Effect of Isoflurane Anesthesia on Circadian Metabolism and Physiology in Rats

Abstract

Isoflurane anesthesia alters the blood levels of several neuroendocrine hormones associated with normal metabolism and physiology and increases stress, but the effect of brief CO₂ anesthesia on these parameters is unknown. In this study, we examined the effects of isoflurane (4%) compared with brief CO₂ (70% CO₂, 30% air) anesthesia on circadian rhythms of plasma measures of physiology and metabolism. Adult male Sprague–Dawley rats (Crl:SD; n = 6 per group) were maintained on a 12:12-h light:dark (300 lx; lights on, 0600) photoperiod. After 1 wk of acclimation, a series of 6 low-volume blood draws were collected by cardiocentesis under anesthesia using isoflurane (10 min or less) compared with CO₂ (1 min or less) at a single circadian time point every 4 d (0400, 0800, 1200, 1600, 2000, or 2400) over 3 wk to assess arterial blood glucose, lactic acid, and potassium and plasma melatonin, leptin, insulin, total fatty acids, and corticosterone concentrations. Results revealed that plasma levels (mean ± SEM) of melatonin were low (11 ± 1 pg/mL) during the light phase in both groups but were significantly lower during the dark phase in the isoflurane group (48 ± 6 pg/mL) compared with the CO₂ group (162 ± 18 pg/mL). In addition, prominent circadian rhythms of arterial plasma levels of corticosterone, glucose, total fatty acids, lactic acid, and potassium were altered in the isoflurane group compared with the CO₂ group. These findings demonstrate that the normal circadian rhythms of endocrine physiology and metabolism observed during brief CO₂ anesthesia in rats are markedly disrupted by isoflurane anesthesia.

In laboratory animal research, investigators and veterinarians use various anesthetics to prevent unnecessary stress, pain, and discomfort.⁴² When determining the best anesthetic for use during a procedure, several factors regarding the anesthetic agent are evaluated. These include ease of administration and monitoring, ability to produce rapid unconsciousness and recovery,⁴⁰ appropriate duration for the proposed procedure, elimination or minimization of pain and stress, no or minimal side effects on normal metabolism and physiology, and no occupational health risks.^14,26 Species differences are often dramatic,^26,44 and can be an overlooked factor when considering anesthetic agents. It is imperative, however, to know the effects of anesthetics commonly approved by IACUC on research outcomes.

Melatonin, a neurohormone that is released from the pineal gland primarily at night, is a key player in animal health maintenance and wellbeing.^{15-17,23,30,31,41,45} Through the nighttime melatonin signal, the daily light–dark cycle resets the circadian clock located in the suprachiasmatic nuclei of the hypothalamus, thereby entraining peripheral clocks in all body tissues, so that core body processes are regulated.¹¹ The nocturnal circadian melatonin signal informs all the cells, tissues, and organs of the body of the time of day, so that they can adjust their functions for optimal homeostasis. Circadian disruption constitutes a significant change in phasing, periodicity, amplitude, or duration of a circadian rhythm from its normal pattern under 12:12-h light:dark conditions, resulting in the interference of the body's normal homeostasis of virtually all biochemical, neurobehavioral, and physiologic events.¹¹ Specifically, melatonin levels have been reported to alter regulation of metabolic hormones, including insulin, leptin, glucose, cholesterol, triglycerides, and corticosterone, among others.⁴⁷ Many pathologic conditions, such as metabolic syndrome and obesity, are known to involve disturbances of this system.^11,46,47 Therefore, protecting this normal biologic timing, particularly the nocturnal melatonin surge, is of paramount importance in mammals used for research.

There is still much to learn about the modes of action of some commonly used anesthetics;⁴² specifically their effect on animals and humans studied without the confounder of preanesthetics and surgical procedures,^{19,32,34,42,55} varying exposure durations, and method of induction and maintenance.^17,42 The mechanism of action and target sites of isoflurane and CO₂ are still largely unknown,¹ however, recent reports implicate activation or inhibition of background K⁺ channels, respectively. Previous studies revealed the adverse effects of isoflurane and CO₂ on animal physiology^8,42,48 and, in turn, the potential effect of their use in investigations using animal models. In contrast, both isoflurane and CO₂ are considered to be superior anesthetics under certain conditions.³⁸

The effects of isoflurane and CO₂ vary depending on the species. Potency (measured as minimum alveolar concentration), anesthetic-sparing effect, and ED₅₀ values^10,26 are all dependent on the species studied. In addition, aspects of normal physiology, including hypoxic ventilatory response, sensitivity of the central respiratory center, and bicarbonate buffering capacity, vary depending on the species (for example, rats, dogs, humans),⁴⁴ as evidenced through increased resting pH and PaCO₂. Genetic adaptions required for burrowing in rodents have been proposed⁴⁴ as an explanation for some of the physiologic differences. Therefore, the effects of anesthesia and analgesia in humans should not be assumed to be similar in other species.^14,42 Species-specific differences are often overlooked in regard to anesthetic mechanisms of action and the resulting effects on normal physiology⁴⁴ and, consequently, scientific outcomes.

In this study we hypothesized that, compared with brief exposure to CO₂ (70% CO₂, 30% air), isoflurane (4%) anesthesia would have detrimental effects on circadian rhythms of physiology and metabolism in male Sprague–Dawley rats, similar to those reported in dogs and humans.^5,30,49 Parameters measured included glucose, lactic acid, blood gases, potassium, total fatty acids (TFA), melatonin, leptin, insulin, and corticosterone, because we routinely measure these markers for our circadian studies. We assessed 6 equispaced circadian time points encompassing the 24-h day over a 3-wk period. In addition, we compared the duration of induction, procedure, and recovery time and noted any behavioral signs exhibited during induction, because all of these factors are important to consider when determining the optimal anesthetic for research studies.

