Pharmacokinetics and Pharmacodynamics of Ketamine As an Analgesic for Rational Intravenous Infusion in Ponies

Pharmacokinetics and Pharmacodynamics of Ketamine as an Analgesic for Rational Intravenous Infusion in Ponies.

Graduate School for Cellular and Biomedical Sciences University of Bern

PhD Thesis

Submitted by

Olivier Louis Raymond LEVIONNOIS from Cancale, France

Thesis advisor

Prof. Dr. Meike Mevissen Division of Veterinary Pharmacology and Toxicology Department of Clinical Research and Veterinary Public Health Vetsuisse Faculty, University of Bern

AUTHOR

OLIVIER LOUIS RAYMOND LEVIONNOIS

Born the 6th September 1977, Paris, France

EDUCATION 2003 Academic degree DVM Ecole Nationale Vétérinaire, Agroalimentaire et de l'Alimentation Nantes-Atlantique, Nantes, France 2006 Academic degree Dr. med.vet. Department of Clinical Veterinary Medicine, Vetsuisse Faculty, University of Bern, Switzerland 2006-2011 PhD-student for the DVM-PhD at the Graduate School for Cellular and Biomedical Sciences, University of Bern, Switzerland

POSTGRADUATE TRAINING 2007 Diplomate of the European College of Veterinary Anaesthesia and Analgesia (ECVAA)

CURRENT POSITION Clinical educator and Research assistant, Division Anaesthesiology, Department Clinical Veterinary Medicine, Vetsuisse Faculty, University of Bern, Switzerland.

LAST PUBLICATIONS Pieper K, Schuster T, Levionnois O, Matis U, Bergadano A. Antinociceptive efficacy and plasma concentrations of transdermal buprenorphine in dogs. Vet J. 2010 Mar 3. [Epub ahead of print] Levionnois OL, Mevissen M, Thormann W, Spadavecchia C. Assessing the efficiency of a pharmacokinetic-based algorithm for target-controlled infusion of ketamine in ponies. Res Vet Sci. 2010 Jun;88(3):512-8. Levionnois OL, Menge M, Thormann W, Mevissen M, Spadavecchia C. Effect of ketamine on the limb withdrawal reflex evoked by transcutaneous electrical stimulation in ponies anaesthetised with isoflurane. Vet J. 2010 Dec;186(3):304-11. Larenza MP, Peterbauer C, Landoni MF, Levionnois OL, Schatzmann U, Spadavecchia C, Thormann W. Stereoselective pharmacokinetics of ketamine and norketamine after constant rate infusion of a subanesthetic dose of racemic ketamine or S-ketamine in Shetland ponies. Am J Vet Res. 2009 Jul;70(7):831-9. Enderle AK, Levionnois OL, Kuhn M, Schatzmann U. Clinical evaluation of ketamine and lidocaine intravenous infusions to reduce isoflurane requirements in horses under general anaesthesia. Vet Anaesth Analg. 2008 Jul;35(4):297-305. Spadavecchia C, Levionnois OL, Kronen P, Andersen OK. The effects of isoflurane minimum alveolar concentration on withdrawal reflex activity evoked by repeated transcutaneous electrical stimulation in ponies. Vet J. 2010 Mar;183(3):337-44. Levionnois OL, Spadavecchia C, Kronen PW, Schatzmann U. Determination of the minimum alveolar concentration of isoflurane in Shetland ponies using constant current or constant voltage electrical stimulation. Vet Anaesth Analg. 2009 Jan;36(1):9-17. Levionnois OL, Kronen P. Development of post-pump syndrome in a sheep after mitral valve stenting. Lab Anim. 2008 Oct;42(4):505-10. Larenza MP, Knobloch M, Landoni MF, Levionnois OL, Kronen PW, Theurillat R, Schatzmann U, Thormann W. Stereoselective pharmacokinetics of ketamine and norketamine after racemic ketamine or S-ketamine administration in Shetland ponies sedated with xylazine. Vet J. 2008 Sep;177(3):432-5. Larenza MP, Landoni MF, Levionnois OL, Knobloch M, Kronen PW, Theurillat R, Schatzmann U, Thormann W. Stereoselective pharmacokinetics of ketamine and norketamine after racemic ketamine or S-ketamine administration during isoflurane anaesthesia in Shetland ponies. Br J Anaesth. 2007 Feb;98(2):204-12. Knobloch M, Portier CJ, Levionnois OL, Theurillat R, Thormann W, Spadavecchia C, Mevissen M. Antinociceptive effects, metabolism and disposition of ketamine in ponies under target-controlled drug infusion. Toxicol Appl Pharmacol. 2006 Nov 1;216(3):373-86. Spadavecchia C, Levionnois OL, Kronen PW, Leandri M, Spadavecchia L, Schatzmann U. Evaluation of administration of isoflurane at approximately the minimum alveolar concentration on depression of a nociceptive withdrawal reflex evoked by transcutaneous electrical stimulation in ponies. Am J Vet Res. 2006 May;67(5):762-9. Pharmacokinetics and Pharmacodynamics of Ketamine as an Analgesic for Rational Intravenous Infusion in Ponies.

Department of Clinical Veterinary Medicine

Division of Anaesthesiology, Vetsuisse Faculty,

University of Berne, Berne, Switzerland

University of Bern | Graduate School for Cellular and Biomedical Sciences

Theodor Kocher Institute | Freiestrasse 1 | CH-3012 Bern

Olivier Louis Raymond LEVIONNOIS

2011

Accepted by the Faculty of Medicine, the Faculty of Science and the Vetsuisse Faculty of the University of Bern at the request of the Graduate School for Cellular and Biomedical Sciences

Bern, Dean of the Faculty of Medicine

Bern, Dean of the Faculty of Science

Bern, Dean of the Vetsuisse Faculty Bern A ma famille, d’ici et de là bas, tous dans mon cœur.

Supervisor: Prof. Meike Mevissen Division of Veterinary Pharmacology and Toxicology Department of Clinical Research and Veterinary Public Health Vetsuisse Faculty, University of Bern, Switzerland

Co-referee: Prof. Claudia Spadavecchia Division of Veterinary Anaesthesiology and Pain Medicine Department of Clinical Veterinary Sciences Vetsuisse Faculty, University of Bern, Switzerland

Secondary Co-referee and Reviewers: Prof. Wolfgang Thormann Department of Clinical Pharmacology, University of Bern, Switzerland

Prof. Christopher Portier Environmental Toxicology Program National Institute of Environmental Health Sciences Research Triangle Park, NC 27709, USA

Thesis Mentor: Prof. Mireille Meylan Ruminant Clinic Department of Clinical Veterinary Sciences Vetsuisse Faculty, University of Bern, Switzerland

Ketamine pharmacokinetic and pharmacodynamic

0 Ketamine pharmacokinetic and pharmacodynamic

Content

1. Abbreviations ...... 2

2. Acknowledgments ...... 3

3. Summary ...... 4

(1) Abstract (English) ...... 4

(2) Abstract (German) ...... 5

(3) Abstract (French) ...... 6

4. List of original publications ...... 7

5. Preface ...... 9

(1) Preliminary notes ...... 9

(2) Background ...... 9

6. Introduction ...... 13

7. Aims of the PhD project ...... 15

(1) Characterization of ketamine pharmacokinetics ...... 15

(2) The effect of isoflurane on the Withdrawal Reflex ...... 16

(3) Effects of ketamine on the Withdrawal Reflex ...... 16

8. Methods and Results ...... 17

(1) Animals ...... 17

(2) Induction of isoflurane general anaesthesia ...... 17

(3) Ketamine: stereoselective pharmacokinetics ...... 17

(4) Electrophysiological recordings of the Withdrawal Reflexes ...... 27

9. Conclusion and Future applications ...... 31

10. References ...... 35

Declaration of Originality ...... 39

Appendix: Publications ...... 41

1 Ketamine pharmacokinetic and pharmacodynamic

1. Abbreviations

CC Constant current CE Capillary electrophoresis CV Constant voltage EMG Electromyographic ETT Endotracheal tube HPLC High pressure liquid chromatography ID Internal diameter MAC Minimum alveolar concentration NMDA N-methyl-D-aspartate NWR Nociceptive withdrawal reflex PBPk Physiologically-based pharmacokinetic R-ket R (-) ketamine R-nor R (-) norketamine RMS Root mean square S-ket S (+) ketamine S-nor S (+) norketamine TCI Target-controlled infusion TS Temporal summation WDR Wide dynamic range WR Withdrawal reflex

2 Ketamine pharmacokinetic and pharmacodynamic

2. Acknowledgments

The present study has been carried out at the Anaesthesiology division, Department of Clinical Veterinary Medicine, Vetsuisse Faculty, University of Berne, Switzerland. I wish to sincerely thank Prof. Urs Schatzmann who initially received me in this team and continuously supported me in the development of my career, as well as providing location and experimental animals.

I wish to express my deepest and sincere gratitude to the whole team of the Anaesthesiology division where I grew up together with many collaborators: Prof. Claudia Spadavecchia, for her continuous support from the first day until the last one, for her to trust me, and of course for her best-quality scientific supervision; Dr. Paula Larenza, for sharing interests and joining forces at the time we both started our PhD work, and to all my other colleagues for sharing idea, time and help by the experiments as well as supporting my agenda between clinics and office.

My sincere thanks to Prof. Meike Mevissen for overtaking the supervision of this PhD and providing organizational support in financing these years, in writing and in reviewing pharmacokinetic issues as well as leading the investigation on the PBPk model.

I am in debt to Dr. Monika Knobloch who, besides being a warm friend since many years, invested time of her life to start this PhD and to help with unfettered motivation to set the capillary electrophoresis method. This would not have been possible without the full support of Prof. Wolfgang Thormann, Regula Theurillat, and their all laboratory team at the Institute of Clinical Pharmacology and Visceral Research. Many thanks for their help and wonderful work.

Thanks to Dr. Luciano Spadavecchia who developed the devices and software. The ketamine and S- ketamine were obtained from the laboratory company Dr. E. Graeub AG. I thank them for their support. Thanks to Prof. André Jaggy for its support at combining clinical duty with experimental work.

Finally my gratefulness and love to my wife Simone, who provided me the most important in life – Immense love, blind confidence and full of time!

The studies of this PhD have been supported financially by Vetsuisse and the Swiss National Science Foundation. The grants are gratefully acknowledged.

3 Ketamine pharmacokinetic and pharmacodynamic

3. Summary

(1) Abstract (English)

The research presented in this thesis focused on experimental studies on the pharmacokinetics and the antinociceptive activity of ketamine in ponies anaesthetized with isoflurane. The latter was investigated by the modulation of the nociceptive withdrawal reflex (NWR).

In a first step, an analytical method using capillary electrophoresis was developed to quantify both ketamine enantiomers and their metabolites S- and R-norketamine. Secondly, the stereospecific pharmacokinetics of ketamine and norketamine was studied after administration of a single bolus. Slight but significant differences were found between S-norketamine and R-norketamine levels. A target- controlled infusion (TCI) of ketamine was administered successfully. Subsequently, the pharmacokinetics of the ketamine infusion was studied and integrated, together with previous data, in a physiologically-based pharmacokinetic (PBPk) model.

In order to investigate the modulation of the NWR by ketamine, the minimal alveolar concentration (MAC) of isoflurane was determined, and the effect of isoflurane on the NWR was studied below and up to the MAC when single stimulation as well as repeated stimulations was applied. Isoflurane dose- dependently depressed the NWR but did not completely inhibit it. Repeated stimulation elicited stronger reflexes, but temporal summation (TS) was inhibited at peri-MAC concentration of isoflurane. Finally, the TCI of ketamine significantly abolished the NWR with the effect being more pronounced when compared to the administration of isoflurane alone. In absence of TS, a specific effect of the NMDA- antagonist ketamine on central sensitization could not be investigated.

In conclusion, the work presented in this PhD thesis has provided a new method for enantioselective quantification of ketamine in plasma, large evidence for the stereoselective pharmacokinetics of ketamine in anaesthetized ponies, and its integration in a PBPk model, as well as modulatory effects of isoflurane and ketamine on the NWR. At the plasma levels observed in the present work, ketamine was found to modulate the sensory-motor processing.

4 Ketamine pharmacokinetic and pharmacodynamic

(2) Abstract (German)

Die vorliegende Arbeit befasst sich mit experimentellen Studien zur Pharmakokinetik und der antinocizeptiven Aktivität von Ketamin bei Ponies unter Isoflurananästhesie. Letztere wurde über die Regulierung des sogenannten ‚nociceptive withdrawal reflex„ (NWR) erforscht.

Im ersten Schritt wurde eine analytische Methode etabliert, die es erlaubte beide Ketaminenantiomere sowie deren Metaboliten S- und R-Norketamin mittels Kapillarelektrophorese zu analysieren. Anschliessend wurde die stereospezifische Pharmakokinetik von Ketamin und Norketamin nach Verabreichung einer Ketamin-Bolus-Infusion untersucht. Es konnten geringe, aber signifikante Unterschiede zwischen den S-Norketamin und R-Norketamin Level nachgewiesen werden. Eine ‚target- controlled„ Ketamininfusion (TCI) wurde erfolgreich verabreicht. Die Pharmakokinetik der Ketamininfusion wurde zusammen mit vorher erhobenen Daten in einem Pharmakokinetik-Modell untersucht, welches die Physiologie der zu untersuchenden Spezies miteinbezieht (PBPk) pharmakokinetischen (PBPk) Model untersucht.

Um die Modulation des NWR durch Ketamin zu untersuchen, wurde die minimale alveoläre Konzentration (MAC) von Isofluran bestimmt und der Einfluss von Isofluran auf den NWR unterhalb und oberhalb der MAC bei einmaligen und sich wiederholenden Stimulationen erforscht. Isofluran hemmte den NWR dosisabhängig, aber nicht vollständig. Wiederholte Stimulationen riefen stärkere Reflexe hervor, allerdings wurde die ‚temporal summation„ (TS) bei der MAC Konzentration von Isofluran gehemmt. Die Ketamin TCI hemmte den NWR stärker im Vergleich zu Isofluran allein. In Abwesenheit von TS konnte kein Effekt des NMDA Antagonisten Ketamin auf die zentrale Sensitivierung untersucht werden.

Zusammenfassend kann gesagt werden, dass eine neue Methode für die enantioselektive Quantifikation von Ketamin in Plasma etabliert wurde. Weiterhin konnte gezeigt werden, dass die Pharmakokinetik von Ketamin in anästhesierten Ponies stereoselektiv ist. Die erhobenen Daten der experimentellen Studie erlaubten die Erstellung eines PBPk Model und die regulierende Wirkung von Isofluran und Ketamin auf den NWR wurde nachgewiesen. Es konnte gezeigt werden, dass Ketamin, bei den in unserer Studie vorliegenden Plasmaspiegeln, neuromuskuläre Prozesse reguliert.

5 Ketamine pharmacokinetic and pharmacodynamic

(3) Abstract (French)

La collection d‟articles présentée ici offre l‟investigation des propriétés pharmacocinétiques ainsi que l‟activité antinociceptive de la kétamine chez des poneys sous anesthésie générale à l‟isoflurane. Les propriétés anti nociceptives furent étudiées grâce au modèle du reflexe de retrait à la douleur.

Dans un premier temps, une méthode analytique basée sur l‟electrophorèse capillaire fut développée pour permettre la quantification des deux énantiomères de la kétamine et de ses métabolites issus de la N-déméthylation (les énantiomères de la nor-kétamine). Puis, la pharmacocinétique stéréospécifique de ces molécules fut étudiée in vivo après l‟administration d‟un bolus intraveineux unique chez les poneys. Les différences entre énantiomères de la nor-kétamine furent légères mais présentes et significatives. Une perfusion intraveineuse à objectif de concentration plasmatique fut administrée ensuite avec succès. Le profil pharmacocinétique de la kétamine a été étudié et avec l‟ensemble des données recueillies, un modèle pharmacocinétique de type physiologique fut élaboré.

Afin de caractériser l‟effet modulatoire de la kétamine sur le modèle du reflexe de retrait à la douleur, la concentration minimale alvéolaire fut déterminée, et l‟influence de l‟isoflurane sur le modèle fut étudiée autour ce cette concentration lorsque des stimulations électriques uniques, ou répétées, ont été appliquées. L‟isoflurane a démontré un effet dépresseur dose-dépendant du reflexe mais n‟a pas pu l‟inhiber complètement. Les stimulations répétées ont évoqué des réflexes plus marqués, mais la sommation temporelle fut inhibée lorsque l‟isoflurane était administré à environ la concentration minimale alvéolaire. La perfusion intraveineuse de kétamine a finalement montré un effet depresseur supplémentaire du reflexe. Etant donné l‟absence de sommation temporelle par l‟isoflurane, cet effet de la kétamine ne put être démontré.

En conclusion, le travail présenté dans cette thèse procure une nouvelle méthode de quantification énantioselective de la kétamine dans le plasma sanguin, une évidence pour la pharmacocinétique stéréoselective de la kétamine chez les poneys sous anesthésie générale à l‟isoflurane, le développement d‟un modèle pharmacocinétique physiologique, ainsi que les effets modulatoire de l‟isoflurane et de la kétamine sur le modèle du reflexe de retrait à la douleur. Aux niveaux plasmatiques étudiés dans les articles présents, un effet modulateur du système sensoriel et moteur par la kétamine a pu être démontré.

6 Ketamine pharmacokinetic and pharmacodynamic

4. List of original publications

The present thesis is based on the following original publications, referred to in the text by their roman numerals (I-VIII).

I. Theurillat R, Knobloch M, Levionnois O, Larenza P, Mevissen M, Thormann W. Characterization of the stereoselective biotransformation of ketamine to norketamine via determination of their enantiomers in equine plasma by capillary electrophoresis. Electrophoresis. 2005 Oct;26(20):3942-51.

II. Larenza MP, Landoni MF, Levionnois OL, Knobloch M, Kronen PW, Theurillat R, Schatzmann U, Thormann W. Stereoselective pharmacokinetics of ketamine and norketamine after racemic ketamine or S-ketamine administration during isoflurane anaesthesia in Shetland ponies. Br J Anaesth. 2007 Feb;98(2):204-12.

III. Levionnois OL, Mevissen M, Thormann W, Spadavecchia C. Assessing the efficiency of a pharmacokinetic-based algorithm for target-controlled infusion of ketamine in ponies. Res Vet Sci. 2010 Jun;88(3):512-8.

IV. Knobloch M, Portier CJ, Levionnois OL, Theurillat R, Thormann W, Spadavecchia C, Mevissen M. Antinociceptive effects, metabolism and disposition of ketamine in ponies under target-controlled drug infusion. Toxicol Appl Pharmacol. 2006 Nov 1;216(3):373-86.

V. Levionnois OL, Spadavecchia C, Kronen PW, Schatzmann U. Determination of the minimum alveolar concentration of isoflurane in Shetland ponies using constant current or constant voltage electrical stimulation. Vet Anaesth Analg. 2009 Jan;36(1):9-17.

VI. Spadavecchia C, Levionnois O, Kronen PW, Leandri M, Spadavecchia L, Schatzmann U. Evaluation of administration of isoflurane at approximately the minimum alveolar concentration on depression of a nociceptive withdrawal reflex evoked by transcutaneous electrical stimulation in ponies. Am J Vet Res. 2006 May;67(5):762-9.

VII. Spadavecchia C, Levionnois O, Kronen P, Andersen OK. The effects of isoflurane minimum alveolar concentration on withdrawal reflex activity evoked by repeated transcutaneous electrical stimulation in ponies. Vet J. 2010 Mar;183(3):337-44.

VIII. Levionnois OL, Menge M, Thormann W, Mevissen M, Spadavecchia C. Effect of ketamine on the limb withdrawal reflex evoked by transcutaneous electrical stimulation in ponies anaesthetised with isoflurane. Vet J. 2010 Dec;186(3):304-11.

7 Ketamine pharmacokinetic and pharmacodynamic

8 Ketamine pharmacokinetic and pharmacodynamic

5. Preface

(1) Preliminary notes

The manuscript presented here is based on experimental studies conducted by our research group. This thesis has been written in 2010. Reported experimental studies were performed between 2005 and 2007, and published between 2005 and 2010. The author of the present thesis (OL Levionnois) was the main investigator conducting study design, data collection, data analysis and article writing with help from collaborators and co-authors in the publications III., V. and VIII. In the publications VI. and VII., the author of the thesis (OL Levionnois) was a main investigator (study design, data collection, data analysis and article writing) together with C Spadavecchia who directed the experiments and the article writing. In the publication IV., the author of the present thesis (OL Levionnois) was a main investigator (study design, data collection, data analysis and article writing) together with M Knobloch who led the data analysis and the article writing under the direction of M Mevissen and C Spadavecchia. The development of the PBPk model was under the direction of CJ Portier. In the publications I. and II., the author of the present thesis (OL Levionnois) has been involved only partially to the experimental design and the data collection. The main investigators were R Theurillat and MP Larenza, respectively, under the direction of W Thormann and M Mevissen. These two publications are included in the description of this thesis but are not joined to the appendix.

(2) Background

In an editorial, contemporary to the writing of this thesis, J.F. Antognini wrote that “there are at least three desired end-points for general anaesthesia, namely, (1) amnesia, (2) unconsciousness and (3) immobility. Some clinicians and investigators would also include analgesia and perhaps cardiovascular stability” (Antognini 2010). From the clinical perspective, all anaesthesiologists would agree to state that their primary goal is getting the best possible conditions to safely perform surgical, therapeutic or diagnostic procedures, with the lowest possible morbidity rate. Under these circumstances, amnesia, unconsciousness and immobility may appear as means, rather than main endpoints, to serve the primary objectives: blunt the stress response to the (physical and psychic) surgical trauma and provide the best conditions for the success of the required procedure. Shafer and Stanski (Shafer and Stanski 2008) recently pointed out that analgesia is in fact at least as important as hypnosis during general anaesthesia. Since the early development of general anaesthesia, Crile (1901) defined the concept of anociassociation: “In conscious individuals, all noxious stimuli reach the brain. During general anaesthesia only the traumatic stimuli are perceived centrally while with complete anociassociation all stimuli are blocked” (Vandam 2000). Once again, general anaesthesia is defined as amnesia, unconsciousness and immobility, but not analgesia, which could be provided at this time by loco- regional anaesthetic techniques to reach anociassociation. Shortly after, Cushing (1902) considered also

9 Ketamine pharmacokinetic and pharmacodynamic general anaesthesia combined to opioid analgesics as anociassociation (Vandam 2000). With developments of better techniques and drugs, these first fruits of “balanced anaesthesia” were then complemented by the association of muscle relaxant and premedication to improve the triad immobility, unconsciousness and amnesia of the hypnosis, then named “combined anaesthesia” (Griffith and Johnson 1942). At this time, there was no drug able to produce all these effects adequately without undesirable over-dosage and the association of several drugs owing these respective properties was necessary. Of further interest is the approach of Little and Stephen (1954). They distanced themselves from the means to perform a satisfactory anaesthesia, rather defining basic goals to be achieved: “The patient should undergo surgery safely and pleasantly; The surgeon’s work should be made expeditious; Prompt return to pre-operative physiological Status Quo” (Little and Stephen 1954). The clinical endpoints to reach these goals were described, and included oxygenation, analgesia, removal of waste product, maintenance of circulatory, acid-base and electrolyte equilibrium, the control of undesirable reflex activity, optimal working conditions, and the triad of hypnosis, muscle relaxation and amnesia. As the title of the article indicated in 1954, these tenets for “modern” balanced anaesthesia are holding until today. Nowadays, some single anaesthetic drugs cover all three aspects of the triad and allow for good quality of unconsciousness, muscle relaxation, and amnesia. Therefore, the distinction between each of these parameters is not as important anymore when setting up a clinical anaesthetic plan for standard procedures. Moreover, advanced loco-regional techniques and sedative drugs highlight that adequate analgesia may be enough to allow for performing procedures with neither loss of consciousness nor immobility, providing the patient is cooperative. Recent advances in the understanding of the physiology of nociception and antinociception confirmed these hypotheses, mentioned since the early twentieth century, that hypnosis is not all and anociassociation is superior. If emotional pain may not be integrated and actually felt in absence of consciousness, noxious stimulation, nociception and hyperalgesia still take place and induce the stress response, responsible for physiological instability, metabolic alterations and potentially morbidity. Based on actual knowledge, it would be appropriate to reformulate the definition of anaesthesia and state that there are at least four desired end-points for general anesthesia in animals, namely, (1st.) antinociception and antihyperalgesia (i.e. depression of nociceptive pathways and subsequent consequences), more than analgesia sensu stricto, (2nd.) physiological protection (through stability of cardio-respiratory, acid-base and electrolyte equilibrium, or via controlled depression of potential stress response like sympathetic activity, oxygen demand), (3rd.) vigilance/arousal depression, from anxiolysis to unconsciousness, and (4th.) movement depression (from mild myorelaxation to complete immobility). However, the last two are more means than primary objectives.

This point of view predicts very well the clinical impression that an anaesthetic protocol focusing on amnesia, immobility and unconsciousness (e.g. acepromazine premedication, propofol induction and isoflurane maintenance) may be appropriate for non noxious procedures, but will reveal to be very poor at maintaining a stable and safe peri-anaesthetic phase - even though still better than without general anaesthesia at all - and will provide only minimal amelioration during the post-operative phase. Indeed, a very large clinical interest developed over the last decades for peri-operative analgesia. This is to be understood on the top of adequate hypnosis (“combined anaesthesia”). Nowadays, the term “general anaesthesia” automatically implies hypnosis (unconsciousness and muscle relaxation) and analgesia (Shafer and Stanski 2008), together with general patient safety. In this regard, several analgesic drugs have been described to be administered intravenously concomitantly to hypnotics. Ketamine is one of

10 Ketamine pharmacokinetic and pharmacodynamic those drugs and can be administered at doses which do not target loss of consciousness (subanaesthetic) but elicit antinociception. Ketamine is in use by clinicians since decades in all commonly treated animal species and has been reported in clinical studies to decrease isoflurane requirement, while reducing its undesirable effects and provide partial antinociception. Some clinical regimens of administration have even been proposed. However, only little information is available on its efficacy and potency as an analgesic. This is the focus of the present thesis.

11 Ketamine pharmacokinetic and pharmacodynamic

12 Ketamine pharmacokinetic and pharmacodynamic

6. Introduction

Ketamine is an anaesthetic drug widely used in many animal species since several decades to induce and maintain general anaesthesia (Wright 1982). Its mechanism of action is still unclear. However, beside interactions with multiple receptor types, ketamine antagonistic action at the N-methyl-D- aspartate (NMDA) receptor has been described (Anis, Berry et al. 1983) and appears to play a major role in its anaesthetic properties (Flohr, Glade et al. 1998). The key contribution of NMDA receptors within the nociceptive physiology has also been reported (Haley, Sullivan et al. 1990; Dickenson and Aydar 1991; Murray, Cowan et al. 1991), with particular emphasis on its interplay with central sensitization and hyperalgesia (Woolf and Thompson 1991; Coderre and Melzack 1992; Coderre and Melzack 1992). At this point, it was straightforward to propose that ketamine could have clinically relevant analgesic properties. As expected, this was demonstrated in rats (Alam, Saito et al. 1996). Ketamine is in fact a non-competitive antagonist at the NMDA subtype of glutamate receptor at the phencyclidine binding site, modulating the duration and the frequency of the calcium channel opening followed by a depression of the long term potentiation of C-fibers-evoked potentials (Liu and Sandkuhler 1995). This mechanism is responsible for avoiding wind-up and subsequent central sensitization leading to hyperalgesia and potentially chronic pain states (Orser, Pennefather et al. 1997).

One limitation for the use of ketamine as an analgesic, is the panel of side-effects observed at anaesthetic dosage (i.e. hallucination, psychodyslepsia, excitation, muscle tremors). Its use requires association to sedatives and/or muscle relaxants, and leads to complete loss of consciousness. However, some investigations raised the hypothesis that lower “sub-anaesthetic” doses of ketamine could decrease hyperalgesia and allodynia contributing in a dose-dependent manner to its antinociceptive properties with minimal side-effects. Soon thereafter, clinical investigations demonstrated the analgesic effect of low-dose of ketamine in man (Schmid, Sandler et al. 1999; Snijdelaar, Cornelisse et al. 2004; Richebe, Rivat et al. 2005). Similar trials followed in veterinary medicine (Wagner, Walton et al. 2002; Fielding, Brumbaugh et al. 2006) based on the same assumptions. However, dose and efficacy were extrapolated from human studies and there is good evidence that the relationship between administered dose and obtained plasma level (pharmacokinetic) as well as between plasma level and analgesic effect (pharmacodynamic) vary among species. Moreover, all authors agree on a continuum for ketamine activity from anti-hyperalgesic properties at a very low dose, till deep depression of the central nervous system with profound general anaesthesia at high dose (Richebe, Rivat et al. 2005). More information is available on the effect of anaesthetic dose of ketamine as reported by studies aiming at reducing the minimal alveolar concentration (MAC) of volatile anaesthetic agents (Muir, Wiese et al. 2003; Boscan, Pypendop et al. 2005; Solano, Pypendop et al. 2006). In 2006, only scarce information was available about ketamine‟s analgesic efficacy and adequate dosing in animals. It has to be noted that the pharmacokinetic-pharmacodynamic profile of ketamine is rendered more complicated by its racemic nature. Ketamine is a racemic mixture of two optical isomers R ( - ) and S ( + ) ketamine (Figure 1).

13 Ketamine pharmacokinetic and pharmacodynamic

Both enantiomers are active but S ( + ) ketamine (S-ket) was found to be four-times more potent than R ( - )-ketamine (R-ket) according to its superior binding affinity to the NMDA receptor (Orser, Pennefather et al. 1997). Moreover, even though metabolism pathways are similar, each enantiomer may have a different rate of metabolism depending on species, and concomitant therapy (i.e. administration of premedication or hypnotics) (Geisslinger, Hering et al. 1993; Henthorn, Krejcie et al. 1999; Ihmsen, Geisslinger et al. 2001).

More difficult than the identification of species-specific pharmacokinetic profiles, is the quantification of the pharmacodynamic effects of ketamine. Usually, experimental pain models are invasive, allow only for measurement of thresholds and are not sensitive enough to detect anti-hyperalgesic properties of a drug. Moreover, the experimental testing of analgesics should be performed under circumstances as close as possible to the clinical ones and should allow to study antinociception regardless of potential drug-related behavioural depression. In humans, an objective non-invasive method to study nociception has been developed. The electromyographic analysis of the withdrawal reflex (WR) in response to electrical noxious stimuli has even been described as a means to study temporal summation, a model of wind-up and central sensitization (Andersen, Gracely et al. 1995; Andersen, Jensen et al. 1995; Arendt- Nielsen, Sonnenborg et al. 2000). This model was able to quantify the hypoalgesic profile of ketamine in humans (Arendt-Nielsen, Petersen-Felix et al. 1995; Guirimand, Dupont et al. 2000; You, Morch et al. 2003). Recently, it has been applied and validated in the equine species (Spadavecchia, Spadavecchia et al. 2002; Spadavecchia, Arendt-Nielsen et al. 2003; Spadavecchia, Andersen et al. 2004; Spadavecchia, Arendt-Nielsen et al. 2005).

14 Ketamine pharmacokinetic and pharmacodynamic

7. Aims of the PhD project

In order to gain species-specific knowledge about the effects of ketamine on spinal nociceptive processing in ponies anaesthetized with isoflurane, a set of experiments was planned and two main goals were targeted.

To obtain insights in the stereoselective pharmacokinetics of ketamine in ponies and allow for maintenance of a constant plasma level of ketamine while measuring its analgesic activity. To demonstrate the feasibility of evoking the WR from the forelimb in ponies anaesthetised with isoflurane and study the modulation of the reflex after single and repeated electrical stimulations (model of temporal summation) as a first step before studying its further modulation by ketamine To investigate the pharmacological activity of ketamine of the spinal nociceptive processing (through the WR and temporal summation) during a CRI at sub-anaesthetic dose in ponies anaesthetised with isoflurane. The experimental work has been published in eight (I.-VIII.) different publications. This thesis presents and discusses the experimental work and the results obtained.

(1) Characterization of ketamine pharmacokinetics

The first aim of the study was to maintain ketamine plasma concentration constant over a 2-hours period in ponies anaesthetized with isoflurane, in order to allow for comparable data over time and between subjects with regards to ketamine modulation of the WR. Several steps were required to reach this aim. Firstly, ketamine plasma levels were measured to obtain a dose-concentration time course of ketamine and extract its pharmacokinetic parameters. The first publication (I.) describes an original technique, which was developed to separate and measure plasma levels of ketamine enantiomers and their respective metabolites in plasma samples with minimal blood volume. We tested the hypothesis that capillary electrophoresis is able to separate and quantify both S- and R-ket enantiomers as well as their noketamine metabolites (S-nor and R-nor).

The second publication (II.) describes investigations on the pharmacokinetics of ketamine enantiomers in ponies under isoflurane anaesthesia. Ketamine was administered as a single bolus, and a standard compartmental analysis of the time course was conducted.

Once ketamine pharmacokinetic parameters were obtained, it became possible to predict which administration regimen was required to target a constant plasma concentration at the desired subanaesthetic level. The third publication (III.) describes the technique employed to maintain the plasma level of S-ket constant over time. An algorithm used the pharmacokinetic parameters obtained in publication V to calculate automatically the infusion rate needed to maintain a constant plasma concentration of ketamine in blood. This allowed obtaining comparable results for the pharmacodynamic evaluation step.

15 Ketamine pharmacokinetic and pharmacodynamic

Finally, the fourth publication (IV.) presents an original attempt to fit the complete set of pharmacokinetic data obtained for ketamine enantiomers and their metabolites from publications II. and III. within a Physiologically-Based Pharmacokinetic (PBPk) recirculatory model. In several species, enantioselective pharmacokinetics of ketamine have been observed, with most often S-norketamine (S- nor) being present in blood at higher concentrations than R-nor. This model offers the possibility to test some hypotheses on metabolic pathways in ponies in different situations.

(2) The effect of isoflurane on the Withdrawal Reflex

Before the antinociceptive activity of ketamine could be investigated using the WR model in ponies anaesthetized with isoflurane, the baseline effect of isoflurane had to be characterized. The MAC is admitted to represent a standard of potency for volatile anaesthetics. In the fifth publication (V.), the MAC of isoflurane was determined for each experimental subject in order to provide standardization of the isoflurane dose in further studies. However, the mode of stimulation may influence the MAC. A constant-current transcutaneous electrical stimulation over the palmar nerve is used for the WR model, but has not been investigated previously in a MAC study. Therefore, we also intended to compare different mode of stimulation, and aimed at verifying the hypothesis that a constant-current electrical stimulation over the palmar nerve is comparable to previously described modes. The sixth publication (VI.) investigated the modulation of the WR under isoflurane anaesthesia in ponies for the first time. The WR was elicited by single electrical stimulus as investigated previously (V.). The hypothesis that isoflurane would dose-dependently depress the WR but not inhibit it completely at concentrations around MAC was investigated. The WR model also offers the possibility to observe the modulatory effects of drugs on temporal summation (TS). This is thought to be a peculiar mechanism of action for ketamine according to its NMDA-antagonist activity. However, TS was also depressed by volatile anaesthetics above or at MAC levels in humans. In order to investigate the effect of ketamine, the hypothesis was tested that isoflurane does not depress TS significantly at MAC-level when investigated by the WR model. In the seventh publication (VII.), the facilitation of the WR by repeated electrical stimuli was reported as a measure of temporal summation under isoflurane anaesthesia. The influence of different stimulus intensities and frequencies on temporal summation was evaluated. This was an important basis to understand later in the further experiment to which extent ketamine is able to depress these parameters.

(3) Effects of ketamine on the Withdrawal Reflex

The last goal of the project was to study the activity of ketamine when administered at a subanaesthetic dose in ponies anaesthetized with isoflurane. As presented in the eighth publication (VIII.), the antinociceptive activity of ketamine was determined in ponies receiving isoflurane by analyzing WR responses to single and repeated stimuli. Based on results obtained with isoflurane alone (VI. and VII.), the specific effect of ketamine on WR intensity and latency as well as on the stimulation-response relationship with increasing stimulation intensities, and finally on the depression of temporal summation in response to repeated stimulation was investigated.

16 Ketamine pharmacokinetic and pharmacodynamic

8. Methods and Results

The methods used to perform the different steps of the project are described here. The methods were developed to obtain a high-quality experimental design, and confidence in the obtained results.

(1) Animals

For all steps of this project, eight ponies were used over the period of 3 years of experimental data collection. The ponies were owned by our group (Anaesthesiology Section, Department for Veterinary Clinical Sciences, Vetsuisse Faculty, University of Berne) housed at the Vetsuisse Faculty. The ponies were healthy, underwent regularly a clinical examination, and maintained a good body condition over the whole experimental phase. The animals were all castrated males. All the ponies were familiar with the experimental personnel and the experimental procedures, such that anxiety and stress were kept as low as possible. Prior to the first experimental study, the right carotid artery had been surgically elevated to a subcutaneous position in order to facilitate its catheterization for arterial pressure measurement and arterial blood sampling. Food was withheld in the morning of the experimental sessions. Only one pony at a time was present in the laboratory. The laboratory room was kept at constant temperature (21°C). The ponies were continuously controlled after the experiments for possible complications, or skin alterations at the site of electrode application. No adverse events occurred during the experiments.

(2) Induction of isoflurane general anaesthesia

In all studies presented here, the ponies were anaesthetized with isoflurane. Anaesthesia was induced in a quiet and darkened room. The ponies were positioned in right lateral recumbency by gentle manual restraint, without any medication (V.). Both investigators and ponies were trained for this technique and acquainted. Oxygen (6 L minute-1, 100%) and isoflurane (delivered in a stepwise increase of 0.5% every 30 seconds up to 3%) were administered through a standard anaesthetic circle system via an air-tight face mask. The latter was specifically designed for this purpose (Figure 2). Once the palpebral reflex was absent, the trachea was intubated [14 or 16 mm ID, cuffed endotracheal tube (ETT)] and isoflurane end-tidal concentration maintained at 1.3%. The use of gentle manual restraint and apparently stress- free induction in our Shetland ponies may have minimized the influence of stress on our results.

(3) Ketamine: stereoselective pharmacokinetics

Determination of ketamine concentrations using enantioselective capillary electrophoresis

During all steps of the experiments, plasma samples were obtained from the animals to determine the concentration of ketamine and norketamine enantiomers. A P/ACETM MDQ capillary electrophoresis (CE) system (Beckman Coulter, Fullerton, CA, USA), equipped with an untreated fused-silica capillary (Polymicro Technologies, Phoenix, AZ, USA) of 50 m ID and 38.5 cm (28 cm to the detector) total length, was used for quantification (I.). The enantioselective CE assay was robust and provides an analytical method to monitor ketamine and norketamine enantiomer concentrations >0.01 g/mL in liquid–liquid extracts of 1 mL aliquots of equine plasma. Compared to HPLC methods described in the literature, sample extraction in the CE based assay is simpler and the running costs are considerably lower. Only small amounts of chemicals

17 Ketamine pharmacokinetic and pharmacodynamic and solvents are required. Furthermore, total ketamine and norketamine concentrations monitored with the enantioselective CE assay compare well to those determined by CE in absence of the chiral selector.

Figure 2. Tight respiratory mask used to induce general anaesthesia with isoflurane in unpremedicated ponies.

Figure 3. Position of electrodes for electrical stimulations.

18 Ketamine pharmacokinetic and pharmacodynamic

Stereoselective pharmacokinetics of ketamine as single bolus

Plasma concentrations for the enantiomers of ketamine and norketamine were determined over time after administration of a single bolus of ketamine racemic mixture as well as S-ket alone (II., Figure 8). The pharmacokinetics for all ketamine enantiomers could be well described with a two-compartment model. No significant differences in the pharmacokinetic variables between and within the treatment groups could be detected for the ketamine enantiomers. However, after administration of racemic ketamine, the area under the curve (AUC) and the maximal concentration (Cmax) of R-nor were found to be significantly smaller (P=0.002 and P=0.009, respectively) than those of S-nor showing an enantioselective metabolism of ketamine under these experimental settings (II.).

Figure 8. Mean and SD of plasma concentrations of (A) S-ket and R-ket and (B) S-nor and R-nor after racemic ketamine administration to seven ponies anaesthetized with isoflurane in oxygen. *P<0.05 within treatment group. **P<0.05 between the two treatment groups.

The concentrations of the two ketamine enantiomers in plasma are equal, whereas S-nor is typically found in a larger amount than R-nor. The statistically significant differences found for the S-ket plasma concentrations 2 min after administration of either racemic ketamine or S-ket had no significant impact on the calculated pharmacokinetic variables. However, the statistically significant differences found in the plasma disposition of S and R-nor after administration of the racemate, mirrored by their calculated pharmacokinetic parameters, suggested a highly stereoselective metabolism in ponies anaesthetized with 1 iMAC isoflurane (II.).

The Target-Controlled Infusion

During the experiments for studying the antinociceptive effects of a ketamine intravenous infusion in anaesthetized ponies, a target-controlled infusion (TCI) device was used to maintain a constant plasma level of ketamine over the whole study time in order to obtain comparable results (III.).

Ketamine was administered intravenously as a continuous infusion using a peristaltic pump. The pump was driven by a computer calculating the infusion rate necessary to maintain the plasma level of ketamine constant over time for the duration of the experiment and at a target level set by the anaesthetist. The computer-driven peristaltic pump has been purposely built for the experiment, was calibrated and had a high precision. The infusion rate was based on a prediction according to the pharmacokinetics of ketamine obtained in a previous experiment in similar conditions on the same experimental subjects.

19 Ketamine pharmacokinetic and pharmacodynamic

The classical tri-exponential equation that defines the time course of the molar concentration of a compound in plasma displaying linear pharmacokinetic is given by: - 1t - 2t - 3t C(t) = Y1·e + Y2·e + Y3·e (1) where C(t) is the plasma concentration ( M) at time t, Yi ( M) represents a fraction of the initial dose that is eliminated from plasma with the first-order rate constant i (per hour). When the compound of interest is best represented by a two-compartment model, the third term is set to zero. Assuming a mammillary arrangement of the three compartments and elimination only from the central compartment (Figure 9) , the model has been described by Maitre and Shafer (Maitre and Shafer 1990) using first-order transport rate constants (per minute) for tissue uptake (k12, k13), recycling (k21, k31) and elimination (k10).

Figure 9. A three compartment mammillary pharmacokinetic model. The parameters kij represent the intercompartmental transport rate constant (min-1) between the compartments i and j. The parameter k10 represents the elimination rate constant. The parameters V1 and C(t) represent the volume of the central compartment (compartment 1) and the plasma concentration in this compartment at the time t, respectively.

