Main

Several techniques are currently available to investigators for measuring CBF. These include xenon clearance, radionucleotide angiography, the Kety-Schmidt/nitrous oxide method, hydrogen ion washout(1–4), and radioactive- or fluorescent-labeled microspheres(5). Among these, microspheres are generally considered to provide the most accurate measurements of actual CBF. However, the microsphere technique suffers from two related but distinguishable drawbacks. First, only a small number of CBF measurements (usually four or five) can be determined during the course of one experimental protocol. Second, because microsphere measurements are necessarily intermittent, phasic changes of CBF over short time intervals (i.e. seconds to minutes) are undetectable by this technique.

Flow probe technology (either electromagnetic or ultrasonic) has the potential to overcome both of these limitations. Flow probes provide continuous measurements on a beat-to-beat, on-line basis over long periods (at least hours) of time. In addition, flow probe measurements can be electronically and mechanically zeroed and signals can be compared during brief periods of mechanical occlusion (stopped-flow) to electronic zero flow signals and even, after the experimental animal is killed, to postmortem absolute no-flow signals. This zero flow capability, unique to flow probes, provides additional assurance that assessments of relative changes in blood flow, as opposed to or in addition to absolute changes, can be determined accurately.

However accurate the flow probe signal may be, appropriate interpretation of flow probe measurements depends, necessarily, on knowledge of the distribution of blood flow distal to the probe. For some organs (e.g. kidney or gut), interpretation of flow probe signals (placed upon the renal or mesenteric artery) is relatively straightforward. For other organs, particularly the brain, it may not be.

A flow probe placed around the common carotid artery may measure blood flow to the face, the brain, or some combination of the two. Attempts to disentangle intracerebral versus extracerebral blood flow often involve occlusion of either the external or internal carotid circulations, for example, by ligation or thrombosis(6,7). However, the success of these attempts to isolate CBF varies from species to species, in large part as a function of the relative sizes of the internal versus external carotid arteries and the potential for anatomic communication distal to the experimental occlusion.

In some species (e.g. sheep, ox, goat, cat), blood flow from the common carotid artery passes almost entirely to the external carotid circulation and from there to both the face and brain(8–10). In other species (e.g. man, rhesus monkey, horse, pig), the internal carotid artery dominates the afferent cerebral circulation(8,10–12). The presence of a rete mirabile may, but does not necessarily(6,7,12), further confound correlations of carotid artery blood flow and CBF. At a minimum, however, these considerations should raise serious doubts about the validity of a facile inference that blood flow measured in the common carotid artery necessarily reflects blood flow directed to the brain.

To address these concerns in piglets, we investigated whether blood flow determined by a flow probe placed on the right common carotid artery provided an accurate estimation of CBF determined by microspheres. We correlated blood flow through the right common carotid artery versus CBF under two separate conditions: 1) before ligation of the right external carotid artery, where right common carotid blood flow represents the sum of flow through the right external and internal carotid circulations (CCABF condition); and 2) after ligation of the right external carotid artery (ipsilateral to the flow probe) with residual right-sided CCABF representing only flow through the right internal carotid artery (ICABF condition). The left carotid artery was not manipulated in any way in either protocol. We studied these correlations under conditions of normal CBF and during high CBF conditions induced by hypercarbia or hypoxia.

METHODS

Anesthesia and ventilation. Eight piglets [3-wk-old, 4.0 ± 0.1 kg (SD)] received ketamine (25 mg/kg) with xylazine (5 mg/kg) intramuscularly. A tracheostomy tube was placed by cutdown; air leak around the tracheostomy tube was prevented by a snug circumferential tie at the membrane between tracheal rings. An ear vein was cannulated for venous access. Mechanical ventilation (Baby Bird Inc., Palm Springs, CA) was used to produce initial arterial PCO2 (PaCO2)value of 30-40 torr with FiO2 of 0.30-0.50. The piglets received fentanyl (150 µg/kg) and pancuronium bromide (0.1 mg/kg). Throughout the experiment, the level of anesthesia was assessed by response of heart rate and blood pressure to noxious stimuli and was adjusted by intermittent bolus of fentanyl (approximately 5 µg/kg) and pancuronium (0.1 mg/kg). Overhead heating lamps were used to maintain rectal temperature at 37-38° C. The femoral artery was catheterized by cutdown, and the catheter was advanced into the abdominal aorta to collect microspheres. A polyethylene catheter was placed by cutdown in the left common carotid artery, was attached to a transducer, and was advanced into the left ventricle for the purpose of injection of microspheres. Accurate placement of the left ventricular catheter was determined by assessing the left ventricular pressure tracing and was confirmed at the end of the experimental protocol.

