Effects of Anesthesia and Recovery from Ketamine Racemate and Enantiomers on Regional Cerebral Glucose Metabolism in Rats

Experiments were performed in 3- to 4-month-old (average weight, 225–275 g) male Fisher-344 rats (Charles River Italia, Como, Italy). The study was conducted according to a protocol approved by the Animal Care Committee of the Anesthesiology Department of Padova University (Italy) and the guidelines for animal care of the National Institutes of Health. 28[¹⁴C]2-deoxy-d-glucose (specific activity, 50 –55 mCi/ mmol) was purchased from Amersham International (Arlington Heights, IL). Ketamine racemate and enantiomers were kindly provided by Parke Davis (Freiburg, Germany) and were dissolved in saline (1 ml/kg body weight).

Drug doses, timing, and routes of administration for the [¹⁴C]2-deoxy-d-glucose study were chosen on the base of preliminary behavioral studies. To study rCMRglc during anesthesia, doses of ketamine racemate and enantiomers were sought that ensured a deep anesthesia for the entire duration of the [¹⁴C]2-deoxy-d-glucose experiment, i.e. , 50 min, and prevented confounding from postanesthetic behavioral activation. To study rCMRglc during recovery from anesthesia, doses of ketamine racemate and enantiomers were sought that induced an anesthesia of similar duration and minimized potential confounding from different anesthetic potencies. The righting reflex was used as an index of duration of anesthesia. 12,29A single observer evaluated reflex changes after intraperitoneal (dose range, 80 –170 mg/ kg) or intravenous (dose range, 10 –100 mg/kg) administration of ketamine to separate groups of four rats each. After drug injection, rats were positioned in a dorsal recumbent position every minute or until the righting reflex was lost, i.e. , the animals remained in the recumbent position for at least 30 s; animals were then observed until the righting reflex was regained, i.e. , animals returned to their normal prone position. Times elapsing from drug administration to loss and to regaining of the righting reflex were recorded and considered as the onset and end of anesthesia. The lowest dose of the least active agent, i.e. , R  (−)-ketamine, that was required for the anesthesia to endure for the [¹⁴C]2-deoxy-d-glucose experiment corresponded to an intraperitoneal dose of 170 mg/kg. Such an anesthesia could not be achieved by bolus intravenous injection of ketamine without significant toxicity. An anesthesia of similar duration was instead obtained most reliably (i.e. , least intergroup variations) with intravenous doses. Hence, to study rCMRglc during recovery, we chose doses of 20, 12.5, and 30 mg/kg S , R  (±)-ketamine, S  (+)-ketamine, or R  (−)-ketamine, which had also been used in previous behavioral studies. 29 

In the pilot study, after an intraperitoneal equimolar dose of 170 mg/kg S , R  (±)-ketamine, S  (+)-ketamine, or R  (−)-ketamine, rats became rapidly anesthetized and remained so for 106 ± 9, 111 ± 10, and 57 ± 5 min (mean ± SD); anesthesia by R  (−)-ketamine was significantly (P < 0.01) shorter than by congeners but exceeded the duration of the [¹⁴C]2-deoxy-d-glucose experiment. After equianesthetic doses of 20, 12.5, and 30 mg/kg S , R  (±)-ketamine, S  (+)-ketamine, and R  (−)-ketamine, rats were anesthetized for 12 ± 2, 11 ± 3, and 10 ± 3 min (mean ± SD) but were fully awake by 20 min. Hence, anesthesia times from the pilot study indicated that S , R  (±)-ketamine, S  (−)-ketamine, and R  (+)-ketamine differed in anesthetic potency with a ratio that was approximately 2:1.2:3 at low doses and approached to 2:2:1 at higher doses. At the end of anesthesia, S , R  (±)-ketamine– and R  (−)-ketamine–treated rats presented ataxia, hyperreactivity to sensory stimuli (i.e. , noise and touch), and intense stereotyped movements (not quantified); S  (+)-ketamine–treated rats were awake but not agitated.

