A randomised comparison of two intranasal dexmedetomidine doses for premedication in children


V. M. Yuen
Email: vtang131@hku.hk


We compared sedation levels in children following administration of intranasal dexmedetomidine. One hundred and sixteen children aged between 1 and 8 years were enrolled in this prospective, randomised trial. Children were assigned to receive either intranasal dexmedetomidine 1 μg.kg−1 (Group 1) or 2 μg.kg−1 (Group 2). Thirty-one (53%) patients from Group 1 and 38 (66%) patients from Group 2 were satisfactorily sedated at the time of anaesthetic induction. Logistic regression showed a significant interaction effect (p = 0.049), with the odds ratio between Group 2 over Group 1 estimated as 1.1 (95% CI 0.5–2.7) for the 1–4 year age group, and 10.5 (95% CI 1.4–80.2) for the 5–8 year age group. Both doses produced a similar level of satisfactory sedation in children aged 1–4 years, whereas 2 μg.kg−1 resulted in a higher proportion of satisfactory sedation in children aged 5–8 years. There were no adverse haemodynamic effects. We conclude that intranasal dexmedetomidine in a premedication dose of 2 μg.kg−1 resulted in excellent sedation in children.

Dexmedetomidine is an α2-adrenergic receptor agonist that provides sedation without respiratory depression. Clinical trials have demonstrated that intranasal dexmedetomidine in a dose of 1 μg.kg−1 produces satisfactory sedation in between 53% and 57% of children at anaesthetic induction [1, 2]. There are also reports using higher doses of intranasal dexmedetomidine [3]. We compared the sedative effect of 1 μg.kg−1 intranasal dexmedetomidine with those of 2 μg.kg−1 for premedication in children aged 1–8 years and hypothesised that the higher dose would produce satisfactory sedation in more children at the time of anaesthetic induction.


This study was approved by the local research ethics committee and written informed consent was obtained from the parents or legal guardians of all the patients. Children weighing between 7 and 17 kg, of ASA physical status 1 or 2, aged 1–8 years, who were scheduled for elective surgery at either Queen Mary Hospital, Hong Kong (QMH) or the University of Malaya, Kuala Lumpur Malaysia (UM) were enrolled in the study. If the child understood and discussed the need for premedication, then his/her assent was also obtained. Exclusion criteria included known allergy or hypersensitivity to dexmedetomidine or glucose-6-phosphate dehydrogenase deficiency.

All patients were taken to a holding area approximately 30 min before surgery and baseline measurements of non-invasive blood pressure, oxygen saturation and heart rate were recorded.

Children were randomly allocated to one of two groups; intranasal dexmedetomidine 1 μg.kg−1 (Group 1) or intranasal dexmedetomidine 2 μg.kg−1 (Group 2). Two randomisation lists were produced for the two centres, stratifying patients according to age: preschool (1–4 years) and school (5–8 years), to help ensure similar age distribution of children within Group 1 and Group 2. The randomisation list for each age group was generated by an independent statistician who was not involved in data collection or analysis. Block randomisation was used to ensure a balanced allocation between treatment groups.

Preservative-free dexmedetomidine (PrecedexTM; Hospira Inc., Lake Forest, IL, USA) in a concentration of 100 μg.ml−1 was used. The study drug was prepared in a 1-ml syringe by an independent anaesthetist. If recruited children were assigned to Group 1, dexmedetomidine was diluted with an equal volume of 0.9% saline; if assigned to Group 2 then dexmedetomidine was not diluted. The final volume of administered solution was 0.02 ml.kg−1. An equal volume of drug was dropped into each nostril by a blinded research assistant.

Vital signs were measured every 15 min after intranasal drug administration until transfer to the operating theatre. Following intranasal drug administration, sedation and behaviour scores were assessed every 5 min by a blinded observer using a four-point scale (Table 1). When the patient was transferred to the operating theatre, the study was concluded.

