Until recently, there was no a priori reason to think that physiologically well managed general anesthesia adversely affects longevity. Yet, a recent large study demonstrated an association between depth of anesthesia, as measured by a processed electroencephalographic (EEG) variable (bispectral index [BIS]), and 1-yr mortality in both middle-aged and elderly patients (1,2). The reasons for reduced life expectancy in association with anesthesia at low BIS values are not clear but the implications are considerable, including introducing the potential for a heretofore unrecognized sickness bias into studies that follow the performance of anesthetized subjects over time. Indeed, we became concerned that our studies showing persistent cognitive impairment in aged rats, followed for weeks to months after general anesthesia (3–5), could have been confounded by subtle, subclinical illness or premature aging in the anesthetized animals. Therefore, we designed this study to test the hypothesis that general anesthesia with isoflurane-nitrous oxide shortens life expectancy in aged rats.
The protocol was approved by the Harvard Medical Area Standing Committee on Animals. Sixteen 22-mo-old, male Fischer-344 rats were acquired from the National Institutes of Health Aged rat colony at Harlan (Harlan Sprague Dawley, Inc., Indianapolis, IN). Animals were housed in a climate- and humidity-controlled room on a 12-h light–dark cycle with continuous access to food and water. After a 1-wk acclimation period, rats were randomly assigned (n = 8/group) to receive 1.2% isoflurane-70% nitrous oxide-30% oxygen (Anesthesia) or 30% oxygen alone (control) for 2 h in an anesthetizing chamber, as described in our previous behavioral studies (3–5). Anesthetic and oxygen concentrations were measured continuously (Datex, Tewksbury, MA; Ohmeda, Madison, WI) and temperature of the chamber was controlled to maintain rat body temperature at 37°C ± 0.5°C. Arterial oxygen saturation and mean arterial blood pressure (MAP) were measured noninvasively in anesthetized rats using a pulse oximeter and rat tail cuff, respectively. At the end of treatment rats were removed from their anesthesia or control chamber and allowed to recover for 30 min in a box flushed with 40% oxygen and then placed in their home cage. Each rat was housed in an individual cage in the hospital animal care facility, where they received daily care from an animal care technician and intermittent visits from a veterinarian. The date of death and time from anesthesia were recorded for each animal.
Because the longevity effects of general anesthesia may be related to its brain effect, we conducted a second experiment in a separate group of aged Fisher 344 rats (n = 3) to estimate the depth of anesthesia achieved with 1.2% isoflurane. In lieu of BIS, which has not been validated in rats, we used EEG burst suppression as a measure of isoflurane’s brain/EEG potency. For this purpose, rats were anesthetized with 3% isoflurane and the tracheas were intubated with a 14-gauge angiocath. The lungs were then mechanically ventilated with 1.2% isoflurane-100% oxygen and the femoral artery and vein cannulated with PE-50 tubing as described previously for monitoring of MAP and blood gases and administration of IV fluids, respectively (6). Bipolar platinum needle electrodes were inserted over each hemisphere for acquisition of the EEG signal (Cadwell Laboratories, Kennewick, WA). The animal was then left undisturbed for 15 min and a baseline EEG recorded. The isoflurane concentration was then increased or decreased in steps, with a 15-min equilibration period between changes, until a stable burst suppression (40%-80% suppression) pattern was obtained. Rectal temperature was maintained at 37°C ± 0.5°C throughout the period of EEG measurement and MAP was maintained above 75 mm Hg by bolus administration of normal saline (up to 3.0 mL total in increments).
Data on MAP and oxygen saturation were analyzed with Student’s t-test. The Kaplan-Meier method was used for constructing survival curves and the log-rank test was applied to determine statistical significance. Because of the possibility of a type II error resulting from the small sample size, the survival curves were also compared using the bootstrap method based on 1000 random samples obtained by resampling with replacement from the empirical data. In this regard, two different types of bootstrap confidence intervals, adjusted percentile and Studentized, were computed to estimate the difference between the survival curves. Among accepted bootstrap statistics, the Studentized and adjusted percentile methods are considered to be most accurate because they adjust best for effects of bias, nonconstant variance, and skewness (7). The decision rule for these statistics is straightforward: if the confidence intervals cover zero, the curves overlap and are not statistically significantly different. All computations were performed with S-Plus software and library boot (a set of functions in S language for bootstrap) (7,8).
Anesthesia with 1.2% isoflurane – 70% nitrous oxide was well tolerated physiologically, with oxygen saturation and MAP remaining within the normal physiologic range. However, there was a statistically significant difference in MAP between the control and anesthetized rats (mean ± sem, 121 ± 1 versus 107 ± 2) mm Hg, respectively; P < 0.001). Oxygen saturation could not be measured in awake rats because of movement artifact but it was well within the physiologic range in anesthetized rats (97.4% ± 0.5%, mean ± sem). In the aged, mechanically ventilated rats, burst suppression was achieved at an expired isoflurane concentrations of 2.1% ± 0.1% (mean ± sem). Therefore, 1.2% isoflurane corresponds to approximately 60% of the burst suppression dose in these animals.
