Ketamine, but Not S  (+)-ketamine, Blocks Ischemic Preconditioning in Rabbit Hearts In Vivo 

The current study conforms to the Guiding Principles in the Care and Use of Animals as approved by the Council of the American Physiologic Society and was approved by the local institutional Animal Care Committee.

After local anesthesia with lidocaine–prilocaine cream, a marginal ear vein was cannulated in 48 male New Zea-land White rabbits weighing 2.6–3.6 kg (mean, 3.0 kg). The animals were anesthetized with intravenous propofol (10 mg/kg) followed by a continuous infusion of α-chloralose 40 mg · kg⁻¹· h⁻¹. The trachea was intubated (endotracheal tube, 3.0 mm ID) and ventilation was controlled using a Starling pump (Type 874/052; Braun Melsungen AG, Melsungen, Germany). Ventilatory frequency was set at 30–35 breaths/min and tidal volume at 15–20 ml to maintain end-expiratory partial pressure of carbon dioxide (Pco²) at approximately 35 mmHg (measured by Datex Capnomac Ultima; Division of Instrumentarium Corp., Helsinki, Finland). Arterial blood gas tensions, hemoglobin concentration, and packed cell volume were assessed at regular intervals (before surgery, before the study, and at the end of the study). Adequacy of the anesthesia regimen was shown by lack of muscle movement and hemodynamic responses during surgery. Additional bolus doses of α-chloralose were administered as needed during surgery.

For measurement of aortic pressure, a 20-gauge catheter was advanced from the left carotid artery into the aortic arch and connected to a Statham transducer (PD23; Gould, Cleveland, OH). After cannulation of the external jugular vein, animals received a continuous infusion of normal saline 15 ml·kg⁻¹·h⁻¹to compensate for fluid loss. After median sternotomy and pericardiotomy, an ultrasonic flow probe was placed around the ascending aorta to measure cardiac output (CO) minus coronary flow volume (4S ultrasonic flow probe, T 208; Transonic Systems Inc., Ithaca, NY). Left ventricular (LV) pressure (LVP) was monitored using a catheter tip manometer (Sensodyn S PO SF-1; Braun Melsungen AG) introduced into the LV via  the left atrium. A ligature snare was passed around a major coronary artery for later occlusion. Temperature was measured inside the pericardial cradle (GTH 1160; Digital Thermometer, Geisinger Electronic, Germany) and maintained between 38.3 and 38.7°C by adjusting a heating pad and an infrared lamp.

Fifteen minutes after completion of the surgical preparation, baseline measurements were performed. All rabbits were subjected to 30 min of coronary artery occlusion by tightening of the snare. The effectiveness of this maneuver was verified by the appearance of epicardial cyanosis and changes in the surface electrocardiogram. Ventricular fibrillation during coronary artery occlusion was treated by electrical defibrillation (5 J, DCS261 Defibrillator; Piekser, Ratingen, Germany). After 30 min of occlusion, the snare occluder was released and reperfusion was verified by the disappearance of epicardial cyanosis. After 2 h of reperfusion, the heart was arrested by injection of 20 ml potassium chloride solution, 16 mm, into the left atrium, quickly excised, and mounted on a modified Langendorff apparatus for perfusion with ice-cold normal saline via  the aortic root at a perfusion pressure of 40 cm H²O to wash away intravascular blood. After 5 min of perfusion, the coronary artery was occluded again and the remaining myocardium was perfused through the aortic root with 0.2% Evans blue in normal saline for 10 min. Intravascular Evans blue was then washed away by perfusion with normal saline for 5 min. This treatment identifies the ischemic-reperfused area (area at risk) as unstained. The heart was then cut into transverse slices, 2 mm thick, and stained for 15 min in buffered 0.75% triphenyltetrazolium chloride solution at 38°C to identify viable and necrotic tissue within the area at risk. The area at risk and the infarcted area were determined by planimetry. The slices were dried and the weight of each slice was measured.