Table 1.   Sedation and behaviour scores.
Sedation score
 1Alert, awake
 2Drowsy, sleepy, lethargic
 3Asleep but responds to mild prodding or shaking
 4Asleep and does not respond to mild prodding or shaking
Behaviour score
 1Crying or resisting
 2Anxious and not reassurable
 3Anxious but reassurable
 4Calm and cooperative

For statistical analysis, sedation and behaviour status were categorised as being satisfactory when the sedation or behaviour scores were 3 or 4 and unsatisfactory when they were 1 or 2. The proportion of patients with satisfactory sedation was classified by dose, age and hospital. From the results of a previous study, using 1 μg.kg−1 intranasal dexmedetomidine, satisfactory sedation was obtained in 57% of subjects. If this proportion was to increase by 25% after premedication with 2 μg.kg−1, then 52 children per group were required to achieve 80% power of detecting a statistical difference with a false positive rate of less than 5%. The difference between the two doses was analysed by logistic regression adjusting for age and hospital. The relationship between age and dose effect was also analysed by logistic regression. Sedation onset was defined as the time from drug administration until the onset of satisfactory sedation. The duration of sedation was defined as the onset of satisfactory sedation until the time when the patient either woke up naturally or following stimulation, such as transfer to the operating theatre, the moment when intravenous cannulation was performed or the time when a facemask was applied.

Sedation onset time and duration of sedation were compared between groups using a log-rank test. Comparisons between hospitals were with Fisher's exact test. The statistical software used was SPSS for Windows version 18.0 (SPSS Inc., Chicago, IL, USA). Logistic analysis and Kaplan–Meier analysis were performed using SAS System for Windows Release 9.2 (SAS Institute Inc., Cary, NC, USA). A value of p < 0.05 was taken as denoting statistical significance.


One hundred and twenty-three patients were initially recruited to the study (Fig. 1). Three children had intercurrent upper respiratory tract infection, three arrived too late for premedication and one parent refused to enter their child into the study on the day of surgery. The results for 116 children were analysed. Patients’ characteristics and surgical data are shown in Table 2. There was a significantly higher proportion of older children from UM (37%) compared with QMH (13%), p = 0.006. Sedation characteristics for the children are shown in Table 3. The median (IQR [range]) time from administration of premedication to anaesthetic induction was 56 (40–65 [20–165]) min. There were no differences in sedation onset time or duration of sedation between the different age groups or the two drug groups.

Figure 1.

 CONSORT flow diagram. URTI, upper respiratory tract infection.

Table 2.   Baseline characteristics of patients receiving intranasal dexmedetomidine 1 μg.kg−1 or 2 μg.kg−1. Values are median (range [IQR]), number or number (proportion).
 1 μg.kg−12 μg.kg−1
(n = 58)(n = 58)
  1. QMH, Queen Mary Hospital, Hong Kong; UM, University of Malaya, Kuala Lumpur; EUA, examination under anaesthesia.

Age; years2 (1–7 [1–4])2 (1–8 [1–4])
Weight; kg13 (7–43 [11–16])13 (7–38 [11–17])
Age distribution
 QMH (n = 78)
    1–4 years3533
    5–8 years46
 UM (n = 38)
    1–4 years1212
    5–8 years77
Type of surgery
 Hydrocele/testis/hernia20 (35%)19 (33%)
 Excision lymph nodes or lumps7 (12%)13 (22%)
 Circumcision/other penile surgery11 (19%)12 (21%)
 Cystoscopy/colonoscopy/EUA3 (5%)1 (2%)
 Laparoscopic surgery12 (21%)10 (17%)
 Ophthalmic surgery3 (5%)0
 Orthopaedic surgery2 (4%)2 (4%)
 Tonsillectomy01 (2%)
Table 3.   Sedation characteristics of children following intranasal dexmedetomidine 1 μg.kg−1 or 2 μg.kg−1. Values are median (95% CI) or number (proportion).
 1 μg.kg−12 μg.kg−1
  1. QMH, Queen Mary Hospital, Hong Kong; UM, University of Malaya, Kuala Lumpur; EUA, examination under anaesthesia.

Sedation onset time; min
 Overall30 (25–30)
   (n = 58)
25 (20–25)
(n = 58)
 1–4 years30 (25–30)
   (n = 47)
25 (20–30)
(n = 45)
 5–8 years30 (20–45)
   (n = 11)
25 (20–30)
(n = 13)
Duration of sedation; min
 Overall45 (40–70)
   (n = 47)
95 (45–135)
(n = 50)
 1–4 years45 (30–n/a)
   (n = 40)
95 (45–135)
(n = 38)
 5–8 years58 (15–70)
   (n = 7)
n/a (20–n/a)
(n = 12)
Satisfactory sedation
  1–4 years27 (43%)27 (60%)
  5–8 years4 (36%)11 (85%)
  1–4 years24 (69%)22 (67%)
  5–8 years2 (50%)6 (100%)
  1–4 years3 (25%)5 (42%)
  5–8 years2 (29%)5 (71%)

