Pharmacokinetics of Lidocaine, Bupivacaine, and Levobupivacaine in Plasma and Brain in Awake Rats

Yuko Ikeda, M.D.,* Yutaka Oda, M.D., Ph.D.,† Taketo Nakamura, M.D., Ph.D.,‡ Ryota Takahashi, M.D., Ph.D.,‡ Wakako Miyake, M.D.,* Ichiro Hase, M.D., Ph.D.,‡ Akira Asada, M.D., Ph.D.§ Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/112/6/1396/250464/0000542-201006000-00019.pdf by guest on 01 October 2021

ABSTRACT Conclusions: Although the ratio of total brain concentra- Background: We have compared the pharmacokinetics tion to total plasma concentrations of lidocaine, bupivacaine, and brain distribution of lidocaine, racemic bupivacaine and levobupivacaine was similar, concentration ratio of bu- (bupivacaine), and levobupivacaine in awake, spontane- pivacaine in the cerebral extracellular fluid to plasma un- ously breathing rats. bound fraction was significantly higher than lidocaine and Ϫ Ϫ Methods: Lidocaine (0.5 mg ⅐ kg 1 ⅐ min 1), bupiva- levobupivacaine. Ϫ Ϫ caine (0.1 mg ⅐ kg 1 ⅐ min 1), or levobupivacaine (0.1 Ϫ Ϫ mg ⅐ kg 1 ⅐ min 1) was continuously administered to rats for 2h(nϭ 12, each anesthetic). Blood samples and cerebral dia- What We Already Know about This Topic lysate were collected during infusion and for 2 h after termina- ❖ Although local anesthetics can cause central nervous system toxicity, the distribution of bupivacaine from plasma to brain tion of infusion. Concentrations of anesthetics in the cerebral has been incompletely studied extracellular fluid were measured by microdialysis using the retrodialysis calibration method. Tissue-to-plasma partition coef- What This Article Tells Us That Is New ficients calculated from the total (protein-bound and unbound) ❖ In awake, spontaneously breathing rats receiving intravenous and unbound concentrations in plasma and brain as well as infusion, the ratio of the cerebral extracellular fluid concentra- pharmacokinetic parameters in plasma and cerebral extracellu- tion to the plasma unbound fraction was greater for bupivacaine than for levobupivacaine or lidocaine lar fluid were compared among the three anesthetics. Results: There were no differences in plasma total or unbound concentrations between bupivacaine and levobupivacaine. ENTRAL nervous system (CNS) and cardiovascular Concentrations of bupivacaine in the cerebral extracellular fluid toxicities are life-threatening side effects of local anesthet- Ͻ C were significantly higher than levobupivacaine (P 0.001). ics. Lidocaine and bupivacaine are commonly used for epidural Despite no differences in the ratio of total brain concentration to anesthesia/analgesia and nerve blockade. Bupivacaine exerts total plasma concentration among the three anesthetics, the ra- higher toxicity as well as stronger anesthetic potency compared tio of cerebral extracellular fluid concentration to plasma un- with lidocaine, and differences exist in the toxicity between its bound fraction of bupivacaine was significantly higher than li- dextro and levorotatory stereoisomers.1 Although CNS plays an Ϯ Ϯ docaine and levobupivacaine (0.58 0.09, 0.47 0.18, and important role in inducing cardiovascular toxicity of bupiva- Ϯ ϭ 0.40 0.09, respectively; P 0.03 and 0.003, respectively). caine2 and an increase in the cerebral extracellular fluid (ECF) concentrations across the blood–brain barrier secondary to the * Graduate Student, † Associate Professor, ‡ Assistant Professor, increase in plasma concentrations may contribute to induce § Professor and Chairman, Department of Anesthesiology, Osaka CNS toxicity, relationships between plasma and cerebral extra- City University Graduate School of Medicine, Osaka, Japan. cellular concentrations of bupivacaine are not known. We have Received from Department of Anesthesiology, Osaka City Uni- versity Graduate School of Medicine, Osaka, Japan. Submitted for examined the relationships of the concentrations of intrave- publication July 12, 2009. Accepted for publication December 16, nously administered lidocaine, racemic bupivacaine (bupiva- 2009. Supported, in part, by Grant-in-Aid for Research from the caine), and levobupivacaine between plasma and cerebral ECF Ministry of Education, Culture, Sports, Science and Technology of Japan (Tokyo, Japan) no. 20591813. in awake, spontaneously breathing rats. Address correspondence to Dr. Oda: Department of Anesthe- siology, Osaka City University Graduate School of Medicine, Materials and Methods 1-5-7 Asahimachi, Abeno-ku, Osaka 545-8586, Japan. odayou@ msic.med.osaka-cu.ac.jp. Information on purchasing reprints may be Animal Preparation found at www.anesthesiology.org or on the masthead page at the All the study protocols were approved from the Institutional beginning of this issue. ANESTHESIOLOGY’s articles are made freely ac- cessible to all readers, for personal use only, 6 months from the cover Animal Care and Use Committee (Osaka, Japan). In the first date of the issue. experiment, 36 male Sprague–Dawley rats aged 8–10 weeks

