Dexmedetomidine, a selective α2 agonist, has a sedative action via the locus coeruleus with EEG activity similar to natural sleep. It produces analgesia as well as sedation and has minimal or no effect on respiratory rate or tidal volume [1, 2]. There is increasing evidence that dexmedetomidine is an effective and safe sedative in children in the critical care setting [3–6], for various diagnostic procedures [7–13] and pre-operatively [14–16]. Previous investigations have shown that 1 μg.kg−1 intranasal dexmedetomidine produces significant sedation in healthy adults  and in children aged between 2 and 12 years . Intranasal application is a relatively non-invasive and easy route of administration which has the additional benefit of not requiring patient cooperation as would be the case for swallowing the medication or retaining it sublingually. It is well tolerated and does not have an unpleasant taste or pungency. Although it is established that intranasal dexmedetomidine is an effective sedative for premedication in children [14–16], no data have been published on its onset time or duration of action in the paediatric population. In healthy adult volunteers the onset time was about 45 min . The primary aim of this investigation was to evaluate the optimal time to perform intravenous cannulation in children after 1 μg.kg−1 intranasal dexmedetomidine. Intravenous cannulation was chosen as an endpoint because intravenous induction is commonly employed in children  and because the timing could be controlled, whereas the time of anaesthetic induction, which is a more commonly used endpoint in paediatric premedication studies, is too dependent on the operating theatre schedule. Secondary outcomes of this investigation included onset time and duration of sedation.
Previous studies have shown that 1 μg.kg−1 intranasal dexmedetomidine produces significant sedation in children aged between 2 and 12 years. This investigation was designed to evaluate the onset time. One hundred children aged 1–12 years of ASA physical status 1–2 undergoing elective surgery were randomly allocated to five groups. Patients in groups A to D received intranasal dexmedetomidine 1 μg.kg−1. Patients in Group E received intranasal placebo (0.9% saline). Children from groups A, B, C, D and E had intravenous cannulation attempted at 30, 45, 60, 75 and 45 min respectively after intranasal drug or placebo administration. Vital signs, behaviour and sedation status of the children were assessed regularly until induction of anaesthesia. More children from groups A to D achieved satisfactory sedation at the time of cannulation when compared to group E (p < 0.001). The proportion of children who achieved satisfactory sedation was not significantly different among groups A to D. Overall 62% of the children who received intranasal dexmedetomidine had satisfactory sedation at the time of cannulation. The median (95% CI) time for onset of sedation was 25 (25–30) min. The median (95% CI) duration of sedation was 85 (55–100) min.
After approval from the our local Institutional Review Board and written informed consent from the parents or legal guardian, 100 children of ASA physical status 1 or 2, aged between 1 and 12 years, scheduled to undergo elective surgery at Queen Mary Hospital (Hong Kong) were enrolled in this prospective, randomised, double-blind controlled trial. When the child was mature enough to understand and discuss the need for premedication, patient assent was also obtained. Exclusion criteria included allergy or hypersensitivity to dexmedetomidine, G6PD deficiency, organ dysfunction, cardiac arrhythmia, congenital heart disease or mental retardation.
The primary objective of this study was to identify the optimal time for intravenous cannulation after 1 μg.kg−1 intranasal dexmedetomidine. Patients were randomly assigned into groups each of which had a different time for the cannulation. The relationship between time and the proportion of children with satisfactory sedation and behaviour at the time of cannulation was then studied. Twenty patients were randomly allocated to each of the treatment groups A, B, C, D and E. Children from groups A, B, C and D had an attempt at cannulation by a paediatric anaesthetist at 30, 45, 60 and 75 min, respectively, after 1 μg.kg−1 intranasal dexmedetomidine. From our previous experience of 32 patients aged between 2 and 12 years , the median onset time for satisfactory sedation was about 45 min. Therefore, 20 additional patients were enrolled in a placebo group, group E. Children in this group had an attempt at cannulation 45 min after an equivalent volume of intranasal placebo (0.9% saline). Although not the primary objective, with 20 patients per group we calculated that we could detect a difference in the proportion of children with satisfactory sedation between group B (treatment) and group E (placebo) of 50% or more with 80% power and a 5% false positive rate.
