Acceptance of general anesthesia is predicated on the assumption that its effects are entirely reversible. However, studies indicate that anesthesia and surgery are associated with cognitive impairment lasting ≥3 mo in 10%–14% of elderly patients (1). There has been much speculation about the causes of such impairment, but the etiology remains unknown. One potential candidate mechanism is general anesthesia itself. General anesthesia affects brain function at all levels, including neuronal membranes, receptors, ion channels, neurotransmitters, and cerebral blood flow and metabolism (2). Moreover, the aged brain is more susceptible to anesthetic effects and has greater sensitivity to nonanesthetic drugs (3,4). The aged brain is also different from the younger brain in several important respects, including size, distribution and type of neurotransmitters, metabolic function, and capacity for plasticity, suggesting that it might be more susceptible to anesthetic-mediated changes (5). Nevertheless, the possibility that general anesthesia contributes to cognitive deterioration in the elderly has not been directly tested, in part because it is difficult in clinical studies to differentiate between the effects of general anesthesia and those of surgery and hospitalization. Accordingly, we hypothesized that general anesthesia itself can cause prolonged cognitive alterations in aged subjects, and we tested this hypothesis in rats exposed to general anesthesia without surgery.
This study was approved by the Standing Committee on the Use of Animals in Research and Teaching, Harvard University/Faculty of Arts and Sciences. Young (6 mo old;n = 12) and aged (18 mo old;n = 13) Fischer 344 rats were acquired from the National Institutes of Health aged rat colony. After a 1-wk acclimation period, rats were food-restricted to 85% of free-feeding body weight and trained in a 12-arm radial arm maze (RAM). Fischer 344 rats were chosen because they have a median life expectancy of 26 mo, are frequently used to study both aging and cognitive impairment, and develop progressive cognitive impairment with age but are not so impaired that ceiling and floor effects are a problem (6–8).
Testing of cognitive function was performed in a 12-arm RAM. This procedure tests spatial working and reference memory and assesses the integrity of the frontal cortex, entorhinal cortex, and hippocampus. We chose the RAM because it allows for repetitive testing and can detect subtle differences in learning and memory caused by aging or sedative medications (9–11). The maze consists of a central platform that communicates with 12 arms, each of which was baited with a hidden food reward. The walls of the maze display simple geometric designs that provide fixed, extramaze cues to assist in spatial navigation. To ensure motivated performance, rats were food-restricted but had free access to water in the home cage.
Rats were adapted to the maze for 10 min daily over 3 days. During this interval, the rat was able to freely explore the maze, in which food rewards were scattered randomly. Initial training consisted of a daily 10-min session in which the rat was placed on the central platform of the maze and all arms were baited. The rat was allowed to choose arms in any order until all 12 arms were visited or 10 min elapsed. A correct choice was defined as one in which the rat entered a baited arm not previously explored, whereas an incorrect choice was scored when the rat entered and proceeded more than 80% down an arm it has previously visited or failed to enter the arm in 10 min. Formal training was concluded when all rats met standardized performance criteria for 2 days, defined as 11 correct choices with 1 or fewer errors in <10 min. The number of days required to meet standardized performance criteria was recorded, as were error rate and time to complete the maze during initial training. Furthermore, to complicate the maze, rats were trained on delay trials. These consisted of removing the rat from the maze for 30 s or 2 h between the first and last 6 correct arm choices.
After this initial training, we excluded two aged rats that never learned the maze and one young adult rat whose performance was more than 2 sd below that of the other young rats. The remaining rats were randomized to an anesthesia or control group. Rats randomized to the anesthesia group (n = 5 young and 6 aged) received 1.2% isoflurane in 70% nitrous oxide/30% oxygen for 2 h in a Plexiglas anesthetizing chamber, whereas the control group (n = 6 young and 5 aged) received air/oxygen (fraction of inspired oxygen, 0.3) at identical flow rates for 2 h. These anesthetics were selected because isoflurane and nitrous oxide are commonly used anesthetics; dosages are an extrapolation from halothane and nitrous oxide minimum alveolar anesthetic concentration (MAC) studies and represent 1.2 and 1.0 MAC in aged and adult rats, respectively (12). Anesthetic (Datex, Tewksbury, MA) and oxygen (Ohmeda, Madison, WI) concentrations were measured continuously, and the temperature of the Plexiglas anesthetizing chamber was controlled to maintain rat temperature at 37°C ± 0.5°C. Anesthesia was terminated by discontinuing the anesthetics; all rats received 100% oxygen for 5 min before removal from the chamber. Rats were allowed to recover for 24 h to avoid the confounding influence of residual anesthetic and then were retested in the maze during postanesthesia weeks 1, 3, and 8. This testing was conducted over six consecutive days, with 30-s and 2-h delay trials on alternate days, and the results of the three trials were averaged.
