Cognitive Impairment after Small-dose Ketamine Isomers in Comparison to Equianalgesic Racemic Ketamine in Human Volunteers

Twenty-four healthy young adults (American Society of Anesthesiologists physical status I; 7 women, 17 men; age, 25 ± 3 yr; height, 178 ± 9 cm; weight, 70 ± 12 kg) served as paid volunteers. After Human Subjects Committee approval (Ulm University, Ulm, Germany) was obtained, all participants granted written, informed consent. They were free to withdraw at any time. Exclusion criteria included known idiosyncrasies, allergy to ketamine, pregnancy, substance abuse, or use of any medication except oral contraceptives within 2 weeks before the study. Volunteers with a history of emotional difficulty, psychiatric illness, or dementia in first-degree relatives were excluded.

To minimize confounding factors by lack of experience and to achieve a ceiling effect by learning from repetition, 6participants were exposed to three preliminary test series.

All subjects completed a three-session, randomized, double-blind, crossover protocol comparing racemic ketamine, S  (+)-ketamine, and R  (−)-ketamine. To ensure drug clearance from the body to avoid carryover effects, sessions were spaced by 1-week intervals. The night before the study, volunteers were instructed to fast after midnight. On the study day, participants were placed in a quiet room of the departmental step-down unit. They were monitored with a three-lead electrocardiograph, a peripheral pulse oximeter, and an automated blood pressure device. In the presence of an anesthetist who monitored subjects’ well-being, a catheter was inserted in the nondominant arm to administer intravenous fluid and drugs, and another catheter was placed into the contralateral elbow to collect blood samples. Once all devices had been instituted, a 20-min resting period was provided. Predrug baseline readings were then obtained. Ketamine was administered intravenously over 30 s. To control for diurnal variations, subjects received each study drug at the same time of day. The volunteers, the anesthetist, and the persons providing the tests were blinded to the respective drug applied. Extraordinary events were noted. Subsequent to complete recovery, but not earlier than 6 h after tests, subjects were considered ready for discharge when they had attained alertness, normal vital signs, and motor function. Figure 1shows the study timeline and measurements.

Ketamine was administered intravenously at 0.5 mg/kg racemic ketamine, 0.25 mg/kg S  (+)-ketamine, and 1 mg/kg R  (−)-ketamine. This regimen was based on doses used in analgesic therapy and the potency relation of ketamine isomers. 1–5,9–12Equianalgesic efficacy of the 1:0.5:2 dose ratio of racemic:S  (+)-:R  (−)-ketamine used was confirmed by measuring pain tolerance thresholds in pilot studies (data not shown). Ampoules containing the drugs at equivolumetric amounts (chemical and optical purities > 98.5%) were provided by Parke-Davis Pharmaceuticals (Freiburg, Germany).

Venous plasma concentrations of (nor-)ketamine and (nor-)epinephrine concentrations were measured by high-performance liquid chromatography with ultraviolett 14or electrochemical detection, 15as described.

Before and 20 min after drug injection, participants’ mood was assessed using visual analog scales. 16,17At the end of each study at 60 min, subjects were asked about feelings they recalled retrospectively for the study session. The sets of adjectives presented were as follows: afraid–brave, tense–relaxed, agitated–calm, happy– sad, lethargic–quick-witted, alert–drowsy, well–poorly concentrated, detached–moved, depressed–euphoric, peaceful–aggressive, small–big, interoceptive emptiness–eventfulness, lucky–unlucky, threatened–superior, out-of-control–inhibited, teared down–high floating. The differences between baseline and 20-min scores and between baseline and 60-min scores were calculated. The 16 items were additionally categorized into four clinically derived functional clusters consisting of mood, affect–emotion, cognitive capacity, and ego perception. 16,17Changes between evaluation time points were calculated by analyzing differences in clustered scores (see  Appendixfor details).

To assess short-term memory, a subtest of the German Hamburg-Wechsler Intelligence Scale for Adults was used (similar to Digit Span used in the United States). 18,19Three unrelated digits were read to subjects with a rate of 1 s per number. Immediately after presentation, subjects were asked to repeat the digit series in the order given. If performed correctly, a new sequence was generated by increasing the digit number by one digit up to a final of nine digits. The last series recalled with digits in their proper position was scored using an eight-point scale (0 = no recall, 7 = correct recall of nine digits).

To assess the ability to form new memory, subjects were shown 1 of 10 memory cards with simple line sketches of a symbol, a letter, or a number at each evaluation point. 20A different card was presented for 5 s at baseline and 5, 10, 15, 30, and 45 min after drug injection in random order. At 60 min, subjects were asked to recall spontaneously the figures shown. The number of correctly recalled sketches was counted for each subject.