Carotid artery dissection. We have previously described our surgical technique for isolating the common carotid and internal carotid arteries in piglets(13,14). This procedure represents a modification of the methods of Buckley et al. and Scremin et al.(15–18) and facilitates continuous measurement of phasic CCABF or ICABF with a flow probe. Two separate dissections were made in the right neck region. Through the proximal dissection, at a level midway between the root of the aorta and the angle of the jaw, the right common carotid artery was identified. The vagus nerve was carefully freed from the vessel at this level, and an external ultrasonic flow probe (#2R, Transonics Systems Inc., Ithaca, NY) was placed on the right common carotid artery.

A separate distal dissection was made at the level of the angle of the jaw, leaving approximately 2 cm of tissue intact between dissections to isolate the carotid artery blood flow probe from any physical manipulation during the mechanical zero procedure. Through the distal dissection, the right external carotid artery was identified at its origin immediately cephalad to the carotid bifurcation. A purse-string suture was loosely placed around the external carotid artery at that site. At a designated point in the course of each protocol, the purse-string suture around the external carotid artery was pulled tight, effectively redirecting all blood flow through the common carotid artery into the internal carotid artery.

In this article, we will follow the terminology of Stodkilde-Jorgensen et al.(19) who use the name "internal carotid artery" to designate in pigs the large blood vessel (equal in size to the external carotid artery) that originates at the carotid bifurcation, gives rise to the occipital artery, and then penetrates the cranium giving rise to the carotid rete mirabile. Daniel et al.(8) refer to this vessel as the ascending pharyngeal artery and reserve the name "internal carotid artery" for the efferent vessel passing between the carotid rete mirabile and the circle of Willis. The ascending pharyngeal artery and the internal carotid artery (the term used here) are alternative names for the single large blood vessel originating at the carotid bifurcation (as the alternative to the external carotid artery) and serving as the major route of blood supply for the ipsilateral cerebral hemisphere in pigs.

The flow signal from the common carotid artery probe was recorded continuously on a polygraph recorder and digital values were displayed simultaneously on the electromagnetic flow meter (model T 206, Transonic Systems, Inc., Ithaca, NY). Intermittent mechanical stopped-flow signals were produced by briefly (1-2 s) occluding the right common carotid artery with a vascular clamp through the distal dissection without disturbing the carotid artery blood flow probe. At the end of each experiment, the accuracy of the mechanical zero procedure was verified by comparing the flow probe signal during distal occlusion (mechanical zero) with both the signal produced when the probe was turned off (electrical zero) and the flow probe signal after the animal had been killed (absolute zero). As discussed above, verification of the zero flow signal is particularly important in this model because it allows changes in CBF to be expressed as a percentage of initial values.

Microsphere protocol. After vigorous agitation, 15 µm-diameter microspheres labeled with either 51Cr, 141Ce, 85Sr, or 48Sc (New England Nuclear Laboratories, Boston, MA) suspended in 1 mL 0.9 NaCl (350 000-400 000 microspheres/mL) were injected into the left ventricle for a period of 15-20 s. The catheter was then flushed with 1 mL normal saline. Beginning 15 s before microsphere injection, 4.08 mL of reference arterial blood was withdrawn from the descending aorta at a rate of 2.04 mL/min with a Harvard infusion/withdrawal pump. The residual counts in the injection apparatus were determined and subtracted from the preinjection counts to determine the total counts injected. Left ventricular pressure and heart rate were monitored continuously to assure hemodynamic stability during the microsphere infusion.

