Association of Pediatric Postcardiac Arrest Ventilation and Oxygenation with Survival Outcomes

Abstract

Rationale: Adult and pediatric studies provide conflicting data regarding whether post-cardiac arrest hypoxemia, hyperoxemia, hypercapnia, and/or hypocapnia are associated with worse outcomes. Objectives: We sought to determine whether postarrest hypoxemia or postarrest hyperoxemia is associated with lower rates of survival to hospital discharge, compared with postarrest normoxemia, and whether postarrest hypocapnia or hypercapnia is associated with lower rates of survival, compared with postarrest normocapnia. Methods: An embedded prospective observational study during a multicenter interventional cardiopulmonary resuscitation trial was conducted from 2016 to 2021. Patients ⩽18 years old and with a corrected gestational age of ≥37 weeks who received chest compressions for cardiac arrest in one of the 18 intensive care units were included. Exposures during the first 24 hours postarrest were hypoxemia, hyperoxemia, or normoxemia-defined as lowest arterial oxygen tension/pressure (Pa_O2) <60 mm Hg, highest Pa_O2 ⩾200 mm Hg, or every Pa_O2 60-199 mm Hg, respectively-and hypocapnia, hypercapnia, or normocapnia, defined as lowest arterial carbon dioxide tension/pressure (Pa_CO2) <30 mm Hg, highest Pa_CO2 ⩾50 mm Hg, or every Pa_CO2 30-49 mm Hg, respectively. Associations of oxygenation and carbon dioxide group with survival to hospital discharge were assessed using Poisson regression with robust error estimates. Results: The hypoxemia group was less likely to survive to hospital discharge, compared with the normoxemia group (adjusted relative risk [aRR] = 0.71; 95% confidence interval [CI] = 0.58-0.87), whereas survival in the hyperoxemia group did not differ from that in the normoxemia group (aRR = 1.0; 95% CI = 0.87-1.15). The hypercapnia group was less likely to survive to hospital discharge, compared with the normocapnia group (aRR = 0.74; 95% CI = 0.64-0.84), whereas survival in the hypocapnia group did not differ from that in the normocapnia group (aRR = 0.91; 95% CI = 0.74-1.12). Conclusions: Postarrest hypoxemia and hypercapnia were each associated with lower rates of survival to hospital discharge.

Keywords: Hypoxia; child; hypercapnia; hyperoxia; hypocapnia.

Figure 1.

CONSORT diagram. *Number of enrolled index in-hospital cardiac arrests (IHCAs) excluding patients with anomalous pulmonary venous return, aortic stenosis, double-outlet right ventricle, Ebstein’s anomaly, hypoplastic left heart syndrome, interrupted aortic arch, pulmonary atresia, pulmonary valve or artery stenosis, single ventricle that is not hypoplastic left heart syndrome, tetralogy of Fallot, transposition of the great vessels, tricuspid stenosis/atresia, truncus arteriosus, and other cyanotic heart disease. ECMO = extracorporeal membrane oxygenator; ROSC = return of spontaneous circulation.

Figure 2.

Patient and event characteristics by oxygenation group. CICU = cardiac intensive care unit; CPR = cardiopulmonary resuscitation; FSS = functional status scale; PICU = pediatric intensive care unit; PCPC = pediatric cerebral performance category; PRISM = pediatric risk of mortality score. *Oxygenation group: hypoxemia = lowest Pa_O2 <60 mm Hg, normoxemia = all Pa_O2 measurements, 60–199 mm Hg; hyperoxemia = highest Pa_O2 ≥200 mm Hg; hypoxemia and hyperoxemia = lowest Pa_O2 in the hypoxemic range and highest Pa_O2 in the hyperoxemic range. ^†Fisher’s exact test. ^‡Baseline PCPC and baseline FSS represent subject status prior to the event leading to hospitalization. ^§Kruskal-Wallis test. ^‖Weekday = between 7 AM and 11 PM Monday - Friday; weeknight = after 11 PM Monday - Thursday; Weekend = 11 PM Friday through 7 AM following Monday.

Figure 2.