Ten rabbits underwent the ischemia–reperfusion procedure without preconditioning (control group). In the preconditioning group (n = 10), preconditioning was elicited by one 5-min period of coronary artery occlusion followed by 10 min of reperfusion before the 30 min of coronary artery occlusion. In the ketamine groups, ketamine 10 mg/kg (ketamine–preconditioning group, n = 9) or S  (+)-ketamine 10 mg/kg (S  (+)-ketamine–preconditioning group, n = 8) were administered by intravenous bolus injection 5 min before the preconditioning ischemia. Rabbits of the ketamine group (n = 9) received ketamine 10 mg/kg without preconditioning, and animals of the control and the preconditioning group received benzethoniumchloride 0.1 mg/kg as the vehicle of ketamine at the corresponding time point, respectively.

Left ventricular pressure, its first derivative dP/dt, aortic pressure, and stroke volume were recorded continuously on an ink recorder (Recorder 2800; Gould). The data were digitized using an analog-to-digital converter (Data Translation, Marlboro, MA) at a sampling rate of 500 Hz and processed later using a personal computer.

Global systolic function was measured in terms of LV systolic pressure (LVSP) and maximum rate of pressure increase (dP/dtmax). Global LV end-systole  was defined as the point of minimum dP/dt (dP/dtmin), and LV end-diastole  as the beginning of the sharp upslope of the LV dP/dt tracing. The time constant of decrease in LV isovolumic pressure (τ) was used as an index of LV relaxation. 6CO was calculated from stroke volume and heart rate (HR), rate pressure product (RPP) from HR and LVSP, and systemic vascular resistance (SVR) from mean aortic pressure and CO, assuming a right atrial pressure of 0 mmHg in the open chest preparation.

Data are presented as mean and standard deviation (SD). Statistical analysis was performed by two-way analysis of variance for time and treatment (experimental group) effects. If an overall significance between groups was found, a comparison was performed for each time point using one-way analysis of variance followed by the Dunnett post hoc  test, with preconditioning as the reference group. If an overall significance within a group (time effect) was found, one-way analysis of variance followed by the Dunnett post hoc  test, with the baseline value as the reference time point, was used for the assessment of time effects in that group. Changes within and between groups were considered to be statistically significant when the P  value was less than 0.05.

Forty-eight animals were used. Two animals died of ventricular fibrillation during coronary artery occlusion. In the remaining 46 animals, complete data sets were obtained (control group, n = 10; preconditioning group, n = 10; ketamine–preconditioning group, n = 9;S  (+)-ketamine–preconditioning group, n = 8; ketamine group, n = 9).

The hemodynamic variables are summarized in figure 1and table 1. During baseline recordings, there were no significant differences between groups in LVSP, CO, and HR. Consequently, RPP, as a major determinant of myocardial oxygen consumption, was not significantly different between groups during baseline conditions.

The LVSP and CO were reduced after ketamine (LVSP: by 16 ± 18%, P = 0.02 vs.  baseline; CO: by 11 ± 12%, P = 0.7 vs.  baseline) and S  (+)-ketamine application (LVSP: by 17 ± 18%, P = 0.12 vs.  baseline; CO: by 13 ± 11%, P = 0.77 vs.  baseline). The preconditioning ischemia led to a further decrease of CO by 8 ± 12% in the ketamine–preconditioning group and by 12 ± 12% in the S  (+)-ketamine–preconditioning group. In the preconditioning group, the preconditioning ischemia had only a small effect on LVSP (reduced by 4 ± 9%, P = 1.0 vs.  baseline) and CO (reduced by 8 ± 14%, P = 1.0 vs.  baseline). During preconditioning, τ increased (by 21 ± 15% in the preconditioning group [P = 0.79 vs.  baseline], 12 ± 12% in the ketamine–preconditioning group [P = 0.96 vs.  baseline] and 41 ± 18% in the S  (+)-ketamine–preconditioning group [P < 0.01 vs.  baseline]), indicating an impaired diastolic function.

During the 10 min reperfusion period before prolonged ischemia, LVSP recovered to 92 ± 11% in the ketamine–preconditioning group (P = 0.84 vs.  baseline), to 89 ± 13% in the S  (+)-ketamine–preconditioning group (P = 0.73 vs.  baseline), and to 96 ± 9% in the preconditioning group (P = 1.0 vs.  baseline).

