Continuous versus intermittent neuromuscular blockade in patients during targeted temperature management after resuscitation from cardiac arrest

Stöckl et al. Resuscitation 2017;120:14–19. doi: 10.1016/j.resuscitation.2017.08.238

Clinical Question

  • In adult patients post out of hospital cardiac arrest, does a continuous or on demand application of neuromuscular blockade (NMB) reduce the number of shivering episodes during targeted temperature management (TTM)?


  • Current resuscitation guidelines advocate a temperature between 32 and 36 degrees
  • This can lead to shivering which can increase metabolic demand and counteract the beneficial effects of TTM
  • There is little research into the role of NMB in this setting to avoid shivering, and some guidelines have not advocated their use due to increased risk of critical illness polyneuropathy and myopathy


  • Randomized, double blinded, double dummy
  • Treatment assignments randomly generated using computer and sealed envelopes
  • Allocated in a 1:1 ratio
  • Nursing staff not involved in the study or patient’s care prepared the continuous or bolus medication (NMB or saline), which maintained blinding of the bedside clinical staff
  • Local ethical approval sought
  • Requirement of informed consent at time of inclusion waived, however if there was a favourable outcome then the participant asked to provide written consent.
  • Calculated a necessary sample size of 26 patients per group to allow significant difference of 32% probability of at least one shivering episode with 80% power and two-sided type I error of 5%
    • Estimated baseline shivering episode not stated
    • ‘significant difference of 32%’ not clear if this is absolute difference or relative difference


  • Single site (patients admitted to Department of Emergency Medicine, Medical University of Vienna) study
  • November 2010 – September 2013


  • Inclusion Criteria:
    • Adult Patients > 18 years old
    • Resuscitated from non-traumatic out of hospital arrest (OOHA) within 6 hours of arrival to Emergency Department
    • Cardiac Arrest had to be presumed cardio-respiratory in origin and had to be eligible for TTM
    • Temperature had to be > 35˚C on admission
  • Exclusion criteria:
    • Known terminal illness, intoxication, ward of the state, prisoners, known or suspected pregnancy, allergic reaction to rocuronium, myasthenia gravis, and epileptic disease.
  • 782 screened
    • 717 not included (mostly did not meet inclusion / exclusion criteria, 15 excluded due to staff not available)
    • 65 randomised
    • 2 excluded (1 from intracranial haemorrhage, 1 unblinded due to ARDS)
    • 32 in continuous NMB group
    • 31 in bolus NMB group
  • Similar baseline characteristics in each group
    • Age: 62 (continuous NMB) vs. 58 (bolus NMB)
    • Male: 81% vs. 84%
    • BMI: 28 kg/m2 vs. 28kg/m2
    • Witnessed arrest: 91% vs. 90%
    • No BLS: 38% vs. 35%
    • VF: 81% vs. 90%
    • Admission after ROSC: 34 min vs. 34 min
    • Prehospital administration of NMB: 44% vs. 63%
    • pH on admission: 7.18 vs. 7.16
    • Lactate on admission: 7.3 mmol/L vs. 6.9
    • Glucose on admission: 256 mg/dL vs. 252
    • GCS on admission: 3 vs. 3
    • Core temperature on admission: 35.9 vs. 35.8
    • Pittsburgh Brain Stem Scale Score: 8 vs. 9


  • Continuous infusion of NMB drug
    • Rocuronium as initial bolus of 0.25mg/kg
    • Followed by an infusion of 0.25mg/kg/hr
    • If shivering was noticed then a bolus of (placebo) saline was given


  • On demand bolus of NMB drug
    • Initial bolus of (placebo) saline
    • Followed by a continuous infusion of (placebo) saline
    • If shivering was noticed then a 0.25mg/kg bolus of rocuronium was given

Management common to both groups

  • Sedation was intravenous fentanyl (2µg/kg/h) and midazolam (0.125mg/kg/hr)
  • Bolus of sedation allowed if clinically indicated
  • Severity of shivering was graded using “Shivering Assessment Scale” (SAS – 0 to 3 categorical scale)
    • If SAS greater or equal to 1 then a bolus of the intervention was given (either rocuronium or placebo depending on group allocation) and sedation increased as an equivalent of 5kg extra body weight
  • Study medication stopped at 29 hours
  • Sedation and analgesia ceased at 31 hours
  • Arterial oxygen saturation kept > 95%, PaCO2 kept between 35 and 45 mmHg, MAP > 60mmHg using crystalloids +/- vasopressors and serum blood glucose kept between 110 and 180 mg/dl
  • All patients underwent TTM with a cooling catheter into IVC via a femoral puncture
  • Cooling rate set to maximum until oesophageal temperature reached 33°C
  • Cooling period lasted 24 hours from initiation until start of re-warming, which was performed at a rate of 0.4°C/hour until a core temperature of 36°C was reached
  • Standardized intensive care monitoring used on all patients
  • Relaxation measured using TOF watch every three hours; this was performed by an individual not involved in the clinical management of the patient
  • Neurological assessment was done at 1,2,6 and 12 months during the follow up.
  • An independent neurologist evaluated incidence of critical illness polyneuropathy and myopathy at 12 months after cardiac arrest through a clinical exam and nerve conduction studies.


