Preoperative anxiety in preschool children: A randomized clinical trial comparing midazolam, clonidine, and dexmedetomidine

Anxiety in pediatric patients may challenge perioperative anesthesiology management and worsen postoperative outcomes. Sedative drugs aimed to reducing anxiety are available with different pharmacologic profiles, and there is no consensus on their effect or the best option for preschool children. In this study, we aimed to compare the effect of three different premedications on anxiety before anesthesia induction in preschool children aged 2−6 years scheduled for elective surgery. The secondary outcomes comprised distress during peripheral catheter (PVC) insertion, compliance at anesthesia induction, and level of sedation.


| INTRODUC TI ON
Many preschool children experience significant anxiety during the preoperative period. This may cause distress to the child during the preoperative period, which, in turn, may have a negative impact on their postoperative recovery and cause long-term impairment in cognition. 1 Preoperative anxiety can increase stress-induced hemodynamic instability and pose as a challenge to smooth anesthesia induction. 2 Thus, an optimal drug for premedication in young children is crucial.
The benzodiazepine midazolam has long been the most commonly used drug for premedication in children. 3 It has been shown to alleviate anxiety and provide adequate sedation with an onset duration of ~40 min. However, midazolam has untoward side effects with increased risk of respiratory depression, amnesia, and paradoxical reactions. 4 Two alpha-2-agonists, clonidine and dexmedetomidine, have emerged as alternatives to midazolam. These drugs have good sedative, analgesic, and anxiolytic properties with minimal respiratory depression and better perioperative hemodynamic stability. [4][5][6] Among these two alpha-2-agonists, clonidine has been used more frequently. 6 One major downside of clonidine is a longer duration of onset, requiring administration 60 min before anesthesia induction. 7 Compared with clonidine, dexmedetomidine has been suggested to have a more favorable safety profile, with anxiolysis and sedation achieved in 40 min following intranasal administration. 8  that the midazolam was superior in reducing anxiety as compared to clonidine 11 and dexmedetomidine. 12 In sum, the literature on reducing preoperative anxiety in preschool children is sparse with diverging results.
In this study, we aimed to compare the effect of three different premedication regimens on preoperative anxiety and sedation in preschool children scheduled for elective ear-, nose-and-throat (ENT) surgery. We hypothesized that midazolam is superior in reducing preoperative anxiety compared to alpha-2-agonists and that the three groups have equal ability to provide sedation in preschool children.

| Participants
The trial was planned to include 150 children and was conducted at Sunderby Hospital, Luleå, Sweden. The inclusion criteria comprised: age, 2−6 years; both sexes; American Society of Anesthesiologists Physical Status (ASA-PS) I−II; weight, ≤30 kg; scheduled for elective ENT surgery; primarily tonsillotomies/tonsillectomies and adenotomies/adenectomies; and written informed consent obtained from their parents. The exclusion criteria comprised: ASA-PS >II; heart, lung, neurologic, or central nervous system disorders; use of psychotropic medication; or history of recent surgery.

What this article adds
In this randomized clinical trial, using validated instruments for measurement of anxiety and sedation in preschool children aged 2−6 years, showed that midazolam was superior in reducing preoperative anxiety, while the modern alpha-2-agonist, clonidine, and dexmedetomidine, provided deeper sedation. oral clonidine (30−40 min vs. 60 min, respectively). 5,8,14 Therefore, clonidine was administered 20 min earlier than midazolam and dexmedetomidine.

| Consent, randomization, and blinding
Signed informed consent was obtained from the parents upon the child's arrival at the day care unit. All children were accompanied by a parent throughout the preoperative period and one parent was allowed to attend during induction of anesthesia.
Randomization was performed by opening a sequentially numbered envelope containing the group assignment. The envelopes were prepared by an independent statistician using the randomization function in Microsoft Excel, with randomization in blocks of 15.
A nurse, independent from the research team and data collection, administered the study drugs. All children first received oral fluid (clonidine or sterile water); 20 min later, intranasal fluid was administered using a mucosal atomization device (MAD Nasal™, Teleflex, USA; dexmedetomidine or 0.9% NaCl), followed by oral fluid (midazolam or sterile water; Figure 1). The interventions were triple-blinded; that is, blinded for the patient, care providers, and researchers.