In all groups, coronary artery occlusion was accompanied by a reduction of LVSP (by a mean of 17 ± 14%, P < 0.01 vs.  baseline), dP/dtmax (by a mean of 24 ± 21%, P < 0.01 vs.  baseline), and CO (by a mean of 16 ± 19%, P < 0.01 vs.  baseline;table 1and fig. 1). RPP did not significantly differ among groups. With regard to LV relaxation, τ increased by a mean of 18 ± 19% (P < 0.01 vs.  baseline) and LV end-diastolic pressure (LVEDP) by a mean of 75 ± 77% (P < 0.01 vs.  baseline) during coronary artery occlusion (all values at 25 min of coronary artery occlusion). After 2 h of reperfusion, LVSP and dP/dtmax were reduced by 23 ± 19% (LVSP, P < 0.01) and 37 ± 20% (dP/dtmax, P < 0.01) of baseline values in all groups, still reflecting impaired myocardial contractile function in all groups at the end of the experiments. HR decreased by 11 ± 11% in all groups (P < 0.01 vs.  baseline). As a consequence of the reduction in HR and LVSP, RPP was reduced by a mean of 31 ± 22% (P < 0.01 vs.  baseline), with no significant differences between the preconditioning and other groups. Diastolic function also remained depressed during the reperfusion period. τ was prolonged by a mean of 31 ± 33% (P < 0.01 vs.  baseline) and LVEDP remained increased by a mean of 2.3 ± 2.3 mmHg at the end of the experiment (P = 0.09 vs.  baseline).

Mean LV dry weight was 0.77 ± 0.15 g, with no significant differences between groups (data from individual groups are shown in table 2). The ischemic-reperfused area (area at risk) was 0.37 ± 0.14 g and the area at risk constituted 49 ± 16% of the LV, with no significant differences between groups. Preconditioning significantly reduced infarct size from 45 ± 16% of the area at risk (control group) to 24 ± 17% (P = 0.03 preconditioning vs.  control group;fig. 2). Pretreatment with ketamine had no effect on infarct size (48 ± 18%) in animals without preconditioning (ketamine group) but blocked preconditioning, as evidenced by an infarct size of 45 ± 19% in the ketamine–preconditioning group (P = 0.03 vs.  preconditioning group). In contrast, S  (+)-ketamine had no effect on infarct size reduction induced by preconditioning (25 ± 11%;P = 1.0 vs.  preconditioning group).

The main finding of our study is that a single dose of ketamine blocks the protective effect of preconditioning in the rabbit heart in vivo , but S  (+)-ketamine does not. Therefore, the influence of ketamine on preconditioning in vivo  is most likely enantiomer specific.

Variables that are important determinants for development of myocardial infarction are duration of ischemia and myocardial oxygen consumption during ischemia, collateral blood flow toward the ischemic area, and myocardial temperature during ischemia. The rabbit has a consistently small collateral circulation 7and, therefore, it was not necessary to assess collateral blood flow in the ischemic area. Temperature was shown to influence infarct size even in the “normothermic range” to a large extent in the open chest rabbit. 8Therefore, in this study, temperature was measured inside the pericardium and kept within a narrow range of 0.4°C. All experiments were peformed during anesthesia with α-chloralose. This type of anesthesia maintains near-normal cardiovascular reflexes comparable with the awake state and is a classic anesthetic for physiologic and pharmacologic experiments. 9An effect of α-chloralose on preconditioning has not been studied so far.

Ketamine or S  (+)-ketamine 10 mg/kg was administered by intravenous bolus injection 5 min before preconditioning ischemia. In humans, the peak plasma concentration of ketamine is 3–60 μm after intravenous injection of ketamine 2 mg/kg. 10Pedraz et al . 11reported a peak plasma concentration of 63 μm after administration of 10 mg/kg ketamine by intravenous bolus injection in the rabbit. Therefore, the plasma levels of ketamine in the current study can be assumed to be in the same range as those achieved in patients after administration of 2 mg/kg ketamine.