Patient and event characteristics by oxygenation group. CICU = cardiac intensive care unit; CPR = cardiopulmonary resuscitation; FSS = functional status scale; PICU = pediatric intensive care unit; PCPC = pediatric cerebral performance category; PRISM = pediatric risk of mortality score. *Oxygenation group: hypoxemia = lowest Pa_O2 <60 mm Hg, normoxemia = all Pa_O2 measurements, 60–199 mm Hg; hyperoxemia = highest Pa_O2 ≥200 mm Hg; hypoxemia and hyperoxemia = lowest Pa_O2 in the hypoxemic range and highest Pa_O2 in the hyperoxemic range. ^†Fisher’s exact test. ^‡Baseline PCPC and baseline FSS represent subject status prior to the event leading to hospitalization. ^§Kruskal-Wallis test. ^‖Weekday = between 7 AM and 11 PM Monday - Friday; weeknight = after 11 PM Monday - Thursday; Weekend = 11 PM Friday through 7 AM following Monday.

Figure 3.

Patient and event characteristics by carbon dioxide group. CICU = cardiac intensive care unit; CPR = cardiopulmonary resuscitation; FSS = functional status scale; PCPC = pediatric cerebral performance category; PICU = pediatric intensive care unit; PRISM = pediatric risk of mortality score.*Carbon dioxide group: hypocapnia = lowest Pa_CO2 <30 mm Hg; normocapnia = all Pa_CO2 measurements 30–49 mm Hg; hypercapnia = highest Pa_CO2 ≥50 mm Hg; hypocapnia and hypercapnia = lowest Pa_CO2 in hypocapnia range and highest Pa_CO2 in hypercapnia range. ^†Fisher’s exact test. ^‡Baseline PCPC and baseline FSS represent subject status prior to the event leading to hospitalization. ^§Kruskal-Wallis test. ^‖Weekday = between 7 AM and 11 PM Monday - Friday; weeknight = after 11 PM Monday - Thursday; Weekend = 11 PM Friday through 7 AM following Monday.

Figure 3.

Patient and event characteristics by carbon dioxide group. CICU = cardiac intensive care unit; CPR = cardiopulmonary resuscitation; FSS = functional status scale; PCPC = pediatric cerebral performance category; PICU = pediatric intensive care unit; PRISM = pediatric risk of mortality score.*Carbon dioxide group: hypocapnia = lowest Pa_CO2 <30 mm Hg; normocapnia = all Pa_CO2 measurements 30–49 mm Hg; hypercapnia = highest Pa_CO2 ≥50 mm Hg; hypocapnia and hypercapnia = lowest Pa_CO2 in hypocapnia range and highest Pa_CO2 in hypercapnia range. ^†Fisher’s exact test. ^‡Baseline PCPC and baseline FSS represent subject status prior to the event leading to hospitalization. ^§Kruskal-Wallis test. ^‖Weekday = between 7 AM and 11 PM Monday - Friday; weeknight = after 11 PM Monday - Thursday; Weekend = 11 PM Friday through 7 AM following Monday.

Figure 4.

Association of the oxygenation and carbon dioxide groups with survival to hospital discharge. Results are based on Poisson regression with robust error estimates. Models adjust for illness category, age, and duration of CPR. *Oxygenation group: hypoxemia = lowest Pa_O2 <60 mm Hg; normoxemia = all Pa_O2 measurements, 60–199 mm Hg; hyperoxemia = highest Pa_O2 ⩾200 mm Hg; hypoxemia and hyperoxemia=  lowest Pa_O2 in the hypoxemic range and highest Pa_O2 in the hyperoxemic range. ^†Carbon dioxide group: hypocapnia = lowest arterial carbon dioxide tension/pressure (Pa_CO2) <30 mm Hg; normocapnia = all Pa_CO2 measurements, 30–49 mm Hg; hypercapnia = highest Pa_CO2 ⩾50 mm Hg; hypocapnia and hypercapnia = lowest Pa_CO2 in the hypocapneic range and highest Pa_CO2 in the hypercapneic range.

Figure 5.

(A) Spline survival rates by decile of lowest arterial oxygen tension/pressure (Pa_O2; mm Hg) in first 24 hours postarrest. (B) Spline survival rates by decile of highest Pa_O2 (in mm Hg) in first 24 hours postarrest. Survival to hospital discharge probability is based on a spline regression, controlling for illness category, age and duration of cardiopulmonary resuscitation (CPR; in min). Internal knots are at the 5th, 25th, 50th, 75th, and 95th percentiles.

Figure 6.

(A) Spline survival rates by decile of lowest arterial carbon dioxide tension/partial pressure (P_CO2; mm Hg) in first 24 hours postarrest. (B) Spline survival rates by decile of highest arterial P_CO2 (in mm Hg) in first 24 hours postarrest. Survival to hospital discharge probability is based on a spline regression, controlling for illness category, age and duration of CPR (in min). Internal knots are at the 5th, 25th, 50th, 75th, and 95th percentiles.