The time course of the plasma concentration can then be expressed by following equations:

A1(t) = (k21·A2(t-1) + k31·A3(t-1) – (k10 + k12 + k13) ·A1(t-1) + R(t-1)) · t + B(t) (2)

A2(t) = (k12·A1(t-1) – k21·A2(t-1)) · t (3) A3(t) = (k13·A1(t-1) – k31·A3(t-1)) · t (4) where A1(t), A2(t) and A3(t) are the amount of drug ( g) at the time t in compartments 1, 2 and 3, respectively. R(t) is the amount of drug ( g/min) administered intravenously at the time t by continuous infusion during the time interval, and B(t) ( g) the amount administered as a bolus injection. If the drug pharmacokinetic is better described by a two-compartment model, the transport rate constants k13 and k31 have to be set to zero. Eqs. (2)–(4) calculate Ai(t), the approximate change in the amount of drug ( g) in compartment i over the time interval t (min).

20 Ketamine pharmacokinetic and pharmacodynamic

The amount of drug in each compartment at the beginning of the time interval is given by:

Ai(t) = Ai(t-1) + Ai(t) (5) The accuracy of the approximation increases as smaller values for t are chosen (Maitre and Shafer 1990). Finally, the plasma concentration ( g/mL) at the time t is given by: C(t) = A1(t) / (V1·P) (6) where V1 is the volume of the central compartment (mL/kg) and P the body mass (kg) of the individual of interest.

Based on former Eqs. (2)–(6), a TCI algorithm for the prediction of the appropriate infusion rate (IR(t)) starting from a user-defined target plasma level (TC(t), g/mL) was obtained: IR(t) = ((TC(t)·V1) / t) – (A1(t) / (P· t)) – (k21·A2(t) + k31·A3(t) - (k10+k12+k13)·A1(t)) / P (7) The prediction error (PE) (Varvel, Donoho et al. 1992) was calculated for each time point as follows: PE = (Cm-Cp)/Cp x 100 (8)

Based on the two-compartmental model obtained, a TCI aiming at maintaining a constant plasma level of S-ket at 1 g/mL over 2 hours was realized. Based on prediction accuracy, the plasma concentration was satisfactorily maintained in all ponies as show in Figure 10. However, the targeted level was not always reached (Publication III.).

Figure 10. Infusion rates of S-ket and time course of S-ket and S-nor plasma concentrations measured during and after the target-controlled infusion in ponies (n = 6) anesthetized with isoflurane. S-ket was administered intravenously from 0 to 120 min to maintain a target plasma concentration of 1 g/mL. The continuous line represents the individual expected plasma concentrations of S-ket during the infusions. To ease the reading, a dotted line was added to follow the actually measured plasma concentrations of S-ket during the infusion.

21 Ketamine pharmacokinetic and pharmacodynamic

Four variables were calculated from the PE: 1) the median predictive error (MDPE) is a measure of bias per individual and constitutes the median of the PE. 2) The median absolute predictive error (MDAPE) is a measure of the error per individual and constitutes the median of the absolute values of the PE. 3) The divergence estimates the trend away from or toward the targeted concentration and constitutes the slope of the linear regression line when |PE| is plotted against time. Units of divergence are percentage per hour. 4) The wobble is a measure of the prediction instability per individual and is calculated as the median of the absolute value of the difference between MDPE and each PE. If the divergence and the wobble are low (close to zero), the prediction error defined by MDAPE is stable over time and samplings, respectively. Performance parameters are presented in Table 1.

Table 1. Performance parameters of the target-controlled infusion of S-ket (target of 1 g/mL over 120 min) in ponies (n = 6) anaesthetized with isoflurane at 0–10 min and 30–120 min after the beginning of the intravenous infusion of racemic ketamine.

The Physiologically-Based Pharmacokinetic (PBPk) model

Figure 11 illustrates the six-compartment flow-limited PBPk model used to analyze the plasma concentration data on R-ket, S-ket, R-nor and S-nor (publication IV.).

Figure 11. A schematic representation of the six-compartment physiologically based pharmacokinetic (PBPk) model describing the absorption, distribution, metabolism, and elimination of R-ket, S-ket and R-nor, S-nor in ponies.

22 Ketamine pharmacokinetic and pharmacodynamic

The plasma concentration was obtained from three different studies as described previously: two bolus studies (2.2 mg/kg racemic ketamine iv and 1.1 mg/kg S-ket iv) under isoflurane anaesthesia and the target-concentration infusion study in the same animals were fit to the model. Ketamine, as a 50:50 racemic mixture of R-ket and S-ket, was administered via iv dose into the saphenous vein over varying time periods directly into the venous blood where it is considered to become rapidly mixed as it moves through the lung into the arterial blood. From the arterial blood, distribution of R-ket and S-ket to the liver, kidney, fat, rapidly perfused tissues (a grouping of organs such as pancreas and heart) and the slowly perfused tissues (a grouping of organs such as muscle and skin) takes place. The equations used to model the physiology and the distribution of ketamine were given in the publication IV. and closely follow the flow-limited models used by numerous other authors for different compounds (Kohn and Melnick 1996; Reddy, Andersen et al. 2003).

In liver and lung, R-ket and S-ket are assumed to be metabolized through Hill kinetics to R-nor and S- nor, respectively. The Hill kinetic equation used has the general form:

nct Vxct Cct

nct nct (9) kmct Cct

where Cct is the concentration of compound c (R-ket or S-ket) in tissue t (this subscripting is used for all values), Vxct is the maximal rate of reaction (Vmax) for metabolism, kmct is the concentration yielding half of the maximal rate and nct is the Hill coefficient governing the shape of the reaction as C increases. In addition, both R-ket and S-ket are assumed to have a secondary first-order metabolic pathway in both the liver and the lung (IV.). The R-nor and S-nor are followed explicitly through a parallel PBPk model (IV.) with further metabolism via Hill kinetics in both the liver and lung.

The individual pony body weights were used in the model with tissue volumes (VT) calculated as a proportion of mean body weight using literature values. Blood flow rates to the tissues (Qt) were calculated as a proportion of total cardiac output using literature values. Partition coefficients for R-ket and S-ket were calculated based upon the octanol–water partition coefficient for ketamine (3.1) and an allometric relationship for calculating partition coefficients linked to lipid content in tissue. The partition coefficients for R-nor and S-nor were assumed to be 1 based upon the hydrophilicity of norketamine. The partition coefficients are presented in publication IV.

Metabolic parameters were estimated by maximum likelihood estimation where it was assumed that the data for R-ket, S-ket, R-nor and S-nor were log-normally distributed with constant variance over time. Separate variances were estimated for ketamine and norketamine enantiomers. Data below the limit of detection were treated as censored observations. Estimated parameters are presented in the article IV. They included the two variances, the metabolic constants (Vxct, kmct, nct and kect) and the elimination rate constant for R-nor and S-nor from the kidney (kect) where c is the compound (ketamine or norketamine). Confidence bounds for the estimated parameters and statistical tests based on the model were all done using the likelihood ratio test. Standard deviation for the model for ketamine (R or S) and for norketamine (R or S) were 3.03 and 1.00, respectively. The modeling and parameter estimation were done using MATLAB language version 6.5 (The Mathworks Inc., 2002, Natick, MA, USA).

23 Ketamine pharmacokinetic and pharmacodynamic

A PBPk recirculatory model was then built to predict the behavior of S- and R-ket as well as S- and R-nor after single bolus injections and continuous intravenous infusion (Figure 12, IV.). The use of recirculatory kinetic models is also expected to better estimate the early behaviour of rapid-acting intravenous drugs. Moreover, it allows testing hypotheses regarding the reason for ketamine stereospecific kinetics. Two hypotheses were tested to explain the consistently higher plasma concentrations of S-nor in comparison to R-nor. The first hypothesis assumed different biotransformation rates from ketamine enantiomers to R- and S-nor referring to Delatour et al. (Delatour, Benoit et al. 1990) followed by identical rates of biotransformation for norketamine enantiomers to further metabolites. The second hypothesis included different rates of biotransformation for R-nor and S-nor to further metabolites (Delatour, Jaussaud et al. 1991; Schmitz, Theurillat et al. 2009). Trevor et al. found human liver microsomes to exhibit selectivity with respect to the formation of hydroxylated norketamine metabolites, with R- and S-nor undergoing preferential hydroxylation at different positions at the cyclohexanone ring, termed product selectivity (Trevor, Woolf et al. 1983). However, identical biotransformation rates for R- and S-ket are assumed in the second hypothesis. Therefore, it did not consider differences in enzymatic reactions for R- and S-ket to norketamine enantiomers, termed substrate selectivity. Kharasch and Labroo concluded that human ketamine metabolism exhibits moderate degrees of both substrate and product stereoselectivity (Kharasch and Labroo 1992). After simulation of both hypotheses using the PBPk model, we could reject the first hypothesis due to a statistically significant difference (p<0.01) compared to the likelihood value of an overall and unrestricted model, but not the second hypothesis (p>0.05).

24 Ketamine pharmacokinetic and pharmacodynamic

Figure 12. Predicted (dashed line) plasma concentrations and 95% confidence intervals (dotted lines) for R-ket, S-ket, R- nor, and S-nor in one pony following ketamine TCI (A), intravenous bolus administration of 2.2 mg/kg racemic ketamine (B) and intravenous bolus administration of 1.1 mg/kg S-ketamine (C). Open circles represent measured plasma concentrations.

25 Ketamine pharmacokinetic and pharmacodynamic

26 Ketamine pharmacokinetic and pharmacodynamic

(4) Electrophysiological recordings of the Withdrawal Reflexes

The nociception tests used in this project consisted of electrical stimulations applied transcutaneously to a sensory nerve of the limb to elicit a withdrawal reflex response. This electrical stimulation modality was different from the ones presented in previously published equine MAC-determination studies. Therefore, we first intended to compare both techniques. During the MAC determination study (V.), three different modes of electrical stimulation (Figure 3) were applied sequentially: two used constant-voltage and one used constant-current. Details on electrodes and stimulation mode are presented in publications V. to VIII. Placement of the electrodes is presented in Figure 3.

As main findings, it could be well proven that MAC did not differ when the constant-current mode of stimulation was used and compared to constant-voltage mode and it can be applied to improve repeatability (V.).

Figure 4. Electrodes and cables in place ready for the recording session.

27 Ketamine pharmacokinetic and pharmacodynamic

Figure 5. Examples of electromyographic deflection recorded at the deltoid muscle elicited by the withdrawal reflex after single (1 x 5 pulses) electrical stimulation at an intensity of 40 mA.

Figure 6. Examples of electromyographic deflection recorded at the deltoid muscle elicited by the withdrawal reflex after repeated (10 x 5 pulses) electrical stimulation at intensity of 5 (A.), 20 (B.) and 40 (C.) mA.

28 Ketamine pharmacokinetic and pharmacodynamic

The WR model records via electromyography the efferent muscular activity in response to an electrical stimulation of an afferent sensitive nerve (publication II). This model of nociception has been previously described in horses as the nociceptive withdrawal reflex (NWR) (Spadavecchia, Spadavecchia et al. 2002; Spadavecchia, Arendt-Nielsen et al. 2003; Spadavecchia, Andersen et al. 2004; Spadavecchia, Arendt-Nielsen et al. 2005; Spadavecchia, Arendt-Nielsen et al. 2007). The reflex nature of the response is confirmed by its short latency after stimulation. The EMG activity of the WR was expected within the period of 20 to 70 milliseconds after stimulation (VI.). The EMG activity of the WR in response to stimulations was recorded for the common digital extensor and deltoid muscles of the ipsilateral limb (Figure 4).

Evidence of a reflex movement in response to each electrical stimulations was assessed by visual observation, and responses were recorded. To quantify EMG reflex activity in response to single stimulation, the EMG recording was conducted from 100 milliseconds before until 400 milliseconds after a stimulus, which resulted in a total recording time of 500 milliseconds with 512 sample points (sampling frequency, 1,024 Hz). To be considered a reflex response, the EMG burst following stimulation had to be at least 3 times the amplitude of the background activity with duration of at least 10 milliseconds within the period from 20 to 70 milliseconds after stimulation onset (Figure 5), to target the response to A -fibers activity (nociceptive activation of the WR) (Spadavecchia, Spadavecchia et al. 2002).

To record and quantify the EMG reflex activity in response to repeated stimulation (Figure 6), EMG activity was stored from 500 ms before until 1500 ms after the stimulation ended, resulting in a total recording time of 4000 ms (sampling frequency, 1 kHz). Details on the effect of isoflurane aloneor during ketamine administration are given in publications VI, VII and VIII, respectively.

Although isoflurane abolishes the electrically induced NWR in humans at concentrations substantially lower than 1.0 MAC (Petersen-Felix, Arendt-Nielsen et al. 1996), it did not occur to be the same in our ponies, even at approximately MAC, suggesting species-specific anesthetic modulation of spinal withdrawal reflexes despite homogeneous interspecific MAC values (Sani and Shafer 2003). At all tested MAC multiples, amplitude of the reflex response increased with stimulus intensity, whereas the slope of the stimulus-response function was reduced with increasing isoflurane concentrations (VI.).

When repeated stimulations were applied to evoke temporal summation (TS), reflexes were not abolished at isoflurane concentrations that were able to prevent purposeful movements in response to supra-maximal noxious stimulations (VII.). This is consistent with the results reported by Petersen-Felix et al. (Petersen-Felix, Arendt-Nielsen et al. 1995; Petersen-Felix, Arendt-Nielsen et al. 1996) who found that TS, but not reflexes to single stimuli, could still be evoked at isoflurane concentrations around MAC in humans.

Racemic ketamine significantly and dose-dependently depressed the reflex EMG activity (VIII.) when added to isoflurane anaesthesia administered at iMAC (Figure 7). The reflex activity recovered rapidly after the end of ketamine infusion. However, it did not return to pre-infusion values within the study time (90 min after cessation of ketamine administration), despite the rapid decrease of ketamine plasma concentrations. In the present study, no facilitation of the WR, as depicted by a lower threshold compared to single stimulations or increasing response amplitude throughout repeated stimulations could be observed after repeated stimulations during baseline measurements under isoflurane anaesthesia. Thus, no significant effect of ketamine on these parameters was determined. Ketamine equally decreased the responses to single and repeated stimulations (VIII.).

29 Ketamine pharmacokinetic and pharmacodynamic

Figure 7. Examples of electromyographic deflection recorded at the deltoid muscle elicited by the withdrawal reflex after single (A, 1 x 5 pulses) or repeated (B, 10 x 5 pulses at 5 Hz) electrical stimulation at increasing intensities (5, 20 and 40 mA) before (T0) and during (T2) R- /S-ket target-controlled infusion in one pony.

30 Ketamine pharmacokinetic and pharmacodynamic

9. Conclusion and Future applications

In the present work, we were able to develop an analytical assay to quantify plasma concentrations of ketamine‟s enantiomers and their metabolites based on capillary electrophoresis. This method has advantages over high-pressure liquid chromatography regarding sample amount and costs. The pharmacokinetic profile of ketamine‟s enantiomers could be studied and a TCI algorithm based on these results allowed for maintaining satisfactorily constant plasma levels over two hours. The complete set of data from both studies led to the elaboration of a physiologically-based pharmacokinetic recirculatory model giving further insights into stereoselective pharmacokinetics of ketamine. In parallel, the nociceptive withdrawal reflex model was used to define spinally-modulated antinociception of isoflurane and the additional effect of ketamine in ponies after single or repeated stimulation. A species- specific modulatory effect of isoflurane was observed with a significant but incomplete depression of both responses to single and to repeated stimulation. These observations were compared and linked to the usual MAC determination method, showing that isoflurane concentrations required to abolish purposeful movement in response to supramaximal stimulation are at about the level initiating spinal nociceptive reflex depression in ponies. Ketamine dose-dependently abolished the nociceptive withdrawal reflex in anaesthetized ponies indicating antinociceptive properties at sub-anaesthetic levels. It was hypothesized that the antagonistic activity of ketamine at NMDA receptors would results in a depression of the temporal summation, and this phenomenon could be evaluated by the nociceptive withdrawal reflex model. This could not be verified because the indicators of temporal summation were already inhibited by isoflurane during baseline recordings. Therefore, specific effect of ketamine on temporal summation could not be evaluated. As an overall conclusion, we succeeded to provide further insights within pharmacokinetic-pharmacodynamic profile of ketamine as an analgesic for rational intravenous infusion in ponies thanks to a series of original research studies. Both metabolism and distribution of drugs may be partially driven by active binding mechanism to membrane proteins (transport proteins or enzymes) and can therefore exhibit stereospecificity (Renkin and Curry 1982). However, several authors failed to demonstrate stereoselective ketamine distribution in humans (Geisslinger, Hering et al. 1993; Ihmsen, Geisslinger et al. 2001; Persson, Hasselstrom et al. 2002) and in dogs (Henthorn, Krejcie et al. 1999). In the present study, indicators of drug distribution like half-lives and Vc did not differ for the ketamine enantiomers between and within groups. These results do not support a stereoselective distribution for ketamine in the equine species when isoflurane is administered.

However, there are probably some differences with regard to the metabolic pathway of R- and S-nor, as shown by the presented results in the racemic ketamine treatment group. The hydroxylation of norketamine already exhibited enantioselectivity at different positions on the cyclohexanone ring. Therefore, it is possible that R-nor-hydroxylation prevails in one stereospecific pathway more than in the other. This hypothesis could be the consequence of substrate enantioselectivity of the cytochrome P450-dependent N-demethylation. In the present study, no R-ket was detected after administration of the pure S-enantiomer, and after administration of the racemate, the mean plasma concentrations of R- and S-ket were similar, suggesting no inversion between the R- and S-enantiomers. This has been supported by in vitro findings with liver microsomes (Schmitz, Portier et al. 2008).

The development of a recirculatory model for the pharmacokinetic behavior of ketamine‟s enantiomers and their respective metabolites helped to investigate hypotheses on the metabolic pathways responsible for their stereoselective disposition. However the conclusion still results from a

31 Ketamine pharmacokinetic and pharmacodynamic model and we could not conclude on real in vivo mechanisms. As a continuation of our studies, in vitro studies investigating the chemical transformation of ketamine‟s enantiomers and further degradation products by liver and lung cytochrome enzymes have been performed by collaborators of our team (Schmitz, Portier et al. 2008; Capponi, Schmitz et al. 2009; Schmitz, Thormann et al. 2010). In the future, such data could be introduced within the physiologically-based pharmacokinetic model to refine it. Similarly, further in vivo studies could be performed to provide additional insights and finally better understand pharmacokinetic differences associated to each ketamine enantiomers. A TCI for ketamine in anaesthetized ponies was achieved. In some subjects, the plasma level was higher than expected. Retrospectively, the main source of error was likely initiated by overdosing of the loading bolus (publication VI). In a standard compartmental model, the initial bolus dose calculation is based on the volume of the central compartment (V1) (Schwilden, Stoeckel et al. 1986). This parameter has been identified to be a weak point of such polyexponential models, which misspecifies the early time course of drug concentration assuming several simplifications far from reality (immediate distribution in the blood compartment, independent of cardiac output) (Fisher 1996; Krejcie and Avram 1999). Thus, the overestimated V1 was found to be responsible for common overdosing at the beginning of TCIs based on traditional kinetic models and its adjustment has been proposed to improve their performance (Avram and Krejcie 2003). A better understanding could also be obtained from the recirculatory model. The PBPk model could also further help to investigate the variations of the TCI regimen in different situations.

One aspect of the results presented here, is the application of the NWR model during MAC determination for an inhalant anaesthetic. The first study which was essential for the thesis discussed the validity of using the observation of a complex purposeful movement in response to supramaximal stimulation as end-point for MAC determination and thus as a measure of anaesthetic potency. Inhalation anaesthetics are mainly known to exert their effects at spinal level (Antognini 2010). Therefore, a method that allows more precise quantification of anaesthetic spinal depressant activity like the NWR seems to offer significant advantages over the classical MAC determination method. Indeed, not only could we describe the use of the NWR model to more precisely standardize, observe and quantify the response to noxious stimulation under different levels of isoflurane anaesthesia in equine, but we could also provide evidence of specific isoflurane effects on temporal summation, which is known to play an important but hidden role in the MAC determination process (Dutton, Zhang et al. 2003). This work opens a new method to better understand mechanism and quantification of anaesthetic effects like discrimination between peripheral, spinal and cerebral depression (Haga, Ranheim et al. 2009). Low dose ketamine administration to conscious patients has been reported to have little influence on the WR in response to single stimulation in humans (Arendt-Nielsen, Petersen-Felix et al. 1995) and in dogs (Bergadano, Andersen et al. 2009). In ponies, comparable plasma concentrations of ketamine (50 ng/mL) were associated with a mild but significant depression of the reflex amplitudes (Peterbauer, Larenza et al. 2008). If such low doses are generally not exceeded in conscious patients to avoid undesirable psychomimetic side effects (Schmid, Sandler et al. 1999), higher doses can be administered under general anaesthesia. In the present study, higher plasma levels of ketamine were reached (2– 4 g/mL) in our anaesthetised ponies. The responses to single stimulations were strongly depressed or even completely abolished in several ponies, and a flattening of the recruitment curves was observed in all individuals. This outcome might be explained by a dose-dependent modulation of the WR by ketamine and/or by a synergistic interaction with isoflurane. Owing to its antagonistic property at NMDA receptors, ketamine has been described to affect both the spinal „wind-up‟ and the development of central sensitization (Woolf and Thompson 1991). Low doses of ketamine limit the development of central sensitization in animal and human pain models, as well as in the post-operative period in the

32 Ketamine pharmacokinetic and pharmacodynamic clinic, where it has been described to be anti-hyperalgesic more than analgesic (Richebe, Rivat et al. 2005). The specific effect of ketamine on the WR is most likely due to inhibiting the activation of the NMDA receptors on wide dynamic range (WDR) neurons after sustained discharge to the dorsal horn from A fibres, with amelioration of the subsequent facilitation of the polysynaptic reflex (i.e. expansion of the receptive field, decreased stimulation threshold, increased response amplitude and duration). Unfortunately, because no facilitation of the WR could be observed after repeated stimulations during baseline measurements under isoflurane anaesthesia, no significant effect of ketamine on these parameters was determined. The biggest research field opened by this series of experimental work is the species-specific spinal modulation of nociception by ketamine. Our collaborators could already follow on further experiments based on the knowledge acquired here as the effect of ketamine intravenous infusion in awake ponies (Peterbauer, Larenza et al. 2008) and in dogs (Bergadano, Andersen et al. 2009). Ketamine also needs to be investigated in interaction with other commonly used analgesics, with a large interest on its complementary activity with opioids as NMDA antagonists may inhibit hyperalgesia and tolerance (Pascual, Goicoechea et al. 2010). Besides ketamine, its metabolite norketamine has also been proven to elicit antinociceptive activity. Norketamine was found to reduce pain in rodent models (Holtman, Crooks et al. 2008). Prolonged antinociceptive effect of ketamine could partly rely on norketamine enantiomers (Huge, Lauchart et al. 2010). As plasma concentrations of norketamine are maintained longer than those of ketamine, their role in prolonged analgesia after recovery has been hypothesized (Sakai, Mi et al. 1998). Even though we could provide some insights on the pharmacokinetic of norketamine‟s enantiomers, our studies could not discriminate the relative role of ketamine and norketamine on analgesia, but the NWR depression did not return to baseline value when ketamine plasma concentrations decreased while norketamine levels were still observable. Further investigations may be required to better define the role of norketamine enantiomers in clinical analgesia. Ketamine is an old drug in the anaesthetic pharmacy, but requires further investigations to argue at an evidence-based level on the large range of properties which are newly attributed to it. This does not only include the field of pain therapy as discussed in the present thesis, but also finds further applications as an anaesthetic for newborn, in head-trauma patients, as well as a new antidepressant drug (Persson 2008; aan het Rot, Collins et al. 2010; Persson 2010).

33 Ketamine pharmacokinetic and pharmacodynamic

34 Ketamine pharmacokinetic and pharmacodynamic

10. References

aan het Rot, M., K. A. Collins, et al. (2010). "Safety and efficacy of repeated-dose intravenous ketamine for treatment-resistant depression." Biol Psychiatry 67(2): 139-45. Alam, S., Y. Saito, et al. (1996). "Antinociceptive effects of epidural and intravenous ketamine to somatic and visceral stimuli in rats." Can J Anaesth 43(4): 408-13. Andersen, O. K., R. H. Gracely, et al. (1995). "Facilitation of the human nociceptive reflex by stimulation of A beta-fibres in a secondary hyperalgesic area sustained by nociceptive input from the primary hyperalgesic area." Acta Physiol Scand 155(1): 87-97. Andersen, O. K., L. M. Jensen, et al. (1995). "Modulation of the human nociceptive reflex by cyclic movements." Eur J Appl Physiol Occup Physiol 70(4): 311-21. Anis, N. A., S. C. Berry, et al. (1983). "The dissociative anaesthetics, ketamine and phencyclidine, selectively reduce excitation of central mammalian neurones by N-methyl-aspartate." Br J Pharmacol 79(2): 565- 75. Antognini, J. F. (2010). "Anesthetic action: The importance of the spinal cord to immobility." Vet J Epub[Ahead of print]. Arendt-Nielsen, L., S. Petersen-Felix, et al. (1995). "The effect of N-methyl-D-aspartate antagonist (ketamine) on single and repeated nociceptive stimuli: a placebo-controlled experimental human study." Anesth Analg 81(1): 63-8. Arendt-Nielsen, L., F. A. Sonnenborg, et al. (2000). "Facilitation of the withdrawal reflex by repeated transcutaneous electrical stimulation: an experimental study on central integration in humans." Eur J Appl Physiol 81(3): 165-73. Avram, M. J. and T. C. Krejcie (2003). "Using front-end kinetics to optimize target-controlled drug infusions." Anesthesiology 99(5): 1078-86. Bergadano, A., O. K. Andersen, et al. (2009). "Plasma levels of a low-dose constant-rate-infusion of ketamine and its effect on single and repeated nociceptive stimuli in conscious dogs." Vet J 182(2): 252-60. Boscan, P., B. H. Pypendop, et al. (2005). "Cardiovascular and respiratory effects of ketamine infusions in isoflurane-anesthetized dogs before and during noxious stimulation." Am J Vet Res 66(12): 2122-9. Capponi, L., A. Schmitz, et al. (2009). "In vitro evaluation of differences in phase 1 metabolism of ketamine and other analgesics among humans, horses, and dogs." Am J Vet Res 70(6): 777-86. Coderre, T. J. and R. Melzack (1992). "The contribution of excitatory amino acids to central sensitization and persistent nociception after formalin-induced tissue injury." J Neurosci 12(9): 3665-70. Coderre, T. J. and R. Melzack (1992). "The role of NMDA receptor-operated calcium channels in persistent nociception after formalin-induced tissue injury." J Neurosci 12(9): 3671-5. Delatour, P., E. Benoit, et al. (1990). "Chirality of the sulphoxide metabolites of fenbendazole and albendazole in sheep." J Vet Pharmacol Ther 13(4): 361-6. Delatour, P., P. Jaussaud, et al. (1991). "Enantioselective N-demethylation of ketamine in the horse." J Vet Pharmacol Ther 14(2): 209-12. Dickenson, A. H. and E. Aydar (1991). "Antagonism at the glycine site on the NMDA receptor reduces spinal nociception in the rat." Neurosci Lett 121(1-2): 263-6. Dutton, R. C., Y. Zhang, et al. (2003). "Temporal summation governs part of the minimum alveolar concentration of isoflurane anesthesia." Anesthesiology 98(6): 1372-7. Fielding, C. L., G. W. Brumbaugh, et al. (2006). "Pharmacokinetics and clinical effects of a subanesthetic continuous rate infusion of ketamine in awake horses." Am J Vet Res 67(9): 1484-90. Fisher, D. M. (1996). "(Almost) everything you learned about pharmacokinetics was (somewhat) wrong!" Anesth Analg 83(5): 901-3. Flohr, H., U. Glade, et al. (1998). "The role of the NMDA synapse in general anesthesia." Toxicol Lett 100-101: 23-9.

35 Ketamine pharmacokinetic and pharmacodynamic

Geisslinger, G., W. Hering, et al. (1993). "Pharmacokinetics and pharmacodynamics of ketamine enantiomers in surgical patients using a stereoselective analytical method." Br J Anaesth 70(6): 666-71. Griffith, H. R. and G. E. Johnson (1942). "The use of curare in general anesthesia." Anesthesiology 3: 418. Guirimand, F., X. Dupont, et al. (2000). "The effects of ketamine on the temporal summation (wind-up) of the R(III) nociceptive flexion reflex and pain in humans." Anesth Analg 90(2): 408-14. Haga, H. A., B. Ranheim, et al. (2009). "Effects of isoflurane upon minimum alveolar concentration and cerebral cortex depression in pigs and goats: An interspecies comparison." Vet J. Haley, J. E., A. F. Sullivan, et al. (1990). "Evidence for spinal N-methyl-D-aspartate receptor involvement in prolonged chemical nociception in the rat." Brain Res 518(1-2): 218-26. Henthorn, T. K., T. C. Krejcie, et al. (1999). "Ketamine distribution described by a recirculatory pharmacokinetic model is not stereoselective." Anesthesiology 91(6): 1733-43. Holtman, J. R., Jr., P. A. Crooks, et al. (2008). "Effects of norketamine enantiomers in rodent models of persistent pain." Pharmacol Biochem Behav 90(4): 676-85. Huge, V., M. Lauchart, et al. (2010). "Effects of low-dose intranasal (S)-ketamine in patients with neuropathic pain." Eur J Pain 14(4): 387-94. Ihmsen, H., G. Geisslinger, et al. (2001). "Stereoselective pharmacokinetics of ketamine: R(-)-ketamine inhibits the elimination of S(+)-ketamine." Clin Pharmacol Ther 70(5): 431-8. Kharasch, E. D. and R. Labroo (1992). "Metabolism of ketamine stereoisomers by human liver microsomes." Anesthesiology 77(6): 1201-7. Kohn, M. C. and R. L. Melnick (1996). "Effects of the structure of a toxicokinetic model of butadiene inhalation exposure on computed production of carcinogenic intermediates." Toxicology 113(1-3): 31-9. Krejcie, T. C. and M. J. Avram (1999). "What determines anesthetic induction dose? It's the front-end kinetics, doctor!" Anesth Analg 89(3): 541-4. Little, D. M., Jr. and C. R. Stephen (1954). "Modern balanced anesthesia; a concept." Anesthesiology 15(3): 246- 61. Liu, X. G. and J. Sandkuhler (1995). "Long-term potentiation of C-fiber-evoked potentials in the rat spinal dorsal horn is prevented by spinal N-methyl-D-aspartic acid receptor blockage." Neurosci Lett 191(1-2): 43-6. Maitre, P. O. and S. L. Shafer (1990). "A simple pocket calculator approach to predict anesthetic drug concentrations from pharmacokinetic data." Anesthesiology 73(2): 332-6. Muir, W. W., 3rd, A. J. Wiese, et al. (2003). "Effects of morphine, lidocaine, ketamine, and morphine-lidocaine- ketamine drug combination on minimum alveolar concentration in dogs anesthetized with isoflurane." Am J Vet Res 64(9): 1155-60. Murray, C. W., A. Cowan, et al. (1991). "Neurokinin and NMDA antagonists (but not a kainic acid antagonist) are antinociceptive in the mouse formalin model." Pain 44(2): 179-85. Orser, B. A., P. S. Pennefather, et al. (1997). "Multiple mechanisms of ketamine blockade of N-methyl-D- aspartate receptors." Anesthesiology 86(4): 903-17. Pascual, D., C. Goicoechea, et al. (2010). "Antinociceptive effect of three common analgesic drugs on peripheral neuropathy induced by paclitaxel in rats." Pharmacol Biochem Behav 95(3): 331-7. Persson, J. (2008). "The ketamine enigma." Acta Anaesthesiol Scand 52(4): 453-5. Persson, J. (2010). "Wherefore ketamine?" Curr Opin Anaesthesiol. Persson, J., J. Hasselstrom, et al. (2002). "Pharmacokinetics and non-analgesic effects of S- and R-ketamines in healthy volunteers with normal and reduced metabolic capacity." Eur J Clin Pharmacol 57(12): 869-75. Peterbauer, C., P. M. Larenza, et al. (2008). "Effects of a low dose infusion of racemic and S-ketamine on the nociceptive withdrawal reflex in standing ponies." Vet Anaesth Analg 35(5): 414-23. Petersen-Felix, S., L. Arendt-Nielsen, et al. (1996). "The effects of isoflurane on repeated nociceptive stimuli (central temporal summation)." Pain 64(2): 277-81. Petersen-Felix, S., L. Arendt-Nielsen, et al. (1995). "Analgesic effect in humans of subanaesthetic isoflurane concentrations evaluated by experimentally induced pain." Br J Anaesth 75(1): 55-60. Reddy, M. B., M. E. Andersen, et al. (2003). "Physiological modeling of inhalation kinetics of octamethylcyclotetrasiloxane in humans during rest and exercise." Toxicol Sci 72(1): 3-18.

36 Ketamine pharmacokinetic and pharmacodynamic

Renkin, E. M. and F. E. Curry (1982). "Endothelial permeability: pathways and modulations." Ann N Y Acad Sci 401: 248-59. Richebe, P., C. Rivat, et al. (2005). "[Low doses ketamine: antihyperalgesic drug, non-analgesic]." Ann Fr Anesth Reanim 24(11-12): 1349-59. Sakai, T., W. D. Mi, et al. (1998). "[Pharmacokinetics of propofol and ketamine during and after total intravenous anesthesia with propofol, fentanyl and ketamine for pediatric patients]." Masui 47(3): 277- 80. Sani, O. and S. L. Shafer (2003). "MAC Attack?" Anesthesiology 99(6): 1249-50. Schmid, R. L., A. N. Sandler, et al. (1999). "Use and efficacy of low-dose ketamine in the management of acute postoperative pain: a review of current techniques and outcomes." Pain 82(2): 111-25. Schmitz, A., C. J. Portier, et al. (2008). "Stereoselective biotransformation of ketamine in equine liver and lung microsomes." J Vet Pharmacol Ther 31(5): 446-55. Schmitz, A., R. Theurillat, et al. (2009). "CE provides evidence of the stereoselective hydroxylation of norketamine in equines." Electrophoresis 30(16): 2912-21. Schmitz, A., W. Thormann, et al. (2010). "Enantioselective CE analysis of hepatic ketamine metabolism in different species in vitro." Electrophoresis 31(9): 1506-16. Schwilden, H., H. Stoeckel, et al. (1986). "Pharmacological models and their use in clinical anaesthesia." Eur J Anaesthesiol 3(3): 175-208. Shafer, S. L. and D. R. Stanski (2008). "Defining depth of anesthesia." Handb Exp Pharmacol(182): 409-23. Snijdelaar, D. G., H. B. Cornelisse, et al. (2004). "A randomised, controlled study of peri-operative low dose s(+)- ketamine in combination with postoperative patient-controlled s(+)-ketamine and morphine after radical prostatectomy." Anaesthesia 59(3): 222-8. Solano, A. M., B. H. Pypendop, et al. (2006). "Effect of intravenous administration of ketamine on the minimum alveolar concentration of isoflurane in anesthetized dogs." Am J Vet Res 67(1): 21-5. Spadavecchia, C., O. K. Andersen, et al. (2004). "Investigation of the facilitation of the nociceptive withdrawal reflex evoked by repeated transcutaneous electrical stimulations as a measure of temporal summation in conscious horses." Am J Vet Res 65(7): 901-8. Spadavecchia, C., L. Arendt-Nielsen, et al. (2003). "Comparison of nociceptive withdrawal reflexes and recruitment curves between the forelimbs and hind limbs in conscious horses." Am J Vet Res 64(6): 700- 7. Spadavecchia, C., L. Arendt-Nielsen, et al. (2005). "Effect of romifidine on the nociceptive withdrawal reflex and temporal summation in conscious horses." Am J Vet Res 66(11): 1992-8. Spadavecchia, C., L. Arendt-Nielsen, et al. (2007). "Effects of butorphanol on the withdrawal reflex using threshold, suprathreshold and repeated subthreshold electrical stimuli in conscious horses." Vet Anaesth Analg 34(1): 48-58. Spadavecchia, C., L. Spadavecchia, et al. (2002). "Quantitative assessment of nociception in horses by use of the nociceptive withdrawal reflex evoked by transcutaneous electrical stimulation." Am J Vet Res 63(11): 1551-6. Trevor, A. J., T. F. Woolf, et al. (1983). Stereoselective metabolism of ketamine enantiomers, phencyclidine and related arylcyclohexamines: present and future applications. NPP Books. J. M. Kamenka, E. F. Domino and P. Geneste, Ann Arbor: pp. 279-289. Vandam, L. D. (2000). History of anesthetic practice. Anesthesia. R. D. Miller. Philadelphia, Churchill Livingston. 1: 1-14. Varvel, J. R., D. L. Donoho, et al. (1992). "Measuring the predictive performance of computer-controlled infusion pumps." J Pharmacokinet Biopharm 20(1): 63-94. Wagner, A. E., J. A. Walton, et al. (2002). "Use of low doses of ketamine administered by constant rate infusion as an adjunct for postoperative analgesia in dogs." J Am Vet Med Assoc 221(1): 72-5. Woolf, C. J. and S. W. Thompson (1991). "The induction and maintenance of central sensitization is dependent on N-methyl-D-aspartic acid receptor activation; implications for the treatment of post-injury pain hypersensitivity states." Pain 44(3): 293-9.

37 Ketamine pharmacokinetic and pharmacodynamic

Wright, M. (1982). "Pharmacologic effects of ketamine and its use in veterinary medicine." J Am Vet Med Assoc 180(12): 1462-71. You, H. J., C. D. Morch, et al. (2003). "Role of central NMDA versus non-NMDA receptor in spinal withdrawal reflex in spinal anesthetized rats under normal and hyperexcitable conditions." Brain Res 981(1-2): 12- 22.

38 Ketamine pharmacokinetic and pharmacodynamic

Declaration of Originality

Last name, first name: LEVIONNOIS, Olivier

Matriculation number: 05-118-690

I hereby declare that this thesis represents my original work and that I have used no other sources except as noted by citations.

All data, tables, figures and text citations which have been reproduced from any other source, including the internet, have been explicitly acknowledged as such.

I am aware that in case of non-compliance, the Senate is entitled to divest me of the doctorate degree awarded to me on the basis of the present thesis, in accordance with the “Statut der Universität Bern (Universitätsstatut; UniSt)”, Art. 20, of 17 December 1997.

Place, date Signature

…..Bern, 15th December 2010……………

39 Ketamine pharmacokinetic and pharmacodynamic

40 Ketamine pharmacokinetic and pharmacodynamic

Appendix: Publications

1. (III.) Levionnois OL, Mevissen M, Thormann W, Spadavecchia C. Assessing the efficiency of a pharmacokinetic-based algorithm for target-controlled infusion of ketamine in ponies. Res Vet Sci. 2010 Jun;88(3):512-8.

2. (IV.) Knobloch M, Portier CJ, Levionnois OL, Theurillat R, Thormann W, Spadavecchia C, Mevissen M. Antinociceptive effects, metabolism and disposition of ketamine in ponies under target-controlled drug infusion. Toxicol Appl Pharmacol. 2006 Nov 1;216(3):373-86.

3. (V.) Levionnois OL, Spadavecchia C, Kronen PW, Schatzmann U. Determination of the minimum alveolar concentration of isoflurane in Shetland ponies using constant current or constant voltage electrical stimulation. Vet Anaesth Analg. 2009 Jan;36(1):9-17.

4. (VI.) Spadavecchia C, Levionnois O, Kronen PW, Leandri M, Spadavecchia L, Schatzmann U. Evaluation of administration of isoflurane at approximately the minimum alveolar concentration on depression of a nociceptive withdrawal reflex evoked by transcutaneous electrical stimulation in ponies. Am J Vet Res. 2006 May;67(5):762-9.

5. (VII.) Spadavecchia C, Levionnois O, Kronen P, Andersen OK. The effects of isoflurane minimum alveolar concentration on withdrawal reflex activity evoked by repeated transcutaneous electrical stimulation in ponies. Vet J. 2010 Mar;183(3):337-44.

6. (VIII.) Levionnois OL, Menge M, Thormann W, Mevissen M, Spadavecchia C. Effect of ketamine on the limb withdrawal reflex evoked by transcutaneous electrical stimulation in ponies anaesthetised with isoflurane. Vet J. 2010 Dec;186(3):304-11.

41 Research in Veterinary Science 88 (2010) 512–518

Contents lists available at ScienceDirect

Research in Veterinary Science

journal homepage: www.elsevier.com/locate/rvsc

Assessing the efﬁciency of a pharmacokinetic-based algorithm for target-controlled infusion of ketamine in ponies

O.L. Levionnois a,*, M. Mevissen b, W. Thormann c, C. Spadavecchia a a Division of Veterinary Anesthesiology, Department of Clinical Veterinary Medicine, Vetsuisse Faculty, University of Bern, Switzerland b Division of Veterinary Pharmacology and Toxicology, Department of Clinical Research and Veterinary Public Health, Vetsuisse Faculty, University of Bern, Switzerland c Department of Clinical Pharmacology and Visceral Research, University of Bern, Switzerland article info abstract

Article history: The objective of this study was to assess a pharmacokinetic algorithm to predict ketamine plasma con- Accepted 6 December 2009 centration and drive a target-controlled infusion (TCI) in ponies. Firstly, the algorithm was used to simulate the course of ketamine enantiomers plasma concentrations after the administration of an intravenous bolus in six ponies based on individual pharmacokinetic Keywords: parameters obtained from a previous experiment. Using the same pharmacokinetic parameters, a TCI Target-controlled infusion of S-ketamine was then performed over 120 min to maintain a concentration of 1 lg/mL in plasma. Pharmacokinetic algorithm The actual plasma concentrations of S-ketamine were measured from arterial samples using capillary Predictive efﬁcacy electrophoresis. Ketamine Equine The performance of the simulation for the administration of a single bolus was very good. During the TCI, the S-ketamine plasma concentrations were maintained within the limit of acceptance (wobble and divergence <20%) at a median of 79% (IQR, 71–90) of the peak concentration reached after the initial bolus. However, in three ponies the steady concentrations were signiﬁcantly higher than targeted. It is hypothesized that an inaccurate estimation of the volume of the central compartment is partly responsible for that difference. The algorithm allowed good predictions for the single bolus administration and an appropriate maintenance of constant plasma concentrations. Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction the drug concentration, a process referred to as ‘‘Titration of Intra- venous Agent by Computer” (TIAC, Absalom and Struys, 2007). Fur- When administering analgesic or anesthetic drugs intrave- thermore, this software can drive an infusion pump by frequently nously, it is of considerable advantage to maintain a stable efficacy adjusting the administered dose at its minimum necessary to target over time. This is ideally achieved with a steady-state blood con- and maintain in plasma a drug concentration as defined by the centration. For this purpose, Kruger-Thiemer (1968) described the anesthetist. This process is called target-controlled infusion (TCI, use of a classical pharmacokinetic compartmental model to opti- Absalom and Struys, 2007). When the plasma concentration is only mize drug administration. Standard pharmacokinetic parameters predicted by an algorithm, but not adjusted to actually measured including volume of distribution, clearance and elimination half- plasma samples, the TCI targeting a given plasma concentration is time allow a prediction of approximate doses for bolus and basic defined as ‘‘open-loop” (Absalom and Struys, 2007). Even though infusion schemes. However, accurate prediction of plasma levels better control of a plasma level can be achieved using TCI (Servin, after multiple dosing and prolonged infusions at various rates, as 1998), the clinical endpoints observed at the patient side by the it occurs during the course of a typical anesthesia, requires a more anesthetist are still the main determinants of the clinical effect of complex algorithm (Maitre and Shafer, 1990). Advanced technolo- the administered regimen. In complement to constant-rate infusion gies for intra-operative computer-assisted prediction of plasma lev- (CRI), TCI allows to reach more rapidly the desired drug concentra- els based on pharmacokinetic equations have been developed in tion in the blood while reducing drug accumulation and subsequent human anesthesia over the last years (Schwilden and Schuttler, hangover (Russell, 1998; Hu et al., 2005). To date, despite the equa- 2008). Specific softwares allow the clinician to predict and titrate tions being known and drug-specific software having been developed for use in humans, no simple software is commercially available for the incorporation of veterinary pharmacokinetic data. * Corresponding author. Address: Division of Veterinary Anesthesiology, Depart- The purpose of this study was to generate a spreadsheet to pro- ment of Clinical Veterinary Medicine, Vetsuisse Faculty, University of Bern, P.O. 8466, 3001 Bern, Switzerland. Tel.: +41 316312287; fax: +41 316312620. vide a simulation of the dose-concentration course of ketamine E-mail address: [email protected] (O.L. Levionnois). enantiomers in ponies under isoflurane anesthesia (Larenza et al.,

2007). The algorithm is based on the pharmacokinetic equations used in traditional compartmental models. Based on this algorithm, an open-loop, plasma-targeted TCI of ketamine, a common component of veterinary balanced anesthesia (Knobloch et al., 2006; Solano et al., 2006; Bettschart and Larenza, 2007), was delivered to ponies under isoﬂurane anesthesia using a computer-driven pump to test the in vivo performance of the predictions.

