Cerebral blood flow sensitivity to CO₂ measured with steady-state and Read's rebreathing methods

Abstract

The ventilatory response to carbon dioxide (CO₂) measured by the steady-state method is lower than that measured by Read's rebreathing method. A change in end-tidal P_CO2 (Pet_CO2) results in a lower increment change in brain tissue P_CO2 (Pt_CO2) in the steady-state than with rebreathing: since Pt_CO2 determines the ventilatory response to CO₂, the response is lower in the steady-state. If cerebral blood flow (CBF) responds to Pt_CO2, the CBF–CO₂ response should be lower in the steady-state than with rebreathing. Six subjects undertook two protocols, (a) steady-state: Pet_CO2 was held at 1.5 mmHg above normal (isocapnia) for 10 min, then raised to three levels of hypercapnia, (8 min each; 6.5, 11.5 and 16.5 mmHg above normal, separated by 4 min isocapnia). End-tidal P_O2 was held at 300 mmHg; (b) rebreathing: subjects rebreathed via a 6 L bag filled with 6.5% CO₂ in O₂. Transcranial Doppler-derived CBF yielded a higher CBF–CO₂ sensitivity in the steady-state than with rebreathing, suggesting that CBF does not respond to Pt_CO2.

Introduction

Read, 1967, Read and Leigh, 1967 described a rebreathing technique to measure the central chemoreceptor ventilatory response to CO₂. Their method is still used widely in basic science and clinical research (Saito et al., 1995). Read (1967) used a small bag (about 6 L) filled with a gas mixture of about 6–7% CO₂ in O₂. Under these conditions, rebreathing was initiated at a CO₂ tension close to that of mixed venous blood (Pv_CO2) and the subsequent increases in end-tidal P_CO2 (Pet_CO2) and ventilation ($\text{V}\text{̇}$e) were linear in time. It is generally thought that initiating rebreathing close to Pv_CO2 leads to a rapid equalisation of Pet_CO2, arterial P_CO2 (Pa_CO2), Pv_CO2, and presumably also brain tissue (extracellular) P_CO2 (Pt_CO2), and that all these variables increase at the same rate (Rebuck and Slutsky, 1981).

Read's method differs from the alternative, steady-state method in which inspired P_CO2 is manipulated to increase Pet_CO2 in a step to a new level, and then held constant at this new, higher level until the ventilatory response reaches a steady-state. With this method (and in contrast with rebreathing) there is always a gradient between Pet_CO2 and Pt_CO2.

Some studies have found good agreement between Read's method and the steady-state method for the central chemoreflex ventilatory CO₂-sensitivity (Read, 1967, Clark, 1968), but most investigators have concluded that the ventilatory CO₂-sensitivity measured by Read's method was greater than that measured by the steady-state method (Tenney et al., 1963, Honda and Miyumara, 1972, Linton et al., 1973, Jacobi et al., 1989, Berkenbosch et al., 1989, Bourke and Warley, 1989, Lumb and Nunn, 1991, Mohan et al., 1999). Berkenbosch et al. (1989) have argued that this more usual result was predictable from first principles: since Pet_CO2 and Pt_CO2 are always equal during rebreathing, the corresponding rise in Pt_CO2 must be the same as a unit rise in Pet_CO2. However, since a gradient always exists between Pet_CO2 and Pt_CO2 in the steady-state, a unit rise in Pet_CO2 will yield a smaller corresponding rise in Pt_CO2 (and as the CO₂ perturbation increases in the steady-state, the difference between Pet_CO2 and Pt_CO2 will decrease). Poulin and Robbins (1998) have estimated that a step increase in Pa_CO2 of 7.5 mmHg results in a corresponding rise in Pt_CO2 of only 5.5 mmHg in the steady-state. Because Pt_CO2 in the region of the central chemoreceptor ultimately determines the ventilatory response to CO₂, the slope of the $\text{V}\text{̇}$e–Pet_CO2 relationship will, therefore, be greater with rebreathing than with the steady-state method. Additionally (because of the gradient between Pet_CO2 and Pt_CO2 in the steady-state), it is expected that the x-axis intercept of the $\text{V}\text{̇}$e versus Pet_CO2 plot will be shifted to the left as compared with the rebreathing plot. These predictions have been confirmed experimentally (Berkenbosch et al., 1986, Berkenbosch et al., 1989, Dahan et al., 1990).

Read's method would seem to provide a relatively simple means of effecting a predictable change in CO₂ tensions, whilst at the same time allowing continuous estimation of the cerebral blood flow (CBF). However, to our knowledge, Read's method and the steady-state methods have not previously been directly compared with regard to CBF sensitivities to CO₂.

We made two assumptions. First, we assumed that brain extracellular Pt_CO2 is the prime determinant of the CBF response to CO₂. This view is widely-held (at cellular level, Pt_CO2 might exert its effect through changes in local pH), and Pt_CO2 is thought to be one of the main mediators of the increase in blood flow in response to increased local metabolic brain activity (Busija and Heistad, 1984, Brian, 1998). Second, we assumed that if this is the case, the same argument employed by Berkenbosch and colleagues regarding the reasons for the rebreathing ventilatory CO₂ sensitivity being higher than steady-state sensitivity, would also apply to the CBF response to CO₂. We consequently predicted that (a) the CBF sensitivity to CO₂ would be higher when measured by Read's technique as compared with the steady-state method; (b) the x-axis intercept of the plot of CBF versus Pet_CO2 will be shifted to the left for the steady-state as compared with rebreathing. The purpose of this study was to test these predictions and compare CBF sensitivities to CO₂ measured by rebreathing and steady-state methods.