Because hypotension, hypercarbia, and hypoxia are potential causes of cognitive deterioration, we assessed physiologic status, including mean arterial blood pressure (MAP) and arterial blood gases, in a separate group of young (n = 3) and aged (n = 6 or 9) Fischer 344 rats. These rats were anesthetized with isoflurane 1.2% for insertion and externalization of a femoral arterial catheter, as described previously (13). Nitrous oxide 70% was added thereafter, and rectal temperature, MAP, and arterial blood gases were measured after a 2-h equilibration period, as well as 2 and 24 h after recovery.
Initial training-trial variables for young and aged rats were analyzed with Student’s t-test. Physiologic data were analyzed by one-way analysis of variance (ANOVA), followed by the Student-Newman-Keuls test. For within-group analysis of delay trials, the average group score for the last three pretreatment trials was taken as the baseline, and the result was compared with the average score for each posttreatment time point by a two-way ANOVA, with time and treatment as the two factors. Significant differences on the ANOVA were subjected to the Student-Newman-Keuls test to clarify significant effects. The same statistics were used for between-group comparisons of age-matched control and anesthetized groups at each time point.
Before randomization, aged rats required more training trials to meet standardized performance criteria (13.8 ± 1.2 versus 9.1 ± 0.7;P ≤ 0.01; Student’s t-test), made more errors, and took longer to complete the maze than young rats (Figs. 1 and 2). Spontaneous ventilation with isoflurane/nitrous oxide/oxygen was well tolerated physiologically in both age groups (Table 1). Anesthesia was associated with a small but statistically significant decrease in MAP in young and aged rats (−14% and −9%, respectively, versus the corresponding 2h postanesthesia value;P < 0.01). In addition, Pao2 was higher and pH lower during anesthesia in young rats. However, MAP and arterial blood gases remained well within physiologic limits in both age groups during and after anesthesia.
One of the aged anesthetized rats died before the 8-wk testing period; data from this animal were included in the 1- and 3-wk results, but a post hoc analysis demonstrates that the findings do not differ whether this animal is included or excluded. On the 2-h delay trials, there were no differences in any of the groups, regardless of age or anesthesia condition, in time to complete the maze or error rate (data not shown), suggesting that the task was too difficult. On the 30-s delay trials, in contrast, anesthesia had a differential effect with age. In the young control rats, time to complete the maze and error rate remained stable relative to the pretreatment baseline throughout the experiment, except at 8 wk, when they made more errors (Figs. 3 and 4). Prior anesthesia in young rats did not affect the time to complete the maze but reduced the error rate compared with the group’s preanesthesia baseline at 1, 3, and 8 wk after anesthesia (Fig. 4;P < 0.05). However, it is possible that this improvement could partially reflect the relatively higher error rate in this group at baseline. Indeed, there were no differences between young control and young anesthetized rats in time or error rate except at 2 mo after anesthesia, when previously anesthetized rats made fewer errors than the controls. With respect to errors in aged rats, there were no differences within the aged control or anesthetized groups compared with baseline and no differences between the groups at any testing interval (Fig. 4). However, aged control rats ran the maze significantly faster 1 and 3 wk after treatment than at baseline, signifying improvements in performance with repeated testing, but returned to the pretreatment baseline by 8 wk (P < 0.05;Fig. 3). In comparison, aged rats that received anesthesia made no such improvement. These rats appeared to be indecisive; they often looked down an arm repeatedly, and sometimes looked down several arms, before making a choice, whereas unanesthetized aged and young rats were much more deliberate. Consistent with this apparent indecision, the time required for aged anesthetized rats to complete the maze did not improve with repeated testing and, at 1 and 3 wk after anesthesia, was significantly worse than that of age-matched control rats (P < 0.05;Fig. 3).