For both the doses and both the age groups, children from QMH had a higher proportion of satisfactory sedation than children from UM; overall 69% of children from QMH and 40% of children from UM were satisfactorily sedated at anaesthetic induction. Logistic regression revealed a significant interaction effect (p = 0.049), with the odds ratio (95% CI) between Group 2 and Group 1 estimated to be 1.1 (0.5–2.7) for the 1–4 year age group, and 10.5 (1.4–80.2) for the 5–8 year age group. The satisfactory sedation odds ratio (95% CI) for QMH over UM was 4.3 (1.734–10.6), p = 0.002. Figure 2 shows the proportion of patients attaining satisfactory sedation with time after administration of the test drug and Fig. 3 shows the proportion of patients who remained sedated over time once satisfactory sedation had be recorded. The median (95% CI) duration of sedation was 45 (40–70) min and 95 (45–135) min for Groups 1 and 2, respectively, but this did not reach statistical significance.

Figure 2.

 Time after intranasal dexmedetomidine administration and the proportion of children attaining satisfactory sedation after dexmedetomidine 1 μg.kg−1 (inline image) or 2 μg.kg−1 (inline image). Each ‘+’ indicates a child was not yet sedated at the observation point.

Figure 3.

 The proportion of children who remained sedated once onset of sedation had been recorded after intranasal dexmedetomidine 1 μg.kg−1 (inline image) or 2 μg.kg−1 (inline image). Each ‘+’ indicates a child was still sedated at the time of anaesthetic induction.

Dexmedetomidine had no effect on oxygen saturation levels, the lowest recording being 95% during the observation period. Changes in systolic blood pressure and heart rate from baseline (before administration of premedication) are shown in Fig. 4. The maximum mean reductions in systolic blood pressure were 15.7% and 11.9% for Groups 1 and 2, respectively. The maximum mean reductions in heart rates were 19.1% and 14.1% for Groups 1 and 2, respectively.

Figure 4.

 Mean percentage change in systolic blood pressure (SBP) and heart rate (HR) from baseline in children following premedication with intranasal dexmedetomidine 1 μg.kg−1(inline image) or 2 μg.kg−1 (□). Error bars indicate SEM.


Satisfactory pre-operative sedation in this study was defined as a child that was asleep with a sedation score of 3 or 4 when transferred from the holding area to the operating theatre in preparation for induction of anaesthesia. Whilst this degree of sedation is necessary for young children, it may not be so for older children who usually remain calm and cooperative to induction of anaesthesia when their sedation level is lighter, probably because they have a better understanding of the anaesthetic process. A relatively deep level of sedation was chosen as the endpoint in this study because we were investigating the effects of different intranasal doses of dexmedetomidine based on the results of previous studies [1, 2].

Previous studies have shown a dose-dependent increase in sedation levels in adults when dexmedetomidine is given intravenously [4–6], therefore we expected an increase in the proportion of patients with satisfactory sedation when we doubled the dose of intranasal dexmedetomidine. This was true in older children aged between 5 and 8 years, but not so in those aged 4 years or less. Petroz et al. [7] found that when intravenous dexmedetomidine was administered to children in doses of 0.33, 0.66 and 1 μg.kg−1 over 10 min, there were no differences in sedation levels. Although the authors stated that the number of patients were too few to detect significant differences, the true explanation may be that inadequate plasma concentrations were attained at the doses administered. The absence of a dose response in children less than 4 years of age in the present study may be due to inadequate dosing. The bioavailability of intranasal dexmedetomidine has been estimated at 65% in adults when administered as a concentrated veterinary formulation (84 μg in 0.2 ml) [8]. The bioavailability of intranasal dexmedetomidine was probably less in our group because we administered a more dilute intravenous formulation (100 μg in 1 ml). In younger children the intranasal surface area is less than that of older children, which may result in less systemic drug absorption.