| Background and baseline
Immediately after inclusion, the patients' background information was collected, and baseline assessments were performed. No questions were addressed directly toward the child. The parents were asked about their child's anesthetic experience, in what way they had prepared their child for anesthesia (preoperative information, dichotomized into "yes" or "no"), and to grade their child's anxiety level at home due to surgery, using a self-devised fourgrade scale (not worried, somewhat worried, worried, and very worried; dichotomized into "not worried" or "worried" ["somewhat" to "very"]).

| Endpoints and timing
The primary endpoint was anxiety at anesthesia preparation, and the secondary endpoints were distress during peripheral venous catheter (PVC) insertion, compliance at induction, and level of sedation. The timing of the measurement is shown in Figure 1.

| Data collection and tools
To minimize interrater variability, one person in the research team (ÅB) collected all data. The researcher was not responsible for administering clinical anesthesia or providing care for the patients. The researcher was trained through reading, learning, and testing the instruments during normal clinical encounters, followed by pilot runs of the complete protocol.

| Anxiety as measured by mYPAS
Anxiety was measured at the time of arrival in the day care unit (baseline) and during anesthesia preparation in the operating theater. The level of anxiety was assessed using the validated Swedish version of the modified Yale Preoperative Anxiety Scale (mYPAS), containing 27 items in five categories (activity, emotional, expressivity, state of arousal, vocalization, and use of parents), representing five domains of anxiety. 17,18 The mYPAS scale ranges from 0 to 100 points, with higher scores signifying higher levels of anxiety. The score is normalized for the different number of steps in the items and also accounts for nonobservable items. The instrument assesses the children's anxiety during the perioperative period, has good-to-excellent observer reliability, with high concurrent and construct validity. 17,18 Before the mYPAS was developed, the State-Trait Anxiety Inventory for Children (STAI-CH) was used as the golden standard. A validation study showed that in children aged 5−12 years, an mYPAS score >30 for anxiety had a positive predictive value of 79%, which corresponded to an STAI-CH score of >37, with only 6% false-positives and 4% false-negatives. 18 F I G U R E 1 Setting and timeline 2.9 | Distress during PVC insertion as measured by BDS Distress during PVC insertion was assessed on a five-point behavioral distress scale (BDS), with zero denoting no response. 19 To minimize the pain associated with PVC insertion, all children received a topical anesthetic (EMLA ® 25 mg/25 mg, Astra Zeneca, Stockholm) that was applied 60 min before needle insertion. 20

| Compliance at induction as measured using the induction compliance checklist (ICC)
The child's compliance to the face mask during preoxygenation before anesthesia induction was measured using the ICC. The ICC observational scale has shown good reliability, containing 10 negative behavioral items, with 0 p representing a perfect induction with no negative behaviors due to fear or distress. 21

| Induction of anesthesia
Anesthesia was induced using a standardized protocol. First, a bolus of atropine (0.01 mg/kg) was administered, continued by a 1-min infusion of remifentanil (2-3 µg/kg), and a 2-min infusion of propofol 3-5 mg/kg infused for 2 min until the child was asleep. If an intravenous line could not be established, either for technical reasons, or due to child discomfort (BDS >3), an alternative rescue anesthesia induction was initiated with inhalation of oxygen/nitrous oxide (1:1) mixed with 8% sevoflurane using a face mask.

| Statistical analysis
A sample size of minimum 102 participants was calculated using the mean (27.30) and standard deviation (6.24) from the method reported by Wright et al., 24 and was modeled to detect a mean difference of ≥4 in the mYPAS among the three groups (F-test, omnibus, one-way, alpha = 0.05%; power =80%) using the G*Power software. 25 To adjust and compensate for uncertainties in the power calculation and to account for drop-outs, the study was designed to include 150 participants. As the initial power calculation was based upon uncertain data, the protocol included an interim analysis to analyze the variance in the primary outcome variable after more than half of the patients were included. During the study period, the procedure was relocated to another hospital, which severely impeded recruitment for the study. After the inclusion of 90 patients (six randomization blocks), an interim analysis was performed in collaboration with an independent statistician and the three study groups were compared. The variance in the primary endpoint variable was <70% of the variance on which the power analysis was based upon. The analysis concluded that the primary study objectives could be answered, and after a discussion with the monitoring authority, a decision was taken to conclude the study.
After all study data were entered in the database and had been validated according to the GCP guidelines, the study database was locked, and the complete randomization code was broken.
Data are reported as numbers, means ± standard deviation,

| Primary outcome: anxiety
There were no differences in the mYPAS score at baseline between the groups (Figure 3, Table 2). During anesthesia preparation, the mYPAS score increased compared to the baseline in both the CLO (p = 0.016) and DEX (p = .007) groups, while no change was observed in the MID group. None of the five categories/domains of anxiety in the mYPAS showed any significant differences when comparing the groups at baseline or at anesthesia preparation (Table S1).
In the entire study population (n = 84), a negative correlation was observed between age and mYPAS during anesthesia preparation (r = −.26, p = .019), indicating that the high mYPAS scores were predominately among the younger children.

