Regional brain blood flow in man during acute changes in arterial blood gases

Abstract

Despite the importance of blood flow on brainstem control of respiratory and autonomic function, little is known about regional cerebral blood flow (CBF) during changes in arterial blood gases.We quantified: (1) anterior and posterior CBF and reactivity through a wide range of steady-state changes in the partial pressures of CO2 (PaCO2) and O2 (PaO2) in arterial blood, and (2) determined if the internal carotid artery (ICA) and vertebral artery (VA) change diameter through the same range.We used near-concurrent vascular ultrasound measures of flow through the ICA and VA, and blood velocity in their downstream arteries (the middle (MCA) and posterior (PCA) cerebral arteries). Part A (n =16) examined iso-oxic changes in PaCO2, consisting of three hypocapnic stages (PaCO2 =∼15, ∼20 and ∼30 mmHg) and four hypercapnic stages (PaCO2 =∼50, ∼55, ∼60 and ∼65 mmHg). In Part B (n =10), during isocapnia, PaO2 was decreased to ∼60, ∼44, and ∼35 mmHg and increased to ∼320 mmHg and ∼430 mmHg. Stages lasted ∼15 min. Intra-arterial pressure was measured continuously; arterial blood gases were sampled at the end of each stage. There were three principal findings. (1) Regional reactivity: the VA reactivity to hypocapnia was larger than the ICA, MCA and PCA; hypercapnic reactivity was similar.With profound hypoxia (35 mmHg) the relative increase in VA flow was 50% greater than the other vessels. (2) Neck vessel diameters: changes in diameter (∼25%) of the ICA was positively related to changes in PaCO2 (R2, 0.63±0.26; P<0.05); VA diameter was unaltered in response to changed PaCO2 but yielded a diameter increase of +9% with severe hypoxia. (3) Intra- vs. extra-cerebral measures: MCA and PCA blood velocities yielded smaller reactivities and estimates of flow than VA and ICA flow. The findings respectively indicate: (1) disparate blood flow regulation to the brainstem and cortex; (2) cerebrovascular resistance is not solely modulated at the level of the arteriolar pial vessels; and (3) transcranial Doppler ultrasound may underestimate measurements of CBF during extreme hypoxia and/or hypercapnia.

Figure 1. Relationship between arterial () and end-tidal () partial pressures of carbon dioxide during targeted steady state changes in

A, Regression equation: = 0.882 () + 2.47 (R² = 0.98). B, Bland–Altman plot of differences between and , and the mean value of both. Dashed lines represents the 95% confidence intervals and the mean bias.

Figure 2. Blood flow (top panels; ICA and VA) and blood velocity (bottom panels; MCA and PCA) during steady state changes in arterial CO₂ (left panels) and oxygen (right panels)

*Difference from baseline (40 mmHg or 100 ;), P < 0.05. †Differences between vessel flow (ICA vs. VA) or velocity (MCA vs. PCA) at a given stage; : P < 0.006; : P < 0.012. All values are means ± SD. ICA, internal carotid artery; VA, vertebral artery; MCA, middle cerebral artery; PCA, posterior cerebral artery. Note: the number of subjects comprising each mean value is different between stages and vessels. Please refer to Tables 1 and 2 for these values.

Figure 3. Percentage change from baseline in blood flow (; ICA and VA) and blood velocity (CBV; MCA and PCA) during steady state changes in arterial CO₂ (top panels) and oxygen (bottom panels)

All values are mean ± SD. ICA, internal carotid artery; VA, vertebral artery; MCA, middle cerebral artery; PCA, posterior cerebral artery. The relative flow change in the neck arteries were ∼50% greater than velocity change in the intracranial vessels during hypercapnia, suggesting either anatomical flow redistribution or dilatation of the MCA and PCA – or both. Note the much greater increase in VA flow during hypoxia compared to the other vessels; see Fig. 4 for comparative CO₂ and O₂ reactivities. Note: the number of subjects comprising each mean value is different between stages and vessels. Please refer to Tables 1 and 2 for these values.

Figure 4. Mean (SD) cerebral blood flow responses to CO₂ and hypoxaemia

The top row depicts the slope of the normalized CBF change (A%, relative to baseline) and or ; the bottom row depicts absolute flow (; ICA and VA) or blood velocity (CBV; MCA and PCA) and or . Horizontal bars indicate significant relationships, P < 0.05.

Figure 5. Mean (squares; ±SD) and individual (circles) percentage change in luminal diameters of the internal carotid artery during steady state change in

Linear regression of individual data y = 3.6 × 10⁻¹ () – 13.3; R² = 0.63. Through a large range, internal carotid diameter changes by ∼20%, indicating that cerebrovascular resistance is not solely modulated at cerebral arterioles with changes in the partial pressure of arterial blood gases.

Figure 6. Percent change from baseline in MAP(A) and relationship between ΔMAP and ΔCBF or ΔCBV (B) during hypercapia and hypoxia

A, percentage change in (ΔMAP) relative to baseline for (top row) and hypoxaemia (bottom row) trials. Right hand eight plots (B) depict ΔMAP in the hypercapnic or hypoxic ranges (shown boxed in A). ICA, internal carotid artery; VA, vertebral artery; MCA, middle cerebral artery; PCA, posterior cerebral artery. The hypercapnia related hypertension was positively related (P < 0.05) to elevations in CBF/CBV in all vessels; in contrast, hypoxia induced increases in MAP were not related to CBF/CBV.

Figure 7. Bland–Altman plots of differences between neck artery blood flows and the respective downstream intracranial vessel, and the mean value of both

Values are percentage change from baseline. Dashed line represents the 95% confidence intervals, dotted line the mean bias. All plots show systematic error proportional to the increase in CBF; A and B suggest systematic and random error. Data show that at lower CBF values TCD estimates of intracranial blood velocity accurately reflect CBF for both the vertebral–PCA systems and ICA–MCA systems. With high CBF with hypoxia or hypercapnia MCAv and PCAv are likely to underestimate CBF.