A commercial eutectic mixture of lignocaine and prilocaine cream (EMLA®; AstraZeneca, Hong Kong, China) was applied to the dorsum of both hands at least 30 min before administration of intranasal medication, so that when the intravenous cannula was inserted a further 30 min after intranasal solution (intranasal dexmedetomidine or placebo), the EMLA would be effective.
All recruited patients were sent to the pre-operative holding area approximately 1 h before surgery. After the first recording of blood pressure, oxygen saturation and heart rate, the intranasal solution was administered by a trained research assistant.
The research assistant was blinded to the study drug administered. Intranasal drug was dripped into both nostrils using a 1-ml syringe with the child in the recumbent position. Blood pressure, oxygen saturation and heart rate were measured before and every 15 min after intranasal drug administration until transfer to the operating theatre. Sedation status was assessed every 5 min using a four-point sedation scale (Table 1) and behaviour was evaluated every 5 min with a four-point behaviour scale by the research assistant until transfer to the operating theatre (Table 1). The research assistant was also responsible for calling the anaesthetist to perform intravenous cannulation and data collection. Three paediatric anaesthetists were responsible for the cannulations, which were performed in the pre-operative holding area. They were blinded both to the study drug administered and the timing of cannulation. The sedation and behaviour status were observed and recorded by the research assistant after needle puncture, whether the subsequent cannula insertion was successful or not. If the child refused or was already in distress before any attempt at cannulation, the sedation and behaviour status were recorded and the anaesthetists did not attempt cannulation. The mode of induction of anaesthesia was determined by the attending anaesthetist. A parent was allowed to accompany the child at induction if the child refused to be separated from his/her parent.
|2||Drowsy, sleepy, lethargic|
|3||Asleep but responds only to mild prodding or shaking|
|4||Asleep and does not respond to mild prodding or shaking|
|1||Crying or resisting|
|2||Anxious and not reassurable|
|3||Anxious but reassurable|
|4||Calm and cooperative|
Randomisation was stratified by age: preschool (1–4 years); lower (5–8 years); and upper elementary school (9–12 years). The randomisation list for each age group was generated by a statistician (TJY) before commencement of the study. Block randomisation with blocks of 10 was used to ensure a balanced allocation among treatment groups.
The study drug was prepared in a 1 ml syringe by an anaesthetist not involved in administering anaesthesia to the children. For children who weighed less than 40 kg, the final volume of the study drug was 0.4 ml. Dexmedetomidine was diluted to 0.4 ml with 0.9% saline. For those weighing more than 40 kg, the final volume was equivalent to 0.01 ml.kg−1. An equal volume was dripped into each nostril.
Sedation scores and behaviour status were categorised as being satisfactory when rated between 3 and 4 and unsatisfactory when 1 or 2. Usually, when a child had a sedation score between 3 and 4, they had minimal or no reaction to external stimulation.
The demographic data were analysed by one-way ANOVA and the chi-squared test. The proportion of children with satisfactory sedation and behaviour at each time point was estimated using descriptive statistics. The chi-squared test was used to compare the proportion of children attaining satisfactory sedation or behaviour among all five groups at the time of cannulation, and between the collective treatment and placebo groups at different times. For children in the treatment groups, the relationship between time and the proportion of children attaining satisfactory sedation or behaviour at the time of cannulation was analysed by logistic regression. The sedation onset time was defined as the time from drug administration to the observed onset time of satisfactory sedation. The proportion of children attaining satisfactory sedation over time was estimated with the Kaplan-Meier method. If a patient had never been sedated at the last observation time, the time from drug administration to the last observation of that patient was included in the analysis but marked as censored. The duration of sedation was defined as the time from sedation onset to the time the patient woke up naturally or by stimulation (with a sedation score of 2 or below), e.g. from being transferred to the operating theatre, cannulation or application of the facemask. The proportion of patients remaining sedated over time was also estimated with the Kaplan-Meier method using the patients who reached satisfactory sedation. If a patient remained sedated when they were transferred to the operating theatre, the time from onset of sedation to the time for induction of anaesthesia for that patient was included in the analysis but marked as censored. The mean percentage changes of the vital signs from baseline to different time points were estimated and plotted with standard errors. Due to the exploratory nature of this study, p values were not adjusted for multiple testing. The statistical software used was spss for Windows version 16.0 (SPSS Inc., Chicago, IL, USA).
