Association between continuous hyperosmolar therapy and survival in patients with traumatic brain injury- a multicentre prospective cohort study and systematic review

Asehnoune K, Lasocki S, Seguin P et al. Critical Care, 2017; 21: 320. DOI 10.1186/s13054-017-1918-4
Background Traumatic brain injury (TBI) in people is associated with a high in-hospital mortality (33%) and poor neurological recovery (33%). Hyperosmolar agents are used to treat increased intracranial hypertension (ICH); whilst boluses of hyperosmolar agents have a transient effect on decreasing the intracranial pressure, the effects of continuous hyperosmolar therapy (CHT) on outcomes and survival have not been fully elucidated.
Rational and Objective

Given findings of previous prospective and retrospective studies have given contradicting results on association of CHT on outcomes in TBI, the authors investigated the effects of early CHT in patients with TBI on:

  • Mortality (primary objective)
  • Long-term outcome (secondary objective)
Trial Design Prospective multicenter cohort study performed using pooled individual patient data included in three previous different prospective trials (CORTI-TC, BI-VILI and ATLAN-REA studies). The authors also performed a systematic review of the existing literature, including the result of this trial.
Patient population and location A multicentre study  of people aged between 15-75 years with a duration of mechanical ventilation >24 hours and moderate/severe TBI (defined as GCS £12).   
Methods Patients receiving early CHT were compared to patients receiving boluses of hyperosmolar agents.   In the control group, a bolus of hyperosmolar therapy (mannitol 0.25-1 g/kg or hypertonic saline 250 mOsm dose) was used as first line treatment and repeated in case of recurrence of ICH as required.  In the study group (at one single center), early continuous hypertonic saline therapy (1 hour bolus of 20% NaCl) was initiated as first line treatment followed by 24 hours or more infusion of hypertonic saline titrated to target natremia according to the evolution of ICP.   Systematic review was performed following MOOSE guidelines.
Endpoints (if applicable)

Primary endpoint: risk of survival at 90 days

Secondary endpoint: dichotomized Glasgow Outcome Scale (GOS) at 90 days (GOS 1-3 vs 4-5)

The outcomes selected for the systematic review were:

  • Number of in-ICU deaths
  • Rate of ICH
  • Rate of hypernatremia (Na >160 mmol/L)
Results
  • 1086 patients were included of which 26% received CHT.
  • CHT was continued for a median duration of 5 (3-8) days.
  • Patients treated with CHT required less frequently application of hypocapnia or decompressive craniectomy.
  • Patients treated with CHT had a higher survival at 90 days (74.1%) compared to the control group (65.9%) (p = 0.001).
  • Following propensity score analysis, the adjusted hazard ratio for 90 days survival for CHT was 1.74 (C.I 1.36-2.23, P<0.001).
  • No significant difference was detected on favourable outcomes based on GOS at 90 days between the groups.
  • Sodium concentrations were higher in patients treated with CHT with severe hypernatremia (>160 mmol/L) being more frequent in patients with CHT (9.1%) compared to patients treated with conventional hyperosmolar therapy (2.2%) (p< 0.001).
  • Following the systematic review, a lower mortality was associated with treatment with CHT vs control (23.6% vs 31.2%, OR 1.42, C.I 1.04-1.95, p=0.03).
Limitations

This study only assesses associations but not causality in relation to reduced mortality.

Early CHT was employed in one single center compared to standard treatment performed in other centers, therefore center-effects on outcomes cannot be excluded.

Clinical Relevance Conclusions

In this trial and following a systematic review, early use of CHT was associated with survival at 90 days and reduced mortality respectively, suggesting it could be used in people with TBI at high risk of or as a first line treatment of ICH.

Risk of severe hypernatremia exists, and close monitoring should be performed to avoid adverse events.