This prospective, equally randomized (1:1), double-blind, parallel group trial was conducted in a large pediatric ED. This study was monitored by an independent regulatory unit, Karolinska Trial Alliance. Ethical approval was obtained from the Regional Ethical Review Board Stockholm. And this study was registered with European Clinical Trial Registry. The trial protocol is presented as Additional file 1.

The CONSORT guidelines [21] were used for reporting our data.

Children 1–3 years old who presented to the ED with a laceration in need of suturing or a burn covering less than 4% of the body surface area and required PSA were eligible for enrolment. Injury was assessed and the need for PSA determined by the ED physician (mainly physicians in training in pediatrics, emergency medicine or general medicine) according to local guidelines. The trial physician was contacted, and the inclusion and exclusion criteria were revised. Patients with American Society of Anesthesiologist physical status classification (ASA) [22] ≥ III, current respiratory tract infection, impaired level of consciousness or any other neurologic symptoms as well as hypersensitivity to the trial drugs were excluded. In addition, children of parents with insufficient understanding of the Swedish language could not be enrolled in this study as written information of the trial was provided in Swedish. Signed informed consent was given by the parents.

This study was conducted in the pediatric ED at Astrid Lindgren Children’s hospital (ALB), Karolinska University Hospital in Stockholm, Sweden. ALB ED has 50 500 annual visits and offers medical care to children and adolescents with all levels of trauma, injuries, and sickness.

The primary outcome was pain, measured as the highest level of pain during the procedure. Procedure was defined for lacerations as suturing of the wound and for burns as wound debridement and dressing, local guidelines for the procedures were followed. Pain was assessed with Face, Legs, Activity, Cry, Consolability scale (FLACC) [23], which has been validated for procedural pain assessment by Nilsson et al. [24]. FLACC scores were classified as following: 0 = no pain, 1–3 = mild discomfort, 4–6 = moderate pain and 7–10 = severe pain [25, 26]. We considered a change of two points on the FLACC as clinically significant.

Secondary outcomes were sedation depth, parents’ satisfaction, and ED physician’s assessment of the feasibility of the procedure. Ramsay sedation scale [27] was used to evaluate sedation depth, as it is one of the most widely used tools for observationally based sedation assessment and it has been used in studies assessing sedation with intranasal dexmedetomidine [12, 14]. Ramsay score 1 = awake, 2 = awake; co-operative, orientated and tranquil, 3 = awake; responds to commands only 4 = asleep; reacts with a brisk response to a light glabellar tap or a loud auditory stimulus, 5 = asleep; reacts with a sluggish response to a light glabellar tap or a loud auditory stimulus, 6 = asleep; does not respond to pain. We considered a change of one point as clinically significant.

Parent/parents who were with the child from administration of the trial drug until recovery, received a questionnaire with four questions: (1) in your opinion; how much pain did your child have during the procedure on a scale of 0–10 (the revised Faces Pain Scale (FPS-R) [28] was shown to the parents), (2) what was your opinion about the sedation and the procedure on a scale of 1–5 (1 = not satisfied, 5 = very satisfied). (3) if your child needs procedural sedation again in the future, would you prefer the same management? (yes/no), (4) if not what would you wish to be different?

The ED physician graded the feasibility of performing the procedure on a scale of 1–5 (1 = very easy, 5 = very difficult).

All patients included in the study received oral paracetamol (40 mg/kg) no later than 1–1.5 h before the procedure.

Patients were enrolled and randomized before the administration of local anesthesia. Buffered lidocaine (10 ml 1% lidocaine + 2 ml NaHCO3 6 M) was used for local anesthesia. The burn area was covered with lidocaine-soaked gauze for 20–30 min before the procedure. Wounds were infiltrated with buffered lidocaine for a minimum of 5 min prior to the procedure. The maximum lidocaine dose without adrenaline was 5 mg/kg and with adrenaline 7 mg/kg.

DEX 100 mcg/ml and sKET 25 mg/ml were used without dilution. A 1 ml syringe with a nasal atomizer was used for drug administration. 2.0 mcg/kg DEX and 1.0 mg/kg sKET were used following the local and national guidelines [29]. The dose was equally divided between both nostrils when the recommended volume per nostril (0.3 ml/nostril) [30] was exceeded.

Patients were monitored from the administration of the trial drug until the patient had recovered, and Ramsay score was 1. Pain (FLACC score) and sedation (Ramsay score) were assessed at least every 5 min before and during the procedure and every 10 min after the procedure. Oxygen saturation (SpO2) and heart rate (HR) were monitored continuously. SpO2 and HR were recorded at the same timepoints with FLACC and Ramsay unless significant changes occurred at other times. Any aberration from normal values described in Pediatric Advanced Life Support (PALS) [31] were considered significant. Assessment and monitoring were done by two experienced pediatricians (AN, KS) who were familiar with both the FLACC and Ramsay scales. Five patients were initially evaluated together to ensure uniform assessment.

The procedure was started when Ramsay score 2 was reached. If it was not reached within 30 min and the child was not co-operative the procedure was not carried through within trial protocol and other sedation was used for the treatment. Patients were able to leave the ED when Ramsay score was 1 and the patient had returned to the habitual condition.

Patients were randomized to two groups: 1) IN DEX 2.0 mcg/kg, 2) IN sKET 1.0 mg/kg. We used block randomization (blocks of 10 subjects (5 from both arms), except one block with 12 subjects). A physician not participating in the trial created a randomization list after a random draw and filled the opaque envelopes with information (trial drug and dose table) and numbered them according to the list. Envelopes were then used in number order. Randomization was kept in a sealed envelope during trial period and the seal was intact at the time of trial closure. The blocks were used in order and the order was not known by the physician performing the assessment of the patient.

The trial physician, ED physician and nurses caring for the patient, as well as the patient and parents were blinded to the trial drug, as were all other staff working in the ED. Trial drug was prepared and administered by a nurse who was not involved in the care of the patient otherwise. Trial physician, ED physician and nurses caring for the patient were not present at the time of drug administration. Patients and their parents did not have knowledge about the trial drugs (e.g., smell or taste, nasal irritation) or the difference between the volume of the two drugs.

An a priori power analysis was conducted to test the null hypothesis of this superiority trial. We aimed for a mean difference of 2 on FLACC which would be the smallest clinically relevant difference. Assuming a within-group standard deviation of 2.5 we would need n = 26 patient per group to obtain 0.80 power for the test.

Continuous variables are presented using medians and interquartile ranges (IQRs) and tested with Mann–Whitney U test and tested correlations using the Spearman method. Categorical variables are presented using counts and percentages and tested using Fisher’s test. We considered p-values below 0.05 significant. All analyses were done using R version 4.2.2. [32].