The main findings of this study are that general anesthesia produces long-lasting but reversible impairment in aged rats on a previously learned spatial memory task, whereas it appears to improve maze performance in young rats. This is unlikely to be a consequence of physiologic changes associated with general anesthesia, because blood pressure and arterial blood gases remained within the physiologic range during and after anesthesia and were similar in both age groups. During weeks 1 and 3 after anesthesia, impairment in the aged rats was manifested not as deterioration from baseline performance but rather as failure of previously anesthetized aged rats to improve at the same rate or to the same degree as unanesthetized controls. This difference was evident in the time to complete the maze, but not the error rate, and had resolved by eight weeks after anesthesia, suggesting a persistent change in memory function rather than a toxic effect of isoflurane/nitrous oxide anesthesia. In contrast, young rats made fewer errors after anesthesia, and this improvement lasted for two months. No changes were noted in the time to complete the maze, however, most likely because they had reached maximal performance (i.e., a floor) before anesthesia and further improvement could not be detected. Although the learning impairment in aged rats was limited to the time to complete the maze, it is unlikely to be the result of isolated motor impairment, because all rats ambulated normally and otherwise navigated and explored the maze and cage without difficulty. Accordingly, we infer that anesthesia with isoflurane and nitrous oxide produces a sustained, differential effect on established spatial memory in young and aged rats, with improvement in the former and impairment in the latter, and that these effects last considerably longer than previously realized.
Short-term impairment of cognitive and psychomotor performance is common after general anesthesia and is typically attributed to incomplete drug clearance (14). It is unlikely that incomplete clearance explains the results observed here because behavioral testing did not resume until 24 hours after anesthesia, opposite effects were seen in young and aged rats, and the changes lasted 3–8 weeks. Although long lasting memory impairment after uncomplicated anesthesia has not previously been reported, there is some evidence that general anesthetics can produce sustained effects. Halothane and nitrous oxide anesthesia during the perinatal period leads to learning deficits and delayed behavioral development (15), and N-methyl-d-aspartate receptor blockade, which is a property of nitrous oxide, can produce long lasting memory deficits (16). In addition, nitrous oxide produces a distinctive, but apparently reversible, neurotoxic reaction in the cerebral cortex of adult rats at concentrations within the range used for human anesthesia (17,18). Also, isoflurane-induced burst suppression is more effective and longer lasting than electroconvulsive therapy for the treatment of refractory depression, implying that sustained brain changes result from the treatment (19). The relevance of these observations to our results is unclear, but it is not surprising that postanesthetic memory impairment occurs in aged rats. The aged brain is different from the young brain in most respects, including size, neurotransmitter levels, and capacity for neuroplasticity. In addition, the fact that aged rats are more susceptible to the amnesic effects of anticholinergic drugs than young rats suggests that neural systems underlying spatial learning may be more fragile in aged rats (20,21). Indeed, aging itself is associated with an impairment in memory (22–24).
Interest in postoperative cognitive dysfunction has been fueled recently by prospective clinical studies showing that general anesthesia and surgery are associated with long lasting cognitive impairment in elderly patients (1). The largest study prospectively entered more than 1000 elderly patients (median age, 68 years) and demonstrated deterioration lasting at least 3 months on a battery of cognitive tests in nearly 10% of those who underwent surgery and general anesthesia, whereas only approximately 3% of age-matched controls (median age, 67 years) got worse (1). Among patients older than 75 years, 14% were worse 3 months after surgery and anesthesia. This impairment seems to resolve over time, however, because there was no difference in the incidence of cognitive deterioration between small subgroups of control and surgery/anesthesia patients followed up for one to two years (25). This suggests a lasting, but ultimately reversible, functional change in learning and memory after general anesthesia and surgery in aged persons. The cause has not been established, however, in part because clinical studies have not controlled for the anesthetics used and cannot differentiate among the effects of illness, hospitalization, surgery, and anesthesia. What seems relatively clear is that physiologic changes typical of anesthesia are not sufficient cause; in fact, episodes of significant intraoperative or perioperative arterial hypotension (MAP <60% of baseline for >30 minutes) or hypoxemia (arterial oxygen saturation <80% for >2 minutes) do not correlate with the development of cognitive impairment in elderly patients (1). Indeed, our results show that postanesthetic cognitive impairment can occur without systemic physiologic abnormalities and, for the first time, implicate general anesthesia itself in sustained memory impairment in aged subjects.
