Optimal Targets of the First 24-h Partial Pressure of Carbon Dioxide in Patients with Cerebral Injury: Data from the MIMIC-III and IV Database

Abstract

It is generally believed that hypercapnia and hypocapnia will cause secondary injury to patients with craniocerebral diseases, but a small number of studies have shown that they may have potential benefits. We assessed the impact of partial pressure of arterial carbon dioxide (PaCO2) on in-hospital mortality of patients with craniocerebral diseases. The hypothesis of this research was that there is a nonlinear correlation between PaCO2 and in-hospital mortality in patients with craniocerebral diseases and that mortality rate is the lowest when PaCO2 is in a normal range. We identified patients with craniocerebral diseases from Medical Information Mart for Intensive Care third and fourth edition databases. Cox regression analysis and restricted cubic splines were used to examine the association between PaCO2 and in-hospital mortality. Nine thousand six hundred and sixty patients were identified. A U-shaped association was found between the first 24-h PaCO2 and in-hospital mortality in all participants. The nadir for in-hospital mortality risk was estimated to be at 39.5 mm Hg (p for nonlinearity\u2009<\u20090.001). In the subsequent subgroup analysis, similar results were found in patients with traumatic brain injury, metabolic or toxic encephalopathy, subarachnoid hemorrhage, cerebral infarction, and other encephalopathies. Besides, the mortality risk reached a nadir at PaCO2 in the range of 35–45 mm Hg. The restricted cubic splines showed a U-shaped association between the first 24-h PaCO2 and in-hospital mortality in patients with other intracerebral hemorrhage and cerebral tumor. Nonetheless, nonlinearity tests were not statistically significant. In addition, Cox regression analysis showed that PaCO2 ranging 35–45 mm Hg had the lowest death risk in most patients. For patients with hypoxic-ischemic encephalopathy and intracranial infections, the first 24-h PaCO2 and in-hospital mortality did not seem to be correlated. Both hypercapnia and hypocapnia are harmful to most patients with craniocerebral diseases. Keeping the first 24-h PaCO2 in the normal range (35–45 mm Hg) is associated with lower death risk.