Share this article on:

The Effects of Lidocaine and Adrenergic Agonists on Rat Sciatic Nerve and Skeletal Muscle Blood Flow In Vivo

Palmer, Greta M. MBBS, FANZCA, FFPMANZCA*,; Cairns, Brian E. PhD*,; Berkes, Steven L. MD*,; Dunning, Patricia S. BSc, RT(R)†,; Taylor, George A. MD†,; Berde, Charles B. MD, PhD*

doi: 10.1213/00000539-200210000-00054

It has been proposed that epinephrine prolongs lidocaine nerve blockade duration by exerting a local vasoconstrictive effect on tissues at the injection site, slowing lidocaine’s local clearance. However, previous studies have failed to demonstrate consistent effects of lidocaine and epinephrine, injected alone and in combination, on vascular tone or regional blood flow. To reinvestigate this idea, in this study we used the radiolabeled microsphere technique to measure in vivo tissue blood flow before and at several time points after perisciatic nerve and intramasseter muscle injection of lidocaine alone, epinephrine, the selective α1-adrenergic receptor agonist phenylephrine, or lidocaine combined with these adrenergic receptor agonists. Repeated-measures analyses of variance were used to assess significant changes in blood flow over time. Lidocaine (2, 10, and 20 mg/mL) and epinephrine (10 μg/mL or 1:100,000) injected alone did not alter blood flow in sciatic nerve, perisciatic muscle, or masseter muscle. Injections of lidocaine (10 mg/mL) combined with epinephrine (10 μg/mL) did not affect adjacent muscle blood flow but caused a mild reduction in sciatic nerve blood flow, which was significant 30 min after injection. However, phenylephrine (10 μg/mL), a potent vasoconstrictor, combined with lidocaine (10 mg/mL) significantly reduced blood flow in all three tissues. Our findings suggest that mechanisms other than local vasoconstriction may contribute to the prolongation of lidocaine nerve blocks by epinephrine.

Departments of *Anesthesia and †Radiology, Children’s Hospital, and Harvard Medical School, Boston, Massachusetts

Supported by Purdue Pharma and the Boston Children’s Hospital’s Anesthesia/Pain Research Endowment Fund.

June 19, 2002.

Address correspondence and reprint requests to Charles B. Berde, MD, PhD, Pain Treatment Service, Children’s Hospital, 333 Longwood Ave., 5th Floor, Boston, MA 02115. Address e-mail to

Epinephrine prolongs lidocaine nerve block duration in concentrations of 1.25–20 μg/mL (1,2). It is thought that the mechanism for prolongation of anesthetic block is due to epinephrine’s vasoconstrictive action. Evidence to support this mechanism includes blood flow reduction seen when epinephrine is used alone (3,4) and in combination with lidocaine (1,3–6), as well as decreased lidocaine clearance and reduced systemic lidocaine levels when it is coadministered with epinephrine (6–8). These studies support the aforementioned proposal that epinephrine prolongs lidocaine neural blockade duration by local vasoconstriction of injection site tissues, which slows the local clearance of lidocaine and provides a larger concentration gradient, facilitating diffusion of lidocaine from its deposition site into the nerves. Nevertheless, in prior studies using laser Doppler blood velocity assessment, epinephrine, when combined with lidocaine, was not consistently associated with vasoconstriction (3,4). In addition, conflicting results have been reported for the effect of lidocaine on vascular tone, nerve, and other tissue blood flows. With different blood flow measurement techniques, these studies assessed the effect of varying volumes of very small to clinically used lidocaine concentrations (e.g., 0.0001% to 2%) (1,3–5,9) after either surgical exposure (3,4,10) or sectioning of vessels (9). For lidocaine 10 mg/mL (or 1.0%), some authors found vasoconstriction of forearm skin (9) and sciatic nerve vasculature (4), whereas others reported vasodilation of the cremaster arterioles (10) and abdominal skin vessels (1) and a biphasic response of the sciatic nerve vasculature (3).

