Elsevier

Resuscitation

Volume 83, Issue 4, April 2012, Pages 417-422
Resuscitation

The effect of hyperoxia following cardiac arrest – A systematic review and meta-analysis of animal trials

https://doi.org/10.1016/j.resuscitation.2011.12.021Get rights and content

Abstract

Aim

There are conflicting findings from observational studies regarding the nature of the association between hyperoxia and risk of mortality in patients admitted to intensive care following cardiac arrest. This systematic review and meta-analysis evaluates animal data investigating the effect of administration of high concentrations of oxygen following cardiac arrest on neurological outcome and the clinical applicability of this data.

Methods

A systematic search of Medline and Embase identified controlled animal studies modelling cardiac arrest with subsequent cardiopulmonary resuscitation that compared ventilation with 100% oxygen to lower concentrations following return of spontaneous circulation. Eligible studies were included in a meta-analysis in which the inverse variance weighted differences were calculated for the standardised mean difference of the primary outcome measure, the neurological deficit score.

Results

Ten studies met the criteria for inclusion in the systematic review. In a meta-analysis of six studies, with 95 animals, treatment with 100% oxygen resulted in a significantly worse neurological deficit score than oxygen administered at lower concentrations, with a standardised mean difference of −0.64 (95% CI −1.06 to −0.22). In four of five studies, histological evidence of increased neuronal damage was present in animals that received 100% oxygen therapy.

Conclusions

The administration of 100% oxygen therapy is associated with worse neurological outcome than lower oxygen concentrations in animal models of cardiac arrest. However, due to limitations in study design and poor generalisability of the animal models to the situation of post cardiac arrest resuscitation in humans, the clinical applicability of this data is uncertain.

Introduction

For over 40 years guidelines have recommended the provision of high concentration oxygen therapy in the setting of cardiac arrest.1, 2 However, there has been recent debate over this practice, with concerns that inadvertent hyperoxia may potentiate brain injury and worsen myocardial dysfunction through adverse effects on the systemic ischaemia/reperfusion response.3, 4, 5, 6, 7 Despite these theoretical concerns, there has only been one small randomised controlled trial of oxygen administration in the setting of adult cardiac arrest.7 This study compared the administration by ambulance staff of an inspired oxygen concentration of 30% versus 100% following return of spontaneous circulation (ROSC) in patients with out-of-hospital ventricular fibrillation cardiac arrest. The study lacked the statistical power to detect clinically important differences in survival or functional outcomes. However, in a secondary analysis the trial found significant increases in neuron-specific enolase, a surrogate serum marker of neurological damage, in a subset of patients receiving 100% oxygen therapy who did not receive therapeutic hypothermia.

Recently, two groups have performed retrospective cohort studies investigating the association between mortality and arterial hyperoxia measured by arterial blood gas analysis (ABG) in intensive care unit (ICU) patients. The first study reported an association between hyperoxia (determined by the first ABG measurement after ICU admission) and in-hospital mortality, compared to patients with either normoxia or hypoxia on primary analysis.4 Secondary analysis demonstrated a linear relationship between arterial oxygen tension (PaO2) and mortality.5 However, a larger study in which hyperoxia, determined by the worst ABG in the first 24 h after ICU admission, did not find robust or consistently reproducible associations with mortality after adjustment for illness severity and inspired oxygen concentration.6

In light of the limited and conflicting data available from human studies, this systematic review and meta-analysis was undertaken to evaluate the pre-clinical data available from animal models of adult cardiac arrest that compared administration of 100% oxygen therapy to lower concentrations following ROSC.

Section snippets

Search strategy

Controlled trials investigating the effects of different fractions of inspired oxygen in animal models of cardiac arrest with cardiopulmonary resuscitation (CPR) were identified by a search of the following databases (search date 15 February 2011): Medline 1950 to present with daily update, Medline pending, Old Medline and EMBASE. Key words used were: ‘cardiac arrest’, ‘heart arrest’, ‘cardiopulmonary resuscitation’, ‘hyperox$’, ‘normox$’, ‘oxygen concentration’, ‘oxygen ventilation’, ‘oxygen

Description of studies

The search identified 194 papers of which ten met the inclusion criteria (Fig. 1).10, 11, 12, 13, 14, 15, 16, 17, 18, 19 Table 1 demonstrates the characteristics of the studies included in this review including species, mechanism and duration of cardiac arrest, and details of the normoxic protocol. Therapeutic hypothermia was not used. In all studies animals were ventilated prior to cardiac arrest. Most of the studies reported administration of room air or normoxic PaO2 values at baseline prior

Statement of key findings

This systematic review and meta-analysis has shown that animals resuscitated after cardiac arrest with 100% oxygen for 60 min following CPR had a significantly worse NDS than those treated with lower concentrations of oxygen. In keeping with this, continuous 100% oxygen regimens increased neuronal damage based on both histology and indirect markers of cerebral metabolic function.

Significance of findings

We have shown that hyperoxic regimes are potentially harmful in animal models of cardiac arrest and resuscitation.

Conclusions

This study demonstrates the potential for 100% oxygen administration following cardiac arrest in animal models to cause neurological harm when compared to lower levels of oxygen therapy. While neurological injury is a major cause of morbidity and mortality in humans after cardiac arrest, there are major limitations preventing the generalisability of this animal data to the clinical setting. The conflicting data reported in recent observational studies of ICU patients in the post cardiac arrest

Conflicts of interest

No conflicts of interest.

Funding

The Medical Research Institute of New Zealand has previously received funding for other projects by the Health Research Council of New Zealand, AstraZeneca, GlaxoSmithKline, Novartis and Chiesi.

References (21)

There are more references available in the full text version of this article.

Cited by (0)

A Spanish translated version of the summary of this article appears as Appendix in the final online version at doi:10.1016/j.resuscitation.2011.12.021.

View full text