. 7 Therefore, various drugs have been showed effective for paradoxical reaction, such as physostigmine, 8 haloperidol, 9 flumazenil 10-14 and ketamine. 2 But the potential adverse effects limit the application and those agents are only recommended to rescue the most serious side effects of midazolam in children. 14 The mechanism of PR is still unclear.
Several hypotheses have been proposed. It includes cortical de-suppression, diversity of GABA receptor, central cholinergic effect. 15,16 Interestingly, recently studies observed excitatory effect of sevoflurane, which are epileptic spikes on the electroencephalographic (EEG) during anesthesia induction. 17 The possible mechanism is that overloaded GABAergic activity inhibits neuronal chloride channels and eventually leads to inhibition potentials being converted to excitatory potentials. 18 Furtherly, McCarthy et.al indicated interneuron anti-synchrony might provide an explanation for paradoxical reactions generated by propofol. 19 With experience, changing administration order during anesthesia induction seemed to be effective for preventing paradoxical reaction. Therefore, the present study was designed to investigate whether the dosing sequence could affect the incidence of PR.

Study population
We undertook a double-blind, parallel-group, randomised controlled trial of single center.
We included patients aged 6 months to 12 years who were scheduled general anesthesia, 4 but we excluded individuals who had a history of psychological disorders, mental retardation, hypoxia and hypotension before surgery, or gave consent for another interventional study or declined to participate. Written informed consent was obtained from a parent or guardian for participants under 16 years old.
The study was approved by the research ethics boards of Shanghai Children's Hospital, Shanghai, China the protocol number that was attributed by this ethics committee was 2018R012-F01 on 15 March 2018 and the name of the Chairperson of the ethics committee was Guangjun Wu .

Randomisation and masking
Randomization was done by using computer generated random numbers. Patients were randomly assigned in a 1:1:1 ratio to one of three parallel arms to receive different administration order during induction of anesthesia. Patients were blinded to group assignment.
A research nurse placed the random numbers in sealed envelopes and one anesthetist who did not follow the patient, opened the closed envelop to know the random allocation scheme, prepared the drugs, and performed intravenous drug administration.
Judgment of paradoxical reaction was performed by two anesthesiologists, who were engaged in pediatric anesthesia for more than 10 years. Additionally, they remained unaware of the random scheme. The random allocation was also concealed from patients, research staff, and the independent statistician.

Outcomes
The primary outcome was the incidence of the paradoxical reactions. Secondary endpoints included bispectral index values, extubation duration, recovery duration and safety endpoint, which included hemodynamics, fentanyl-induced cough and respiratory depression during the recovery period. 5

Procedures
All the patients were placed an intravenous line before they were brought into the operation room. electrocardiogram (ECG), noninvasive blood pressure, heart rate, temperature, oxygen saturation, exhaled CO² (end tidal CO²) and BIS were monitored. Two investigators performed intravenous induction according the random allocation. The administration order for the three study groups were as following: The fentanyl group: fentanyl 2µg.kg-1, midazolam 0.1mg.kg-1, atropine 0.01 mg.kg-1, propofol 3 mg.kg-1.
Two anesthesiologists, who were engaged in pediatric anesthesia for more than 10 years judged the presence of a paradoxical reaction. Additionally, they remained unaware of the random scheme and did not participant in other parts of the study. Sudden burst of abnormal symptoms were defined as paradoxical reactions, which include extremely panic, rage, violent and self-mutilating behavior. These symptoms are different from separation anxiety and the frightened reaction caused by encountering unfamiliar situation 1 .
Additionally, patients showed temporary calmness before paradoxical reactions burst 1 .
Paradoxical reaction occurred at about 3 minutes after administration of midazolam, 2, 3 so investigators observed the patients for 5 min after administration of midazolam and then given other drug according to the plan. Additionally, the injection rate significantly influences the incidence of paradoxical reactions, so we dosed midazolam at a constant rate of 0. 1mg.s-1 by an intravenous pump. 3 Furthermore, it takes 15 second for plasma 6 concentration of fentanyl reaching the threshold inducing cough 20 , so the duration of administrating fentanyl was more than 30 seconds 21 . Finally, propofol is pumped at a rate of 20 mg·kg -1 ·h -1 . When paradoxical reaction happened, an anesthesia nurse dosed flumazenil 0.01 mg.kg-1 intravenously.

Statistical analysis
The incidence in the midazolam group was reported ranged from 1.4% to 11.1% and it was estimated as 5.4% in our pre-test and 1.5% in the fentanyl group. 1,2,[22][23][24][25][26][27] The sample size was adequate (302 patients for each group) with a = 0.05 and a power of 0.8.
Assuming the drop rate was 10%, 1200 patients were enough. Continuous data were expressed as mean, and frequencies and percentages (%) for categorical data. ANOVA test was used for continuous data (including age, weight, extubation and recovery period,BIS value) and fisher's exact test was used for categorical data (including incidence of PR, sex, patients with Chest wall rigidity, cough and SpO 2 < 95%). All comparisons were two-

Results
The study was reported in accordance with CONSORT statement. From March, 2018, to January, 2019, 1320 patients were randomly assigned, received intended treatment, and fourteen patients were excluded due to missing data.
So 1306 were analysed for the primary outcome including 436 individuals in the fentanyl group, 432 participants in the midazolam group and 438 patients in the saline group.