Baseline characteristics were similar for all patients between groups and are summarised in Table 2.
|Group A (n = 19)||Group B (n = 20)||Group C (n = 21)||Group D (n = 19)||Group E (n = 21)||Overall|
|Age; years||3.7 (3.3 [1–11])||3.9 (3.4 [1–12])||4.2 (2.9 [1–11])||4.0 (2.7 [1–11])||4.0 (2.8 [1–10])||4.0 (3.0 [1–12])|
|Body weight; kg||18.3 (8.3 [10.6–41.2])||17.1 (9.5 [8.6–45.2])||17.6 (7.7 [10.7–40.2])||18.6 (7.4 [10.6–39.9])||19.1 (10.6 [9.1–51.6])||18.1 (8.7 [8.6–51.6])|
|Age distribution; years|
More children from groups A to D achieved satisfactory sedation at the time of intravenous cannulation when compared to group E (p < 0.001, Fig. 1). The proportion of children who achieved satisfactory sedation was not significantly different among groups A to D. Although there appears to be an increasing trend over time in the proportion of patients with satisfactory sedation (univariate logistic regression, p = 0.2), the study may have been underpowered to detect this.
Seventy-four out of the 79 children (93.7%) who received intranasal dexmedetomidine had satisfactory sedation at some time during the premedication period. The median onset time of satisfactory sedation was 25 min (95% CI 25–30 min) (Fig. 2); 91% (95% CI 85–98%) of the subjects had achieved satisfactory sedation 45 min after intranasal dexmedetomidine (Fig. 2).
For those children who had attained satisfactory sedation, the duration of sedation was also analysed. Fouty-five out of the 74 sedated patients were censored in this analysis as they remained sedated when they were transferred to the operating theatre for induction of anaesthesia. The proportion of patients remaining sedated over time is shown in Fig. 3. The median duration of sedation was 85 min (95% CI 55–100 min). By 30 min, about 12% (95% CI 5–20%) of children had woken up from satisfactory sedation.
The proportion of children who had satisfactory behaviour at the time of intravenous cannulation among groups A to E was not statistically different (Fig. 1). However, when the proportion of children who had satisfactory behaviour was compared between the treatment groups collectively (i.e. groups A to D, n = 79) to that of the placebo group (group E, n = 21), significantly more children in the combined treatment group had satisfactory behaviour at the time of cannulation when compared with the placebo group (82.3% vs 57.1% respectively, p = 0.02) (Table 3).
|Dexmedetomidine (n = 79)||Placebo (n = 21)||p value|
|Satisfactory sedation at cannulation||49 (62.0%)||2 (9.5%)||< 0.001*|
|Satisfactory sedation at parental separation||49 (62.0%)||3 (14.3%)||< 0.001*|
|Satisfactory sedation at induction||45 (57.0%)||2 (9.5%)||< 0.001*|
|Satisfactory behaviour at cannulation||65 (82.3%)||12 (57.1%)||0.021*|
|Satisfactory behaviour at parental separation||65 (82.3%)||12 (57.1%)||0.021*|
|Satisfactory behaviour at induction||56 (70.9%)||11 (52.4%)||0.109|
All the recruited children arrived in the waiting area of the operating theatre approximately 60–75 min before the planned surgery. Once intravenous cannulation had been attempted, surgery could commence when the operating theatre was ready. Hence, the time of parental separation and anaesthetic induction from the time of study drug administration mainly depended on the operating theatre schedule. Consequently, analysis of sedation and behaviour at parental separation and anaesthetic induction were based on treatment (groups A-D) and placebo group (group E).