There was no decrease in life expectancy and no difference in the survival curves after isoflurane–nitrous oxide anesthesia (Fig. 1) regardless of the statistics used. The mean survival times of control and anesthetized rats were 140 and 158 days, with 95% bootstrap Studentized confidence intervals of (69.0, 219.2) and (85.7, 214.7), and 95% bootstrap adjusted percentile confidence intervals (78.7, 201.3) and (100.3, 202.0). Comparing the survival curves from control and anesthetized rats using the log-rank test yielded a test statistic of 0.27, meaning the curves were not statistically significantly different (P > 0.8). Likewise, there was no difference between the control and anesthetized rats using either the Studentized (−0.0222, 0.1041) or adjusted percentile (0.0000, 0.0893) bootstrap 95% confidence intervals for the difference between the curves (i.e., both covered zero, meaning the curves are not statistically different with P > 0.2).
This primary finding of this study is that 2 hours of general anesthesia with 1.2% isoflurane − 70% nitrous oxide does not decrease the life expectancy of aged Fischer 344 rats. The life expectancy of our rats was 27 ± 4 (mean ± sem) months, which is similar to that reported in the literature (9,10), with a trend for increased longevity in the rats that had been previously anesthetized. Therefore, generally poorer health is not likely to be a confounder in longitudinal studies of previously anesthetized rats and is unlikely to account for the persistent postanesthetic cognitive impairment we reported previously in these animals (3–5).
No other study has investigated the effect of general anesthesia without surgery on life expectancy. This is probably attributable, in part, to the fact that, absent major physiologic abnormalities during or immediately after anesthesia, there has been no reason to think that anesthesia affects longevity. However, a prospective cohort study of 907 surgical patients reported in preliminary form an association between anesthetic depth and 1-year mortality (1). In this report, anesthetic depth was assessed by a processed EEG variable, the BIS, using a commercial device (Aspect Medical Systems, Newton, MA) and the percentage of time spent during the maintenance phase within various BIS ranges was determined. A predominant BIS <40, which is considered “deep” anesthesia, was associated with 9.3% and 16.7% 1-year mortality among middle-aged and aged patients, respectively; whereas the mortality among young patients at the same predominant BIS was 3.1% (which was the same as the mortality rate among patients of all ages maintained at a BIS >60). In the elderly, even a predominant BIS of 40–60, which is considered appropriate for surgery and anesthesia in humans (11), was associated with a threefold more frequent 1-year mortality than those maintained at a BIS >60 (12.6% versus 4.2%, respectively) but the effect did not achieve statistical significance. A full report of these data has recently been published but without 1-year mortality stratified by age (2).
BIS is a variable derived from the human EEG and there is no counterpart in rats so comparisons are difficult. However, the depth of anesthesia in our rats was probably not too different from the lower levels achieved in the clinical study. EEG burst suppression occurs in humans at 1.5%–1.8% isoflurane (12), whereas 1.8%–2.3% (2.1% in our aged Fisher 344 animals) is required in rats (13). Sixty percent of the burst suppression dose of isoflurane (i.e., 1%) produces BIS values of 25–50 in humans (14) and both nitrous oxide and advanced age reduce anesthetic requirements (15). Therefore, it seems likely that our anesthetic regimen of 1.2% isoflurane, which is 60% of the burst suppression dose in these animals, together with 70% nitrous oxide would produce a depth of anesthesia that approximates the level associated with decreased longevity in humans. Based on such logic, we find no evidence to support the idea that “deep” general anesthesia itself shortens life expectancy.
Differences between this study and the clinical report could be explained by numerous factors. Species differences are the most obvious. Also, unlike clinical studies, where patients have surgery, anesthesia, coexisting disease processes, and hospitalization, our study isolated the effect of general anesthesia itself. Furthermore, comorbid conditions in surgical patients not present in rats that have not had surgery could influence mortality. For instance, β-adrenergic receptor blockers, widely used in the treatment of common diseases of the elderly, such as hypertension and heart disease, increase depth of anesthesia as measured by BIS monitoring (16,17). Likewise, patients with dementia have a lower resting BIS value (18), suggesting that they may have an exaggerated EEG response to general anesthetics. Thus, coexisting disease processes or the medications used to treat them may be responsible for an enhanced brain and BIS response to general anesthetics. If so, sustained unexpectedly deep anesthesia and low BIS values may be a marker for patients already at high risk for a shortened lifespan rather than a cause of it.