After the animal was killed, the brain was removed and the cerebral hemispheres were dissected from the cerebellum and brainstem structures. Total wet cerebral weight was obtained with a microanalytical balance. In five piglets, the left versus right cerebral hemispheres were separated. No other anatomic distinctions were made within cerebral structures. Radioactivity counts were obtained with a gamma counter (Nuclear Chicago, model #1185). By use of previously described methodology(5,20), CBF (expressed as mL/min) was calculated from the following formula: CBF = reference sample withdrawal rate (mL/min) × (cpm of microspheres in the brain/cpm of microspheres in the reference sample). Unilateral CBF was calculated as one-half total CBF. CBF was also normalized as mL/min per 100 g of wet cerebral tissue.

Blood gas monitoring. Blood specimens were taken from the aorta during the stabilization period after surgery and during the hypercarbia or hypoxia phases of each protocol. From these blood samples, pH, PO2, PCO2, and base excess were determined with a blood gas analyzer (ABL II, Radiometer, Copenhagen, Denmark).

Experimental protocols

Baseline CBF versus right CCABF. For clarity, all flow probe measurements obtained before ligation of the ipsilateral right external carotid artery will be referred to as CCABF. After obtaining an arterial blood gas during baseline conditions (normoxia/eucarbia), an initial injection of microspheres was performed. Simultaneously, right-sided CCABF was recorded continuously during the 2-min microsphere infusion, and the average value was determined. Piglets were then subjected to hypercarbia or hypoxia.

Elevated CBF versus right CCABF. To produce hypercarbia, the inspired gas circuit was instantly switched to 10% CO2, 21% O2, 69% N2. An identical protocol was used to induce hypoxia, except that an 8% O2, 92% N2 gas mixture was used. After 7.5 min, a time previously determined by us to produce steady state responses in cerebral hemodynamics (see Fig. 2), an arterial blood gas was obtained to document the extent of either hypercarbia or hypoxia. The second injection of microspheres was then performed. CCABF was recorded continuously during the 2-min microsphere infusion, and the average value was determined.

Figure 2
figure 1

Tracing of flow probe signal in the ICABF condition during the switch from eucarbia to hypercarbia. The ICABF signal responded within 1 min of the change in PCO2 (arrow).

Full size image

Elevated CBF versus right ICABF. While hypercarbia or hypoxia continued, the purse-string suture around the right external carotid artery was pulled tight. After ligation of the external carotid artery, right CCABF represents flow only to the right internal carotid artery. For clarity, all flow probe measurements obtained after ligation of the right external carotid artery will be referred to as ICABF. After a brief stabilization period (2-5 min), the third injection of microspheres was performed. ICABF was recorded continuously during the 2-min microsphere infusion, and the average value was determined.

Baseline CBF versus right ICABF. The ventilator settings and gas mixture were then returned to baseline values. After 7.5 min, an arterial blood gas was obtained, and the fourth (final) microsphere injection was performed. ICABF was recorded continuously during the 2-min microsphere infusion, and the average value was determined.

In sum, flow probe measurements of blood flow through the right common carotid artery were correlated with four microsphere injections: two with the right external and internal carotid artery circulation intact (CCABF condition) and two after ligation of the right external carotid artery with only the right internal carotid artery circulation intact (ICABF condition). For each of these two conditions, one flow probe/microsphere correlation was determined at baseline CBF and one at elevated CBF. Under no condition was the left-sided carotid artery manipulated in any way.

Statistics

Values of right carotid artery blood flow determined by flow probe and CBF determined by microspheres were correlated independently for the CCABF condition (both right external and internal carotid circulations patent) and ICABF condition (right external carotid artery ligated; right internal carotid artery patent). For each condition, the Pearson product moment correlation (r value) was determined, as were the mean and SEM of the slope, the SEM of the estimate, and the y intercept. Comparisons of slopes in the two conditions and of the y intercept with the origin were performed with t test for unpaired samples(21).