General anesthesia has memory-enhancing effects in young rats, but ours is not the first study to demonstrate this. In studies in young adult mice, the volatile anesthetics halothane, enflurane, and isoflurane have been shown to enhance memory (26). In fact, 4 consecutive days of enflurane anesthesia for one hour immediately after RAM training reduced the error rate by 60%–70% in mice (27). This improvement occurred on a novel task, and the anesthetic was administered during the consolidation phase (immediately after learning). We too observed a decrease in errors on a spatial memory task, but the task was familiar, and the improvement lasted two months. Our study is therefore unique in demonstrating sustained memory enhancement after general anesthesia, but the mechanism is unknown and the phenomenon has not been described clinically, perhaps because it is too subtle or is negated by other effects of illness/surgery.
Our study has several important limitations. First, our model does not reproduce the clinical situation, in which multiple factors are likely to contribute to postoperative cognitive dysfunction. Second, because we studied a combination of anesthetics at a single dose, it is impossible to say whether the effects are dose dependent or drug specific. Third, we did not correct for age-related decreases in MAC and, thus, cannot exclude the possibility that similar memory impairment would be detected in young rats after deeper anesthesia or that aged rats would be unaffected by lighter anesthesia. Another consideration is that the experimental design tested the effect of general anesthesia on prelearned behavior, not a novel task. This is relevant because although the young and aged rats received equivalent numbers of training trials before anesthesia, the young rats reached the standardized performance criteria more quickly. Thus, at the time of anesthesia, young rats were effectively overtrained relative to aged rats. This does not negate the essential results of the study, however, because performance was compared with aged-matched control rats in each case. Moreover, the testing procedures we used are reliable indicators of age-related learning impairments in that they demonstrate test-retest reliability and consistency and, unlike the Morris water maze, permit repeated testing over time. Finally, our results could underestimate the severity of age-related postanesthetic cognitive impairment because, at 18 months of age, Fischer rats are only at late middle age. Clinical data, for example, demonstrate a more frequent incidence of cognitive deterioration among patients older than 75 years. In practice, however, working with older rats is difficult because some cannot learn and because baseline performance is so poor that a decline is difficult to detect.
This study raises at least as many questions as it answers. It is difficult to understand how general anesthesia ablates memory during the time it is administered, regardless of age, and yet subsequently can enhance memory performance in young animals and impair it in old ones. However, memory itself is a complicated and poorly understood phenomenon that has multiple temporal phases, involves widely distributed neuronal circuits, and ultimately requires new gene expression, protein synthesis, and structural changes within neurons (28). Only further research will help determine how general anesthetics affect these processes.
In the meantime, we can draw several conclusions from these results. First, this paradigm is useful for studying anesthesia-induced cognitive deterioration in aging because it is possible to minimize potential confounders in a way that is not possible in a clinical situation. Second, general anesthesia produces sustained impairment in spatial memory performance in aged animals, which suggests that it may contribute to the cognitive dysfunction observed in some aged patients after anesthesia and surgery. Moreover, such learning impairment can occur in the absence of appreciable systemic physiologic changes. Third, sustained impairment of cognitive function after general anesthesia in an aged animal model, and improvement in young animals, provides a basis for examining the neurobiological substrates of sustained anesthesia-related alterations in learning and memory in humans. Finally, it appears that general anesthesia affects learning and memory longer than previously recognized.
1. Moller JT, Cluitmans P, Rasmussen LS, et al. Long-term postoperative cognitive dysfunction in the elderly ISPOCD1 study: ISPOCD investigators—International Study of Post-Operative Cognitive Dysfunction. Lancet 1998; 351: 857–61.
2. Franks NP, Lieb WR. Molecular and cellular mechanisms of general anaesthesia. Nature 1994; 367: 607–14.
3. Magnusson KR, Scanga C, Wagner AE, Dunlop C. Changes in anesthetic sensitivity and glutamate receptors in the aging canine brain. J Gerontol A Biol Sci Med Sci 2000; 55: B448–54.
4. Ingram DK, Garofalo P, Spangler EL, et al. Reduced density of NMDA receptors and increased sensitivity to dizocilpine-induced learning impairment in aged rats. Brain Res 1992; 580: 273–80.
5. Mrak RE, Griffin ST, Graham DI. Aging-associated changes in human brain. J Neuropathol Exp Neurol 1997; 56: 1269–75.
6. Frick KM, Baxter MG, Markowska AL, et al. Age-related spatial reference and working memory deficits assessed in the water maze. Neurobiol Aging 1995; 16: 149–60.