  • Primary Outcome: shivering episodes were less frequent in the continuous infusion group compared to the on demand bolus group
    • Continuous infusion: 25%
    • On demand bolus: 94%
    • Absolute risk reduction (ARR): 68.55% (95% CI 51.2% to 85.9%; P = 0.0001)
    • Number needed to treat (NNT): 2
    • Fragility Index (FI): 16
  • Secondary Outcomes:
    • Cumulative drug doses were different between groups
      • Continuous infusion group receiving more rocuronium
      • Continuous infusion group received less midazolam and fentanyl
    • Mortality was not different between groups
      • Continuous infusion: 47%
      • On demand bolus: 39%
      • Absolute risk increase (ARI): 8.1% (95% CI -16% to + 32%; P = 0.61)
    • Favourable neurological outcome at one year was not different between groups


Primary Outcome


Continuous NMB Bolus NMB P Value

 Shivering Episodes Detected

8/32 (25%)

29/31 (94%)

Median Shivering Episodes

0 (IQR 0 – 0.5)

8 (IQR 4 – 12)


ARR 68.5%, NNT 2 (95% CI 1.2 – 2)

Fragility Index 16

Secondary Outcomes


Continuous NMB

Bolus NMB

P Value

Cumulative Midazolam Dose

4.3 +/- 0.8 mg/kg

5.1 +/- 0.9 mg/kg

< 0.01

Cumulative Fentanyl Dose

0.062 +/- 0.014 mg/kg

0.071 +/- 0.007 mg/kg

< 0.01

Cumulative Rocuronium Dose

7.8 +/- 1.8 mg/kg

2.3 +/- 1.6 mg/kg

< 0.01

Median TOF Ratio

(i.e. TOF Count < 4)

(i.e. TOF Count = 4)


Time to waking

2 (IQR 2-3) days

4 (IQR 2 – 7.5) days


ICU Length of Stay

6 (IQR 3 – 5.9) days

10 (IQR 5 – 15) days



15/32 (47%)

12/31 (39%)


Hazard Ratio 1.34 (95% CI 0.63 – 2.87), ARI 8.2% (95% CI -16.2 – 32.5%)

Favourable Neurological Function at 12 months

17/32 (53%) 17/31 (55%)


Risk Ratio 0.97 (95%CI 0.61 – 1.53), ARI 1.7% (95% CI -22.9 – 26.3%)

Authors’ Conclusions

  • The use of a continuous neuromuscular blockade in patients undergoing TTM following an OOHA when compared to on demand bolus neuromuscular blockade reduced shivering, time until awakening and length of intensive care stay without significant harm.


  • Pragmatic design to answer a clinical question with limited evidence base
    • Theory concerning risks of shivering such as increased metabolic demand must be balanced against potential for occurrence of critical illness myopathy with use of NMB
    • Appropriate RCT methodology
  • Good randomisation and concealment strategy, which reduces change of systematic bias
  • Blinding of bedside clinicians reduces the chance of performance bias
    • See note below regarding possible unintentional unblinding
  • Balanced baseline characteristics results improves internal validity
  • Recruited more patients than calculated sample size to allow for potential missing data such as early mortality.
  • Attempted to collect patient-centred data for secondary and tertiary outcomes


  • Single centre with a small cohort of patients
  • Basis for expected difference in power calculation not explained
  • Not analysed on an intention to treat basis
    • Two patients excluded following randomization
    • However fragility index of 16 suggests the loss of these two patients does not weaken the conclusion
  • The primary outcome, shivering assessment, was subjective and there is the risk of inter-observer variation
    • The impact of this is reduced by adequate blinding of observers
    • The Shivering Assessment Scale has been validated and demonstrated excellent interrater reliability
  • The primary outcome measure – occurrence of shivering – was also the trigger for administering NMB drugs in the control group
    • It is therefore of no surprise that, if we wait for the primary outcome to occur before instigating a treatment, this treatment is associated with the primary outcome occurring more frequently!
    • The primary outcome is of limited clinical relevance and of no relevance to the patient
    • Important outcomes were only planned as secondary outcomes and therefore not powered for any conclusion
  • Midazolam based sedation regimes are not used in some centres due to concerns around delayed waking; this will affect the external validity of the trial.
  • Although initially double blinded (medications drawn up by nurse not involved in patients care) the clinical effects of saline or rocuronium may become apparent following administration resulting in an unintentional unblinding
  • Significant numbers of patients were lost to final follow up – only 14% (9/63) had a neurological examination at one year
    • Therefore the incidence of polyneuropathy can not be assessed and no conclusions can be drawn surrounding the safety of continuous NMB infusion
    • Given this and small sample sizes further work surrounding this should be carried to assess risk-benefit profile of continuous NMB.
  • There was no assessment for awareness with those having a continuous NMB infusion
  • Depending on results of TTM2 this question may not be as relevant in the future

The Bottom Line

  • When cooling patients after cardiac arrest, if you wait until a patient shivers before administering a neuromuscular blocking drug rather than commencing an infusion from the time of admission, then they will shiver more
  • This trial may indirectly show that by stopping people from shivering, then less sedation is required leading to earlier awakening and potential earlier ICU discharge, however these secondary outcomes should be considered hypothesis generating and further research is required

External Links


Summary author: George Walker
Summary date: 4 May 2018
Peer-review editor: Segun Olusanya

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.