A Method for Differential Delivery of Intravenous Drugs to the Head and Torso of the Goat

We previously developed a model for differential delivery of inhaled anesthetics to the in situ goat brain ^[1,2] to investigate the involvement of the brain and spinal cord in anesthetic action. This model exploits the unique anatomy of the goat, which has two external carotid arteries and two external jugular veins; there are no internal jugular veins or extracranial internal carotid arteries ^[3,4]. The vertebral arteries do not normally contribute to the cerebral circulation. Cranial venous blood is drained into a bubble oxygenator and infused back into a carotid artery, thus permitting selective and differential delivery of inhaled anesthetics ^[1,2]. Although some venous admixture can occur with this experimental preparation, this is not problematic with inhaled anesthetics. Small amounts of inhaled anesthetic in blood that crosses from the torso to the head are eliminated via the oxygenator exhaust, and any inhaled anesthetic in blood that crosses over from the head to the torso is exhaled via the lungs ^[1]. We previously documented that an IV drug (metocurine) administered to the torso accumulates in the head and bypass unit ^[1]. Thus, the preparation, as originally developed, is not optimal for differential delivery of IV drugs. We have modified the preparation by ligating the neck muscle and tissue. We hypothesized that this would minimize the amount of venous admixture between the head and torso during bypass, thereby permitting differential delivery of an IV anesthetic.

Methods

Our institutional animal care and use committee approved this study. Six female goats (52 +/- 17 kg) were anesthetized with isoflurane via a mask, and their tracheas were intubated. A rumen tube was passed via the esophagus to drain rumen contents. Neck dissections were performed to isolate the carotid arteries and jugular veins. The occipital artery, which is an anastomosis between the carotid and vertebral arteries, was ligated on both sides. We modified the surgical preparation by ligating the tissue and musculature of the neck. This was accomplished by passing two strings through the neck wounds and tying them tightly around the posterior and anterior neck tissues. The string around the posterior neck tissues included the spinal column. Umbilical tapes were also passed around the trachea and esophagus (with the endotracheal and rumen tubes in situ) and tightly secured to eliminate blood flow and to further minimize contamination. Only the jugular veins and carotid arteries were excluded from these ties.

A peripheral IV catheter was placed for the administration of fluids, and a carotid artery catheter (directed toward the heart) was placed for measurement of systemic blood pressure and withdrawal of blood to ensure normal hematocrit, glucose, acid-base balance, and blood gases. A catheter (18-gauge) was inserted into the other carotid artery and directed toward the head, thereby permitting measurement of cranial blood pressure during bypass. After the administration of heparin (4 mg/kg), two Y cannulae were placed into each jugular vein, and a single cannula was placed into a carotid artery. The Y cannulae permitted diversion of cranial venous outflow either to the torso or a bubble oxygenator ^[1,2]. Blood (500 mL) was withdrawn from the goat and drained into a bubble oxygenator (Bentley B-10Plus; Baxter Healthcare, Irvine, CA). Gas flow to the oxygenator was O₂ 95%/CO₂ 5% at 4-5L/min, which maintained normal pH, PCO₂, and PO₂ >400 mm Hg. A vaporizer filled with isoflurane was placed in line with the gas flow to the oxygenator. Cranial venous blood was diverted from the head into the oxygenator, and clamps were placed on the venous cannulae going into the body. A clamp was also placed on the carotid artery that was perfusing the head via the systemic circulation. Cranial bypass was initiated with flows at 500-600 mL/min. Bypass flows were stabilized by adjusting pump flow, reservoir height, and partial occlusion of the venous return as needed. During bypass, torso (systemic) and cranial blood pressures were 96 +/- 26 mm Hg and 43 +/- 13 mm Hg, respectively. Torso (rectal) and cranial (nasopharyngeal) temperatures were 37.5 +/- 1.0[degree sign]C and 37.3 +/- 0.9[degree sign]C, respectively. Glucose was infused (10-20 mg/min) into the bypass unit to maintain euglycemia. Isoflurane was maintained at 1.3% +/- 0.1% to both head and torso. Isoflurane concentrations were determined from analysis of the oxygenator exhaust and end-tidal samples from the lungs using a calibrated anesthetic analyzer.

Propofol (4 mg/kg) was injected into the peripheral IV catheter, and blood samples were withdrawn from the carotid arterial catheter in the torso and from the arterial and venous limbs of the bypass unit 0, 1, 5, 10, and 15 min after injection. When needed, systemic blood pressure was maintained by intermittent injections of phenylephrine (100 [micro sign]g).

To facilitate the detection of admixture, a second propofol injection (4 mg/kg) was made in two goats approximately 45-60 min after the first injection, and an infusion was started at 200 [micro sign]g [center dot] kg^-1 [center dot] min^-1. Blood was withdrawn from the torso and bypass unit at 0, 5, 15, and 30 min, after which the infusion was stopped. Forty-five minutes after completing the infusion and 45 min after the first injection in the remaining four goats, propofol (0.2 mg/kg) was administered via the venous limb of the bypass circuit. Blood samples from the torso (systemic arterial blood) and arterial/venous limbs of the bypass unit were withdrawn at 0, 1, 5, 10, and 15 min. The animal was then killed with isoflurane and potassium chloride.

