2. Dosage study outcomes.

Study ID	Sample	Intervention	Outcomes	Outcome results	Treatment completed
Hydroxyzine
Faytrouny 2007	n = 30 14 females, 16 males Mean age (SD) months: Group 1 61.9 (11.9) Group 2 53.7 (12.8)	Group 1: hydroxyzine (20 mg 24 hours before) + hydroxyzine (3.7 mg/kg at the appointment) Group 2: hydroxyzine (3.7 mg/kg at the appointment) All oral	Houpt. Analysed using ANOVA and Mann‐Whitney	No significant differences at any time point. At 20 minutes Houpt Group 1, 5.2 (SD 1.5) and Group 2, 4.6 (SD 1.6) No adverse effects reported in either group Monitoring: pulse oximeter	‐
Midazolam (intranasal)
Al‐Rakaf 2001	n = 38 children Mean age (SD) in years and gender: Group 1 (n = 12) 3.75 (0.75), 6 males, 6 females Group 2 (n = 13) 4.3 (0.65), 6 males, 7 females Group 3 (n = 13) 4 (0.71), 6 males, 7 females	Group 1: midazolam (0.3 mg/kg) Group 2: midazolam (0.4 mg/kg) Group 3: midazolam (0.5 mg/kg) All intranasal	Houpt. Analysed using Tukey's range test and non‐parametric 2‐factor ANOVA Duration of sedation	Significant difference in Houpt behavioural scores between Group 3 and Group 1 (P < 0.0001) and between Group 3 and Group 2 (P < 0.01) No difference in outcomes between fasting and no‐fasting in each group (P = 0.8286) None of the children were asleep Adverse effects: sneezing and coughing during administration and drowsiness (Groups 1 and 2), diplopia (only in Group 3) Monitoring: pulse oximeter	79% 0.3 mg/kg, 96% 0.4 mg/kg and 100% 0.5 mg/kg completed treatment
Lam 2005	n = 23 (12 Group 1, 11 Group 2) Mean age (range) in years: 5.13 (2‐9) Mean weight (range) in kg: 21.74 (12‐30) 15 males, 7 females	Group 1: midazolam (0.2 mg/kg) (intramuscular) Group 2: midazolam (0.2 mg/kg) (intranasal) All used as premed for unspecified intravenous sedation drug	Houpt. Analysed using Mann‐Whitney. Inter‐examiner variability assessed using Spearmans rank correlation Good/excellent sedation levels in each group	Patients more deeply sedated in Group 1 at time of local anaesthesia administration and venepuncture (P < 0.048, P < 0.015 respectively). 1 observer found Group 1 (intramuscular) significantly more effective than Group 2 (P < 0.04). Not significant for the second observer (P = 0.056). Individual outcomes for each observer not reported Good/excellent sedation 12/12 (100%) in intramuscular group and 6/11 (54%) in the intranasal group Children more likely to be drowsy in Group 1 Adverse effects: none reported Monitoring: heart rate, respiratory rate, blood pressure, oxygen saturation	Treatment completed by all participants
Lee‐Kim 2004	n = 40 Mean age (unclear, possibly SD) in months; mean weight (unclear, possibly SD) in kg; gender: Group 1 (n = 20) 40.8 (11), 17 (3.6), 11 males, 9 females Group 2 (n = 20) 38.5 (9.8), 16.2 (4), 10 males, 10 females	Group 1: midazolam (0.7 mg/kg) (oral) Group 2: midazolam (0.3 mg/kg) (nasal)	Modified Houpt ‐ domains of sleep movement and crying but no overall measure. Analysed with ANOVA, Chi2 statistic and t‐test Mean time of onset and mean working time in each group	Data presented on graphs only and text states no significant differences in Houpt using multivariate ANOVA (P = 0.749) Onset of sedation mean 5.55 minutes (SD 2.2) for nasal and mean 15.5 minutes (SD 5) for oral (P < 0.001) Mean working time was 29.3 minutes (SD 11.6) for nasal and & 38.1 min (SD 7.58) for PO (P = 0.007) Adverse effects: none reported, no differences between groups. Monitoring: oxygen saturation, heart rate, respiratory rate	All participants completed
