IntroductionIn former studies, the arterio-alveolar carbon dioxide gradient (ΔCO2) predicted in-hospital mortality after initially survived cardiac arrest. As early outcome predictors are urgently needed, we evaluated ΔCO2 as predictor for good neurological outcome in our cohort. MethodsWe retrospectively analyzed all patients ≥18 years of age after non-traumatic in- and out of hospital cardiac arrest in the year 2018 from our resuscitation database. Patients without advanced airway management, incomplete datasets or without return of spontaneous circulation were excluded. The first arterial pCO2 after admission and the etCO2 in mmHg at the time of blood sampling were recorded from patient's charts. We then calculated ΔCO2 (pCO2 – etCO2). For baseline analyses, ΔCO2 was dichotomized into a low and high group with separation at the median. Good neurological outcome on day 30, expressed as Cerebral Performance Category 1–2, defined our primary endpoint. Survival to 30 days was used as secondary endpoint. ResultsOut of 302 screened patients, 128 remained eligible for analyses. ΔCO2 was lower in 30-day survivors with good neurological outcome (12.2 mmHg vs. 18.8 mmHg, p = 0.009) and in 30-day survivors (12.5 mmHg vs. 20.0 mmHg, p = 0.001). In patients with high ΔCO2, a cardiac etiology of arrest was found less often. They had a higher body mass index, longer duration of resuscitation, higher amounts of epinephrine, lower pO2 levels but both higher pCO2 and blood lactate levels, resulting in lower blood pH and HCO3− levels at admission.In a crude binary logistic regression analysis, ΔCO2 was associated with 30-day neurological outcome (OR = 1.041 per mmHg of ΔCO2, 95% CI 1.008–1.074, p = 0.014). This association persisted after the adjustment for age, sex, witnessed arrest and shockable first rhythm. However, after addition of the duration of resuscitation or the cumulative epinephrine dosage to the model, ΔCO2 lost its association. ConclusionΔCO2 at admission after a successfully resuscitated cardiac arrest is associated with 30 days survival with good neurological outcome. However, a higher ΔCO2 may rather be a surrogate for unfavorable resuscitation circumstances than an independent outcome predictor.