RESULTS

Figure 1 displays a representative tracing of phasic signal from a flow probe placed on the right common carotid artery during normoxia/eucarbia before and after ligation of the ipsilateral right external carotid artery. Reliability of the mechanical zero and electrical zero procedures are shown before ligation of the external carotid artery (a and b) and after ligation (d and e). The flow probe signal fell promptly upon ligation of the external carotid artery (c), reaching a new steady state value reflecting ICABF that was approximately 40% of the previous value of CCABF.

Figure 1
figure 2

Representative tracing of phasic signal from flow probe placed on the right common carotid artery during normoxia/eucarbia before and after ligation of ipsilateral right external carotid artery: a, mechanical zero before ligation; b, electrical zero before ligation; c, external carotid ligation; d, mechanical zero after ligation; e, electrical zero after ligation. Flow probe signal fell promptly upon ligation, reaching a new steady state value for ICABF at approximately 40% of the previous level for CCABF.

Full size image

Figure 2 displays a representative tracing of the signal from a flow probe placed upon the right common carotid artery (in the ICABF condition) during the switch from eucarbia to hypercarbia. The signal responded within 1 min of the change in PCO2 and reached a new steady state value within 7.5 min (not shown).

Figure 3 displays the correlation of unilateral CCABF and ICABF (determined by flow probe) with unilateral CBF (determined by microspheres). Both correlations are highly significant (r = 0.94 and 0.92, respectively; both p < 0.001), suggesting that 88 and 85% of the variance in flow probe signal is accounted for by variations in true CBF, respectively. The slope of the CCABF correlation [1.68 ± 0.19 (SEM)] differs significantly from unity, suggesting that CCABF overestimated CBF by approximately 68% over this range of CBF. The slope of the ICABF versus CBF correlation [1.06 ± 0.15 (SEM)] does not differ significantly from unity, and the y intercept [-1.3 ± 5.1 (SEM) mL/min] does not differ significantly from zero, suggesting that flow probe measurements of unilateral ICABF accurately reflected microsphere measurements of unilateral CBF under these conditions.

Figure 3
figure 3

Correlations of unilateral right-sided CCABF and ICABF (determined by flow probe) vs unilateral CBF (determined by microspheres). CCABF vs CBF (open circles): Regression equation is y = 1.68× + 4.4 mL/min; SEM of the slope = 0.19; SEM of the estimate = 9.64; r = 0.94, p < 0.001. The slope differs significantly from unity, suggesting that unilateral CCABF overestimates unilateral CBF by 68% over this range of CBF. ICABF vs CBF (closed squares): Regression equation is y = 1.06× - 1.3 mL/min; SEM of the slope = 0.15; SEM of the estimate = 7.09; r = 0.92, p < 0.001. The slope does not differ significantly from unity, nor does the y intercept differ significantly from zero, suggesting that unilateral ICABF accurately reflects unilateral CBF over this range of CBF.

Full size image

Figure 4 displays the dependence of unilateral CCABF and ICABF (determined by flow probe) versus PaCO2. The regression of ICABF versus PCO2 was steeper and more tightly correlated than CCABF versus PCO2 (r = 0.94, p < 0.001 versus r = 0.68, p < 0.020).

Figure 4
figure 4

Correlation of unilateral CCABF and ICABF (determined by flow probe) vs PCO2. CCABF vs PCO2 (open circles): Regression equation is y = 0.56x + 18.7 mL·min-1·torr-1; SEM of the slope = 0.20; SEM of the estimate = 13.61; r = 0.68, p < 0.02. ICABF vs PCO2 (closed squares): Regression equation is y = 0.76x - 10.0 mL·min-1·torr-1; SEM of slope = 0.10; SEM of the estimate = 4.87; r = 0.94, p < 0.001.