Comparison of the incidence of PR in the three groups
There were no detectable differences between the three groups in terms of demographic 7 data, table1. Paradoxical reactions were recorded in 5 (1.15%) of 436 patients in the fentanyl group versus 26 (5.98%) of 432 individuals in the midazolam group (odds ratio (OR) 0.18; 95% confidence interval (CI): 0.015-0.22); p < 0.00001, table 2). Once the paradoxical reaction happened, it could not be eliminated by fentanyl but responded rapidly to flumazenil.

Paradoxical reaction and age
A total of 31 individuals experienced paradoxical reactions, among them, 26 patients were aged 8 months to 2 years old (83.9%), two were younger than 8 months (6.5%) and three were older than two years old (9.7%). So, children with paradoxical reactions were most aged 8 months to 2 years old; p< 0.05.

Changes in BIS values
BIS electrodes could not be placed before induction owing to agitation in 106 children in the midazolam group, 114 in the fentanyl group and 157 in the blank group. There was no difference in BIS before and after saline injection (92.1±4.9 vs.91.4±5.7; p = 0.24).

Safety and side-effects
Hemodynamics was stable and there were no fentanyl induced cough and respiratory depression seen in each of the groups. Additionally, there was no significant difference in terms of time to extubation, recovery duration in the three groups (Table 2).

Discussion
In this randomized blind controlled study, we found decreased incidence of PR by changing 8 the order of midazolam and fentanyl administration. However, administration of fentanyl seems invalid to relieve when these reactions have already burst out. Additionally, we found that BIS was significantly higher when children displayed PR.
Paradoxical reactions following midazolam were reported as an incidence varies from 1.4% to 11.1% in different studies. 1,2,[22][23][24][25][26][27] This may be related to heterogeneity in terms of study design, dose and route of administration, purpose of medication, judgment criteria for PR, age, and selectivity bias. 2,14,28 In our study, 5.98% patients occurred PR in the midazolam group and the incidence was significant reduced by preferentially and slowly injecting fentanyl. The mean time of onset of the reaction occurred at 1.8 minutes (standard deviation ±1.4 minutes) after administration of midazolam in our study. The mechanism of paradoxical reactions has not yet been elucidated. Various mechanisms proposed to explain paradoxical reactions. Our study found that BIS was significantly higher when children developed PR and it is in fact with that paradoxical reactions occurred after an initial sedative period of following midazolam administration.
As a GABA receptor agonist, midazolam has sedative and anti-epileptic effects. 29 Activation of the GABAA receptor induces inward chloride currents, hyperpolarizing postsynaptic neurons, ultimately leading to pyramidal neuronal suppression and GABAA receptor-mediated inhibition is essential for inhibition of brain excitability. 30 However, the midazolam-induced excitatory response seems to contradict the GABAA receptor-mediated inhibition of midazolam. Interestingly, sevoflurane and propofol can also induce excitatory response, which are epileptic spikes on the electroencephalogram during anesthesia induction. 17 Similarly, excitability in the cerebral cortex appears to be conflict with the strong GABAergic effect of sevoflurane and propofol. Recently studies showed that all anesthetics that enhance GABAergic neurotransmission have both inhibitory and 9 convulsive effects because overloaded GABAergic activity inhibits neuronal chloride channels and eventually leads to inhibition potentials being converted to excitatory potentials, which we named the double-sided effect of GABAergic neurotransmission. 18 This double-sided effect may explain why midazolam causes contradictory reactions. Even more interesting is that epileptiform discharges during anesthesia induction are similarly to the PR of midazolam. On one hand, we found that PR usually occurs after a period of sedation in our study. BIS usually drops first and then rises sharply when PR occurs. This is very similar to the appearance of suppression with spike in EEG, which is outbreak of brain waves after a period of brain wave suppression. 19 On the other hand, suppression with spikes is only seen during anesthesia induction, which varies from other types of epileptic discharges, such as interictal spike events, also seen in other functional brain disorder. (E.g. ADHD, migraine, Epilepsy). 27,31,32 Finally, PR and epileptiform discharges both belong to hyperexcitability. 28 Hence the mechanism of PR might be further clarified by EEG.
Our study has several limitations. First, the judgment of PR is objective and might lead to measurement bias. Fortunately, judgment of paradoxical reaction was performed by two anesthesiologists, who were engaged in pediatric anesthesia for more than 10 years.
Secondly, most anesthesiologists do not induce with midazolam, however, many sites still premedication midazolam, which might also cause paradoxical reactions. In the present study, preferentially and slowly administrated fentanyl was helpful to prevent occurrence of PR, which was beneficial to patients with malignant hyperthermia or susceptibility, patients allergic to fluoride, and hospitals in developing countries.

Conclusion
In summary, we observed that preferentially and slowly injecting fentanyl during anesthesia induction resulted in lower incidence of paradoxical reactions following midazolam. BIS was significantly higher when children displayed PR. There were no significant differences in terms of extubation duration, recovery duration and safety endpoint between fentanyl group and midazolam group.

Acknowledgements
The study was carried out at Department of Anesthesiology, Shanghai Children's Hospital (Shanghai, China). The authors acknowledge the support of Shanghai Children's Hospital for providing infrastructure and funding to ensure that all patients at this center are treated free of costs with no commercial or financial gains to any member of the team.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Availability of data and materials
The datasets during the current study are available from the corresponding author on reasonable request.  Age data is expressed as means ± SD. There were no statistically significant differences in the demographic data between the three groups. Fentanyl group received fentanyl followed by midazolam; Midazolam group received midazolam followed by fentanyl; Blank group received saline followed by fentanyl.