This study is limited in some important ways. First, the sample size was small for a mortality study. This was necessary because of the expense involved in purchasing aged rats and housing them for up to 9 months. Accordingly, we used resampling statistics to analyze the data. Although the bootstrap procedure does not fully overcome sample size limitations, it has the benefits of making no assumptions about the distribution of the population of interest and providing a more accurate estimate of error than traditional methods. Thus, it is more robust, enhancing confidence in our conclusion that longevity was not different in our study populations. In addition, we have excluded female rats from our studies to eliminate the potential effects of the estrus cycle on behavioral performance and thus cannot comment on the impact of general anesthesia on their longevity. Another limitation to this study is that we used a single depth of anesthesia and one drug condition (i.e., isoflurane–nitrous oxide). Moreover, our estimate of the depth of anesthesia is based on extrapolation from a raw EEG variable (i.e., burst suppression), as described previously, rather than BIS values. Consequently, although the dosage of isoflurane used represents approximately 60% of the burst suppression dose (not considering the 70% nitrous oxide), we cannot exclude the possibility that different results would be obtained at a larger dosage of isoflurane or with other anesthetics. Burst suppression dosages of isoflurane produce marked hypotension in aged rats, however, which itself could adversely influence longevity.
In conclusion, at a depth corresponding to at least 60% of the burst suppression dose for isoflurane, we find no evidence that general anesthesia with isoflurane–nitrous oxide reduces life expectancy in aged rats. As such, the data indicate that subtle subclinical illness or premature aging do not explain the lasting alterations in spatial memory that we described previously in aged rats after general anesthesia (3–5) and are unlikely to be confounders in longitudinal outcome studies in these animals. Moreover, the results challenge the concept that general anesthesia itself is a risk factor for increased 1-year mortality in elderly humans.
1. Weldon BC, Mahla ME, van der Aa MT, Monk TG. Advancing age and deeper intraoperative anesthetic levels are associated with increased first year death rates. Anesthesiology 2002;97:A–1097.
2. Monk TG, Saini V, Weldon BC, Sigl JC. Anesthetic management and one-year mortality after noncardiac surgery. Anesth Analg 2005;100:4–10.
3. Culley DJ, Baxter MG, Crosby CA, et al. Impaired acquisition of spatial memory two weeks after isoflurane and isoflurane-nitrous oxide anesthesia in aged rats. Anesth. Analg 2004;99:1393–7.
4. Culley DJ, Baxter MG, Yukhananov RY, Crosby G. Long-term impairment of acquisition of a spatial memory task following isoflurane-nitrous oxide anesthesia in rats. Anesthesiology 2004;100:309–14.
5. Culley DJ, Yukhananov RY, Baxter MG, Crosby G. Memory effects of general anesthesia persist for weeks in young and aged rats. Anesth. Analg 2003;96:104–9.
6. 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.
7. Davison AC, Hinkley DV. Bootstrap methods and their application. Cambridge: Cambridge University Press, 1999.
8. Venables WN, Ripley BD. Modern applied statistics with S-plus. 3rd ed. New York: Springer, 1999.
9. 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.
10. Chesky JA, Rockstein M. Life span characteristics in the male Fischer rat. Exp Aging Res 1976;2:399–407.
11. Gan TJ, Glass PS, Windsor A, et al. Bispectral index monitoring allows faster emergence and improved recovery from propofol, alfentanil, and nitrous oxide anesthesia. Anesthesiology 1997;87:808–15.
12. Hartikainen KM, Rorarius M, Perakyla JJ, et al. Cortical reactivity during isoflurane burst-suppression anesthesia. Anesth Analg 1995;81:1223–8.
13. Akrawi WP, Drummond JC, Kalkman CJ, Patel PM. A comparison of the electrophysiologic characteristics of EEG burst-suppression as produced by isoflurane, thiopental, etomidate, and propofol. J Neurosurg Anesthesiol 1996;8:40–6.
14. Olofsen E, Dahan A. The dynamic relationship between end-tidal sevoflurane and isoflurane concentrations and bispectral index and spectral edge frequency of the electroencephalogram. Anesthesiology 1999;90:1345–53.
15. Cole DJ, Kalichman MW, Shapiro HM, Drummond JC. The nonlinear potency of sub-MAC concentrations of nitrous oxide in decreasing the anesthetic requirement of enflurane, halothane, and isoflurane in rats. Anesthesiology 1990;73:93–9.
16. Auerbach AD, Goldman L. Beta-blockers and reduction of cardiac events in noncardiac surgery: scientific review. JAMA 2002;287:1435–44.
17. Johansen JW. Esmolol promotes electroencephalographic burst suppression during propofol/alfentanil anesthesia. Anesth Analg 2001;93:1526–31.
18. Renna M, Handy J, Shah A. Low baseline Bispectral Index of the electroencephalogram in patients with dementia. Anesth Analg 2003;96:1380–5.