ALS 3305 Use of supplemental oxygen during cardiopulmonary resuscitation: TF SR

CoSTR Citation

Skrifvars MB, Ohshimo S, Grunau B, Crow C, Scquizzato T, D´Arrigo S, Jakobsen J, Kamp C, Silassen C, Niemelä V, Soar J on behalf of the International Liaison Committee on Resuscitation Basic Life Support and Advanced Life Support Tasks Forces. Oxygen dose during cardiopulmonary resuscitation. Advanced Life Support Task Forces, 2026. Available from: http://ilcor.org

Methodological Preamble and Link to Published Systematic Review

The previous update of this topic was conducted as an evidence update in 2019 (ALS889). For the current COSTR a PICOST template was conceived and approved by the ALS Task Force in August 2025. A search strategy was developed, and a search was conducted in October-November 2025 and completed on November 17^th 2025.

Systematic Review

Skrifvars MB, Ohshimo S, Grunau B, Crow C, Scquizzato T, D´Arragio S, Jakobsen J, Kamp C, Silassen C, Niemelä V, Soar J on behalf of the International Liaison Committee on Resuscitation Basic Life Support and Advanced Life Support Tasks Forces. Oxygen dose during cardiopulmonary resuscitation. Advanced Life Support Task Forces, 2026. Submission pending.

PICOST

Submitted for PROSPERO registration on November 18^th 2025.

Article Selection: ALS 3305 Figure Article selection;

ROB: ALS 3305 Table Ro B Robins I

Table: ALS 3305 Table Included studies

Consensus on Science

The initial search identified 38,528 studies out of which 28,428 were screened by title and abstract, 47 were retrieved for assessment of eligibility, and six were included. Characteristics of the six included studies are shown in Table 1. The Risk Of Bias In Non-Randomized Studies - of Interventions (ROBINS-I) tool was used to assess of bias in the included studies.

The search identified no studies that compared different doses of administered oxygen during cardiopulmonary resuscitation (CPR). The identified six studies were observational, either retrospective or prospective, included an arterial blood gas (ABG) measurement obtained during CPR and compared the oxygen levels in patients with good or poor outcome. All patients in these trials were treated, as recommended in the Guidelines with the highest possible fraction of oxygen. All studies were judged to be at critical risk of bias according to the ROBINS-I tool.

These six studies can be briefly summarized as follows. Since all had critical risk of bias and were observational, formal GRADE evaluation was not completed because these factors alone already result in very low-certainty evidence.

Spindelboeck 2013 (critical risk of bias): A study of arterial blood gas (ABG) measurements were compared in 145 out-of-hospital cardiac arrest (OHCA) patients undergoing CPR. All patients received a 100% fraction of inspired oxygen (FiO₂). The levels of oxygen in the arterial blood gas samples were divided into hypoxia (≤60 mmHg [8 kPa]), normoxia (61–300 mmHg [8.1–40 kPa] or hyperoxia (>300 mmHg [>40 kPa]). The proportions of patients surviving to hospital admission were 19%, 51% and 83%, respectively (p<0.001). There was no statistically significant difference in the proportion of patients surviving with good functional outcome (p=0.06). In a multivariable logistic regression model a higher partial pressure of oxygen in arterial blood (PaO₂₎ was associated with a higher likelihood of survival to hospital admission.

Spindelboeck 2016 (critical risk of bias): In a prospective study 83 ABGs were obtained during ongoing CPR. Arterial oxygen was higher in patients surviving to hospital admission (median 85 mmHg compared to 66 mmHg, p=0.05). In a multivariable model, the PaO₂ was not associated with hospital admission.

Patel 2018 (critical risk of bias): A retrospective cohort study of 167 in-hospital cardiac arrest patients in whom an ABG was obtained. All patients received 100% FiO₂ either with bag or mechanical ventilation. Intra-arest oxygen levels were categorized as PO₂<60 mmHg (n=38), PaO₂ of 60-92 mm Hg (n=44), PaO2 of 93 to 159 mm Hg (n=43), PaO₂ of 160 to 299 mm Hg (n=24), and PaO₂ 300 mm Hg (n=18). Higher intra-arrest PaO₂ levels had progressively higher rates of ROSC (58% vs 71% vs 72% vs 79% vs 100%, P=0.021) and survival to discharge (16% vs 23% vs 30% vs 33% vs 56%, P = .031). In multivariate analysis, a PaO₂ of higher than 300 mm Hg was independently associated with higher survival to discharge (odds ratio 60.68; 95% confidence interval: 3.04-1210.28; P=0.007; reference PaO2 < 60 mm Hg).

Hong 2021 (critical risk of bias): A prospective study of 217 patients undergoing CPR in the emergency department in whom an arterial line was placed during CPR. This study focused on the 80 patients who had more than one ABG measured during CPR. Out of these 80 patients, 13 achieved sustained ROSC. The difference between the second and the first ABG was higher in the patients with return of spontaneous circulation (ROSC) (median 90.2 [51.2–122.7] mmHg compared to median 16.5 [−0.4 to 42.2] mmHg, p<0.001). On multivariable logistic analysis the ΔPaO₂ (odds ratio [OR] = 1.023; 95% confidence interval [CI] = 1.004–1.043, p = 0.020) was a significant predictor of sustained ROSC.