Given the inconsistent results of the previous studies, we decided to reinvestigate the effects of clinically used concentrations of lidocaine, with and without adrenergic agonists, to investigate the hypotheses that 1) lidocaine alone causes neither vasoconstriction nor vasodilation of the neural sheath or adjacent skeletal muscle, which serves as a drug deposition site, but that 2) co-injection of epinephrine with lidocaine causes vasoconstriction in these tissues. To decrease the potential confounds associated with surgical exposure of these tissues, which is required for laser Doppler and videomicroscopy, and to study blood flow in multiple tissues simultaneously, we measured tissue blood flow with the radiolabeled microsphere technique in rats anesthetized with isoflurane. The effect of perisciatic nerve injection was studied, because this large peripheral nerve is reliably blocked in the rat (2,11–13) and has been used in prior studies of blood flow effects (3,4). In addition, in humans, it is a nerve blocked during general anesthesia for lower-limb surgery. However, because one study suggested that isoflurane may alter skeletal muscle blood flow (14), we also examined the effect of intramasseter injection, because this muscle’s blood flow is unaffected during isoflurane anesthesia (14).

Back to Top | Article Outline


The Boston Children’s Hospital Animal Care and Use Committee approved all surgeries and procedures. Seventy adult male Sprague-Dawley rats weighing 400–500 g were studied. Rats were anesthetized with isoflurane(1.8%–3.0% in oxygen). A 3F (internal/external diameter, 0.56:1.0 mm) was introduced into the right carotid artery, and arterial blood pressure was monitored (Model 76; Hewlett-Packard, Palo Alto, CA). This catheter was advanced through the carotid artery into the left ventricle and used to deliver radiolabeled microspheres for systemic distribution according to tissue blood flow. Placement of the catheter within the left ventricle was confirmed by the observation of a typical ventricular blood pressure wave form (Fig. 1). A second polyethylene catheter (internal/external diameter, 0.58:0.965 mm) was inserted into the right femoral artery, and the arterial blood pressure was monitored. This catheter was used to collect blood during radiolabeled microsphere injections. After surgical tracheotomy, a shortened 14-gauge IV cannula was secured in the trachea, through which the lungs were ventilated (rodent ventilator SAR-830/P; Argmore, PA). Ventilation was adjusted to produce an end-tidal CO2 of 28 mm Hg (range, 22–34 mm Hg over all experiments). This target value for end-tidal CO2 was chosen, on the basis of pilot experiments examining the range of end-tidal to arterial differences by using our experimental setup, to give arterial CO2 values in the range of 35–40 mm Hg. Throughout the experiment, rectal temperature was maintained in the range of 36.5°C–37.5°C, and mean arterial blood pressures of 60–100 mm Hg and heart rates of 280–330bpm were maintained through isoflurane titration and fluid supplementation. Removed blood samples were replaced in a 3:1 ratio with heparinized (4 U/mL) normal saline flush of the arterial lines.

Figure 1

Figure 1

Rats were divided into nine groups:

  1. Normal Saline Control (n = 6).
  2. 2, 3, and 4. Lidocaine (Astra Pharmaceuticals, Westborough, MA; pH tested 6.5) 2, 10, and 20 mg/mL (respectively 0.2%, 1%, and 2%;n = 5, 6,and 6).
  3. 5 and 6. Phenylephrine (Astra Pharmaceuticals) 10 and 50 μg/mL (n = 7 and 6, respectively).
  4. Epinephrine (Abbott Laboratories, North Chicago, IL) 10 μg/mL (n = 6).
  5. Lidocaine 10 mg/mL combined with phenylephrine 10 μg/mL (n = 6).
  6. Lidocaine 10 mg/mL combined with epinephrine 10 μg/mL (n = 6).

Drug dilutions were made with normal saline 0.9% immediately before injection. Injectates were dyed with fluorescein for identification of injection site and spread at dissection. For intramasseter injection, a 25-gauge 1-in. needle with a 1-mL syringe was inserted percutaneously into the belly of the masseter muscle, and 0.1 mL of solution was injected—drug into the left masseter and normal saline into the right masseter (control). For left perisciatic nerve injection, a 25-gauge 1-in. needle with a 1-mL syringe was inserted percutaneously, from the posterior aspect of the upper leg, and advanced anterolaterally toward the greater trochanter. On bony contact, the needle was withdrawn 1 mm, and then 0.1 mL of drug was injected.

The radiolabeled microsphere technique (15,16) was used with 15 ± 0.1 μm glass microspheres labeled separately with 5 isotopes—146Ce, 51Cr, 103Ru, 95Nb, and 46Sc—suspended in normal saline with the surfactant Tween-80 0.01% (Perkin Elmer Life Sciences, Boston, MA) and continually ultrasonically agitated, until immediately before injection, to prevent aggregation. These isotopes each have distinct gamma emission spectra, which enabled blood flow calculation for each tissue bed at five serial time points.