In the rCMRglc study, four groups of rats were injected intraperitoneally with saline or 170 mg/kg S , R  (±)-ketamine, S  (+)-ketamine, or R  (−)-ketamine, 10 min before [¹⁴C]2-deoxy-d-glucose, and four groups of rats were injected intravenously with saline or 20, 12.5, or 30 mg/kg S , R  (±)-ketamine, S  (+)-ketamine, or R  (−)-ketamine, 20 min before [¹⁴C]2-deoxy-d-glucose.

Rats were allowed free access to water until the day of the experiment, and food was withheld the night before to provide steady plasma glucose concentration. Under 1.5% isoflurane anesthesia, polyethylene catheters (PE-50) were introduced into a femoral artery (for blood pressure measurements and collection of arterial blood samples) and a vein (for injection of drug and of the isotope). The incision site was infiltrated with 0.2 ml bupivacaine hydrochloride, 0.25%. Then, a loose-fitting plaster cast was applied around the lower abdomen, and the rats were allowed to recover for 3 h in a temperature-controlled, sound-insulated wooden box.

Throughout the experimental period, arterial blood pressure and heart rate were measured by connecting the arterial catheter with a pressure transducer, and body temperature was measured with a rectal thermo-probe; both were connected to a monitor (model PM-2A; Honeywell, Minneapolis, MN). Arterial oxygen tension (PaO²) and arterial carbon dioxide tension (PaCO²) were determined before and 40 min after [¹⁴C]2-deoxy-d-glucose administration. A thermostatic device activated an ambient heating element when body temperature decreased below 35.5°C. rCMRglc was quantitatively determined as described previously. Briefly, a bolus of 125 μCi/kg [¹⁴C]2-deoxy-d-glucose was injected. Timed arterial samples were drawn to measure plasma glucose (Glucose Analyzer II; Beckman Instruments, Fullerton, CA) and [¹⁴C]2-deoxy-d-glucose concentrations (Liquid Scintillation Spectrometer model B2450; Packard, Downers Grove, IL). At the end of the 45-min period, the rats were killed by an overdose of thiopental, and the brains were removed, frozen in isopentane at −50°C, and cut in a cryostat (Cryocut E; Reichert-Jung, Heidelberg, Germany) maintained at −22°C. Coronal sections (20 μm thick) were exposed to Kodak SB-5 films (Eastman Kodak Company, Rochester, MN) for 7 days, together with a set of precalibrated ¹⁴C-methacrilate standards (Amersham International). Local tissue ¹⁴C concentrations were determined in 40 brain regions using a computer-based densitometer (BRS2 System; MCID, Ontario, Canada) comparing the optical densities of the autoradiographic sections with those of the calibrated [¹⁴C]2-deoxy-d-glucose standards. Twelve determinations were made for each region in the left and right sides of the brain, and the means were averaged. Mean rCMRglc values from all brain regions investigated were calculated. Each anatomic brain region evaluated was defined by comparison with an atlas of the rat brain. 30rCMRglc values were calculated from brain and plasma radioactivities and plasma glucose concentrations according to equations and constants given by Sokoloff et al.  23,24 

Data were analyzed for statistical significance by one-way analysis of variance and the Bonferroni multiple comparison test. Mean values of physiologic parameters and rCMRglc of each ketamine-treated group were compared with controls; the mean times of ketamine enantiomer anesthesia were compared with those of S , R  (±)-ketamine anesthesia. Statistical significance was taken in all cases as P < 0.01.

