Carbon Dioxide Management in TBI: From Theory to Practice

Abstract

Hyperventilation is a double-edged sword strategy for controlling intracranial volumes and therefore reducing intracranial pressure (ICP) after acute brain damage. The effect is linked to the induced local alkalosis producing a reduction in cerebral blood flow (CBF). Although this effect is short-lasting, hyperventilation carries a potential risk of cerebral ischemia. Although all patients with severe traumatic brain injury (TBI) are mechanically ventilated, the target for the partial pressure of carbon dioxide in arterial blood (PaCO₂) remains poorly defined and there is insufficient evidence to support any recommendation. A target PaCO₂ of ≈36–40 mmHg has been reported by clinicians and suggested by consensus statements. In presence of raised ICP, this is usually lowered to ≈30–35 mmHg. In this chapter, starting from physiological concepts, we will review the evidence around PaCO₂ management in TBI and will present some data on current practice of the use of hyperventilation in TBI. We observed considerable variability in CO₂ management between centers in Europe. Further research is needed to identify optimal ventilatory strategies in the management of severe TBI.

1 Introduction

Hyperventilation is a double-edged sword strategy for controlling intracranial volumes and therefore reducing intracranial pressure (ICP) after acute brain damage. The effect of hyperventilation is due to perivascular alkalosis, producing vasoconstriction and, therefore, reduced cerebral blood flow (CBF). Although this effect is short-lasting, hyperventilation carries a potential risk of cerebral ischemia. Although all patients with severe traumatic brain injury (TBI) are treated with mechanical ventilation, the target of the partial pressure of carbon dioxide in arterial blood (PaCO₂) remains poorly defined and there is insufficient evidence to support any recommendation. Even the latest guidelines and consensus documents state that in patients with severe TBI, normocapnia should be maintained (PaCO₂ 35–45 mmHg) and that, with a low level of evidence, prolonged prophylactic profound hyperventilation is not recommended. A target PaCO₂ of ≈36–40 mmHg has been reported by clinicians and, in the presence of raised ICP, this is usually lowered to ≈30–35 mmHg. In this chapter, starting from physiological concepts, the evidence around PaCO₂ management in TBI will be reviewed and some data on current practice of use of hyperventilation in TBI will be presented.

2 Cerebral Blood Flow and Cerebrovascular CO₂ Reactivity

Brain oxygen consumption is very high, i.e., about 3.5 ml per 100 g/min, which is equivalent to 20% of the oxygen consumption of the whole human body. This high metabolic energetic demand requires a fine-tuned CBF to avoid ischemic conditions, i.e., situations in which the metabolic demand is not fulfilled. Under normal conditions, the CBF is maintained at a constant flow rate of 50–60 ml per 100 g/min in young adults, and 50 ml of oxygen is drawn from 700 to 800 ml of blood per minute. The extraction rate of oxygen by the brain is very high: the arteriovenous blood difference of the brain is 6.3 ml per 100 ml of blood. Considering the high metabolic demand of the brain and the limited storage of substrates, it is necessary to maintain CBF levels within the normal range. Under physiological conditions, this is achieved through a variety of mechanisms, commonly referred to as autoregulation.

The CBF depends on the diameter of the cerebral arterioles (resistance vessels), increasing with vasodilation and decreasing with vasoconstriction. To maintain CBF constant, these vessels physiologically respond to changes in systemic blood pressure, blood viscosity, and metabolic requirements. CBF is functionally linked to regional brain metabolism as expressed in the Fick equation:

$$ {\mathrm{CMRO}}_2= CBF\times {\mathrm{avDO}}_2 $$

where CMRO₂ is the cerebral metabolic rate of oxygen and avDO₂ is the cerebral arterio-venous oxygen difference. Metabolic activity generates CO₂ and the CO₂ reactivity relates to the response of the cerebral vessels, and consequently CBF, to local changes in CO₂.