1396 Anesthesiology, V 112 • No 6 • June 2010 PERIOPERATIVE MEDICINE and weighing 350–400 g (CLEA Japan, Inc., Tokyo, Japan) vitro and in vivo using 3-hydroxybupivacaine (0.5 ␮g/ml) for were included. The guide cannulas for microdialysis (GI-12; measuring lidocaine or ropivacaine (0.2 ␮g/ml) for measur- Eicom, Kyoto, Japan) were inserted so that those tips were ing bupivacaine and levobupivacaine as internal standards placed into the nucleus accumbens 2–3 days before experi- following a previously reported method.3 The in vivo K fac- ments following the previously reported method.3 On the tor, defined as the ratio of RL of 3-hydroxybupivacaine Ϯ day of experiments, the carotid artery and the cervical vein (RLOHBUP) to RL of lidocaine (RLLID) was 1.05 0.01, and were cannulated with polyethylene catheters for measuring the ratios of RL of ropivacaine (RLROP) to RL of bupivacaine blood pressure, heart rate, blood sampling, and infusing (RLBUP) and to RL of levobupivacaine (RLLEVO) were drugs, besides the microdialysis probe (A-UI-12-02; Eicom), 1.11 Ϯ 0.05 and 1.09 Ϯ 0.04, respectively (n ϭ 5), and the which was inserted through the guide cannula and was per- mean values were used for calculation. Concentrations of fused with the artificial cerebrospinal fluid containing inter- lidocaine in the cerebral ECF (LID ) were calculated as

ex Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/112/6/1396/250464/0000542-201006000-00019.pdf by guest on 01 October 2021 ϫ nal standards. LIDdialysate K/RLOHBUP, where LIDdialysate is the concentration of lidocaine in the dialysate.6 Concentrations of bu- Experimental Protocol pivacaine and levobupivacaine in the cerebral ECF were also After full recovery from anesthesia, animals were randomly calculated using the same equation with their concentrations divided into three groups receiving lidocaine, bupivacaine, in the dialysate, K factor, and RLROP. Relative recovery of ϭ and levobupivacaine (n 12, each group). After confirming lidocaine and RLOHBUP, bupivacaine and RLROP, and the stable rate of loss of the internal standards in the cerebral levobupivacaine and RLROP were almost equal and not af- dialysate and baseline hemodynamic measurement, contin- fected by the extracellular concentrations of the local anes- Ϫ Ϫ uous infusion of lidocaine (0.5 mg ⅐ kg 1 ⅐ min 1), bupiva- thetics of interest. Ϫ Ϫ caine (0.1 mg ⅐ kg 1 ⅐ min 1), or levobupivacaine (0.1 Ϫ Ϫ mg ⅐ kg 1 ⅐ min 1) via the intravenous catheter was started, Analysis of Plasma and Intracerebral Local Anesthetic which lasted for 2 h. Infusion rates of local anesthetics were Concentrations determined based on our previous reports, and preliminary Concentrations of total (protein bound and unbound) and studies not to induce hemodynamic changes or the CNS unbound lidocaine in plasma as well as total lidocaine in the toxicity such as excitation, hyperventilation, or convul- brain were determined by liquid chromatography–mass sions.4,5 Mean arterial blood pressure and heart rate were spectrometry (4000QTRAP, Applied Biosystems, Foster continuously monitored, and