We hypothesized that the effect of ketamine and S  (+)-ketamine on preconditioning is a receptor-mediated mechanism. Therefore, we did not use “equianesthetic” concentrations but the same dose of both drugs to achieve identical plasma concentrations. However, the anesthetic potency of ketamine and S  (+)-ketamine is different, and differences in anesthetic depth might have influenced the severity of myocardial ischemia.

Brief periods of myocardial ischemia followed by reperfusion provide strong endogenous myocyte protection from a subsequent ischemic injury. This concept, known as “ischemic preconditioning,” has been shown to occur in all animals and humans studied. 2Although the signal transduction pathway of preconditioning is not fully understood, overwhelming evidence suggests that the K^ATPchannel is an important component of this phenomena and may serve as an end effector. 12–14Ko et al . 4reported a concentration-dependent inhibitory effect of ketamine on K^ATPchannel activity in isolated cardiomyocytes. Based on this finding, we hypothesized that ketamine might block preconditioning. A single bolus dose of ketamine was sufficient to block the protective effects of preconditioning in the rabbit heart in vivo , whereas the administration of the same dose of ketamine in the absence of preconditioning did not influence infarct size. Ketamine is a 1:1 racemic mixture of two optically active isomers with different effects on cardiac 15and central nervous system receptors. 16In general, stereoselectivity is a function of the potency or affinity of the more potent isomer to specific macromolecular receptors, which are themselves stereospecific. Therefore, we hypothesized that the blockade of preconditioning by ketamine may be stereospecific. Indeed, administration of S  (+)-ketamine before the preconditioning ischemia did not alter the protective effect of preconditioning in vivo  because infarct size in this group was not different compared with the preconditioning group. Therefore, the blockade of ischemic preconditioning by racemic ketamine is most likely caused by the R  (−)-isomer. This is in accordance with the results of Molojavyi et al . in this issue, where R  (−)-ketamine blocked the cardioprotective effects of ischemic preconditioning in isolated rat hearts. 5In the current study, the isolated effects of the R  (−)-isomer were not tested.

Based on our findings, it is not possible to determine the site of action of ketamine on preconditioning. However, the stereoselectivity of this effect suggests a receptor-mediated mechanism. A viable candidate for such a receptor may be the sulfonylurea receptor of the K^ATPchannel. Binding of an agonist prevents opening of this channel during ischemia.

The result of the current study is in accordance with a recent in vitro  study from our labaratory. 5In that study, the effects of ketamine, R  (−)- and S  (+)-ketamine on preconditioning were investigated in isolated rat hearts subjected to 30 min of global ischemia. The improved functional recovery and reduced creatine kinase release seen in that model after preconditioning was blocked by 20 μg/ml but not by 2 μg/ml ketamine. The blockade was caused by the R  (−)-enantiomer, whereas S  (+)-ketamine had no effect on preconditioning. The current study confirms these findings for a single bolus injection of ketamine before preconditioning.

However, in vivo , the pathophysiology of preconditioning and of ischemia–reperfusion is more complex and other effects of ketamine must be considered, i.e.,  hemodynamic side effects, effects on anesthetic depth, or effects on the mechanism of reperfusion injury.

In vitro  data suggests that ketamine and the two optical isomers exert direct cardiac depressive effects in a dose-dependent manner, 15with S  (+)-ketamine inducing less cardiac depression because of an increased availability of catecholamines. 15,17A reduction of myocardial inotropy increases ischemic tolerance. However, it is unlikely that the cardiodepressive effects of ketamine and S  (+)-ketamine contributed to differences in infarct size. Administration of both drugs induced a similar decrease in LVSP and CO, but both parameters recovered to values not significantly different from baseline values before the 30-min ischemia. RPP, a major determinant of myocardial oxygen demand, was not different between the groups during the 30 min of ischemia.

Sympathetic activation caused by surgical trauma may lead to reduced ischemic tolerance. S  (+)-ketamine is a more potent anesthetic agent, 18but a less cardiodepressant drug than is ketamine. 17We observed similar hemodynamic depressant effects caused by S  (+)-ketamine in comparison with ketamine. Therefore, the more cardiodepressant effect of S  (+)-ketamine caused by an increase in anesthetic depth might be “counterparted” by a less direct negative inotropic effect. In addition, stimulation of cardiac adrenergic receptors can induce preconditioning because of the potential link of adrenergic receptors to K^ATPchannels. 19Therefore, we cannot rule out the possibility that an increased catecholamine level induced by S  (+)-ketamine might have contributed to the maintained cardioprotection.