Psychomotor coordination was studied using a test of the procedure of Newman et al.  21( Appendix).

Participants were asked to count the appearance of a target symbol out of groups of symbols as quickly as possible. 22The time required for the task was scored (see  Appendix).

Based on analytical principles that apply in information psychology, several variables of central information processing were studied.

The time required for reading 20 independent characters presented in random order was noted. 23,24Central Information Flow, 23–26Retention Time in Primary Memory, 23,24,26Short-term Storage Capacity, 23,24,26and Intelligence Quotient 23,25–27were calculated as reported previously ( Appendix).

After completion of each session and 24 h after the tests, subjects were asked for adverse events, including nausea, dizziness, and unusual sensations. The occurrence and quality of “dreams,” hallucinations (altered consciousness states), and other phenomenon was noted.

After completion of each session, subjects and examiners noted their personal assessment of which of the three drugs they thought had been examined in the respective study session.

Where appropriate, data are reported as mean ± SD after verification of normal distribution with the Kolmogorov-Smirnov test. For statistical analysis, the Friedman test was used for comparison between all three drugs. In case of significance, the Wilcoxon signed rank post hoc  test with Bonferroni correction was applied to further analyze differences between two drugs. The anterograde amnesia data are reported as counts. They were evaluated by contingency table analysis over all three agents, followed by chi-square test between two drugs. P  values (two-tailed) < 0.05 were considered significant. The software programs Statistical Package for the Social Sciences (SPSS Inc., Chicago, IL) and StatView 4.5 (Abacus Concepts, Berkeley, CA) were used for the analysis.

Blood pressure and heart rate increases were comparable within 1 min after injection of all drugs (fig. 2A), confirming hemodynamic stimulation after ketamine. 10,14The dose ratio of 1:0.5:2 of racemic ketamine:S  (+)-ketamine:R  (−)-ketamine was mirrored in parallel increases and decreases in ketamine and (nor)-ketamine plasma concentrations (fig. 2B). All drugs induced a similar, small transient increase in norepinephrine and epinephrine plasma concentrations, with peaks at 5 min followed by a gradual decline over the course of time (fig. 2C).

The most prominent changes in mood and individual psychic items are shown in table 1. Subjects who received the S  (+)-isomer rated their concentration to be less impaired (P = 0.024) and felt more brave (P = 0.046) than after racemic ketamine at 20 min. They indicated they were less lethargic (P = 0.014) but more out-of-control (P = 0.046) after R  (-)- than racemic ketamine. Subjects rated themselves as less drowsy after R  (−)- (P = 0.002) or S  (+)-ketamine (P = 0.044) and as less lethargic after R  (−)- (P = 0.010) or S  (+)-ketamine (P = 0.042) than racemic ketamine after 60 min. Whereas cluster analysis for mood, affect–emotion, and ego perception (data not shown) revealed no difference between drugs, clustered cognitive capacities were less limited after R  (−)- (P = 0.040) or S  (+)-ketamine (P = 0.042) than racemic ketamine after 60 min (fig. 3).

Immediate digit recall was impaired until 10 min after drug injection (fig. 4). The most marked, 50% decrease occurred at 1 min after administration of racemic ketamine. After administration of S  (+)- (P = 0.010) or R  (−)-ketamine (P = 0.020), digit recall was less limited than after racemic ketamine at 1 min. Anterograde amnesia was less pronounced after S  (+)- (P = 0.009) or R  (−)-ketamine (P = 0.028) than after racemic ketamine at 5 min after injection (fig. 5).

Performance in Trieger test deteriorated after administration of all three drugs, with markedly poor values at 1 and 5 min after ketamine injection (fig. 6A). Test scores returned to predrug levels within 30 min. The changes caused by R  (−)-ketamine almost paralleled those induced by S  (+)-ketamine. Concentration capacity analysis showed an early, transient ketamine-related decrease and then a return to predrug values within 30 min (fig. 6B). Concentration capacity declined less after S  (+)-ketamine than either racemic (P = 0.010) or R  (−)-ketamine (P = 0.022) at 1 min after drug injection.