In comparison with control rats, R  (−)-ketamine anesthesia increased heart rate significantly (P < 0.01) by 3 min (i.e. , 440 ± 6 beats/min, +18%) and decreased heart rate and mean arterial blood pressure by 12 min (i.e. , 263 beats/min, −28%; 91 mmHg, −19%) and 30 min (i.e. , 272 beats/min, −27%; 82 mmHg, −18%) after [¹⁴C]2-deoxy-d-glucose administration. During recovery from S , R  (±)-ketamine, S  (+)-ketamine, and R  (−)-ketamine, arterial blood pressure increased significantly (P < 0.01) by 3 min (i.e. , 154, 156, and 146 mmHg; +26, 27, and 21%) and 12 min (i.e. , 140, 135, and 133 mmHg; +24, 19, and 18%) after [¹⁴C]2-deoxy-d-glucose administration. Reportedly, ketamine has predominant stimulatory effects on cardiovascular parameters 1,6that were apparent in this study during recovery; when a deeper anesthesia is achieved, however, cardiovascular parameters are more commonly depressed. 6 

S , R  (±)-ketamine and S  (+)-ketamine anesthesia significantly (P < 0.01) decreased rCMRglc in 22 and 14 brain areas (average reductions, 23 and 15%) and increased rCMRglc in two limbic regions (table 1). R  (−)-ketamine anesthesia decreased rCMRglc in 11 and increased rCMRglc in 14 brain regions (average decrease, 4%) (fig. 1). Ketamine racemate and enantiomer anesthesias all decreased rCMRglc in the sensorimotor and auditory cortices, the superior and inferior colliculi, the interpositus, the cochlear and vestibular nuclei, the medial and lateral habenula, the median raphe, and the mesencephalic reticular formation; in addition, anesthesia with S , R  (±)-ketamine and S  (+)-ketamine decreased rCMRglc in the frontal, visual, and posterior cingulate cortices; anesthesia with S , R  (±)-ketamine and with R  (−)-ketamine decreased rCMRglc in the cerebellar vermis; anesthesia with S , R  (±)-ketamine decreased rCMRglc in the mediodorsal thalamic nucleus, the medial and lateral geniculate nuclei, the substantia nigra and the subthalamic nucleus, the anterior cingulate cortices, and the locus ceruleus; anesthesia with S  (+)-ketamine decreased rCMRglc in the interpeduncular nucleus; and anesthesia with R  (−)-ketamine decreased rCMRglc in the pontine reticular formation (fig. 2). Ketamine racemate and enantiomer anesthesias all increased rCMRglc in the dentate gyrus; S , R  (±)-ketamine anesthesia increased rCMRglc in the fasciculus retroflexus; S  (+)-ketamine anesthesia increased rCMRglc in the lateral septum; and R  (−)-ketamine anesthesia increased rCMRglc in the basal ganglia (i.e. , caudate-putamen and globus pallidus nuclei) and the limbic regions (i.e. , accumbens nucleus, anterior cingulate and entorhinal cortices, presubicular and hippocampal fields, and lateral amygdala).

Postanesthetic recovery from S , R  (±)-ketamine, S  (+)-ketamine, and R  (−)-ketamine anesthesia increased rCMRglc significantly (P < 0.01) in 28, 16, and 20 cortical, basal ganglia, limbic, and brainstem regions (average increases, 35, 13, and 20%). In addition, recovery from S , R  (±)-ketamine decreased rCMRglc in 1 brain region (i.e. , inferior colliculus), recovery from S  (+)-ketamine decreased in 6 brain regions (i.e. , inferior colliculus, medial geniculate and auditory cortex, posterior cingulate cortex and habenular nuclei), and recovery from R  (−)-ketamine decreased in 5 brain regions (i.e. , sensorimotor cortex, medial geniculate and inferior colliculus, globus pallidus and posterior cingulate cortex) (fig. 2).

Ketamine anesthesia decreased and postanesthetic recovery increased rCMRglc. S , R  (±)-ketamine and S  (+)-ketamine determined deep anesthesia and rCMRglc depression, and S , R  (±)-ketamine and R  (−)-ketamine determined large postanesthetic rCMRglc activations.

