C57BL/6 male mice (7 weeks; weight range, 22–25 g) were ordered from Beijing Vital River Laboratory Animal Technology Co., Ltd., Beijing, China. Mice were bred in a temperature- and humidity-controlled room with a 12-h light/dark cycle. All mice had free access to standard mouse chow and tap water before the experiments. The animals used for the behavioral experiments were housed in the room for a week.

Sleep deprivation in mice was achieved by modified multiple platform method [4]. The mice were placed on 9 circular and 3.5 cm-diameter platforms in a ventilated transparent cage (50 $\times$ 30 $\times$ 17 cm, available food and water, 4 mice/cage) filled by water up to 1 cm of the platforms’ surface enabling the mice to jump between the platforms. Whenever the mice reached the rapid eye movement (REM) sleep, they fall into the water because of muscle atonia and then wake up and tried to climb back to the platforms to avoid being drowned. The water in the cage was refreshed every day due to the excreta of mice. The sleep deprivation periods were timed to begin and end at 8:00 AM. The non-sleep deprived mice were maintained under normal feeding conditions.

The LORR ED${}_{50}$ of sevoflurane in mice was determined according to our previously described method [5] with slight modifications. Briefly, all the behavioral interventions LORR ED${}_{50}$ of each mouse was determined, and all the values were compared. For either the control group (ad libitum activity) or the sleep deprivation group LORR ED${}_{50}$ determination, mice were individually placed in independent plastic grid V-shaped trough fixed in a transparent plastic chamber (205 $\times$ 134 $\times$ 69 mm) with an electrical fan to mix gases. One side of the chamber was connected to a sevoflurane vaporizer (Aika, Ichikawa Shiseido, Tokyo, Japan). The other side was connected to an infrared gas monitor (BeneView T5, Mindray Bio-Medical Electronics, Shenzhen, China) to measure the sevoflurane, oxygen, and carbon dioxide concentrations in real time. The monitor can monitor the sevoflurane concentration with a precision of 0.01%.

When a mouse was placed in the chamber, pure oxygen was immediately supplied at a rate of 600 mL min${}^{-1}$. When the chamber’s oxygen concentration increased to 99%, sevoflurane gas mixed in pure oxygen was provided by the vaporizer. The initial sevoflurane concentration in the chamber was 1.00%, which was maintained for 15 minutes to equilibrate the mouse with sevoflurane gas. Then, the chamber was rotated 180${}^{\circ }$ to place the mouse on its back in the V-shaped trough, and its righting reflex was observed. LORR was defined as the supine mouse unable to turn itself onto all 4 paws three times within 1 min. According to the mouse’s righting reflex, a stepwise increase or decrease of 0.10% sevoflurane in the chamber was applied. Specifically, if the mouse’s righting reflex disappeared, the sevoflurane concentration was decreased 0.10%; otherwise, it was increased 0.10%. After 15 minutes of equilibration at each sevoflurane concentration, the mouse’s righting reflex was observed again. The LORR ED${}_{50}$ was the average of the two critical sevoflurane concentrations at which the mouse either lost or regained its righting reflex. All determinations were made between 8:00 and 18:00. For the sleep deprivation group, LORR ED${}_{50}$ was determined immediately after sleep deprivation.

Sample sizes were predetermined according to our previous study [5]. GraphPad Prism software (version 8.0.2 for Windows, GraphPad Software Inc., San Diego, CA, USA) was used for statistical analyses. The acquired values of LORR ED${}_{50}$ were expressed as mean $\pm$ SD, repeated-measures analysis of variance (RMANOVA) was used to determine the significance of our behavioral interventions. Post hoc multiple comparisons were made using the Bonferroni test. A p value less than 0.05 (two-tailed) was considered to be statistically significant.

The LORR ED${}_{50}$ of sevoflurane was 1.08 $\pm$ 0.11% (95% CI, 1.01%–1.16%), 0.88 $\pm$ 0.12% (95% CI, 0.80%–0.96%), 0.89 $\pm$ 0.08% (95% CI, 0.83%–0.95%) and 0.88 $\pm$ 0.12% (95% CI, 0.80%–0.96%) for the control, sleep deprivation (24 h), sleep deprivation (48 h) and sleep deprivation (72 h) group, respectively. By comparing the LORR ED${}_{50}$ of different groups using RMANOVA, we found a statistically significant effect of different durations of sleep deprivation on LORR ED${}_{50}$ (F = 10.6, p = 0.0003).