All drugs induced a rapid decrease in information processing that was most prominent at 1 and 5 min after injection (table 2). This was followed by a gradual return to predrug levels, which were reached after approximately 30 min. Speed reading was markedly reduced at 5 min. The 100% increase in reading time after racemic ketamine was higher than the 50% increase after the S  (+)-isomer at 5 min (P < 0.001). Central information flow declined to a smaller extent after S  (+)- (30%;P = 0.018) than racemic ketamine (40%) at 5 min. Retention in primary memory was less impaired after S  (+)-ketamine than either racemic (P = 0.040) or R  (−)-ketamine (P = 0.044) at 1-min study time. At 5 min, retention in primary memory showed a greater decrease after racemic than S  (+)- (P = 0.010) or R  (−)-ketamine (P = 0.022). Short-term storage capacity was more limited after racemic than S  (+)- (P = 0.003) or R  (−)-ketamine (P = 0.008) at 5 min. Figure 6Cillustrates the marked decrease in intelligence quotients early after drug injection. S  (+)- (P = 0.002) or R  (−)-ketamine (P = 0.006) induced a smaller reduction in the intelligence quotient than racemic ketamine at 5-min study time.

Six subjects indicated “dreams” after R  (-)- and racemic ketamine, and five after S  (+)-ketamine. There was no difference in dream quality between drugs. For the test and 6 h after the study observation period, two female subjects reported feelings of helplessness with death threats after R  (-)-ketamine; they wanted to withdraw from further study. One female had a “bad nightmare” after racemic ketamine; she experienced that session as a very undesirable event. Mild sensory illusions occurred comparably after all drugs during the 60-min study period. In the 24-h follow-up interview, approximately half of all subjects stated unusual dreams following the night after the study. No recurrent or residual events were reported.

Neither subjects nor investigators were able to discriminate between drugs. The best result with 49% correct guess was observed for S  (+)-ketamine based on the investigators’ opinions (data not shown).

Our study is the first to disclose divergent effects of nonequimolar, equianalgesic small-dose S  (+)-ketamine, R  (−)-ketamine, and racemic ketamine on mood and intellectual functions in volunteers: early after intravenous injection, subjects felt less tiredness and less intellectual decline after ketamine isomers than racemic ketamine, and for a few minutes, subjects who received S  (+)-ketamine also showed less impairment in concentration capacity and primary memory. Although several mechanisms may account for these effects, to the best of our knowledge, there is no direct explanation for the different responses to the drugs.

Our data extend early work on human pharmacology of racemic ketamine and its isomers. 8,13After equianesthetic doses of the three drugs, patients treated with S  (+)-ketamine felt less disorientation, fear, and pain in the postoperative period and showed a more general sense of well-being and happiness. 13In volunteers, White et al.  8found more rapid recovery of psychomotor skills after high-dose ketamine isomers than racemic ketamine, and Arendt-Nielsen et al.  12reported shorter reaction times after equianalgesic dose S  (+)- than racemic ketamine at plasma concentrations of 180 or 350 ng/ml, respectively.

There is no universally accepted analysis of cognitive decline. A spectrum of psychological tests to screen drug effects was therefore selected. Although our paradigm was compiled of relatively simple, paper–pencil measures with moderate sensitivity to discriminate between cognitive domains, 28reliability and validity of the tests used has been shown in studies of drug-related psychic events 6,7,13,29and neurocognitive decline. 8,13,18–22,26–29Analysis of information processing and psychological intelligence may be considered as “rule of thumb” estimates, but they are used to assess context-related behavioral performance 18,22–27,30–32based on neuroanatomy and neurotransmitter disposition.

Fitting a two-compartment open model, 10,14ketamine distribution and plasma elimination shows an α half-life of approximately 10 min and a β phase of 1–2 h after intravenous use. We therefore studied a 60-min time period after one drug injection, where humans experience a maximum of brain effects. 33For data interpretation, it should be kept in mind that the differences we observed may relate to specific differences in drug potencies, pharmacokinetic profiles, pharmacodynamic characteristics, or a combination of these mechanisms.

Our choice of dose regimen was guided by analgesic plasma concentrations of racemic ketamine, which range within 100–200 ng/ml. 1–5,10There is no way to know that the doses studied are equipotent for nonanalgesic effects of ketamine. Inhibition of excitatory N -methyl-d-aspartate (NMDA) glutamate receptors is believed to be the predominant mode of action of small-dose ketamine, which has been related to substantial blockade of NMDA receptor–mediated synaptic transmission 9–11,33–35and dysfunctions of linked pathways. 10,34,36 