arterial blood samples (0.5 ml) City, CA), according to previously reported methods.5 For were drawn before infusion of drugs (baseline), 15, 30, 45, measuring bupivacaine and levobupivacaine, midazolam (3 60, 90, 120, 135, 150, 165, 180, 210, and 240 min after ␮M, 100 ␮l) was added to the samples as an internal standard starting intravenous infusion for measuring plasma concen- and measured by mass spectrometry at m/z 288.8, 288.8, and trations, with the same volume of drug-free blood obtained 326.7, respectively. Unbound fraction in plasma was mea- from another rat being transfused to prevent blood dilution sured after ultrafiltration using a membrane (YM-30; Milli- and hemodynamic changes. Blood gas was measured at base- pore Corporation, Billerica, MA). Concentrations of un- line, at the end of the infusion of anesthetics (120 min), and bound anesthetics in the cerebral dialysate were measured by at the end of experiments (240 min). a previously reported method with small modifications.3 In the second experiment, total contents of local anesthet- Briefly, indwelling microdialysis probes were perfused with ics in the whole brain tissue were measured in 15 male Spra- the cerebrospinal fluid at a rate of 2 ␮l/min, and the dialysate gue–Dawley rats. Catheters were inserted into the cervical was collected for every 5 min. Dialysates were kept at 4°C vein under general anesthesia with sevoflurane. After recov- and injected onto the high-performance liquid chromato- ery from anesthesia, they received the same solution of lido- graph equipped with electrochemical detector (HTEC-500; caine, bupivacaine, or levobupivacaine at the same infusion Eicom) on the same day of experiments. Calibration curves rate as in the first experiment (n ϭ 5 for each anesthetic). for lidocaine were constructed for each run over the range of They were euthanized with 100 mg/kg thiopental 120 min 0.02–20 ␮g/ml, and calibration curves for both bupivacaine after starting the infusion and killed by decapitation. After and levobupivacaine were 0.02–2.0 ␮g/ml. The values of r2 immediate removal of the brain, surface blood vessels were were greater than 0.999 for all anesthetics, and the limit of cleaned and dissected on moist filter paper and kept at detection was 0.01 ␮g/ml. Within-day and day-to-day coef- Ϫ70°C until analysis. ficients of variation of lidocaine, bupivacaine, and levobupivacaine were less than 6%. We have examined the concor- In Vitro and In Vivo Microdialysis Probe Calibration dance of the concentrations of local anesthetics measured by Study electrochemical detection and mass spectrometry; the details For quantitative measurement of the extracellular concentra- are described in the appendix. tion of local anesthetics in the brain, we used a retrodialysis technique based on the principle that relative loss (RL) of an Pharmacokinetic Analysis internal standard is related to the relative recovery of the Pharmacokinetic parameters of lidocaine, bupivacaine, and solutions of interest. The dialysis probe was calibrated in levobupivacaine in plasma and in the cerebral ECF were