Full size image

Figure 5 presents the correlation of right versus left cerebral hemisphere blood flow (determined by microspheres) before and after ligation of the right external carotid artery. Both correlations are highly significant (p < 0.001) with r values of 0.99. There is no significant difference in the slopes of the two lines, suggesting that ligating the right external carotid artery as described in the ICABF protocol did not affect the hemispheric symmetry of CBF. Similarly, total CBF (determined by microspheres) did not change significantly on the basis of values before and after ligation of the external carotid artery (ΔCBF = 5.9 ± 8.7 mL/min, p = NS; %ΔCBF = 7.3 ± 11.2%, p = NS), suggesting that ligation of the right external carotid artery did not quantitatively affect CBF.

Figure 5
figure 5

Correlation of right vs left hemisphere CBF (both determined by microspheres) before and after ligation of right external carotid artery. Before ligation (open circles): Regression equation is y = 0.86x + 12.1; SEM of the slope = 0.05; SEM of the estimate = 8.07; r = 0.99, p < 0.001. After ligation(closed squares): Regression equation is y = 0.81x + 17.5; SEM of the slope = 0.04; SEM of the estimate = 7.26; r = 0.99, p < 0.001. The two slopes do not differ significantly from each other, suggesting that ligation of the right external carotid artery does not affect the hemispheric symmetry of CBF.

Full size image

The Table displays systolic blood pressure and blood gas values at normoxia/eucarbia, hypercarbia, and hypoxia. During the elevated CBF conditions, ligation of the right external carotid artery significantly reduced right carotid artery blood flow to 63 ± 8% of its previous value. During the 2-min microsphere infusion, neither blood pressure (0.1 ± 11.4 mm Hg), heart rate (-2.0 ± 11.2 bpm), nor carotid artery flow probe signal (-1.3 ± 4.3 mL/min) changed significantly. The average value for 21 postmortem absolute zero flow probe signals was 1.5 ± 0.5 mL/min, equivalent to 4.8 ± 1.7% of the ICABF signal immediately before the animal was killed. CBF during normoxia/eucarbia was 91 ± 29 mL/min per 100 g. Wet brain weight was 43.7 ± 3 g (SD), representing 1.05 ± 0.11% of total body weight.

Table 1 Hemodynamic and blood gas values
Full size table

DISCUSSION

We have presented correlations of continuous on-line flow probe measurements of unilateral (right) CCABF with CBF determined by radioactive-labeled microspheres under two experimental conditions: 1) with both the right internal and external carotid circulations intact (CCABF), and 2) with the right external carotid artery ligated and the subsequent right CCABF reflecting flow only through the right internal carotid artery (ICABF). There was no manipulation or measurement of left-sided carotid blood flow during any of these protocols. Several findings are apparent from these studies.

First, under both experimental conditions, flow probe measurements and microsphere measurements were tightly and linearly correlated. Not only were the correlations highly significant (p < 0.001) but also the regression coefficients were high (0.90-0.94). Next, in absolute terms, flow probe measurements in the unilateral CCABF condition seemed to overestimate microsphere determinations of unilateral CBF by approximately 68%. This overestimation likely reflects blood flowing through the right common carotid artery to the right external carotid artery to the scalp, face, and snout. In contrast, after ligation of the right external carotid artery, flow probe measurements of unilateral ICABF accurately reflected microsphere determinations of unilateral CBF on a milliliter-to-milliliter basis.

The values of CBF during normoxia/eucarbia, expressed either as mL/min (39 ± 5) or normalized to brain weight (91 ± 29 mL/min per 100 g), are comparable to those reported in piglets by others using a variety of experimental protocols and anesthetic techniques(15,22–28). The rise in CBF induced by hypercarbia reported here agrees well with previously published findings for comparable elevations of PCO2 in other experimental settings, expressed either as percentage rise per torr [3.7%/torr reported here versus 2.0-3.6%/torr in lambs(29), 1.5%/torr in rabbits(17,18), 2.9-3.6%/torr in baboons(30), and 3%/torr in human infants(31)] or as absolute increase in flow/min per 100 g brain tissue per torr [2.4 mL·100 g-1·min-1·torr-1 noted here compared with 2.0 mL·100 g-1·min-1·torr-1 in lambs(32) and 1.3 mL·min-1·100 g-1·torr-1 in humans(33)].