PaCO₂ can be manipulated with ventilation. Changes in PaCO₂ will elicit movement of CO₂ across the blood-brain barrier, and consequent vascular changes. An increase in PaCO₂ will produce acidosis and consequently vasodilation. A reduction in PaCO₂ will produce alkalosis and vasoconstriction. This effect is short-lasting due to subsequent re-equilibration of the following reaction, catalyzed by carbonic anhydrase:

$$ {CO}_2+{\mathrm{H}}_2\mathrm{O}\rightleftharpoons {\mathrm{H}}_2{CO}_3\rightleftharpoons {HCO_3}^{-}+{\mathrm{H}}^{+} $$

The data obtained from in vitro and in vivo studies suggest that the endothelium and smooth muscles, as well as perivascular nerve cells, neurons, and glia, may be involved in the CO₂ reactivity and this complex mechanism seems to be correlated to changes in perivascular pH. The effect of pH variation on smooth muscle tone can be direct, modifying intracellular calcium concentration, or mediated by a second messenger system, such as nitric oxide (NO), potassium and calcium, prostaglandins, and cyclic nucleotides. Prostaglandins are effective vasodilators that can activate adenylate cyclase and increase cyclic adenosine monophosphate (cAMP). NO produced by the NO synthase (NOS) family in cerebral vascular endothelial cells, perivascular nerves, neurons, and glial cells will increase the intracellular concentration of cyclic guanosine monophosphate, thereby causing vasodilation [1]. Cyclic nucleotides reduce the entry of calcium into vascular smooth muscle and cause vasodilation directly or in a permissible manner, thereby allowing hypercapnia to exert its vasodilation effect. The opening of potassium channels decreases the influx of extracellular calcium into cells in an indirect way, thereby reducing the tension of the vascular smooth muscle. In a clinical setting, every mmHg change in PaCO₂ from 20 to 60 mmHg in patients with TBI produces a CBF change of ≈3% [2] Hypoventilation leading to hypercarbia causes vasodilation and an increase in CBF, while hyperventilation causes vasoconstriction and a decrease in CBF.

3 Effects of Hyperventilation on Intracranial Pressure and CBF

The changes in PaCO₂ modify the intracranial blood volume_, and therefore the pressure of the intracranial compartment (ICP). The relationship between PaCO₂ and ICP is not linear. In experimental studies over wide ranges of PaCO₂, an S-shaped relationship has been described between ICP and PaCO₂ [3]. Stocchetti et al. [4] calculated that for each millimeter of mercury of change in PaCO₂ the blood volume changed by 0.72 ± 0.42 ml. Similarly, Yoshihara et al. [5] demonstrated that, in patients with severe TBI, a change in blood volume of 0.5 ml could produce an ICP change of 1 mmHg. However, it is still unclear if the effect of hyperventilation on ICP remains during prolonged hyperventilation. After 24 h, as the pH of the perivascular spaces normalizes, the vasoconstrictive effect is reduced.

One of the main concerns in hyperventilating patients with TBI to reduce raised ICP is the risk of ischemia-induced by CBF reduction. One of the first descriptions of the therapeutic use of hyperventilation to treat elevated ICP was published by Lundberg et al. [6] in 1959. These authors stated that hyperventilation did not cause ischemia thanks to compensatory mechanisms that act to maintain tissue oxygenation. In fact, in normal conditions, cerebral oxygen delivery exceeds the brain’s oxygen consumption, and this mechanism leaves an important reserve that allows the brain to tolerate CBF reduction, as occurs in hyperventilation. One year later, in 1960, Meyer et al. [7] confirmed this hypothesis, demonstrating that in healthy volunteers, hyperventilation produced no change in the CMRO₂. In 2002, Coles et al. [8] demonstrated that moderate hypocapnia could significantly reduce global CBF and result in significant increases in the volume of critically hypoperfused tissue in the injured brain even when improved cerebral perfusion pressure (CPP) and ICP values were recorded. These authors used positron emission tomography (PET) to quantify regional CBF and metabolism in response to CO₂ changes in 33 patients with TBI. This effect was not limited to the first 24 h after brain injury and the authors suggested that even brief periods of hyperventilation may cause a harmful reduction in CBF when PaCO₂ is reduced below 33 mmHg. The immediate effect on CBF of hyperventilation is clear, but the consequences of it, and the link with ischemia, is somehow controversial. Diringer et al. [9], using PET during hyperventilation in 13 patients with severe TBI in the first 8–14 h after TBI, found that CBF decreased, oxygen extraction fraction increased, and CMRO₂ was unchanged, suggesting that low CBF can be explained in the setting of patients with severe TBI treated with sedatives, with a primary reduction in CMRO₂ and a secondary passive fall in CBF.