The median (range) time to parental separation from the administration of the study drug was 75 (30–135) min and 70 (45–145) min for groups A-D and group E respectively. Anaesthetic induction usually took place 5–10 min after parental separation. The median (range) time to induction was 80 (35–150) min and 80 (45–155) min for groups A-D and group E, respectively. At the time of parental separation, significantly more children from groups A–D achieved satisfactory sedation compared to those receiving placebo (62% vs 14.3% respectively, p < 0.001) (Table 3). Similarly, significantly more children from groups A-D achieved satisfactory sedation at the time of anaesthetic induction when compared to placebo (57% vs 9.5% respectively, p < 0.001) (Table 3). At the time of parental separation, significantly more children from groups A-D had satisfactory behaviour compared to children from the placebo group (82.3% vs 57.1%, p = 0.021) (Table 3). Although there was a higher percentage of children from groups A-D than those from group E with satisfactory behaviour at the time of anaesthetic induction (70.9% vs 52.4%), this difference was not statistically significant (Table 3). For the children treated with intranasal dexmedetomidine, no significant trend over time was indicated in the proportion of satisfactory sedation or behaviour at parental separation and anaesthetic induction.
Dexmedetomidine had no effect on arterial saturation and no child had a reduction of arterial saturation < 95% during the observation period after premedication. The changes in systolic blood pressure and heart rate from baseline (before administration of premedication) are shown in Figs 4 and 5 respectively. The maximum mean reduction in systolic blood pressure of children from groups A-D was 13.2% and this occurred at 60 min. The maximum reduction in heart rate in this group was 14.9% at 75 min.
The main objective of this investigation was to identify the optimal time to perform intravenous cannulation in children who had been given intranasal dexmedetomidine. An increasing trend of satisfactory sedation from 30 to 75 min was noted. As EMLA had been applied for what is generally considered the optimal length of time in this patient population , it is unlikely that the increasing trend of the proportion of children with satisfactory sedation at the time of intravenous cannulation was related to any further increased length of application. We have no information on the time when the maximum number of children will be sedated at the time of intravenous cannulation, as we did not investigate the rate of satisfactory sedation if this procedure was performed at a time later than 75 min after intranasal dexmedetomidine. In clinical practice it may not be feasible to wait for longer than 75 min to perform intravenous cannulation after premedication.
Since it is unethical to insert intravenous cannulae repeatedly in each patient, it is impossible to design a study that evaluates the sedation onset time in each patient with this methodology. Despite the limitation of the design of this investigation, we still have approximation of time to sedation of most patients and the Kaplan-Meier analysis revealed quite a precise estimation. The median onset time of sedation was 25 min with a 95% CI of 25–30 min, and this probably implies a sufficient sample size to evaluate the onset of sedative effect of this dose of intranasal dexmedetomidine in children. The duration of sedation was also estimated by the Kaplan-Meier method and it was revealed that the median duration of sedation was 85 min (95% CI 55–100 min). As it can be difficult accurately to coordinate premedication with the time of surgery, our results suggest that this route of dexmedetomidine administration provides some flexibility as long as it is given at least 30, and preferably 45 min, in advance.
In this investigation we have shown that intranasal dexmedetomidine 1 μg.kg−1 produced satisfactory sedation in 62% of children at the time of intravenous cannulation following application of topical local anaesthetic cream (EMLA®). Fifty-seven percent (95% CI 46–67%) of the children remained sedated when they were transferred from the pre-operative holding area to the operating theatre. A sedation rate of 62% at the time of intravenous cannulation and 57% at the time of anaesthetic induction may appear suboptimal. However, in this investigation we aimed to look for the proportion of children who had attained a deep level of sedation; therefore ‘satisfactory sedation’ was defined as a child who remained asleep even with external stimulation (sedation score of 3 or 4). Therefore, if a child was drowsy and appeared sleepy, the sedation status was still classified as ‘unsatisfactory’. Nonetheless, the depth of sedation required for children of different ages would be different. For some children a sedation score of 2 may be adequate for pre-operative preparation.