The blood samples were analyzed for propofol using high-pressure liquid chromatography with a protocol modified from two previous studies. ^[5,6]. Briefly, after centrifuging, the plasma was stored at -70 [degree sign]C until analysis. Propofol was extracted from the plasma using solid-phase extraction and injected onto a Waters high-pressure liquid chromatography column with electrochemical detection. Peaks were compared with standard curves, which demonstrated high correlation coefficients (r = 0.98-1.0). The lower limit of detection was 25 ng/mL. Propofol concentrations in the systemic arterial, arterial bypass, and venous bypass blood were compared by using analysis of variance to detect overall differences, followed by t-tests. P < 0.05 was considered significant.

Results

Due to technical problems, propofol concentrations in the torso after injection were not obtained in one animal and at 1 min in two other animals. When it was injected into the bypass unit, propofol was not detected in the torso at any time (Figure 1A). Peak propofol concentrations occurred at 1 min; at 10 and 15 min, there were still small but measurable amounts of propofol in the bypass unit.

Figure 1: A, Propofol plasma concentrations in the arterial and venous limbs of the bypass unit and torso after propofol injection (0.2 mg/kg) into the bypass unit. Mean +/- SD. The propofol concentration in the torso never exceeds zero. The average peak concentration exceeds that after injection into the torso (see B); at 10 and 15 min, there are still low but detectable concentrations. *P < 0.05 propofol concentration in arterial-bypass blood compared with propofol concentration in torso (systemic) arterial blood. B, Propofol plasma concentrations in the torso and arterial and venous limbs of the bypass unit after propofol injection (4 mg/kg) into the torso. The propofol concentration in the bypass unit (and, hence, the head) never achieves clinically significant levels. *P < 0.05 propofol concentration in torso (systemic) arterial blood compared with propofol concentration in arterial-bypass blood. The rapid decrease in propofol concentration in the torso is consistent with the pharmacokinetic characteristics of propofol.

When propofol was injected into the torso, only trace amounts were detected in the bypass unit. Peak propofol concentrations occurred at 1 min; at 10 min, propofol concentrations had decreased to near zero (Figure 1B). The peak propofol concentrations in the torso (systemic) circulation after torso injection were half of those in the arterial limb of the bypass unit after injection into the bypass unit (Figure 1A). During propofol infusion to the torso in two goats, there was minimal cross-over to the bypass unit, as seen in Figure 2. In these goats, the propofol concentration in the arterial limb of the bypass unit was 7% +/- 2% of that in the torso (systemic) arterial blood.

Figure 2: Propofol plasma concentrations in two goats that received propofol (4 mg/kg) to the torso followed by 200 [micro sign]g [center dot] kg^-1 [center dot] min^-1. The propofol concentrations in the head (all times combined and averaged) are 7% +/- 2% of those in the torso.

Discussion

Our data show that IV drugs (specifically propofol) can be administered selectively to the head and torso with minimal mixing between the respective circulations. We chose propofol because it is representative of commonly used anesthetic drugs (e.g., rapid onset, rapid distribution phase). We were able to occlude most of the collateral circulation between the head and torso, including venous collaterals through muscle. This desirable additional step adds little time to the surgical preparation but does not occlude the venous plexus that surrounds the spinal cord. Because this plexus is valveless, blood can flow in either direction and therefore represents a possible source of mixture. In our previous study ^[1], we found that systemic arterial blood perfused the caudal brainstem or upper cervical cord, with the level depending on the blood pressure difference between the systemic (torso) and cranial circulations. In the present study, systemic blood pressure greatly exceeded cranial blood pressure; therefore, it is unlikely that any cranial arterial blood perfused the spinal cord via collateral arteries (e.g., the basilar artery).

This preparation is particularly good for drugs administered to the head. If a small amount of drug passed from the head to the torso, significant blood levels would not result. Cross-contamination is more likely with torso injection of a drug, but with a drug that is quickly eliminated and distributed (e.g., propofol), there is minimal opportunity for cross-over. However, infusing propofol in two goats maintained clinically relevant blood concentrations for 30 min but did not appreciably alter the amount of mixture (Figure 2). Thus, we believe that drugs that have a slower distribution phase than propofol and maintain significant blood concentrations for a longer time could be successfully used in this preparation.

The peak and decline of the propofol concentration determined in the present study are similar to those found by Reid et al. ^[7]. Because our primary interest was to determine the degree of mixture, we did not draw blood samples at times that would permit rigorous pharmacokinetic analysis. To match the concentration profile seen in the torso, a propofol dose smaller than the one used should be injected into the bypass unit.

Although the amount of torso-head contamination is small and likely occurs primarily via the spinal cord venous plexus, drug transfer from torso to head and vice versa is limited when the distribution half-life is brief. Drugs with longer half-lives may be more likely to cross-over. In any case, with this preparation, measurement of blood concentrations is necessary to ensure comparable torso and head drug concentrations. In summary, because of the rapid distribution characteristic of many anesthesia-related drugs, this preparation can be used to investigate the pharmacology of anesthetic action in the brain and spinal cord.