Similarly, Letarte et al. [10], using a microdialysis probe placed adjacent to injured brain in 8 patients with severe TBI, concluded that hyperventilation that lowered PaCO₂ by 10 mmHg for 30 min (baseline PaCO₂ 35 ± 2 mmHg) reduced local CBF by 20%, while the lactate/pyruvate ratio did not change.

Ischemia is defined as a mismatch between metabolic requirements and CBF. Hyperventilation causes a reduction in CBF, but it does not always seem to be associated with a reduction in oxygen and metabolic supply. Finally, in 2006, Marion [11] explained that these discrepancies in findings concerning low PaCO₂ levels could be due to a loss of integrity of local CO₂ autoregulation, which may be impaired after TBI, and suggested that in patients with raised ICP responding to short-term hyperventilation, it can be considered safe.

4 Current Recommendations

Only one randomized controlled trial (RCT) is available about hyperventilation in TBI. In this trial, Muizelaar et al. [12] randomized 113 patients into three groups: normal ventilation, defined as PaCO₂ 35 ± 2 mmHg; prophylactic hyperventilation, defined as PaCO₂ 25 ± 2 mmHg; and prophylactic hyperventilation + THAM (tris(hydroxymethyl)aminomethane), which was added to compensate for the loss of HCO₃⁻ buffer from cerebrospinal fluid (CSF) that is responsible for the short-lived effect of hyperventilation on cerebral vasoconstriction. Patients were stratified based on the motor component of the Glasgow Coma Scale (GCS; 1–3 and 4–5). The outcome was assessed using the Glasgow Outcome Scale (GOS) at 3, 6, and 12 months. For patients with the higher motor GCS (motor score 4–5), the 3 and 6 month GOS scores were lower in the hyperventilated patients than in the control or THAM group but the effect was not confirmed at 12 months. This discrepancy between 3–6- and 12-month outcome may be due to a too small sample size; a direct correlation between hyperventilation and worse outcome has not been demonstrated. Moreover, CBF was lower in the hyperventilation + THAM group than in the control and hyperventilation groups, indicating a prolongation of the hyperventilation effect. There was no evidence of cerebral ischemia in any of the three groups, using CBF or avDO₂ data. In addition, in this trial, hyperventilation was used as a prophylactic maneuver and not as a treatment strategy and the course of ICP was most stable in the hyperventilation + THAM group.

Putting all these elements together, the fourth edition of the Brain Trauma Foundation guidelines [13] states that, in patients with TBI, there is insufficient evidence to support a strong recommendation in PaCO₂ management. Prolonged prophylactic hyperventilation with PaCO₂ of 25 mmHg or less is not recommended, but the optimal PaCO₂ range in these patients is still uncertain.

The recent Seattle International Severe Traumatic Brain Injury Consensus Conference [14] recommended mild hyperventilation, i.e., a PaCO₂ 32–35 mmHg, as a tier-two treatment if ICP remains resistant to first-line treatments, such as analgesia and sedation, osmotherapy, CPP maintenance and CSF removal. The Seattle Consensus did not support lower PaCO₂ levels and recommended against routine hyperventilation to below 30 mmHg. In the same direction, a recent consensus on ventilation [15] in acute brain damage suggests that hypercapnia should be avoided in patients with acute brain injury, aiming for a physiologic range of PaCO₂ of between 35 and 45 mmHg. Short term hyperventilation in patients with acute brain injury and brain herniation can be considered as a therapeutic option. No consensus was reached regarding hyperventilation as a therapeutic option in patients with ICP elevation.