Reperfusion itself can cause additional cellular damage and may increase the extent of myocardial infarction. Some anesthetics interfere directly with the mechanisms of reperfusion injury at the myocardial level. 20,21For ketamine, direct effects on myocardial reperfusion injury have not been investigated. However, postischemic endothelial adherence of neutrophils in the coronary system is a major secondary mediator of reperfusion injury. 22Szekely et al . reported a reduction of postischemic adherence of neutrophils in the coronary system of isolated guinea pig hearts by S  (+)-ketamine, whereas R  (−)-ketamine worsened coronary vascular leak. 23Therefore, a reduction in reperfusion injury by S  (+)-ketamine might have contributed to the infarct size reduction in S  (+)-ketamine–treated animals in vivo .

Several previous studies used ketamine as part of the anesthetic regime for investigating preconditioning, and it is surprising that preconditioning was still possible. Cason et al . 24and Ismaeil et al . 25investigated anesthetic-induced preconditioning in rabbits sedated preoperatively with 70 mg/kg intramuscular ketamine while anesthesia was maintained with intravenous propofol during the remainder of the experiment. Haessler et al . 26used 67 mg/kg intramuscular ketamine for induction of anesthesia followed by a continuous infusion of 20 mg · kg⁻¹· h⁻¹during the experiment. Although the plasma concentrations of ketamine were neither measured in these nor in the current study, it is likely that ketamine plasma concentration was higher in the current study 5 min after intravenous bolus administration of 10 mg/kg ketamine, when preconditioning was elicited.

Walsh et al.  13reported that glibenclamide has antipreconditioning effects in a rabbit model during combination ketamine (1.5 mg·kg⁻¹·min⁻¹) and xylazine (0.75 mg·kg⁻¹·min⁻¹) anesthesia, but not during pentobarbital anesthesia. Thornton et al.  27found that the K^ATPchannel blocker glibenclamide was proischemic, but did not eliminate preconditioning in rabbits anesthetized with pentobarbital. In contrast, Toombs et al.  28reported that glibenclamide blocked preconditioning in a nearly identical rabbit model, with the exception that anesthesia was maintained with repeated intramuscular injections of ketamine (10 mg/kg) and xylazine (2 mg/kg). These seemingly conflicting results were explained by the study of Morita et al. , 29who reported that ketamine anesthesia enhanced the blocking effect of glibenclamide on K^ATPchannels. Taken together, these findings suggest that the glibenclamide dose used in some of the studies only blocked preconditioning in addition with ketamine as a basal anesthetic. The current study showed also a direct blocking effect of ketamine on preconditioning in a similar rabbit model. The role of anesthesia may be critical in determining, evaluating, and reevaluating the results of studies that investigate myocardial preconditioning.

Preconditioning is still a laboratory-based phenomena that has not been conclusively documented in patients. However, some in vitro  evidence 2of preconditioning in humans exists, and there are several possible clinical scenarios in which preconditioning might occur, e.g. , unstable angina preceding myocardial infarction, percutaneous transluminal coronary angioplasty, and warm-up angina. It is therefore of great practical interest to avoid the use of anesthetic agents that block this strongest endogenous cardioprotective mechanism against myocardial ischemia. In contrast to the detrimental effects of K^ATPchannel blockade, use of K^ATPagonists, including anesthetics such as isoflurane 14or opiods, 30is suggested to provide organ protection intraoperatively during cardiac, vascular, or neurologic surgery. However, clinical evidence for the role of anesthetic agents in improving morbidity and mortality in patients at risk for ischemic or hypoxic injury is lacking.

In summary, the current data indicate that a single bolus dose of ketamine blocks the cardioprotective effect of preconditioning in vivo , but S  (+)-ketamine does not. Therefore, the influence of ketamine on preconditioning is most likely enantiomer specific.

The authors thank T. Wietschorke, cand.med., Institut für Herz- und Kreislaufphysiologie, for excellent technical assistance.