Ikeda et al. Anesthesiology, V 112 • No 6 • June 2010 1397 Pharmacokinetics of Local Anesthetics in the Brain calculated by noncompartmental analysis based on the data Statistical analysis was performed using SigmaStat 3.0 (Systat obtained from the end of infusion until the last sampling Software, Inc., Richmond, CA). Differences in mean arterial time using WinNonlin Professional 5.1 (Pharsight Corpora- blood pressure, heart rate, hemoglobin content, and blood gas tion, Mountain View, CA) as reported previously.3 We esti- data during the whole experiments were examined by two-factor mated the three different tissue-to-plasma partition coeffi- repeated-measure ANOVA using the agents and measurement cients: the ratio of the total brain concentration to total times as factors, followed by Student–Newman–Keuls test for plasma concentration (Kp), the ratio of the total brain con- multiple comparisons. Differences in the concentrations of bu- centration to unbound plasma concentration (Kp,u), and the pivacaine and levobupivacaine in plasma and in the cerebral ratio of the unbound brain concentration to unbound ECF during the whole experiments were examined by one-fac- 7 plasma concentration (Kp,uu) as reported by Gupta et al. tor repeated-measure ANOVA. Pharmacokinetic parameters of using the following equations: anesthetics in plasma, cerebral ECF, tissue-to-plasma partition Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/112/6/1396/250464/0000542-201006000-00019.pdf by guest on 01 October 2021 ϭ ϭ coefficients (Kp, Kp,u, and Kp,uu), and the unbound volume of Kp AUCtot,br/AUCtot,pl (AUCu,brECF/fu,brECF)/AUCtot,pl distribution in the brain (Vu,br) were compared with one-factor (1) ANOVA, followed by Student–Newman–Keuls test for multi- K ϭ AUC /AUC ϭ (AUC /f )/AUC ple comparisons. Values of P less than 0.05 were considered p,u tot,br u,pl u,brECF u,brECF u,pl significant. (2) ϭ Kp,uu AUCu,brECF/AUCu,pl (3) Results where AUCtot,br is area under the concentration–time curve During experiments, symptoms of CNS toxicity such as ex- (AUC) from time 0 to infinity of total lidocaine, bupiva- citation, depression, or convulsions were not observed while caine, or levobupivacaine in the whole brain; AUCtot,pl and all animals were awake at the end of experiments. There were AUCu,pl are AUC of the total and unbound concentrations no significant differences in baseline, end-infusion, and end- in plasma, respectively. AUCu,brECF denotes the AUC of the experiment values for mean arterial blood pressure, heart unbound concentrations in the cerebral ECF obtained in the rate, hemoglobin content, blood gas data, or between-drug microdialysis study. AUCtot,br is not able to be directly mea- groups for these parameters (data not shown). 7 sured in vivo and expressed as AUCu,brECF/fu,brECF, where Plasma concentrations of total and unbound lidocaine, fu,brECF is the ratio of unbound concentrations in cerebral bupivacaine, and levobupivacaine reached the highest levels ECF to the total amount per gram of brain tissue and is also at the end of infusion (120 min; figs. 1A and B). Peak plasma 7 expressed as 1/Vu,br. Vu,br stands for the unbound volume of concentration divided by dose (Cmax/dose) of both total and distribution in the brain and is calculated using the following unbound lidocaine was significantly higher than total and equation8: unbound bupivacaine and levobupivacaine, respectively Ͻ ϭ Ϫ ⅐ (P 0.05 for both). There were no differences in the elim- Vu,br (Abr Vbl Cbl)/Cu,brECF (4) ination half-time (t1/2) or mean residence time among total where Abr is the total amount of the local anesthetics per lidocaine, bupivacaine, and levobupivacaine or among their gram of the brain, which was measured in the second exper- unbound fractions in plasma (table 1). There were no differ- iment. Vbl is the volume of blood per gram of brain, which ences in clearance or volume of distribution at steady state ␮ 9 was assumed to be 15 l/g brain. Cbl and Cu,brECF are total (Vdss) among total lidocaine, bupivacaine, and levobupivacaine; concentrations in plasma and unbound concentration in the however, both these parameters of unbound lidocaine were sig- cerebral ECF, respectively, measured at the end of infusion nificantly lower than those of bupivacaine and levobupivacaine Ͻ (120 min) in the first experiment. Calculated Vu,br and (P 0.001 for all). Plasma concentrations of total bupivacaine fu,brECF were assumed to be constant with time (0–240 and levobupivacaine and of unbound bupivacaine and min).7 The fraction of local anesthetics bound to erythro- levobupivacaine were comparable, and there were no differences cytes was ignored because the content of anesthetics in the in the pharmacokinetic parameters between these two an- blood vessels in the brain expressed as the volume of blood in the esthetics (fig. 1B and table 1). Protein-binding ratios of lido- brain (Vbl) multiplied by total plasma concentrations (Cbl)is caine, bupivacaine, and levobupivacaine in plasma calculated as Ϫ Ϯ Ϯ much smaller than the total amount in the brain (Abr). (AUCtot,pl AUCu,pl)/AUCtot,pl, were 41 14%, 86 3%, and 86 Ϯ 4%, respectively (n ϭ 12). Statistical Analysis Concentrations of lidocaine, bupivacaine, and The number of animals in each group was determined based on levobupivacaine in the cerebral ECF were also highest at ϭ our preliminary study (n 6) in which AUC of levobupiva- the end of infusion (fig. 2). Cmax/dose of lidocaine in the Ϫ caine was 5.0 Ϯ 0.9 min ⅐ ␮g ⅐ ml 1. On the basis of the for- cerebral ECF was significantly higher than those of bupiv- mula for normal theory and assuming a type I error protection acaine and levobupivacaine (P Ͻ 0.001 for both). In con- of 0.05 and a power of 0.80 allowing us to detect a 25% change trast to plasma, Cmax/dose, AUC0-ϱ/dose, and the concen- in the cerebral extracellular concentrations, 12 animals were trations of bupivacaine in the cerebral ECF throughout required for each group. All values are expressed as means Ϯ SD. the experiments were significantly higher than levobupi-