Shashikiran 2006	n = 40 Group 1: 11 males, 9 females, mean age (SD) in years 3.4 (0.6), mean weight (SD) in kg 12.2 (1.2) Group 2: 8 males, 12 females, mean age (SD) in years 3.5 (0.7), mean weight (SD) in kg 12.6 (1.4)	Group 1: midazolam (0.2 mg/kg) (intramuscular) Group 2: midazolam (0.2 mg/kg) (intranasal)	Houpt and Fukuta. Analysed using Chi2 statistic and Mann‐Whitney U test	No difference in behaviour between groups (Chi2 = 0.37, P = 0.83), but both groups showed improvement from baseline. Intranasal midazolam was significantly faster acting at all time points and allowed a shorter treatment time overall (P < 0.001) Mean onset times 15.7 ± 2.0 minutes intramuscular versus 10.8 ± 2.0 minutes intranasal Adverse effects: 2 patients in the intramuscular group and 6 patients in the intranasal group showed instances of sneezing/coughing/hiccups after the administration of the sedative (difference not statistically significant) No fasting pretreatment and no vomiting in either group Monitoring: heart rate, respiratory rate	Score 3 given to excellent, 2 to satisfactory, 1 to unsatisfactory
Shanmugaavel 2016a	n = 20 Age range = 4‐7 years Group 1 (n = 10) Group 2 (n = 10)	Group 1: midazolam (0.2 mg/kg) (intranasal) Group 2: midazolam (0.2 mg/kg) (sublingual)	Venham's Clinical Anxiety Scale Salivary cortisol level	Significant decrease in anxiety in Group 1 (P = 0.004) and Group 2 (P = 0.0.003) 20 minutes after drug administration Group 1 showed statistically significant decrease in anxiety at each of the 4 points of measurement during operative procedure (T1, T2, T3, T4), whereas Group 2 did not show statistically significant change at T1, T2 and T3 No significant difference in salivary cortisol levels before and after drug administration in Group 1 and Group 2 (P = 0.07, P = 0.38 respectively) No significant correlation between decrease in clinical anxiety and salivary cortisol level in Group 1 and Group 2 (P = 0.554, P = 0.457 respectively) Adverse effects: not reported	‐
Shanmugaavel 2016b	n = 40 Mean age (SD) in years, gender, weight (SD) in kg: Group A: 5.1 (1.07), 12 males and 8 females, 17.5 (4.39) Group B: 5.2 (1.15), 12 males and 8 females, 17.4 (4.33)	Group A: midazolam (0.2 mg/kg) (intranasal) Group B: midazolam (0.2 mg/kg) (sublingual)	Venham's Clinical Anxiety Scale Acceptance (Al‐Rakaf 2001)	No statistically significant difference in Venham's anxiety score between groups at baseline or at the end time point (T4) Mean (SD) Venham's score at T4 in Group A 0.35 (0.59) and Group B 0.45 (1.10) P = 0.001 Statistically significant difference in acceptance with better acceptance in Group B compared to Group A (95% versus 40%, P = 0.001) Adverse effects not reported	‐
Midazolam (oral)
Aydintug 2004	n = 50 Mean age (unclear, possibly SD) in years; mean weight (unclear, possibly SD) in kg; gender: Group 1 (n = 25), 5.36 (1.7), 19.068 (3.43), 18 males and 7 females Group 2 (n = 25), 4.96 (1.513), 17.804 (3.08), 12 males and 13 females	Group 1: midazolam (0.5 mg/kg) (oral) Group 2: midazolam (0.35 mg/kg) (rectal)	Ramsay Sedation Score, acceptance of application, acceptance of local anaesthesia, operating conditions, state of amnesia Analysed using Chi2 test	Acceptance of application significantly better in oral group (72% excellent in oral group compared to 20% excellent in rectal group, P < 0.05) No significant difference seen (P > 0.05) between acceptance of local anaesthesia, state of amnesia or operating conditions. Ramsay's Sedation Scores not reported Adverse effects: no significant difference (P > 0.05) in adverse effects between groups (56% oral, 44% rectal, included hypoxaemia, vomiting and nausea, disinhibition) Monitoring: oxygen saturation, heart rate, blood pressure	Treatment completed by all participants