The results of Bonferroni multiple comparisons test showed a decrease in LORR ED${}_{50}$ after 24-h (p = 0.0174), 48-h (p = 0.0043) and 72-h (p = 0.0023) sleep deprivation compared with the control group. Nevertheless, there are no significant statistical differences in LORR ED${}_{50}$ between sleep deprivation groups of different durations (p $>$ 0.9999) (Fig. 2). In other words, the sleep deprivation duration did not affect the LORR ED${}_{50}$. Because a decreased ED${}_{50}$ means increased potency of anesthetic, these findings indicated that sleep deprivation can increase anesthetic potency.

Fig. 2.

Effects of sleep deprivation on the LORR ED${}_{\mathrm{50}}$ of sevoflurane in mice. The LORR ED${}_{50}$ of sevoflurane was 1.08% (95% CI, 1.01%–1.16%), 0.88% (95% CI, 0.80%–0.96%), 0.89% (95% CI, 0.83%–0.95%) and 0.88% (95% CI, 0.80%–0.96%) for the control, sleep deprivation (24 h), sleep deprivation (48 h) and sleep deprivation (72 h) group, respectively. Data are shown as mean $\pm$ SD (n = 10/group).

The LORR ED${}_{50}$ of sevoflurane was 1.12 $\pm$ 0.08% (95% CI, 1.03%–1.20%), 1.12 $\pm$ 0.08% (95% CI, 1.03%–1.20%), 1.15 $\pm$ 0.09% (95% CI, 1.06%–1.24%) for the control, 3-day recovery from sleep deprivation (24 h) and 3-day recovery from sleep deprivation (48 h) group, respectively (Fig. 3). By comparing the LORR ED${}_{50}$ of different groups using RMANOVA, we found there was no statistically significant difference in LORR ED${}_{50}$ after 3-day recovery from sleep deprivation (F = 0.6250, p = 0.5549).

Fig. 3.

LORR ED${}_{\mathrm{50}}$ determination after 3-day recovery from sleep deprivation. SD means sleep deprivation. There was no statistically significant difference in LORR ED${}_{50}$ after 3-day recovery from sleep deprivation (p $>$ 0.9999). Data are shown as mean $\pm$ SD (n = 6/group).

Previous study had shown that circadian clock genes can affect sleep homeostasis [6] and there are circadian differences in the minimal alveolar concentration (MAC) for recovery of righting reflex (MAC${}_{\text{RORR}}$) and the time to recovery of righting reflex (Time${}_{\text{RORR}}$) [7], REM sleep deprivation can induce circadian clock gene abnormalities of rats’ hippocampus after sevoflurane inhalation [8]. So, there may be a shared component between sleep, anesthesia and circadian rhythms. We designed the above experiment to explore the relationship between sleep and anesthesia, in this animal experiment, we found an approximate 20% reduction in the LORR ED${}_{50}$ after a behavioral intervention of sleep deprivation, however the reduction did not change with the duration of sleep deprivation.

