The present study is an adaptive, multicenter, randomised, double blind, placebo-controlled, non-profit, superiority clinical trial with two-parallel groups. The study will be conducted in adherence to the principles of the World Medical Association’s Declaration of Helsinki. An internal Data and Safety Monitoring Board (DSMB) has been nominated to monitor data and safety and it will decide on the continuation, modification or termination of the trial. The Study Protocol Final Version 2.0 (18 September 2016) was approved by the Ethics Committee of the Coordinating Center (University-Hospital S.Orsola-Malpighi of Bologna) on 22 March 2017. All centers received the approval from the local Ethical Committee. The study was authorized by the Italian Medicines Agency (AIFA, ID TIP-15-01) and registered in the National Monitoring Center for Clinical Trial (OsSC) and successively in the Eudra CT Register (Identification Number 2015-002114-80). In addition, the study was prospectively registered on the ClinicalTrial.gov Registry (registration date 24 August 2018) with the Identification Number NCT03645603. The protocol has been designed following the SPIRIT international guidelines: the Figure 1 shows the SPIRIT-schedule of enrolment, interventions and assessments, and a populated SPIRIT Checklist is attached as Additional File 1.
Setting
The study will involve three PICUs belonging to three tertiary-care pediatric academic centers (University-Hospital S.Orsola-Malpighi, Policlinic, Bologna, Italy; Catholic University of Rome A. Gemelli, Policlinic, Rome, Italy; University-Hospital of Padua, Padua, Italy).
Study population
The study population will involve patients admitted to PICU who meet the following criteria (Figure 2). Inclusion Criteria: (1) age from 0 to 18 years, (2) post-natal age ≥ 7 days and post-menstrual age (gestational age at birth (weeks) plus weeks since birth) ≥ 37 weeks, (3) having received continuous intravenous analgosedation with opioids and/or benzodiazepines for at least 5 days, (4) having required invasive or non-invasive mechanical ventilation, (5) presence of clinical conditions that allow the treating physician to start the analgosedation weaning, including absence of signs and symptoms of WS, (6) parents’ written consent obtained.
Exclusion Criteria: (1) presence of hemodynamic instability according to the treating physician judgement; (2) receiving inotropic or antihypertensive treatments (ß-blockers, calcium antagonists, ACE inhibitors, digoxin, nicardipine, nitroglycerin), (3) presence of II or III degree cardiac atrio-ventricular (AV) block; (4) known or suspected hypersensitivity to alpha-agonists; (5) presence of persistent unknown-origin fever or history of malignant hyperthermia; (6) use of alpha-agonist (clonidine or dexmedetomidine) in the 30 days preceding the study enrolment.
Definitions
Withdrawal syndrome: WS is defined as iatrogenic clinical syndrome that manifests when the administration of a sedative or analgesic agent is abruptly discontinued or too rapidly weaned in a patient who is physically tolerant [2].
Withdrawal Assessment Tool version 1 (WAT-1): WAT-1 is a validated assessment tool for the monitoring of withdrawal symptoms in pediatric patients. This twice-daily assessment consists of 11 items, determined by the following components: a review of the patient's record for the past 12 hours, a direct observation of the patient for 2 minutes, a patient assessment using a progressive stimulus and an assessment of post-stimulus recovery [5]. The score ranges from 0 to 12 and a score ≥3 indicates the presence of signs or symptoms of WS. The severity of WS is higher as the score increases, as defined by the WAT-1 official definition [5].
Recruitment and Consent
Comprehensive information will be provided by each Center Principal Investigator to parents of children potentially involved. A detailed information sheet has been designed to support the oral communication. A written informed consent will be obtained from both parents for each involved child. Even when appropriate for age, a child’s consent will be not needed because of the sedation status. A guarantee of optimal children’s care will be assured independently of the study involvement. If present, a consent refusal will be recorded.
Randomization
Each patient will be randomly assigned to one of the 2 treatment groups: treatment-A group (receiving dexmedetomidine) or treatment-B group (receiving placebo). An identification code will be individually assigned to each patient. The Investigational Drug Service of the Coordinating Center has generate a block randomisation scheme on 11 June 2017 using the Web site Randomization.com (available online http://radomization.com). This confidential document will be available only to the non-blinded staff, who will carry out the preparation of the treatments. Thus, the allocation sequence and the treatment administration will be unknown for the blinded researchers, including the study Principal Investigator. During the study, two sealed copies of the randomisation list that clearly show the treatment attributed to the patient will be available for emergencies. A sealed list will be kept in the archive of the Investigational Drug Service of the Coordinating Center and the others in the archive of each Principal Investigator. If an opening is needed, the Investigator will be asked to report the reason, the date/time of the opening and to immediately notify the Project Principal Investigator.