Solutions of one of the different isotopes (10 μCi, 0.2 mL) were administered through the left ventricular catheter at the following times: baseline (immediately before perisciatic and masseter drug injection at Time 0) and then 5, 15, 30, and 60 min after drug injections. Beginning 10 s before the microsphere injection and continuing for 75 s, a reference blood sample (∼1 mL) was withdrawn from the femoral catheter at a rate of 0.6 mL/min (Harvard Apparatus pump). At the completion of the experiment, the animals were killed with injection of pentobarbital through the intracardiac catheter (100 mg/kg, Euthasol; Diamond Animal Health, Des Moines, IA). Then, the fluorescein-labeled segments of muscle adjacent to the sciatic nerve (including the pyriformis, deep portions of the gluteus, and origins of the biceps femoris or adductor muscles), masseter muscle, and sciatic nerve, as well as abdominal wall muscle, brain, and kidneys, were harvested. The reference blood samples and harvested tissues and organs were weighed and placed in 5-mL vials for analysis of radioisotope content in an automatic gamma-counting system (Cobra II, Model D5003; Packard Instrument Co., Downers Grove, IL). Tissue blood flows were calculated with the following equation:MATH

Calculated blood flow values on the left (injected) side were subtracted from the right (control) side values to give a side-to-side difference in blood flow. This was done to compensate for any potential systemic fluctuations in blood flow. Significant alterations in the side-to-side difference in blood flow over time were determined with a one-way repeated analysis of variance. The minimum sample size required for each group of rats was estimated to be five, on the basis of a blood flow change of at least 30% ± 15%, a power of 0.8, and an α of 0.05. Where significance differences were found, a post hoc analysis using Dunnett’s method was applied. A P value of <0.05 was considered significant. The Pearson product-moment correlation was used to determine significant relationships between pairs of tissue blood flows.

Back to Top | Article Outline


Baseline blood flows for different tissues were tabulated (Table 1). Flows in peripheral muscular tissue, such as masseter, leg, and abdominal wall, were significantly lower than flows in the rat’s brain, kidney, and sciatic nerve. A side-to-side difference was apparent, with baseline right-sided blood flow being reduced by 39.7% for the sciatic nerve, 35.5% for the masseter, and 11% for the adjacent muscle, presumably because of carotid and femoral arterial catheterization.

Table 1

Table 1

Linear regression analyses of right-sided blood flow (control) were undertaken to determine whether blood flow changes in sciatic nerve or masseter muscle linearly correlated with brain or kidney blood flow changes. Changes in masseter muscle blood flow exhibited a significant linear relationship with changes in both kidney (r = 0.31) and brain (r = 0.62) blood flow. In contrast, sciatic nerve blood flow failed to correlate with kidney and only weakly correlated with brain blood flow (Fig. 2), and adductor muscle blood flow showed no correlation with either kidney or brain blood flow. As a criterion for physiologic stability, a ≥25% reduction in brain blood flow at any time point resulted in exclusion of the data for that time point from analysis.

Figure 2

Figure 2

Injection of saline had no significant effect on side-to-side differences in blood flow in the sciatic nerve or the perisciatic or masseter muscles. Injection of lidocaine at all three concentrations was also without significant effect on these tissues’ blood flow (Fig. 3).

Figure 3

Figure 3

Epinephrine 10 μg/mL alone had no consistent effect on side-to-side differences in sciatic nerve or perisciatic or masseter muscle blood flow. Phenylephrine 10 μg/mL caused no significant change in sciatic nerve or perisciatic muscle blood flow but did significantly decrease side-to-side differences in blood flow in the masseter muscle from 5 to 30 min. Phenylephrine 50 μg/mL significantly reduced side-to-side differences in masseter muscle and sciatic nerve blood flow for 60 min, with no significant effect on adjacent leg muscle (Fig. 4).