Materials and Methods

Reagents.

HPLC-grade chloroform, ethyl ether, methanol, glacial acetic acid, heptane, and hexane were purchased from Fisher Scientific (Pittsburg, PA). Free fatty acid, rapeseed oil methyl ester standards, boron–trifluoride–methanol, potassium chloride, sodium chloride, and perchloric and trichloroacetic acids were purchased from Sigma Scientific (St Louis, MO). Ultrapure water (catalog no. 400000) was purchased from Cayman Chemical (Ann Arbor, MI).

Animals, housing conditions, and diet.

Male Sprague–Dawley rats (weight, 350 to 500 g) used in this study were purchased from Charles River (Crl:CD), Kingston, NY) and were certified by the vendor to be pathogen-free for all known rodent bacterial, viral, and parasitic pathogens. Rats were maintained in the Tulane University School of Medicine facility, which has been AAALAC-accredited since 1962. All animal use and accompanying procedures were in accordance with the Guide for the Care and Use of Laboratory Animals³⁴ and approved by the IACUC.

Rats were housed in IVC containing hardwood maple bedding (catalog no. 7090, Sanichips, Harlan Teklad, Madison, WI; 2 bedding changes weekly). Serum samples from sentinel animals on combined soiled bedding from project rats were tested quarterly (multiplex fluorescent immunoassays for rat corona virus, Sendai virus, pneumonia virus of mice, sialodacryoadentis virus, Kilham rat virus, Toolan H1 virus, reovirus type 3, Mycoplasma pulmonis, lymphocytic choriomeningitis virus, mouse adenovirus 1 and 2, Hantaan virus, Encephalitozoon cuniculi, cilia-associated respiratory bacillus, parvovirus NS1, rat parvoviruses, rat murine virus, and rat theilovirus (Idexx Research Animal Diagnostic Laboratory, Colombia, MO) and for external and internal parasites to ensure that rats remained infection-free; all test results were negative during the course of the experiment. Rats had free access to a commercial diet (no. 5053 Irradiated Laboratory Rodent Diet, Purina, Richmond, IN) and acidified water. Quadruplicate determinations of this diet contained 3.9 g TFA per 100 g of diet, composed of 1.06% myristic (C14:0), 15.94% palmitic (C16:0), 1.47% palmitoleic (C16:1n7), 3.97% stearic (C18:0), 22.22% oleic (C18:1n9), 54.54% linoleic (C18:2n6), and 0.26% arachidonic (C20:4n6) acids; minor amounts of other fatty acids comprised 0.54%. Food and water intake was measured every 3 to 4 d; premeasured amount of waters (500 mL) and food (500 g) were provided in the cage on cage change day. On subsequent cage change days, the remaining food and water was removed and measured by using a digital scale or graduated cylinder, respectively, and recorded. These measurements were then divided by the product of the number of days since change and the number of rats per cage (that is, 2) to give the amount of food and water consumed per animal daily. Individual rats were marked with a sharpie for identification, and body weights were measured weekly. The overall body weight of all rats at the start of experiments after acclimation to the facility and room lighting was 432 ± 30 g.

Caging, lighting regimens, and spectral transmittance measurements.

On arrival, a Department of Comparative Medicine laboratory technician randomly assigned each rat to a standard ventilated polysulfone laboratory rat cage (n = 6 per group, 2 per cage; 10.5 in. × 19 in. × 8 in.; wall thickness, 0.1 in., catalog no. 10198, Allentown Caging, Allentown, PA) for a 1-wk acclimation period. All cages had identical stainless steel lids (catalog no. 10SS, Ancare, Bellmore, NY) for cradling food and the water bottle and were covered with polysulfone translucent microfilter tops (catalog no. N10MBT, Ancare). The rats were maintained in climate- and light-controlled rooms (21 to 25 °C; 50% to 55% relative humidity), with half of the animals receiving diurnal lighting of 12 h light followed by 12 h dark (lights on, 0600), and the other half set on reverse-phase lighting of 12 h dark followed by 12 h light (lights on, 1800 h) so that 2 circadian time point blood collections could be done in the same day from different animal groups. Animal rooms were lighted with overhead white fluorescent lamps as described previously¹⁸ and were completely devoid of light contamination during the dark phase. Daily (0800) during the course of this experiment, the lighting intensity in the animal room was measured at 1 m above the floor in the center of the room by using a radiometer–photometer and radiance detector with filter and diffuser, as previously reported,¹⁸ which were calibrated regularly. All cages were rotated daily on the room shelf in 8 preestablished, specific positions to ensure that there were no significant differences between cages in lighting intensity inside the cages (at rodent eye level). To this end, lighting intensity was measured inside each cage (with no animals inside) while it was on the shelf rack (model no. RS10198U30MVFLG, Allentown). Once all intensities at each of the 8 positions were approximately equal, cage positions on the shelf were labeled 1 through 8. During the experiment, cages were rotated one position to the right on the shelf daily; the cage in position 8 was rotated to position 1. Cage cleaning procedures, as previously reported,¹⁸ did not cause any variations in light intensity measurements throughout the course of the study.

Blood collection.