Interventions
Twenty-four hours before the start of the analgosedation weaning, an intravenous infusion of dexmedetomidine or placebo (i.e. normal saline) will be started according to the following schedule (Figure 2). The starting dose will be 0.4 mcg/kg/h. No loading dose will be administered. If the infusion will be well-tolerated (i.e. without the occurrence of adverse effects), the dose will be increased of 0.2 mcg/kg/h per hour up to 0.8 mcg/kg/h. Given the pharmacological peculiarities of the neonatal period [15], newborns will receive a starting dose of 0.2 mcg/kg/h, which will be increased of 0.1mcg/Kg/h up to 0.4mcg/Kg/h. At 24 hours of dexmedetomidine infusion, the analgosedation weaning process will be started, consisting in a 10% reduction of one of the drugs every 12 hours. If requested, a switch from opioid and/or benzodiazepine to an equipotent drug of the same pharmacological class but longer half-life will be allowed (including enteral methadone, morphine, lorazepam). The switch should aim to facilitate the patient’s management. In the same way as iv drugs, enteral drugs will be weaned with 10% reduction every 12 hours.
The WAT-1 scale will be administered every 12 hours of treatment infusion. If WS will be diagnosed, the clinician will administer a rescue dose of the using opioid and/or benzodiazepine, repeatable until resolution of the crisis, and will increase the dexmedetomidine/placebo dose by 0.2 mcg/kg/h (0.1 mcg/Kg/h in neonates). If the following WAT-1 score shows a decrease by at least 1 point compared with the previous one, the weaning program will be restarted (by 10% of reduction) and the current dexmedetomidine/placebo dosage will be maintained. If the WS symptoms persist, dexmedetomidine/placebo will be increased by 0.2 mcg/kg/h (0.1 mcg/Kg/h in neonates) according to the WS score, up to a maximum of 1.6 mcg/Kg/h (0.8 mcg/kg/h in neonates).
Once the analgosedation weaning will be completed, dexmedetomidine will be weaned or discontinued. A gradual reduction of the dexmedetomidine dose is strongly recommended to prevent the risk of dexmedetomidine withdrawal [16, 17], but it is not mandatory. Since the analysis of dexmedetomidine weaning is not a specific aim of the present study, no specific protocol will be recommended. Time and modality of dexmedetomidine weaning will be recorded.
A follow-up visit will be performed at five days after PICU discharge, with the aim to collect the following data: (1) actual duration of the analgosedation weaning when longer than 5 days, (2) values of WAT-1 scores collected every 12 hours up to 72 hours after the analgosedation discontinuation, (3) length of dexmedetomidine weaning (hours), (4) occurrence of signs and symptoms of dexmedetomidine withdrawal.
Outcome measures
Primary outcome measure
The primary outcome measure of our study is the efficacy of the treatment efficacy in the prevention of WS, i.e. the reduction of the WS rate in the DEX arm comparing with the placebo arm.
Secondary outcome measures
Treatment safety
The safety of the treatment will be assessed: (1) with a strict monitoring of hemodynamic parameters (heart rate, systolic and diastolic blood pressure) which are considered altered if their values differ more than 20% comparing with the patient’s baseline values, (2) collecting every Adverse Reactions (ARs), Adverse Events (AEs), Serious AEs and Suspected Unexpected Serious Adverse Reactions (SUSARs). Every AR, AE or SUSAR potentially-related to dexmedetomidine will be summarized separately.
Secondary outcome measures related to Efficacy
Secondary outcome measures evaluated to confirm the efficacy of the treatment will be: (1) trend of WAT-1 score; (2) number of rescue doses required for WS symptoms; (3) number of temporary discontinuations of the analgosedation weaning due to the presence of WS, (4) duration of analgosedation weaning (days); (5) length of mechanical ventilation (days); (6) PICU length of stay (days).
Data collection and management
The blinded investigators will collect data by means of a standardized paper case-report form (CRF). Paper CRFs will be stored in accordance with national regulations. Paper CRFs will have an identifiable patient code in order to allow a clinical follow-up and a data monitoring by national coordinators or regulatory committees. Investigators will transcribe patient’s data into an electronic CRF using the identification code. No patients’ identifiable data will be directly accessible from the electronic CRF. Data recorded on each CRF will be entered in a dedicated database, checked and subsequently processed.
Sample size
The sample size has been calculated in respect with our primary outcome measure, that is the reduction of the WS rate. A recent multicenter national study reported a WS incidence of 64.6% among PICU patients receiving more than 5 days of analgosedation with opioids and/or benzodiazepines [4]. Given the small level of available evidence, to have a clearer picture of the potential reduction of WS, we decided to apply a classical prior elicitation process in four separate stages [18, 19]: (i) selecting the experts and identifying the aspects of the problem to elicit; (ii) proceeding with the elicitation process, that is, interaction with the experts; (iii) fitting the probability distributions to the expert's summaries; and (iv) including the information from the elicitation process in the evaluation of the sample size requirements.