2. Methods

Fig. 1. A three compartment mammillary pharmacokinetic model. The parameters 2.1. Spreadsheet for prediction of plasma concentration curve 1 kij represent the intercompartmental transport rate constant (min ) between the

compartments i and j. The parameter k10 represents the elimination rate constant.

2.1.1. Equations for the algorithm of the spreadsheet The parameters V1 and C(t) represent the volume of the central compartment This part aims at obtaining an algorithm to predict the time (compartment 1) and the plasma concentration in this compartment at the time t, course of the plasma concentration of a drug with known pharma- respectively. cokinetic parameters. The classical tri-exponential equation that deﬁnes the time course of the molar concentration of a compound Table 1 in plasma displaying linear pharmacokinetic is given by: Individual pharmacokinetic parameters of arterial plasma S- and R-ketamine obtained from ponies (n = 6) anaesthetized with isoﬂurane after intravenous racemic k1t k2t k3t CðtÞ ¼ Y1 e þ Y2 e þ Y3 e ð1Þ ketamine (2.2 mg/kg) or S-ketamine (1.1 mg/kg) administration (Larenza et al., 2007).

Pony ID 1 2 3 4 5 6 where C(t) is the plasma concentration (lM) at time t,Yi (lM) represents a fraction of the initial dose that is eliminated from plasma S-ketamine V (L/kg) 0.257 0.214 0.276 0.127 0.026 0.388 with the ﬁrst-order rate constant ki (per hour). When the compound 1 of interest is best represented by a two-compartment model, the k10 (per min) 0.264 0.322 0.116 0.533 0.647 0.09 k (per min) 0.164 0.197 0.115 0.341 0.874 0.18 third term is set to zero. Assuming a mammillary arrangement of 12 k21 (per min) 0.12 0.078 0.046 0.314 0.103 0.032 the three compartments (Fig. 1) and elimination only from the cen- Racemic ketamine tral compartment, the model has been described by Maitre and Sha- S-ketamine fer (1990) using ﬁrst-order transport rate constants (per minute) for V1 (L/kg) 0.147 0.181 0.179 0.215 0.114 0.293 tissue uptake (k12, k13), recycling (k21, k31) and elimination (k10). The k10 (per min) 0.218 0.280 0.200 0.247 0.271 0.174 time course of the plasma concentration can then be expressed by k12 (per min) 0.146 0.190 0.274 0.115 0.198 0.173 k (per min) 0.051 0.078 0.097 0.082 0.034 0.057 following equations: 21 R-ketamine

DA1ðtÞ ¼ðk21 A2ðt1Þ þ k31 A3ðt1Þ ðk10 þ k12 þ k13ÞA1ðt1Þ V1 (L/kg) 0.145 0.184 0.173 0.217 0.139 0.280 k10 (per min) 0.209 0.269 0.221 0.215 0.26 0.161 þ Rðt1ÞÞDt þ BðtÞ ð2Þ k12 (per min) 0.145 0.182 0.297 0.129 0.172 0.201

k21 (per min) 0.043 0.079 0.14 0.048 0.037 0.054 DA2ðtÞ ¼ðk12 A1ðt1Þ k21 A2ðt1ÞÞDt ð3Þ V1, volume of the central compartment; kij, transfer rate constant between the ith and the jth compartment. DA3ðtÞ ¼ðk13 A1ðt1Þ k31 A3ðt1ÞÞDt ð4Þ where A1(t), A2(t) and A3(t) are the amount of drug (lg) at the time t tions of ketamine enantiomers after a single intravenous bolus. in compartments 1, 2 and 3, respectively. R(t) is the amount of drug In a previous study (Larenza et al., 2007), S-ketamine was admin- (lg/min) administered intravenously at the time t by continuous istered intravenously at 1.1 mg/kg at two occasions in five-year- infusion during the time interval, and B(t) (lg) the amount adminis- old healthy gelding Shetland ponies (n = 6), between 92 and tered as a bolus injection. If the drug pharmacokinetic is better de- 158 kg body mass (mean ± SD, 123 ± 22), anesthetized with isoflu- scribed by a two-compartment model, the transport rate constants rane. Once S-ketamine was administered alone and the second k13 and k31 have to be set to zero. Eqs. (2)–(4) calculate DAi(t), the time together with the same amount of R-ketamine in a racemic approximate change in the amount of drug (lg) in compartment i mixture. The plasma concentrations of S- and R-ketamine were over the time interval Dt (min). The amount of drug in each com- determined by capillary electrophoresis (Theurillat et al., 2005) partment at the beginning of the time interval is given by: at 0, 1, 2, 4, 8, 16, 32, 64, 128 min after bolus administration and three set of individual pharmacokinetic parameters were obtained Ai t ¼ Ai t 1 þ DAi t ð5Þ ð Þ ð Þ ð Þ (Table 1): S-ketamine alone, S-ketamine in the racemic mixture The accuracy of the approximation increases as smaller values and R-ketamine in the racemic mixture (Larenza et al., 2007). for Dt are chosen (Maitre and Shafer, 1990). Finally, the plasma These pharmacokinetic parameters were inserted in the spread- concentration (lg/mL) at the time t is given by: sheet to obtain a simulation of the individual plasma levels. The predictive accuracy of the simulations was then assessed. For each CðtÞ ¼ A1ðtÞ=ðV 1 PÞð6Þ curve, every single measured plasma concentration (Cm) corre- where V1 is the volume of the central compartment (mL/kg) and P sponded to a prediction (Cp) obtained from the algorithm. The dif- the body mass (kg) of the individual of interest. The equations pre- ference between Cp and Cm was then analyzed with the method sented above ((2)–(6)) were implemented in an excel spreadsheet described by Varvel et al. (1992). (standard Microsoft Office software for Windows), such that a pre- The prediction error (PE) was calculated for each time point as dicted plasma concentration curve was generated from a simulated follows: dose regimen. PE ¼ðCm CpÞ=Cp 100 ð7Þ 2.1.2. Evaluation of the predictive performance of the equations Four variables were calculated from the PE: (1) the median pre- This part aims at assessing if the algorithm implemented within dictive error (MDPE) is a measure of bias per individual and consti- the spreadsheet simulates appropriately the plasma concentra- tutes the median of the PE. (2) The median absolute predictive error 514 O.L. Levionnois et al. / Research in Veterinary Science 88 (2010) 512–518

(MDAPE) is a measure of the error per individual and constitutes 130, 150, 170, 190 min after starting the infusion. All samples were the median of the absolute values of the PE. (3) The divergence esti- immediately put on ice, centrifuged and the plasma was stored at mates the trend away from or toward the targeted concentration 80 °C for the determination of ketamine and norketamine enan- and constitutes the slope of the linear regression line when |PE| tiomers within 2 months (Knobloch et al., 2006). These compounds is plotted against time. Units of divergence are percentage per were measured in plasma using capillary electrophoresis (Theuril- hour. (4) The wobble is a measure of the prediction instability per lat et al., 2005). The detection limit for the enantiomers of keta- individual and is calculated as the median of the absolute value mine was 10 ng/mL. of the difference between MDPE and each PE. If the divergence For each sampling time point, the differences between actual and the wobble are low (close to zero), the prediction error defined and predicted plasma concentrations were evaluated by PE values. by MDAPE is stable over time and samplings, respectively. The parameters MDPE, MDAPE, Wobble and divergence were calculated. Therefore, three different time intervals were defined from 2.2. Target-controlled infusion of S-ketamine 0 to 30 min after the beginning of the infusion, from 30 to 120 min and from 120 to 190 min after the beginning of the infusion. The 2.2.1. Equation for TCI actual and estimated context-sensitive decrement time 50% This part aims at obtaining an algorithm to predict the infusion (CSDT50, time required for the plasma concentration to decrease rate necessary to drive a TCI of a drug with known pharmacoki- to 50% of its starting value when the IV infusion is stopped) were netic parameters. Based on former Eqs. (2)–(6), a TCI algorithm calculated. for the prediction of the appropriate infusion rate (IR(t)) starting from a user-defined target plasma level (TC(t), lg/mL) was 2.3. Statistical analysis obtained: The anesthetic data were described as parametric data by mean IR TC V = t A = P t k A k A ðtÞ ¼ðð ðtÞ 1Þ D ð 1ðtÞ ð D ÞÞð 21 2ðtÞ þ 31 3ðtÞ (±SD). The plasma concentrations, bolus dose, infusion rates, PE,

ðk10 þ k12 þ k13ÞA1ðtÞÞ=P ð8Þ MDPE, MDAPE, wobble, divergence, actual and estimated CSDT50 were described as non-parametric data by median and interquar- This equation was also implemented in an excel spreadsheet tile range (IQR). A statistical difference between S-ketamine and (standard Microsoft Office software for Windows), such that a pre- R-ketamine plasma concentrations, as well as between actual dicted infusion rate was generated from a targeted plasma and estimated CSDT50 were analyzed with a Wilcoxon Signed rank concentration.1 test for paired data defining significance at P < 0.05.

2.2.2. Anesthesia and data collection 3. Results All animals in this study received care according to the laws on care and use of experimental animals in Switzerland. Experiments 3.1. Predictive accuracy for single bolus administration were approved by the Bernese state committee for animal experimentation. The same ponies were used as in the former study (Lar- The prediction curves obtained by the algorithm were very enza et al., 2007) and the individual pharmacokinetic parameters close to the curves of the actually measured plasma concentrations for S-ketamine under isoflurane anesthesia were known (Table (Fig. 2). Very low values for MDPE, MDAPE, Wobble and Divergence 1). The ponies were fasted during 12 h with permanent access to were obtained (Table 2). water before each experiment. On the study day, ponies were physically examined and complete blood analysis was performed. Anesthesia was induced with isoflurane (Abbott AG, Baar, Switzer- 3.2. Anesthetic period during the ketamine TCI land) in oxygen by mask and, after endo-tracheal intubation, further maintained throughout the whole experiment at individual Total duration of anesthesia was 332 (±46) min. Normocapnia, minimal alveolar concentration (iMAC) for isoflurane (0.7–1.2%) normotension and normothermia were maintained throughout as determined in a previous study (Spadavecchia et al., 2006). Full anesthesia in all ponies, with arterial carbon dioxide partial pres- instrumentation allowed for continuous recording of temperature, sure of 5.85 (±0.2) kPa, mean arterial blood pressure of 86.3 heart rate, arterial blood pressure, pulse oxymetry, end-tidal con- (±7.3) mmHg and body temperature of 37.1 (±0.5) °C. All ponies recovered uneventfully from anesthesia. centration of O2,CO2, and isoflurane. All physiological values were kept within normal range. Saphenous vein catheterization was performed for administration of lactated Ringer solution at 10 mL/kg/ 3.3. Predictive accuracy for ketamine TCI h(Larenza et al., 2007) and racemic ketamine. Sixty minutes after anesthesia induction, a TCI of S-ketamine was initiated to achieve a Ketamine and norketamine enantiomers were measured in all constant plasma level of 1 lg/mL by IV administration of racemic plasma samples obtained during the racemic ketamine infusion ketamine (Graeub AG, Bern, Switzerland). Initially, racemic keta- (Fig. 3). Each pony received an intravenous bolus of racemic ketamine was administered as a bolus, followed by an infusion over mine at 360 (IQR, 310–413) lg/kg followed by an exponentially 120 min using a purposely built high-precision calibrated peristal- decreasing infusion rate ranging from 162 (IQR, 120–169) to 87 tic infusion pump. This pump was driven by a computer running a (IQR, 70–102) lg/kg/min (Fig. 3). As plasma concentrations for R- specifically designed ad hoc routine incorporating the TCI algo- ketamine were similar to S-ketamine at each time point, they were rithm (Eq. (8)). The total volume of racemic ketamine infused not represented on Fig. 3. At the beginning of the infusion (time was continuously controlled visually and weighted to confirm ade- interval from 0 to 10 min), all measured plasma concentrations quate infusion rate over the whole experiment. for S-ketamine were higher than targeted and the prediction error Arterial blood samples were taken from the left carotid artery in was high (|PE| > 40%) in four ponies (Fig. 4). During the intravenous heparinised tubes shortly before the racemic ketamine administra- infusion of racemic ketamine in the time interval from 30 to tion and at 0.5, 1, 2, 3, 6, 10, 30, 50,70, 90, 110, 120, 121, 123, 126, 120 min, the plasma concentration of S-ketamine was then maintained around 79% (IQR, 71–90) of the peak plasma concentration 1 The files can be obtained at http://www.medvet.umontreal.ca/4avet/pharmaco- measured just after the bolus administration. The measured S-ket- kinetic/pharmacokinetic.html. amine plasma concentrations during the TCI were significantly O.L. Levionnois et al. / Research in Veterinary Science 88 (2010) 512–518 515

start, as shown by |Divergence| and |Wobble| < 20% in all ponies (Table 3). After the end of racemic ketamine administration, the

predicted CSDT50p (4.6, 3.6, 8.4, 3.5, 3.95, 7.5 min) with a median of 4.28 (IQR, 3.6–7.5) were not signiﬁcantly different (P = 0.563) from the actual values (4, 6, 4.5, 3, 5, 4 min, respectively) with a median of 4.25 (IQR, 4–5).

4. Discussion

The aim of this study was to evaluate the performance of a pharmacokinetic algorithm to optimize the clinical administration of intravenous drugs. This algorithm is limited to drugs fitting adequately to a classical mammillary multi-compartmental model for linear pharmacokinetic. The predicted plasma concentrations for S- ketamine obtained by the simulation of a single intravenous bolus were very close to actual values. The algorithm allowed also to reach rapidly a plasma concentration for S-ketamine and maintained it at a pseudo-steady-state over 120 min during a TCI. High- er plasma levels than initially targeted were obtained in three ponies. During the maintenance phase of the TCI, MDPE and MDA- PE were over 60% in these three individuals. The performance of the predictions was assessed by calculation and derivation of PEs. The MDAPE is a measure of the size of the typical error. Values below 20% represent a low error and maximal values of 40% have been defined an acceptable limit for a TCI system (Swinhoe et al., 1998). This value must be adapted to the therapeutic index of the drug of interest. In analogy to inhalational anesthesia, this is equivalent to a difference of approximately 0.5 vol% between the vaporizer setting and the actual concentration of isoflurane at the effect site (central nervous system). MDPE estimates the bias and is considered acceptable if below 20%. When wobble (instability of the prediction) and divergence (trend away from or toward the targeted concentration) are low (within ±20%), the user can assume that the error (MDAPE) remains constant over time and within subjects (Swinhoe et al., 1998). This minimizes the importance of this prediction error in a clinical sit- uation as the clinical endpoints remain the main determinants for the choice of a particular dose. This principle – that an error is acceptable but should not vary within the course of an anesthetic – supports the clinical use of TCI systems (Schwilden and Schuttler, 2008). The predictions calculated in this study based on the measured plasma levels from single bolus studies of Larenza et al. (2007) are very good, as confirmed by the performance parameters. As the equations used derived from those defining standard mammillary multi-compartmental pharmacokinetic models, it is not surprising that the curves were nearly identical. This holds true as long as the drug behaviour fits well to this pharmacokinetic model (linear kinetic). The user can simulate different dosing regimens with single bolus or continuous rate infusions. The excel interface allows an easy handling of the prediction software and the user is able to visualize graphically the predicted plasma concentrations as well as of the infusion rates to be administered.

Fig. 2. Time course of ketamine enantiomers disposition (median, IQR) after the In the second part of the present work, the maintenance of a administration of a single intravenous bolus of racemic ketamine or S-ketamine in constant plasma concentration for S-ketamine at 1 lg/mL over ponies (n = 6). (A) S-ketamine plasma concentrations after administration of 120 min by TCI of racemic ketamine was targeted. Therefore, an racemic ketamine. (B) R-ketamine plasma concentrations after administration of initial loading bolus followed by a continuous infusion at exponen- racemic ketamine. (C) S-ketamine plasma concentrations after administration of S- tially decreasing rate was actually administered to six ponies based ketamine alone. Filled circles illustrate analytically determined plasma concentrations while open circles represent plasma concentrations predicted by the on their individual pharmacokinetic parameters. After the initial algorithm. loading bolus, the ketamine plasma concentrations were too high compared to the targeted levels in all ponies, as illustrated by MDPE over 20% in the time interval from 0 to 30 min after infusion higher than the target (|PE| > 40%) in 3 ponies (Fig. 4, Table 3). start. During the administration of ketamine in the time interval However, the S-ketamine plasma concentrations were maintained from 30 to 120 min, the plasma concentrations were maintained constant over the time interval from 30 to 120 min after infusion about 79% of the initial concentrations. The low values for diver- 516 O.L. Levionnois et al. / Research in Veterinary Science 88 (2010) 512–518

Table 2 Median and interquartile range (25–75%) of performance parameters for the prediction of the plasma concentrations of ketamine enantiomers in ponies(n = 6) anaesthetized with isoﬂurane after intravenous racemic ketamine (2.2 mg/kg) or S-ketamine (1.1 mg/kg) administration.

S-ketamine Racemic ketamine S-ketamine R-ketamine

MDPE (bias) 0.15 (1.21–0.80) 0.00 (0.27–0.93) 0.00 (1.58–1.88) MDAPE (inaccuracy) 8.85 (6.75–11.54) 13.31 (6.38–23.30) 10.07 (7.09–11.94) Wobble (variability) 9.04 (5.56–12.66) 15.47 (6.38–23.20) 9.13 (6.86–10.19) Divergence 5.23 (3.72–7.36) 9.48 (2.98–9.67) 6.04 (3.38–7.19)

MDPE, Median Performance Error; MDAPE, Median Absolute Performance Error.

Fig. 3. Infusion rates of S-ketamine and time course of S-ketamine and S-nor-ketamine plasma concentrations measured during and after the target-controlled infusion in ponies (n = 6) anesthetized with isoﬂurane. S-ketamine was administered intravenously from 0 to 120 min to maintain a target plasma concentration of 1 lg/mL. The continuous line represents the individual expected plasma concentrations of S-ketamine during the infusions. To ease the reading, a dotted line was added to follow the actually measured plasma concentrations of S-ketamine during the infusion. gence and wobble during this period of infusion demonstrate that lus was maintained over the whole experiment and led to high val- the variability of the plasma concentrations during TCI was small. ues for MDAPE (>40%). This illustrates the importance of regular Thus, in three individuals, the error obtained by overdosing the bo- observation of clinical endpoints. An error between the targeted and the actual concentrations can and needs to be corrected by the anesthetist when the effect is judged inadequate based on clinical observations. In the case of ketamine used as adjunct to inhalational anesthesia, higher ketamine concentrations than predicted would decrease the isoﬂurane iMAC (Muir and Sams, 1992) and the clinical signs of a more profound anesthesia depth would be detected by the anesthetist. Nevertheless, the main goal of computer-assisted TCI is to reach rapidly a pseudo-steady-state blood concentration and a stable clinical effect, together with predicting and minimizing drug accumulation in the body over long administration time. The latter appears particularly relevant because prolonged ketamine elimination may worsen the quality of recovery from general anesthesia in horses, a major determinant of perioperative morbidity in this species (Johnston et al., 2002). In the pres-

ent study, the CSDT50 were always short and both ketamine enantiomers were rapidly eliminated after cessation of the infusion. However, residual effects after plasma clearance of the drug Fig. 4. Prediction error (PE) for S-ketamine during target-controlled infusion of are possible and active metabolites like norketamine enantiomers racemic ketamine over 120 min in ponies (n = 6) anesthetized with isoﬂurane. The best prediction accuracy (PE = 0) is represented by a thick solid line, and the limit of may prolong potentially undesirable effects beyond the elimina- tolerance (±40%) is represented by the dashed lines. tion of the parent drug. O.L. Levionnois et al. / Research in Veterinary Science 88 (2010) 512–518 517

Table 3 Performance parameters of the target-controlled infusion of S-ketamine (target of 1 lg/mL over 120 min) in ponies (n = 6) anaesthetized with isoﬂurane at 0–10 min and 30– 120 min after the beginning of the intravenous infusion of racemic ketamine.

Time interval 0–10 min 30–120 min Pony ID 1234561 2 34 56 MPL (lg/mL) 1.23 1.19 1.87 2.04 1.79 2.48 0.9 1.01 1.07 2.03 1.65 1.74 MDPE (%) 23 19 87 104 79 148 11 1 6.5 103 64.5 73.5 MDAPE (%) 23 19 90 104 79 148 10.5 7 9 103 64.5 73.5 Wobble (%) 28.3 18.5 70.1 11 28.9 65.2 6 14.5 13.4 11.5 5.64 10 Divergence (%/h) NC 8.4 3.6 18 7.8 16 14

MPL, Median Plasma Level; MDPE, Median Performance error; MDAPE, Median Absolute Performance Error; NC, Not Calculated.

The main source of error for TCI systems is the inter-individual for common overdosing at the beginning of TCIs based on tradi- variation of pharmacokinetic parameters when a population model tional kinetic models (Avram and Krejcie, 2003a) and its adjust- is used (Schwilden and Schuttler, 2008). Despite that we used indi- ment has been proposed to improve their performance (Avram vidual pharmacokinetic parameters obtained in a similar experi- and Krejcie, 2003b). The use of recirculatory kinetic models is also mental setting, prediction errors were high in three individuals expected to better estimate the early behaviour of rapid-acting in the present study (Fig. 3). The main hypothesis to explain this intravenous drugs (Avram and Krejcie, 2003b). In the present discrepancy include a modification of the individual pharmacoki- study, a post hoc correction of V1 was performed based on plasma netic parameters between the initial bolus administration study concentrations actually obtained shortly after the initial bolus of when they were determined (Larenza et al., 2007), and the present the TCI. The new V1 was obtained by dividing the dose adminis- one when they were used (bolus followed by TCI). Typical covari- tered intravenously by the concentration of ketamine actually ates of pharmacokinetic parameters are for example age or gender, measured at the end of the bolus. For each pony, a prediction curve which were controlled in the present study. As the plasma concen- was then obtained from the algorithm when simulating the same trations were well maintained over 2 h despite the varying infu- dose regimen as administered during the TCI but with the new va- sion rates and the elimination was well predicted, it is likely that lue for V1. Only by adjusting the V1, the predictions were then very the intercompartmental transport parameters (k10, k12, k13, k21, close to the concentrations actually obtained (Fig. 5). It is hypoth- k31) were maintained and the algorithm was appropriate. Having esized that most of the prediction errors observed in this study used a CRI, the steady-state is generally reached after five times very likely issued from inappropriate values of V1 for TCI, as part the effective half-life (Schwartz, 2004). According to Larenza of the intrinsic inaccuracy of the pharmacokinetic model used. et al. (2007), this would have mean at least 60 min when a pseu- However, despite similar experimental settings among the differ- do-steady-state was reached in all patients within less than ent studies, an extrinsic factor of variation cannot be excluded 30 min. Retrospectively, the main error was likely initiated by (e.g. cardiac output). In conclusion, the algorithm was imple- overdosing of the loading bolus. In a standard compartmental mented in a software and allowed to study and predict plasma con- model, the initial bolus dose calculation is based on the volume centration curves of intravenous drugs fitting to a traditional of the central compartment (V1)(Schwilden et al., 1986). This multi-compartmental model. This opens the field of TIAC in veter- parameter has been identified to be a weak point of such polyexpo- inary applications. The use of the TCI equations is limited by inad- nential models, which misspecifies the early time course of drug equate pharmacokinetic parameters. However, the present study concentration assuming several simplifications remote from the supports the fact that TCI offers a better control than standard con- reality (immediate distribution in the blood compartment, inde- stant-rate infusions (Hu et al., 2005). pendent of cardiac output) (Fisher, 1996; Krejcie and Avram, 1999). Thus, the overestimated V1 has been found to be responsible Conflict of interest statement

None declared.

Acknowledgements

Racemic and S-ketamine for IV administration was kindly supplied by Dr. E. Gräub AG (Bern, Switzerland). This study was funded by Vetsuisse and by the Swiss National Science Foundation (analytical part).

References

Absalom, A.R., Struys, M.M.R.F., 2007. Target-controlled infusions. In: Overview of Target Controlled Infusions and Total Intravenous Anesthesia. Academia Press, Gent, pp. 9–14. Avram, M.J., Krejcie, T.C., 2003a. Initial distribution volume estimates affect the kinetic model and targeted infusions. Clinical Pharmacology and Therapeutics 73 (2), 85. Avram, M.J., Krejcie, T.C., 2003b. Using front-end kinetics to optimize target- controlled drug infusions. Anesthesiology 99 (5), 1078–1086. Fig. 5. Time course of the measured and the predicted S-ketamine plasma Bettschart, R.W., Larenza, M.P., 2007. Balanced anesthesia in the equine. Clinical concentrations in one pony receiving a target-controlled infusion over 120 min. Techniques in Equine Practice 6 (2), 104–110. The ﬁlled circles represent the plasma concentrations as analytically determined. Fisher, D.M., 1996. (Almost) everything you learned about pharmacokinetics was The dashed line represents the initially predicted plasma concentrations. The (somewhat) wrong! Anesthesia and Analgesia 83 (5), 901–903. dotted line represents the predicted levels with a corrected value for the volume of Hu, C., Horstman, D.J., Shafer, S.L., 2005. Variability of target-controlled infusion is the central compartment (V1 = Dose/Cm = 0.106 instead of 0.215 L/kg). less than the variability after bolus injection. Anesthesiology 102 (3), 639–645. 518 O.L. Levionnois et al. / Research in Veterinary Science 88 (2010) 512–518

Johnston, G., Eastment, J., Wood, J., Taylor, P., 2002. The confidential enquiry into Schwilden, H., Schuttler, J., 2008. Target controlled anaesthetic drug dosing. perioperative equine fatalities (CEPEF): mortality results of phases 1 and 2. Handbook of Experimental Pharmacology 182, 425–450. Veterinary Anaesthesia and Analgesia 29 (4), 159–170. Schwilden, H., Stoeckel, H., Schuttler, J., Lauven, P.M., 1986. Pharmacological models Knobloch, M., Portier, C.J., Levionnois, O.L., Theurillat, R., Thormann, W., and their use in clinical anaesthesia. European Journal of Anaesthesiology 3 (3), Spadavecchia, C., Mevissen, M., 2006. Antinociceptive effects, metabolism and 175–208. disposition of ketamine in ponies under target-controlled drug infusion. Servin, F.S., 1998. TCI compared with manually controlled infusion of propofol: a Toxicology and Applied Pharmacology 216, 373–386. multicentre study. Anaesthesia 53 (Suppl 1), 82–86. Krejcie, T.C., Avram, M.J., 1999. What determines anesthetic induction dose? It’s the Solano, A.M., Pypendop, B.H., Boscan, P.L., Ilkiw, J.E., 2006. Effect of intravenous front-end kinetics, doctor! Anesthesia and Analgesia 89 (3), 541–544. administration of ketamine on the minimum alveolar concentration of Kruger-Thiemer, E., 1968. Continuous intravenous infusion and multicompartment isoflurane in anesthetized dogs. American Journal of Veterinary Research 67 accumulation. European Journal of Pharmacology 4 (3), 317–324. (1), 21–25. Larenza, M.P., Landoni, M.F., Levionnois, O.L., Knobloch, M., Kronen, P.W., Theurillat, Spadavecchia, C., Levionnois, O., Kronen, P.W., Leandri, M., Spadavecchia, L., R., Schatzmann, U., Thormann, W., 2007. Stereoselective pharmacokinetics of Schatzmann, U., 2006. Evaluation of administration of isoflurane at ketamine and norketamine after racemic ketamine or S-ketamine approximately the minimum alveolar concentration on depression of a administration during isoflurane anaesthesia in Shetland ponies. British nociceptive withdrawal reflex evoked by transcutaneous electrical stimulation Journal of Anaesthesia 98 (2), 204–212. in ponies. American Journal of Veterinary Research 67 (5), 762–769. Maitre, P.O., Shafer, S.L., 1990. A simple pocket calculator approach to predict Swinhoe, C.F., Peacock, J.E., Glen, J.B., Reilly, C.S., 1998. Evaluation of the predictive anesthetic drug concentrations from pharmacokinetic data. Anesthesiology 73 performance of a ‘Diprifusor’ TCI system. Anaesthesia 53 (Suppl. 1), 61–67. (2), 332–336. Theurillat, R., Knobloch, M., Levionnois, O., Larenza, P., Mevissen, M., Thormann, W., Muir 3rd, W.W., Sams, R., 1992. Effects of ketamine infusion on halothane minimal 2005. Characterization of the stereoselective biotransformation of ketamine to alveolar concentration in horses. American Journal of Veterinary Research 53 norketamine via determination of their enantiomers in equine plasma by (10), 1802–1806. capillary electrophoresis. Electrophoresis 26 (20), 3942–3951. Russell, D., 1998. Intravenous anaesthesia: manual infusion schemes versus TCI Varvel, J.R., Donoho, D.L., Shafer, S.L., 1992. Measuring the predictive performance of systems. Anaesthesia 53 (Suppl. 1), 42–45. computer-controlled infusion pumps. Journal of Pharmacokinetics and Schwartz, S., 2004. Toxicokinetics. In: Evans, G. (Ed.), A Handbook of Bioanalysis and Biopharmaceutics 20 (1), 63–94. Drug Metabolism. CRC Press, Boca Raton, pp. 142–156. Toxicology and Applied Pharmacology 216 (2006) 373–386 www.elsevier.com/locate/ytaap

Antinociceptive effects, metabolism and disposition of ketamine in ponies under target-controlled drug infusion

M. Knobloch a, C.J. Portier b, O.L. Levionnois c, R. Theurillat d, W. Thormann d, ⁎ C. Spadavecchia c, M. Mevissen a,

a Division Veterinary Pharmacology and Toxicology, University of Bern, 3012 Bern, Switzerland b Environmental Toxicology Program, National Institute of Environmental Health Sciences, Research Triangle Park, NC 27709, USA c Division Anesthesiology, Department of Clinical Veterinary Medicine, University of Bern 3012, Bern, Switzerland d Department of Clinical Pharmacology, University of Bern, 3010 Bern, Switzerland Received 25 April 2006; revised 2 June 2006; accepted 7 June 2006 Available online 3 July 2006

Abstract

Ketamine is widely used as an anesthetic in a variety of drug combinations in human and veterinary medicine. Recently, it gained new interest for use in long-term pain therapy administered in sub-anesthetic doses in humans and animals. The purpose of this study was to develop a physiologically based pharmacokinetic (PBPk) model for ketamine in ponies and to investigate the effect of low-dose ketamine infusion on the amplitude and the duration of the nociceptive withdrawal reflex (NWR). A target-controlled infusion (TCI) of ketamine with a target plasma level of 1 μg/ml S-ketamine over 120 min under isoflurane anesthesia was performed in Shetland ponies. A quantitative electromyographic assessment of the NWR was done before, during and after the TCI. Plasma levels of R-/S-ketamine and R-/S-norketamine were determined by enantioselective capillary electrophoresis. These data and two additional data sets from bolus studies were used to build a PBPk model for ketamine in ponies. The peak-to-peak amplitude and the duration of the NWR decreased significantly during TCI and returned slowly toward baseline values after the end of TCI. The PBPk model provides reliable prediction of plasma and tissue levels of R- and S-ketamine and R- and S-norketamine. Furthermore, biotransformation of ketamine takes place in the liver and in the lung via first-pass metabolism. Plasma concentrations of S- norketamine were higher compared to R-norketamine during TCI at all time points. Analysis of the data suggested identical biotransformation rates from the parent compounds to the principle metabolites (R- and S-norketamine) but different downstream metabolism to further metabolites. The PBPk model can provide predictions of R- and S-ketamine and norketamine concentrations in other clinical settings (e.g. horses). © 2006 Elsevier Inc. All rights reserved.

Keywords: Physiologically based pharmacokinetic (PBPk) model; Ketamine; Norketamine; Nociceptive withdrawal reflex (NWR); Enantioselective metabolism; Target-controlled infusion (TCI)

Introduction racemic ketamine (Arendt-Nielsen et al., 1996) and four times greater than that of R-ketamine (Klepstad et al., Ketamine, a non-competitive antagonist of the N-methyl- 1990). D-aspartate (NMDA) receptor, has been used as anesthetic in Besides its affinity to the NMDA receptor, ketamine interacts humans for over 30 years. The racemic compound consists with other receptors, including non-NMDA glutamate recep- of two optical isomers: R- and S-ketamine. Major character- tors, opioid receptors, nicotinic and muscarinic acetycholine istics of the pure S-enantiomer are its fourfold greater receptors and GABA receptor type A. However, most of its affinity for NMDA receptors in comparison to R-ketamine analgesic, amnesic, psychomimetic, and neuroprotective effects (Oye et al., 1992). S-ketamine's analgesic potency has been (Chang et al., 2002), are mediated via the NMDA receptor (for a reported to be approximately two times greater than that of review, see Kohrs and Durieux, 1998). Besides ketamine's characteristic advantages of producing ⁎ Corresponding author. Fax: +41 31 53122630. indirect cardiovascular stimulation (Zielmann et al., 1997) and E-mail address: [email protected] (M. Mevissen). only mild respiratory depression (Werner et al., 1997),

0041-008X/$ - see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.taap.2006.06.011 374 M. Knobloch et al. / Toxicology and Applied Pharmacology 216 (2006) 373–386 numerous studies demonstrated the analgesic properties of dose schedules for different indications and create subsequent ketamine when administered in sub-anesthetic doses (Laulin et extrapolation of the model for horses. Given the promising al., 2002; Menigaux et al., 2001). Ketamine is used in patients results of ketamine treatment in chronic pain, experiments on suffering from hemodynamic shock, active asthmatic disease, long-term infusion versus experiments on bolus administration and anesthesia in children and intensive care settings (Kohrs are of interest. Quantitative assessment of the nociceptive and Durieux, 1998). However, phencyclidine-like psychomi- withdrawal reflex (NWR) and its facilitation as a measure of metic adverse effects have restricted its use in humans, and temporal summation has been previously described in horses abuse is common for this compound (Jansen, 2000). Recently, (Spadavecchia et al., 2002, 2004). A second goal of our study ketamine has been shown to induce neuronal death in rats was to investigate the amplitude and the duration of the NWR during early development (Scallet et al., 2004). before, during and after target-controlled infusion of ketamine. In veterinary medicine, racemic ketamine is one of the most frequently used anesthetic agents in a wide range of Materials and methods dosages in almost all species (Wright, 1982). A variety of pharmacokinetic studies on ketamine using Animals. The study was performed on 5-year-old healthy gelding Shetland ponies (n=6), weighing between 92 and 158 kg (mean±SD, 123.33±22.15). conventional two or three compartment models are available For each pony, the left carotid artery was surgically elevated to subcutaneous for humans (Hijazi et al., 2003), rats (White et al., 1976), tissue 2 years before the experiment. The in vivo experiments were approved by pigs (Loscher et al., 1990), cats (Waterman, 1983), dogs the committee for animal experimentation, Kanton Bern, Switzerland. The (Schwieger et al., 1991) and horses (Waterman et al., 1987; ponies were fasted 24 h with access to water prior to experiments. On the study Muir and Sams, 1992). These studies do not consider day, ponies were physically examined and complete blood analysis was performed. individual enantiomers, and only a few studies consider the stereoselectivity of ketamine (White et al., 1985; Delatour et Anesthesia. After preoxygenation, anesthesia was induced with isoflurane1 in al., 1991; Geisslinger et al., 1993; Henthorn et al., 1999; oxygen via a face mask that was connected to a conventional circle anesthetic Ihmsen et al., 2001; Edwards and Mather, 2001). Significant system.2 Gradual increment of inspired isoflurane concentration and physical differences in the pharmacokinetic profile, e.g. clearance, of restraint allowed stress-free induction and oro-tracheal intubation. Anesthesia was maintained with isoflurane in oxygen throughout the whole experiment at both enantiomers were seen in these experiments. PBPk individual minimal alveolar concentration that had been determined in a models for ketamine are not available in the literature. These previous study (Spadavecchia et al., 2005). Saphenous vein and carotid artery models were developed in the early 1980s (Lutz et al., catheterization were performed for Ringer lactate, ketamine administration and 1980) and offer advantages for extrapolation between blood pressure monitoring. Sixty minutes after induction of anesthesia, a target- 3 species, estimation of tissue concentrations and changing controlled infusion (TCI) of racemic ketamine was initiated. The TCI system was developed in a pilot study according to an algorithm of Maitre and Shafer dosing regimens. Currently, PBPk models are commonly (1990). To achieve a constant plasma level of 1 μg/ml of S-ketamine, a used for the analysis of absorption, distribution, metabolism computer-controlled infusion pump was used. Racemic ketamine was given as a and elimination data (ADME) for pharmaceuticals and bolus followed by an infusion over 120 min. The volume of the bolus was toxics. 10 ml, and the vehicle was sodium chloride (0.9%). Dosing for each animal was Animal studies on the pharmacodynamic effects of ketamine based on results obtained from a previous single dose study using a two- ‘ ’ compartment model. A target plasma level of 1 μg/ml (4.2 μM) of S-ket over have shown that NMDA antagonists inhibit the wind-up 120 min was anticipated by administration of an iv bolus dose (mean (SD)=376 phenomenon (Woolf and Thompson, 1991), which plays a (±123) μg/kg) followed by a linearly decreasing infusion rate of racemic major role in the development of chronic pain. The term ‘wind- ketamine (mean (SD) values at 0 and 120 min of 152 (±38) and 85 (±21) μg/kg/ up’ was defined by Mendell and Wall (1965) who observed that min, respectively). Mechanical ventilation was performed to maintain end-tidal repetition of a fixed stimulus at low frequency resulting in an CO2 in between 35 and 45 mm Hg during the entire experiment. Esophageal body temperature, ECG, pulsoxymetry, arterial blood pressure, inspired and activation of C fibers leads to a progressive build up in the end-tidal concentration of O2,CO2 and isoflurane were continuously monitored amplitude of the response, recorded as action potential with a calibrated unit.4 Mean arterial blood pressure was always maintained discharge in dorsal horn neurons. Temporal summation, defined above 70 mm Hg with iv dobutamine administration if necessary. Arterial blood by the increase of the perceived pain during repetitive gas analysis was performed every hour. After the end of each experiment, application of stimuli of constant intensity, has been proposed ponies were assisted for recovery. as a psychophysical correlate of the early phase of ‘wind-up’ in Blood sample collection. Arterial blood samples were taken from the left humans (Price et al., 1994). Studies in humans clearly carotid artery in heparinized tubes shortly before bolus ketamine application and demonstrate a reduction of temporal summation under ketamine at 0, 0.5, 1, 2, 3, 6, 10, 30, 50, 70, 90, 110, 120, 121, 123, 126, 130, 150, 170 and administration (Arendt-Nielsen et al., 1996; Guirimand et al., 190 min after starting the ketamine infusion. All samples were immediately put 2000). on ice, centrifuged and the plasma was stored at −80 °C for the determination of In veterinary medicine, clinical studies on the use of ketamine and norketamine enantiomers. ketamine for pain management are also available (Joubert, Enantioselective analysis of ketamine and norketamine in plasma. Enantio- 1998; Wagner et al., 2002). A dose-dependent reduction in mers of ketamine and its active metabolite norketamine were measured in minimal alveolar concentration (MAC) of halothane of up to 37% during low-dose ketamine infusion in horses is reported 1 Isoflo® Abbott AG, Baar, Switzerland. (Muir and Sams, 1992). 2 Electronic respirator 3100, F. Hoffmann-La Roche, Basel, Switzerland. One main goal of our study was the development of a PBPk 3 Ketasol 100©, Dr. E. Gräub, Bern, Switzerland. model for ketamine in ponies to predict tissue levels, estimate 4 S/5 compac Datex-Ohmeda, Helsinki, Finland. M. Knobloch et al. / Toxicology and Applied Pharmacology 216 (2006) 373–386 375

Fig. 1. Electropherograms of samples from one pony that were collected before TCI (A), during TCI (90 min after start of TCI) (B) and 3 min after end of TCI (C). Concentrations of S-ket, R-ket, S-nor, and R-nor in the plasma of panel B were determined to be 1.39, 1.42, 0.85, and 0.41 μg/ml, respectively and in panel C, 0.78, 0.79, 0.85, and 0.40 μg/ml, respectively. Asterisks in (B) and (C) mark peaks of unidentified ketamine metabolites. (D) depicts the temporal behaviour of the current of (C). plasma using capillary electrophoresis (Theurillat et al., 2005). Briefly, the assay same animals [data from a different study] were fit to the model. Ketamine, as a is based upon liquid–liquid extraction of ketamine and norketamine from 1 ml 50–50 racemic mixture of R-ket and S-ket, was administered via iv dose into the plasma followed by analysis of the reconstituted extract by capillary saphenous vein over varying time periods directly into the venous blood where it electrophoresis in the presence of a phosphate buffer (pH 2.5) containing is considered to become rapidly mixed as it moves through the lung into the 10 mg/ml highly sulfated β-cyclodextrin5 as a chiral selector. For each ketamine arterial blood. From the arterial blood, R-ket and S-ket distributed to the liver, enantiomer, the calibration range was between 0.04 and 2.17 μg/ml and between kidney, fat, rapidly perfused tissues (a grouping of organs such as pancreas and 0.05 and 2.5 μg/ml for the enantiomers of norketamine,6 respectively. heart) and the slowly perfused tissues (a grouping of organs such as muscle and Lamotrigine7 was used as an internal standard (IST). Analyses were performed skin). The equations used to model the physiology and the distribution of on a capillary electrophoresis analyzer8 using a 50 μm ID fused-silica capillary9 ketamine are given in Appendix A and closely follow the flow-limited models of 28 cm effective length, an applied voltage of 20 kV and a cartridge used by numerous other authors for different compounds (Kohn and Melnick, temperature of 30 °C. The detection wavelength was 195 nm. The detection limit 1996; Reddy et al., 2003). for the enantiomers of ketamine and norketamine was 0.01 μg/ml. Typical In liver and lung, R-ket and S-ket are assumed to be metabolized through electropherograms are presented in Fig. 1. Hill kinetics to R-nor and S-nor, respectively. The Hill kinetic equation used has the general form: Physiologically based pharmacokinetic (PBPk) model. Fig. 2 illustrates the six-compartment flow-limited PBPk model used to analyze the blood nct VxctCct concentration data (described above) on R-ketamine (R-ket), S-ketamine (S- ð1Þ knct þ Cnct ket), R-norketamine (R-nor) and S-norketamine (S-nor) obtained during TCI in mct ct the present study. Furthermore, plasma concentrations of R-ket, S-ket, R-nor and S-nor obtained from two additional bolus studies (2.2 mg/kg racemic where Cct is the concentration of compound c (R-ket or S-ket) in tissue t (this ketamine iv and 1.1 mg/kg S-ketamine10 iv) under isoflurane anesthesia in the subscripting is used for all values), Vxct is the maximal rate of reaction (Vmax) for metabolism, kmct is the concentration yielding half of the maximal rate and nct is the Hill coefficient governing the shape of the reaction as C increases. In 5 Sulfated-β-cyclodextrin, Sigma Aldrich Chemie, Schnelldorf, Germany. addition, both R-ket and S-ket are assumed to have a secondary first-order 6 Racemic norketamine hydrochloride, Cerilliant, Round Rock, USA. metabolic pathway in both the liver and lung (see Appendix A). The R-nor and 7 Lamotrigine, The Wellcome Foundation, London, UK. S-nor are followed explicitly through a parallel PBPk model (see Appendix A) 8 ProteomeLab™ PA 800 CE system, Beckman Coulter, Fullerton, CA, USA. with further metabolism via Hill kinetics in both the liver and lung. 9 Fused-silica capillary, Polymicro Technologies, Phoenix, AZ, USA. 10 S-ketamine, not commercially available in Switzerland, Dr. E. Gräub AG, Parameter estimation. The individual pony body weights were used in the

Bern, Switzerland. model with tissue volumes (VT) calculated as a proportion of mean body weight 376 M. Knobloch et al. / Toxicology and Applied Pharmacology 216 (2006) 373–386

Fig. 2. A schematic representation of the six-compartment physiologically based pharmacokinetic (PBPk) model describing the absorption, distribution, metabolism, and elimination of R-ket, S-ket and R-nor, S-nor in ponies. using literature values (Manohar et al., 1987; Deavers et al., 1973; Baggot, 1977; based upon the hydrophilicity of norketamine. The partition coefficients are Barone, 1984a,b, 1989, 1990, 1996; Berg, 1995; Engelhardt, 2000; Nickel et al., presented in Table A2.