Figure 4

Figure 4

Epinephrine 10 μg/mL in combination with lidocaine 10 mg/mL had no effect on side-to-side differences in masseter or perisciatic muscle blood flow. However, there was a small but significant reduction in side-to-side differences in sciatic nerve blood flow at 30 min postinjection. Phenylephrine 10 μg/mL combined with lidocaine 10 mg/mL significantly reduced side-to-side differences in blood flows at 15 min in all three tissues (Figure 5).

Figure 5

Figure 5

Back to Top | Article Outline


In this study, we found that clinically-used concentrations of lidocaine alone did not significantly alter injection site tissue blood flow. We were surprised to find that the mixed α/β-adrenergic receptor agonist epinephrine, alone or in combination with lidocaine, administered in a concentration often used to enhance the duration of lidocaine blocks, also did not significantly alter tissue blood flow at the injection site (muscle). As a result of this unexpected finding, we investigated the effect of phenylephrine, a selective α1-adrenergic receptor agonist with potent vasoconstrictive actions, alone and in combination with lidocaine. Phenylephrine did produce a dose-dependent decrease in injection site nerve and muscle blood flow. Therefore, our results do not support the hypothesis that clinically relevant concentrations of epinephrine, administered either alone or with lidocaine, prolong neural blockade by a local vasoconstrictive mechanism in the skeletal muscle surrounding the injection site. Conversely, there was a small, but significant, decrease in sciatic nerve blood flow when epinephrine was combined with lidocaine. This finding supports the idea that vasoconstriction of the epineural tissues by the combination of epinephrine and lidocaine may contribute to an increase in intraneural lidocaine content and the prolongation of nerve block associated with this combination, possibly by decreasing local anesthetic clearance from the nerve.

The radiolabeled microsphere technique offers advantages over some other techniques because it measures regional blood flow in multiple tissues simultaneously and allows multiple repeated measurements in each tissue sample. In addition, this technique permits retrospective confirmation of systemic physiologic stability through determination of blood flow in the vital organs. However, our own results, as well as those of previous studies, indicate that the intrinsic variability between successive blood flow measurements is approximately 15%, which serves to limit the sensitivity of the microsphere technique to the measurement of relatively large, i.e., 30% or more, blood flow changes (17). Rats were studied, despite the increased technical challenge of catheterization and maintenance of homeostasis compared with larger animals, because most behavioral, electrophysiological, and drug uptake/nerve content studies have been performed in this species (2,11–13).

There are, however, several possible shortcomings associated with the use of the radiolabeled microsphere blood flow measurement technique in rats. Blood loss during line placement or blood sampling must be minimized to limit reductions in hematocrit and circulating blood volume, which can alter hemodynamics. Although crystalloid replacement can maintain blood volume and preload, the associated decrease in hematocrit and blood viscosity can alter peripheral blood flow dynamics. In addition, a change in blood oxygen content because decreased hematocrit will lead to compensatory reflex vasodilation to maintain oxygen delivery, with a reduction in peripheral vascular resistance (18). Support for the physiologic stability of our preparation is found in the observation that in the saline-injected control animals, tissue blood flows remained nearly constant over time, suggesting that blood loss and hemodilution played a minimal role in our measurements.

The preparation required carotid and femoral arterial cannulation for microsphere injection, arterial blood pressure monitoring, and blood sampling. Cannulation of these vessels per se produced significant baseline differences in right- and left-sided blood flows in masseter muscle, sciatic nerve, and perisciatic muscle. This factor made a direct comparison of side-to-side flows inappropriate. In some cases, blood flow increases were observed on the catheterized (control) side after contralateral drug injection. These may be interpreted as resulting from systemic absorption and distribution of vasoactive drugs. In an effort to minimize artifact due to systemic effects of the injected drugs, the blood flow data were reported as side-to-side differences in blood flow. Statistical analyses tested for statistically significant changes in the difference between right- and left-sided flows over time, before and after drug injections. Because some previous studies (3,4) measured local nerve blood flow without simultaneous measurement of systemic hemodynamics or major organ blood flows, it is conceivable that systemic actions of the drugs were responsible for some of their observed effects.

Masseter muscle injection of lidocaine 10 mg/mL (10 μL) blocks muscle afferent nerve fibers for 5–15 minutes in the rat (19). In awake, behaving rats, sciatic nerve blockade with lidocaine 10 mg/mL (100 μL) alone lasted 35–60 minutes and, with the addition of epinephrine 10 μg/mL, approximately 80 minutes (2,11,12). In this study, we used the same volume (100 μL) and concentration of lidocaine and epinephrine as those used in awake rat studies, but we found no effect on injection site muscle blood flow and found comparatively small effects on sciatic nerve blood flow.

