Maintaining proper cerebral oxygen delivery is crucial for preventing secondary ischaemic injury in patients with acute brain injuries (ABI). Arterial carbon dioxide tension (PaCO₂) significantly influences cerebral blood flow (CBF). When intracranial pressure (ICP) is not elevated, adjusting PaCO₂ to an optimal level can support sufficient cerebral oxygenation and reduce ischaemic damage.
A recent article discusses the relationship between PaCO₂ and CBF and highlights evidence-based strategies for managing PaCO₂ in common ABI conditions.
The relationship between PaCO₂ and CBF is nearly linear within the range of 20–80 mmHg, with each 1 mmHg increase in PaCO₂ causing a 3–6% rise in CBF, and each 1 mmHg decrease leading to a 1–3% reduction in CBF. CO₂ reactivity, or the degree to which PaCO₂ influences CBF, can be reduced in older individuals or in severe ABI. The effects of PaCO₂ on CBF are regionally varied, with grey matter showing two- to three-fold greater CO₂ reactivity than white matter.
The CBF response to CO₂ is driven by changes in cerebrospinal fluid (CSF) pH. CO₂ has a stronger and faster effect on CBF than systemic arterial pH because it easily crosses the blood-brain barrier and directly impacts perivascular pH. Hypercapnia causes CSF acidification, leading to vasodilation of pial resistance arterioles, increased CBF, and blood volume, which can raise intracranial pressure (ICP) in patients with limited compensatory reserve. In contrast, hypocapnia causes CSF alkalosis, vasoconstriction of pial arterioles, and reduced CBF, lowering blood volume and ICP, which is why hyperventilation is used clinically to control elevated ICP.
Sustained, profound hyperventilation can lead to cerebral ischaemia by reducing CBF. PaCO₂ levels of 25 mmHg within the first 24–36 hours after ABI significantly increase metabolic stress markers in the cerebrospinal fluid. Imaging studies show a substantial reduction in CBF when PaCO₂ is lowered from 36 to 29 mmHg. Profound hyperventilation is only recommended briefly for refractory intracranial hypertension or imminent herniation. While robust data is lacking, mild hypocapnia (PaCO₂ 32–35 mmHg) appears to pose less risk, supporting cautious, short-term hyperventilation when elevated ICP is a concern.
Clinical guidelines recommend maintaining PaCO₂ between 35–45 mmHg in patients with ABI who do not have elevated ICP, with allowable reductions in cases of elevated ICP. A sub-analysis of the ENIO observational study (1476 patients) found a U-shaped relationship between PaCO₂ and outcomes, with significantly higher in-hospital mortality at PaCO₂ levels below 32 mmHg and above 45 mmHg compared to normal levels (35–45 mmHg). The highest mortality was observed with PaCO₂ levels below 26 mmHg.
The Seattle International Brain Injury Consensus Consortium recommends a tiered approach to targeting PaCO₂ in patients with severe traumatic brain injury (TBI) to manage ICP: tier 1 (low-normal PaCO₂: 35–38 mmHg) and tier 2 (mild hypocapnia: 32–35 mmHg). The Brain Trauma Foundation advises against routinely lowering PaCO₂ below 30 mmHg, based on a 30-year-old randomised trial showing worse long-term outcomes with prolonged prophylactic hyperventilation compared to normocapnia. However, mean PaCO₂ levels vary across hospitals. Brain oxygen monitoring, such as jugular venous oximetry or brain tissue probes, is recommended to guide therapeutic hyperventilation and prevent CBF attenuation and hypoxia. In the ongoing BONANZA and BOOST-III trials, mild hypercapnia (45–50 mmHg) is being tested to improve brain tissue oxygen by increasing CBF when ICP is normal, though this strategy remains investigational and cannot be recommended without invasive intracranial monitoring.
Contemporary guidelines do not specifically address PaCO₂ targets for patients with ischaemic or haemorrhagic stroke, and recommendations are often drawn from TBI literature. Mechanical ventilation guidelines for ABI suggest PaCO₂ targets of 35–45 mmHg, including for stroke patients. However, several disease-specific factors should be considered when managing PaCO₂ in these cases.
In patients with subarachnoid haemorrhage, spontaneous hyperventilation and hypocapnia are common. While hypocapnia can reduce CBF, especially in those with vasospasm or delayed cerebral ischaemia, normalising PaCO₂ may not always be advisable due to risks like sedation, paralysis, and loss of neurological assessment. An individualised approach is often needed. Conversely, controlled hypercapnia (up to 60 mmHg) has been shown to improve CBF and brain oxygenation without raising ICP in a study of six patients with high-grade subarachnoid haemorrhage. A subgroup analysis of the ENIO sub-study found mild hypocapnia was linked to higher in-hospital mortality following subarachnoid haemorrhage, while hypercapnia was not. Controlled hypercapnia may offer a therapeutic option for managing delayed cerebral ischaemia in certain patients, though further research is needed.
In patients with intracranial haemorrhage, PaCO₂ management follows general guidelines. Avoiding hypercapnia is especially important for those with large hematomas causing mass effect, as even small increases in CBF and blood volume can raise ICP in cases of reduced intracranial compliance.
Current guidelines for patients with post-cardiac arrest brain injury recommend adjusting ventilation targets to maintain PaCO₂ between 35–45 mmHg. The TAME trial, which included 1700 patients with out-of-hospital cardiac arrest, found that targeting a PaCO₂ of 50–55 mmHg did not improve neurological function or reduce mortality at 6 months compared to normocapnia (35–45 mmHg). This neutral outcome may be due to impaired CO₂ reactivity in many patients following cardiac arrest.
After the return of spontaneous circulation, CBF typically experiences a brief hyperemic phase (10–30 minutes), followed by a prolonged low-flow phase. The risk of hypocapnia-induced ischaemia is particularly high during the low-flow phase, emphasising the importance of avoiding PaCO₂ levels below 35 mmHg once circulation is restored.
Normocapnia is recommended for most patients with ABI, especially in the early phase, as PaCO₂-induced changes in CBF can significantly affect ICP and the risk of secondary brain injury. Mild hypocapnia achieved through hyperventilation can be useful for managing elevated ICP and acute ICP crises but should be used cautiously to avoid ischaemic complications. Controlled hypercapnia is still under research, and future studies should focus on personalised PaCO₂ management strategies.
Source: Intensive Care Medicine
Image Credit: iStock
References:
Taran S, Sekhon M, Robba C (2025) Carbon dioxide pathophysiology and targets following acute brain injury. Intensive Care Med.