Our findings indicated that sleep deprivation can increase the anesthetic potency of sevoflurane. Previous research has been done to evaluate the time to loss of righting reflex and recovery by propofol and isoflurane after sleep deprivation, the results also suggested that sleep deprivation can potentiate the effect of anesthetics. Then how might sleep deprivation affect the anesthetic potency? Although the underlying mechanism of sleep deprivation is still unclear, there have been some discoveries about the role of neurotransmitters. For example, the concentration of adenosine in the basal forebrain of cats increased after 6 hours of sleep deprivation and decreased after recovery [9]. However, administration of adenosine antagonist cannot completely reverse the effects of sleep deprivation on righting reflex, suggesting that the effects of sleep deprivation may not be mediated by adenosine [10]. In addition, another widely concerned neuropeptides produced by the posterior hypothalamic region, which plays a key role in the maintenance of sleep and arousal, include orexin and melanin concentrating hormone (MCH). Orexin is a neuropeptide family secreted by hypothalamus that promotes appetite and regulates sleep and wakefulness, it consists of two peptides orexin-A and orexin-B. Dong et al. [11] found that intrabasalis microinjection of orexin-A can shorten the emergence time to sevoflurane anesthesia and the orexin receptor antagonist (SB-334867A) prolonged the emergence time to sevoflurane anesthesia. The decrease of orexin-A is also the reason for the delayed recovery of sleep deprived rats under isoflurane anesthesia [12]. Whereas MCH perform opposite roles to orexin in sleep and wakefulness [13]. In the rebound phase of REM sleep after 72 hours of REM sleep deprivation, MCH${}^{+}$ neurons were strongly active, and rapid eye movement sleep was prolonged after lateral ventricular injection of MCH [14]. Systemic application of MCHR-1 antagonist resulted in a dose-dependent reduction of slow wave (SW) sleep and REM sleep, and a corresponding increase in wakefulness [15]. Orexin and MCH were also found to be related with the effects of ketamine and propofol on sleep structure in the period of postanethesia [16]. Other neurotransmitters include glutamate (Glu), and $\gamma$-amino butyric acid (GABA) also play an important role in the wake/sleep cycle, Xie et al. [17] found that 24 h sleep deprivation significantly increased the concentration of Glu and GABA in the rat’s hippocampus and propofol anesthesia can normalized the upregulated GABA and Glu levels like natural sleep.

We chose LORR ED${}_{50}$ as an observed index because this index was more stable compared with the time to loss of righting reflex or the emergence time from anesthesia, and it only takes 0.5–1 h to complete a LORR ED${}_{50}$ test, which is not too long. Induction time is not an index to evaluate the potency of inhaled anesthetics, for example, sevoflurane induces anesthesia faster than isoflurane, but sevoflurane is less potent than isoflurane [18]. LORR ED${}_{50}$ has been commonly used to evaluate the potency of inhaled anesthetics in inducing unconsciousness in mice [5, 19]. Furthermore, although studies demonstrated that prolonged sedation with propofol can discharge the sleep debt [20, 21], when exposed to the inhaled anesthetics such as sevoflurane, isoflurane and halothane, rapid eye movement sleep debt accrues in mice [22, 23]. Pal et al. [24] also found that sevoflurane induction time was shortened after sleep deprivation and REM sleep could not be restored during sevoflurane anesthesia. Therefore, we did not need to consider the recovery of REM sleep that may be caused by the duration of LORR ED${}_{50}$ test. We did not use the minimal alveolar concentration (MAC) as a measure to evaluate the anesthetic potency for following reasons. At first, MAC involves the painful stimulation during tail-clamp, while sleep loss would increase the pain sensitivity in mice [25, 26, 27]. In addition, MAC was used to evaluate the anesthetic potency of inducing immobility, and the value was not altered after spinal cord transection in rats which means that MAC associated body movements may be in the spinal cord [28], nevertheless, the potency of anesthetics to induce hypnosis is measured by loss of righting reflex and hypothalamic nuclei and cerebral cortex may be involved in this process [29].

There are many ways to model sleep deprivation, in this study we chose the modified multiple platform method for several reasons. Above all, we want to explore the effects of REM sleep deprivation on anesthetic potency as the determination of LORR ED${}_{50}$ required a period of 0.5–1 h though it is not too long, we do not know that if the sleep debt can be discharged during anesthesia, whereas as mentioned above, inhaled anesthetics cannot satisfy the homeostatic need for REM sleep, so we can exclude the effect of anesthetic time on our results. Also, modified multiple platform method was widely used for REM sleep deprivation [4, 30, 31] and it is convenient to model.

Our results demonstrated that preoperative sleep disorders may affect perioperative anesthetic managements. Nowadays, sleep disorders are regarded as an independent risk factor of postoperative cognitive dysfunction (POCD) which is a severe postoperative neurological sequela. So, for the anesthesia of such patients, we can consider reducing the dose of anesthetics during operation, this may help to reduce the incidence of POCD, however, this hypothesis needs to be further verified.

This study only presents a behavioral finding and does not further investigate the mechanism. Therefore, the next step is to explore the basic mechanism of sleep deprivation leading to the increased anesthetic potency and try to uncover the relationship between sleep and anesthesia.

In addition, it can also be clinically explored whether reducing the dosage of anesthetics in patients with sleep disorders can reduce the incidence of POCD.