1995). Blood flow rates to the tissues (Qt) were calculated as a proportion of Metabolic parameters were estimated by maximum likelihood estimation total cardiac output using literature values (McConaghy et al., 1996; Parks and where it was assumed that the data for R-ket, S-ket, R-nor and S-nor were log- Manohar, 1983). These are summarized in Table A1. Partition coefficients for R- normally distributed with constant variance over time. Separate variances were ket and S-ket were calculated based upon the octanol–water partition coefficient estimated for ketamine and norketamine enantiomers. Data below the limit of for ketamine (3.1) and an allometric relationship (Poulin and Krishnan, 1996) detection were treated as censored observations using the method of Koo et al. for calculating partition coefficients linked to lipid content in tissue (Haddad et (2002). Estimated parameters are presented in Table 1. They included the two al., 2000). The partition coefficients for R-nor and S-nor were assumed to be 1 variances, the metabolic constants (Vxct, kmct, nct and kect) and the elimination

Table 1 Estimated metabolic parameters (ketamine (subscript ket) or norketamine (subscript nor)) used in the PBPk model for enantiomers of ketamine and norketamine in ponies Parameter Description Liver Lung Kidney Units RSR S RS

Vxkett Maximal rate of metabolism 71.7556 71.7556 304.6433 304.6433 – mmol norketamine/mmol ketamine→norketamine ketamine/min kmkett Dissociation constant for ketamine 0.020965 0.020965 23.5331 23.5331 – mmol ketamine nkett Hill Coefficient for metabolism 9.8481 9.8481 5.284 5.284 – Unitless ketamine→norketamine kekett The first-order rate of elimination 0.001255 0.001255 0.0010614 0.0010614 – l/min of ketamine

Vxnort Maximal rate of metabolism 1.1025 0.24569 31.0721 69.5397 – mmol other/mmol norketamine→other norketamine/min kmnort Dissociation constant for 0.0023946 0.0020467 1.6809 6.9547 – mmol norketamine norketamine nnort Hill coefficient for metabolism 9.8377 9.9771 1.5661 2.6643 – mmol norketamine norketamine→other kenort The first-order rate of elimination –– 0.090819 of norketamine l/min M. Knobloch et al. / Toxicology and Applied Pharmacology 216 (2006) 373–386 377

rate constant for R-nor and S-nor from the kidney (kect) where c is the compound ments. Mean (SD) total anesthesia duration was 345 (±52) (ketamine or norketamine). min. All animals recovered uneventfully from anesthesia. Confidence bounds for the estimated parameters and statistical tests based on the model were all done using the likelihood ratio test as described by Koo et al. (2002). Standard deviation for the model for ketamine (R or S) and for Time courses of blood concentrations of R-/S-ketamine and norketamine (R or S) were 3.03 and 1.00, respectively. The modeling and R-/S-norketamine parameter estimation were done using MATLAB language version 6.5 (The Mathworks Inc., 2002, Natick, MA, USA). A target plasma level of 1 μg/ml (4.2 μM) of S-ket over 120 min was anticipated by administration of an iv bolus Electrophysiological recordings and quantification. The method used to μ assess NWR in horses has been described in detail by Spadavecchia et al. dose (mean (SD)=376 (±123) g/kg) followed by a linearly (2002). Briefly, NWR was evoked by a transcutaneous electrical stimulus that decreasing infusion rate of racemic ketamine (mean (SD) consisted of a train-of-five 1-ms constant-current square wave pulses values at 0 and 120 min of 152 (±38) and 85 (±21) μg/kg/ delivered at a frequency of 200 Hz. For stimulation, two self-adhesive 11 min, respectively). During TCI, plasma levels of ketamine surface electrodes were applied to the clipped skin over the palmar lateral μ Ω enantiomers were determined to be at 6.81 M±1.19 digital nerve. Resistance of stimulation electrodes had to be lower than 2 k . – μ The ground electrode was placed on the back of each pony. For (mean±SD) (range 5.47 9.35 M). Values obtained for electromyographic (EMG) recordings, a pair of self-adhesive electrodes12 arterial blood concentrations of S- and R-ket were similar at was placed over the deltoid muscle. Stimulations and recordings were all time points (Fig. 3A). performed through a specially designed computerized system. EMG signals R-nor and S-nor were detected 60 s after ketamine were amplified with an overall gain of 5000 and bandpass filtered (7 to administration. Plasma concentrations for norketamine enantio- 200 Hz; first-order active filters with 6 dB/octave slope). In order to record and quantify the EMG reflex activity in response to stimulation, the EMG mers were determined to be lower compared to those of the was recorded from 100 ms prior to until 400 ms after the stimulus, resulting parent compound in all but one of the animals and during the in a total recording time of 500 ms with 512 sampling points at a sampling entire sampling period (see Fig. 4A). frequency of 1024 Hz. Electrical stimulations were performed before TCI under isoflurane anesthesia (Time Point 1=TP 1), 25 min (TP 2) and 105 min (TP 3) after start of ketamine TCI and 5 min (TP 4) and 65 min (TP 5) after end of ketamine TCI under isoflurane alone. Each stimulation consisted of 60 consecutive stimuli at the fixed intensity of 44 mA and a frequency of 1 Hz. The EMG responses obtained were averaged, and their morphology was analyzed to quantify the peak-to-peak amplitude and the reflex duration. To be considered a reflex response, the EMG burst following stimulation had to be at least 3 times the amplitude of the background activity with a duration of at least 10 ms within the epoch occurring 20 to 70 ms after the stimulation onset. To quantify the muscular response, the root-mean-square (RMS) value for reflex amplitude was calculated for the 20–70 ms epoch after stimulation. The background EMG amplitude was calculated as the RMS amplitude during the 100 ms interval before stimulation. To minimize the influence of possible variability among animals, the relative amplitude of the reflex in the range from 20 to 70 ms was calculated as the ratio between the RMS amplitude detected during such epoch and the RMS of the background EMG in the prestimulation interval.

Statistical analysis. Peak-to-peak amplitude and reflex duration of the NWR were analyzed by Friedman repeated measures ANOVA on ranks followed by the post hoc Tukey test for multiple comparisons. A value of p<0.05 was considered significant. The statistical analysis was performed using SigmaStat software (Systat Software inc., version 3.10, Point Richmond, CA, USA). Statistical evaluations within the context of the PBPk model were performed using a likelihood ratio test (Bard, 1974—Bard, J. Nonlinear Parameter Estimation, Academic Press, Inc., New York City).

Results

Physiological parameters

The infusion regime was well tolerated and no adverse effects were observed. All continuously monitored data stayed within physiological ranges throughout the experi- Fig. 3. Plasma concentrations of R-ket and S-ket (not discernably different) in ponies (n=6) during and after TCI (presented as mean±SD) (A). Proportions of R-ket (open circles), S-ket (filled circles), R-nor (open triangle), and S-nor 11 Neuroline 7 00 02-J, Medicotest, Olstykke, Denmark. (filled triangle) in percent in plasma of ponies (n=6) during and after TCI (data 12 Synapse, Ambu A/S, Ballerup, Denmark. are presented as mean) (B). 378 M. Knobloch et al. / Toxicology and Applied Pharmacology 216 (2006) 373–386

Interestingly, the plasma levels of S-nor exceeded those of provides an excellent fit for plasma concentrations of R- and S- R-nor in all samples later than 3 min after start of the TCI nor with respect to the remarkable differences seen between (Fig. 3B). measured plasma levels. Additional peaks that were detected in plasma samples (see IV bolus administration of racemic ketamine (2.2 mg/kg) peaks marked with an asterisk in Fig. 1) were found to increase and S-ket (1.1 mg/kg) are shown in Figs. 4BandC, over time. Comparison of the UV spectra of those peaks respectively. In general, a rapid decline was seen after iv revealed spectra similar to those of ketamine and norketamine. bolus injection. This rapid decline is driven by the metabolism. Therefore, they represent enantiomers of metabolites of Because of the high values for the Hill coefficients, the decline ketamine and norketamine. Due to lack of standards, however, becomes much less rapid as tissue concentrations go below these compounds could not be identified. approximately 23 mM (see Fig. 5) due to the highly nonlinear metabolism in the lung. In the liver, due to the low km value PBPk model and the large Hill coefficient, the metabolism from racemic ketamine to norketamine is effectively zero order. Both The model was developed using the physiological para- measured and predicted data show no difference in S-ket meters given in Table A1, and the partition coefficients given in levels for administration of the racemic mixture (Fig. 4B) and Table A2. The value for cardiac output obtained from the S-ket alone (Fig. 4C). literature (Manohar et al., 1987) was reduced by 60% under isoflurane anesthesia at 1 MAC. Biotransformation The fits of R- and S-ket and R- and S-nor during and after TCI using the PBPk model are given for one pony in Fig. 4A. The liver has been reported to be the main organ of Note that all ponies were used in the estimation of model biotransformation for ketamine (White et al., 1982). In order parameters, this figure is representative of the quality of the fit to improve the fit of the model to the data, metabolism in for all of the ponies. The time course for the metabolites R- and the lung was included in the model. This resulted in a S-nor during and after TCI is characterized by an initial rise significant improvement to the fit of the data (p<0.001). immediately after the bolus application (shortly before initiation The data for one pony are shown in Fig. 4A. Because of the of TCI) and a rapid decline at the end of TCI. The model initial passage of ketamine from the venous compartment

Fig. 4. Predicted (dashed line) plasma concentrations and 95% confidence intervals (dotted lines) for R-ket, S-ket, R-nor, and S-nor in one pony following ketamine TCI (A), intravenous bolus administration of 2.2 mg/kg racemic ketamine (B) and intravenous bolus administration of 1.1 mg/kg S-ketamine (C). Open circles represent measured plasma concentrations. M. Knobloch et al. / Toxicology and Applied Pharmacology 216 (2006) 373–386 379

rejected due to a statistically significant difference (p<0.01), but not the second hypothesis (p>0.05). Figs. 5 and 6 show the final model predictions for distribution and metabolism of ketamine and norketamine enantiomers, respectively, during and after TCI for a typical pony. Predicted tissue concentrations of ketamine and norketamine enantiomers in slowly and rapidly perfused tissue, fat, kidney, liver and lung are presented. Similar to the measured concentrations in plasma, R- and S-ket show equal tissue levels according to model prediction at all time points and in all tissues. Ketamine distribution in kidney tissue is characterized by an initial rise during the first third of TCI. A plasma concentration of 72.5 μM is predicted for the end of TCI (120 min, Fig. 5A). Thereafter, an exponential decline is observed with the tissue concentration being 4.8 μM at 4.25 days after starting TCI (Fig. 5B). The highest tissue concentrations of ketamine enantiomers are predicted to be in the kidneys. Following TCI, ketamine enantiomers are predicted to accumulate slightly in rapidly perfused tissue and strongly in fat, due to redistribution from other tissues. The highest tissue concentrations in rapidly perfused tissue are predicted to be reached around 10 h after initiation of TCI followed by a slow reduction. Levels of ketamine enantiomers in fat still continue to increase 4 days after starting TCI according to the model, showing concentrations of 32.4 μM at this time point (Fig. 5B). Peak concentrations of R- and S-ket in fat are reached around 7 days after TCI followed by a slow decrease in levels of ketamine enantiomers. The model predicts concentrations of about 20 μM for R- and S-ket in fat 42 days after the start of TCI. Ketamine levels in slowly perfused tissue and lung show a rapid decline immediately after TCI with slowly perfused Fig. 5. Predicted tissue concentrations of R-ket and S-ket for the pony whose tissue showing the most gradual reduction (Fig. 5B). Predicted plasma data are given in Fig. 4 during TCI (A) and after TCI (B). Equal values values in blood match those that have been determined in for tissue concentration of R-ket and S-ket are predicted for all time points. plasma using capillary electrophoresis (prediction for blood not shown). In lung, the time course of R- and S-ketamine is predicted through the lung before entering other parts of the body, to be similar to R- and S-ketamine in plasma (Fig. 4A), there was a marked first-pass metabolic effect to reduce the whereas the concentration of ketamine enantiomers in the lung levels of ketamine in plasma. is predicted to be about 5 times greater than the plasma levels Two hypothetical metabolic schemes were compared using during and after TCI. The liver (data not shown) is predicted the PBPk model to explain the difference of about 50% to contain the lowest concentrations (consistently below between the concentrations of R- and S-nor as illustrated in 0.02 μM) of ketamine enantiomers at virtually all times. In Fig. 4A. The first hypothesis allows different biotransforma- all tissues, concentrations of the metabolites are predicted to tion rates from R- and S-ket to R- and S-nor but no differences remain at much lower levels than those of the parent concerning biotransformation of norketamine enantiomers to compounds. further metabolites. In mathematical terms, this would assume Similar to their time course in plasma, norketamine equal Hill kinetics for R- and S-nor (VxRnt =VxSnt, kmRnt =kmSnt, enantiomers are characterized by an initial rise at the and nRnt =nSnt). The second hypothesis assumes equal rates of beginning of TCI and a rapid decline after TCI in all tissues biotransformation for ketamine to norketamine followed by (Fig. 6). further transformation of norketamine to other metabolites that Norketamine enantiomers are much more hydrophilic than could differ between R- and S-nor. In mathematical terms, this their lipophilic parent compounds. Therefore, they are predicted would assume equal Hill kinetics for R- and S-ket (VxRkt = to be rapidly eliminated from fat in contrast to R- and S-ket. In VxSkt, kmRkt =kmSkt, and nRkt =nSkt) into R- and S-nor. Compar- general, norketamine enantiomers are eliminated rapidly from ing likelihood values of these two different simulations with all tissues in comparison to S-andR-ket. the likelihood value of an overall and unrestricted model using Similar to the determined plasma levels, the model all six ponies simultaneously, the first hypothesis could be consistently predicts higher amounts for S-nor than for R-nor 380 M. Knobloch et al. / Toxicology and Applied Pharmacology 216 (2006) 373–386

Fig. 6. Predicted tissue concentrations (same pony as for Fig. 5)ofS-nor during TCI (A) and after TCI (B); and of R-nor during TCI (C) and after TCI (D). in all tissues and at all time points and the ratio between S- and rences in the relative amplitude of the reflex in the epoch of 20 R-nor concentrations differs between tissues. In liver, this to 70 ms after stimulation were detected between pre-drug difference is most prominent. It reflects the different biotrans- exposure (TP 1) and ketamine exposure (TP 3) (p=0.004) formation rates for S- and R-nor to further metabolites. The and between post-ketamine exposure (TP 5) and ketamine model predicts the rate of biotransformation to downstream exposure (TP 3) (p=0.004). The averaged NWR slowly metabolites of R-nor in the liver to be 8 times higher than the returned toward baseline values after the end of TCI as biotransformation rate for S-nor. Therefore, in liver tissue, the illustrated in Fig. 7 although it remained altered at 65 min post concentration for S-nor rises (Fig. 6A) during TCI, whereas R- TCI (TP 5). nor is rapidly metabolized and does not accumulate (levels of about 0.2 μM only, Fig. 6). This predicted difference in Discussion metabolism of R- and S-nor in the liver may explain the difference between R- and S-nor concentrations in plasma and Pharmacokinetics other tissues. In the current study, a PBPk model for ketamine was built Nociceptive withdrawal reflex (NWR) from plasma concentrations obtained in ponies after continuous infusion or bolus administration. The model considers a first- A statistically significant difference in peak-to-peak ampli- pass metabolism in lung to provide a precise prediction of the tude of the NWR was found between pre-ketamine exposure rapid rise and decline of the curve of norketamine enantiomers. (TP 1) and 25 min after starting ketamine TCI (TP 2) (p=0.001) Metabolism of ketamine in lung does occur as reported for in and 105 min after starting the ketamine TCI (TP 3) (p=0.001). vitro ketamine biotransformation in microsomes isolated from Reflex duration was significantly decreased 105 min after the lung tissue of rabbits (Pedraz et al., 1986). beginning of ketamine TCI (TP 3) (p=0.014) in comparison to An interesting finding was the continuously obtained higher pre-ketamine exposure (TP 1). Statistically significant diffe- concentrations of S-nor compared to R-nor in all plasma M. Knobloch et al. / Toxicology and Applied Pharmacology 216 (2006) 373–386 381

In vivo studies in humans have shown higher values for the clearance of S-ket than for R-ket either after administration of the individual enantiomers (Persson et al., 2002; White et al., 1985) or after administration of the racemic mixture (Henthorn et al., 1999; Geisslinger et al., 1993; Ihmsen et al., 2001). Our results also clearly show differences in biotransformation of ketamine and/or norketamine enantiomers resulting in consistently higher plasma concentrations of S-nor than of R-nor. Ihmsen et al. (2001) reported that the clearance of S-ket, when administered as pure enantiomer, was significantly larger than the clearance of S-ket when administered in the racemate. They suggested an inhibition of the metabolism of S-ket by R-ket referring to an in vitro study of Kharasch and Labroo (1992). Using human liver microsomes, the rate of N-demethylation of the racemate was found to be significantly smaller than the sum of the rates of the individual enantiomers. Therefore, a metabolic enantiomeric interaction was suggested by Kharasch and Labroo (1992), whereby S-ket inhibits the metabolism of R-ket and R-ket inhibits the metabolism of S-ket termed as relative enantiomeric selectivity. In the current study, two hypotheses were tested to explain the consistently higher concentrations of S-nor in comparison to Fig. 7. Electromyographic recording depicted for one animal. The averaged R-nor detected in plasma. The first hypothesis assumed different response of the deltoid muscle to electrical stimulation (60 consecutive stimuli, biotransformation rates from ketamine enantiomers to R- and S- 44 mA, 1 Hz) of the palmar lateral digital nerve is presented. The time points nor referring to Delatour et al. (1991) followed by identical rates (TP) indicate predrug (without ketamine) under isoflurane anesthesia (TP 1), of biotransformation for norketamine enantiomers to further 25 min (TP 2) and 105 min (TP 3) after the start of ketamine TCI, 5 min (TP 4) metabolites. The second hypothesis included different rates of and 65 min after the end of ketamine TCI (TP 5) under isoflurane anesthesia. biotransformation for R- and S-nor to further metabolites, taking a study of Trevor et al. (1983) in account. They found samples after equal amounts of ketamine enantiomers were human liver microsomes to exhibit selectivity with respect to administered to the ponies. This is in agreement with results the formation of hydroxylated norketamine metabolites, with R- obtained in an earlier study (Delatour et al., 1991). After iv and S-nor undergoing preferential hydroxylation at different bolus administration of 6 mg/kg of racemic ketamine, positions at the cyclohexanone ring, termed product selectivity. proportions of S- and R-nor of about 3:1 after 5 min and of However, identical biotransformation rates for R- and S-ket are about 9:1 after 40 min were reported, whereas no difference in assumed in the second hypothesis. Therefore, it did not consider plasma concentrations of S- and R-ket was obtained at all time differences in enzymatic reactions for R-andS-ket to points. The authors hypothesized a substrate enantioselectivity norketamine enantiomers, termed substrate selectivity. Khar- of the cytochrome-P450-dependent N-demethylation of keta- asch and Labroo (1992) concluded that human ketamine mine (Delatour et al., 1991). This is contrary to our hypothesis metabolism exhibits moderate degrees of both substrate and that there are identical rates of biotransformation to norketamine product stereoselectivity. After simulation of both hypotheses enantiomers but differences in downstream metabolism to using the PBPk model, we could reject the first hypothesis due further hydroxylated and glucuronidated compounds. to a statistically significant difference (p<0.01) compared to the Ketamine is metabolized extensively by the hepatic likelihood value of an overall and unrestricted model, but not cytochrome P450 system to norketamine, an active metabolite the second hypothesis (p>0.05). Further investigations in vitro with a potency one-third to one-fifth when compared to the using equine microsomes are required and are already under parent compound (Kohrs and Durieux, 1998). Norketamine is way. hydroxylated to form hydroxy–norketamine compounds that Edwards and Mather (2001) investigated the effect of two can be conjugated to more water-soluble glucuronide deriva- different infusion regimes of racemic ketamine in rats by tives (White et al., 1982) and further metabolized to performing either a ‘washin infusion’ of 6 mg/kg until lethality dehydronorketamine which appears to be the most abundant or a ‘washout infusion‘ of 20 mg/kg over 5 min. Plasma and ketamine metabolite other than norketamine (Williams and tissue concentrations of ketamine and norketamine were Wainer, 2002). Contrary ring hydroxylation of ketamine determined enantioselectively. During ‘washin infusion’, the without prior N-demethylation occurs, but this pathway appears authors reported consistently higher plasma levels of R- than S- to be quantitatively of minor importance (White et al., 1982). In nor. During ‘washout infusion’, a period with higher concentra- humans, CYP3A4, CYP2B6 and CYP2C9 have been reported tions of R- than S-nor was followed by a phase during which the to be involved in ketamine biotransformation (Hijazi and S-nor levels significantly exceed those of R-nor. The authors Boulieu, 2002; Yanagihara et al., 2001). proposed two possible reasons: concentration-dependent N- 382 M. Knobloch et al. / Toxicology and Applied Pharmacology 216 (2006) 373–386 demethylation and concentration-independent metabolic inver- animal. Especially for this application, the therapeutic window sion from R-toS-nor or from R-toS-ket and subsequent is very narrow. Exact prediction of plasma and tissue levels can metabolism to S-nor. We consider the second suggestion to be be estimated using a PBPk model to achieve therapeutic effects rather unlikely because no R-enantiomers were detected after without adverse effects. administration of S-ket (Geisslinger et al., 1993; Ihmsen et al., 2001). Pharmacodynamics

Tissue distribution of ketamine and norketamine enantiomers In the chosen route of administration, ketamine resulted in a decreased peak-to-peak amplitude and a shorter duration of the The PBPk model predicts identical tissue concentration NWR. The NWR is a polysynaptic, spinal reflex that is curves in all compartments for S-andR-ket. However, responsible for the escape reaction from a damaging stimulus to consistently higher concentrations for S- than for R-nor in all protect the integrity of the body (nocifensive reaction). Due to tissues are predicted due to differences in metabolism leading to advantages like non-invasiveness, relative reproducibility and a differences in plasma. good correlation between muscle response and subjective pain Our prediction for tissue distribution of ketamine is in intensity in humans, the quantitative assessment of the NWR agreement with those of Henthorn et al. (1999). They used a has become an important tool in pain research (Arendt-Nielsen recirculatory pharmacokinetic model to analyze arterial blood and Petersen-Felix, 1995). Our results are consistent with concentrations that they obtained after administration of studies in rats and cats that clearly have shown that ketamine racemic ketamine intravenously in dogs. Because of the lack selectively reduces spinal reflexes to noxious stimuli (Hartell of stereoisomeric effects for ketamine distribution in pulmonary and Headley, 1990; Headley et al., 1987). They report the tissue, identical distribution of S- and R-ket in other tissues was inhibition of spinal responses to be dose-dependent. Depression suggested. This is contrary to Edwards and Mather (2001), who of spinal reflexes consistently was found to correlate with the detected significant differences between concentrations of S- degree of nociception caused by stimulus intensity and amount and R-ket and S- and R-nor in different tissues after either of tissue damage. ‘washin’ or ‘washout’ infusion in rats. The predictions of our In a previous study, an inhibition of the NWR already was PBPk model analysis do not confirm these results. Due to confirmed in ponies under isoflurane anesthesia (Spadavecchia identical partition coefficients for optical isomers, our model et al., 2005). In comparison to these results, our results showed a does not predict differences in tissue uptake. more pronounced inhibition of the NWR under ketamine In the present study, kidneys were estimated to provide infusion during isoflurane anesthesia. Therefore, isoflurane highest tissue levels of ketamine enantiomers, which is in might have contributed to the observed inhibition of the NWR. agreement with results from studies in rats (White et al., 1976; There is also evidence from human studies showing that Edwards and Mather, 2001). ketamine provokes a depression of the NWR evoked by The PBPk model predicted the biotransformation rate for the mechanical or strong electrical stimuli (Arendt-Nielsen et al., formation of norketamine enantiomers to be 2.8 times higher in 1995). Moreover, temporal summation threshold was signifi- the liver compared to the lung during steady state. This cantly increased when repeated electrical stimuli (five pulses at assumption is supported by results from a study in horses 2 Hz) were applied during administration of racemic or S-ket (Larsson et al., 2003). The authors report that the capacity of the (Arendt-Nielsen et al., 1996). Guirimand et al. (2000) observed lung to activate aflatoxin B1 (AFB1) by cytochrome P450 this phenomenon when repeated stimuli were applied at the enzymes to the reactive intermediate AFB1-8,9-epoxide was frequency of 1 Hz in humans. The same stimulation frequency clearly shown to be lower than those of the liver using was used in the current study. While averaging the responses to microsomal preparations from 5 different horses. 60 stimuli, temporal summation effects contribute to the Due to the high lipid solubility of ketamine that is reported to amplitude of the recorded NWR. The reduction of the peak- be 5 to 10 times higher than thiopental (Cohen and Trevor, to-peak and the duration observed during ketamine TCI 1974), our PBPk model predicted accumulation of ketamine in possibly indicates inhibition of the single nociceptive reflex, fat. Especially in horses, adverse effects of ketamine like of temporal summation or both. Further investigation will be excitement, delirium and severe muscle rigidity complicate required to more specifically discriminate the mechanisms of recovery phase and should be prevented by reducing ketamine action in similar experimental settings. administration early enough before recovery. The PBPk model Our results show that continuous iv administration of considers accumulation of ketamine in fat and other tissues and racemic ketamine over 120 min with S-ketamine plasma level is, therefore, an important tool to create dosing regimes for of around 7 μM is well-tolerated and resulted in no adverse ponies as well as other equines or different animals. Extrapola- effects in ponies. The developed PBPk model allows estimation tion to other species can be performed by changing the of levels of ketamine and norketamine enantiomers in plasma physiological parameters in the model if tremendous differences and other tissues using different administration schedules. The in biotransformation can be precluded. The PBPk model allows model confirmed the main properties of ketamine: initial rapid dose estimation schedules for ketamine administration at even distribution followed by accumulation in fat. Furthermore, it lower dosages than those used in this study. Lower dosage estimated different levels of R- and S-nor resulting from provides administration to the standing non-anesthetized different metabolism to further metabolites. M. Knobloch et al. / Toxicology and Applied Pharmacology 216 (2006) 373–386 383

Using physiological parameters for horses or other species, The equation describing the change in the amount of R-nor the model can be used to predict dosing regimes for iv long-term and S-nor in kidney capillary space is identical to Eq. (A2) with ketamine administration. The depression of the NWR during the exception that they are assumed to be eliminated by first- ketamine infusion demonstrates the depressive spinal effect of order kinetics from kidney tissue (see Fig. 1) so the equation this NMDA antagonist. However, more data are needed to governing kidney tissue concentration is given by: confirm the preliminary evidence we obtained in our study on a dAcK ðtÞ AcK ðtÞ AcK ðtÞ AcK ðtÞ possible antinociceptive effect of ketamine at low doses. ¼ QK kecK : ðA3Þ dt VK PcK VK VK Acknowledgments where kecK is the first-order rate of elimination of compound from the kidney (μmol/min) and keRkK =keSkK =0. The authors gratefully acknowledge the support provided by Finally, to maintain mass balance, the amount of R-ket, S-ket, Prof. U. Schatzmann, Prof. M.F. Landoni and Dr. M.P. Larenza. R-nor, and S-nor entering and leaving the venous blood must Racemic ketamine and S-ketamine for iv administration was balance with the amount entering and leaving the combined kindly supplied by Dr. E. Gräub AG (Bern, Switzerland). This tissues (excluding lung). Hence, the equations describing the study was funded by Vetsuisse, Switzerland and by the Swiss kinetics of compound in the venous blood are given by: National Science Foundation (analytical part). This research was supported [in part] by the Intramural Research Program of ð Þ ð Þ ð Þ ð Þ ð Þ dAV t ¼ QF AF t þ QSAS t þ QRAR t þ QK AK t the NIH and NIEHS. dt VF PF VSPS VRPR VK PK Conflict of Interest Statement: the authors have no conflict of ð Þ ð Þ þ QLAL t QV AV t ð Þ interest of any kind related to the work presented in this A4 VLPL VV publication. and for arterial blood is given by: Appendix A dA ðtÞ Q A ðtÞ Q A ðtÞ A ¼ A L A A : ðA5Þ From the arterial blood, R-ket, S-ket, R-nor and S-nor dt VL VA equilibrate with tissue via tissue partition coefficients. For four tissues, kidney (t=K), fat (t=F), rapidly perfused tissues (t=R) and slowly perfused tissues (t=S), the rate of transfer in and out Table A1 of the tissue is modeled using flow-limited kinetics leading to Physiological parameters used in the PBPk model for R-ket, S-ket, R-nor and the equations: S-nor in ponies Parameter Value Description Reference dActðtÞ AcAðtÞ ActðtÞ name ¼ Qt ðA1Þ dt VA VtPct co 5% of bw Cardiac output rate (Manohar et al., μ (l/min) at which blood 1987) where Act(t) is the amount ( mol) of compound (c) in tissue (t) flows through the heart space at time t, Qt is the rate (l/min) at which blood enters the bw Individual Body weight in kg tissue t, AcA is the amount of compound in arterial blood animal vF 18% % of body mass that is (Deavers et al., (μmol), VA is the volume of the arterial blood (l), Act(t) is the amount (μmol) of compound in tissue t, V is the volume (l) of fat (unitless but results 1973) t in a volume in liters tissue t and Pt is the partition coefficient (unitless) for when multiplied by compound between blood and tissue for tissue t. body weight) The movement of compound in and out of the liver vS 100%— % of body mass that (Baggot, 1977, (t=L) and lung (t=P) tissue is modeled by the equation: sum is slowly perfused tissue Barone 1984a,b, (all others) (unitless but results in a 1989; Berg, 1995) ð Þ ð Þ ð Þ nct ð Þ= nct volume in liters when dAct t AcA t Act t VxctAct t Vt ¼ Q multiplied by body weight) t nct þ nct ð Þ= nct dt VA PctVt kmct Act t Vt ð Þ vR 28% % of body mass that is (Barone, 1996; Act t rapidly perfused tissue Berg, 1995) kect ðA2Þ Vt (unitless but results in a volume in liters when where Vxct is the maximal rate of reaction (Vmax) for multiplied by body weight) metabolism, kmct is the concentration yielding half of the vk 0.35% % of body mass that is (Barone, 1990; kidney (unitless but results Berg, 1995) maximal rate, nct is the Hill coefficient governing the shape in a volume in liters when of the reaction as C increases, and kect is the first order rate μ multiplied by body weight) of elimination for all other metabolites of compound ( mol vL 2% % of body mass that is liver (Barone, 1984a) metabolites/μmol c/min). For lung, the rate of blood flow (unitless but results in a into the tissue is the entire cardiac output (QP =QV =QA). volume in liters when For R-nor and S-nor, no first order rate is assumed so multiplied by body weight) keRnL =keSnL =keRnP =keSnP =0. (continued on next page) 384 M. Knobloch et al. / Toxicology and Applied Pharmacology 216 (2006) 373–386

Table A1 (continued) Table A2 (continued) Parameter Value Description Reference Parameter Value Partition coefficient (unitless) name name vX 1% % of body mass that is lung (Nickel et al., 1995) pmK 1 For norketamine (R or S) between blood (unitless but results in a and kidney volume in liters when pmL 1 For norketamine (R or S) between blood multiplied by body weight) and liver vA 5.4% % of body mass that is (Engelhardt, pmX 1 For norketamine (R or S) between blood arterial blood (unitless 2000) and lung but results in a volume in liters when multiplied by body weight) References qF 6.50% % (unitless) of cardiac (McConaghy output (co) which, et al., 1996) Arendt-Nielsen, L., Petersen-Felix, S., 1995a. Wind-up and neuroplasticity: is when multiplied by co there a correlation to clinical pain? Eur. J. Anaesthesiol. Suppl. 10, 1–7. yields rate (l/min) at Arendt-Nielsen, L., Petersen-Felix, S., Fischer, M., Bak, P., Bjerring, P., which blood flows through Zbinden, A.M., 1995b. The effect of N-methyl-D-aspartate antagonist the fat (ketamine) on single and repeated nociceptive stimuli: a placebo- qS 31.50% % (unitless) of co which, (McConaghy controlled experimental human study. Anesth. Analg. 81, 63–68. when multiplied by co et al., 1996) Arendt-Nielsen, L., Nielsen, J., Petersen-Felix, S., Schnider, T.W., Zbinden, A. yields rate (l/min) at M., 1996. Effect of racemic mixture and the (S+)-isomer of ketamine on which blood flows temporal and spatial summation of pain. Br. J. Anaesth. 77, 625–631. through the slowly Baggot, J.D., 1977. Drug distribution. In: Baggot, J.D. (Ed.), Principles of Drug perfused tissue Disposition in Domestic Animals: The Basis of Veterinary Clinical qR 100%-sum % (unitless) of co which, (McConaghy Pharmacology. W.B. Saunders, Philadelphia, pp. 48–72. (rest) when multiplied by co et al., 1996) Barone, R., 1984a. Appareil digestif, glandes annexes de l'intestin, foie. In: yields rate (l/min) at Barone, R. (Ed.), Anatomie comparée des mammifères domestiques; tome 3: which blood flows Splanchnologie I. Editions Vigot, Paris, pp. 507–561. through the rapidly Barone, R., 1984b. Appareil respiratoire, poumons. In: Barone, R. (Ed.), perfused tissue Anatomie comparée des mammifères domestiques; tome 3: Splanchnologie qK 13.10% % (unitless) of co which, (Parks and I. Editions Vigot, Paris, pp. 756–800. when multiplied by co Manohar, 1983) Barone, R., 1989. Myologie, caractères généraux des muscles. In: Barone, R. yields rate (l/min) at which (Ed.), Anatomie comparée des mammifères domestiques; tome 2: blood flows through the kidney Arthrologie et Myologie. Editions Vigot, Paris, pp. 357–375. qL 2.10% % (unitless) of co which, when (McConaghy Barone, R., 1990. Apparail uro-génital, reins. In: Barone, R. (Ed.), Anatomie multiplied by co yields rate et al., 1996) comparée des mammifères domestique; tome 4: Splanchnologie II. Editions (l/min) at which blood Vigot, Paris, pp. 19–45. flows through the liver Barone, R., 1996. Cœur, caractères généraux et conformation extérieure. In: qX 100% % (unitless) of co which, Barone, R. (Ed.), Anatomie comparée des mammifères domestique: tome when multiplied by co 5: Angiologie. Editions Vigot, Paris, pp. 14–25. yields rate (l/min) at which Berg, R., 1995. Angewandte Anatomie der Schlachttierkörper. In: Berger, R. blood flows through the lung (Ed.), Angewandte und topographische Anatomie der Haustiere. G. Fischer Verlag, Stuttgart, pp. 475–495. Chang, M.L., Yang, J., Kem, S., Klaidman, L., Sugawara, T., Chan, P.H., Adams Table A2 Jr., J.D., 2002. Nicotinamide and ketamine reduce infarct volume and DNA Partition coefficients for R- and S-ket and R- and S-nor fragmentation in rats after brain ischemia and reperfusion. Neurosci. Lett. Parameter Value Partition coefficient (unitless) 322, 137–140. name Cohen, M.L., Trevor, A.J., 1974. On the cerebral accumulation of ketamine and the relationship between metabolism of the drug and its pharmacological ppF 332 For ketamine (R or S) between effects. J. Pharmacol. Exp. Ther. 189, 351–358. blood and fat Deavers, S., Rosborough, J.P., Garner, H.E., Huggins, R.A., Amend, J.F., 1973. ppS 4.83 For ketamine (R or S) between Blood volumes and total body water in the domestic pony. J. Appl. Physiol. blood and slowly perfused tissue 34, 341–343. ppR 35.3 For ketamine (R or S) between blood Delatour, P., Jaussaud, P., Courtot, D., Fau, D., 1991. Enantioselective N- and rapidly perfused tissue demethylation of ketamine in the horse. J. Vet. Pharmacol. Ther. 14, ppK 31.4 For ketamine (R or S) between blood 209–212. and kidney Edwards, S.R., Mather, L.E., 2001. Tissue uptake of ketamine and norketamine ppL 16.7 For ketamine (R or S) between enantiomers in the rat: indirect evidence for extrahepatic metabolic blood and liver inversion. Life Sci. 69, 2051–2066. ppX 10 For ketamine (R or S) between Engelhardt, W., 2000. Kreislauf. In: Engelhardt, W., Breves, G. (Eds.), blood and lung Physiologie der Haustiere. Enke Verlag, Stuttgart, pp. 171–189. pmF 1 For norketamine (R or S) between Geisslinger, G., Hering, W., Thomann, P., Knoll, R., Kamp, H.D., Brune, K., blood and fat 1993. Pharmacokinetics and pharmacodynamics of ketamine enantiomers in pmS 1 For norketamine (R or S) between surgical patients using a stereoselective analytical method. Br. J. Anaesth. blood and slowly perfused tissue 70, 666–671. pmR 1 For norketamine (R or S) between Guirimand, F., Dupont, X., Brasseur, L., Chauvin, M., Bouhassira, D., blood and rapidly perfused tissue 2000. The effects of ketamine on the temporal summation (wind-up) of M. Knobloch et al. / Toxicology and Applied Pharmacology 216 (2006) 373–386 385