After the 1-wk acclimation to the facility and then a 2-wk exposure under the described lighting conditions, blood was collected from rats from each group at one circadian time point (0400, 0800, 1200, 1600, 2000, or 2400) every 4 d by cardiocentesis under either isoflurane or CO₂ gas anesthesia. These low-volume collections (0.5 to 1.0 mL, depending on body weight) were spaced 4 d apart to minimize adverse effects of serial blood collection, as previously reported.¹⁸ Briefly, 4% isoflurane (rate, 2 L per minute) was passed into the bottom of a acrylic gas anesthesia chamber (10 in. × 8 in. × 8 in.), displacing air at a rate of 20% volume per minute. On loss of righting reflex, each rat was removed from the chamber and placed in a nose cone connected by a T-piece to the same isoflurane supply line while blood (0.5 to 1 mL; less than 5% total blood volume) was withdrawn by using a 3-mL luer-lock syringe (Tyco Healthcare, Mansfield, MA) and 25-gauge, 3/8-in. needle (Tyco) moistened with sodium heparin (5000 USP units/mL, Sagent Pharmaceuticals, Schaumburg, IL). Over the last decade in our laboratory, mortality and morbidity of rats due to cardiocentesis has been less than 5%.²¹

For CO₂ anesthesia, 100% CO₂ (rate, 2 L per minute) was passed into the same acrylic gas anesthesia chamber, displacing air at a rate of 20% volume per minute, to keep concentrations constant between animals. On loss of righting reflex, each rat was removed and allowed to breathe room air while blood was withdrawn, as opposed to requiring additional anesthesic via a nosecone as in the isoflurane anesthesia group. CO₂ and O₂ concentrations inside the anesthesia chamber when no animals were present were monitored (model no. 8762, serial no. 01120587, IAQ-Calc Indoor Air Quality Meter, TSI, Shoreview, MN, and Type 401, serial no. E065052 Eagle Series Portable Multigas Detector, RKI Instruments, Union City, CA).

Dark-phase sampling was performed in less than 2 min for CO₂ anesthesia and in less than 10 min for isoflurane anesthesia under 1 or 2 safelight red lamps (Kodak 1A, model B, catalog no. 152 1517, 120 V, 15 W, Kodak, Rochester, NY) to preserve the normal nocturnal melatonin surge. Exposure at rodent eye level was 45 to 179.5 µW/cm² (18.4 to 73.3 lx), depending on the use of 1 or 2 lights and their position during sampling.

Induction, procedure, and recovery.

For both CO₂ and isoflurane anesthesia protocols, the induction, procedure, and recovery times were recorded for each rat by using benchtop timers (catalog no. 14-648-34, Traceable Benchtop Timer, Fisher Scientific). Induction time was defined as the interval from placement of the rat into the anesthesia chamber to the onset of unconsciousness, defined as the absence of the righting reflex and the presence of relaxed body condition. Toe pinch (pedal withdrawal reflex) was used to evaluate anesthetic depth (surgical plane), along with recumbency and muscle relaxation on loss of consciousness and decreased respiratory rate. In addition, the absence of purposeful movement during the procedure indicated appropriate anesthetic depth for animal wellbeing.²⁶ Recovery time was defined as the time interval from the end of the procedure to sternal recumbency or spontaneous return to righting. Observation notes were used to record behavior during these phases.

Melatonin analysis.

Plasma melatonin levels were measured by using a melatonin ¹²⁵I-radioimmunoassay kit (Bühlmann, Switzerland) and analyzed in an automated gamma counter (Cobra 5005, Packard, Palo Alto, CA) as previously described.¹⁸ Briefly, C18 reverse-phase extraction columns included in the kit were used to extract the melatonin from the samples by using 0.125 mL of plasma for 2400 and 0400 time points. For the 0800, 1200, 1600, and 2000 time points, 0.25 mL of plasma within the same group were pooled to make a total of 0.5 mL, because of the low levels of melatonin in these samples (that is, prior to the normal nighttime surge). The functional sensitivity for the assay was 0.9 pg/mL. Intraassay precision of the column extraction and radioimmunoassay procedures combined was 8.2%.

ELISA analysis of corticosterone, insulin, and leptin.

Plasma samples were prepared in duplicate for measurement of corticosterone (catalog no. 55-CORMS-E01), insulin (catalog no. 80-INSRTH-E01, rat high range), and leptin (catalog no. 22-LEPMS-E01, mouse or rat) levels by using chemiluminescent ELISA diagnostic kits (ALPCO, Salem, NH) and read on a microplate reader (450 nM; VersaMax Microplate Reader, Molecular Devices, Sunnyvale, CA) as previously described.¹⁸ Detection sensitivity for corticosterone, insulin, and leptin were 4.5 ng/mL, 0.124 ng/mL, and 10 pg/mL, respectively. Lower limits of detection of corticosterone, inulin, and leptin were 15 ng/mL, 0.15 ng/mL, and 10 pg/mL, respectively, and the coefficient of variation of all assays was less than 4.0%.

Fatty acid extraction and analysis.

Plasma free fatty acids were extracted as previously described¹⁸ from 0.05-mL samples in duplicate by using an internal standard consisting of heptadecanoic acid (100 µg) dissolved in chloroform. A gas chromatograph (model 58990A, Hewlett Packard, Palo Alto, CA) fitted with a flame ionization detector, auto injector (both adjusted to 220 °C), and integrator was used to analyze the methyl esters of fatty acids according to their retention time compared with known standards. A 0.25 mm × 30 cm capillary column using helium as the carrier gas was used for separations. The minimal detection level for the assay was 0.05 µg/mL.