Prior elicitation process
Twenty-eight pediatric intensivists and nurses expert in the field of analgesia and sedation were asked via an email survey about: (i) if they believe that DEX has some efficacy in reducing WS prevalence and (ii) if yes, what is the expected reduction in WS prevalence via DEX compared to standard care (placebo in the terms of this trial). All of the experts replied to the survey, and 25 (92%) of them declared to expect DEX to have an effect in WS prevalence: 23 of them declared that DEX is able to reduce the WS prevalence, 2 of them declared that DEX is not able to prevent WS and the last 3 were not able to provide an certain answer. Among those who replied affirmatively (n=23), 18 provided an estimated percentage of WS reduction in DEX arm vs. placebo. The median expected reduction in WS prevalence was 47.5% (I quartile 33.75% and III quartile 51.25%).
Adaptive design
We assumed that the WS prevalence, according to literature [4], amounts at about πPlacebo= 64.6% and this value was used for sample size calculation. According to the prior elicitation process, we conservatively adopted the most pessimistic evaluation provided by the experts, assuming a reduction in WS prevalence of 33.75 points in the DEX group compared to Placebo. The sample size has been computed using an adaptive approach based on a two-stage Group Sequential Design with an interim Sample Size Reassessment in order to compensate for discrepancies between expected and observed incidence of the primary endpoint at the first stage [20].
The following assumptions were considered:
- Two-sided superiority Z test without continuity correction for rejecting the null hypothesis H0: πPlacebo – πDEX = 0;
- α = 0.05;
- Power 0.90;
- Efficacy bounds are derived using an O'Brien-Fleming boundary;
- No futility bounds;
- An incidence rate of πPlacebo = 0.646 in Placebo arm and πDEX = 0.375 in DEX arm (corresponding to a 33.75% reduction out of the 64.6%);
- An allocation rate of 1:1;
Based on the above assumptions, this yields to 138 patients overall. To account for a possible R=5% dropout rate, the sample size n has been increased [21] to total patients. Thus, the trial has been designed as 20 + 20 = 40 patients at first stage and 154 overall (Supplemental Table 1).
Sample Size Reassessment
A promising zone design has been considered. Following steps have been provided at the first interim analysis step:
1. If computed z test statistic crosses the O'Brien-Fleming Bounds, then study will be terminated early for efficacy (ie );
2. If computed z test statistic does not cross the O'Brien-Fleming Bounds different solutions are possible:
a) If interim conditional power (CP) is in the promising zone, say comprised between 0.3 and 0.9, then a sample size reassessment will be performed;
b) Otherwise, the study will continue until the second stage;
c) If the CP is critically lower than 0.3, then the internal DSMB will be consulted to discuss a termination of the trial for futility reasons.
A simulation has been performed to evaluate the extent of a sample size reassessment assuming that the null hypothesis will not be rejected at the first interim analysis, and assuming that the interim conditional power is in the promising zone.
To this purpose, reductions in the expected difference in the primary endpoint at interim, are investigated. A range of interim event rates say p interim/placebo , from 0.5 to 0.75 following a pace of 0.01, has been hypothesized in the Placebo arm. The event rate at interim in DEX arm is supposed equal to pinterimto/DEX = 0.6625 * p interim/placebo , where 0.6625 corresponds to the most pessimistic reduction hypothesized above.
For each scenario, the number of patients to add to the sample size at the second stage (nadd), given the interim results, is computed to achieve a conditional statistical power of at least 0.8 (Supplemental Table 2).
Sample size estimation has been performed using the gsDesign [22] and the R-System [23].
Statistical analysis
The primary endpoint will be analyzed for the Intent-to-Treat (ITT) population in terms of WS rate (placebo vs. dexmedetomidine). If the interim conditional power will be not in the promising zone, no modifications to the design will be performed.
The final efficacy of the treatment will be evaluated using a Z test statistics; after having reached the sample size foreseen for each of the stages, an evaluation using a 95% repeated confidence interval will be performed.
Demographics (age, sex, and race) and other baseline characteristics will be summarized using descriptive statistics. As for secondary outcome measures, no formal testing procedure will be adopted, to avoid the inflation of type-I error. Differences in distributions will be evaluated on a clinical reasoning basis.
An analytical detailed list of patients who will discontinue the study for AR, AE or SUSARs will be collected.
All analyses will be performed using the R-system [23].
Trial status
The present trial (Study Protocol Final Version 2.0, approved on 18 September 2016) is currently ongoing. All centers are actively recruiting patients. From the beginning of the enrolment (30 August 2018) to date (22 July 2019), 34 of 160 patients have been recruited. The period for the whole population enrolment has been estimated as a three-year period (estimated end-of-enrolment date: August 2021).