Oxygen targets following cardiac arrest: A meta-analysis of randomized controlled trials

Introduction The appropriate oxygen target post-resuscitation in out-of-hospital cardiac arrest (OHCA) patients is uncertain. We sought to compare lower versus higher oxygen targets in patients following OHCA. Methods We searched MEDLINE, Embase, the Cochrane Library, and ClinicalTrials.gov until January 2023 to include all randomized controlled trials (RCTs) that evaluated conservative vs. liberal oxygen therapy in OHCA patients. Our primary outcome was all-cause mortality at 90 days while our secondary outcomes were the level of neuron-specific enolase (NSE) at 48 h, ICU length of stay (LOS), and favorable neurological outcome (the proportion of patients with Cerebral Performance Category scores of 1–2 at end of follow-up). We used RevMan 5.4 to pool risk ratios (RRs) and mean differences (MDs). Results Nine trials with 1971 patients were included in our review. There was no significant difference between the conservative and liberal oxygen target groups regarding the rate of all-cause mortality (RR 0.95, 95% CI: 0.80 to 1.13; I2 = 55%). There were no significant differences between the two groups when assessing favorable neurological outcome (RR 1.01, 95% CI: 0.92 to 1.10; I2 = 4%), NSE at 48 h (MD 0.04, 95% CI: −0.67 to 0.76; I2 = 0%), and ICU length of stay (MD −2.86 days, 95% CI: −8.00 to 2.29 days; I2 = 0%). Conclusions Conservative oxygen therapy did not decrease mortality, improve neurologic recovery, or decrease ICU LOS as compared to a liberal oxygen regimen. Future large-scale RCTs comparing homogenous oxygen targets are needed to confirm these findings.


Introduction
Cardiac arrest often results in hypoxic-ischemic brain injury with less than 10% of patients achieving meaningful neurologic recovery [1]. Comatose patients following out-of-hospital cardiac arrest (OHCA) frequently require mechanical ventilation with supplemental oxygen. However, there is equipoise regarding the optimal oxygen target in such patients. Studies have shown that hyperoxia might cause exacerbation of neurological injury by producing excessive oxygen free radicals [2][3].
Moreover, liberal oxygenation has also been associated with an increased risk of ischemic encephalopathy and death [4]. On the other hand, conservative oxygen therapy may be further detrimental to already hypoxic tissue. Therefore, determining an appropriate oxygen target post-resuscitation in OHCA patients is of paramount importance. Recently, there has been an increased focus on different suggested oxygen targets in randomized clinical trials (RCTs), with the largest of these trials published recently [5]. Hence, the present paper sought to review the current literature and conduct a contemporary meta-analysis to compare lower versus higher oxygen targets in patients following OHCA.

Methods
We registered our protocol with PROSPERO (CRD42022383931) and conducted this meta-analysis following the guidance presented in the Cochrane Handbook for Systematic Reviews of Intervention. We performed a systematic search on MEDLINE (Ovid), the Cochrane Library, Embase, and ClinicalTrials.gov from inception to January 2023 to retrieve relevant studies. Additionally, we manually searched the reference lists of relevant studies. The detailed search strategy is given in Supplementary Table 1. The screening process was carried out independently by two authors based on the following inclusion criteria: (1) population: adults with OHCA; (2) interventions: conservative oxygen therapy versus liberal oxygen therapy as defined by the individual RCTs; and 3) type of study: RCTs only. We defined all-cause mortality at 90 days as our primary outcome. If a study did not report 90-day mortality, we used the endpoint closest to it. Our secondary outcomes were: (1) the level of neuron-specific enolase (NSE) at 48 h; (2) ICU length of stay (LOS); and (3) favorable neurological outcome: the proportion of patients with Cerebral Performance Category (CPC) scores of 1-2 at end of follow-up.
We assessed the risk of bias for each trial using the Cochrane Risk of Bias version 2 (RoB 2.0) tool. We used a random-effects model to pool risk ratios (RRs) for dichotomous outcomes and mean differences (MDs) for continuous outcomes with their corresponding 95% confidence intervals (CIs). We chose the random-effects model because we anticipated our included studies to be considerably heterogeneous. The I 2 statistic was used to evaluate heterogeneity. All statistical analyses were performed using R software version 4.1.0 (the meta package) [6]. We were unable to assess publication bias as the number of included studies was less than 10.

Results
Nine trials with 1971 patients were included in our review [2][3]5,[7][8][9][10][11][12]. The detailed selection process is depicted in a PRISMA flowchart (Supplementary Fig. 1) The detailed characteristics of each trial are presented in Table 1. All trials were ascertained to have some concerns of bias except two which were rated to be at low risk of bias (Supplementary Figure 2) [5,12].

Discussion
To the best of our knowledge, this is the largest meta-analysis on this topic to date. In this study, we observed no difference between conservative oxygen therapy and liberal oxygen therapy regarding mortality or any other studied clinical outcome.
The idea that proposed significant adverse effects of high oxygen therapy was first derived from experimental studies [13]. A metaanalysis on animal trials by Pilcher et al. showed worse neurological outcomes after the administration of 100% oxygen compared to the group which received restrictive oxygen therapy [13]. These findings were further corroborated in humans by observational studies [14]. However, evidence from RCTs is needed to make any explicit recommendations for clinical practice.
Until 2020, several small RCTs had been performed. An individuallevel patient data meta-analysis of RCTs published by Young et al. showed a significant reduction in mortality with conservative oxygen therapy after pooling data from 429 patients [15]. This meta-analysis did not include the data from HOT-ICU [9], EXACT [12], and the largest trial to date on this topic which has been recently published by the investigators of the BOX trial [5]. In our meta-analysis, after pooling the results from these trials, we found no benefit of conservative oxygen therapy in cardiac arrest patients. Our results suggest that the focus needs to be shifted to therapeutic interventions other than different oxygen, blood pressure, and carbon dioxide targets as these measures have shown no benefits [16,17].
Our study has several strengths. This is the largest meta-analysis to date including the data of 1971 patients from 9 RCTs. Moreover, our outcomes had low heterogeneity. The main limitation of our study was the variability of the study design, follow-up periods, and outcome definitions of the included trials.
In conclusion, conservative oxygen therapy did not decrease mortality, improve neurologic recovery, or decrease ICU LOS as compared to a liberal oxygen regimen. Future large-scale RCTs comparing homogenous oxygen targets are needed to confirm these findings.
Statements and Declarations. Financial support. No financial support was received for this study.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.