Until now, ketamine anesthesia had induced a pattern of cerebral metabolic decreases in few sensory areas and large metabolic increases in basal ganglia and limbic regions. 3,8–10Such findings on ketamine were in contrast with reports of diffuse cerebral hypometabolism during general anesthesia by intravenous or gaseous anestetics. 2,4–6Cerebral hypometabolism, however, attenuates during postanesthetic recovery in the case of most anesthetics but reverses into a marked cerebral hypermetabolism in the case of ketamine. 9Therefore, a possible cerebral hypometabolism during ketamine anesthesia could have been overshadowed by metabolic activation of an initial recovery whose onset is a critical methodologic issue when ketamine anesthesia is being studied. Duration and depth of ketamine anesthesia are dose dependent, 7–9and in previous studies, ketamine doses may have been barely sufficient or insufficient to ensure an anesthetic state for the entire [¹⁴C]2-deoxy-d-glucose experiment. 7–9In this study, large doses of S , R  (±)-ketamine induced a long-lasting anesthesia and widespread rCMRglc decreases in the neocortical, sensory thalamic, and brainstem regions and rCMRglc increases in the dorsal hippocampus. This rCMRglc pattern is consistent with animal and human reports of diffuse rCMRglc decreases and few rCMRglc increases during intravenous or gaseous anesthesia. 3–5Specifically, benzodiazepine and propofol anesthesia determined rCMRglc decreases only. 5,6,10Barbiturate and halogenated gas anesthesia determined widespread rCMRglc decreases but also rCMRglc sparing or increase in few brain regions (i.e. , hippocampal regions, habenulointer-peduncular and brainstem nuclei). 3,4,6Hence, rCMRglc increases in limbic regions have not been observed universally during general anesthesia and likely reflect the property of some anesthetics, 6including ketamine, 2of increasing electrical activity of hippocampal neurons.

S  (+)-ketamine anesthesia decreased cerebral metabolism to a larger extent than R  (−)-ketamine but less than S , R  (±)-ketamine anesthesia. Because S  (+)-ketamine is known to be approximately twofold to fivefold more potent as an anesthetic than both R  (−)-ketamine and S , R  (±)-ketamine, 12the first finding was expected, but the latter was less so. S  (+)-ketamine is cleared from the bloodstream at the same 29or higher 31rates than R  (−)-ketamine; therefore, larger rCMRglc decreases by S  (+)-ketamine anesthesia do not depend on kinetic factors. In addition, we chose a high dose that ensured R  (−)-ketamine anesthesia to last for the entire [¹⁴C]2-deoxy-d-glucose experimental without the confounding of recovery. To do so, however, we used a behavioral indicator (i.e. , loss of righting reflex) that may not adequately inform of anesthetic depth. In spite of an apparently identical anesthetic state, in fact, R  (−)-ketamine anesthesia slowed electroencephalographic activity to a substantially lesser degree than S , R  (±)-ketamine and S  (+)-ketamine anesthesia in animals 13and humans. 2Consistently, our study indicates that R  (−)-ketamine anesthesia does not reach the marked and widespread rCMRglc depression that is typical of more potent general anesthetics 3–6and congeners. In contrast and unexpectedly, S  (+)-ketamine anesthesia determined smaller rCMRglc decreases than S , R  (±)-ketamine, which is supposedly less potent and is being used clinically in larger doses. 1Data in the literature, however, are not completely consistent with this notion. In rat behavioral studies, S  (+)-ketamine was more potent as an anesthetic at low-intermediate doses (i.e. , below 60 mg/kg) 12,29but equivalent with S , R  (±)-ketamine at doses higher than 80 mg/kg. 12Similarly, in our dose-finding study, S  (+)-ketamine anesthesia had a longer duration than S , R  (±)-ketamine anesthesia at low doses but a similar duration at higher doses. Further, in neurophysiologic studies, S  (+)-ketamine was equipotent with S , R  (±)-ketamine in slowing electroencephalographic activity in cats 12and was stronger 32or equivalent 33with S , R  (±)-ketamine in humans. Hence, the current data suggest that the anesthetic potency ratio between S  (+)-ketamine and S , R  (±)-ketamine is probably not linear and that, at high doses, S  (+)-ketamine may be equivalent with or even weaker than S , R  (±)-ketamine.