Isik 2008b	Mean age (SD) in years, gender, mean weight (SD) in kg: Group 1 (n = 14), 4.6 (1.2), 7 males and 7 females, 15.6 (2.8) Group 2 (n = 13), 4.4 (1.0), 8 males and 5 females, 16.2 (2.4) Group 3 (n = 13), 4.4 (0.9), 6 males and 7 females, 16.1 (2.4) Group 4 (n = 13), 4.3 (0.9), 5 males and 8 females, 15.8 (2.6)	Group 1: midazolam (0.2 mg/kg) Group 2: midazolam (0.5 mg/kg) Group 3: midazolam (0.75 mg/kg) Group 4: midazolam (1 mg/kg) All oral	Ramsay Sedation Score Analysed using Kruskal‐Wallis, Mann‐Whitney U test	At the 20 minute time point, mean Ramsay Sedation Score was 1.2 (0.4), 1.6 (0.5), 2.2 (0.6), 2.7 (1.0) in Groups 1‐4 respectively. Children in Groups 3 and 4 were more sedated than children in Groups 1 and 2 (P < 0.05) Sedation considered inadequate in 12/14, 5/13, 3/13 and 5/13 in Groups 1‐4 respectively Mean recovery time was 45 (0), 45 (0), 47.3 (8.3), and 57.7 (42.8) minutes in Groups 1‐4 respectively Adverse effects: most of the adverse effects were seen in Group 4 with 1 patient desaturating and 3 presenting with delayed recovery, no adverse effects in Group 1 and "very few" in Groups 2 and 3 Monitoring: pulse oximeter	‐
Somri 2012	n = 90 (30 per group) Age range = 3‐10 years Mean age (SD) in years, weight (SD) in kg: Group 1: 5.6 + 1.85 years, 19.2 + 3.68 kg Group 2: 5.6 + 1.67 years, 19.7 + 3.38 kg Group 3: 6.2 + 2.00 years, 20.3 + 3.65 kg	Group 1: midazolam (0.5 mg/kg) Group 2: midazolam (0.75 mg/kg) Group 3: midazolam (1 mg/kg)	Wisconsin Sedation Scale Houpt behavioural rating scale Parent satisfaction	Sedation and behaviour co‐operation scores were noted at baseline, 15 minutes, 30 minutes and 45 minutes. Significant difference with sedation scores in Group 1 lower than Group 2 and Group 3 (P < 0.001) No significant difference (P < 0.001) in sedation score of Group 2 and Group 3 except at baseline Behavioural co‐operation score was significantly lower (P < 0.001) in Group 1 compared to Group 2 and Group 3. Significant difference (P < 0.001) in behaviour scores of Group 2 and Group 3 with Group 2 having lower scores at baseline and 45 minutes Significant difference (P = 0.025) in completion scores between Group 1 and Group 3 No significant difference (P = 0.43) in duration between the groups Significant difference in discharge time between the groups. Group 1 had shorter mean inpatient stay (85 minutes + 18.5, P < 0.001) compared to Group 2 (103.7 + 13.3 minutes) and Group 3 (137 + 14.7 minutes) Significant difference in parent satisfaction (P < 0.001) where Group 1 is lower than Group 2 and Group 3. No significant difference (P = 0.147) between Group 2 and Group 3 Adverse effects: respiratory events Group 2 (3/30), Group 3 (10/30) Nausea and drowsiness Group 1 (3/30), Group 2 (7/30), Group 3 (12/30)	Group 1 20% (n = 6), Group 2 6.7% (n = 2) did not complete treatment

ANOVA = analysis of variance; n = number; SD = standard deviation.