the R(III) nociceptive flexion reflex and pain in humans. Anesth. Analg. Muir, W.W.3., Sams, R., 1992. Effects of ketamine infusion on halothane 90, 408–414. minimal alveolar concentration in horses. Am. J. Vet. Res. 53, Haddad, S., Poulin, P., Krishnan, K., 2000. Relative lipid content as the sole 1802–1806. mechanistic determinant of the adipose tissue:blood partition coefficients of Nickel, R., Schummer, A., Seiferle, E., 1995. Lunge. In: Habermehl, K.H., highly lipophilic organic chemicals. Chemosphere 40, 839–843. Vollmerhaus, B., Wilkens, H. (Eds.), Lehrbuch der Anatomie der Hartell, N.A., Headley, P.M., 1990. Spinal effects of four injectable anaesthetics Haustiere; Band II: Eingeweide. Blackwell Wissenschafts-Verlag, Berlin- on nociceptive reflexes in rats: a comparison of electrophysiological and Wien, pp. 255–265. behavioural measurements. Br. J. Pharmacol. 101, 563–568. Oye, I., Paulsen, O., Maurset, A., 1992. Effects of ketamine on sensory Headley, P.M., Parsons, C.G., West, D.C., 1987. The role of N-methylaspartate perception: evidence for a role of N-methyl-D-aspartate receptors. J. receptors in mediating responses of rat and cat spinal neurones to defined Pharmacol. Exp. Ther. 260, 1209–1213. sensory stimuli. J. Physiol. 385, 169–188. Parks, C.M., Manohar, M., 1983. Distribution of blood flow during moderate Henthorn, T.K., Krejcie, T.C., Niemann, C.U., Enders-Klein, C., Shanks, C.A., and strenuous exercise in ponies (Equus caballus). Am. J. Vet. Res. 44, Avram, M.J., 1999. Ketamine distribution described by a recirculatory 1861–1866. pharmacokinetic model is not stereoselective. Anesthesiology 91, Pedraz, J.L., Lanao, J.M., Hernandez, J.M., Dominguez-Gil, A., 1986. The 1733–1743. biotransformation kinetics of ketamine “in vitro" in rabbit liver and lung Hijazi, Y., Boulieu, R., 2002. Contribution of CYP3A4, CYP2B6, and CYP2C9 microsome fractions. Eur. J. Drug Metab. Pharmacokinet. 11, 9–16. isoforms to N-demethylation of ketamine in human liver microsomes. Drug Persson, J., Hasselstrom, J., Maurset, A., Oye, I., Svensson, J.O., Almqvist, O., Metab. Dispos. 30, 853–858. Scheinin, H., Gustafsson, L.L., 2002. Pharmacokinetics and non-analgesic Hijazi, Y., Bodonian, C., Bolon, M., Salord, F., Boulieu, R., 2003. effects of S- and R-ketamines in healthy volunteers with normal and reduced Pharmacokinetics and haemodynamics of ketamine in intensive care metabolic capacity. Eur. J. Clin. Pharmacol. 57, 869–875. patients with brain or spinal cord injury. Br. J. Anaesth. 90, 155–160. Poulin, P., Krishnan, K., 1996. A tissue composition-based algorithm for Ihmsen, H., Geisslinger, G., Schuttler, J., 2001. Stereoselective pharmacoki- predicting tissue: air partition coefficients of organic chemicals. Toxicol. netics of ketamine: R(−)-ketamine inhibits the elimination of S(+)-ketamine. Appl. Pharmacol. 136, 126–130. Clin. Pharmacol. Ther. 70, 431–438. Price, D.D., Mao, J., Frenk, H., Mayer, D.J., 1994. The N-methyl-D-aspartate Jansen, K.L., 2000. A review of the nonmedical use of ketamine: use, users and receptor antagonist dextromethorphan selectively reduces temporal summa- consequences. J. Psychoactive Drugs 32, 419–433. tion of second pain in man. Pain 59, 165–174. Joubert, K., 1998. Ketamine hydrochloride—An adjunct for analgesia in dogs Reddy, M.B., Andersen, M.E., Morrow, P.E., Dobrev, I.D., Varaprath, S., with burn wounds. J. S. Afr. Vet. Assoc. 69, 95–97. Plotzke, K.P., Utell, M.J., 2003. Physiological modeling of inhalation Kharasch, E.D., Labroo, R., 1992. Metabolism of ketamine stereoisomers by kinetics of octamethylcyclotetrasiloxane in humans during rest and exercise. human liver microsomes. Anesthesiology 77, 1201–1207. Toxicol. Sci. 72, 3–18. Klepstad, P., Maurset, A., Moberg, E.R., Oye, I., 1990. Evidence of a role for Scallet, A.C., Schmued, L.C., Slikker Jr., W., Grunberg, N., Faustino, P.J., NMDA receptors in pain perception. Eur. J. Pharmacol. 187, 513–518. Davis, H., Lester, D., Pine, P.S., Sistare, F., Hanig, J.P., 2004. Develop- Kohn, M.C., Melnick, R.L., 1996. Effects of the structure of a toxicokinetic mental neurotoxicity of ketamine: morphometric confirmation, exposure model of butadiene inhalation exposure on computed production of parameters, and multiple fluorescent labeling of apoptotic neurons. Toxicol. carcinogenic intermediates. Toxicology 113, 31–39. Sci. 81, 364–370. Kohrs, R., Durieux, M.E., 1998. Ketamine: teaching an old drug new tricks. Schwieger, I.M., Szlam, F., Hug Jr., C.C., 1991. The pharmacokinetics and Anesth. Analg. 87, 1186–1193. pharmacodynamics of ketamine in dogs anesthetized with enflurane. J. Koo, J.W., Parham, F., Kohn, M.C., Masten, S.A., Brock, J.W., Needham, L.L., Pharmacokinet. Biopharm. 19, 145–156. Portier, C.J., 2002. The association between biomarker-based exposure Spadavecchia, C., Spadavecchia, L., Andersen, O.K., Arendt-Nielsen, L., estimates for phthalates and demographic factors in a human reference Leandri, M., Schatzmann, U., 2002. Quantitative assessment of nociception population. Environ. Health Perspect. 110, 405–410. in horses by use of the nociceptive withdrawal reflex evoked by Larsson, P., Persson, E., Tyden, E., Tjalve, H., 2003. Cell-specific activation of transcutaneous electrical stimulation. Am. J. Vet. Res. 63, 1551–1556. aflatoxin B1 correlates with presence of some cytochrome P450 enzymes in Spadavecchia, C., Andersen, O.K., Arendt-Nielsen, L., Spadavecchia, L., olfactory and respiratory tissues in horse. Res. Vet. Sci. 74, 227–233. Doherr, M., Schatzmann, U., 2004. Investigation of the facilitation of the Laulin, J.P., Maurette, P., Corcuff, J.B., Rivat, C., Chauvin, M., Simonnet, G., nociceptive withdrawal reflex evoked by repeated transcutaneous electrical 2002. The role of ketamine in preventing fentanyl-induced hyperalgesia and stimulations as a measure of temporal summation in conscious horses. Am. subsequent acute morphine tolerance. Anesth. Analg. 94, 1263–1269 J. Vet. Res. 65, 901–908. (table). Spadavecchia, C., Levionnois, O., Kronen, P.W., Leandri, M., Spadavecchia, Loscher, W., Ganter, M., Fassbender, C.P., 1990. Correlation between drug and L., Schatzmann, U., 2005. Peri-MAC isoflurane depression of a metabolite concentrations in plasma and anesthetic action of ketamine in nociceptive withdrawal reflex evoked by transcutaneous electrical swine. Am. J. Vet. Res. 51, 391–398. stimulation in ponies. Am. J. Vet. Res. 66, 1992–1998. Lutz, R.J., Dedrick, R.L., Zaharko, D.S., 1980. Physiological pharmacoki- Theurillat, R., Knobloch, M., Levionnois, O., Larenza, P., Mevissen, M., netics: an in vivo approach to membrane transport. Pharmacol. Ther. 11, Thormann, W., 2005. Characterization of the stereoselective biotransfor- 559–592. mation of ketamine to norketamine via determination of their Maitre, P.O., Shafer, S.L., 1990. A simple pocket calculator approach to predict enantiomers in equine plasma by capillary electrophoresis. Electrophor- anesthetic drug concentrations from pharmacokinetic data. Anesthesiology esis 26, 3942–3951. 73, 332–336. Trevor, A.J., Woolf, T.F., Baillie, T.A., Adams, J.D., Castagnoli, N., 1983. Manohar, M., Gustafson, R., Goetz, T.E., Nganwa, D., 1987. Systemic Stereoselective metabolism of ketamine enantiomers, phencyclidine and distribution of blood flow in ponies during 1.45%, 1.96%, and 2.39% related arylcyclohexamines: present and future applications. In:

end-tidal isoflurane–O2 anesthesia. Am. J. Vet. Res. 48, 1504–1510. Kamenka, J.M., Domino, E.F., Geneste, P. (Eds.), NPP Books. Ann McConaghy, F.F., Hodgson, D.R., Rose, R.J., Hales, J.R., 1996. Redistribution Arbor, pp. 279–289. of cardiac output in response to heat exposure in the pony. Equine Vet. J. Wagner, A.E., Walton, J.A., Hellyer, P.W., Gaynor, J.S., Mama, K.R., 2002. Use Suppl. 42–46. of low doses of ketamine administered by constant rate infusion as an Mendell, L.M., Wall, P.D., 1965. Responses of single dorsal cord cells to adjunct for postoperative analgesia in dogs. J. Am. Vet. Med. Assoc. 221, peripheral cutaneous unmyelinated fibres. Nature 206, 97–99. 72–75. Menigaux, C., Guignard, B., Fletcher, D., Sessler, D.I., Dupont, X., Chauvin, Waterman, A.E., 1983. Influence of premedication with xylazine on the M., 2001. Intraoperative small-dose ketamine enhances analgesia after distribution and metabolism of intramuscularly administered ketamine in outpatient knee arthroscopy. Anesth. Analg. 93, 606–612. cats. Res. Vet. Sci. 35, 285–290. 386 M. Knobloch et al. / Toxicology and Applied Pharmacology 216 (2006) 373–386

Waterman, A.E., Robertson, S.A., Lane, J.G., 1987. Pharmacokinetics of Williams, M.L., Wainer, I.W., 2002. Role of chiral chromatography in intravenously administered ketamine in the horse. Res. Vet. Sci. 42, therapeutic drug monitoring and in clinical and forensic toxicology. Ther. 162–166. Drug Monit. 24, 290–296. Werner, C., Reeker, W., Engelhard, K., Lu, H., Kochs, E., 1997. Ketamine Woolf, C.J., Thompson, S.W., 1991. The induction and maintenance of racemate and S-(+)-ketamine. Cerebrovascular effects and neuroprotection central sensitization is dependent on N-methyl-D-aspartic acid receptor following focal ischemia. Anaesthesist 46 (Suppl. 1), 55–60. activation; implications for the treatment of post-injury pain hypersensi- White, P.F., Pudwill, C.R., Johnston, R.R., Way, W.L., Trevor, A.J., Marietta, M. tivity states. Pain 44, 293–299. P., 1976. Effects of halothane anesthesia on ketamine biotransformation. Wright, M., 1982. Pharmacologic effects of ketamine and its use in veterinary Proc. West. Pharmacol. Soc. 19, 227–231. medicine. J. Am. Vet. Med. Assoc. 180, 1462–1471. White, P.F., Way, W.L., Trevor, A.J., 1982. Ketamine–its pharmacology and Yanagihara, Y., Kariya, S., Ohtani, M., Uchino, K., Aoyama, T., Yamamura, therapeutic uses. Anesthesiology 56, 119–136. Y., Iga, T., 2001. Involvement of CYP2B6 in n-demethylation of keta- White, P.F., Schuttler, J., Shafer, A., Stanski, D.R., Horai, Y., Trevor, A.J., 1985. mine in human liver microsomes. Drug Metab. Dispos. 29, 887–890. Comparative pharmacology of the ketamine isomers. Studies in volunteers. Zielmann, S., Kazmaier, S., Schnull, S., Weyland, A., 1997. S-(+)-ketamine and Br. J. Anaesth. 57, 197–203. circulation. Anaesthesist 46 (Suppl. 1), 43–46. Veterinary Anaesthesia and Analgesia, 2009, 36, 9–17 doi:10.1111/j.1467-2995.2008.00420.x

RESEARCH PAPER Determination of the minimum alveolar concentration of isoﬂurane in Shetland ponies using constant current or constant voltage electrical stimulation

Olivier L Levionnois* DVM, Dr Med Vet, Diplomate ECVAA, Claudia Spadavecchia* DVM, Dr Med Vet, PhD, Diplomate ECVAA, Peter W Kronen DVM, Dr Med Vet, Diplomate ECVAA & Urs Schatzmann* Dr Med Vet, Diplomate ECVAA *Department of Clinical Veterinary Sciences, Vetsuisse Faculty, University of Berne, Berne, Switzerland Veterinary Anaesthesia Services International, Winterthur, Switzerland

Correspondence: Olivier L Levionnois, Anaesthesiology section, Department of Clinical Veterinary Sciences, Vetsuisse Faculty, University of Berne, Postfach 8466, 3001 Berne, Switzerland. E-mail: [email protected]

was accompanied by variable inter-electrode resis- Abstract tances resulting in uncontrolled stimulus intensity. Objective To determine the minimum alveolar con- At 0.9 MAC, the third stimulation induced more centration (MAC) of isoflurane in Shetland ponies positive responses than the first stimulation, inde- using a sequence of three different supramaximal pendent of the mode of stimulation used. noxious stimulations at each tested concentration of isoflurane rather than a single stimulation. Conclusions The MAC of isoflurane in the Shetland ponies was lower than expected with considerable Study design Prospective, experimental trial. variability among individuals. Constant current surface electrode stimulations were the most repeatable. Animals Seven 4-year-old, gelding Shetland ponies. A summation over the sequence of three supramaximal stimulations was observed around 0.9 MAC. Methods The MAC of isoflurane was determined for each pony. Three different modes of electrical Clinical relevance The possibility that Shetland stimulation were applied consecutively (2 minute ponies require less isoflurane than horses needs intervals): two using constant voltage (90 V) on the further investigation. Constant current surface- gingiva via needle- (CVneedle) or surface-electrodes electrode stimulations were the most repeatable. (CVsurface) and one using constant current (CC; Repetitive supramaximal stimuli may have evoked 40 mA) via surface electrodes applied to the skin movements at isoflurane concentrations that over the digital nerve. The ability to clearly interpret provide immobility when single supramaximal the responses as positive, the latency of the evoked stimulation was applied. responses and the inter-electrode resistance were Keywords anaesthesia, isoflurane, minimum alveo- recorded for each stimulus. lar concentration, noxious stimulation, Shetland pony. Results Individual isoflurane MAC (%) values ranged from 0.60 to 1.17 with a mean (±SD) of 0.97 (±0.17). The responses were more clearly Introduction interpreted with CC, but did not reach statistical significance. The CVsurface mode produced The minimum alveolar concentration (MAC) responses with a longer delay. The CVneedle mode concept was developed by Merkel, Eger and

9 Isoflurane MAC in Shetland ponies OL Levionnois et al. co-workers to compare inhalational anaesthetic Such CV modes of stimulation depolarize the nerves potencies and was defined as the lowest alveolar with an intensity that depends on inter-electrode concentration of anaesthetic (at one atmosphere) electrical resistance (Kimura 1989). Thus, varia- required to prevent gross purposeful movement in tions in inter-electrode electrical resistance over response to a supramaximal noxious stimulus in time and within the study lead to inconsistent 50% of the population (Merkel & Eger 1963; Eger stimulation intensity. Conversely, application of a et al. 1965a,b). This technique applied repetitive constant current (CC) compensates for variations of stimuli at different isoflurane concentrations to inter-electrode resistances and therefore may elicit determine individual MAC and has been used fewer fluctuations in stimulation than a CV (Stecker extensively in veterinary research (Steffey et al. 2004). Constant current stimulation has never been 1977). described for animal MAC-studies, but has Individual differences in MAC values are com- been used to improve reproducibility and sensi- monly assumed to reflect different sensitivities to tivity in nerve stimulation studies in horses inhalational anaesthetic agents (Sonner 2002). (Spadavecchia et al. 2002). Attempts to discrimi- However, methods of stimulation can also increase nate the differences in physical response to diverse variability and alter results. Various types of noxious electrical stimulation modes have not been reported stimuli such as tail clamping, paw pressure or in equids. electrical nerve stimulation have been applied to The individual isoflurane MAC was determined in determine MAC (Steffey et al. 1977; Eger et al. 1988; seven Shetland ponies, comparing the movements Valverde et al. 2003). Eger et al. (1965a) reported induced in response to three different electrical that electrical stimulation, applied at adequate inten- stimulation modes: a CV applied via needle electrodes sity, consistently provided a supramaximal noxious in the gingiva; a CV applied via surface electrodes on input. Compared to other techniques, electrical stim- the gingiva; a CC applied via surface electrodes on the ulation is totally reversible despite the high intensity distal digital nerve. The hypothesis that a CC mode delivered and maintains an intact neurophysiology could evoke more repeatable stimulation, with and tissue integrity (Le Bars et al. 2001). An impor- clearer and quicker responses, was tested. tant question is whether the stimulus is ‘supramaximal’ or not (Antognini & Carstens 2005). The Materials and methods application of a submaximal stimulus requires lower anaesthetic concentrations than MAC to induce Animals immobility. For MAC studies, a noxious stimulus is considered supramaximal when an increase in The study was performed on seven healthy geld- intensity does not require a higher anaesthetic ing Shetland ponies, 4 years old, with body concentration to prevent purposeful movements. weight of 122 ± 16 kg (mean ± SD) whose left The responses to supramaximal stimulations have carotid artery was surgically elevated earlier to a been described as ‘all-or-none’. However, increasing subcutaneous position. The committee for animal inhalational anaesthetic concentration depresses experimentation, County of Berne, Switzerland, progressively – rather than abruptly – the movements approved the study. Each pony was individually in response to nociception (Antognini & Carstens housed and fasted with permanent access to 2005). Furthermore, investigators must state water during the 24 hours prior to the experi- subjectively if the animal did or did not display ‘gross ment. On the morning of the experiment, the and purposeful movement’. The variability in the ponies were physically examined and a complete interpretation of the stimulation-evoked responses blood cell count and serum biochemistry analyses during MAC determination may alter the final results were performed. Ponies were used in the experi- and deserves further investigation. ment if these tests proved to be within the In horses, a square-shaped constant voltage (CV) reference ranges. electrical stimulation at 50 V and 5 Hz, applied through needle electrodes inserted deeply in the Anaesthetic technique gingiva, is commonly used for MAC determination (Steffey et al. 1977; Pascoe et al. 1993; Bennett Anaesthesia was induced in a quiet and darkened et al. 2004). Surface electrodes have also been used room using isoflurane (Isoflo; Abbott A.G., Baar, over the distal digital nerve (Doherty et al. 1997a). Switzerland) in oxygen (O2). The ponies were posi-

10 2009 The Authors. Journal compilation 2009 Association of Veterinary Anaesthetists, 36, 9–17 Isoflurane MAC in Shetland ponies OL Levionnois et al. tioned in right lateral recumbency by gentle manual isoflurane was reduced to 0.9%. Thirty minutes restraint, without any medication. The investigators later (approximately 1 hour of anaesthesia), the and the ponies were trained in this technique. Oxygen three modes of electrical stimulation were applied (6 L minute)1, 100%) and isoflurane (delivered in a once, sequentially, to verify the proper function of stepwise increase of 0.5% for every 30 seconds up to the equipment. The responses to this first stimu- 3%) were administered through a standard anaes- lation series were ignored. thetic circle system via an air-tight face mask. Once the palpebral reflex was absent, the trachea was Determination of MAC intubated [14 or 16 mm ID, cuffed endotracheal tube (ETT)] and isoflurane end-tidal concentration main- A sequence of three different electrical stimulations, tained at 1.3%. Intermittent positive-pressure venti- instead of a single noxious stimulation usually lation was initiated with a respiratory rate (fR)of8 described for MAC determination (Steffey et al. breaths minute)1 and a tidal volume (TV) of 1977), was applied at each concentration of isoflu- 12 mL kg)1. Lactated Ringer’s solution was admin- rane tested. The end-tidal isoflurane concentration istered throughout anaesthesia at 5 mL kg)1 hour)1 was kept constant (reading precision of 0.01%, via a 16-SWG 8.2-cm catheter inserted in a medial within ±0.03% of the target value) for a period of saphenous vein. The left carotid artery was cathe- 30 minutes (result averaged) before a stimulus was terized with a 20-SWG, 4.8-cm catheter and applied. The application order of the three stimuli connected to a calibrated pressure transducer was rotated. A minimum interval of 2 minutes after (Angiokard one-use transducer; Medizintechnik complete return to immobility and cardiovascular GmbH, Friedeburg, Germany) for measurement of stabilisation was allowed between subsequent systolic, mean (MAP), and diastolic arterial blood stimuli. Each stimulus was applied until a positive pressures and heart rate (HR). The electrocardiogram response was observed or for a maximum of (base-apex ECG), distal oesophageal temperature (T), 60 seconds. A response was considered positive inspired and end-tidal fractions (sampled by side when a gross purposeful movement of a nonstimu- stream at ETT end) of O2 and isoflurane (FIO2,FE¢Iso), lated body area was observed. Slight movements, fR, tidal volume (V_ T), minute ventilatory volume and muscle tremors, swallowing, nystagmus and phys- peak inspiratory pressure (PIP) were continuously iological parameter modifications were considered recorded. All parameters were displayed on a moni- as negative responses. If no positive response toring unit (S/5 Compact; Datex-Ohmeda, Helsinki, occurred after the three stimuli, the FE¢Iso was Finland) and recorded every 5 seconds on a con- reduced by 0.1% (approximately 10% of its value); nected laptop computer by specially designed this concentration was kept constant for 30 min- software (S/5 Collect; Datex-Ohmeda, Helsinki, utes, and then the three stimuli were applied again. Finland). Barometric pressure was measured hourly When at least one of the three stimuli elicited a in kPa (PATM). The gas-monitoring module was cali- positive response, the FE¢Iso was increased by 0.1% brated according to the manufacturer’s recommen- (approximately 10% of its value), the new concen- dations using a commercial calibration gas tration was kept constant for 30 minutes, and then containing 3% desflurane (Quick Cal; Datex-Ohmeda) the three stimuli were applied again. If one stimulus before each pony was anaesthetized. To avoid elicited a negative response when another one influences on MAC, the MAP was maintained elicited a positive response at the same end-tidal over 70 mmHg by intravenous infusion of concentration, the negative response was recorded dobutamine, (Dobutamin; Fresenius Kabi A.G., as a false-negative response (FNR). The MAC was Stans, Switzerland) as required, in a range of 0.25– individually calculated as the average of two suc- 1 lgkg)1 minute)1. A forced warm air device cessive concentrations (mean over 30 minutes), one (Bair Hugger 505; Carbamed, Berne, Switzerland) allowing and one preventing gross purposeful was used to maintain body temperature stable movement in response to the series of three stimuli. around 36 C throughout, and mechanical venti- The individual MAC was determined in triplicate, lation settings were adapted to maintain an and the average (iMACAmb) reported after com- end-tidal CO2 of 4.7 kPa (35 mmHg). Arterial pensation for barometric pressure at sea level blood-gas analysis was first performed after placing [iMAC = iMACAmb(PATM/100)]. The population the arterial catheter and then repeated hourly. MAC was calculated as the arithmetic mean of the Upon completion of instrumentation, the end-tidal seven individual MACs.

2009 The Authors. Journal compilation 2009 Association of Veterinary Anaesthetists, 36, 9–17 11 Isoﬂurane MAC in Shetland ponies OL Levionnois et al.

scored from 1 to 3 (1 = >45 seconds, 2 = between Electrical stimulation 10 and 45 seconds, 3 = <10 seconds). Median and For each MAC determination, three different modes range of scores for each mode of stimulation were of electrical stimulation were applied sequentially: calculated. two using CV and one using CC. Gingival stimulations were delivered by needle electrodes (CVneedle; Statistical analysis Stimulation platinum needles 8 cm; Comepa, Bagnolet, France) or surface electrodes (CVsurface; Statistical analysis of all data was performed using Neuroline 70002-J; Medicotest, Olstykke, Denmark) standard statistical software (NCSS-2004 statistical positioned on the rostral gingiva. The two needle software; NCSS, Kaysville, UT, USA). Parametric electrodes were inserted 2 cm deep in the maxillary data (HR, MAP, fR, V_ T, PIP, PE¢CO2, SpO2, SaO2, gingiva, each 1.5 cm from midline. Just lateral to PaCO2, PaO2, pH, T, duration of anaesthesia) were each, the surface electrodes were applied. Each pair expressed as mean (±SD) of all recording points for of electrodes was intermittently connected to an the seven ponies. Differences in the frequency of electrical stimulator unit (Grass S88; Grass medical unclear responses and in the scores of latency instruments, MA, USA) to deliver 10-msecond CV among modes of stimulation were tested with a chi- (90 V) square-wave pulses at a frequency of 5 Hz. squared test. The hypothesis that the occurrence of The upper lip was elevated with an adhesive strip to FNR was correlated with the mode of stimulation minimize interferences with the stimulation elec- used or with the rank of the stimulation in the trodes. The electrical resistance between the elec- sequence of three was tested by a chi-squared trodes was measured with an amperometer test. Differences in inter-electrode resistances for (Supertester 680; I.C.E., Milano, Italy) prior to each CVneedle and CVsurface were tested using an ANOVA stimulation and recorded. To deliver the CC for repeated measures. Statistical significance was stimulation, two surface electrodes were applied set at p < 0.05 and at p < 0.10 for the presence of a (inter-electrode distance of 1 cm) to the shaved and trend. degreased skin over the lateral palmar digital nerve between the coronary band and the fetlock joint, Results and secured with adhesive bandages. Stimuli consisted of a 25 msecond train of five 1-msecond CC The overall acceptance of restraint before anaesthetic (40 mA) square-wave pulses. The trains of five were induction by mask was good, as only one pony dis- delivered at a frequency of 5 Hz and were delivered played slight struggling. For the other ponies, lateral from a battery-powered CC stimulator purposely recumbency was achieved in a smooth and appar- designed, and controlled by a computer. As the ently stress-free manner. Immobility occurred after maximum voltage delivered by the stimulator was the ponies assumed lateral recumbency, character- 200 V, the resistance between the electrodes was ized by general relaxation. Throughout the periods of regularly measured. If above 3 kW, the electrodes observation, HR, MAP, fR, V_ T, PIP, PE¢CO2, SaO2, )1 were replaced to ensure discharge of a current of PaCO2, PaO2, pH and T were 37 ± 5 beats minute , 40 mA. Further details of this equipment have been 93 ± 7 mmHg, 8 ± 1 breaths minute)1, 1.5 ± reported (Spadavecchia et al. 2002, 2003). 0.2 L, 3.11 ± 0.51 (31.7 ± 5.2 cm H2O) kPa, 4.51 ± 0.22 kPa (34 ± 2 mmHg), 99.8 ± 0.2%, 4.92 ± 0.2 kPa (37 ± 2 mmHg), 58.25 ± 7.85 kPa Analysis of responses to stimulations (437 ± 59 mmHg), 7.4 ± 0.03 and 35.9 ± 0.6 C, The quality of the responses to stimulation was respectively. Duration of anaesthesia was 471 evaluated by two parameters: clearness and latency. (±37) minutes. Dobutamine was administered to all Each response was judged clear or unclear. A posi- of the ponies for a mean time of 194 minutes per pony tive response was judged unclear when the origin of and at a mean dose of 0.45 lgkg)1 minute)1. the gross movements was doubtful. A negative The iMAC values were all below 1.17% with a response was judged unclear when slight move- mean (±SD) of 0.97 (±0.17) (Table 1). Unclear ments of reflex origin were doubtful. The latency of responses were not statistically significant (p = the response reflected the time between onset of the 0.54) between the three modes of stimulation (10/ stimulus and occurrence of the first purposeful 43; 23%, 14/41; 34% and 12/43; 28% for CC, movement. Each positive response latency was CVsurface and CVneedle, respectively. The scores for

12 2009 The Authors. Journal compilation 2009 Association of Veterinary Anaesthetists, 36, 9–17 Isoﬂurane MAC in Shetland ponies OL Levionnois et al.

Table 1 Individual and mean (±SD) isoﬂurane MAC val- with the CC stimuli and three with the CVneedle ues (%) stimuli (Table 2). No statistical difference was observed in the occurrence of FNR between modes

Ponies iMACAmb iMAC of stimulation. Considering the rank of stimulation independently to the mode used, seven FNR were

1 1.12 1.05 associated with the first of the three stimuli, four 2 0.64 0.6 with the second and two with the third stimulius 3 1.06 1 (Table 2). This trend did not reach statistical signifi- 4 1.15 1.09 cance with the 42 observations. According to the 5 0.96 0.91 chi-value, a correlation between the occurrence of 6 1.04 0.96 7 1.25 1.17 FNR at 0.9 iMAC and the rank of stimulation would reach significance (p < 0.05) if the trend would be Mean 1.03 0.97 conserved over more than sixty observations (for ±SD 0.18 0.17 each stimulation mode, six stimuli more than in the present study). iMACAmb were measured at daily ambient barometric pressure. iMAC were calculated by correction for barometric pressure at Positive responses to stimulation were mostly sea level (100 kPa). head movements. Some weak limb movements also occurred in response to electrical stimulation in 29.4%, 29.4% and 13.3% for CVneedle, CC and the latency of responses (Fig. 1) to CVsurface stimuli CVsurface stimulations respectively. Only in one were found to be lower than to CVneedle pony, CVsurface-stimulation elicited limb move- (p = 0.028) and to CC (p = 0.089). No significant ments without head movement, but a tonic con- differences were found between the latencies from traction of neck muscles occurred, which could have CVneedle and CC (p = 0.85). Latencies increased masked movement of the head. The inter-electrode with increasing isoflurane concentration with the resistance was found to be significantly higher and proportion of ‘immediate’ responses decreasing from more variable for needle- than for surface electrodes 100% at 0.6 iMAC to 68%, 77%, 58% and 0% at (22.28 ± 11.7 and 2.82 ± 1.9 kW respectively; 0.7, 0.8, 0.9 and 1.1 iMAC respectively. Fig. 2). All ponies recovered quickly (16 ± 5 min- False negative response occurred with each pony utes) and smoothly after discontinuation of isoflu- but only at 0.9 iMAC. Over the 127 responses rane administration. observed, 42 responses were recorded at 0.9 iMAC (14 groups of 3) from which 13 were FNR. Five FNR were associated with the CVsurface stimuli, five Table 2 Occurrence of false negative responses (FNR) and positive responses (PR) to stimulations at 0.9 iMAC: percentages of FNR for each category over the total number of FNR; Number of FNR and PR to electrical stimuli with CVneedle, CVsurface and CC or with first, second and third stimuli

FNR PR

Percentage Number Number

Mode of stimulation CVneedle 23% 3 11 CVsurface 38% 5 9 CC 38% 5 9 Total 100% 13 29 Rank Figure 1 Evaluation of the response latencies to electrical CVneedle 54% 7 7 stimuli using constant voltage with needle- (CVneedle) or CVsurface 31% 4 10 surface electrodes (CVsurface), or constant current (CC): CC 15% 2 12 percentage of immediate, late and very late responses over Total 100% 13 29 the total number of stimuli.

2009 The Authors. Journal compilation 2009 Association of Veterinary Anaesthetists, 36, 9–17 13 Isoﬂurane MAC in Shetland ponies OL Levionnois et al.

mechanical ventilation), O2 (by high-inspired O2 fraction) and arterial pressure (by dobutamine), as used in this study, are important factors known to reduce bias. The low mean body temperature observed in the present study may have contributed to decrease MAC up to about 10% (Antognini 1993). However, individuals with the lowest body temperature did not have the lowest MAC values. The use of end-tidal isoflurane concentration may further increase the difference between alveolar and effect site gas concentration when ventilation to perfusion (V_ =Q)_ mismatch exists. However, a difference between inspired and expired isoflurane Figure 2 Mean, standard deviation, maximum and mini- concentration of <10% – as was always the case in mum of the electrical resistances measured between the present study – has been associated with a needle- (CVneedle) and surface-electrodes (CVsurface). negligible difference between end-tidal and arterial partial pressures of isoflurane (Eger & Bahlman 1971). Moreover, the end-tidal isoflurane concen- Discussion trations were measured by infrared technology at a Our findings suggest that the Shetland ponies in this long wavelength instead of short wavelength to study had relatively low MAC values, and one decrease measurement errors (Dujardin et al. individual was remarkably more sensitive to the 2005). As inhalational anaesthetic requirements anaesthetic effects of isoflurane than the rest. The may decrease in the first hour and after a first strong individual MAC variability is thought to represent noxious stimulation, the first series of stimuli were primarily genetic sensitivity to inhalational anaes- applied only after 1 hour of anaesthesia and the thetics (Sonner 2002). This may account for 10– results discarded (Eger et al. 1965a; Petersen-Felix 15% of differences within a population and between et al. 1993; Higuchi & Adachi 2002). Manual species (de Jong & Eger 1975; Travis & Bowers restraint and inhalational mask induction in adult 1991; Sonner et al. 2000). However, the reasons horses may induce severe stress and increase for MAC variability within a population are still endogenous catecholamine release, which increased poorly understood. In Equidae, for instance, pub- MAC (Steffey & Eger 1975; Antognini & Carstens lished values of halothane MAC have a variability of 2005). The use of gentle manual restraint and more than 20% (0.88–1.07%) in small horse pop- apparently stress-free induction in our Shetland ulations (Steffey et al. 1977; Bennett et al. 2004). ponies may have reduced the influence of stress on In nine Shetland ponies, individual halothane MAC isoflurane potency in comparison with other MAC values varied between 1% and 2% (Matthews & studies in horses and may have contributed to the Lindsay 1990). In several other horse studies, var- low MAC values observed. Duration of anaesthesia iability >10% was observed when iMACs were in this study was between 7 and 8 hours. Quasha measured repeatedly in the same animals and with et al. (1980) showed that, inhalational anaesthesia identical methodology (Matthews & Lindsay 1990; of up to 10 hours did not modify anaesthetic Steffey et al. 2000, 2003). If this large MAC vari- requirements. ability reflected differences in individual sensitivity Values for MAC can be underestimated when to inhalational anaesthetics, it appears important – purposeful movements are not very obvious and and particularly in Equids – to determine the indi- therefore interpreted as negative responses. The vidual baseline MAC for each subject rather than positive responses were defined as in other MAC selecting a MAC value from the literature where studies. Although other MAC studies only described MAC is expected to influence the studied para- all-or-none responses, unclear and delayed meters. responses to stimulations were reported in this The MAC can also be lowered by methodological study. In fact, purposeful movements are uncoordi- factors. The determination in triplicate, the use of nated and weak during inhalational anaesthesia, so 10% incremental steps and the maintenance within that responses to noxious stimulations are not physiological temperature ranges, arterial CO2 (by necessarily gross and obvious. Antognini et al.

14 2009 The Authors. Journal compilation 2009 Association of Veterinary Anaesthetists, 36, 9–17 Isoﬂurane MAC in Shetland ponies OL Levionnois et al.

(1999) reported that isoflurane progressively A voltage of 90 was chosen for this study as 50-V depressed movements in response to supramaximal stimulation was insufficient to elicit effective stim- stimulation in a dose-dependent manner. The ulation in preliminary trials. Despite this, CVsurface results presented here confirm the observation of probably failed to reach supramaximal intensity in Dutton et al. (2003) that a delay appears between some cases, producing the longer latencies and the noxious stimuli and the evoked movement unclear responses. The voltage can hardly be as the inhalational anaesthetic concentration compared to other studies as resistances between increases. The interpretation of movements can be electrodes in previous studies have not been considered more difficult and subjective around 0.9 reported. Needle electrodes had very high and MAC, where gradual muscle relaxation occurs. variable inter-electrodes resistances. However, this Interestingly, the ponies in this study, showing the type of electrode can be inserted closer to the nerve less clear responses (pony nos 1 and 4), had the fibres and the field of electrical stimulation is highest MAC values. Thus, some difficulties in stronger between the electrode tips (Kimura discriminating presence or absence of movement 1989). Thus, lower intensities are required to evoke and long latencies were present but did not seem to maximal effect (Polhill et al. 1998). Despite having lead to underestimation of MAC. lower intensity, the electrical stimuli generated by No statistical differences were found, but the needle electrodes may still have been supramaximal, observer-judged responses to CC stimuli were more evoking a more immediate response. The stimulus often clearer than CV stimuli. Noxious stimuli intensity, provided by CV devices, may also vary generally induce a nonpurposeful (reflex) motion during the stimulation event according to the tissue of the stimulated body area which is not depressed impedance, and may become submaximal (Kimura at MAC level (Doherty et al. 1997b; Antognini & 1989). The application of a CC instead of a CV Carstens 2005). As the head appeared to be the should further improve the control of the stimula- prevailing body area moving in the ponies, stimu- tion efficacy and its repeatability (Kimura 1989; lation at the limb (CI mode) may have helped to Stecker 2004). The CC applied close to the digital discriminate between withdrawal reflex motions nerve via surface electrodes provided a repeatable and purposeful movements. It appeared particu- supramaximal stimulus and evoked the clearest larly, in one pony (no. 1), that the neck muscles responses in this study. The stimulation mode CC contracted strongly in response to stimuli applied to and to a lesser extent CVneedle appeared more the gingiva (CVneedle and CVsurface). The neck appropriate than CVsurface to evoke supramaximal contraction can be interpreted as a marked noci- stimulation in ponies. ceptive reflex which provoked contraction of the A striking finding was that the FNR at 0.9 MAC digastric muscle (Brown et al. 2002). This could occurred mostly (seven times over 14) at the first have masked an eventual purposeful movement of stimulus – rather than at the third – independently the head, leading to unclear responses. Despite the of the stimulation mode. Although the application use of two stimulus locations we were unable to find of more than one supramaximal noxious stimulus a significant difference in the determined MAC did not seem to increase anaesthetic requirements values. With our results we cannot state whether (Eger et al. 1965a), it is not known whether this lack of difference was due to the mode of application of two or more submaximal, or different stimulation or its location. noxious stimuli requires more anaesthesia com- Studies of responses to noxious stimuli must pared with applying either one separately (Antog- ensure the quantification and consistency of the nini & Carstens 2005). According to the concept of stimulus intensity throughout the study (Zbinden the diffuse noxious inhibitory control (DNIC), a first et al. 1994). The use of surface electrodes on the noxious stimulus is expected to depress the response gingiva, in this study, provided stable and low inter- to a second one (Antognini & Carstens 2005). As electrode resistances predisposing to accurate high- multiple stimuli were used in our study, a larger stimulation intensity. However, the thick mucosal number of negative responses following the second tissue has high-water content and could have and third stimuli could have been expected. On the facilitated an undesirable superficial conduction of other hand, it is known that the DNIC is depressed the electrical current. The actual current reaching by isoflurane above 0.8 MAC (Jinks et al. 2003). In the target nerves could then be insufficient to induce our case, all the FNR occurred at 0.9 MAC. This depolarisation (Kimura 1989; Polhill et al. 1998). supported the hypothesis that in the absence of

2009 The Authors. Journal compilation 2009 Association of Veterinary Anaesthetists, 36, 9–17 15 Isoﬂurane MAC in Shetland ponies OL Levionnois et al.

DNIC (between 0.8 and 1 MAC), a temporal of intravenous and intrathecal administration of mor- summation occurred when repetitive stimuli were phine in anesthetized dogs. Am J Vet Res 63, 1349– applied, increasing the probability of movement 1353. (and MAC) in response to the third stimulus. Our Doherty TJ, Geiser DR, Rohrbach BW (1997a) Effect of understanding is that some summation occurred at high volume epidural morphine, ketamine and butorphanol on halothane minimum alveolar concentra- 0.9 MAC despite the use of at least a 2-minute tion in ponies. Equine Vet J 29, 370–373. interval between stimuli. This raises the question Doherty TJ, Geiser DR, Frazier DL (1997b) Comparison of whether, at higher isoflurane concentrations, stim- halothane minimum alveolar concentration and mini- uli may need to be of higher intensity to be mum effective concentration in ponies. J Vet Pharmacol supramaximal; or whether supramaximal stimuli Ther 20, 408–410. when repeated, may undergo temporal summation. Dujardin CL, Gootjes P, Moens YPM (2005) Isoflurane This study reports low MAC values for isoflurane measurement error using short wavelength infrared in Shetland ponies but previous studies have used techniques in horses: influence of fresh gas flow and pre- horses (Steffey et al. 1977; 2000 & 2003). The use anaesthetic food deprivation. Vet Anaesth Analg 32, of CC electrical stimulation for MAC determination 101–106. allowed repeatable supramaximal intensity and Dutton RC, Zhang Y, Stabernack CR et al. (2003) Tem- poral summation governs part of the minimum alveolar clear reactions, similar to the application of concentration of isoflurane anesthesia. Anesthesiology constant voltage stimuli through needle electrodes. 98, 1372–1377. The CV mode of stimulation applied through surface Eger EI II, Bahlman SH (1971) Is the end-tidal anesthetic electrodes at the gingiva in the ponies appeared to partial pressure an accurate measure of the arterial increase methodological variability, unclear and anesthetic partial pressure? Anesthesiology 35, 301– delayed responses but the results were not signifi- 303. cantly different then the other modes of stimulation. Eger EI II, Brandstater B, Saidman LJ (1965a) Equipotent Interestingly, the longer latencies and harder inter- alveolar concentrations of methoxyflurane, halothane, pretations of purposeful movements with increasing diethyl ether, fluroxene, cyclopropane, xenon and isoflurane concentration pointed to the difficult and nitrous oxide in the dog. Anesthesiology 26, 771–777. subjective discrimination between positive and Eger EI II, Saidman LJ, Brandstater B (1965b) Minimum alveolar anesthetic concentration: a standard of anes- negative movement patterns at anaesthetic concen- thetic potency. Anesthesiology 26, 756–763. trations around MAC. The summation evoked by Eger EI II, Johnson BH, Weiskopf RB et al. (1988) Mini- the three consecutive stimuli rather than by a single mum alveolar concentration of I-653 and isoflurane in standard stimulus revealed the unexpected result pigs: definition of a supramaximal stimulus. Anesth that using the three stimulus sequence tended to Analg 67, 1174–1176. increase the probability of evoking a positive Higuchi H, Adachi Y (2002) Decrease in minimum response (i.e. increase the MAC). alveolar concentration of sevoflurane during anaesthesia and arthroscopy. Eur J Anaesthesiol 19, 600–603. Jinks SL, Antognini JF, Carstens E (2003) Isoflurane References depresses diffuse noxious inhibitory controls in rats Antognini JF (1993) Hypothermia eliminates isoflurane between 0.8 and 1.2 minimum alveolar anesthetic requirements at 20 degrees C. Anesthesiology 78, concentration. Anesth Analg 97, 111–116. 1152–1156. de Jong RH, Eger ED (1975) MAC expanded: AD50 and Antognini JF, Carstens E (2005) Measuring minimum AD95 values of common inhalation anesthetics in man. alveolar concentration: more than meets the tail. Anesthesiology 42, 384–389. Anesthesiology 103, 679–680. Kimura J (1989) Electrodiagnosis in Diseases of Nerve and Antognini JF, Wang XW, Carstens E (1999) Quantitative Muscle: Principles and Practice. (2nd edn), Oxford Uni- and qualitative effects of isoflurane on movement versity Press, Philadelphia, PA, pp. 50. occurring after noxious stimulation. Anesthesiology 91, Le Bars D, Gozariu M, Cadden SW (2001) Animal models 1064–1071. of nociception. Pharmacol Rev 53, 597–652. Bennett RC, Steffey EP, Kollias-Baker C et al. (2004) Matthews NS, Lindsay SL (1990) Effect of low-dose butor- Influence of morphine sulfate on the halothane sparing phanol on halothane minimum alveolar concentration effect of xylazine hydrochloride in horses. Am J Vet Res in ponies. Equine Vet J 22, 325–327. 65, 519–526. Merkel G, Eger EI II (1963) A comparative study of halo- Brown DC, Bernier N, Shofer F et al. (2002) Use of non- thane and halopropane anesthesia including method for invasive dental dolorimetry to evaluate analgesic effects determining equipotency. Anesthesiology 24, 346–357.

16 2009 The Authors. Journal compilation 2009 Association of Veterinary Anaesthetists, 36, 9–17 Isoﬂurane MAC in Shetland ponies OL Levionnois et al.