Postanesthetic recovery from ketamine racemate and enantiomers determined dramatic rCMRglc increases in most neocortical, thalamic, limbic, and basal ganglia regions and rCMRglc sparing or decrease in few areas. Patterns of rCMRglc during postanesthetic recovery from ketamine are virtually indistinguishable from those of low-dose ketamine 6,8,9or subanesthetic phencyclidine and MK801, 34two other noncompetitive NMDA antagonists, and suggest that they all may depend on low NMDA antagonistic activity. Intriguingly, postanesthetic patterns also resemble those of glutamate agonists 6and may reflect a paradoxical increase of glutamatergic neurotransmission. At subanesthetic doses (i.e. , 10 –30 mg/kg intraperitoneally), in fact, ketamine increases extracellular brain glutamate, 35which in turn may activate non-NMDA glutamate receptors and unblocked NMDA receptors and may result in a net increase of glutamatergic neurotransmission. 35Hence, behavioral and rCMRglc activation during recovery likely reflect ketamine concentrations abating to subanesthetic levels 36and a subsequent brain glutamate increase. Postanesthetic rCMRglc activation was smaller after S  (+)-ketamine than after S , R  (±)-ketamine or R  (−)-ketamine, and these findings contrast with neuroimaging and behavioral reports of more pronounced rCMRglc increases and side effects by S  (+)-ketamine than by R  (−)-ketamine in humans 32; they are consistent instead with reports of larger behavioral activation by R  (−)-ketamine in animals 12,13and suggest possible interspecies differences.

Reportedly, ketamine binds to NMDA, 14μ- and κ-opioid, nicotinic, muscarinic, dopamine D², and serotonin S²receptors; ketamine anesthesia increases brain extracelular concentrations of norepinephrine, 37dopamine, 38and acetylcholine 39and decreases those of serotonin, 40glutamate, 35and the endogenous opioid metenkephalin 41but has no effect on brain γ-aminobutyric acid concentrations and receptors. Hence, ketamine anesthesia markedly perturbs several neurochemical parameters, which, in turn, may be involved in ketamine anesthesia. However, the relation between the dose of ketamine and behavioral and rCMRglc effects (i.e. , activation at low doses, inhibition at high doses) differs from those of aminergic agonists and antagonists 25–27; hence, neurochemical changes may be not relevant for or may be obscured by ketamine anesthesia. Ketamine-induced rCMRglc changes are instead similar during anesthesia to those of general anesthetics and during recovery to those of noncompetitive NMDA antagonists such as MK801 and phencyclidine, which also possess anesthetic properties. 42Some reports contrast with an NMDA hypothesis of ketamine anesthesia: NMDA receptors are already blocked at nonanesthetic doses 43; NMDA-induced swallowing responses are not blocked at anesthetic doses 44; and some competitive NMDA antagonists, such as 3-((R)-2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid, may be not anesthetic. 45On the other hand, ketamine interacts with NMDA receptors in fashions that are not shared by other NMDA antagonists; ketamine anesthesia is stereospecifically counteracted by NMDA 6; and several noncompetitive or functional NMDA receptor antagonists, including MK801, dextrorphan, (+)-N-allylnormetazocine, and riluzole, increase the potency of general anesthetics and have analgesic and anesthetic properties themselves. 45Evidence also links general anesthetics to an inhibition of glutamatergic neurotransmission: enflurane blocks glutamate-induced MK801 binding to NMDA receptors 46; halothane and isoflurane block glutamate- and NMDA-induced intraneuronal entry of calcium 47; and ketamine blocks non-NMDA, glutamatergic receptor second messenger. 48Finally, it has been recently reported that glutamatergic synaptic transmission and rCMRglc have a tight stoichiometric relation, and rCMRglc would be a direct index not only of neuronal metabolism but also of glutamate neurotransmission. If this holds true, a diffuse cerebral hypometabolism may be a common feature of gaseous and intravenous anesthesia that reflects a diffuse depression of glutamatergic neurotransmission.