One example of a potential advantage of the flow probe technique reported here devolves directly from considerations of the PCO2 responsiveness of the cerebral circulation. By use of flow probes, the time course of the change in CBF can be readily determined (cf. Fig. 2). During the eucarbia-to-hypercarbia protocol, the flow probe signal revealed a change in CBF within 1 min of the change in the inhaled gas mixture. This value represents an outside limit on the reactivity of the cerebral circulation to PCO2, as the ventilation protocol used here was not designed to effect extremely rapid changes (i.e. within seconds) of arterial blood gases. Elsewhere, using a specially designed ventilator circuit, we have increased PaCO2 from 40 to 70 torr with a half-time of 17 s and have determined that changes in CBF occur within 30 s of changes in PCO2(34), consistent with other more indirect measures of CBF responsiveness to changes in PCO2 [e.g. changes in rate of venous effluent(35)]. CBF responses within this time frame are necessarily invisible when viewed by microsphere methodology.

In some species (e.g. sheep, ox, goat, cat), blood flow from the common carotid artery passes almost entirely to the external carotid circulation and from there to both the face and brain via the internal maxillary artery(8–10). These animals have essentially no ICABF under physiologic circumstances. Consequently, in these species, it seems inappropriate on anatomic grounds to attempt the ICABF/CCABF distinction drawn here. We note, however, that a flow probe protocol for determination of CBF has been successfully accomplished in the goat by use of extensive ablation of the proximal external carotid circulation and a flow probe placed distally on the carotid artery(6,7). Alternatively, in sheep, CCABF has been correlated significantly (but weakly when compared with the ICABF correlations reported here) with CBF(36–38). In contrast, Dunnihoo and Quilligan(39), also working with sheep, were unable to show any significant correlation of CCABF with CBF.

In other species (e.g. man, rhesus monkey, horse, pig), the internal carotid artery dominates the afferent cerebral circulation(8,10–12). In ponies, Orr et al.(12) have taken advantage of this configuration and have developed a technique to assess CBF by placing a flow probe directly on the internal carotid artery. In piglets, approximately one twentieth the size of a pony, direct placement of a flow probe on the internal carotid artery proved impractical.

However, we were able to take advantage of three propitious anatomic features found in piglets: 1) the external carotid artery is easy to identify and ligate at its origin, just distal to the carotid bifurcation; 2) the vertebral artery system is almost entirely limited to supplying extracranial structures, particularly the cervical spinal cord(19); and 3) pigs normally receive almost all CBF from the internal carotid arteries. Daniel et al.(8) have determined that "… the circle of Willis derives its blood supply almost entirely from the ascending pharyngeal artery [called here the internal carotidartery] via the intracranial carotid rete. The arteria anastomotica and ramus anastomoticus [arterial connections between the external and internal carotid circulations] were found to be insignificant vessels, and thus the contribution to the circle of Willis provided by the internal maxillary artery [derived from the external carotid circulation, and the major source of external carotid blood to the brain in the other species studied] would appear to be negligible."

In sum, in piglets, we have correlated flow probe determinations of right CCABF with microsphere determinations of unilateral CBF under conditions in which the entire right-sided carotid artery circulation (i.e. both external and internal) was intact (CCABF) and in which the right external carotid circulation was ligated and right-sided CCABF reflected flow only through the right internal carotid artery (ICABF). Both CCABF and ICABF correlated well with CBF, and ICABF but not CCABF provided an accurate estimate of absolute CBF. We suspect that these observations will be useful to investigators who are interested in determining dynamic aspects of vascular control of the cerebral circulation that have heretofore proven difficult because of the limited number of intermittent samples available with the microsphere technique.

Acknowledgments. The authors thank the editor and reviewers for their contributions to this work.