Prior studies of vascular caliber with videomicroscopy (10) and tissue blood flow velocity with laser Doppler (1,3,4) found that clinically used lidocaine concentrations of 5–20 mg/mL (or 0.5%–2%) had inconsistent and often opposite effects on different tissues. For example, in a previous study of cremaster muscle, arteriolar constriction was seen on videomicroscopy with lidocaine ≤1 mg/mL, and vasodilation was seen with topical lidocaine 10 mg/mL (10). In our study, there was no significant change in the perisciatic or masseter muscle blood flow with lidocaine 2, 10, or 20 mg/mL. With laser Doppler assessment of changes in nerve blood flow velocity, topical lidocaine 10 and 20 mg/mL caused dose-dependent decreases in sciatic nerve blood flow velocity by 12%–61%(3,4). No significant effect of lidocaine on the nerve blood flow was demonstrated in our study. Additionally, in the laser Doppler studies, epinephrine 2.5–10 μg/mL alone resulted in 20%–45% decreases in sciatic nerve blood flow velocity (3,4), whereas 1.25–20 μg/mL caused decreases in skin blood flow velocity (1). Skeletal muscle effects were not specifically assessed. Our findings appear to differ from these previous studies: epinephrine 10 μg/mL failed to significantly alter either skeletal muscle or nerve blood flow. After subcutaneous injection (1) or topical administration (3,4) of the combination of epinephrine and lidocaine, laser Doppler assessments of blood flow velocity also varied, with reports of both increases (4) and decreases (1,5). In our study, when epinephrine 10 μg/mL was combined with lidocaine 10 mg/mL, a comparatively small, but statistically significant, reduction in side-to-side difference in nerve blood flow occurred, with no change in skeletal muscle blood flow. Some factors that may contribute to the differences in results between our study and previous studies include

  1. Different type of general anesthetic (e.g., pentobarbital versus isoflurane), with resulting differences in vascular effects.
  2. Vascular reflex responses to surgical exposure (20), which is required for videomicroscopy (10) and laser Doppler (3,4).
  3. Changes in baseline regional blood flow rates and vascular reactivity in the contralateral control sides in our model because of the requirement for indwelling femoral and carotid arterial catheters.

Epinephrine’s combined α-mediated vasoconstriction and β-mediated vasodilation could, in principle, cancel each other, resulting in minimal changes in muscle blood flows. For this reason, we chose the α1-adrenergic receptor agonist phenylephrine to demonstrate vasoconstriction with this technique in the tissues under study. Phenylephrine’s local effect on sciatic nerve and skeletal muscle blood flow has not been reported previously. Phenylephrine, administered as 10 μg/mL (clinically used concentrations, 5 μg/mL to 2.5 mg/mL), showed a significant side-to-side difference in blood flow only in the masseter muscle. Phenylephrine 50 μg/mL caused a significant decrease in both masseter muscle and sciatic nerve blood flow, but it did not significantly alter blood flow in the perisciatic leg muscle. The difference in the two skeletal muscle bed responses could be a result of the significantly higher baseline blood flow in the masseter muscle; in that, low baseline blood flow in the adjacent muscle may have limited detection of small decreases in blood flow. Decreased flow was seen in all three tissues after the injection of phenylephrine combined with lidocaine.

There appears to be a positive correlation between intraneural lidocaine content and functional block duration (11,13). In a study of human subjects, it was found that the coadministration of epinephrine and lidocaine slowed clearance of lidocaine from the superficial peroneal nerve and prolonged nerve block (6). Furthermore, this increase was associated with a 50% decrease in blood flow of the overlying skin, although nerve and injection site blood flow were not assessed in this study (6). Our results, which indicate that the combination of lidocaine and epinephrine significantly decreased rat sciatic nerve blood flow, are compatible with the idea that epinephrine prolongs lidocaine blocks, in part, by vasoconstriction within the neural sheath.

In summary, doses of epinephrine that produce substantial prolongation in lidocaine-induced nerve blockade clinically and experimentally (2) produced only small changes in sciatic nerve blood flow and no demonstrable changes in surrounding muscle blood flow. More substantial reductions in both nerve and muscle blood flow are seen with phenylephrine added to lidocaine. Further studies and combined pharmacokinetic/pharmacodynamic modeling may help to clarify the mechanisms underlying block prolongation by adrenergic receptor agonists.