| Secondary outcome: sedation
The baseline assessment of the RSS showed no significant differences between the groups. Sixty minutes after the first study intervention (clonidine or placebo), with all participants given an active premedication, the CLO and DEX groups had a higher RSS score compared to the MID group (MID, 2.26 ± 0.45; CLO, 3.56 ± 1.12; DEX, 4.03 ± 0.72; p < .001; Figure 4). Compared to the observations at 60 min, seven (26%) children in the CLO group and four (13%) in the DEX group had a decreased RSS score from 4 to 2 during anesthesia preparation in the operating theater. These children had a higher median mYPAS score compared with those who did not have a decrease in the RSS score however, no correlation was observed at 50 min, 60 min, or at anesthesia preparation.

| Rescue induction of anesthesia
In 75 children, anesthesia was induced intravenously according to the study protocol, while in nine children, the rescue inhalational method was required (MID, n = 2; CLO, n = 3; and DEX, n = 4; not significant). As noted by Doughty in 1959, a smooth anesthesia induction depends on the likelihood of the establishment of a rapport between the anesthesiologist and patient. 26

| Strength, limitations, and future research
To increase the generalizability of our results, we included children from among the normal patient influx at a county hospital.
Furthermore, we used block randomization; therefore, the study groups had similar characteristics ( Table 1). The strengths of our study also include the narrow age spectrum of population and the use of validated instruments, adapted and developed for use in children, to validate the interventions in the perioperative period.
Further, we used the mYPAS as it is an observation instrument developed specifically to enable measurement of preoperative anxiety in children. 18 Kain et al. compared the mYPAS against the STAI-CH, which is the golden standard, in children aged 5−12 years and showed an acceptable concurrent validity between the two (p = .01; r = .79) and good inter-and intra-observer reliability. In addition to being reliable and validated, the mYPAS has several other important features. First, it can be applied to all children aged >2 years.
Second, because it is a structured instrument that comprises five domains of anxiety, it is much more sensitive to changes in anxiety levels than global instruments.
The limitations of this study include its single-center study design, early termination, and a sample size of only 90 children.
However, there were uncertainties regarding the calculated sample size prior to the start of the study, and an interim analysis was incorporated in the study protocol. We observed significant differences between the groups in this analysis and combined with limited resources, we decided to terminate the study early. It would have been appropriate to complete the study; however, we were unable to do so. Nonetheless, as we were able to draw conclusions from the patients included, we believe that our findings should be reported.
Furthermore, the parents were not separated from the child until the child was anesthetized. The transfer from the daycare unit to the operating room, and being moved to the operation table aroused some of the children, as noted with the increase in the mYPAS scores and decrease in the RSS scores. An optimal premedication would help overcome these challenges.
As the perioperative settings differ between hospitals and countries, a more optimal premedication routine can be established in the specific context through an increased knowledge of the drugs used.
The results from our study contribute to this knowledge by adding information from a Swedish perspective. Further, international multi-center studies on the optimal timing and doses for sedation and anxiety relief in children are warranted.

| CON CLUS IONS
In preschool children aged 2−6 years, premedication with midazolam resulted in a more effective anxiolysis, thereby requiring less sedation compared with premedication with clonidine and dexmedetomidine. If deeper sedation is required, dexmedetomidine, due to its easy administration and faster duration of onset, might be preferable to clonidine in clinical use.

ACK N OWLED G EM ENTS
We would like to thank Robert Lundqvist for statistical support,

CO N FLI C T O F I NTE R E S T
The study was conducted with public funding from Region Norrbotten and Region Västerbotten. JW has received lecture fees from AbbVie Sweden AB. All authors declare no conflicts of interest.

E TH I C A L A PPROVA L
The study was approved by the Regional Ethic Review Board in

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author, MH, upon reasonable request.