Pascoe PJ, Steffey EP, Black WD et al. (1993) Evaluation of Stecker MM (2004) Nerve stimulation with an electrode of the effect of alfentanil on the minimum alveolar con- finite size: differences between constant current and centration of halothane in horses. Am J Vet Res 54, constant voltage stimulation. Comput Biol Med 34, 51– 1327–1332. 94. Petersen-Felix S, Zbinden AM, Fischer M et al. (1993) Steffey EP, Eger EI (1975) The effect of seven vasopres- Isoflurane minimum alveolar concentration decreases sors on halothane MAC in dogs. Br J Anaesth 47, 435– during anesthesia and surgery. Anesthesiology 79, 438. 959–965. Steffey EP, Howland D Jr, Giri S et al. (1977) Enflurane, Polhill SL, Clewlow F, Smith DC (1998) Are changes in the halothane, and isoflurane potency in horses. Am J Vet evoked electromyogram during anaesthesia without Res 38, 1037–1039. neuromuscular blocking agents caused by failure of Steffey EP, Pascoe PJ, Woliner MJ et al. (2000) Effects of supramaximal nerve stimulation? Br J Anaesth 81, xylazine hydrochloride during isoflurane-induced anes- 902–904. thesia in horses. Am J Vet Res 61, 1225–1231. Quasha AL, Eger EI II, Tinker JH (1980) Determination Steffey EP, Eisele JH, Baggot JD (2003) Interactions of and applications of MAC. Anesthesiology 53, 315–334. morphine and isoflurane in horses. Am J Vet Res 64, Sonner JM (2002) Issues in the design and interpretation 166–175. of minimum alveolar anesthetic concentration (MAC) Travis CC, Bowers JC (1991) Interspecies scaling of anes- studies. Anesth Analg 95, 609–614. thetic potency. Toxicol Ind Health 7, 249–260. Sonner JM, Gong D, Eger EI II (2000) Naturally occurring Valverde A, Morey TE, Hernandez J et al. (2003) Valida- variability in anesthetic potency among inbred mouse tion of several types of noxious stimuli for use in deter- strains. Anesth Analg 91, 720–726. mining the minimum alveolar concentration for Spadavecchia C, Spadavecchia L, Andersen OK et al. inhalation anesthetics in dogs and rabbits. Am J Vet Res (2002) Quantitative assessment of nociception in horses 64, 957–962. by use of the nociceptive withdrawal reflex evoked by Zbinden AM, Maggiorini M, Petersen-Felix S et al. (1994) transcutaneous electrical stimulation. Am J Vet Res 63, Anesthetic depth defined using multiple noxious stimuli 1551–1556. during isoflurane/oxygen anesthesia. I. Motor reactions. Spadavecchia C, Arendt-Nielsen L, Andersen OK et al. Anesthesiology 80, 253–260. (2003) Comparison of nociceptive withdrawal reflexes and recruitment curves between the forelimbs and hind Received 23 October 2006; accepted 19 December 2007. limbs in conscious horses. Am J Vet Res 64, 700–707.

2009 The Authors. Journal compilation 2009 Association of Veterinary Anaesthetists, 36, 9–17 17 05-07-0357r.qxp 4/17/2006 2:59 PM Page 762

Evaluation of administration of isoflurane at approximately the minimum alveolar concentration on depression of a nociceptive withdrawal reflex evoked by transcutaneous electrical stimulation in ponies

Claudia Spadavecchia, DVM, PhD; Olivier Levionnois, DVM; Peter W. Kronen, DVM; Massimo Leandri, DM, PhD; Luciano Spadavecchia, DP, PhD; Urs Schatzmann, DVM, PhD

Objective⎯To investigate effects of isoflurane at ABBREVIATIONS approximately the minimum alveolar concentration MAC Minimum alveolar concentration (MAC) on the nociceptive withdrawal reflex (NWR) of NWR Nociceptive withdrawal reflex the forelimb of ponies as a method for quantifying iMAC Individual MAC anesthetic potency. EMG Electromyographic Animals⎯7 healthy adult Shetland ponies. RMS Root mean square Procedure⎯Individual MAC (iMAC) for isoflurane was IQR Interquartile range determined for each pony. Then, effects of isoflurane administered at 0.85, 0.95, and 1.05 iMAC on the NWR were assessed. At each concentration, the NWR thresh- understood. Analysis of results of studies1-5 in which old was defined electromyographically for the common investigators used the MAC technique of evaluating digital extensor and deltoid muscles by stimulating the anesthetic potency through observation of gross pur- digital nerve; additional electrical stimulations (3, 5, 10, 20, poseful movement in response to noxious stimuli sug- 30, and 40 mA) were delivered, and the evoked activity was recorded and analyzed. After the end of anesthesia, gests that volatile anesthetics block movement largely the NWR threshold was assessed in standing ponies. by means of action at the spinal cord. However, MAC is a nonquantitative, all-or-none measure of motor output Results⎯Mean ± SD MAC of isoflurane was 1.0 ± 0.2%. The NWR thresholds for both muscles increased signifi- that does not allow for investigation of subtle, graded cantly in a concentration-dependent manner during anes- anesthetic effects on depression of motor function. The thesia, whereas they decreased in awake ponies. MAC is assessed by use of a widely used but arbitrary Significantly higher thresholds were found for the deltoid criterion to check for evidence of complex movements muscle, compared with thresholds for the common digi- after supramaximal noxious stimulation, whereas sim- tal extensor muscle, in anesthetized ponies. At each pler and potentially more quantitative variables, such as iMAC tested, amplitudes of the reflex responses from the reflex withdrawal of a stimulated extremity, are usu- both muscles increased as stimulus intensities increased ally neglected. from 3 to 40 mA. A concentration-dependent depression Evidence is lacking to support the contention that of evoked reflexes with reduction in slopes of the stimu- a supramaximal noxious stimulation of a limb system- lus-response functions was detected. atically induces flexion in animals anesthetized with Conclusions and Clinical Relevance⎯Anesthetic- approximately the MAC of anesthetic agents. The induced changes in sensory-motor processing in ponies motor response evoked during MAC assessment con- anesthetized with isoflurane at concentrations of approximately 1.0 MAC can be detected by assessment of sists of 2 patterns (flexion withdrawal vs complex limb NWR. This method will permit comparison of effects of movements) that use differing neural circuits, which inhaled anesthetics or anesthetic combinations on spinal possibly undergo differing modulation for volatile processing in equids. (Am J Vet Res 2006;67:762–769) anesthetics.6 Studies7,8 in human volunteers revealed that the electrically induced NWR of the lower limb disappears he mechanisms by which anesthetics act on senso- at sub-MAC end-tidal concentrations of isoflurane. Try-motor processing to cause immobility are poorly Analysis of results of these studies indicates an early depression of simple reflexes, whereas complex move- Received September 12, 2005. Accepted November 23, 2005. ments can still be elicited. Conversely, rats anes- From the Anesthesiology Section, Department of Clinical Veterinary thetized with approximately the MAC of volatile agents 9,10 Sciences, Vetsuisse Faculty, University of Berne, Langassstrasse 124, still have limb withdrawal reflexes. On the basis of 3012 Berne, Switzerland (C. Spadavecchia, Levionnois, Kronen, the aforementioned evidence, important interspecific Schatzmann); Interuniversitary Center for the Study of Pain differences in the depressive action of anesthetics on Neurophysiology, University of Genova, 16146 Genova, Italy (Leandri); nociceptive reflexes can be hypothesized. and Institute of Biophysics, National Research Council, Via De Marini 6, Limb NWRs evoked by transcutaneous electrical 16149 Genova, Italy (L. Spadavecchia). Dr. C. Spadavecchia’s present 11,12 address is the Norwegian School of Veterinary Science, Department of stimulation have been described in conscious hors- Companion Animal Clinical Science, PO Box 8146, 0033 Oslo, Norway. es. The stimulation and quantification protocols Address correspondence to Dr. C. Spadavecchia. applied in these studies were extremely similar to those

762 AJVR, Vol 67, No. 5, May 2006 05-07-0357r.qxp 4/17/2006 2:59 PM Page 763

described for humans.7 Evaluating the effects of The final iMAC was the mean of 3 crossover points. The approximately the MAC of isoflurane on the NWR in mean of the iMAC values was designated as the group MAC. equids could provide a quantitative method for assess- Electrophysiologic recordings and quantification of ment of anesthetic-induced sensory-motor depression EMG reflex activity—After each pony was anesthetized and in these animals and provide interesting data for inter- positioned in right lateral recumbency, skin over the palmar specific comparisons. lateral digital nerve and over the common digital extensor The objective of the study reported here was to and deltoid muscles of the left forelimb was clipped and cleaned. Pairs of self-adhesive surface electrodes were placed assess the effects of isoflurane administered at concen- f trations of approximately 1.0 MAC on the NWR. The 20 mm apart and used for transcutaneous nerve stimulation and EMG recordings.g Resistance of stimulation electrodes working hypothesis was that isoflurane would begin to had to be < 2 KΩ. The ground electrode was placed on the depress the NWR at the same concentrations that inhib- dorsum of each pony at a location immediately caudal to the ited gross purposeful movements of pain avoidance. point of the shoulders. Stimulations and recordings were performed by use of a computerized system, as described else- Materials and Methods where.12 The EMG signals were amplified with an overall gain Animals—Experiments were conducted on 7 gelding of 5,000 and band-pass filtered (7 to 200 Hz; first-order Shetland ponies. The ponies were 4 years old with a mean ± active filters with a slope of 6 dB/octave). Signals were sub- SD body weight of 121 ± 25 kg. Ponies were judged to be sequently passed through an analogue-to-digital converter healthy on the basis of results of physical, biochemical, and and stored on a computer for additional processing. hematologic examinations. Food was withheld from the Each stimulus consisted of train-of-five, 1-millisecond, ponies for 24 hours before the experiments, but they had ad constant-current, square-wave pulses delivered at a fre- libitum access to water. Two years before the study reported quency of 200 Hz. This configuration has been used for here was conducted, the left carotid artery of each pony was studies11,14-17 conducted in conscious humans and horses. surgically translocated to a subcutaneous position. The The final stage of the stimulation was provided by a battery- Committee for Animal Experimentation, County of Berne, powered, optoisolated, constant-current stimulator with a Switzerland, approved the study, which was part of a larger maximum supply voltage of 100 V. investigation on determination of isoflurane MAC in ponies. For each pony, 3 series of stimulations were performed for isoflurane administered at concentrations of approxi- Induction and monitoring of anesthesia—Oxygen was mately 1.0 MAC (2 concentrations less than the isoflurane administered to each pony. Anesthesia was then induced by MAC [0.85 and 0.95 iMAC] and 1 concentration greater than administration of isofluranea in oxygen via a face mask by use the isoflurane MAC [1.05 iMAC] in end-tidal concentration of a conventional circle anesthetic system.b The isoflurane increments of 0.1%), always after an equilibration period of vaporizer setting was 1% for 1 minute and then increased to at least 30 minutes. Reflex EMG responses to stimulations 3% until endotracheal intubation could be performed. For were recorded for the common digital extensor and deltoid the first 30 minutes, the end-tidal concentration of isoflurane muscles of the ipsilateral limb. Evidence of a reflex move- was set at approximately 1.3%, which is the established MAC ment in response to each electrical stimulation was assessed for equids,13 to permit instrumentation. Mechanical ventila- by visual observation, and responses were recorded. tion (ie, intermittent positive-pressure ventilation) was start- To record and quantify EMG reflex activity in response ed, with initial settings of 8 breaths/min and a tidal volume to stimulation, the EMG recording was conducted from 100 of 12 mL/kg. milliseconds before until 400 milliseconds after a stimulus, Catheters were inserted into a saphenous vein and which resulted in a total recording time of 500 milliseconds carotid artery; these catheters were used for administration of with 512 sample points (sampling frequency, 1,024 Hz). To lactated Ringer’s solution and monitoring of arterial blood be considered a reflex response, the EMG burst following pressure. Esophageal body temperature, ECG, pulse oxime- stimulation had to be at least 3 times the amplitude of the try, arterial blood pressure, inspired and end-tidal oxygen background activity with a duration of at least 10 millisec- concentration, end-tidal carbon dioxide concentration, and onds within the period from 20 to 70 milliseconds after stim- end-tidal anesthetic concentration were continuously moni- ulation onset. tored by use of a calibrated unit.c Mean arterial pressure was At each MAC, the lowest stimulation intensity able to maintained at > 70 mm Hg by IV administration of dobuta- evoke 2 consecutive EMG reflex responses was defined as the mine.d Arterial blood gas analysis was performed after inser- reflex threshold. Intensity of the current was initially set at tion of the catheter into the carotid artery, and the analysis 3 mA and increased in increments of 1 mA until a reflex was then repeated at intervals of 1 hour. response could be detected for each muscle. Additional elec- MAC measurement—After instrumentation, the end- trical stimulations at 3, 5, 10, 20, 30, and 40 mA were deliv- tidal isoflurane concentration was set at 0.9% and maintained ered in ascending order at 1-minute intervals to assess the constant for at least 30 minutes. Thereafter, the iMAC for intensity-response function. Finally, 60 consecutive stimuli each pony was determined by applying a supramaximal elec- were administered at a fixed intensity of 40 mA and frequen- trical stimulatione (90 V and 5 Hz) on the oral mucous mem- cy of 1 Hz. The EMG responses obtained were averaged and branes, as described elsewhere.13 Stimulation was applied for quantified. 60 seconds or until gross purposeful movement was Latency of the reflex response was defined as the amount observed. Lifting of the head or limb movement was inter- of time that elapsed between the onset of the stimulus and preted as a positive response, whereas tonic extensions of the onset of the EMG reflex (deflection from baseline). To quan- limbs or neck were interpreted as a negative response. tify the muscular response, the RMS value for reflex ampli- Depending on the response, the anesthetic concentration was tude was calculated for the period from 20 to 70 milliseconds increased or decreased by 0.1% end-tidal concentration. After after stimulation. The background EMG amplitude was calcu- an equilibration period of 30 minutes, electrical stimulation lated as the RMS amplitude during the 100-millisecond inter- was again applied. This process was continued until anes- val before stimulation. To minimize the influence of possible thetic concentrations were detected that barely permitted and variability among ponies, the relative amplitude of the reflex barely prevented, respectively, purposeful movement. The in the period from 20 to 70 milliseconds after stimulation was mean of these concentrations constituted a crossover point. calculated as the ratio between the RMS amplitude detected

AJVR, Vol 67, No. 5, May 2006 763 05-07-0357r.qxp 4/17/2006 2:59 PM Page 764

during such periods and the RMS of the background EMG in Reflex responses at threshold stimulation inten- the 100-millisecond period before stimulation. During record- sities—Reflex threshold intensities after anesthesia ings performed as part of preliminary experiments, it was were always significantly (P < 0.001) less than thresh- observed that a burst of EMG reflex activity in response to old intensities during anesthesia for both muscles electrical nerve stimulation could appear within the period (Figure 1). Median reflex thresholds in awake ponies from 20 to 70 milliseconds after stimulation in ponies at a light plane of isoflurane-induced anesthesia. were 3 mA (IQR, 3 to 4 mA) for the common digital extensor muscle and 4.5 mA (IQR, 4.0 to 6.5 mA) for Recovery and reflex recordings in standing ponies— the deltoid muscle. During isoflurane-induced anes- After completion of MAC determination and administration thesia, reflex threshold intensities for both muscles of the series of electrical stimulations, ponies were allowed to increased significantly (P = 0.007 for the common dig- breath pure oxygen until the return of the swallowing reflex ital extensor muscle and P = 0.014 for the deltoid mus- was detected. They were then assisted for recovery. Stimulation and recording electrodes applied during anesthe- cle) in a concentration-dependent manner (Table 1). sia were left in place throughout the recovery period. One Median maximal values of 8 mA (IQR, 7.2 to 9.5 mA) hour after disconnection from the anesthetic circuit, with the for the common digital extensor muscle and 24 mA ponies in standing position and residual ataxia resolved, the (IQR, 19 to 41 mA) for the deltoid muscle were found reflex threshold was again assessed. Intensity of electrical at the highest isoflurane concentration tested (1.05 stimulation was initially set at 1 mA and gradually increased iMAC). Significantly higher reflex threshold intensities in increments of 0.5 mA until a reflex response was detected were found for the deltoid muscle, compared with during the period from 20 to 70 milliseconds after stimula- threshold intensities for the common digital extensor tion, as described previously. muscle, at all tested isoflurane concentrations. Statistical analysis—Nonparametric analysis of data However, no significant differences were found for was chosen on the basis of tests for normal distribution. threshold intensities between the 2 muscles after anes- Group MAC was expressed as mean ± SD. Other results were thesia in awake ponies. During anesthesia, stimula- reported as median and IQR (25% to 75%) values. Relative amplitudes, latencies, and durations of the reflexes at various stimulus intensities and MAC concentrations were analyzed by use of Friedman repeated-measures ANOVA on ranks, with post hoc Tukey tests for multiple comparisons. Values obtained for the 2 muscles were compared by use of the Mann-Whitney rank sum test. Values of P < 0.05 were considered significant. Statistical analyses were performed by use of commercially available software.h

Results Anesthetic period—Mean ± SD isoflurane MAC was 1.0 ± 0.2%, with values of iMACs < 1.2% (corrected for a barometric pressure of 760 mm Hg). Mean total duration of anesthesia was 471 ± 37 minutes. Normocapnia and normotension were maintained throughout anesthesia in all ponies, with mean end- tidal carbon dioxide concentration of 33.9 ± 1.6 Figure 1—Median and IQR NWR thresholds for the common ± digital extensor muscle (black circles) and deltoid muscle (white mm Hg and mean arterial blood pressure of 92.5 6.8 circles) of 7 ponies during anesthesia at 0.85, 0.95, and 1.05 mm Hg. Mean PaCO2 was 37.2 ± 1.5 mm Hg. All ponies iMAC of isoflurane and in standing ponies after recovery from recovered uneventfully from anesthesia. anesthesia (awake).

Table 1—Median (IQR) values of relative amplitude, latency, and duration of the NWR obtained during EMG recordings for the common digital extensor (CDE) and deltoid muscles in 7 ponies when stimulated at NWR threshold intensities during anesthesia induced by administration of isoflurane at concentrations of 0.85, 0.95, and 1.05 iMAC and in standing ponies after recovery from anesthesia (awake).

Variable Muscle 0.85 iMAC 0.95 iMAC 1.05 iMAC Awake P value* NWR threshold intensity (mA) CDE 6 (5.0–8.0)a.A 8 (7.2–9.5)b.A 8 (7.2–9.5)b,A 3 (3.0–4.0)c 0.001 Deltoid 15 (8.5–19.5)a,B 18 (11.2–21.5)a,B 24 (19.0–41.2)b,B 4.5 (4.0–6.5)c 0.001 Relative amplitude† CDE 4.4 (3.4–7.3) 8.4 (4.5–9.2) 7.7 (5.1–8.6) 6.4 (3.4–6.5) 0.400 Deltoid 4.3 (3.7–5.3) 4.4 (3.4–11) 3.8 (2.4–11.9) 12.7 (3.6–28.6) 0.610 NWR latency (ms) CDE 20 (19.2–21.5)A 20 (19.2–21.7)A 20 (19.2–21.7)A 20 (17.7–21.7) 0.600 Deltoid 33 (31.5–35.5)a,B 32 (27.2–33)a,B 33 (31.2–36.0)a,B 19 (16.2–32.2)b 0.008 NWR duration (ms) CDE 25 (20.5–25.0) 22 (20.2–30.2) 23 (20.5–31.0) 24 (20.7–32.5) 0.310 Deltoid 25 (21.7–28.7)a 28 (24.2–30.5)a 25 (21.2–25.7)a 38 (30.5–45)b 0.005 *Values represent results for the Friedman test. †Ratio between the RMS amplitude of the EMG activity in the period 20 to 70 milliseconds after stimulus and the RMS amplitude of the EMG activity in the 100-millisecond interval before stimulus. a,bWithin a row, values with different superscript letters differ significantly (P 0.05; Tukey test for multiple comparisons). A,BWithin a variable within a column, values with different superscript letters differ significantly (P 0.05; Mann-Whitney rank sum test).

764 AJVR, Vol 67, No. 5, May 2006 05-07-0357r.qxp 4/17/2006 2:59 PM Page 765

tions at intensities that would barely evoke an EMG reflex response from the common digital extensor muscle were not accompanied by any visible movement of the stimulated limb, whereas stimulations that evoked reflexes from the deltoid muscle were always accompanied by a visible flexion-protraction movement of the stimulated limb. In 2 ponies at 1.05 iMAC, the maximal stimulation intensity of 44 mA was not sufficient to evoke a reflex response and the reflex threshold intensities could not be determined. At threshold intensities, relative amplitudes of the reflex responses were similar for both muscles during and after anesthesia and no concentration-related Figure 2—Representative NWR threshold recordings obtained for the deltoid muscle of a pony during anesthesia at 0.85, 0.95, and changes were observed (Figure 2). Latency of the reflex 1.05 iMAC of isoflurane and after anesthesia (awake). Total recording responses remained stable and did not change signifi- time was 500 milliseconds. The stimulus onset is indicated (arrow). cantly (P = 0.600) for the common digital extensor muscle, with a median overall value of 20 milliseconds (IQR, 19 to 22 milliseconds). Conversely, latency of the reflex response for the deltoid muscle had significant changes, decreasing from a median value of 33 milliseconds (IQR, 31 to 36 milliseconds) during anesthesia to 19 milliseconds (IQR, 16 to 32 milliseconds) after anesthesia. In awake ponies, no significant difference was found for latency of the reflex response between the 2 muscles. Duration of the reflex response did not change during and after anesthesia for the common digital extensor muscle, with a median overall value of 23 milliseconds (IQR, 20 to 25 milliseconds). Conversely, duration of the reflex response for the deltoid muscle increased significantly from a median value of 25 milliseconds (IQR, 22 to 29 milliseconds) during anesthesia to a median value of 38 milliseconds (IQR, 30 to 45 milliseconds) in awake ponies (Table 1). Reflex responses to stimulations of increasing intensity—During anesthesia and at each MAC tested, amplitudes of the reflex responses from both muscles increased significantly (P < 0.001) as stimulus intensity increased from 3 to 40 mA (Figure 3). Reflex responses were consistently recorded for the common digital extensor muscle but not for the deltoid muscle when stimulations of 5 mA were applied (Table 2). The reflex responses usually consisted of a burst of EMG activity during the period from 20 to 70 milliseconds after stimulation. Graded Figure 3—Median and IQR relative amplitudes of EMG activity for the common digital electrical-evoked reflexes during extensor muscle (A) and deltoid muscle (B) of 7 ponies calculated for the period 20 to 70 isoflurane-induced anesthesia were milliseconds after stimulus with 3, 5, 10, 20, 30, and 40 mA during anesthesia induced by administration of isoflurane concentrations of 0.85 iMAC (circles and solid line), 0.95 iMAC detected at 0.85, 0.95, and 1.05 iMAC (squares and dashed line), and 1.05 iMAC (inverted triangles and dotted line). Relative (Figure 4). Between 0.85 and 1.05 amplitude is the ratio between the RMS amplitude of the EMG activity in the period 20 iMAC, there was concentration-depen- to 70 milliseconds after stimulus and the RMS amplitude of the EMG activity in the 100- millisecond interval before stimulus. Notice that the scale of the y-axis differs between dent depression of electrically evoked panels. The reflex threshold (thin horizontal dashed line) is indicated for each muscle. reflexes with a reduction in the slopes

AJVR, Vol 67, No. 5, May 2006 765 05-07-0357r.qxp 4/17/2006 2:59 PM Page 766

Table 2—Median (IQR) values for relative amplitude, latency, and duration of the NWR obtained during EMG recordings for the CDE and deltoid muscles in 7 ponies when stimulated at 3, 5, 10, 20, 30, and 40 mA during anesthesia achieved by administration of isoflurane at concentrations of 0.85, 0.95, and 1.05 iMAC.

Variable Muscle iMAC 3 mA 5 mA 10 mA 20 mA 30 mA 40 mA P value* Relative Deltoid 0.85 1.1 (0.9–1.2) 1.2 (0.9–1.4) 1.7 (1.2–3.0) 10.9 (3.1–45.9)A 57.7 (17.3–89.0)A 43.7 (19.7–74.0)A 0.001 amplitude† 0.95 1.1 (0.8–1.2) 1.1 (0.8–1.1) 1.6 (1.4–15.6) 2.9 (2.0–48.4)A,B 9.1 (3.2–53.3)A 31.0 (4.0–50.9)A,B 0.001 1.05 1.1 (1.0–1.5) 0.9 (0.6–1.1) 1.0 (0.8–1.5) 1.7 (1.5–3.7)B 3.1 (1.8–4.0)b 13.0 (3.8–21.0)B 0.001

CDE 0.85 1.3 (1.0–1.7 4.4 (1.4–7.3) 9.3 (5.1–13.6) 21.1 (10.0–28.6) 21.9 (13.8–28.9)A 25.0 (12.5–37.8) 0.001 0.95 1.1 (1.0–1.4) 3.2 (2.1–4.1) 8.9 (5.1–14.1) 9.2 (5.4–19.6) 11.9 (6.0–20.1)B 13.3 (10.4–16.9) 0.001 1.05 1.3 (1.0–1.4) 2.4 (1.1–4.6) 8.4 (3.7–15.1) 7.8 (6.7–17.0) 7.2 (5.4–23.3)B 8.2 (6.9–31.5) 0.001

NWR latency Deltoid 0.85 NN NN 33 (27–37) 22 (21–32) 21 (21–25)A 21 (21–26)A 0.430 (ms) 0.95 NN NN 31 (27–35) 26 (23–31) 24 (22–28)B 23 (21–28)A 0.370 1.05 NN NN 28 (27–30) 31 (28–33) 33 (29–38)C 31 (29–35)B 0.210

CDE 0.85 NN 23 (20–23) 21 (20–22)A 20 (19–22)A,B 20 (19–21)A,B 19 (19–21)A 0.240 0.95 NN 22 (19–23) 20 (19–23)A 19 (19–21)A 19 (19–21)A 19 (19–20)A 0.010 1.05 NN 24 (21–27) 23 (21–25)B 21 (21–22)B 21 (21–22)B 21 (21–22)B 0.017

NWR duration Deltoid 0.85 NN NN 24 (21–31) 34 (23–38) 40 (36–42)A 41 (38–45)A 0.140 (ms) 0.95 NN NN 29 (28–30) 34 (25–40) 36 (27–38)A 41 (31–43)A 0.370 1.05 NN NN 20 (14–26) 25 (22–26) 20 (18–26)B 22 (20–28)B 0.950

CDE 0.85 NN 23 (19–24) 24 (21–32) 29 (25–37) 35 (32–38) 33 (32–35) 0.008 0.95 NN 20 (11–24) 23 (20–27) 25 (22–38) 32 (28–39) 32 (26–40) 0.001 1.05 NN 20 (16–25) 22 (15–30) 23 (22–34) 29 (23–35) 29 (22–34) 0.150 A,BWithin a variable within a column, values with different superscript letters differ significantly (P 0.05; Mann-Whitney rank sum test). NN = No NWR evoked. See Table 1 for remainder of key.

extensor muscle. Duration of the reflex responses tended to increase with increasing stimulation intensities, with significant differences between responses for the common digital extensor muscle at 0.85 and 0.95 iMAC. At 1.05 iMAC, duration of the reflex responses remained stable regardless of the stimulation intensity. Averaged reflex responses—The averaged reflex responses to 60 stimuli administered at 40 mA and 1 Hz during isoflurane-induced anesthesia permitted visual interpretation of 2 reflex compo- Figure 4—Representative EMG recordings obtained from the deltoid muscle of a nents with differing latencies. Median pony during anesthesia at isoflurane concentrations of 0.85, 0.95, and 1.05 iMAC after stimulus with 3, 5, 10, 20, 30, and 40 mA. The onset of stimulus is indicated latencies of the 2 reflex components were for each iMAC value (vertical dotted lines). 19 milliseconds (IQR, 17 to 20 milliseconds) and 39 milliseconds (IQR, 39 to 40 of stimulus-response functions. For the common digi- milliseconds) for the common digital extensor muscle tal extensor muscle, amplitudes of the reflex response and 19 milliseconds (IQR, 17 to 21 milliseconds) and decreased significantly for increasing anesthetic con- 36 milliseconds (IQR, 36 to 40 millisecond) for the centrations when stimuli were administered at 30 mA. deltoid muscle. Averaged reflex responses appeared to Similarly, a concentration-dependent decrease was be unaffected by isoflurane concentration because observed for stimulations at 20, 30, and 40 mA for the there was no change in latency, amplitude, or duration. deltoid muscle. Significant differences were always Conversely, purposeful movement during the series of detected between the lowest and the highest isoflu- 60 stimuli was a concentration-dependent event. At rane concentrations. 0.85 iMAC, 4 of 5 ponies had purposeful movement In general, latencies of the reflex responses during (the other 2 ponies were judged to be at too light of a isoflurane-induced anesthesia decreased with increas- plane of anesthesia to tolerate stimulation), whereas at ing stimulation intensities (Table 2). Significance of 0.95 iMAC, 2 ponies had purposeful movements and 5 this pattern was detected at each isoflurane concentra- did not. None of the ponies had purposeful movement tion for the reflex response of the common digital at 1.05 iMAC. In contrast, reflex limb withdrawals in

766 AJVR, Vol 67, No. 5, May 2006 05-07-0357r.qxp 4/17/2006 2:59 PM Page 767

response to stimulation were always evident, begin- The concentration-dependent increase of the ning with the second or third stimulus. NWR threshold was a muscle-specific phenomenon, with lowest values found for the common digital Discussion extensor muscle (a muscle of the forelimb), compared The study reported here revealed that subtle anes- with values for the deltoid muscle (a muscle of the thetic-induced changes in sensory-motor processing in shoulder). Visible flexion of the stimulated limb ponies anesthetized with isoflurane at concentrations accompanied NWRs of the deltoid muscle, whereas of approximately 1.0 MAC can be detected by assess- limb flexion was mostly lacking when NWRs were ment of NWR. In fact, in ponies anesthetized at a light recorded for the common digital extensor muscle. The plane of anesthesia, NWR thresholds increased with common digital extensor muscle consistently was acti- increasing isoflurane concentrations. Furthermore, vated at lower stimulation intensities, compared with reflex activities in response to single electrical stimuli values for the deltoid muscle. Reflex activity of the were largely depressed at isoflurane concentrations common digital extensor muscle had only a moderate able to prevent purposeful movements in response to increase in threshold, even for the highest isoflurane supramaximal noxious stimulations, which indicated concentrations, which indicated a low sensitivity to the parallel inhibition of the 2 evoked motor patterns depressant effects of the anesthetic. Muscle-specific (flexion withdrawal and complex limb movements) in NWR thresholds were not observed in conscious hors- ponies. es in another study,11 in which other muscles of the Although isoflurane abolishes the electrically forelimb were examined, nor were they observed after induced NWR in humans at concentrations substan- recovery from anesthesia in the ponies of the study tially < 1.0 MAC,8 it did not do so in our ponies, even reported here in which results suggested that isoflu- at approximately MAC, which suggested species-spe- rane, lateral recumbency, or both must account for the cific anesthetic modulation of spinal withdrawal reflex- difference in patterns of reflexive muscle activation es despite homogeneous interspecific MAC values.18 At during anesthesia. all tested MAC multiples, amplitude of the reflex The reflex response recorded from both forelimb response increased with stimulus intensity, whereas the muscles was detected during the period from 20 to 70 slope of the stimulus-response function was reduced milliseconds after stimulus. The early part of the reflex with increasing isoflurane concentrations. Similarly, was probably attributable to activation of Aβ fibers. concentrations of isoflurane at approximately 1.0 MAC With a mean distance of 85 cm between the stimula- decreased the force of reflexive limb withdrawals in tion electrodes and dorsal point of the shoulders, and response to noxious thermal stimuli in a concentra- considering a peripheral conduction velocity of 75 m/s tion-dependent manner in rats, with the largest reduc- for ponies,23 an afferent time of 11 milliseconds was tion between 0.9 and 1.1 MAC.9 It can be concluded expected. Adding an approximate central delay of 5 that abolishment of gross purposeful movement is par- milliseconds and neglecting the amount of time need- alleled by inhibition of withdrawal reflexes. ed for the motor component of the reflex,15 the Determination of MAC depends on detection of observed minimal latency of 16 milliseconds appears complex gross purposeful movements, which typically reasonable. The late part of the reflex, which always involve the limbs and head. These complex move- terminated before 70 milliseconds after the stimulus, ments are probably dependent on the central pattern could have reflected activation of A∆ fibers, with a con- generator, which is the neural network involved in production velocity between 15 and 35 m/s, which corre- cessing this behavior.6 The central pattern generator sponds to an afferent time between 25 and 55 millisec- receives afferent inputs from the spinal cord and brain onds. In the study reported here, it was not possible to and acts to initiate and terminate complex move- separate reflex components originating from Aβ fibers ments.19 Isoflurane-induced disruption of activity and from those originating from A∆ fibers when examining coordination of the central pattern generator has been a single reflex. When 60 consecutive stimuli at the described20 in spinal cords isolated from lampreys. intensity of 40 mA were applied during 1 minute and Unfortunately, only scarce data are available about the reflex responses were averaged, 2 peaks of differing interspecies differences in organization and control of latencies appeared for both muscles, which confirmed the central pattern generator21 and its sensitivity to that the 2 expected components exist in ponies (simi- anesthetics.6 lar to results reported for humans24 and horses11,12) but Withdrawal reflexes in awake animals incorporate probably have more intrasubject variation. flexion of the stimulated limb with extension of the The method applied to determine MAC for each contralateral limb.22 It is not currently known whether pony in the study reported here is considered the stan- this reflex remains intact during inhalant anesthesia.6 dard for equids and was described for the first time in The study reported here provides evidence of reflex 1977.13 It consists of supramaximal electrical stimula- activity in ponies anesthetized with isoflurane concen- tion of the oral mucous membranes; the stimulation is trations of approximately 1.0 MAC that is consistent applied at a fixed voltage and frequency during 1 with the withdrawal reflex pattern described in con- minute or until a gross purposeful movement is scious horses.12 When single stimuli of increasing observed. Repetition of the stimulus causes a temporal intensity were applied during anesthesia, latency of the summation of the nociceptive input, which influences NWR decreased but duration of the NWR increased, MAC values of the volatile agent.25 A similar temporal confirming that the reflex undergoes intensity-depen- summation effect was evident when 60 transcutaneous dent modulation. stimuli at 40 mA were applied for 1 minute over the

AJVR, Vol 67, No. 5, May 2006 767 05-07-0357r.qxp 4/17/2006 2:59 PM Page 768

digital nerve and was responsible for the purposeful f. Neuroline 7 00 02-J, Medicotest, Olstykke, Denmark. movements observed during the stimulation series in g. Synapse, Ambu A/S, Ballerup, Denmark. ponies receiving isoflurane at concentrations < 1.0 h. Sigma Stat, version 3.10 for Windows, Systat Software Inc, Point Richmond, Calif. MAC. As expected, none of the ponies had purposeful movements when anesthetized at 1.05 iMAC, whereas References NWRs still were facilitated as a result of stimulus rep- 1. Antognini JF, Carstens E, Buzin V. Isoflurane depresses etition. The MAC values determined for each of the motoneuron excitability by a direct spinal action: an F-wave study. Shetland ponies of the study were less than the MAC Anesth Analg 1999;88:681–685. values reported for horses.13 Similarly, lower MAC val- 2. Antognini JF, Carstens E, Tabo E, et al. Effect of differential ues for ponies, compared with MAC values for horses, delivery of isoflurane to head and torso on lumbar dorsal horn activ- have been found for desflurane,26 and large differences ity. Anesthesiology 1998;88:1055–1061. in halothane MAC have been reported for horses,13,27 3. Antognini JF, Carstens E, Atherley R. Does the immobiliz- ing effect of thiopental in brain exceed that of halothane? even when the same standard stimulation method was Anesthesiology 2002;96:980–986. applied. Among the possible causes for MAC varia- 4. Rampil IJ. Anesthetic potency is not altered after hypothermic tions, it is likely that variation in sensitivity among spinal cord transection in rats. Anesthesiology 1994;80:606–610. subjects, genetic influences,28 failure of the noxious 5. Quasha AL, Eger EI II, Tinker JH. Determination and appli- stimulus to generate a supramaximal effect,29,30 and the cations of MAC. Anesthesiology 1980;53:315–334. point in time that the stimuli were applied in relation- 6. Antognini JF, Wang XW, Carstens E. Quantitative and qual- 31 itative effects of isoflurane on movement occurring after noxious ship to the time of induction of anesthesia played a stimulation. Anesthesiology 1999;91:1064–1071. role in these effects. 7. Petersen-Felix S, Arendt-Nielsen L, Bak P, et al. Analgesic Quantification of NWR during anesthesia provides effect in humans of subanaesthetic isoflurane concentrations evaluat- information that sensory-motor depression as a whole ed by experimentally induced pain. Br J Anaesth 1995;75:55–60. is evident at various isoflurane concentrations but does 8. Petersen-Felix S, Arendt-Nielsen L, Bak P, et al. The effects not permit differentiation of the effects of anesthesia on of isoflurane on repeated nociceptive stimuli (central temporal summation). Pain 1996;64:277–281. afferent and efferent reflex pathways. An invasive 9. Jinks SL, Martin JT, Carstens E, et al. Peri-MAC depression approach in laboratory animals has been used to exam- of a nociceptive withdrawal reflex is accompanied by reduced dorsal ine the selective effects of inhaled anesthetics on neu- horn activity with halothane but not isoflurane. Anesthesiology rons in the dorsal horn32 and motoneurons.10 2003;98:1128–1138. Depression of neurons in the dorsal horn of rats is 10. You HJ, Colpaert FC, Arendt-Nielsen L. Nociceptive spinal mainly at concentrations of isoflurane < 1.0 MAC.9 withdrawal reflexes but not spinal dorsal horn wide-dynamic range neuron activities are specifically inhibited by halothane anaesthesia Analysis of such findings suggests that the immobiliz- in spinalized rats. Eur J Neurosci 2005;22:354–360. ing action of isoflurane is not mediated by depression 11. Spadavecchia C, Spadavecchia L, Andersen OK, et al. of nociceptive transmission through the dorsal horn, Quantitative assessment of nociception in horses by use of the noci- but is mediated instead by depression at a more ventral ceptive withdrawal reflex evoked by transcutaneous electrical stimula- location. Isoflurane modulation of the F wave and H tion. Am J Vet Res 2002;63:1551–1556. reflex as noninvasive measures of motoneuron 12. Spadavecchia C, Arendt-Nielsen L, Andersen OK, et al. excitability has been investigated in humans33,34 and Comparison of nociceptive withdrawal reflexes and recruitment 35 curves between the forelimbs and hind limbs in conscious horses. rats. These methods have revealed that a prevalent Am J Vet Res 2003;64:700–707. inhibition of the efferent nociceptive pathways is prob- 13. Steffey EP, Howland D Jr, Giri S, et al. Enflurane, halothane, ably responsible for isoflurane-induced immobility evi- and isoflurane potency in horses. Am J Vet Res 1977;38:1037–1039. dent at concentrations of approximately 1.0 MAC. 14. Arendt-Nielsen L, Sonnenborg FA, Andersen OK. In addition to theoretic considerations on the Facilitation of the withdrawal reflex by repeated transcutaneous electrical stimulation: an experimental study on central integration in mechanisms of action of isoflurane on the motor sys- humans. Eur J Appl Physiol 2000;81:165–173. tem, results of the study reported here provide a prac- 15. Andersen OK, Sonnenborg FA, Arendt-Nielsen L. Modular tical tool to be applied in veterinary anesthesia for organization of human leg withdrawal reflexes elicited by electrical research purposes that will complement studies of stimulation of the foot sole. Muscle Nerve 1999;22:1520–1530. MAC. Analysis of our results indicates that a quanti- 16. Rossi A, Decchi B. Flexibility of lower limb reflex respons- tative noninvasive assessment of sensory-motor es to painful cutaneous stimulation in standing humans: evidence of load-dependent modulation. J Physiol 1994;481:521–532. depression of a simple NWR of a limb at isoflurane 17. Le-Bars D, Willer JC, De-Broucker T. Morphine blocks concentrations of approximately 1.0 MAC is feasible descending pain inhibitory controls in humans. Pain 1992;48:13–20. in ponies. In the future, it will be possible to apply 18. Sani O, Shafer SL. MAC attack? Anesthesiology this method to compare the effects of various inhalant 2003;99:1249–1250. agents or anesthetic combinations on spinal process- 19. Yuste R, MacLean JN, Smith J, et al. The cortex as a central ing at concentrations related to MAC in equines and pattern generator. Nat Rev Neurosci 2005;6:477–483. 20. Jinks SL, Atherley RJ, Dominguez CL, et al. Isoflurane dis- other species, thus improving our knowledge of rupts central pattern generator activity and coordination in the lam- agent-specific and species-specific mechanisms of prey isolated spinal cord. Anesthesiology 2005;103:567–575. anesthetic action. 21. Cazalets JR, Borde M, Clarac F. Localization and organization of the central pattern generator for hindlimb locomotion in a. Isoflo, Abbott AG, Baar, Switzerland. newborn rat. J Neurosci 1995;15:4943–4951. b. Roche electronic respirator 3100, F. Hoffmann-La Roche, Basel, 22. Sherrington CS. Flexion-reflex of the limb, crossed exten- Switzerland. sion-reflex and reflex stepping and standing. J Physiol (Lond) c. S/5 compac, Datex-Ohmeda, Helsinki, Finland. 1910;40:28–121. d. Dobutrex, Eli Lilly & Co, Vernier, Geneva, Switzerland. 23. Blythe LL, Kitchell RL, Holliday TA, et al. Sensory nerve con- e. Grass S88, Grass Instruments, Quincy, Mass. duction velocities in forelimb of ponies. Am J Vet Res 1983;44:1419–1426.