1398 Anesthesiology, V 112 • No 6 • June 2010 Ikeda et al. PERIOPERATIVE MEDICINE Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/112/6/1396/250464/0000542-201006000-00019.pdf by guest on 01 October 2021

Fig. 1. Plasma concentration–time profiles of total and unbound lidocaine, bupivacaine, and levobupivacaine. (A) Plasma concentrations of total (protein bound and unbound) and unbound lidocaine. (B) Plasma concentrations of total and unbound racemic bupivacaine (bupivacaine) and levobupivacaine. Data are expressed as the mean Ϯ SD of 12 experiments. There were no differences between total bupivacaine and levobupivacaine or between unbound bupivacaine and levobupivacaine concentrations. vacaine (P ϭ 0.006, P ϭ 0.04, and P Ͻ 0.001, respectively; cantly larger than lidocaine and bupivacaine (P Ͻ 0.001 for table 2 and fig. 2B). Apparent clearance (clearance divided both). by bioavailability, Cl/F) and apparent volume of distribution (volume of distribution divided by bioavailability, Discussion Vd/F) of lidocaine were significantly smaller relative to those of both bupivacaine and levobupivacaine (P Ͻ 0.001 for all). In this study, we measured the concentrations of lidocaine, Both Cl/F and Vd/F of bupivacaine were significantly bupivacaine, and levobupivacaine in plasma, brain, and ce- smaller compared with those of levobupivacaine (P Ͻ 0.001 rebral ECF. Although there were no differences between the ϭ and P 0.04, respectively). There were no differences in t1/2 concentrations of bupivacaine and levobupivacaine in or mean residence time among lidocaine, bupivacaine, and plasma, levels of unbound bupivacaine in the cerebral ECF levobupivacaine. were significantly higher than levobupivacaine. Pharmacoki-

There were no differences in Kp among the three anesthetics netic analysis has also revealed differences between bupiva- ϭ (P 0.9, fig. 3). On the other hand, Kp,u of both bupivacaine caine and levobupivacaine only in the cerebral ECF and not and levobupivacaine was significantly higher than lidocaine in plasma. For elucidating these pharmacokinetic differences (P Ͻ 0.001 for both), with no differences being observed be- between plasma and cerebral ECF, we have calculated the tween bupivacaine and levobupivacaine. Kp,uu of bupivacaine tissue-to-plasma partition coefficients. was also significantly higher than that of lidocaine and There were no differences in Kp, the ratio of AUC of total levobupivacaine (P ϭ 0.03 and 0.003, respectively). Contents brain concentrations to AUC of total plasma concentrations, of lidocaine, bupivacaine, and levobupivacaine in the whole among the three anesthetics. Kp of these anesthetics was ap- brain at 120 min after starting infusion (Abr) measured in our proximately 2–3, being comparable with the ratio of total second experiment were 23.1 Ϯ 2.6, 6.8 Ϯ 1.5, and 6.6 Ϯ 1.8 concentrations in the brain homogenates against the concen- ␮ 4,5 g/g brain, respectively. Vu,br of levobupivacaine was signifi- trations in plasma in our previous studies, suggesting that

Table 1. Pharmacokinetic Parameters of Lidocaine, Bupivacaine, and Levobupivacaine in Plasma

Total Unbound Lidocaine Bupivacaine Levobupivacaine Lidocaine Bupivacaine Levobupivacaine Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Cmax/dose (mg/ml) 174 30 141 16* 155 18† 95 23 30 6‡ 25 3‡ Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ t1/2 (min) 43 12 50 64934515 45 5416 ⅐ ⅐ Ϫ1 Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ AUC0-ϱ/dose (min g ml ) 8.1 2.9 7.8 1.1 7.7 0.8 4.6 1.4 1.1 0.3‡ 1.0 0.3‡ Clearance 154 Ϯ 117 129 Ϯ 16 131 Ϯ 13 236 Ϯ 114 955 Ϯ 226‡ 1032 Ϯ 310‡ (ml ⅐ kgϪ1 ⅐ minϪ1) MRT (min) 64 Ϯ 20 77 Ϯ 873Ϯ 568Ϯ 22 66 Ϯ 860Ϯ 11 ϫ 3 Ϯ Ϯ Ϯ Ϯ Ϯ Ϯ Vdss ( 10 ml/kg) 2 1212131136‡ 11 4‡

Values are the mean Ϯ SD of 12 experiments. * P Ͻ 0.01, † P Ͻ 0.05, and ‡ P Ͻ 0.001 compared with lidocaine. ϭ ϭ ϭ AUC0-ϱ area under the plasma concentration versus time curve from zero to infinity; Cmax peak plasma concentration; dose total ϭ ϭ ϭ dose administered during experiments; MRT mean residence time; t1/2 elimination half-time; Vdss volume of distribution at steady state.