Glucose, lactate, blood gases, and potassium measurements.

After each blood collection, a portion (less than 95 µL) of the total sample was used to measure pH, pO₂, pCO₂, glucose, and lactic acid (iSTAT1 Analyzer with CG4+ and CG8+ cartridges, Abbott Laboratories, East Windsor, NJ). The remaining blood sample was centrifuged at 9660 × g for 10 min at 4 °C (accuSpin Micro17R centrifuge, Fisher Scientific). Plasma was transferred to a new microtube (Sigma-Aldrich) and stored at –20 °C until assayed for melatonin, TFA, corticosterone, insulin, and leptin.

Statistical analysis.

All data are presented as mean ± SEM unless otherwise noted and were compared by using one-way ANOVA followed by the Bonferroni multiple-comparison test or 2-tailed t test (Prism, GraphPad Software, La Jolla, CA). Due to assay errors, protocol sample requirements (noted earlier for melatonin analysis), or sample volume limitations, sample numbers vary but are 6 or greater, unless otherwise noted. This experiment was repeated once. Differences between the group means were considered statistically significant at a P value of less than 0.05.

Results

Measurement of caging illumination.

Light illumination measurements were 75.6 ± 9.8 µW/cm² (184.4 ± 67.9 lx; mean ± SEM, n = 8) inside the cage at rat eye level. As mentioned previously, cage positions were rotated daily among presetablished positions to ensure equal luminance between groups.

Dietary and water intakes and body growth rate.

Food consumption, water intake, and body weight did not differ between groups; therefore, these data were combined. Food consumption (mean ± 1 SD) was 6.7 ± 0.5 g per 100 g body weight, and water intake among subjects was 10.6 ± 1.4 mL per 100 g body weight. Body growth rate averaged 24.3 ± 12.9 g weekly. These values are comparable to those reported by the stock producer.³⁰

Anesthesia chamber measurements.

Atmospheric CO₂ and O₂ gases inside the anesthesia chamber were measured while no animals were present and validated in a subsequent session by simulating the CO₂ anesthesia methodology. Baseline atmospheric O₂ measured inside the chamber with the lid fully open prior to any gas introduction was 19.5% ± 0.1% (n = 6). Within 1 min of turning on the CO₂ gas, atmospheric CO₂ exceeded 6000 ppm, the monitors’ upper limit. To simulate removal and replacement of successive rats for anesthesia, the top was opened for 5 s and then closed, and O₂ and CO₂ again were measured at the front of the chamber, 2 in. from the bottom and 2 in. from the top of the chamber. This process was repeated 23 times. Because no significant difference was found between trials, data were averaged (top CO₂, 17.8% ± 0.4%; bottom CO₂, 4.4% ± 0.26%, and greater than 6000 ppm, 0.6%).

Induction, procedure, and recovery.

The animal anesthesia induction period for CO₂ was 16.4 ± 3.1 s (mean ± 1 SD, n = 12) compared with isoflurane anesthesia induction time of 222 ± 42 s (n = 38). The cardiocentesis procedure duration for both groups was 19.3 ± 4.0 s (n = 50). Recovery time for the CO₂ anesthesia group was 42.8 ± 33.8 s (n = 12) compared with the isoflurane group recovery time of 294 ± 18 s (n = 38). Behaviors noted during isoflurane induction included nose twitching, raising, paddling after loss of righting reflex, and a period of excitement or staggering for several minutes. Behaviors noted during CO₂ induction included nose twitching, raising, nose touching, and a few seconds of staggering before loss of righting reflex.

Plasma melatonin.

Both the isoflurane and CO₂ anesthesia groups demonstrated prominent circadian rhythms (Figure 1). As expected, daytime melatonin concentrations for both groups were low (isoflurane, 12.0 ± 1.0 pg/mL; CO₂, 10.4 ± 1.4 pg/mL). However, during the dark phase, plasma melatonin was lower in rats exposed to isoflurane anesthesia (48.0 ± 6.5 pg/mL) compared with those exposed to CO₂ anesthesia (162.4 ± 18.5 pg/mL). In addition, the isoflurane group's melatonin concentrations was significantly (P < 0.05) lower than the CO₂ group's at 2400 (isoflurane, 60.9 ± 3.2 pg/mL [n = 6]; CO₂, 208.5 ± 13.2 pg/mL [n = 10]) and 0400 (isoflurane, 55.6 pg/mL ± 13.5 [n = 5]; CO₂, 181.1 ± 24.2 [n = 7] pg/mL). At 0800, plasma melatonin differed significantly (P < 0.05) between the CO₂ (6.7 pg/mL ± 1.4 [n = 6]) and isoflurane (14.5 ± 1.5 pg/mL [n = 3]) groups, but both values were within expected physiologic daytime ranges. No significant differences between groups were found at 1200, 1600, and 2000. At 2400, plasma melatonin concentrations reached their peak for both groups of rats.

Figure 1.

Circadian rhythm of arterial plasma melatonin (pg/mL) concentrations in rats that underwent isoflurane anesthesia (blue) compared with brief CO₂ anesthesia (amber)_. Black bars represent the dark phase; *, significant (P < 0.05) difference between groups. Values are double plotted. The total numbers of samples after pooling (see methods) at time points 0400, 0800, 1200, 1600, 2000, and 2400 were 5, 3, 6, 2, 4, and 6 for the isoflurane group and 7, 6, 6, 6, 4, and 10 for the CO₂ group, respectively. For reference, the peak nighttime (red line) and daytime (gray line) values in decapitated Wistar rats are shown.⁹

Plasma corticosterone.