768 AJVR, Vol 67, No. 5, May 2006 05-07-0357r.qxp 4/17/2006 2:59 PM Page 769

24. Willer JC. Comparative study of perceived pain and noci- 30. Polhill SL, Clewlow F, Smith DC. Are changes in the evoked ceptive flexion reflex in man. Pain 1977;3:69–80. electromyogram during anaesthesia without neuromuscular block- 25. Dutton RC, Zhang Y, Stabernack CR, et al. Temporal sum- ing agents caused by failure of supramaximal nerve stimulation? Br mation governs part of the minimum alveolar concentration of J Anaesth 1998;81:902–904. isoflurane anesthesia. Anesthesiology 2003;98:1372–1377. 31. Petersen-Felix S, Zbinden AM, Fischer M, et al. Isoflurane 26. Bettschart-Wolfensberger R, Jaggin-Schmucker N, Lendl C, minimum alveolar concentration decreases during anesthesia and et al. Minimal alveolar concentration of desflurane in combination surgery. Anesthesiology 1993;79:959–965. with an infusion of medetomidine for the anaesthesia of ponies. Vet 32. Antognini JF, Carstens E. Increasing isoflurane from 0.9 to Rec 2001;148:264–267. 1.1 minimum alveolar concentration minimally affects dorsal horn cell 27. Matthews NS, Lindsay SL. Effect of low-dose butorphanol responses to noxious stimulation. Anesthesiology 1999;90:208–214. on halothane minimum alveolar concentration in ponies. Equine Vet 33. Zhou HH, Jin TT, Qin B, et al. Suppression of spinal cord J 1990;22:325–327. motoneuron excitability correlates with surgical immobility during 28. Sonner JM, Gong D, Li J, et al. Mouse strain modestly influ- isoflurane anesthesia. Anesthesiology 1998;88:955–961. ences minimum alveolar anesthetic concentration and convulsivity of 34. Zhou HH, Mehta M, Leis AA. Spinal cord motoneuron inhaled compounds. Anesth Analg 1999;89:1030–1034. excitability during isoflurane and nitrous oxide anesthesia. 29. Zbinden AM, Maggiorini M, Petersen-Felix S, et al. Anesthesiology 1997;86:302–307. Anesthetic depth defined using multiple noxious stimuli during 35. King BS, Rampil IJ. Anesthetic depression of spinal motor isoflurane/oxygen anesthesia. I. Motor reactions. Anesthesiology neurons may contribute to lack of movement in response to noxious 1994;80:253–260. stimuli. Anesthesiology 1994;81:1484–1492.

AJVR, Vol 67, No. 5, May 2006 769 The Veterinary Journal 183 (2010) 337–344

Contents lists available at ScienceDirect

The Veterinary Journal

journal homepage: www.elsevier.com/locate/tvjl

The effects of isoﬂurane minimum alveolar concentration on withdrawal reﬂex activity evoked by repeated transcutaneous electrical stimulation in ponies

Claudia Spadavecchia a,*, Olivier Levionnois a, Peter Kronen a, Ole K. Andersen b a Anaesthesiology Section, Department of Clinical Veterinary Sciences, Vetsuisse Faculty, Bern University, Bern, Switzerland b Centre for Sensory-Motor Interaction, Aalborg University, Aalborg, Denmark article info abstract

Article history: The aim of this study was to quantify the effects of isoflurane at approximately the minimum alveolar Accepted 22 December 2008 concentration (peri-MAC) on the temporal summation (TS) of reflex activity in ponies. TS was evoked by repeated electrical stimulations applied at 5 Hz for 2 s on the digital nerve of the left forelimb of seven ponies. Surface electromyographic activity was recorded from the deltoid and common digital extensor Keywords: muscles. TS thresholds and amplitude of response to stimulations of increasing intensities were assessed Isoflurane during anaesthesia at 0.85, 0.95 and 1.05 times the individual MAC, and after anaesthesia in standing ani- Withdrawal reflex mals. Under isoflurane anaesthesia, TS thresholds increased significantly in a concentration-dependent Electrical stimulus fashion and at each isoflurane MAC, the responses increased significantly for increasing stimulation Electromyography Equine intensities. A concentration-dependent depression of evoked reflexes with a reduction in the slopes of the stimulus–response function was observed for both muscles. The results demonstrated that with this model it is possible to describe and quantify the effects of anaesthetics on spinal sensory-motor processing in ponies. Ó 2009 Elsevier Ltd. All rights reserved.

Introduction of graded anaesthetic induced changes in sensory-motor processing (Jinks et al., 2003). Attenuation of motor responses induced by anaesthetics is an Despite little research, the systematic evaluation of anaesthetic essential component of the anaesthetic state. Recent evidence depressant effects on the simple withdrawal motor pattern follow- has suggested that inhaled anaesthetics, specifically isoflurane, ing noxious stimulation has been shown to provide interesting act on the spinal cord to suppress the movement that occurs in re- information (Jinks et al., 2003). By simultaneously recording the response to a noxious stimulus, mainly by inhibiting ventral horn sponses of single dorsal horn neurons and hind limb withdrawal activity (Zhou et al., 1998; Antognini et al., 1999a, 2000, 2003; force to a graded noxious thermal hind paw stimulation in rats, Antognini and Wang, 1999). an anaesthetic-specific site of action was demonstrated: halothane The standard method for determining the potency of anaes- suppressed reflex movement mainly by depressing dorsal horn thetic drugs, the minimal alveolar concentration (MAC), is based neurons while isoflurane suppressed movements by an action at on effects on the motor system. The motor response evoked by more ventral sites in the spinal cord. noxious afferent input during MAC assessment consists of two dis- Whereas the withdrawal reflex of the limb evoked by single tinct patterns, namely, a simple flexion withdrawal of the stimu- electrical stimulation disappeared at concentrations of isoflurane lated body part, and complex movements involving multiple well below the MAC in human volunteers, the facilitation of the re- body parts, usually described as gross purposeful movements (Qua- flex following repeated electrical stimulations was still observed at sha et al., 1980). These patterns use differing neural circuits, which anaesthetic levels around the MAC (Petersen-Felix et al., 1996). possibly undergo differing modulation by volatile anaesthetics This suggested that isoflurane alone is not adequate for inhibiting (Antognini et al., 1999b). the central sensitisation that might be evoked by surgical stimuli in Traditionally, only the suppression of the gross purposeful humans. Reflex facilitation following repeated stimulation is gen- movement is considered the end-point in MAC determination. erally attributed to temporal summation (TS) of the action poten- The MAC is therefore a non-quantitative all-or-none measure of tials at spinal level and is classically accompanied by an motor output that does not allow for quantitative detailed analysis amplification of subjective pain perception (Arendt-Nielsen et al., 2000). Further experimental evidence indicates that TS plays a basic role in the classical MAC determination process (Dutton et al., * Corresponding author. Tel.: +41 31 631 27 80; fax: +41 31 631 26 20. E-mail address: [email protected] (C. Spadavecchia). 2003), as is demonstrated by an increase in the delay between

1090-0233/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.tvjl.2008.12.011 338 C. Spadavecchia et al. / The Veterinary Journal 183 (2010) 337–344 stimulus onset and motor response observed at inhalant concen- consisted of a train-of-five, 1 ms, constant-current, square-wave pulses delivered at trations approaching MAC. a frequency of 200 Hz. To record and quantify the EMG reflex activity in response to single stimulation, the EMG recording was conducted from 100 ms before until We have previously reported that isoflurane at approximately 400 ms after the stimulus, which resulted in a total recording time of 500 ms with MAC end-tidal concentrations (peri-MAC) depresses withdrawal 512 sample points (sampling frequency, 1024 Hz). To be considered a reflex re- reflexes evoked by single electrical stimulations in ponies in a con- sponse, the EMG burst following stimulation had to be at least three times the centration-dependent way (Spadavecchia et al., 2006), but does not amplitude of the background activity with a duration of at least 10 ms within the abolish it completely as is seen in humans (Petersen-Felix et al., period 20–70 ms after stimulation onset. Considering a mean afferent distance of approximately 85 cm between the stimulating electrodes and the withers, the early 1996). These results suggested that important inter-specific differ- part of the reflex with a latency around 20 ms can probably be attributed to activa- ences in the depressive action of anaesthetics on withdrawal re- tion of A-beta fibres, with a conduction velocity of 75 m/s (Blythe et al., 1983), while flexes may exist. the late part of the reflex, which always terminates <70 ms after the stimulus, could The primary aim of the present study was to assess the effects of reflect activation of A-delta fibres, with a conduction velocity between 15 and 35 m/ s(Gasser and Erlanger, 1927). At each MAC level, the lowest stimulation intensity isoflurane administered at concentrations of approximately one able to evoke two consecutive EMG reflex responses was defined as the reflex MAC on the TS of reflex activity evoked by repeated transcutaneous threshold. Intensity of the current was initially set at 3 mA and increased in incre- electrical stimulations in ponies. A secondary aim was to compare ments of 1 mA until a reflex response could be detected for each muscle. the effects of isoflurane on TS with results previously reported for When the single stimulation series was over, a repeated stimulation series was reflexes evoked by a single electrical stimulation during the same started. Repeated electrical stimuli were applied at a frequency of 5 Hz over 2 s (total of 10 stimuli). Each of these stimuli consisted of a train-of-five pulses, previously experiment (Spadavecchia et al., 2006). defined as single stimulus. The intensity of the current was initially set at 3 mA and increased in increments of 1 mA until the TS threshold could be defined (see below). Materials and methods Next, repeated electrical stimulations at 3, 5, 10, 20, 30, and 40 mA were delivered in ascending order at 1 min intervals to assess the stimulus–response function. Re- The Committee for Animal Experimentation, County of Berne, Switzerland, ap- flex EMG responses to stimulations were recorded for the common digital extensor proved the study, which was part of a larger investigation on the properties of iso- and deltoid muscles of the ipsilateral limb. Evidence of a reflex movement in re- flurane in ponies. sponse to each electrical stimulation was assessed visually. Experiments were conducted on seven gelding Shetland ponies. The ponies To record and quantify the EMG reflex activity in response to repeated stimula- were 4 years old with a mean ± SD bodyweight (BW) of 121 ± 25 kg. Ponies were tion, EMG activity was stored from 500 ms before until 1500 ms after the stimula- judged to be healthy on the basis of physical, biochemical and haematological tion ended, resulting in a total recording time of 4000 ms (sampling frequency, examinations. Food was withheld for 24 h before the experiments, but water was 1 kHz). At each MAC level, the TS threshold was defined as the stimulation intensity allowed ad libitum. Two years previously, the left carotid artery of each pony had able to evoke at least one reflex during the stimulation series as previously defined. been surgically translocated to a subcutaneous position. To quantify the reflex response, the root-mean-square (RMS) amplitude was Induction and monitoring of anaesthesia, as well as individual MAC determina- calculated for the epochs 20–70 ms after each of the 10 stimuli in the stimulus tion have been described in detail elsewhere (Spadavecchia et al., 2006). Briefly, train. Similarly, the presence of late reflex activity was evaluated by calculating anaesthesia was induced with isoflurane (Isoflo, Abbott AG) in oxygen via a face the RMS amplitude for the epochs 70–200 ms. The background EMG amplitude mask by use of a conventional circle anaesthetic system (Roche electronic respirator was calculated as the RMS amplitude during the 500 ms interval before stimulation. 3100, F. Hoffmann-La Roche). Intermittent positive pressure ventilation was per- At each isoflurane concentration and at each stimulus intensity reflexes were as- formed throughout the experiment. Vital parameters and end-tidal anaesthetic sessed by analysing the amplitude of the response to the first stimulus in each train concentrations were continuously monitored using a calibrated unit (S/5 compac, and by calculating the average amplitude of the 10 responses (epochs 20–70 ms Datex-Ohmeda). Mean arterial pressure was maintained at >70 mm Hg by intrave- after stimulus onset). The absolute TS was calculated by subtracting the RMS ampli- nous (IV) administration of dobutamine (Dobutrex, Eli Lilly). tude of the first response from the first and each subsequent response and by sum- After instrumentation, the end-tidal isoflurane concentration was set at 0.9% ming the residuals. Subtracting the initial response from responses to successive and maintained constant for at least 30 min. Thereafter, the MAC for each pony stimuli removes the baseline response and leaves only the net response that is facil- was determined by applying a supra-maximal electrical stimulation (Grass S88, itated or inhibited during the repetitive stimulus train. The stimulus number evok- Grass Instruments) at 90 V and 5 Hz on the oral mucous membranes. Stimulation ing the maximal response within the stimulation series was recorded. Latency and was applied for 60 s or until gross purposeful movement was observed. Lifting of duration of reflex responses to the first and the last stimulus of the series of 10 were the head or limb movement was interpreted as a positive response, whereas tonic determined for each pony at the maximal stimulation intensity applied. Latency of extensions of the limbs or neck were interpreted as a negative response. Depending the reflex response was defined as the amount of time that elapsed between the on- on the response, the anaesthetic concentration was increased or decreased by 0.1% set of the stimulus and onset of the EMG activity burst (deflection from baseline), end-tidal concentration. After an equilibration period of 30 min, electrical stimula- while duration of the reflex response was defined as the time that elapsed between tion was again applied. This process was continued until two anaesthetic concen- the onset and offset of the EMG activity burst in the predefined post-stimulation trations were detected that barely permitted or prevented purposeful movement. interval. The mean of these concentrations constituted a cross-over point. The final individ- After completion of the stimulations at the three increasing isoflurane concen- ual MAC (iMAC) was the mean of three cross-over points. The mean of the MAC val- trations, ponies were allowed to breathe pure oxygen until the swallowing reflex ues for all ponies was designated as the group MAC. had returned. They were then assisted in recovery. Stimulation and recording elec- The skin over the palmar lateral digital nerve and over the common digital trodes applied during anaesthesia were left in place throughout the recovery peri- extensor and deltoid muscles of the left forelimb was clipped and cleaned. Pairs od. One hour after disconnection from the anaesthetic circuit, with the ponies in of self-adhesive surface electrodes were placed 20 mm apart and used for transcu- standing position and residual ataxia resolved, single reflexes and TS thresholds taneous nerve stimulation (Neuroline 7 00 02-J, Medicotest) and for electromyo- were reassessed. The intensity of electrical stimulation was initially set at 1 mA graphic (EMG) recordings (Synapse, Ambu A/S). Resistance of the stimulation and gradually increased in increments of 0.5 mA until a reflex response could be de- electrodes was <2 kX. The ground electrode was placed on the back of each pony tected in the interval 20–70 ms after each stimulus, as described previously. A sche- at a location immediately caudal to the point of the shoulders. Stimulations and matic diagram of the course of the events during and after anaesthesia is shown in recordings were performed by use of a computerised system, as described previ- Fig. 1. ously (Spadavecchia et al., 2004). The final stage of the stimulation was provided by a battery-powered, optoisolated, constant-current device with a maximum voltage of 100 V and a maximal current of 40 mA. The EMG signals were amplified with an overall gain of 5000 and band-pass filtered (7–200 Hz; first-order active filters with 6 dB/octave slope). Signals were subsequently passed through an analogue- to-digital converter and stored on a computer for additional processing. Once instrumented for electrophysiological recordings, each pony was stabi- 1.05iMAC 0.95iMAC lised at end-tidal isoflurane concentrations of 0.85, 0.95 and 1.05 times the iMAC. 0.85iMAC For all ponies the sequence of isoflurane concentrations tested was the same, from the lowest to the highest. At each isoflurane concentration and after at least 30 min equilibration time, two series of electrical stimulations were applied. In all ponies, a MAC determination Reflex recordings series of single stimulations was first performed and a repeated stimulation series was started when the single stimulation series was over. Anaesthesia Recovery Single stimulations for determination of reflex thresholds and reflex recruitment were firstly delivered and detailed methods and the results of this part of Fig. 1. Schematic diagram of the course of events: single stimulations; the experiment have been reported (Spadavecchia et al., 2006). Each single stimulus repeated stimulations. C. Spadavecchia et al. / The Veterinary Journal 183 (2010) 337–344 339

Nonparametric analysis of data was chosen based on tests for normal distribu- manner. Median maximal values of 8 mA (IQR, 7.2–9.5 mA) for the tion and the results were reported as median and inter-quartile range (IQR, 25–75%) common digital extensor muscle and 24 mA (IQR, 19–41 mA) for values. Average amplitude of the responses and the effect of stimulus number on the position of the maximal response were analysed by Friedman repeated-mea- the deltoid muscle were found at the highest isoflurane concentra- sures ANOVA on ranks, with post hoc Tukey tests for multiple comparisons. Re- tion tested (1.05 iMAC) (Fig. 2)(Spadavecchia et al., 2006). sponse characteristics measured from the two muscles, threshold intensities for TS thresholds in awake ponies were always significantly lower single versus repeated stimulations and first response versus the average response (P < 0.001) than thresholds intensities during anaesthesia for both obtained during the stimulation series were compared by Wilcoxon Signed Rank muscles (Fig. 2). Significantly higher TS thresholds were found for Test. Values of P < 0.05 were considered significant. Statistical analyses were performed using commercially available software (Sigma Stat, version 3.10 for Win- the deltoid muscle, compared to the common digital extensor mus- dows, Systat Software). cle, at all tested isoflurane concentrations (0.85 iMAC, P = 0.018; 0.95 iMAC, P = 0.026; 1.05 iMAC, P = 0.021) and in awake ponies Results (P = 0.05). Median TS thresholds in awake ponies were 2.6 mA (IQR, 1.7–2.6 mA) for the common digital extensor muscle and The mean ± SD isoflurane MAC was 1.0 ± 0.2%, with values of 3 mA (IQR, 1.9–3.4 mA) for the deltoid muscle. During isoflurane MACs for each pony <1.2% (corrected for barometric pressure of anaesthesia, reflex thresholds for both muscles increased signifi- 760 mm Hg). The mean total duration of anaesthesia was cantly (P < 0.001 for the common digital extensor muscle and 471 ± 37 min. Normocapnia and normotension were maintained P = 0.008 for the deltoid muscle) in a concentration-dependent throughout anaesthesia in all ponies that recovered uneventfully manner. Median maximal values of 7 mA (IQR, 5.1–12.1 mA) for from anaesthesia. Details of the anaesthesia and MAC determina- the common digital extensor muscle and 15 mA (IQR, 10.9– tion have been given in our paper reporting the results of the single 21.3 mA) for the deltoid muscle were found at the highest isoflu- stimulation series (Spadavecchia et al., 2006). rane concentration tested (1.05 iMAC). Median single reflex thresholds in awake ponies were 3 mA When reflex thresholds to single stimuli and reflex thresholds (IQR, 3–4 mA) for the common digital extensor muscle and to repeated stimuli were compared, lower thresholds were found 4.5 mA (IQR, 4.0–6.5 mA) for the deltoid muscle. During isoflurane for repeated stimuli compared to single stimuli. Significant differ- anaesthesia, reflex threshold intensities for both muscles increased ences were detected in awake ponies (P = 0.014) and at 0.85 iMAC significantly (P = 0.007 for the common digital extensor muscle (P = 0.003) for the common digital extensor muscle and in awake and P = 0.014 for the deltoid muscle) in a concentration-dependent ponies (P = 0.002), at 0.85 iMAC (P = 0.03) and 1.05 iMAC (P = 0.05) for the deltoid muscle (Fig. 2). Using increasing stimulation intensities up to 40 mA, reflexes could be elicited from both the extensor and the deltoid muscle 25 in all ponies at each isoflurane MAC level. At the maximal stimula- Deltoid * tion intensity applied, the latency of the reflex response to the first 20 stimulus was significantly longer than the latency of the response to the last stimulus for the muscle deltoid, at the three isoflurane Repeated stimuli MAC levels (0.85 iMAC, P = 0.016; 0.95 iMAC, P = 0.03; 1.05 iMAC, Single stimuus 15 P = 0.016) and in awake ponies (P = 0.047). Median values of 24 ms (22.5–27 ms) for the first response and 17.5 ms (16– * 19 ms) for the last response were found for the pooled data. 10 For the muscle common digital extensor no differences in latency were found between responses to the first and the last * 5 stimulus. The observed median latency for pooled data was Intensity threshold [mA] 19 ms (18–21 ms). For both muscles, no differences in reflex duration were found 0 between responses to the first and the last stimulus. Median values ake AC C C Aw 0.85 iM 0.95 iMA 1.05 iMA

25 0 ms 500 ms 2500 ms 4000 ms CDE 20 3 mA

5 mA 15 10 mA * 10 * 20 mA

5 30 mA Threshold intensity [mA]

40 mA 0 500 µV ake C C C Aw 0.85 iMA 0.95 iMA 1.05 i MA 400 ms Fig. 2. Boxplots showing reflex threshold intensities for the deltoid and common digital extensor (CDE) muscle when single and repeated stimuli were applied to the Fig. 3. Representative recordings from the deltoid muscle of a pony anaesthetised left digital nerve in seven ponies, anaesthetised with isoflurane at 0.85, 0.95 and at 0.85 MAC when repeated stimulations (10 stimuli, 5 Hz) were given at 3, 5, 10, 1.05 the individual MAC and after recovery from anaesthesia (awake).*P < 0.05 20, 30 and 40 mA. Total recording time was 4 s. Stimulus onset is indicated by an Lower reflex thresholds for repeated stimuli compared to repeated stimuli. arrow. 340 C. Spadavecchia et al. / The Veterinary Journal 183 (2010) 337–344

0.85 iMAC 0.95 iMAC 1.05 iMAC

3 mA

5 mA

10 mA

20 mA

30 mA

40 mA V µ

500 400ms

Fig. 4. Representative EMG recordings obtained from the deltoid muscle of a pony during anaesthesia at isoﬂurane concentrations of 0.85, 0.95 and 1.05 iMAC after stimulation with 3, 5, 10, 20, 30 and 40 mA.

for pooled data were 38 ms (30–42 ms) and 34 ms (25–40) for the 100 deltoid and the common digital extensor muscles, respectively. During anaesthesia, and at each MAC level, the average reflex Deltoid 0.85 iMAC amplitude increased significantly with stimulation intensities by 80 0.95 iMAC 1.05 iMAC 3–40 mA for both muscles (deltoid muscle: 0.85 iMAC P = 0.006; µ 0.95 iMAC P < 0.001; 1.05 iMAC P < 0.001; extensor muscle: 0.85 iMAC P = 0.004; 0.95 iMAC P < 0.001; 1.05 iMAC P < 0.001). Fig. 3 60 # shows an example of a progressive increase in the reflex amplitude when increasing stimulation intensities are given at 0.85 iMAC. Be- 40 tween 0.85 iMAC and 1.05 iMAC, there was a concentration-depen- # dent reflex depression with a reduction in the slopes of the stimulus–response functions for both muscles (Figs. 4 and 5). The 20 * # highest isoflurane MAC level suppressed the average reflex amplitude compared to the lowest MAC levels in both muscles (P < 0.001). 0 When the average reflex amplitude was compared to the ampli- 3 5 10 20 30 40 tude of the initial response, no significant differences were de- Stimulus intensity [mA] tected. Absolute TS increased significantly with increasing stimulus intensity only for the common digital extensor at 0.85 100 iMAC isoflurane (P < 0.001) (Fig. 6). The position of the maximal response within the stimulation CDE series was influenced by the intensity of stimulation, but not by 80 isoflurane concentration for the deltoid muscle (Fig. 6). In the del- µ # toid muscle, the maximal responses occurred after the first stimulus when the stimulation intensities increased (P = 0.033). 60 No clear bursts of late reflex activity were observed under # anaesthesia when stimulating at intensities up to 40 mA, rather a 40 diffuse increased EMG activity compared to the pre-stimulation * # interval was found. The average amplitude of the late reflex activity of both muscles was significantly increased with increasing 20 stimulation intensities at all isoflurane MAC levels (deltoid muscle: 0.85 iMAC P = 0.016; 0.95 iMAC P < 0.001; 1.05 iMAC P < 0.001; extensor muscle: 0.85 iMAC P = 0.005; 0.95 iMAC P < 0.001; 1.05 0 iMAC P = 0.002), but there were no differences among isoflurane 3 5 10 20 30 40 MAC in responses to stimulations of equal intensities and the ex- Stimulation intensity [mA] tent of increase was much lower than that observed for the epochs 20–70 ms (Fig. 7). Fig. 5. Median and IQR root-mean-square amplitude of averaged reflexes recorded The background EMG amplitude remained stable during anaes- from the deltoid and common digital extensor (CDE) muscle in the 20–70 ms post- thesia and no significant differences were found among different stimulation intervals. Repeated stimulations were given at 5 Hz over 2 s with intensities of 3–40 mA. # P < 0.05: for a constant isoflurane concentration, the isoflurane MAC. Overall, the median amplitude was 4 lV (2– average reflex amplitude increased significantly with increasing stimulation 5 lV) for the deltoid muscle and 3 lV (1–5 lV) for the common intensities. *P < 0.001: the average reflex amplitude was significantly more digital extensor. The amplitude of the post-stimulation interval suppressed at 1.05 iMAC than at 0.85 iMAC. (2500–4000 ms) showed a small but statistically significant trend to rise with increasing stimulation intensities (P < 0.001 for both muscles), but this was not significantly related to isoflurane con- at 0.85 iMAC isoflurane. Reflex withdrawal movements always centration. Complex purposeful movements in response to re- accompanied EMG reflex activity recorded from the deltoid, peated electrical stimulations were observed in 5/7 ponies only whereas if reflex activity was recorded from the common digital C. Spadavecchia et al. / The Veterinary Journal 183 (2010) 337–344 341

Deltoid CDE 100 100

0.85 iMAC 3 mA 5 mA 0.85 iMAC * 80 10 mA 80 20 mA 30 mA 40 mA 60 60

40 40 RMS amplitude [ µ V] 20 RMS amplitude [ µ V] 20

0 0 12345678910 12345678910 Stimulus number Stimulus number

100 100

0.95 iMAC 0.95 iMAC 80 80

60 60

40 40 RMS amplitude [ µ V] 20 RMS amplitude [ µ V] 20

0 0 12345678910 12345678910 Stimulus number Stimulus number

100 100

1.05 iMAC 1.05 iMAC 80 80

60 60

40 40

RMS amplitude [ µ V] 20 RMS amplitude [ µ V] 20

0 0 12345678910 12345678910 Stimulus number Stimulus number

Fig. 6. Median root-mean-square amplitude of reflexes recorded after each stimulus from the deltoid (left column) and from the common digital extensor (CDE, right column) muscles in the 20–70 ms post-stimulation intervals. Ten stimuli were given at 5 Hz over 2 s with intensities of 3–40 mA, at 0.85, 0.95 and 1.05 iMAC. *P < 0.001: absolute temporal summation increased significantly with increasing stimulus intensity. extensor only, as occurred sporadically, it was not accompanied by except that immediately above MAC reflexes to a single stimula- visible reflex movements. tion disappeared in some animals as only 5/7 ponies still showed reflexes at 1.05 MAC (Spadavecchia et al., 2006). On the contrary, Discussion reflexes to repeated electrical stimulation were not abolished at concentrations that were able to prevent purposeful movements The results of the present study indicate that depressive effects in response to supra-maximal noxious stimulations. This is consis- of isoflurane on spinal sensory-motor processing can be described tent with the results reported by Petersen-Felix et al. (1995, 1996) and quantified in ponies using a TS model of withdrawal reflexes. who found that TS, but not reflexes to single stimuli, could still be In ponies anaesthetised with isoflurane at concentrations of evoked at isoflurane concentrations around MAC in humans. As ex- approximately one MAC, TS thresholds increased with increasing pected, none of the ponies in our study showed purposeful move- isoflurane concentrations. ments in response to 2 s repeated electrical stimulations when At all tested MAC multiples, the amplitude of the average reflex anaesthetised at approximately MAC isoflurane concentration response increased with stimulus intensity, whereas the slope of (0.95 and 1.05 iMAC), while spontaneous movements following re- the stimulus–response function was reduced with increasing iso- peated stimulations were observed below MAC in 5/7 ponies. flurane concentrations. These results confirmed our previous re- MAC is the ED50 anaesthetic concentration needed to block port for responses to single stimuli during the same experiment, gross and purposeful movement evoked by a supra-maximal nox- 342 C. Spadavecchia et al. / The Veterinary Journal 183 (2010) 337–344

100 In rats, increasing concentrations of isoﬂurane from 0.7 to 1.4 Deltoid MAC uniformly depressed the motor response but had variable effects on neuronal windup recorded from nociceptive lumbar spinal 80 0.85 iMAC neurons when C-ﬁbre stimulation intensity was used (Jinks et al.,

V] 0.95 iMAC 2004). In that study, a generalised reduction in neuronal excitabil-

[µ 1.05 iMAC ity for increasing isoflurane concentration was hypothesised as the 60 initial C-fibre-evoked responses were reduced in a concentration- dependent manner. In rats, increasing isoflurane from 0.8 to 1.2 40 MAC significantly depressed the initial C-fibre response but en- hanced windup was observed (Cuellar et al., 2005; Ng and Antog- nini, 2006), again demonstrating that TS effects persist above MAC 20 under isoflurane anaesthesia. # As reported for single stimuli, during the 2 s of repeated stimulation, the reflex responses recorded from both forelimb muscles 0 were detected during the epochs 20–70 ms after each stimulus 3 5 10 20 30 40 (Spadavecchia et al., 2006), while almost no activity was recorded Stimulus intensity [mA] in the epochs 70–200 ms. Activation of both non-nociceptive and 100 nociceptive fibres seems to account for the reflex recorded during isoflurane anaesthesia in ponies (Spadavecchia et al., 2004, 2006). CDE The stability of the reflex duration observed at different isoflurane 80 concentrations when stimulations at the maximal intensity were

V] given seems to conﬁrm that when the reﬂex was present both A-

[µ beta and A-delta fibres were activated together. The fact that a vis- 60 ible flexion-protraction movement always accompanied reflexes evoked by stimulation intensities above threshold proves that 40 the reflex was at least partly nociceptive in nature. The shortening of the reflex latency observed within the stimulation series for the muscle deltoid may indicate that central facilitation mechanisms 20 are recruited or that the number of the fast conducting fibres recruited increases when stimulations are repeated. No latency # shortening was observed for the common digital extensor, which 0 maintained stable reflex characteristics during the stimulation 3 5 10 20 30 40 series. Stimulation intensity [mA] We evaluated TS of A fibre activity primarily by calculating the average amplitude of 10 reflex responses obtained during the stim- Fig. 7. Median and IQR root-mean-square amplitude of averaged reflexes recorded ulation series at constant intensity. The average amplitude pro- from the deltoid and the common digital extensor (CDE) muscle in the late reflex vides information about the whole effects of repeated epochs (70–200 ms post-stimulation intervals). Repeated stimulations were given at 5 Hz over 2 s with intensities of 3–40 mA. # P < 0.05: at each isoflurane stimulation, but does not indicate if there has been an increase concentration, average reflex amplitude increased significantly with increasing or decrease in the amplitude of the reflexes recorded during the stimulation intensities. stimulation series. For this reason we compared the average amplitude to the amplitude of the response to the first stimulus in each train. The amplitude of the first response should provide informa- ious stimulus and is considered to be the standard method for tion about the baseline excitability of the reflex pathway (Clarke determining the immobilising potencies of anaesthetics (Quasha et al., 2002). If the average amplitude was always higher than et al., 1980). Traditionally, MAC determination relies on the detec- the amplitude of the first response, it would mean that an increase tion of complex purposeful movements to supra-maximal stimula- in the reflex amplitude during the stimulation series consistently tions involving multiple body parts. These movements are initiated occurs. We did not find any significant difference between the and terminated by the central pattern generator neurons in the average amplitude and the amplitude of the first response, which spinal cord (Jinks et al., 2005). As facilitation of reflex activity per- indicates that in our experiment no consistent increase in reflex sists at isoflurane concentrations which abolish complex move- amplitude occurred for subsequent stimuli. On the other hand, ments, a differential anaesthetic action on the neural networks we looked at the absolute TS as a tool to assess response recruit- involved in processing these two movement patterns must be ment characteristics under different conditions. For this parameter, hypothesised. a significant trend to increase was observed for the common digital From studies in experimental animals it is known that TS gov- extensor muscle only at the lowest isoflurane concentration, while erns part of the MAC of isoflurane and influence anaesthetic at higher concentrations the responses were flattened. This con- requirements (Dutton et al., 2003, 2007). Increasing the duration firms the findings of Dutton et al. (2007) who showed stable but or frequency of stimulation increases the concentration of isoflu- intensity dependent A fibres responses from the sacral dorsal neu- rane required to suppress movement by a 0.4 MAC in rats. From rons in rats when 20 stimuli up to five times the C-fibre threshold an in vitro study it is known that blocking NMDA receptors blocks were given. TS of C-fibre activity, commonly defined as windup (Woolf and Although reflexes of similar size were obtained during the stim- Thompson, 1991). Persistence of TS during isoflurane administra- ulation series, we observed a difference in the reflex threshold to tion might reflect persisting NMDA receptor function (Sonner repeated stimuli compared to the reflex threshold to single stimu- et al., 2003), although some contrasting evidence has been pre- lus (Spadavecchia et al., 2006). When repeated stimuli were given, sented (Yamakura and Harris, 2000). Thus, the immobilising effects a lower stimulation intensity was necessary to evoke at least one of isoflurane seem to be partially independent of its activity on the reflex during the stimulation series, confirming that a facilitation NMDA receptor (Sonner et al., 2003). effect (temporal summation) was actually occurring. In rats anaes- C. Spadavecchia et al. / The Veterinary Journal 183 (2010) 337–344 343 thetised with isoflurane up to 1.2 MAC, A fibre responses evoked by References repeated electrical stimulations and recorded from dorsal horn neurons remained stable during the stimulation series and were Antognini, J.F., Atherley, R., Carstens, E., 2003. Isoflurane action in spinal cord indirectly depresses cortical activity associated with electrical stimulation of not significantly affected by increasing anaesthetic concentration the reticular formation. Anesthesia and Analgesia 96, 999–1003. (Cuellar et al., 2005; Mitsuyo et al., 2006). These latter results are Antognini, J.F., Carstens, E., 1999. Increasing isoflurane from 0.9 to 1.1 minimum in contrast with our findings, demonstrating a significant depres- alveolar concentration minimally affects dorsal horn cell responses to noxious sant effects of A fibre related reflexes when isoflurane was in- stimulation. Anesthesiology 90, 208–214. Antognini, J.F., Carstens, E., Buzin, V., 1999a. Isoflurane depresses motoneuron creased from 0.85 to 1.05 MAC. This discrepancy might be due to excitability by a direct spinal action: an F-wave study. Anesthesia and Analgesia the fact that we looked at the evoked withdrawal reflex response 88, 681–685. and not purely at dorsal horn neuronal activity. Antognini, J.F., Wang, X.W., 1999. Isoflurane indirectly depresses middle latency auditory evoked potentials by action in the spinal cord in the goat. Canadian In general, investigating the modulation of reflexes does not Journal of Anesthesia 46, 692–695. permit differentiation of pharmacological effects on afferent and Antognini, J.F., Wang, X.W., Carstens, E., 1999b. Quantitative and qualitative effects efferent reflex pathways. An invasive approach in laboratory ani- of isoflurane on movement occurring after noxious stimulation. Anesthesiology 91, 1064–1071. mals has been used to examine the selective effects of inhaled ana- Antognini, J.F., Wang, X.W., Carstens, E., 2000. Isoflurane action in the spinal cord esthetics on neurons in the dorsal horn (Antognini and Carstens, blunts electroencephalographic and thalamic-reticular formation responses to 1999; Jinks et al., 2003) and in the ventral spinal cord (Kim noxious stimulation in goats. Anesthesiology 92, 559–566. Arendt-Nielsen, L., Sonnenborg, F.A., Andersen, O.K., 2000. Facilitation of the et al., 2007). Depression of dorsal horn neurons by isoflurane oc- withdrawal reflex by repeated transcutaneous electrical stimulation: an curs mainly well below one MAC (0.4–0.8 MAC) and therefore an experimental study on central integration in humans. European Journal of isoflurane immobilising action is not attributable to depression Applied Physiology 81, 165–173. Blythe, L.L., Kitchell, R.L., Holliday, T.A., Johnson, R.D., 1983. Sensory nerve of nociceptive transmission through the dorsal horn, but rather conduction velocities in forelimb of ponies. American Journal of Veterinary by depression of the ventral horn neurons. It has been recently Research 44, 1419–1426. demonstrated that isoflurane, when administered in the peri- Clarke, R.W., Eves, S., Harris, J., Peachey, J.E., Stuart, E., 2002. Interactions between MAC concentration range (0.8–1.2 MAC), depressed motor neurons cutaneous afferent inputs to a withdrawal reflex in the decerebrated rabbit and their control by descending and segmental systems. Neuroscience 112, 555– and pre-motor interneurons to induce immobility in rats (Kim 571. et al., 2007). Cuellar, J.M., Dutton, R.C., Antognini, J.F., Carstens, E., 2005. Differential effects of We found the concentration-dependent increase of TS thresh- halothane and isoflurane on lumbar dorsal horn neuronal windup and excitability. British Journal of Anaesthesia 94, 617–625. old to be a muscle-specific phenomenon. Consistent with our Dutton, R.C., Cuellar, J.M., Eger 2nd, E.I., Antognini, J.F., Carstens, E., 2007. Temporal findings when single stimuli were applied (Spadavecchia et al., and spatial determinants of sacral dorsal horn neuronal windup in relation to 2006), the common digital extensor muscle was activated at low- isoflurane-induced immobility. Anesthesia and Analgesia 105, 1665–1674. Dutton, R.C., Zhang, Y., Stabernack, C.R., Laster, M.J., Sonner, J.M., Eger 2nd, E.I., 2003. er stimulation intensities than the deltoid muscle, and its activity Temporal summation governs part of the minimum alveolar concentration of was less sensitive to the depressant effects of the anaesthetic. isoflurane anesthesia. Anesthesiology 98, 1372–1377. Muscle-specific TS thresholds were not seen in conscious horses Gasser, H., Erlanger, J., 1927. The role played by the sizes of the constituent fibers of a nerve trunk in determining the form of its action potential wave. American (Spadavecchia et al., 2004) and after recovery from anaesthesia, Journal of Physiology 80, 522–547. which suggests that isoflurane, lateral recumbency, or both must Jinks, S.L., Antognini, J.F., Dutton, R.C., Carstens, E., Eger 2nd, E.I., 2004. Isoflurane account for the differences in patterns of reflexive muscle activa- depresses windup of C fiber-evoked limb withdrawal with variable effects on tion during anaesthesia. Further studies are needed to clarify this nociceptive lumbar spinal neurons in rats. Anesthesia and Analgesia 99, 1413– 1419. issue. Jinks, S.L., Atherley, R.J., Dominguez, C.L., Sigvardt, K.A., Antognini, J.F., 2005. Isoflurane disrupts central pattern generator activity and coordination in the lamprey isolated spinal cord. Anesthesiology 103, 567–575. Conclusions Jinks, S.L., Martin, J.T., Carstens, E., Jung, S.W., Antognini, J.F., 2003. Peri-MAC depression of a nociceptive withdrawal reflex is accompanied by reduced dorsal The present study has shown that depressive effects of isoflu- horn activity with halothane but not isoflurane. Anesthesiology 98, 1128–1138. rane on spinal sensory-motor processing in ponies can be de- Kim, J., Yao, A., Atherley, R., Carstens, E., Jinks, S.L., Antognini, J.F., 2007. Neurons in the ventral spinal cord are more depressed by isoflurane, halothane, and scribed and quantified using the model of TS of withdrawal propofol than are neurons in the dorsal spinal cord. Anesthesia and Analgesia reflexes. Reflexes following electrical stimulations were still 105, 1020–1026. evoked at isoflurane concentrations in the peri-MAC range and a Mitsuyo, T., Dutton, R.C., Antognini, J.F., Carstens, E., 2006. The differential effects of halothane and isoflurane on windup of dorsal horn neurons selected in reflex facilitation was observed following the application of re- unanesthetized decerebrated rats. Anesthesia and Analgesia 103, 753–760. peated stimuli. While a TS paradigm was used as an experimental Ng, K.P., Antognini, J.F., 2006. Isoflurane and propofol have similar effects on spinal tool, repetitive noxious stimuli are often applied to equine patients neuronal windup at concentrations that block movement. Anesthesia and during surgery. So, under isoflurane TS of evoked potentials and Analgesia 103, 1453–1458. Petersen-Felix, S., Arendt-Nielsen, L., Bak, P., Fischer, M., Bjerring, P., Zbinden, A.M., windup might still occur. A multimodal approach for intra-opera- 1996. The effects of isoflurane on repeated nociceptive stimuli (central tive pain control may decrease the likelihood of developing central temporal summation). Pain 64, 277–281. sensitisation and postoperative pain in equine patients. Petersen-Felix, S., Arendt-Nielsen, L., Bak, P., Roth, D., Fischer, M., Bjerring, P., Zbinden, A.M., 1995. Analgesic effect in humans of subanaesthetic isoflurane concentrations evaluated by experimentally induced pain. British Journal of Conflict of interest statement Anaesthesia 75, 55–60. Quasha, A.L., Eger 2nd, E.I., Tinker, J.H., 1980. Determination and applications of MAC. Anesthesiology 53, 315–334. None of the authors of this paper has a financial or personal Sonner, J.M., Antognini, J.F., Dutton, R.C., Flood, P., Gray, A.T., Harris, R.A., Homanics, relationship with other people or organisations that could inappro- G.E., Kendig, J., Orser, B., Raines, D.E., Rampil, I.J., Trudell, J., Vissel, B., Eger 2nd, E.I., 2003. Inhaled anesthetics and immobility: mechanisms, mysteries, and priately influence or bias the content of this paper. minimum alveolar anesthetic concentration. Anesthesia and Analgesia 97, 718– 740. Spadavecchia, C., Andersen, O.K., Arendt-Nielsen, L., Spadavecchia, L., Doherr, M., Acknowledgements Schatzmann, U., 2004. Investigation of the facilitation of the nociceptive withdrawal reflex evoked by repeated transcutaneous electrical stimulations The authors would like to thank Dr. Luciano Spadavecchia for as a measure of temporal summation in conscious horses. American Journal of developing and manufacturing the electrophysiological equip- Veterinary Research 65, 901–908. Spadavecchia, C., Levionnois, O., Kronen, P.W., Leandri, M., Spadavecchia, L., ment and Dr. Marcus Doherr for his assistance in the statistical Schatzmann, U., 2006. Evaluation of administration of isoflurane at analysis. approximately the minimum alveolar concentration on depression of a 344 C. Spadavecchia et al. / The Veterinary Journal 183 (2010) 337–344

nociceptive withdrawal reflex evoked by transcutaneous electrical stimulation Yamakura, T., Harris, R.A., 2000. Effects of gaseous anesthetics nitrous oxide and in ponies. American Journal of Veterinary Research 67, 762–769. xenon on ligand-gated ion channels. Comparison with isoflurane and ethanol. Woolf, C.J., Thompson, S.W., 1991. The induction and maintenance of central Anesthesiology 93, 1095–1101. sensitization is dependent on N-methyl–D-aspartic acid receptor activation; Zhou, H.H., Jin, T.T., Qin, B., Turndorf, H., 1998. Suppression of spinal cord implications for the treatment of post-injury pain hypersensitivity states. Pain motoneuron excitability correlates with surgical immobility during isoflurane 44, 293–299. anesthesia. Anesthesiology 88, 955–961. The Veterinary Journal 186 (2010) 304–311

Contents lists available at ScienceDirect

The Veterinary Journal

journal homepage: www.elsevier.com/locate/tvjl

Effect of ketamine on the limb withdrawal reﬂex evoked by transcutaneous electrical stimulation in ponies anaesthetised with isoﬂurane

Olivier L. Levionnois a,*, Monika Menge b, Wolfgang Thormann c, Meike Mevissen b, Claudia Spadavecchia a a Division of Veterinary Anaesthesiology, Department of Clinical Veterinary Medicine, Vetsuisse Faculty, University of Bern, 3001 Bern, Switzerland b Division of Veterinary Pharmacology and Toxicology, Department of Clinical Research and Veterinary Public Health, Vetsuisse Faculty, University of Bern, Switzerland c Department of Clinical Pharmacology and Visceral Research, University of Bern, Switzerland article info abstract

Article history: The purpose of this study was to evaluate the anti-nociceptive activity of ketamine and isoflurane in Accepted 15 August 2009 horses using a limb withdrawal reflex (WR) model. Single and repeated stimulations were applied to the digital nerve of the left forelimb in ponies anaesthetised with isoflurane before, during and after intravenous administration of racemic ketamine. Surface electromyographic activity was recorded from the Keywords: deltoid muscle. Higher stimulation intensity was required to evoke a reflex during ketamine administra- Ketamine tion. Furthermore, the amplitudes of response to stimulations were significantly and dose-dependently Withdrawal reflex depressed and a flattening of the stimulus–response curves was observed. The reflex activity recovered Electrical stimulus partially once the ketamine infusion finished. The results demonstrated that the limb WR can be used Electromyography Equine to quantify the temporal effect of ketamine on the sensory-motor processing in ponies anaesthetised with isoflurane. Ó 2009 Elsevier Ltd. All rights reserved.