Back to Top | Article Outline


1. Liu S, Carpenter RL, Chiu AA, et al. Epinephrine prolongs duration of subcutaneous infiltration of local anesthesia in a dose-related manner. Reg Anesth 1995; 20: 378–84.
2. Sinnott CJ, Garfield JM, Thalhammer JG, Stichartz GR. Addition of sodium bicarbonate to lidocaine decreases the duration of peripheral nerve block in the rat. Anesthesiology 2000; 93: 1045–52.
3. Myers RR, Heckman HM. Effects of local anesthesia on nerve blood flow: studies using lidocaine with and without epinephrine. Anesthesiology 1989; 71: 757–62.
4. Partridge BL. The effects of local anesthetics and epinephrine on rat sciatic nerve blood flow. Anesthesiology 1991; 75: 243–51.
5. Evans GRD, Gheradini G, Gurlek A, et al. Drug-induced vasodilation in an in vitro and in vivo study: the effects of nicardipine, papaverine and lidocaine on the rabbit carotid artery. Plast Reconstr Surg 1997; 100: 1475–81.
6. Bernards CM, Kopacz DJ. Effect of epinephrine on lidocaine clearance in vivo: a microdialysis study in humans. Anesthesiology 1999; 91: 962–8.
7. Warner MA, Kunkel SE, Offord KO, et al. The effects of age, epinephrine and operative site on duration of caudal analgesia in pediatric patients. Anesth Analg 1987; 66: 995–8.
8. Mazoit JX, Benhamou D, Veillette Y, Samii K. Clonidine and or adrenaline decrease lignocaine plasma peak concentration after epidural injection. Br J Clin Pharmacol 1996; 42: 242–5.
9. Aps C, Reynolds F. The effect of concentration on vasoactivity of bupivacaine and lignocaine. Br J Anaesth 1976; 48: 1171–4.
10. Johns RA, Difazio CA, Longnecker DE. Lidocaine constricts or dilates rat arterioles in a dose dependent manner. Anesthesiology 1985; 62: 141–4.
11. Popitz-Berger FA, Leeson S, Strichartz GR, Thalhammer JG. Relation of functional deficit and intraneural local anesthetic during peripheral nerve block. Anesthesiology 1995; 83: 583–92.
12. Thalhammer JG, Vladimirova M, Bershadadsky B, Strichartz GR. Neurologic evaluation of the rat during sciatic nerve blockade with lidocaine. Anesthesiology 1995; 82: 1013–25.
13. Popitz-Berger FA, Leeson S, Thalhammer JG, Strichartz GR. Intraneural lidocaine uptake compared with analgesic differences between pregnant and non-pregnant rats. Reg Anesth 1997; 22: 363–71.
14. Goetz TE, Manohar M, Nganwa D, Gustafson R. A study of the effect of isoflurane anesthesia on equine skeletal muscle perfusion. Equine Vet J Suppl 1989; 7: 133–7.
15. Prinzen FW, Bassingthwaighte JB. Blood flow distributions by microsphere deposition methods. Cardiovasc Res 2000; 45: 13–21.
16. Soul JS, Taylor GA, Wypij D, et al. Noninvasive detection of changes in cerebral blood flow by near-infrared spectroscopy in a piglet model of hydrocephalus. Pediatr Res 2000; 48: 445–9.
17. Lescanic ML, Miller ED, Difazio CA. The effects of lidocaine on the whole body distribution of radioactively labeled microspheres in the conscious rat. Anesthesiology 1981; 55: 269–72.
18. Stehling L. Autologous transfusion. In: Miller RD, ed. Anesthesia. 5th ed. Philadelphia: Churchill Livingstone, 2000: 1651.
19. Berde CB, Cairns BE. Glutamate-induced sensitization of masseter muscle afferents is associated with a shortened duration of local anesthesia [abstract]. J Pain 2001; 2S: 34.
20. Sugimoto H, Monafo WM, Eliasson S. Regional sciatic nerve and muscle blood flow in conscious and anesthetized rats. Am J Physiol 1986; 251: H1211–6.
© 2002 International Anesthesia Research Society