5 From Guidelines to Clinical Practice

In 2008, Neumann et al. [16] published data obtained from BrainIT (The brain monitoring with information technology) dataset analyzing 7703 blood gas analyses from 151 patients with TBI across 17 centers in Europe. The mean PaCO₂ was 35.8 ± 5.6 mmHg and the PaCO₂ was distributed in the range of normoventilation (PaCO₂ 36–45 mmHg) and moderate hyperventilation (PaCO₂ 35–31 mmHg). Early prophylactic hyperventilation as well as the use of additional cerebral oxygenation monitoring during hyperventilation, suggested by the Brain Trauma Foundation guidelines at that time, were not followed by most of the centers.

In 2018, Huijben et al. [17] performed a survey on treatment strategies before starting the Collaborative European NeuroTrauma Effectiveness Research in Traumatic Brain Injury (CENTER-TBI) trial. The most frequently reported PaCO₂ target was 36–40 mmHg in case of controlled ICP < 20 mmHg (69%), and 30–35 mmHg in case of raised ICP (62%), underlining that, in clinical practice, hyperventilation is used as a therapeutic option in patients with intracranial hypertension.

Some information on current CO₂ management was obtained by the CENTER-TBI study, a longitudinal prospective collection of data from patients with TBI across 65 centers in Europe. The study was conducted between December 19, 2014, and December 17, 2017, and aimed to better describe the incidence, management, and outcomes of patients with TBI in Europe [18]. For each patient enrolled in CENTER-TBI, along with much other information, the daily highest and lowest PaCO₂ values were registered for the first 7 days of admission. In 1100 mechanically ventilated patients with TBI admitted to the ICU for whom more than 2 values were available, the mean daily lowest PaCO₂ was 35.22 mmHg (SD = 5.27), very similar to the BrainIT data. However, there was huge variability in the mean value of the lowest PaCO₂ across centers (ranging from 32.3 to 38.6 mmHg) (Fig. 21.1), highlighting important differences in the way that CO₂ is managed. In patients with ICP monitoring, the observed mean values were lower compared with patients without ICP monitoring, i.e., 34.7 vs. 36.8 mmHg. A total of 397 patients had at least one episode of PaCO₂ <30 mmHg. These data suggest that, with considerable inter-center variability, hyperventilation is still largely used in TBI patients. At this stage, we are still exploring the effects of these management differences.

Fig. 21.1

Lowest mean PaCO₂ value in the CENTER-TBI trial [18]. Data from 1100 patients enrolled in 36 European centers participating in Center-TBI. For each center, the mean values of the lowest PaCO₂ values recorded daily in the first week are represented by a dot, with their respective confidence interval. The solid line is drawn in correspondence with the lowest mean PaCO₂. The mean daily lowest PaCO₂ from different centers ranged from 32.3 to 38.6 mmHg. This result seems to be linked more to the different management strategies of the individual centers, rather than being related to different national PaCO₂ management strategies

6 Conclusion

The relationship between CBF and PaCO₂ is well known. For 60 years, authors have discussed the risk of ischemia related to a reduction in CBF induced by hyperventilation but no strong evidence exists. Although hyperventilation induces a CBF reduction, it has not yet been demonstrated that this reduction always corresponds to an insufficient metabolic supply and therefore to ischemia.

In clinical practice, because of the lack of strong recommendation, the management of PaCO₂ is variable among centers, and hyperventilation is still commonly used, especially in patients with raised ICP. The correlation between PaCO₂ levels and outcome is still under exploration and further studies are required to better characterize ventilator strategies in patients with TBI and their effect on outcomes.