Introduction The basic physiology of the WR and the modulatory effects of different drugs (Spadavecchia et al., 2005; Spadavecchia et al., Ketamine is widely used in veterinary anaesthesia to induce and 2006, 2007) have been reported in conscious horses (Spadavecchia maintain unconsciousness (Wright et al., 1987). It has been also et al., 2002). Recently, dose-dependent WR modulation by isoflu- shown to produce anti-nociception in humans (Arendt-Nielsen rane (Spadavecchia et al., 2006, in press) and ketamine (Peterbauer et al., 1995; Kohrs and Durieux, 1998) and experimental animals et al., 2008) when administered alone have been described in po- (Bergadano et al., 2009) and is recommended for balanced nies. The depression of the amplitude of the reflexes evoked from anaesthesia and post-operative analgesia in clinical veterinary pa- single or repeated electrical stimulations was used to quantify tients (Sarrau et al., 2007; Solano et al., 2006; Wagner et al., 2002; the anti-nociceptive properties of the drugs. Although ketamine Wilson et al., 2008). The anti-nociceptive activity of ketamine is and isoflurane are often used together to produce anaesthesia in mostly attributed to its non-competitive antagonist activity at horses, the combined effect of the two drugs and the mechanism the N-methyl-D-aspartate (NMDA) receptors within the spinal cord of action on the anti-nociceptive system are not known. (Kohrs and Durieux, 1998; Nadeson et al., 2002). Its putative inhib- The purpose of this study was to evaluate the anti-nociceptive itory properties on spinal wind-up and central sensitisation (Woolf activity of ketamine and isoflurane in equids by quantification of and Thompson, 1991) have made it a favoured drug to treat trau- the WR. The hypothesis was that an intravenous (IV) infusion of matic, neuropathic and chronic pain (Woolf and Mannion, 1999; ketamine would depress the WR in ponies anaesthetised with Woolf and Max, 2001). isoflurane. The limb withdrawal reflex (WR) has been investigated and standardised in humans as a polysynaptic spinal response (Sandrini et al., 2005). Electromyographic (EMG) evaluation of the WR reflex Materials and methods permits indirect quantification of the activation of small-diameter, high-threshold nociceptive Ad fibres as the reflex becomes evident Animals at threshold stimulation intensities evoking pain. EMG is therefore The experiments were conducted on five healthy 5-year old Shetland pony geld- a useful tool for investigating pain processes at spinal level. ings, of mean bodyweight (BW) 128 ± 17 kg. The health condition of the animals was based on physical, biochemical, and haematological examinations. Food was withheld for 24 h before the experiments, but the animals had free access to water. * Corresponding author. Tel.: + 41 31 631 2287; fax: +41 31 631 2620. For each pony, the left carotid artery had been surgically elevated to subcutaneous E-mail address: [email protected] (O.L. Levionnois). layers 2 years before the experiment.

All animals received care according to the laws on care and use of experimental formed using self-adhesive surface electrodes placed 20 mm apart. animals in Switzerland and the experiments were approved by the Bernese state The resistance of the stimulation electrodes was <2 kX. The ground committee for animal experimentation. In previous studies, the individual minimum alveolar concentration (iMAC) for electrode was placed on the dorsum of each pony at a location isoflurane (Spadavecchia et al., 2006) and the individual pharmacokinetic parame- immediately caudal to the point of the shoulders. Stimulations ters for ketamine enantiomers (Larenza et al., 2007) had been determined for each and recordings were performed following a technique described pony. by Spadavecchia et al. (2003, 2006). Each single stimulus lasted a total of 25 ms and consisted of a train of 5 Â 1 ms, constant-cur- Induction and monitoring of anaesthesia rent, square-wave pulses delivered at a frequency of 200 Hz. For each pony, five series of stimulations were performed to After pre-oxygenation, anaesthesia was induced with isoflurane evoke the WR, as recorded by EMG responses from the deltoid (Isoflo, Abbott) in oxygen via a face mask, and maintained using a muscle (ipsilateral limb) at the following time points before the conventional circle anaesthetic system (Electronic Respirator 3100, ketamine administration: (T0), 30 (T1) and 110 (T2) min after Roche) following endotracheal intubation. The end-tidal concentra- the infusion started, and 50 (T3) and 70 (T4) min after it stopped. tion of isoflurane was set at iMAC (±0.2%) for each pony and main- At each time point, single stimuli were first delivered with ascend- tained constant (±0.03%) throughout the experiment. Mechanical ing intensity at 30 s intervals from 5–45 mA in 5 mA steps to assess ventilation (i.e. intermittent positive-pressure ventilation) was the stimulus–response curve. Subsequently, a repeated stimulation started, with initial settings of 8 breaths/min and a tidal volume of series with a total of 10 stimuli given over 2 s at a frequency of 12 mL/kg, and further tailored to maintain end-tidal carbon dioxide 5 Hz, was initiated following the same ascending pattern (from (CO2) partial pressure at approximately 40 mm Hg. 5–45 mA). Catheters were inserted into a saphenous vein and a carotid artery to collect blood, for the administration of lactated Ringer’s Quantification of reflexes evoked by single stimulus solution and ketamine, and to monitor arterial blood pressure. Oesophageal body temperature, ECG, pulse oximetry, arterial The EMG activity was recorded from 100 ms before to 400 ms blood pressure, inspired and end-tidal CO2 and anaesthetic gas after the stimulus. The background EMG amplitude, calculated as concentrations were continuously monitored (S/5 Compact, Da- the root-mean-square (RMS) amplitude during the 100 ms interval tex-Ohmeda). The unit measuring breathing gas concentrations before stimulation, was <5 lV. To quantify the reflex response, the was calibrated using a calibration canister (QuickCal Calibration RMS of the EMG amplitude was calculated for the 25–70 ms epoch gas 755583-HEL, Datex-Ohmeda) according to the manufacturer’s after stimulation onset to target the response to Ad-fibres activity instructions before each experiment. Arterial blood gas analyses (nociceptive activation of the WR) (Spadavecchia et al., 2006). For were performed hourly. After the end of each experiment, all po- each EMG recording of a WR, latency defined as the time from nies recovered and were carefully monitored for signs of post- stimulation onset to onset of EMG deflection from baseline, and anaesthetic discomfort over the following days. duration defined as the time from onset to offset of EMG deflection from baseline were characterised. Ketamine administration and plasma level determination Determination of the stimulus–response curve An IV infusion using racemic ketamine (Ketasol, Graeub) was given firstly as an intravenous bolus and then, to maintain the keta- To be considered a WR, the EMG RMS amplitude had to be at mine plasma level constant over the experiment, a purposely built least 50 lV. The minimal current required (mA) to evoke a WR high-precision infusion pump was used driven by a computer run- was defined as threshold for single stimulation series (SS thresh- ning a specifically designed routine. This programme was based on old). If no reflexes were evoked at maximal intensity (45 mA), pharmacokinetic equations from Maitre and Shafer (1990) to the SS threshold was considered 50 mA. The area under the curve administer an open-loop blood-targeted target-controlled infusion (AUC) of the WR recruitment (EMG RMS amplitude for increasing (TCI) of racemic ketamine over 120 min. A recalculation frequency stimulation intensities) was calculated at each measuring time every 5 s was used to obtain a target plasma concentration of 1 lg/ point, in all ponies. mL for S(+)-ketamine. The predicted ketamine plasma concentration for both enantiomers was 2 lg/mL (Knobloch et al., 2006). Quantification of reflexes evoked by repeated stimulation Individual pharmacokinetic parameters for both ketamine enantiomers were considered. The EMG was recorded from 500 ms before, until 1500 ms after Arterial blood samples were taken from the left carotid artery the stimulation (total recording time of 4000 ms, sampling fre- into heparinised tubes shortly before the ketamine administration quency, 1 kHz). The background EMG amplitude, calculated as at 0, 0.5, 1, 2, 3, 6, 10, 30, 50, 70, 90, 110 and 120 min after the infu- the RMS amplitude during the 400 ms interval before stimulation, sion started, and 1, 3, 6, 10, 30, 50 and 70 min after the infusion was <5 lV. The RMS amplitude of the reflex response for the stopped. All samples were immediately put on ice, centrifuged epochs 25–70 ms after each of the 10 stimuli in the stimulus train, and the plasma was stored at À80 °C until analysis. Enantiomers and the average amplitude of the 10 responses were calculated. of ketamine and its active metabolite norketamine were measured Moreover, the absolute Temporal Summation (absTS) was in plasma using capillary electrophoresis, as described previously calculated by subtracting the RMS amplitude of the first response (Theurillat et al., 2005; Knobloch et al., 2006). The detection limit from the first and each subsequent response and by summing for the enantiomers of ketamine and norketamine was 10 ng/mL. the residuals, as described elsewhere (Spadavecchia et al., in press). The threshold for repeated stimulation series (RS threshold) was Stimulations and electrophysiological recordings of EMG reflex activity defined as the stimulation intensity able to evoke at least one reflex during the stimulation series, as previously defined. The AUC After each pony was anaesthetised and positioned in right lat- of the stimulus–response curve (average EMG RMS amplitude for eral recumbency, the skin over the palmar lateral digital nerve increasing stimulation intensities) was calculated at each measur- and over the deltoid muscle of the left forelimb was clipped and ing time point, for all ponies. The presence of late reflex activity cleaned. Transcutaneous nerve stimulation (Neuroline 7 00 02-J, was evaluated by calculating the RMS amplitude for the epochs Medicotest) and EMG recordings (Synapse, Ambu A/S) were per- 70–200 ms. 306 O.L. Levionnois et al. / The Veterinary Journal 186 (2010) 304–311

Statistical analysis rate ranging from 156 (IQR, 108–170) to 102 (IQR, 70–102) lg/ kg/min. Measured plasma concentrations for R/S ketamine and Anaesthetic data were normally distributed and are presented their norketamine metabolites are shown in Fig. 1. The median as means (±SD). Other data including plasma ketamine plasma concentrations of ketamine were maintained constant dur- concentrations are reported as median and inter-quartile ranges ing T1 and T2 and decreased rapidly after ceasing the administra- (IQR, 25–75%) and were analysed with nonparametric tests. Where tion of ketamine (Table 1). appropriate, Wilcoxon signed-rank test for paired data, Friedman analysis of variance on ranks for repeated measures followed by Description of the EMG responses to stimulations Tukey’s test for pair-wise multiple comparison or a Spearman rank order correlation test were used. Values of P < 0.05 were When a WR was evoked, a clear EMG burst was always pres- considered significant. Statistical analyses were performed by use ent within the epoch 25–70 ms after the stimulation onset, with of commercially available software (Sigmastat, Systat Software). similar shapes among ponies, stimulations intensities and over the course of the experiment (Fig. 2). Latencies and durations Results of deltoid EMG activity after single stimulation were decreasing and increasing significantly with rising stimulation intensity Anaesthetic period (P < 0.05), respectively. The RMS amplitude of the WR after single stimulation was positively correlated to stimulation intensity The total duration of anaesthesia was 332 (±46) min. Normo- before (T0) and after (T3, T4) ketamine administration (P < 0.05, capnia, normotension and normothermia were maintained Fig. 3A, Table 1). throughout in all ponies, with arterial CO2 partial pressure of During ketamine administration (T1, T2), the RMS amplitude 5.85 (±0.2) kPa, mean arterial blood pressure of 86.3 (±7.3) mmHg did not increase significantly with ascending stimulation inten- and body temperature of 37.1 (±0.5) °C. In one pony, the end-tidal sity and was significantly different (lower) when compared to isoflurane concentration was increased at 1.1 iMAC because the baseline measurements at T0 (P < 0.05, Fig. 3A). Similarly, the pony showed signs of awareness (nystagmus) and spontaneous average reflex amplitude over the 10 responses for repeated movements during the first stimulations. All ponies recovered stimulation increased significantly with stimulation intensities uneventfully from the anaesthesia. at T0, T3 and T4 (P < 0.05, Fig. 3B, Table 1), but not at T1 and T2. Through the ketamine administration, the average RMS Ketamine infusion rates and plasma levels amplitude was significantly different (lower) than at T0 (P < 0.05, Fig. 3B). The median RS threshold was always lower Each pony received an IV bolus of racemic ketamine at 362 (IQR, than the SS threshold, but no statistical difference was found 294–430) lg/kg followed by an exponentially decreasing infusion (P = 0.174, Fig. 4). Both SS threshold (P = 0.05) and RS threshold

Fig. 1. Median and inter-quartile range of total plasma concentrations of R/S ketamine and R/S norketamine in ﬁve ponies before (T0), during (T1, T2) and after (T3, T4) target- controlled infusion of racemic ketamine over 120 min.

Table 1 Median and inter-quartile range (IQR) for total plasma concentration of R/S ketamine, root-mean-square of the reﬂex amplitude after single and repeated stimulations at 45 mA (SS RMS 45 and RS RMS 45, respectively) and intensity threshold to elicit a reﬂex activity after single and repeated stimulations (SS threshold and RS threshold, respectively) before (T0), during (T1, T2) and after (T3, T4) ketamine administration.

T0 T1 T2 T3 T4 Ketamine plasma 0 (IQR, 0–0) 3.79 (IQR, 2.2–4.07) 3.25 (IQR, 1.87–3.58) 0.52 (IQR, 0.43–0.57) 0.33 (IQR, 0.23–0.48) concentrations (lg/mL) SS RMS 45 (lV) 256 (IQR, 107–413) 20 (IQR, 2.75–52.5) 60 (IQR, 4.25–62.25) 89 (IQR, 6.75–179.5) 123 (IQR, 6.5–219.5) RS RMS 45 (lV) 275.9 (IQR, 257.75–423.1) 37.38 (IQR, 11.62–75.23) 44.79 (IQR, 23.3–70.75) 85.49 (IQR, 28.14–202.94) 95.76 (IQR, 33.99–204.14) SS threshold (mA) 15 (IQR, 10–27.5) 50 (IQR, 40–50) 45 (IQR, 30–50) 20 (IQR13.75–50) 20 (IQR, 10–50) RS threshold (mA) 10 (IQR, 8.75–12.5) 45 (IQR, 23.75–50) 35 (IQR, 20–50) 15 (IQR, 13.75–31.25) 20 (IQR, 13.75–50) O.L. Levionnois et al. / The Veterinary Journal 186 (2010) 304–311 307

Fig. 2. Examples of electromyographic deﬂection recorded at the deltoid muscle elicited by the withdrawal reﬂex after single (A, 1 Â 5 pulses) or repeated (B, 10 Â 5 pulses at 5 Hz) electrical stimulation at increasing intensities (5, 20 and 40 mA) before (T0) and during (T2) R/S ketamine target-controlled infusion in one pony.

Fig. 3. (A) Median of the root-mean-square (RMS) amplitude of the electromyographic deflection recorded at the deltoid muscle elicited by the withdrawal reflex after electrical stimulation at increasing intensities (5–45 mA) before (T0), during (T1, T2) and after (T3, T4) R/S ketamine target-controlled infusion in five ponies. (B) Median of the average root-mean-square (RMS) amplitude of the electromyographic deflections elicited by repeated stimulations (series of 10 Â 5 pulses). §Significantly different from T0 at all stimulation intensities. *Significantly increasing with ascending stimulation intensity.

(P = 0.03) were significantly increased during ketamine adminis- single and repeated stimulation were significantly different at T1 tration compared to before and after (Fig. 4, Table 1). and at T2 compared to T0 (P < 0.01; Fig. 6). In all ponies, AUCs were The first reflex evoked during the repeated stimulation series highest at T0 and smallest at T1 or T2. A significant correlation was was not significantly different from the response to the single estimated between actual ketamine plasma levels and the AUC of stimulation (P = 0.2). Responses to repeated stimulations gener- the EMG RMS amplitude (P < 0.05) as illustrated in Fig. 6. The EMG ally tended to decrease with the number of stimulations but RMS amplitude did not return to initial values when the ketamine no consistent pattern could be identified. No clear burst of late plasma levels decreased, drawing a hysteresis loop. reflex activity was observed during the repeated stimulation series at intensities up to 45 mA, but the diffuse EMG activity was Discussion significantly increased compared to the pre-stimulation interval (P < 0.001). The median RMS amplitude of this diffuse late reflex Racemic ketamine significantly and dose-dependently de- activity increased with rising intensity at T0 (P < 0.05; Fig. 5), pressed the reflex EMG activity when added to isoflurane anaes- and ketamine administration (T1, T2) resulted in a significant thesia administered at iMAC. The reflex activity recovered rapidly depression when compared to T0 (P < 0.05; Fig. 5). after the end of the ketamine infusion. However, it did not return to pre-infusion values within the study time (90 min after Electromyographic responses to stimulations of increasing intensity cessation of ketamine administration), despite the rapid decrease of ketamine plasma concentrations. The stimulus–response curves of the EMG RMS amplitude (lV) The baseline measurements obtained under isoflurane against stimulation intensities (mA) were clearly affected by keta- anaesthesia at iMAC are consistent with previous reports. mine treatment in all ponies (Fig. 3). The AUC of the amplitude for Isoflurane administered alone (at approximately the end-tidal 308 O.L. Levionnois et al. / The Veterinary Journal 186 (2010) 304–311

In the present study, higher plasma levels of ketamine were reached (2–4 lg/mL) in our anaesthetised ponies. The responses to single stimulations were strongly depressed or even completely abolished in several ponies, and a flattening of the recruitment curves was observed in all individuals. This outcome might be explained by a dose-dependent modulation of the WR by ketamine and/or by a synergistic interaction with isoflurane. Owing to its antagonistic property at NMDA receptors, ketamine has been described to affect both the spinal ‘wind-up’ and the development of central sensitisation (Woolf and Max, 2001; Woolf and Thompson, 1991). Low doses of ketamine limit the development of central sensitisation in animal and human pain models, as well as in the post-operative period in the clinic, where it has been described to be anti-hyperalgesic more than analgesic (Richebe et al., 2005). The specific effect of ketamine on the WR is most likely due to inhibiting the activation of the NMDA receptors on wide dynamic range (WDR) neurons after sustained discharge to the dorsal horn from Ad fibres, with amelioration of the subsequent facilitation of the polysynaptic reflex (i.e. expansion of the receptive field, decreased stimulation threshold, Fig. 4. Median threshold intensity (and inter-quartile range) required to evoke a withdrawal reflex of the deltoid muscle after transcutaneous electrical single or increased response amplitude and duration) (Woolf and Chong, repeated stimulation (SS threshold, full circles and RS threshold, open circles, 1993). respectively) before (T0), during (T1, T2) and after (T3, T4) R/S ketamine target- By studying the WR in response to appropriate repeated stimu- § controlled infusion in five ponies. Significantly different from T0. lation patterns, temporal summation of the reflex can be evoked and quantified (Dimitrijevic and Nathan, 1970; Price, 1972; You et al., 2003). Temporal summation is an experimental model of the early process of wind-up and can be used to study and quantify aspects of central integration (Andersen et al., 2005) and its modulation by ketamine. Indeed, ketamine has been shown to exert a selective depression of the WR in response to repeated stimulation in humans (Arendt-Nielsen et al., 1995) and dogs (Bergadano et al., 2009), compared to low concentrations of isoflurane (<0.8 MAC) (Petersen-Felix et al., 1995, 1996). However, isoflurane (at approximately iMAC) was reported to partially depress temporal summation (TS) in ponies (Spadavecchia et al., in press). Temporal and spatial summations have been shown to affect the anaesthetic concentration required to suppress purposeful movement in response to electrical noxious stimuli (Dutton et al., 2007). Temporal summation is thought to influence part of the MAC of isoflurane anaesthesia (Dutton et al., 2003), even if this still remains a contro- versial issue and might be subject to possible species differences (Sonner et al., 2003; Ng and Antognini, 2006). In the present study, no facilitation of the WR, as depicted by a lower threshold compared to single stimulations series or increasing response amplitude throughout repeated stimulations, could be observed after repeated stimulations during baseline measure- Fig. 5. Median RMS amplitude (and inter-quartile range) of the diffuse late reflex ments under isoflurane anaesthesia. Thus, no significant effect of EMG activity (70–200 ms epoch) recorded at the deltoid muscle evoked by repeated ketamine on these parameters was determined. Ketamine equally electrical stimulation before (T0), during (T1, T2) and after (T3, T4) R/S ketamine target-controlled infusion in five ponies. §Significantly different from T0. *Signifi- decreased the responses to single and repeated stimulations. cantly increasing with ascending intensity. Most anti-nociception studies involving ketamine focused on its inhibitory activity on post-synaptic NMDA receptors in WDR neurons at the spinal dorsal horn (Kohrs and Durieux, 1998; Nadeson iMAC) depressed, but did not abolish, the WR in response to a sin- et al., 2002). However, ketamine has been reported to act on a gle stimulation in ponies (Spadavecchia et al., 2006) in a dose- variety of other receptors, such as nicotinic (Scheller et al., 1996), dependent manner. The slope of the stimulus–response curves muscarinic (Hustveit et al., 1995), opioid (Hustveit et al., 1995), was also reduced, but did not flatten completely. Alternatively, monoaminergic (Hirota and Lambert, 1996) and brain NMDA low dose ketamine administration to conscious patients has been (Thomson et al., 1985) receptors, which could potentially influence reported to have little influence on the WR in response to single the WR response. Indeed, a WR is induced when a noxious stimu- stimulation in humans (Arendt-Nielsen et al., 1995) and in dogs lus is transmitted at the spinal level from the dorsal to the ventral (Bergadano et al., 2009). In ponies, comparable plasma horn via a polysynaptic pathway receiving afferent spinal and concentrations of ketamine (50 ng/mL) were associated with a supraspinal modulation to finally activate the motor neuron mild but significant depression of the reflex amplitudes (Peter- (Clarke and Harris, 2004). Therefore, a pharmacological effect at bauer et al., 2008). If such low doses are generally not exceeded each of these levels can affect the EMG output of the WR. in conscious patients to avoid undesirable psychomimetic side ef- Admittedly, the spinal activity of ketamine, classified as a non- fects (Schmid et al., 1999), higher doses can be administered under competitive antagonist at the NMDA receptor, is crucial for its general anaesthesia. analgesic activity at low doses (Ghorpade and Advokat, 1994), with O.L. Levionnois et al. / The Veterinary Journal 186 (2010) 304–311 309

Fig. 6. Median and inter-quartile range of the area under the stimulus–response curve (AUC) against median R/S ketamine total plasma concentrations before (T0), during (T1, T2) and after (T3, T4) ketamine target-controlled infusion in five ponies. (A) Single stimulation series; (B) repeated stimulation series. §Significantly different from T0. the phencyclidine binding site in the dorsal horn being the best A physiologically-based pharmacokinetic model used in the po- candidate (Dickenson, 1995). This hypothesis was supported by nies of the present study receiving the same ketamine TCI Collins (1986) who provided evidence that the analgesic effect of predicted prolonged ketamine concentrations within the ‘rapidly- ketamine may be selective for pain-transmitting systems and perfused’ compartment after termination of ketamine IV infusion may involve the interneurons responsible for communicating that (Knobloch et al., 2006). If the spinal and supraspinal sites of action nociceptive information to the WDR neuron. Conversely, isoflurane for ketamine are assumed to belong to the rapidly-perfused has been found to inhibit spinal nociceptive pathway predomi- compartment rather than the central compartment, this may par- nantly at ventral sites, depressing motor neurons and pre-motor tially explain the hysteresis between the depression of the WR interneurons (Kim et al., 2007). It was originally suggested by and plasma concentrations of ketamine. It is also known that Scheufler et al. (2003) that ketamine also attenuated myogenic norketamine is an active non-competitive antagonist at NMDA motor-evoked potentials (MEP) distal to the spinal level, notably receptor (Ebert et al., 1997) with anti-nociceptive properties under at the neuromuscular junction, although this has since been dis- experimental conditions in rats using the formalin test, constric- proven and ketamine is now commonly used during MEP monitor- tion nerve injury model or tail-flick test (Shimoyama et al., 1999; ing for having little effect on it (Kawaguchi et al., 2000). Holtman et al., 2008a,b). Therefore, it can be hypothesised that Furthermore, additional MEP and H-reflex studies found that keta- norketamine, and particularly S-norketamine (Holtman et al., mine has no inhibitory effect on motor neuron excitability (Chiba 2008b), affected the WR of the ponies in the present study. The rel- et al., 1998; Kakinohana et al., 2000). evance and the extent of this effect according to the circulating A supraspinal hypoalgesic activity has been demonstrated with concentrations of norketamine observed after termination of keta- increasing doses of ketamine leading to general anaesthesia (Col- mine IV infusion remains to be investigated. lins, 1986; Porro et al., 2004). As Arendt-Nielsen et al. (1995) described, the anti-nociceptive effect of ketamine is very probably a continuum ranging from no effect to complete anaesthesia, using Conclusions different modes of action depending on the state of the nociceptive system and the dose. The ability of ketamine to dissociate tactile Plasma concentrations of 2–4 lg/mL of ketamine elicited a po- somatosensory information has been demonstrated to be depen- tent depression of the WR evoked from single and repeated electri- dent upon a site above the level of the spinal cord (Collins, cal stimulations. This effect is consistent with a marked spinal 1986). Moreover, progressive dose-dependent central effects of depressant activity of ketamine when added to isoflurane anaes- ketamine have been reported. Psychomimetic symptoms and seda- thesia. This finding substantiates the known MAC-reduction prop- tion in humans (Klausen et al., 1983), plus MAC reduction (Solano erties of ketamine in equids (Muir and Sams, 1992). The non- et al., 2006) and finally general anaesthesia in dogs (Kaka and Hay- invasive method of the WR allowed us to quantify the temporal ef- ton, 1980), were described when plasma concentrations exceed fect of ketamine on the sensory-motor processing in ponies under approximately 0.5, 1 and 3 lg/mL, respectively. The plasma isoflurane anaesthesia and to describe its relation to the measured concentrations found in the present study are compatible with drug plasma concentration. an action of ketamine at supraspinal sites contributing partially to further inhibition of the WR (Rogers et al., 2004). Conflict of interest statement WRs in response to stimulation significantly recovered after cessation of ketamine administration but did not return to baseline None of the authors of this paper has a financial or personal values despite abruptly decreasing ketamine plasma concentra- relationship with other people or organisations that could inappro- tions. The pharmacodynamic effect of a drug is more closely priately influence or bias the content of the paper. related to its concentration at the site of effect (biophase compartment) than in plasma (central compartment). The difference between plasma and effect site drug concentrations explains in Acknowledgments large part the delay between peak effect and maximal plasma concentration of short acting anaesthetics and analgesics (Minto and Racemic ketamine for IV administration was kindly supplied Schnider, 2008). by Dr. E. Gräub AG. This study was funded by Vetsuisse, Switzer- 310 O.L. Levionnois et al. / The Veterinary Journal 186 (2010) 304–311 land and by the Swiss National Science Foundation (analytical Muir 3rd., W.W., Sams, R., 1992. Effects of ketamine infusion on halothane minimal part). alveolar concentration in horses. American Journal of Veterinary Research 53, 1802–1806. Nadeson, R., Tucker, A., Bajunaki, E., Goodchild, C.S., 2002. Potentiation by ketamine References of fentanyl antinociception. I. An experimental study in rats showing that ketamine administered by non-spinal routes targets spinal cord antinociceptive systems. British Journal of Anaesthesia 88, 685–691. Andersen, O.K., Spaich, E.G., Madeleine, P., Arendt-Nielsen, L., 2005. Gradual Ng, K.P., Antognini, J.F., 2006. Isoflurane and propofol have similar effects on spinal enlargement of human withdrawal reflex receptive fields following repetitive neuronal windup at concentrations that block movement. Anesthesia and painful stimulation. Brain Research 1042, 194–204. Analgesia 103, 1453–1458. Arendt-Nielsen, L., Petersen-Felix, S., Fischer, M., Bak, P., Bjerring, P., Zbinden, A.M., Peterbauer, C., Larenza, P.M., Knobloch, M., Theurillat, R., Thormann, W., Mevissen, 1995. The effect of N-methyl-D-aspartate antagonist (ketamine) on single and M., Spadavecchia, C., 2008. Effects of a low dose infusion of racemic and S- repeated nociceptive stimuli: a placebo-controlled experimental human study. ketamine on the nociceptive withdrawal reflex in standing ponies. Veterinary Anesthesia and Analgesia 81, 63–68. Anaesthesia and Analgesia 35, 414–423. Bergadano, A., Andersen, O.K., Arendt-Nielsen, L., Theurillat, R., Thormann, W., Petersen-Felix, S., Arendt-Nielsen, L., Bak, P., Roth, D., Fischer, M., Bjerring, P., Spadavecchia, C., 2009. Plasma levels of a low-dose constant-rate-infusion of Zbinden, A.M., 1995. Analgesic effect in humans of subanaesthetic isoflurane ketamine and its effect on single and repeated nociceptive stimuli in conscious concentrations evaluated by experimentally induced pain. British Journal of dogs. The Veterinary Journal 182, 252–260. Anaesthesia 75, 55–60. Chiba, A., Nakanishi, H., Hiruma, S., Satou, T., Hashimoto, S., Chichibu, S., 1998. Petersen-Felix, S., Arendt-Nielsen, L., Bak, P., Fischer, M., Bjerring, P., Zbinden, A.M., Magnetically induced motor evoked potentials and H-reflex during nembutal 1996. The effects of isoflurane on repeated nociceptive stimuli (central and ketamine anesthesia administration in rats. Research Communications in temporal summation). Pain 64, 277–281. Molecular Pathology and Pharmacology 101, 43–57. Porro, C.A., Cavazzuti, M., Giuliani, D., Vellani, V., Lui, F., Baraldi, P., 2004. Effects of Clarke, R.W., Harris, J., 2004. The organization of motor responses to noxious ketamine anesthesia on central nociceptive processing in the rat: a 2- stimuli. Brain Research Reviews 46, 163–172. deoxyglucose study. Neuroscience 125, 485–494. Collins, J.G., 1986. Effects of ketamine on low intensity tactile sensory input are not Price, D.D., 1972. Characteristics of second pain and flexion reflexes indicative of dependent upon a spinal site of action. Anesthesia and Analgesia 65, 1123– prolonged central summation. Experimental Neurology 37, 371–387. 1129. Richebe, P., Rivat, C., Rivalan, B., Maurette, P., Simonnet, G., 2005. Low doses Dickenson, A.H., 1995. Spinal cord pharmacology of pain. British Journal of ketamine: antihyperalgesic drug, non-analgesic. Annales Francaises Anaesthesia 75, 193–200. d’Anesthésie et de Réanimation 24, 1349–1359. Dimitrijevic, M.R., Nathan, P.W., 1970. Studies in spasticity in man. 4. Changes in Rogers, R., Wise, R.G., Painter, D.J., Longe, S.E., Tracey, I., 2004. An investigation to flexion reflex with repetitive cutaneous stimulation in spinal man. Brain 93, dissociate the analgesic and anesthetic properties of ketamine using functional 743–768. magnetic resonance imaging. Anesthesiology 100, 292–301. Dutton, R.C., Zhang, Y., Stabernack, C.R., Laster, M.J., Sonner, J.M., Eger 2nd., E.I., Sandrini, G., Serrao, M., Rossi, P., Romaniello, A., Cruccu, G., Willer, J.C., 2005. The 2003. Temporal summation governs part of the minimum alveolar lower limb flexion reflex in humans. Progress in Neurobiology 77, 353–395. concentration of isoflurane anesthesia. Anesthesiology 98, 1372–1377. Sarrau, S., Jourdan, J., Dupuis-Soyris, F., Verwaerde, P., 2007. Effects of Dutton, R.C., Cuellar, J.M., Eger 2nd., E.I., Antognini, J.F., Carstens, E., 2007. Temporal postoperative ketamine infusion on pain control and feeding behaviour in and spatial determinants of sacral dorsal horn neuronal windup in relation to bitches undergoing mastectomy. Journal of Small Animal Practice 48, 670– isoflurane-induced immobility. Anesthesia and Analgesia 105, 1665–1674. 676. Ebert, B., Mikkelsen, S., Thorkildsen, C., Borgbjerg, F.M., 1997. Norketamine, the main Scheller, M., Bufler, J., Hertle, I., Schneck, H.J., Franke, C., Kochs, E., 1996. Ketamine metabolite of ketamine, is a non-competitive NMDA receptor antagonist in the rat blocks currents through mammalian nicotinic acetylcholine receptor channels cortex and spinal cord. European Journal of Pharmacology 333, 99–104. by interaction with both the open and the closed state. Anesthesia and Ghorpade, A., Advokat, C., 1994. Evidence of a role for N-methyl-D-aspartate Analgesia 83, 830–836. (NMDA) receptors in the facilitation of tail withdrawal after spinal transection. Scheufler, K.M., Thees, C., Nadstawek, J., Zentner, J., 2003. S(+)-ketamine attenuates Pharmacology, Biochemistry and Behavior 48, 175–181. myogenic motor-evoked potentials at or distal to the spinal alpha-motoneuron. Hirota, K., Lambert, D.G., 1996. Ketamine: its mechanism(s) of action and unusual Anesthesia and Analgesia 96, 238–244. clinical uses. British Journal of Anaesthesia 77, 441–444. Schmid, R.L., Sandler, A.N., Katz, J., 1999. Use and efficacy of low-dose ketamine in Holtman Jr., J.R., Crooks, P.A., Johnson-Hardy, J.K., Hojomat, M., Kleven, M., Wala, the management of acute postoperative pain: a review of current techniques E.P., 2008a. Effects of norketamine enantiomers in rodent models of persistent and outcomes. Pain 82, 111–125. pain. Pharmacology, Biochemistry, and Behavior 90, 676–685. Shimoyama, M., Shimoyama, N., Gorman, A.L., Elliott, K.J., Inturrisi, C.E., 1999. Oral Holtman Jr., J.R., Crooks, P.A., Johnson-Hardy, J.K., Wala, E.P., 2008b. Interaction ketamine is antinociceptive in the rat formalin test: role of the metabolite, between morphine and norketamine enantiomers in rodent models of norketamine. Pain 81, 85–93. nociception. Pharmacology, Biochemistry, and Behavior 90, 769–777. Solano, A.M., Pypendop, B.H., Boscan, P.L., Ilkiw, J.E., 2006. Effect of intravenous Hustveit, O., Maurset, A., Oye, I., 1995. Interaction of the chiral forms of ketamine administration of ketamine on the minimum alveolar concentration of with opioid, phencyclidine, sigma and muscarinic receptors. Pharmacology and isoflurane in anesthetized dogs. American Journal of Veterinary Research 67, Toxicology 77, 355–359. 21–25. Kaka, J.S., Hayton, W.L., 1980. Pharmacokinetics of ketamine and two metabolites in Sonner, J.M., Antognini, J.F., Dutton, R.C., Flood, P., Gray, A.T., Harris, R.A., Homanics, the dog. Journal of Pharmacokinetics and Biopharmaceutics 8, 193–202. G.E., Kendig, J., Orser, B., Raines, D.E., Rampil, I.J., Trudell, J., Vissel, B., Eger 2nd., Kakinohana, M., Motonaga, E., Taira, Y., Okuda, Y., 2000. The effects of intravenous E.I., 2003. Inhaled anesthetics and immobility: mechanisms, mysteries, and anesthetics, propofol, fentanyl and ketamine on the excitability of spinal minimum alveolar anesthetic concentration. Anesthesia and Analgesia 97, 718– motoneuron in human: an F-wave study. Masui 49, 596–601. 740. Kawaguchi, M., Sakamoto, T., Inoue, S., Kakimoto, M., Furuya, H., Morimoto, T., Spadavecchia, C., Spadavecchia, L., Andersen, O.K., Arendt-Nielsen, L., Leandri, M., Sakaki, T., 2000. Low dose propofol as a supplement to ketamine-based Schatzmann, U., 2002. Quantitative assessment of nociception in horses by anesthesia during intraoperative monitoring of motor-evoked potentials. Spine use of the nociceptive withdrawal reflex evoked by transcutaneous electrical 25, 974–979. stimulation. American Journal of Veterinary Research 63, 1551–1556. Kim, J., Yao, A., Atherley, R., Carstens, E., Jinks, S.L., Antognini, J.F., 2007. Neurons in Spadavecchia, C., Arendt-Nielsen, L., Andersen, O.K., Spadavecchia, L., Doherr, M., the ventral spinal cord are more depressed by isoflurane, halothane, and Schatzmann, U., 2003. Comparison of nociceptive withdrawal reflexes and propofol than are neurons in the dorsal spinal cord. Anesthesia and Analgesia recruitment curves between the forelimbs and hind limbs in conscious horses. 105, 1020–1026. American Journal of Veterinary Research 64, 700–707. Klausen, N.O., Wiberg-Jorgensen, F., Chraemmer-Jorgensen, B., 1983. Spadavecchia, C., Arendt-Nielsen, L., Andersen, O.K., Spadavecchia, L., Schatzmann, Psychomimetic reactions after low-dose ketamine infusion. Comparison with U., 2005. Effect of romifidine on the nociceptive withdrawal reflex and temporal neuroleptanaesthesia. British Journal of Anaesthesia 55, 297–301. summation in conscious horses. American Journal of Veterinary Research 66, Knobloch, M., Portier, C.J., Levionnois, O.L., Theurillat, R., Thormann, W., 1992–1998. Spadavecchia, C., Mevissen, M., 2006. Antinociceptive effects, metabolism and Spadavecchia, C., Levionnois, O., Kronen, P.W., Leandri, M., Spadavecchia, L., disposition of ketamine in ponies under target-controlled drug infusion. Schatzmann, U., 2006. Evaluation of administration of isoflurane at Toxicology and Applied Pharmacology 216, 373–386. approximately the minimum alveolar concentration on depression of a Kohrs, R., Durieux, M.E., 1998. Ketamine: teaching an old drug new tricks. nociceptive withdrawal reflex evoked by transcutaneous electrical Anesthesia and Analgesia 87, 1186–1193. stimulation in ponies. American Journal of Veterinary Research 67, 762–769. Larenza, M.P., Landoni, M.F., Levionnois, O.L., Knobloch, M., Kronen, P.W., Theurillat, Spadavecchia, C., Arendt-Nielsen, L., Spadavecchia, L., Mosing, M., Auer, U., R., Schatzmann, U., Thormann, W., 2007. Stereoselective pharmacokinetics of van den Hoven, R., 2007. Effects of butorphanol on the withdrawal reflex ketamine and norketamine after racemic ketamine or S-ketamine using threshold, suprathreshold and repeated subthreshold electrical administration during isoflurane anaesthesia in Shetland ponies. British stimuli in conscious horses. Veterinary Anaesthesia and Analgesia 34, Journal of Anaesthesia 98, 204–212. 48–58. Maitre, P.O., Shafer, S.L., 1990. A simple pocket calculator approach to predict Spadavecchia, C., Levionnois, O.L., Kronen, P.W., Andersen, O.K., in press. The effects anesthetic drug concentrations from pharmacokinetic data. Anesthesiology 73, of isoflurane minimum alveolar concentrations on withdrawal reflex activity 332–336. evoked by repeated transcutaneous electrical stimulation in ponies. The Minto, C.F., Schnider, T.W., 2008. Contributions of PK/PD modelling to intravenous Veterinary Journal (EPub. ahead of print, PMID 19186084). anesthesia. Clinical Pharmacology and Therapeutics 84, 27–38. O.L. Levionnois et al. / The Veterinary Journal 186 (2010) 304–311 311

Theurillat, R., Knobloch, M., Levionnois, O., Larenza, P., Mevissen, M., Thormann, Woolf, C.J., Mannion, R.J., 1999. Neuropathic pain: aetiology, symptoms, W., 2005. Characterization of the stereoselective biotransformation of mechanisms, and management. Lancet 353, 1959–1964. ketamine to norketamine via determination of their enantiomers in equine Woolf, C.J., Max, M.B., 2001. Mechanism-based pain diagnosis: issues for analgesic plasma by capillary electrophoresis. Electrophoresis 26, 3942–3951. drug development. Anesthesiology 95, 241–249. Thomson, A.M., West, D.C., Lodge, D., 1985. An N-methylaspartate receptor- Woolf, C.J., Thompson, S.W., 1991. The induction and maintenance of central mediated synapse in rat cerebral cortex: a site of action of ketamine? Nature sensitization is dependent on N-methyl-D-aspartic acid receptor activation; 313, 479–481. implications for the treatment of post-injury pain hypersensitivity states. Pain Wagner, A.E., Walton, J.A., Hellyer, P.W., Gaynor, J.S., Mama, K.R., 2002. Use of 44, 293–299. low doses of ketamine administered by constant rate infusion as an adjunct Wright, M., Heath, R.B., Wingfield, W.E., 1987. Effects of xylazine and ketamine for postoperative analgesia in dogs. Journal of the American Veterinary on epinephrine-induced arrhythmia in the dog. Veterinary Surgery 16, 398– Medical Association 221, 72–75. 403. Wilson, J., Doherty, T.J., Egger, C.M., Fidler, A., Cox, S., Rohrbach, B., 2008. Effects You, H.J., Morch, D.C., Chen, J., Arendt-Nielsen, L., 2003. Simultaneous recordings of of intravenous lidocaine, ketamine, and the combination on the minimum windup of paired spinal dorsal horn nociceptive neuron and nociceptive flexion alveolar concentration of sevoflurane in dogs. Veterinary Anaesthesia and reflex in rats. Brain Research 960, 235–245. Analgesia 35, 289–296. Woolf, C.J., Chong, M.S., 1993. Preemptive analgesia – treating postoperative pain by preventing the establishment of central sensitization. Anesthesia and Analgesia 77, 362–379.