Management of arterial partial pressure of carbon dioxide in the first week after traumatic brain injury: results from the CENTER-TBI study

Giuseppe Citerio*, Chiara Robba, Paola Rebora, Matteo Petrosino, Eleonora Rossi, Letterio Malgeri, Nino Stocchetti, Stefania Galimberti, David K. Menon, Caroline M. van Heugten, CENTER-TBI Participants and Investigators

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

13 Citations (Web of Science)


Purpose To describe the management of arterial partial pressure of carbon dioxide (PaCO2) in severe traumatic brain-injured (TBI) patients, and the optimal target of PaCO2 in patients with high intracranial pressure (ICP). Methods Secondary analysis of CENTER-TBI, a multicentre, prospective, observational, cohort study. The primary aim was to describe current practice in PaCO2 management during the first week of intensive care unit (ICU) after TBI, focusing on the lowest PaCO2 values. We also assessed PaCO2 management in patients with and without ICP monitoring (ICPm), and with and without intracranial hypertension. We evaluated the effect of profound hyperventilation (defined as PaCO2 < 30 mmHg) on long-term outcome. Results We included 1100 patients, with a total of 11,791 measurements of PaCO2 (5931 lowest and 5860 highest daily values). The mean (+/- SD) PaCO2 was 38.9 (+/- 5.2) mmHg, and the mean minimum PaCO2 was 35.2 (+/- 5.3) mmHg. Mean daily minimum PaCO2 values were significantly lower in the ICPm group (34.5 vs 36.7 mmHg, p < 0.001). Daily PaCO2 nadir was lower in patients with intracranial hypertension (33.8 vs 35.7 mmHg, p < 0.001). Considerable heterogeneity was observed between centers. Management in a centre using profound hyperventilation (HV) more frequently was not associated with increased 6 months mortality (OR = 1.06, 95% CI = 0.77-1.45, p value = 0.7166), or unfavourable neurological outcome (OR 1.12, 95% CI = 0.90-1.38, p value = 0.3138). Conclusions Ventilation is manipulated differently among centers and in response to intracranial dynamics. PaCO2 tends to be lower in patients with ICP monitoring, especially if ICP is increased. Being in a centre which more frequently uses profound hyperventilation does not affect patient outcomes.

Original languageEnglish
Pages (from-to)961-973
Number of pages13
JournalIntensive Care Medicine
Issue number9
Publication statusPublished - Sept 2021


  • Carbon dioxide
  • Hyperventilation
  • Traumatic brain injury
  • Intracranial pressure
  • Outcome

Cite this