Ikeda et al. Anesthesiology, V 112 • No 6 • June 2010 1399 Pharmacokinetics of Local Anesthetics in the Brain Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/112/6/1396/250464/0000542-201006000-00019.pdf by guest on 01 October 2021

Fig. 2. Cerebral extracellular fluid concentration–time profiles of lidocaine, bupivacaine, and levobupivacaine. (A) Cerebral extracellular fluid concentrations of lidocaine. (B) Cerebral extracellular fluid concentrations of bupivacaine and levobupivacaine. Data are expressed as the mean Ϯ SD of 12 experiments. Overall concentrations of bupivacaine were significantly higher than levobupivacaine (P Ͻ 0.001). the total content of these anesthetics in the brain; that is, sum Among the three parameters, Kp and Kp,u were calculated of the protein-bound and unbound fractions in the cerebral based on the AUC of total concentrations in the brain (AUCtot,br, ECF and fractions bound to brain parenchyma was higher Eqs. 1 and 2), which was not able to be directly measured and than those in plasma. Kp,u, the ratio of AUC of total brain calculated from the AUC of the unbound concentrations in concentrations to AUC of unbound plasma concentrations the ECF (AUCu,brECF) and volume of distribution of the accounts for the differences in binding to blood components. 8 unbound fraction in the brain (Vu,br). Vu,br was calculated Kp,u of both bupivacaine and levobupivacaine was signifi- from the differences in the contents of anesthetics between cantly higher than lidocaine, reflecting the lower unbound the whole brain and blood vessels in the brain, divided by plasma concentrations resulting from higher protein binding concentrations of unbound fraction in the cerebral ECF (Eq. ratio compared with lidocaine. Kp,uu denotes the ratio of 4). The mean V values of lidocaine, bupivacaine, and AUC of unbound concentrations in the cerebral ECF to u,br levobupivacaine were 5.1, 21.6, and 31.3 ml/g brain, respec- AUC of unbound plasma concentrations. K of bupiva- p,uu tively, and approximately 25- to 150-fold larger than the caine was significantly higher than levobupivacaine, resulting volume of the brain extracellular space, which is reported to from higher concentrations in the cerebral ECF than be 0.12–0.20 ml/g brain.10,11 levobupivacaine despite similar plasma concentrations. These results suggest that all these anesthetics are extensively distributed intracellularly or Table 2. Pharmacokinetic Parameters of Lidocaine, bind to proteins in the ECF. Bupivacaine, and Levobupivacaine in the Cerebral There are several studies examining the intracerebral con- Extracellular Fluid centration of anesthetic agents in awake animals3,12–14;un- fortunately, however, tissue-to-plasma partition coefficients Lidocaine Bupivacaine Levobupivacaine have not been examined in those studies. Regarding local Ϯ Ϯ Ϯ 13,14 Cmax/dose 76 12 27 6*† 18 2* anesthetics, previous studies have been predominantly (mg/ml) Ϯ Ϯ Ϯ focused on epidural or intrathecal pharmacokinetics after t1/2 (min) 27 5285228 AUC /dose 2.1 Ϯ 0.4 0.64 Ϯ 0.16*‡ 0.41 Ϯ 0.10* epidural administration, whereas few studies examined the 0-ϱ 3 (min ⅐ g ⅐ mlϪ1) concentrations in the cerebral ECF, diffused across the Cl/F (ml ⅐ kgϪ1 ⅐ 491 Ϯ 82 1656 Ϯ 381*† 2661 Ϯ 1027* blood–brain barrier after intravenous administration. In this Ϫ min 1) study, we have measured the concentrations in the brain Ϯ Ϯ Ϯ MRT (min) 37 7377336 followed by calculation of the tissue-to-plasma partition co- Vd/F (ϫ103 19 Ϯ 667Ϯ 16*‡ 79 Ϯ 17* ml/kg) efficients. Because intracerebral pharmacokinetics is suscep- tible to cerebral blood flow, we administered local anesthetics Values are the mean Ϯ SD of 12 experiments. at rates too slow to induce CNS effects such as excitation and * P Ͻ 0.001 compared with lidocaine, † P Ͻ 0.01 and ‡ P Ͻ 0.05 depression, thereby successfully maintaining blood gas levels compared with levobupivacaine. ϭ during the whole experiments. Although concentrations of AUC0-ϱ area under the concentration vs. time curve from zero to infinity; Cl/F ϭ apparent cerebral extracellular fluid clearance; bupivacaine in the cerebral ECF were significantly higher ϭ Cmax peak concentration in the cerebral extracellular fluid; than levobupivacaine, the total amount (tissue-bound and dose ϭ total dose administered during experiments; MRT ϭ ϭ ϭ unbound) of bupivacaine in the brain was similar to mean residence time; t1/2 elimination half-time; Vd/F apparent volume of distribution, where F is the bioavailability. levobupivacaine, suggesting that a higher ECF concentration

1400 Anesthesiology, V 112 • No 6 • June 2010 Ikeda et al. PERIOPERATIVE MEDICINE Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/112/6/1396/250464/0000542-201006000-00019.pdf by guest on 01 October 2021

