Oxygenation targets in post-resuscitation care: a perspective of the BOX randomized clinical trial

The estimated annual incidence of treated out-of-hospital cardiac arrest (OHCA) ranges from 30.0 to 97.1 individuals per 100 000 population.1 Patient outcomes after OHCA vary substantially by region but are generally poor with hypoxic-ischaemic brain injury as the leading cause of death and disability. Maintenance of adequate brain perfusion pressure and oxygenation may be therapeutic targets to minimize brain ischaemia and related pathophysiological sequalae. Although the optimal PaO2/SaO2 values in OHCA patients remain undefined, prior studies suggest that hypo/hyperoxaemia (PaO2 < 60 and >300 mmHg, respectively) are associated with higher mortality.2,3 A small randomized pilot study of comatose patients admitted to intensive care after OHCA demonstrated no survival and NSE difference between patients targeting PaO2 10–15 kPa to moderate hyperoxia (PaO2 20–25 kPa).4 A subgroup analysis of an OHCA cohort in a larger trial comparing conservative (upper SpO2 alarm level 97%–lower SpO2 limit 90%) vs. usual oxygen therapy (no upper limits–lower SpO2 limit 90%) in mechanically ventilated critically ill patients, found an association with benefit in 90-day survival in patients in whom hyperoxia was avoided (relative risk, 1.39; 95% confidence interval, 1.01–1.92).5 The latest American Heart Association and European guidelines suggest (low quality of evidence) a target of SpO2/SaO2 between 92–98% and 10–13 kPa (75–100 mmHg).6,7 BOX was a multicentre, interventional, randomized clinical trial with 2 × 2 factorial design that allocated comatose OHCA patients in a 1:1 ratio to either a restrictive oxygen target of a partial pressure of arterial oxygen (PaO2) of 9–10 kPa (68–75 mmHg) or a liberal oxygen target of a PaO2 of 13–14 kPa (98–105 mmHg).8 If the patient was unable to achieve the allocated oxygen target by manipulating FiO2 and positive end-expiratory pressure or by supine ventilation or bronchiolar lavage, the highest possible values were considered acceptable. Adjuvant extracorporeal membrane oxygenation was at discretion of the treating physician. The main hypothesis for the oxygen therapy strategy was that restrictive vs. liberal would reduce mortality by reducing the risk and degree of hypoxic brain injury in comatose survivors of OHCA. The primary combined endpoint was death from any cause or discharge from hospital with a poor neurological status [cerebral performance Category (CPC) 3 or 4], whichever occurs first within 90 days after randomization. Patients (age ≥18 years) with OHCA of presumed cardiac cause and Glasgow coma scale <8 (not able to obey verbal commands) admitted to hospital after resuscitated cardiac arrest with sustained spontaneous circulation (ROSC) were eligible and randomized mostly in intensive care units (ICUs). Major exclusion criteria included: >4 h delay from ROSC to screening, unwitnessed asystole, refractory hypotension (despite medical therapy and/or intra-aortic balloon pump/axial flow device), known disease making 180 days survival unlikely, pre-arrest CPC 3 or 4 and/or suspected cerebrovascular acute events. All patients underwent temperature control at 36°C with sedation and mechanical ventilation for at least 24 h. The initial FiO2 was to be set at 0.3 in the restrictive-target group and at 0.6 in the liberal-target group and titrated if arterial blood saturation fell below 93% on peripheral pulse oximetry. A total of 789 patients were included in the intention-to-treat population (394 in the restrictive-target group and 395 in the liberal-target group). On arrival in the ICU, the two groups had similar values of PaO2 (16.1 ± 8.5–122 ± 6.7 vs. 17.1 ± 8.8–128.2 ± 66 mmHg) and fraction of inspired oxygen (FiO2), with a substantial separation of values occurring the second hour. Additionally, the percentage of patients with positive end expiratory pressure <10 cm H2O and with FiO2 < 30% was higher in the restrictive group by the second hour after randomization. Respectively, only 2 and 1 patients in the restrictive group and liberal group underwent extracorporeal membrane oxygenation during the first 36 h. At 90 days, a primary-outcome event had occurred in 126 of 394 patients (32.0% of whom 113 died) in the restrictive-target group and in 134 of 395 patients (33.9% of whom 113 died) in the liberal-target group (adjusted hazard ratio, 0.95; 95% confidence interval, 0.75–1.21; P = 0.69). No differences were found between the two groups for any of the neurocognitive impairment scores (CPC; Rankin scale and Montreal Cognitive Assessment) or neuron-specific enolase level, or pre-specified adverse events (including infection, bleeding, and seizures). These results indicate that there is no difference in mortality or neruological consequences of cardiac arrest by maintaining a restrictive oxygen strategy during the first 48 h of ICU admission when compared with PaO2 up to 105 mmHg. These findings are consistent with a recent systematic review and meta-analysis including 7 trials and 10 observational studies in the pre- and out-of-hospital setting9 and with a trial of patients admitted to ICU with acute respiratory failure, where low rather than higher targeted PaO2 (60 vs. 90 mmHg, respectively) was compared.10 The PaO2 levels before randomization were higher than the cut-off allowed in both arms reaching the desirable target within the first 2 h since randomization. Since the hyperoxaemia has demonstrated to be detrimental for survival, this was not tested by the design of the BOX trial, as the upper limit of PaO2 in the liberal arm (105 mmHg) was close to the cut-off value considered as normal (100 mmHg). Additionally, we note that this was a population with cardiac arrest that had a high proportion of shockable rhythms and rapid initiation of CPR and restoration of spontaneous circulation (mean 21 min in both groups). Moreover, the trial was not well powered to detect clinically meaningful differences <25% relative effect on the primary endpoint. Nevertheless, the BOX trial substantially advances the evidence gathered so far to address oxygen targets in post-cardiac arrest management and together with prior data indicates that maintaining PaO2 levels within the normal range or mildly hypoxemic is a reasonable strategy in patients who experienced cardiovascular arrest and lends support to current guidelines.6,7 Presenter/study group: J. Moller

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.