RESTORE

Protocolized Sedation vs Usual Care in Pediatric Patients Mechanically Ventilated for Acute Respiratory Failure. A Randomized Clinical Trial

RESTORE Study Investigators. JAMA. First published on line Jan 20 2015: doi10.1001/jama.2014.18399

Clinical Question

  • Does a nurse-implemented, goal-directed sedation protocol result in fewer ventilator days in children with acute respiratory failure?

Design

  • Cluster randomised trial
    • units statified by size and, within each stratum, assigned to intervention or control group using computer generated random numbers
  • Unblinded
  • Paediatric Intensive Care Units (PICUS) invited to participate in the study needed to:
    • demonstrate nursing and physician leadership support
    • enroll a minimum of 3 patients per month

Setting

  • 31 PICUs in US
    • 22 PICU’s initially and a further 9 were added at one year
  • 2009-2013

Population

  • Inclusion:
    • 2 weeks to 17 years receiving mechanical ventilation for acute airway and/or parenchymal lung disease
    • PICU had no sedation protocol
  • Exclusion:
    • length of mechanical ventilation unlikely to be altered by sedation management e.g anticipated extubation within 24 hours
  • 2449 patients
    • mean age 4.7 years (2 weeks to 17 years)
    • 52% were between 2 weeks and 2 years

Intervention

  • Protocolised sedation
    • 17 PICUs
    • 1225 patients
  • Included
    • Standard choice of sedative agents.
      • Primary agents: morphine (or fentanyl if hypotensive or reactive airway disease) and midazolam
      • Secondary agents:
        • dexmedetomidine and propofol (to facilitate extubation)
        • clonidine, pentobarbital or ketamine (if unresponsive to primary agents)
        • methadone if Withdrawal Assessment Tool (WAT-1) scores continued to above target
    • Daily team discussions to determine phase of illness and readiness to wean (acute, titration, or weaning phase)
    • prescribing a target State Behavioural Score (SBS)
    • daily extubation readiness test (ERT) when spontaneously breathing with an oxygenation index of 6 or lower
      • if ERT unfavourable -> returned to pretest ventilator settings and ERT repeated next morning
      • if ERT showed failure because of excessive sedation, this was weaned and then retested
      • if ERT favourable -> patient extubated within 6 hours unless needed to remain intubated for non-pulmonary reasons
    • adjustment of sedatives at least 8 hourly
    • discontinuation of opioids and benzodiazepines when no longer necessary
    • written sedation weaning plan when transferred out of PICU

Control

  • Usual care
    • 14 PICUs
    • 1224
  • Sedation managed in usual way for that unit. Sedatives were selected, prescribed, and titrated at the discretion of the medical team. No recommendations were made for extubation readiness testing
  • Control PICUs did not receive a copy of the algorithm for the intervention group

A baseline phase occurred in both groups where all PICUs implemented the same validated pain, sedation and iatrogenic withdrawal assessment instruments

Outcome

  • Primary outcome:
    • median duration of mechanical ventilation
      • 6.5 days in both groups, p=0.61
      • no difference when adjusted for age group, PRISM-III-12 score and POPC score > 1 at enrollment
    • % not successfully extubated by day 28
      • 8% in protocol group vs 11% in usual care group, P=0.19
  • 8 Secondary outcomes: (protocol vs. usual care)
    • Neurological testing to evaluate a change in mental status was the only outcome to be statistically different
      • 14% vs 19%, P=0.03
    • No difference in:
      • median time to recovery from acute respiratory failure
      • median time to wean from mechanical ventilation
      • median time to onset of assisted breathing
      • median length of stay in PICU
      • median length of stay in hospital
      • % in hospital 28 day mortality
      • % in hospital 90 day mortality PICU length of stay
  • Adverse effects
    • Higher % of post-extubation stridor in usual care group 7% vs 4% P=0.03
    • Higher % of immobility-related stage 2 or more pressure ulcers in usual care group 2% vs <1% P-0.01
    • No difference in pain management, extubation failure, unplanned extubations, ventilator associated pneumonia, catheter related blood stream infections or unplanned removal of lines
  • Use of sedative agents
    • dexmedetomidine used twice as much in usual care group (49% vs 23%, P=<0.001)
    • methadone used twice as much in usual care group (30% vs 12%, P=<0.001)
    • chloralhydrate use greater in usual care group (15% vs 3%m P=0.01)
    • use of midazolam was similar in the protocol vs usual care group (82% vs 90%)
    • fentanyl was more commonly used as first line opioid in the usual care group 81% vs 35%

Authors’ Conclusions

  • The use of a sedation protocol compared with usual care did not reduce the duration of mechanical ventilation in critically ill children presenting with acute respiratory failure

Strengths

  • Well powered. Full recruitment for 90% power
  • Large multi-centre paediatric study with useful primary outcome measures
  • The multi-disciplinary staff in units randomised to protocolised weaning received a combination of direct learning and training materials to try and maximise compliance
  • Adherence with protocol was high (71-100%). Prescribed sedation was achieved 98% of the time
  • The use of different sedation protocols for varying stages of the illness is an interesting concept and makes physiological sense

Weaknesses

  • 35% of patients who were eligible were not included as consent was not obtained
    • consent rates were lower in the intervention group (72% vs 84%; p=0.01)
  • Exclusion of patients that they thought would be sedated for short period (e.g. peri-op). This might be appropriate, and does seem to be standard in adult sedation studies
  • The use of dexmedetomidine increased each year throughout this study. It was also used more often than in the control group which may have resulted in less opioid and benzodiazepine requirement
  • Selection bias may have occurred as this was an unblinded study
  • The baseline characteristics were different. The intervention group enrolled more patients < 2 years (58% vs 45%) and more patients with bronchiolitis (typically more difficult to sedate). The predicted risk of mortality based on PRISM III-12 score was also lower in patients < 2 years in the intervention group (1.7% vs. 3.8%)
  • Chloral hydrate was used in 3% of cases in the protocol group. This is a non-protocol sedative. Arguably would not have influenced the overall result greatly

The Bottom Line

  • Clinical outcomes for nurse led sedation protocol for children requiring mechanical ventilation compared with ‘usual care’ was no different. With in hospital mortality rates of around 6% it may just be that ‘usual care’ is actually ‘exemplary care’. ARISE & ProCESS are other examples of this. Further studies perhaps using less complex sedation management strategies, more selective sedative agent choice and refined outcome measures may be helpful in the future.

Leave a Reply