Prospective multicentre randomised, double-blind, equivalence study comparing clonidine and midazolam as intravenous sedative agents in critically ill children: the SLEEPS (Safety profiLe, Efficacy and Equivalence in Paediatric intensive care Sedation) study

Abstract

BackgroundChildren in paediatric intensive care units (PICUs) require analgesia and sedation but both undersedation and oversedation can be harmful.ObjectiveEvaluation of intravenous (i.v.) clonidine as an alternative to i.v. midazolam.DesignMulticentre, double-blind, randomised equivalence trial.SettingTen UK PICUs.ParticipantsChildren (30 days to 15 years inclusive) weighing ≤ 50 kg, expected to require ventilation on PICU for > 12 hours.InterventionsClonidine (3 µg/kg loading then 0–3 µg/kg/hour) versus midazolam (200 µg/kg loading then 0–200 µg/kg/hour). Maintenance infusion rates adjusted according to behavioural assessment (COMFORT score). Both groups also received morphine.Main outcome measuresPrimary end point Adequate sedation defined by COMFORT score of 17–26 for ≥ 80% of the time with a ± 0.15 margin of equivalence.Secondary end points Percentage of time spent adequately sedated, increase in sedation/analgesia, recovery after sedation, side effects and safety data.ResultsThe study planned to recruit 1000 children. In total, 129 children were randomised, of whom 120 (93%) contributed data for the primary outcome. The proportion of children who were adequately sedated for ≥ 80% of the time was 21 of 61 (34.4%) – clonidine, and 18 of 59 (30.5%) – midazolam. The difference in proportions for clonidine–midazolam was 0.04 [95% confidence interval (CI) –0.13 to 0.21], and, with the 95% CI including values outside the range of equivalence (–0.15 to 0.15), equivalence was not demonstrated; however, the study was underpowered. Non-inferiority of clonidine to midazolam was established, with the only values outside the equivalence range favouring clonidine. Times to reach maximum sedation and analgesia were comparable hazard ratios: 0.99 (95% CI 0.53 to 1.82) and 1.18 (95% CI 0.49 to 2.86), respectively. Percentage time spent adequately sedated was similar [medians clonidine 73.8% vs. midazolam 72.8%: difference in medians 0.66 (95% CI –5.25 to 7.24)]. Treatment failure was 12 of 64 (18.8%) on clonidine and 7 of 61 (11.5%) on midazolam [risk ratio (RR) 1.63, 95% CI 0.69 to 3.88]. Proportions with withdrawal symptoms [28/60 (46.7%) vs. 30/58 (52.6%)] were similar (RR 0.89, 95% CI 0.62 to 1.28), but a greater proportion required clinical intervention in those receiving midazolam [11/60 (18.3%) vs. 16/58 (27.6%) (RR 0.66, 95% CI 0.34 to 1.31)]. Post treatment, one child on clonidine experienced mild rebound hypertension, not requiring intervention. A higher incidence of inotropic support during the first 12 hours was required for those on clonidine [clonidine 5/45 (11.1%) vs. midazolam 3/52 (5.8%)] (RR 1.93 95% CI 0.49 to 7.61).ConclusionsClonidine is an alternative to midazolam. Our trial-based economic evaluation suggests that clonidine is likely to be a cost-effective sedative agent in the PICU in comparison with midazolam (probability of cost-effectiveness exceeds 50%). Rebound hypertension did not appear to be a significant problem with clonidine but, owing to its effects on heart rate, specific cardiovascular attention needs to be taken during the loading and early infusion phase. Neither drug in combination with morphine provided ideal sedation, suggesting that in unparalysed patients a third background agent is necessary. The disappointing recruitment rates reflect a reluctance of parents to provide consent when established on a sedation regimen, and reluctance of clinicians to allow sedation to be studied in unstable critically ill children. Future studies will require less exacting protocols allowing enhanced recruitment.Trial registrationCurrent Controlled Trials ISRCTN02639863.FundingThis project was funded by the NIHR Health Technology Assessment programme and will be published in full inHealth Technology Assessment; Vol. 18, No. 71. See the NIHR Journals Library website for further project information.