Dexmedetomidine has unique sedative properties as patients given this drug are clinically sedated, but they can be roused without agitation or anxiety when stimulated . Older children, who have more understanding of anaesthesia and surgery, may remain calm and drowsy when they are roused from stimulation. This, however, may be a disadvantage in young children who could be frightened or excessively anxious when roused and awoken in a strange environment, especially in preschool-age children in whom communication is difficult. Hence, the optimum clinical endpoint of premedication in younger and older children may be different. In younger children deeper sedation may be required to facilitate smooth induction of anaesthesia. Therefore, a relatively deeper level of sedation was chosen as a ‘satisfactory’ sedation endpoint but this may not apply to older age groups.
Since this was an exploratory study to gain further understanding of intranasal dexmedetomidine in children in general, subgroup analysis was not performed as the sample sizes of the age subgroups were small and this study was not powered to compare the difference in sedation and behaviour effects between different age groups. It is possible that the pharmacodynamic and pharmacokinetic response of dexmedetomidine varies with age. Overall, intravenous administration produces consistent and predictable blood concentrations in children. Although there are few reports on the pharmacokinetics of intravenous dexmedetomidine in this group [4, 21, 22], it has been shown that children younger than 2 years of age need a larger initial dose than older children . In a retrospective review of dexmedetomidine infusion in infants and children in intensive care units , it was also noted that there was a trend towards higher dexmedetomidine infusion requirements in patients less than 1 year old compared with older children.
Intranasal dexmedetomidine has been used for premedication in children at higher doses. Talon et al.  compared intranasal dexmedetomidine 2 μg.kg−1 via a Nasal Mucosal Atomizing Device (MAD®; Wolfe Tory Medical Inc., Salt Lake City, UT, USA) with oral midazolam 0.5 mg.kg−1 in children aged between 1 and 18 years. When compared with oral midazolam, 2 μg.kg−1 intranasal dexmedetomidine produced comparable induction conditions, but was more effective for inducing sleep before surgery. No untoward side effects or haemodynamic disturbances were detected with this dose. Mason and colleagues  demonstrated that 3 μg.kg−1 of intravenous dexmedetomidine was required to achieve a high success rate for sedation in children undergoing magnetic resonance imaging. Heart rate and blood pressure stayed within 30% of the patient’s baseline values when this high-dose regimen was used for CT imaging studies . The haemodynamic changes were independent of age, required no pharmacologic intervention and did not result in any adverse events. A mean (SD) dose of 2.1 (0.8) μg.kg−1 intravenous dexmedetomidine was required to achieve adequate sedation in children with pervasive personality disorder undergoing electroencephalography . No significant haemodynamic or respiratory effects were noted. Since the bioavailability of transmucosal dexmedetomidine is about 80% , 1 μg.kg−1 administered intranasally is a relatively low dose. This may explain why the rate of successful sedation in this study was just above 60% and more dose finding studies are warranted in the paediatric population.
Although intranasal dexmedetomidine is a useful sedative in children, more work is required to delineate its use. Firstly, there is no information on the pharmacokinetics of intranasal dexmedetomidine in children. It would also be useful to identify the optimum endpoint of premedication with this drug in different age groups. The optimal dose of intranasal dexmedetomidine as pre-operative sedative is also unknown, although there is evidence that a higher dosing regimen of 2 μg.kg−1 is not associated with any untoward side-effects. In the future a dose comparison study is warranted. In this investigation intranasal dexmedetomidine was given by dripping into the nostrils with a 1-ml syringe. It may be interesting to determine whether there is any pharmacologic advantage of using the more expensive nasal mucosal atomizing device.
In the meantime, one may consider dexmedetomidine a pre-operative sedative when a child refuses oral medication, or has a history of failed sedation with other sedatives. If a dose of 1 μg.kg−1 is chosen, the onset time will range from 25 to 45 min with a median duration of sedative effect of 55–100 min. Although there is no information on the success rate, onset time and duration of sedation with higher doses, there is evidence to suggest that 2 μg.kg−1 intranasal dexmedetomidine is also effective and safe in children.
No external funding and no competing interests declared.