In both groups, plasma corticosterone concentrations had a clear diurnal rhythm, peaking during the dark phase and reaching the nadir during the light phase (Figure 2). In the isoflurane group, peak concentrations were phase-advanced by 4 h, occurring at 2400 (91 ± 3.7 ng/mL), whereas plasma corticosterone in the CO₂ group (42.2 ± 9.2 ng/mL) peaked at 0400 and was significantly lower than the isoflurane level at 1200, 2000, and 2400 (P < 0.05) and tended to be lower at 0400 (P < 0.07) and 0800 (P < 0.06).

Figure 2.

Circadian rhythm of arterial plasma corticosterone (ng/mL) concentrations in rats that underwent isoflurane anesthesia (blue) compared with brief CO₂ anesthesia (amber)_. Black bars represent the dark phase; *, significant (P < 0.05) difference between groups. Values are double plotted. The total numbers of samples at time points 0400, 0800, 1200, 1600, 2000, and 2400 were 5, 3, 3, 6, 4, and 5 for the isoflurane group and 3, 3, 3, 6, 9, and 3 for the CO₂ group, respectively. For reference, corticosterone in trunk blood taken at decapitation of male Sprague–Dawley rats has been reported to be low at the onset of the light phase^5,16 (20 ng/mL; gray line), with a peak (approximately) during the dark phase.¹⁶

Plasma TFA.

Diurnal rhythms of arterial plasma TFA in male rats with free access to food were present for both anesthesia groups (Figure 3), but were lower (P < 0.05) in the isoflurane group than the CO₂ anesthesia group at 0400 (isoflurane, 1666 ± 136 µg/mL [n = 12]; CO₂, 5236 µg/mL ± 37 [n = 11]), 0800 (isoflurane,1369 ± 45 µg/mL [n = 12]; CO₂, 2435 ± 34 µg/mL [n = 12]), 2000 (isoflurane,1102 ± 62 µg/mL [n = 12]; CO₂, 2995 ± 58 µg/mL [n = 12]), and 2400 (isoflurane,1441 ± 6 µg/mL [n = 12]; CO₂, 3730 µg/mL ± 32.2 [n = 12]). The isoflurane anesthesia group had significantly (P < 0.05) higher plasma TFA levels at 1600 (isoflurane,1484 ± 52.0 µg/mL [n = 10]; CO₂, 1346 ± 23 µg/mL [n = 12]). No significant difference was found at 1200. In both the isoflurane and CO₂ anesthesia groups, plasma TFA peaked at 0400.

Figure 3.

Circadian rhythm of arterial plasma total fatty acids (µg/mL) concentrations in rats that underwent isoflurane anesthesia (blue) compared with brief CO₂ anesthesia (amber)_. Black bars represent the dark phase; *, significant (P < 0.05) difference between groups. Black bars represent the dark phase; *, significant (P < 0.05) difference between groups. Values are double plotted. The total numbers of samples at time points 0400, 0800, 1200, 1600, 2000, and 2400 were 6 in all cases for the isoflurane group and 11, 12, 12, 12, 12, and 12 for the CO₂ group, respectively. The current report is the first description of circadian rhythmicity in plasma TFA concentrations of Sprague–Dawley rats; however, the circadian rhythmicity of plasma fatty acids is well known.¹

Blood glucose.

Blood glucose concentrations in the isoflurane group were significantly (P < 0.05) higher than those for the CO₂ group and were at hyperglycemic levels at all time points (Figure 4). Overall blood glucose values (mean ± SEM) for the isoflurane and CO₂ anesthesia groups were 183 ± 3 mg/dL (n = 36) and 132 ± 2 mg/dL (n = 72), respectively.

Figure 4.

Circadian rhythm of arterial blood glucose (mg/mL) concentrations in rats that underwent isoflurane anesthesia (blue) compared with brief CO₂ anesthesia (amber)_. Black bars represent the dark phase; all values differed significantly (P < 0.05) between groups at each time point. Values are double plotted. The total numbers of samples at time points 0400, 0800, 1200, 1600, 2000, and 2400 were 6 in all cases for the isoflurane group and 11, 12, 12, 12, 12, and 12 for the CO₂ group, respectively. For reference, the gray line represents the vendor-reported mean values in Sprague–Dawley rats.³⁰

Plasma insulin.

Plasma insulin concentrations in the isoflurane anesthesia group peaked at 1600, compared with 2400 for the CO₂ group (Figure 5), and reached a nadir at 2400 for the isoflurane group compared with 1200 for the CO₂. Circadian cycling was evident in both groups, but plasma insulin was significantly (P < 0.05) higher in the isoflurane group than the CO₂ group at 0400 (isoflurane, 1.3 ± 0.1 ng/mL [n = 6]; CO₂, 0.8 ± 0.2 ng/mL [n = 7]) and 1200 (isoflurane, 1.2 ng/mL ± 0.2 [n = 5]; CO₂, 0.7 ng/mL ± 0.1 [n = 8]) but lower at 2400 (isoflurane, 0.7 ± 0.0 ng/mL [n = 6]; CO₂, 1.5 ± 0.2 ng/mL [n = 6]). Mean insulin concentration over the 24-h period was higher for the isoflurane group (1.5 ± 0.5 ng/mL) than the CO₂ group (1.1 ± 0.3 ng/mL).

Figure 5.