There is evidence that younger children require larger bolus doses of dexmedetomidine to achieve satisfactory sedation due to pharmacokinetic factors. In a study performed on infants and children aged between 1 month and 11 years undergoing either bronchoscopy or magnetic resonance imaging [9]. All patients were given an intravenous loading dose of dexmedetomidine 1 μg.kg−1. Children younger than 2 years of age had a significantly larger volume of distribution at steady state, hence a larger initial dose was required than in older children to achieve similar plasma concentrations. In two reports where an intravenous dexmedetomidine infusion was the primary sedative following cardiothoracic surgery, neonates and infants required higher drug infusion rates to maintain satisfactory sedation levels compared with older children [10, 11].

Pharmacodynamic differences between younger and older children are also possible; in one study of dexmedetomidine, a variety of simulated infusion regimens demonstrated that children aged less than 1 year required higher plasma concentrations of dexmedetomidine than children aged older than 1 year to achieve similar sedation levels [12]. There are no pharmacokinetic or pharmacodynamic data available following administration of intranasal dexmedetomidine in children and we can only speculate on the reason for a reduced response in younger patients in the present study.

We administered intranasal dexmedetomidine by simply dripping the drug into both nostrils from a 1-ml tuberculin syringe. A Mucosal Atomization Device (MAD NasalTM; Wolfe Tory Medical Inc., UT, USA) has been used to administer intranasal drugs in children [13]. In an animal model, this has not been shown to produce higher plasma levels of midazolam when compared with simple drops [14] and was not associated with improved sedation efficacy in children aged 2 and 3 years who received intranasal midazolam for dental procedures [15]. However, the atomiser has been shown to deliver a consistent volume of drug when compared with nasal drops [15]. Intranasal dexmedetomidine administered via an atomiser in a relatively high dose of 3–4 μg.kg−1 is routine for computerised tomography (CT) scanning in one centre [16]. The author reports a high success rate with minimal haemodynamic changes. Intravenous dexmedetomidine is used routinely as the sole sedative agent for CT scanning in Boston Children’s Hospital with a high success rate. Mason et al. [17] found that, when used as the sole intravenous agent, the mean loading dose of dexmedetomidine required to achieve a minimum Ramsay Sedation Score of 4 was 2.2 μg.kg−1 with a subsequent infusion rate of 1 μg.kg−1.h−1. The success rate with this sedation regimen was 98.9% [18]. The lack of a dose-dependent effect on sedation in our study is unlikely to be due to a ceiling effect and it is possible that if the dose is increased, so will the rate of successful sedation. It has been postulated that intranasal dexmedetomidine administered via an atomiser may produce sedation via a direct effect within the central nervous system. An animal study demonstrated that atomised intranasal midazolam produced higher drug levels within the cerebrospinal fluid than when administered as simple drops [14] and this may also apply to dexmedetomidine.

We found a significant difference in the successful sedation rate between the two centres. Although both are tertiary referral teaching hospitals, the environment and culture is different and the study was not designed to compare the two centres, so we cannot explain the differences.

There are several limitations to this study. First, children were arbitrarily stratified into two age groups during the randomisation process to ensure even distribution of children by age in each group. During the analysis we discovered significant differences in the rate of satisfactory sedation between children in the two age groups. However, since this study was not designed to investigate the effect of intranasal dexmedetomidine in children of different ages, we cannot explain this difference. Second, only those children who were judged to require premedication were recruited and a control group was not included in the study. Third, a relatively subjective sedation and behaviour scale was used because it was simple to use, although these scales are not validated and future studies, perhaps using the University of Michigan Sedation Score [19], could be undertaken. Fourth, a relatively deep sedation level was chosen as ‘satisfactory sedation’ in this investigation. Although such levels are often required in younger children to facilitate smooth anaesthetic induction, they may not be necessary in older children. Finally, these results should not be extrapolated to older children or adults.

In conclusion, dexmedetomidine in doses of 1 and 2 μg.kg−1 administered as intranasal drops produced a similarly high rate of satisfactory sedation in children aged 1–4 years. In children aged 5–8 years, 2 μg.kg−1 was associated with a higher proportion of satisfactory sedation than 1 μg.kg−1 without causing adverse haemodynamic effects, suggesting that the higher dose is more appropriate in this age group.

Acknowledgements and competing interests

We gratefully acknowledge assistance from Ms Jeff SF Man for statistical analysis and Ms Yvonne Lee, research assistant, University of Hong Kong, for data collection. This study was funded by the Department of Anaesthesiology, University of Hong Kong, Hong Kong. No other external funding or competing interests declared.