7. Ghirardi O, Cozzolino R, Guaraldi D, Giuliani A. Within- and between-strain variability in longevity of inbred and outbred rats under the same environmental conditions. Exp Gerontol 1995; 30: 485–94.
8. Baxter MG, Gallagher M. Neurobiological substrates of behavioral decline: models and data analytic strategies for individual differences in aging. Neurobiol Aging 1996; 17: 491–5.
9. Borde N, Jaffard R, Beracochea D. Effects of chronic alcohol consumption or diazepam administration on item recognition and temporal ordering in a spatial working memory task in mice. Eur J Neurosci 1998; 10: 2380–7.
10. Luine V, Rodriguez M. Effects of estradiol on radial arm maze performance of young and aged rats. Behav Neural Biol 1994; 62: 230–6.
11. Decker MW, Gallagher M. Scopolamine-disruption of radial arm maze performance: modification by noradrenergic depletion. Brain Res 1987; 417: 59–69.
12. Loss GEJ, Seifen E, Kennedy RH, Seifen AB. Aging: effects on minimum alveolar concentration (MAC) for halothane in Fischer-344 rats. Anesth Analg 1989; 68: 359–62.
13. Marota JJ, Crosby G, Uhl GR. Selective effects of pentobarbital and halothane on c-fos and jun-B gene expression in rat brain. Anesthesiology 1992; 77: 365–71.
14. Moller JT, Svennild I, Johannessen NW, et al. Perioperative monitoring with pulse oximetry and late postoperative cognitive dysfunction. Br J Anaesth 1993; 71: 340–7.
15. Levin ED, Uemura E, Bowman RE. Neurobehavioral toxicology of halothane in rats. Neurotoxicol Teratol 1991; 13: 461–70.
16. Lukoyanov NV, Paula-Barbosa MM. A single high dose of dizocilpine produces long-lasting impairment of the water maze performance in adult rats. Neurosci Lett 2000; 285: 139–42.
17. Olney JW, Farber NB, Wozniak DF, et al. Environmental agents that have the potential to trigger massive apoptotic neurodegeneration in the developing brain. Environ Health Perspect 2000; 108 (Suppl 3): 383–8.
18. Jevtovic-Todorovic V, Todorovic SM, Mennerick S, et al. Nitrous oxide (laughing gas) is an NMDA antagonist, neuroprotectant and neurotoxin. Nat Med 1998; 4: 460–3.
19. Langer G, Karazman R, Neumark J, et al. Isoflurane narcotherapy in depressive patients refractory to conventional antidepressant drug treatment: a double-blind comparison with electroconvulsive treatment. Neuropsychobiology 1995; 31: 182–94.
20. Poe GR, Teed RG, Insel N, et al. Partial hippocampal inactivation: effects on spatial memory performance in aged and young rats. Behav Neurosci 2000; 114: 940–9.
21. Stemmelin J, Cassel JC, Will B, Kelche C. Sensitivity to cholinergic drug treatments of aged rats with variable degrees of spatial memory impairment. Behav Brain Res 1999; 98: 53–66.
22. Gallagher M, Rapp PR. The use of animal models to study the effects of aging on cognition. Annu Rev Psychol 1997; 48: 339–70.
23. Rapp PR, Amaral DG. Recognition memory deficits in a subpopulation of aged monkeys resemble the effects of medial temporal lobe damage. Neurobiol Aging 1991; 12: 481–6.
24. Gallagher M, Burwell R, Burchinal M. Severity of spatial learning impairment in aging: development of a learning index for performance in the Morris water maze. Behav Neurosci 1993; 107: 618–26.
25. Abildstrom H, Rasmussen LS, Rentowl P, et al. Cognitive dysfunction 1–2 years after non-cardiac surgery in the elderly: ISPOCD group—International Study of Post-Operative Cognitive Dysfunction. Acta Anaesthesiol Scand 2000; 44: 1246–51.
26. Komatsu H, Nogaya J, Anabuki D, et al. Memory facilitation by posttraining exposure to halothane, enflurane, and isoflurane in ddN mice. Anesth Analg 1993; 76: 609–12.
27. Komatsu H, Nogaya J, Kuratani N, et al. Repetitive post-training exposure to enflurane modifies spatial memory in mice. Anesthesiology 1998; 89: 1184–90.
© 2003 International Anesthesia Research Society
28. Bailey CH, Bartsch D, Kandel ER. Toward a molecular definition of long-term memory storage. Proc Natl Acad Sci U S A 1996; 93: 13445–52.