In brain glucose positron emission tomography, different psychological and metabolic effects after S  (+)-versus  equimolar R  (−)-ketamine at average drug plasma values of 385 ng/ml in volunteers have been found. 29While S  (+)-ketamine altered self-perception and increased glucose metabolism in several brain areas, R  (−)-ketamine caused some state of relaxation and reduced glucose metabolism in different areas. Because the change observed after S  (+)-ketamine was similar to that after a 40% higher dose of racemic ketamine in a companion study, S  (+)-ketamine was believed to be responsible for the central effects of racemic ketamine. 37In another positron emission tomography study, 33psychological changes induced by S  (+)-ketamine (0.2 mg/kg) went along with increased drug binding in brain areas with high NMDA receptor density, such as the thalamus, nucleus caudatus, and temporofrontal cortex. Because these regions contribute to conscious behavior and perception, NMDA receptor blockade was judged to underlie the effects found. However, the disparity of responses in our direct comparison of racemic ketamine and its isomers makes pure NMDA receptor–mediated ketamine effects unlikely. Racemic ketamine is a mixture of half S  (+)- and half R  (−)-ketamine. S  (+)-ketamine has a fourfold higher affinity for the phencyclidine site in the NMDA receptor than R  (−)-ketamine, 9–11,13which is believed to relate to the two-times higher analgesic potency of the S  (+)-isomer than the racemic compound. 8–13Because we tested S  (+)-ketamine at half the dose of racemic ketamine, our data suggest that the R  (−)-isomer in racemic ketamine interacts with the S  (+)-isomer. This effect interaction of the isomers in the racemic mixture appears to be nonlinear in nature. Racemic ketamine may have effects that are not confined to stereoselective NMDA receptor antagonism.

Supporting this hypothesis, R  (−)-ketamine elicited changes in attention that were similar to that of S  (+)-ketamine. However, other than after S  (+)-ketamine, subjects reported to be more “out-of-control” after R  (−)-ketamine, and, compared with racemic ketamine, they felt more brave after S  (+)-ketamine. This could relate either to directly different effects of the isomers, to different influences on transmitters linked to glutamatergic pathways, or to combined mechanisms. In favor for the first explanatory approach, R  (−)-ketamine showed higher ς opioid receptor binding than S  (+)-ketamine, which has been related to dysphoric feelings such as disinhibition. 34In favor for the second approach, human positron emission tomography studies found that psychologic effects of low-dose racemic ketamine are paralleled by overall striatal dopaminergic activation, 38whereas heightened mood after S  (+)-ketamine was related to increased dopamine concentrations in the ventral striatum and in the caudate nucleus. 39These effects of S  (+)-ketamine were further shown to be suppressible by the serotonin 2–dopamine 2–1 antagonist sertindole 40and are consistent with mobilization and turnover of limbic dopamine storage pools. 41Because limbic structures and cortico-striato-thalamic feedback have overlapping input with the accumbens dopamine system, 31increased neurotransmitter concentrations may have contributed to the feelings observed after administration of S  (+)-ketamine.

Early after injection, the isomers caused less intellectual decline than the racemic compound. Racemic ketamine–induced impairment in frontal lobe–sensitive tests and decline in explicit memory are well known, but in previous studies, this was mostly related to pure NMDA receptor antagonism. 6,9,28,29,32,33,35,42Recent work on conscious behavior showed close relations between human working memory and executive functions, 31which is supported by the smaller deficit we found in the intelligence quotient after the isomers compared with racemic ketamine. However, although glutamatergic systems exert potent regulatory control over secondary changes in the dopamine system, 31,32,36and dopamine dysregulation contributes to intellectual decline, 38,39our data, as discussed, caution against straightforward interpretation of ketamine drug effect to one discrete localization of NMDA receptor blockade. However, within the nucleus accumbens, the ventral hippocampus has been found to provide a gating influence over information flow from the prefrontal cortex. 43Different blockade of hippocampal NMDA receptors at these gating mechanisms will be followed by altered corticoaccumbens glutamatergic input and different modulation by dopamine, resulting in different intellectual responses. 30,31,36Furthermore, working memory deficits were reported to be reversed by dopamine receptor stimulation after increasing dopamine concentrations. 44At the earliest study time points, the smallest decline in concentration and memory after administration of S  (+)-ketamine thus appears to involve, at least in part, S  (+)-ketamine–mediated influences on the dopaminergic system.

The intensity and timing of the differences found was temporally graded. Drug concentration versus  time profile–mediated action is thus likely. Ketamine tissue distribution was reported not to be stereoselective in dogs. 45Because the drug plasma profiles measured mirrored the dose ratios used, our data are not suggestive for stereoselective drug tissue distribution.

We are well aware that, despite the many tests examined, there were only a few differences at early study time. Some limitations of our work should therefore be considered. Because this study was designed as a first evaluation of potential differences between small-dose racemic ketamine and its isomers, only one single drug injection was examined. However, in clinical pain therapy, repeated use or continuous ketamine infusion may be necessary to reach the targeted pain reduction. This may result in longer-lasting drug plasma concentrations with more pronounced differences in cognitive effects. Because nobody knew whether there would be differences between the ketamine isomers at all, and where those differences would emerge, we performed extensive descriptive tests. Future study should thus concentrate on the disparity in cognitive decline revealed here, using tools appropriate in the clinical setting of pain therapy at higher analgesic dose. This can prove whether the differences we report are truly clinically significant.