Circadian rhythm of arterial plasma insulin (ng/mL) concentrations in rats that underwent isoflurane anesthesia (blue) compared with brief CO₂ anesthesia (amber)_. Black bars represent the dark phase; *, significant (P < 0.05) difference between groups. Values are double plotted to show rhythmicity. The total numbers of samples at time points 0400, 0800, 1200, 1600, 2000, and 2400 were 6, 6, 5, 5, 6, and 6 for the isoflurane group and 7, 7, 8, 5, 5, and 6 for the CO₂ group, respectively. For reference, the gray line represents fasting plasma insulin concentrations reported from tail vein samples in Sprague–Dawley rats.⁵⁴

Blood lactate.

For both isoflurane and CO₂ anesthesia groups, blood lactate concentrations peaked at 2000 (isoflurane, 2.0 ± 0.2 mmol/L [n = 6]; CO₂, 1.9 ± 0.1 mmol/L [n = 12]; P < 0.05) and reached nadir at 1200 (isoflurane, 1.6 ± 0.1 mmol/L [n = 6]; CO₂, 1.2 ± 0.1 mmol/L [n = 12]; P < 0.05), with the circadian rhythms following each other closely (Figure 6). In addition, blood lactate was higher in the isoflurane group than the CO₂ group at 1600 (isoflurane, 2.0 ± 0.1 mmol/L [n = 6]; CO₂, 1.5 ± 0.1 mmol/L [n = 12]; P < 0.05) and 0800 (isoflurane, 2.0 ± 0.1 mmol/L [n = 6]; CO₂, 1.5 ± 0.1 mmol/L [n = 12]; P = 0.064).

Figure 6.

Circadian rhythm of arterial blood lactate (mmol/L) concentrations in rats that underwent isoflurane anesthesia (blue) compared with brief CO₂ anesthesia (amber)_. Black bars represent the dark phase; *, significant (P < 0.05) difference between groups. Values are double plotted to show rhythmicity. The total numbers of samples at time points 0400, 0800, 1200, 1600, 2000, and 2400 were 4, 6, 6, 6, 6, and 6 for the isoflurane group and 11, 12, 12, 12, 12, 12 for the CO₂ group, respectively. For reference, the gray line represents the mean value reported in Sprague–Dawley rats.³⁴

Blood gases.

At the time points of 0400, 0800, 1200, 1600, 2000, and 2400, rats anesthetized with isoflurane had a mean PO₂ of 139.8 ± 41.7, 179.0 ± 48.2, 221.2 ± 31.6, 260.5 ± 25.8, 239.0 ± 52.2, and 203.0 ± 11.5 mm Hg, respectively, and a PCO₂ of 55.7 ± 2.3, 54.4 ± 4.1, 48.9 ± 1.2, 54.4 ± 1.4, 50.7 ± 1.4, and 40.7 ± 0.6 mm Hg, respectively. These measurements are comparable to the control values determined for brief CO₂ anesthesia in previous studies.^18,20

Blood potassium.

Blood potassium concentrations were lower in the isoflurane group than the CO₂ group at 0400 (isoflurane, 4.80 ± 0.16 mEq/L [n = 6]; CO₂, 6.54 ± 0.31 mEq/L [n = 11]; P < 0.05), 0800 (isoflurane, 4.92 ± 0.08 mEq/L [n = 6]; CO₂, 6.37 ± 0.18 mEq/L [n = 13]; P < 0.05), 1200 (isoflurane, 4.35 ± 0.07 mEq/L [n = 6]; CO₂, 5.76 ± 0.14 mEq/L [n = 12]; P < 0.05), and 2000 (isoflurane, 4.55 ± 0.19 mEq/L [n = 6], CO₂ 5.46 ± 0.26 mEq/L [n = 11]; P < 0.05) and approached significance at 2400 (isoflurane, 4.75 ± 0.15 mEq/L [n = 6]; CO₂, 5.45 ± 0.23 mEq/L [n = 12]; P = 0.0638). Circadian rhythmicity was evident in both groups. Blood potassium peaked at 1600 in the isoflurane group but at 0400 in the CO₂ group, with nadirs at 1200 and 2400, respectively (Figure 7).

Figure 7.

Circadian rhythm of arterial blood potassium (mEq/L) concentrations in rats that underwent isoflurane anesthesia (blue) compared with brief CO₂ anesthesia (amber)_. Black bars represent the dark phase; *, significant (P < 0.05) difference between groups. Values are double plotted to show rhythmicity. The total numbers of samples at time points 0400, 0800, 1200, 1600, 2000, and 2400 were 6 in all cases for the isoflurane group and 11, 12, 12, 12, 11, 12 for the CO₂ group, respectively. For reference, the gray line represents the vendor-reported mean value for Sprague–Dawley rats.³⁰

Plasma leptin.

Blood leptin concentrations followed a prominent circadian rhythm in the CO₂ anesthesia group but not in the isoflurane anesthesia group (Figure 8). The isoflurane group demonstrated 2 peaks, with the one at 2400 higher than that at 1200. Two peaks in blood leptin concentration also occurred in the CO₂ anesthesia group, with that at 0800 greater than the second peak at 2400. Leptin concentrations were significantly higher at all time points in the isoflurane group than the CO₂ anesthesia group (P < 0.05 for all time points except 1600 [P = 0.0588]).

Figure 8.