Izawa 2022 (critical risk of bias): An observational registry study including 16,013 patients undergoing CPR on hospital admission after OHCA who had an arterial blood gas obtained. All patients received a FiO₂ of 100%. The intra-arrest PaO₂ level, was divided into three categories: hypoxaemia, PaO₂ < 60 mmHg; normoxaemia, 60–300; or hyperoxaemia, 300. The primary outcome was favourable functional survival at one month or at hospital discharge defined as a cerebral performance category of 1 or 2. The proportion of favourable functional survival increased was 0.5 % (57/11,484) in hypoxaemia, 1.1 % (48/4243) in normoxaemia, and 5.2 % (15/286) in hyperoxaemia (p-value for trend < 0.001). Higher PaO₂ categories were associated with favorable functional survival in a multivariable model: adjusted OR was 2.09 (95 % CI: 1.39–3.14) for normoxaemia and 5.04 (95 % CI: 2.62–9.70) for hyperoxaemia when compared to hypoxaemia as a reference.

Nelskylä 2022 (critical risk of bias): A prospective observational study including 75 patients with OHCA. An attempt was made to insert an arterial line and to measure the PaO₂. Out of 75 patients the PaO₂ was measured in 46 patients. Among these patients the PaO₂ was higher in patients who achieved ROSC compared to those who did not (9.10 kPa, IQR 6.20–11.50 vs. 6.90 kPa, 4.20–8.90, p = 0.02).

Taken together these studies suggest an association between higher measured arterial oxygen levels during CPR and improved outcome including likelihood of ROSC, survival and functional outcome. However, all patients received 100% oxygen and no active intervention was compared in these studies. Some studies compared the association of the PaO2 value after adjustment for other potential confounding factors while other studies only reported unadjusted values or raw study data.

Treatment Recommendations

• We suggest using the highest possible inspired oxygen concentration during CPR (weak recommendation, very-low-certainty of evidence).

Justification and Evidence to Decision Framework Highlights

• This topic was prioritized by the ALS Task Force given that the last review of oxygen use during CPR is from 2020. The 2020 Treatment Recommendation was based three observational studies.

• In making this recommendation we have considered the lack of evidence comparing different fractions of FiO₂ during CPR.

• We found only indirect data that suggest that a higher PaO₂ is associated with better outcomes including higher likelihood of ROSC, survival to hospital admission after OHCA, survival to hospital discharge and functional outcome. Importantly in all these studies patients received 100% FiO₂. Thus, the difference in the PaO₂ is not related to a different administration strategy but due to patient factors including pulmonary function, i.e. factors not open for interventions. Despite this, and the lack of contradictory findings and the high likelihood of intra-arrest hypoxia in patients undergoing CPR supports to the continued recommendation to give the highest feasible amount of oxygen during CPR.
• The recommendation for the highest possible inspired oxygen concentration refers to the highest concentration that can be delivered by the existing approach to ventilation. This recommendation is not intended to specify a specific approach to ventilation during CPR.

• No study showed worse outcome with higher PaO₂ values during CPR.

Knowledge Gaps

• There are no RCTs comparing different fractions of oxygen during CPR and its effect on ROSC, survival and functional outcome.
• There are limited data on the optimal oxygen targets during and immediately after cardiac arrest.
• A high oxygen level during CPR is likely to result in high levels of oxygen immediately after ROSC and whether this worsens the developing reperfusion injury is currently unknown.
• There is limited data on means to monitor oxygenation during CPR. The available evidence does not suggest the value of PaO₂ monitoring with ABG samples during CPR.
• It is unclear if other means to ventilate the patient during CPR will influence oxygenation during CPR. Possibilities may include different compression to ventilation ratios, the use of PEEP and the use of standard and novel means of mechanical ventilation such chest compression synchronized ventilation

References

1. Spindelboeck W, Schindler O, Moser A, Hausler F, Wallner S, Strasser C, Haas J, Gemes G, Prause G. Increasing arterial oxygen partial pressure during cardiopulmonary resuscitation is associated with improved rates of hospital admission. Resuscitation. 2013 Jun;84(6):770-5.

2. Spindelboeck W, Gemes G, Strasser C, Toescher K, Kores B, Metnitz P, Haas J, Prause G. Arterial blood gases during and their dynamic changes after cardiopulmonary resuscitation: A prospective clinical study. Resuscitation. 2016 Sep;106:24-9.

3. Patel JK, Schoenfeld E, Parikh PB, Parnia S. Association of Arterial Oxygen Tension During In-Hospital Cardiac Arrest With Return of Spontaneous Circulation and Survival. J Intensive Care Med. 2018 Jul;33(7):407-414.

4. Hong SI, Kim JS, Kim YJ, Kim WY. Dynamic changes in arterial blood gas during cardiopulmonary resuscitation in out-of-hospital cardiac arrest. Sci Rep. 2021 Nov 30;11(1):23165.

5. Izawa J, Komukai S, Nishioka N, Kiguchi T, Kitamura T, Iwami T. Outcomes associated with intra-arrest hyperoxaemia in out-of-hospital cardiac arrest: A registry-based cohort study. Resuscitation. 2022 Dec;181:173-181.

6. Nelskylä A, Skrifvars MB, Ångerman S, Nurmi J. Incidence of hyperoxia and factors associated with cerebral oxygenation during cardiopulmonary resuscitation. Resuscitation. 2022 Jan;170:276-282.