The authors thank Christine Baier, M.D., and Sven O. Muspach, M.D. (Universitätsklinik für Anästhesiologie, Klinikum der Universität Ulm, Ulm, Germany), for help with performing the experiments; Siegfried Lehrl, M.D. (formerly at the Department of Medical Psychology and Psychopathometry, University of Erlangen-Nuremberg, Erlangen, Germany), for advice on psychological testing; and Prof. Dr. Kurt Ulm, Ph.D. (Institute for Medical Statistics and Epidemiology, Technische Universität München, Munich, Germany), for help with statistical analysis.

Before and 20 min after drug injection, subjects rated their current mood on a paper–pencil questionnaire that contained 16 visual analog scales derived from the multidimensional Freiburg Personality Inventory 16and the self-report assessment of the German Association for Methodology and Documentation in Psychiatry. 17At the end of each session, subjects were asked about feelings they recalled retrospectively for the study period. Each of the 16 visual analog scales was represented by a 100-mm long line, the ends of which were marked by “+50” and “−50” points that indicated adjectives representing the extremes of the dimension being rated. Subjects had to mark each line at the distance that best indicated how they felt at the time. Sets of adjectives were presented as follows: afraid–brave, tense–relaxed, agitated–calm, happy–sad, lethargic–quick-witted, alert–drowsy, well–poorly concentrated, detached–moved, depressed–euphoric, peaceful–aggressive, small–big, interoceptive emptiness–eventfulness, lucky–unlucky, threatened–superior, out-of-control–inhibited, teared down–high floating. To reduce bias in responses, adjectives associated with “positive” emotions were sometimes presented on the right and sometimes on the left side of the lines.

To assess within-subject changes after drug administration, numerical differences between the individual baseline and the 20-min scores, and baseline scores and scores indicating the feelings for the study were calculated for the 16 individual items. In addition, all items were categorized into four clinically derived functional clusters to assess integrated processes of psychological changes 16,17: mood = happy–sad, depressed–euphoric, lucky–unlucky, interoceptive emptiness–eventfulness, teared down–high floating; affect–emotion = afraid–brave, tense–relaxed, small–big, threatened–superior; cognitive capacity = lethargic–quick-witted, alert–drowsy, well–poorly concentrated; and ego perception = tense–relaxed, detached–moved, peaceful–aggressive, out-of-control–inhibited. The composite change between baseline and 20-min scores, and baseline scores and the scores for the 60-min period were calculated by summing across the clustered averaged differences of the 16 scores.

Psychomotor coordination was studied using the procedure described by Newman et al . 21Subjects were asked to exactly connect a series of dots to complete the outlining of a simple geometric form. The sum of the errors (distances from missed target dots and extraneous deviation lines) were scored (maximum, 80 mm).

The time required for correct identification of a preselected symbol chosen out of three different symbols presented on a sheet of paper was recorded. 22Participants were asked to count the appearance of the target symbol out of groups of symbols arranged in rows as quickly as possible; the maximum interval allowed was a 60-s period. If the count was not correct, scores were calculated as the difference between the counted and the correct number of the target added to the time recorded.

To assess the time required for reading a series of independent characters, 20 characters in random order were presented. 23,24Subjects were asked to read the 20 characters with a modestly loud voice as quickly as possible, and the examiner noted the time between the first and last sound of reading (T²⁰).

According to information psychology, reading of characters can be used as a measure of cognitive central information flow. 23–26As shown previously, the time for recognition and loud reading of 1 character in a series of 20 random characters can be transformed into a bit representation of information, 1 character capturing 5 bits. This allows for calculation of central information flow as follows:

Information psychology allows to determine the retention time in primary memory. Briefly, the score obtained for immediate digit recall (DR) is incorporated in the following relation 23,24,26after nonlinear correction according to Lehrl and Fischer 23:

Short-term storage capacity is calculated as the product of central information flow and retention time in primary memory 23,24,26:

Concepts of intelligence related to information psychology have shown a reliable and sensitive correlation between central information flow, retention time in primary memory, and traditional estimates of intelligence, such as in the Hamburg-Wechsler Intelligence Scale for Adults. 23,25–27For practical means, the intelligence quotient can be rapidly assessed using the following equation 23,24: IQ = 0.56 × STSC + 55.