Circadian rhythm of arterial plasma leptin (ng/mL) concentrations in rats that underwent isoflurane anesthesia (blue) compared with brief CO₂ anesthesia (amber)_. Black bars represent the dark phase; *, significant (P < 0.05) difference between groups. Values are double plotted to show rhythmicity. The total numbers of samples at time points 0400, 0800, 1200, 1600, 2000, and 2400 were 6, 6, 5, 5, 6, and 6 for the isoflurane group and 6, 9, 7, 11, 6, and 8 for the CO_2­ group, respectively. For reference, the gray line represents the mean values reported in Sprague–Dawley rats.⁴⁵

Discussion

The interpretation of behavioral signs related to pain and stress during anesthesia remains a topic of debate. Both of our anesthetized groups exhibited nose twitching, rising, and a period of excitement,^2,27,55 which was much longer in the isoflurane group and might have been due to the differences in induction time. Some isoflurane-anesthetized rats exhibited paddling after the loss of the righting reflex, often considered an adverse effect associated with neurologic excitement.²⁶ Some CO₂-anesthetized rats exhibited nose touching, which some authors interpret as an expression of pain.^4,14 Pain is a subjective measurement,⁴ and conflicting reports in humans indicate no pain to a range of pain sensations depending on the CO₂ concentration to which they are exposed_.²⁴ Recent studies advocate nasal application of CO₂ to relieve migraine headaches and symptoms of allergic rhinitis.^12,51 Like all other drugs, CO₂ can have vastly different effects depending on species, total dosage, concentration, and duration of application.^4,12,51 The maintenance of body weight and food and water intake throughout the experiment in both groups give indirect but credible measurements of animal wellbeing.

In addition, isoflurane has a pungent odor.² Inhalation for as few as 10 s at various concentrations produced coughing, burning, and irritation or other discomfort in humans.^2,24,50 Other adverse effects commonly reported include nausea, vomiting, and headaches, which also are associated with occupational exposure to isoflurane.²⁹ Therefore, when comparing CO₂ and isoflurane in regards to pungency, irritability, and possible pain sensations, both gasses induce negative effects.^2,3,23,39 Due to interspecies differences in sensitivity to CO₂, directly studying isoflurane and CO₂ inhalation for their irritant qualities in rats seems prudent.

Although both methods of anesthesia had relatively short induction and recovery times, isoflurane induction was 14-fold slower and recovery 7-fold slower than those for brief CO₂ anesthesia. The duration of action for isoflurane was too short for our blood-draw procedure, because the rats did not stay in a surgical plane long enough after removal from the anesthesia chamber; therefore, we added nosecone maintenance to the isoflurane anesthesia protocol. During brief CO₂ anesthesia, the anesthetic depth gets deeper²⁷ after removal from the anesthetic chamber and before recovery. Because of this difference, CO₂ need not be supplied continually in the chamber or via nosecone for maintenance, and doing so may be the cause of the adverse effects² previously reported for CO₂ anesthesia, presumably due to overdosage and toxicity, resulting in increased morbidity and mortality.

For circadian studies, these seemingly small differences in induction and recovery times are important because each group of rats (n = 6 minimum) must be sampled within 1 h of each other. For example, blood for the 0400 time point had to be collected between 0330 and 0430. As in our previous publications,^18,19,53 4 test groups each with 6 rats means that 24 rats have to be sampled in that hour, or 2.5 min for each procedure. This goal was easily achieved by using brief CO₂ anesthesia and a single investigator; however, isoflurane anesthesia required 2 investigators, who each sampled 12 rats during a 2-h period. Importantly, isoflurane anesthesia enabled only half the number of animals in twice the amount of time by using more staff, compared with CO₂ anesthesia.

Our results show that when using isoflurane anesthesia, blood melatonin levels were more than 70% lower than those in rats exposed to brief CO₂, as demonstrated in previous studies under similar conditions.^18-21 A normal melatonin signal is essential for circadian regulation of physiology and metabolism.^{18,19,22,37,41,43,46,53} Isoflurane clearly disrupts the melatonin signal and thus biochemical homeostasis. Isoflurane may have suppressed the activity of aralalkylamine N-acetyltransferase, the rate-limiting enzyme for the synthesis of melatonin from serotonin.⁴¹ Alternatively, isoflurane may have inhibited the pineal uptake of tryptophan, which is required for pineal melatonin production. Melatonin has been reported to exhibit potent analgesic properties, and its effectiveness is comparable to that of morphine.⁴⁶ The current study, to our knowledge, is the first to examine the effects of isoflurane on the nocturnal circadian melatonin signal in male Sprague–Dawley rats. Other investigators found similar rhythms of melatonin concentration after decapitation in Wistar rats⁹ as those in our Sprague–Dawley rats under CO₂ anesthesia. The preservation of the melatonin signal during CO₂ anesthesia may serve to reinforce some degree of analgesia, which might be lost during isoflurane anesthesia.

At multiple time points, corticosterone levels were higher in the isoflurane-anesthetized group than the CO₂ anesthesia group, indicative of a physiologic stress response. This result has been supported by other studies.⁵ In addition, brief CO₂ anesthesia did not increase other important stress indicators, including prolactin, corticosterone,⁵² and adrenocorticotrophic hormone,⁴⁵ when investigators compared CO₂ with decapitation without anesthesia. In general, corticosterone concentrations can vary considerably depending on animal handling procedures, various stressors, anesthetic used, and time of day;^5,16 therefore the corticosterone results should be interpreted in conjunction with other indices of animal wellbeing.