Fig. 3. Tissue-to-plasma partition coefficients and intracerebral volume of distribution of lidocaine, bupivacaine, and levobupivacaine. (A–C) Tissue-to-plasma partition coefficients (Kp, Kp,u, and Kp,uu) and (D) volume of distribution (Vu,br) of the unbound fraction of lidocaine, bupivacaine, and levobupivacaine in the brain. Data are expressed as the mean Ϯ SD of 12 experiments. * P Ͻ 0.05, ** P Ͻ 0.01 compared with lidocaine, ‡ P Ͻ 0.01 compared with bupivacaine or levobupivacaine. is not directly related to the higher CNS toxicity of bupiva- anesthetics in both plasma and cerebral ECF were unde- caine than that of levobupivacaine reported previously.1,4 tectable in one animal receiving bupivacaine at the end of This study has several limitations. First of all, concentra- experiments, suggesting that their levels in the whole tions of anesthetics in the cerebral ECF were measured by brain would also be low, and such a low level eventually high-performance liquid chromatograph equipped with an made it difficult to perform precise measurement. Despite electrochemical detector. Despite precise quantitation of lo- these limitations, differences of Abr existing between rats cal anesthetics by high-performance liquid chromatography in the first and second experiments would not influence in our previous study as well as in other reports,3,13,14 and the relationships of tissue-to-plasma partition coefficients good concordance with the results measured by mass spec- among the three anesthetics, because Abr was much larger trometry, a similar quantitation method with mass spec- than the amount of anesthetics in the cerebral blood ves- ϫ trometry for measuring the cerebral ECF concentrations in sels (Vbl Cbl) and Vu,br, a determinant of Kp and Kp,u,is the place of electrochemical detection would be more pref- predominantly dependent on Abr and the concentration of erable.7 Second, concentrations of anesthetics in plasma and anesthetics in the cerebral ECF (Eq. 4). in the cerebral ECF were measured only for 2 h after termi- In summary, we have shown that concentrations of nation of the infusion, which may influence calculation of bupivacaine in the cerebral ECF are higher than levobupi- the pharmacokinetic parameters based on the terminal elim- vacaine despite similar concentrations in plasma and in ination phase. the whole brain. Further studies are warranted to elucidate Third, we determined the content of anesthetics in the the mechanism responsible for differences in the toxicity whole brain (Abr) in another group of animals; however, of the stereoisomers. they were not subjected to the microdialysis study in which the pharmacokinetic parameters in the cerebral ECF were measured as reported previously.7,8 We tried to References calculate and compare the pharmacokinetic parameters of 1. Santos AC, DeArmas PI: Systemic toxicity of levobupivacaine, bupivacaine, and ropivacaine during continuous anesthetics both in plasma and cerebral ECF, but sam- intravenous infusion to nonpregnant and pregnant ewes. pling of blood and cerebral dialysate was required until ANESTHESIOLOGY 2001; 95:1256–64 their concentrations decreased to very low levels. In fact, 2. Bernards CM, Artru AA: Hexamethonium and midazolam

Ikeda et al. Anesthesiology, V 112 • No 6 • June 2010 1401 Pharmacokinetics of Local Anesthetics in the Brain

terminate dysrhythmias and hypertension caused by intra- 12. Dagenais C, Graff CL, Pollack GM: Variable modulation of cerebroventricular bupivacaine in rabbits. ANESTHESIOLOGY opioid brain uptake by P-glycoprotein in mice. Biochem 1991; 74:89–96 Pharmacol 2004; 67:269–76 3. Takahashi R, Oda Y, Tanaka K, Morishima HO, Inoue K, 13. Ratajczak-Enselme M, Estebe JP, Rose FX, Wodey E, Mali- Asada A: Epinephrine increases the extracellular lidocaine novsky JM, Chevanne F, Dollo G, Ecoffey C, Le Corre P: concentration in the brain: A possible mechanism for in- Effect of epinephrine on epidural, intrathecal, and plasma creased central nervous system toxicity. ANESTHESIOLOGY pharmacokinetics of ropivacaine and bupivacaine in 2006; 105:984–9 sheep. Br J Anaesth 2007; 99:881–90 4. Tanaka K, Oda Y, Funao T, Takahashi R, Hamaoka N, Asada 14. Rose FX, Estebe JP, Ratajczak M, Wodey E, Chevanne F, Dollo G, A: Dexmedetomidine decreases the convulsive potency of Bec D, Malinovsky JM, Ecoffey C, Le Corre P: Epidural, intrathe- bupivacaine and levobupivacaine in rats: Involvement of cal pharmacokinetics, and intrathecal bioavailability of ropiva- alpha-2-adrenoceptor for controlling convulsions. Anesth caine. Anesth Analg 2007; 105:859–67 Analg 2005; 100:687–96

5. Nakamura T, Oda Y, Takahashi R, Tanaka K, Hase I, Asada Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/112/6/1396/250464/0000542-201006000-00019.pdf by guest on 01 October 2021 A: Propranolol increases the threshold for lidocaine-in- duced convulsions in awake rats: A direct effect on the Appendix: Concordance of the brain. Anesth Analg 2008; 106:1450–5 Concentrations of Local Anesthetics 6. Clement R, Malinovsky JM, Dollo G, Le Corre P, Chevanne F, Measured By Electrochemical Detection Le Verge R: In vitro and in vivo microdialysis calibra- and Mass Spectrometry tion using retrodialysis for the study of the cerebrospinal distribution of bupivacaine. J Pharm Biomed Anal 1998; 17: For examining the correlations between methods for determining local 665–70 anesthetics concentrations, we have prepared standard solutions of li- 7. Gupta A, Chatelain P, Massingham R, Jonsson EN, Ham- ␮ marlund-Udenaes M: Brain distribution of cetirizine enan- docaine (0.02–10.0 g/ml), bupivacaine, and levobupivacaine (0.02– tiomers: Comparison of three different tissue-to-plasma 0.5 ␮g/ml) in the cerebrospinal fluid, measured their concentrations by partition coefficients: K(p), K(p,u), and K(p,uu). Drug electrochemical detection and mass spectrometry, and compared their Metab Dispos 2006; 34:318–23 results using a Spearman rank order test (r2). Agreement was deter- 8. Wang Y, Welty DF: The simultaneous estimation of the mined according to the method described by Bland–Altman plots. We influx and efflux blood-brain barrier permeabilities of calculated mean difference (bias), SD of the difference (precision), and gabapentin using a microdialysis-pharmacokinetic ap- proach. Pharm Res 1996; 13:398–403 limit of agreement (95% confidence interval) for the values mea- 9. Beach VL, Steinetz BG: Quantitative measurement of Evans sured by electrochemical detector and mass spectrometry. The blue space in the tissues of the rat: Influence of 5-hydroxy- correlation coefficients (r2) between methods were more than tryptamine antagonists and phenelzine on experimental in- 0.99 for lidocaine, bupivacaine, and levobupivacaine (P Ͻ 0.01 flammation. J Pharmacol Exp Ther 1961; 131:400–6 for all, fig. 4). Mean (bias) Ϯ SD of the difference (precision) 10. Levin VA, Fenstermacher JD, Patlak CS: Sucrose and inulin were 0.00 Ϯ 0.03, 0.00 Ϯ 0.01, and 0.00 Ϯ 0.01 ␮g/ml for space measurements of cerebral cortex in four mammalian species. Am J Physiol 1970; 219:1528–33 lidocaine, bupivacaine, and levobupivacaine, respectively. Limit Ϫ Ϫ 11. Goodman FR, Weiss GB, Alderdice MT: On the measurement of agreement ranged from 0.06 to 0.07, 0.02 to 0.01, and Ϫ ␮ of extracellular space in slices prepared from different rat 0.01 to 0.01 g/ml for lidocaine, bupivacaine, and levobupi- brain areas. Neuropharmacology 1973; 12:867–73 vacaine, respectively (fig. 5).

Fig. 4. Linear regression plots of the concentrations measured by electrochemical detection and mass spectrometry. Each plot for (A) lidocaine (n ϭ 16), (B) racemic bupivacaine (bupivacaine) (n ϭ 12), and (C) levobupivacaine (n ϭ 12). Solid lines represent the regression lines. P Ͻ 0.001 for significance of the Spearman rank order correlation for lidocaine, bupivacaine, and levobupivacaine.

1402 Anesthesiology, V 112 • No 6 • June 2010 Ikeda et al. PERIOPERATIVE MEDICINE Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/112/6/1396/250464/0000542-201006000-00019.pdf by guest on 01 October 2021

Fig. 5. Bland–Altman plots for describing average and difference of the measured concentrations by electrochemical detection and mass spectrometry. Solid lines represent the bias, the upper 95% confidential interval (ϩ1.96 SD) and the lower 95% confidential interval (Ϫ1.96 SD) for lidocaine (n ϭ 16), racemic bupivacaine (n ϭ 12), and levobupivacaine (n ϭ 12).

Ikeda et al. Anesthesiology, V 112 • No 6 • June 2010 1403