We found that, at all time points, isoflurane-anesthetized rats exhibited marked hyperglycemia compared with the CO₂ group; this finding corroborates the results of our previous studies in Sprague–Dawley rats^19,53 and reported breeder control data (mean, 146.3 mg/dL).²⁸ In addition, the insulin peak in the isoflurane group occurred 8 h prior to the CO₂ group. Isoflurane is well known to inhibit insulin secretion in rats, humans, and dogs, concomitantly causing an increase in glucose concentrations or glucose intolerance by changing normal protein, fatty acid, and glucose metabolism.^5,30,49 The current study also found that rats anesthetized with isoflurane had an abolished TFA rhythm, hyperglycemia, and a shift in the insulin peak.

Plasma TFA levels were lower in isoflurane-anesthetized rats than CO₂-anesthetized rats, and—importantly—the normal circadian rhythm had completely disappeared,⁴¹ although both groups maintained comparable body weights. Neither of our experimental groups was fasted prior to anesthesia, but a previous study found that fasted rats anesthetized with isoflurane had decreased nonesterified fatty acids compared with control animals.⁵ Our results regarding TFA concentrations and rhythmicity in CO₂-anesthetized rats are consistent with other reports^18,27 and those for isoflurane in dogs.³⁰ Isoflurane may act on the suprachiasmatic nuclei, the circadian pacemaker, in a manner that mimics the effects of constant bright light at night,¹⁹ as exhibited by the hyperglycemia, hyperinsulinemia, and the abolished TFA rhythm. These rapid and robust changes in metabolism imply a direct, central action.³⁰ Further investigation is needed in rats to determine the mechanisms of isoflurane-induced changes in the circadian regulation of metabolism.

Isoflurane causes dose-dependent respiratory depression in rats^33,36 and mice,^13,26 which in turn causes hypercapnia, respiratory acidosis, and lactic acidosis.³⁶ The isoflurane-anesthetized rats in our current study exhibited lactic acidosis, validating other studies.^5,31 Interestingly, other authors have found that isoflurane (1%) eliminates CO₂ chemosensation in rat brainstem neurons, thereby blocking the body's natural ventilatory response to hypercapnia⁴² and hypoxia,¹⁵ resulting in alterations of respiratory physiology⁴² and cardiovascular parameters^7,8 that may contraindicate halogenated anesthetics for these studies.

One surprising finding from our study is the significantly lower potassium concentrations for most time points and the 12-h shift in its peak in the isoflurane group compared with the CO₂ group. These dramatic differences in potassium in our study suggests alteration of the K⁺ channel as a viable mechanism of action.²⁵

We noted higher leptin levels in animals undergoing isoflurane anesthesia compared with brief CO₂ anesthesia. Leptin is produced by adipose tissue and acts on metabolic, respiratory, and cardiovascular physiologic functions in addition to its well-known direct influences on obesity, appetite, and metabolic syndrome.^6,47 Increased leptin concentrations typically cause a decrease in appetite in mice for as long as 3 d after anesthesia.¹³ However, we measured food and water intake every 2 to 3 d; this interval is too long for the evaluation of postanesthesia changes in food intake. Serum leptins in rats anesthetized with isoflurane in our study follow the same pattern that emerged in animals fed a high-fat diet, both having significantly reduced melatonin and effects associated with metabolic syndrome, such as hyperinsulinemia and hyperglycemia.⁴⁷ Both isoflurane and high-fat diets cause circadian disruption and increase plasma levels of these hormones, in line with obesity and metabolic syndrome.^6,47 Intense investigation is underway to elucidate the mechanism behind leptins’ actions on blood pressure and ventilatory function. Various authors suggest that leptin depends on the melanocortin system by activation of proopiomelanocortin neurons and melanocortin 4 receptors ⁶ which, on depolarization in multiple hypothalamic nuclei and downstream pathways, modulate metabolism, cardiovascular function, and respiratory function, particularly ventilatory responses to hypercapnia. Melatonin, which was decreased in isoflurane-anesthetized rats, synchronizes circadian rhythms of these physiologic functions.⁴⁶ In addition, high leptin levels abolish the Marey reflex, a baroreceptor response in the carotid sinus that helps regulate high arterial blood pressure by decreasing the heart rate.⁴⁸ When drugs or disease blunt these baroreceptors, the lack of adequate maintenance of systemic arterial blood pressure contributes to morbidity and mortality.⁴⁸ These previous authors concluded isoflurane anesthesia should not be used if lowered blood pressure, a ramped rennin–angiotensin system, and decreases of blood pressure regulation would affect the study.^8,42

In future studies, we plan to introduce an established behavioral ethogram to obtain more precise behavioral assessments between the 2 anesthesia groups. In addition, recording food, dietary and water intake, as well as body weights, at precise intervals in relation to anesthesia events and with greater frequency will allow us to correlate metabolic findings to each anesthesia event. Finally, repeating CO₂ anesthesia chamber measurements with a meter that has a higher range would enable us to quantify the exact dosage of CO₂ required for anesthesia.

The findings of our current study must be differentiated from those of studies that use CO₂ for prolonged anesthesia or euthanasia, given that they report adverse effects that are likely due to overdosage and resulting toxicity. Here we demonstrated that, unlike brief CO₂ anesthesia, isoflurane anesthesia in rats significantly disrupts normal circadian rhythms and blood levels of melatonin, TFA, potassium, glucose, corticosterone, lactate, and leptin. We conclude that brief CO₂ anesthesia is superior to isoflurane anesthesia in rats when used for short procedures, because it preserves normal physiologic circadian function that contributes to the maintenance of animal health and wellbeing and supports unambiguous scientific outcomes.