Exhaled breath is found to be better than blood samples for determining propofol concentrations in the brain tissues of rats

The correlation between propofol concentration in exhaled breath (C E) and plasma (C P) has been well-established, but its applicability for estimating the concentration in brain tissues (C B) remains unknown. Given the impracticality of directly sampling human brain tissues, rats are commonly used as a pharmacokinetic model due to their similar drug-metabolizing processes to humans. In this study, we measured C E, C P, and C B in mechanically ventilated rats injected with propofol. Exhaled breath samples from the rats were collected every 20 s and analyzed using our team’s developed vacuum ultraviolet time-of-flight mass spectrometry. Additionally, femoral artery blood samples and brain tissue samples at different time points were collected and measured using high-performance liquid chromatography mass spectrometry. The results demonstrated that propofol concentration in exhaled breath exhibited stronger correlations with that in brain tissues compared to plasma levels, suggesting its potential suitability for reflecting anesthetic action sites’ concentrations and anesthesia titration. Our study provides valuable animal data supporting future clinical applications.


Introduction
Real-time monitoring the concentration of inhaled anesthetics to adjust the depth of anesthesia is common in clinical practice.Although propofol has also been found detectable in exhaled breath, there are no suitable detection methods in operation rooms.The traditional technique for off-line breath analysis is gas chromatography-mass spectrometry (GC-MS), which is operation-complicated and timeconsuming.[1] In the last two decades, different mass spectrometers have been developed to achieve the goal of on-line propofol monitoring, such as proton transfer reaction mass spectrometry, [2,3] selected ion flow tube mass spectrometry [4] and ion molecule reaction mass spectrometry (IMR-MS) [5].However, these instruments still have disadvantages such as loud noise and large volume [6].Although devices based on ion mobility spectrometry (IMS) are small in size [7,8], such as Edmon, a product based on multi-capillary column-IMS (MCC-IMS) launched in 2017, its sensitivity can be easily affected by moisture [9].
Our team developed an on-line mass spectrometer that combines a vacuum ultraviolet (VUV) ionization source with time-of-flight mass spectrometer (TOF-MS) to analyze propofol concentration in exhaled breath (C E ) in real time.The ionization source is a krypton discharge lamp to produce photons with 10.6 eV energies and most volatile organic compounds in exhaled breath can be ionized directly and simultaneously.When combined with TOF-MS, it enables on-line analysis with high sensitivity down to the parts-per-trillion level.Due to its miniaturization and ease of movement, the instrument holds great potential for utilization in surgical settings.Furthermore, studies have demonstrated that similar mass spectrometry techniques exhibit the capability to detect inhaled anesthetics, thereby expanding the future clinical applications of this instrument [10][11][12].
Previous studies have shown that C E correlates well with plasma propofol concentration (C p ) [5,13].However, the anesthetic sites of propofol are in the brain and not in the plasma.Therefore, it is crucial to investigate the correlations between C p and propofol concentration in brain tissues (C B ).Studies have shown that in human studies, the bispectral index (BIS), which represents the anesthetic effect of propofol, has a similar time to C E for time to detection and time to peak indices [14,15].An animal study by Müller-Wirtz et al in rats initially showed that C B of propofol (R 2 = 0.75) correlated better with C E after 6 h of continuous infusion than with C p (R 2 = 0.71) [16].However, the presumed delay between C p and C B caused by the continuous infusion route of administration cannot be determined.Other studies have suggested that there is a delay in C E , [14,17] and whether it is more related to C B remains uncertain.Furthermore, Müller-Wirtz et al concluded that the time deviation of the measurements was up to several minutes, resulting in the significant difference in C B values, which should be further investigated and analyzed.
Given the impracticality of directly sampling human brain tissues, rats are commonly used as a pharmacokinetic model due to their similar drugmetabolizing processes to humans.Therefore, we analyzed the C E of propofol in rats after the bolus injection using the VUV TOF-MS, and compared the relationships between C E , C p and C B to explore whether the propofol concentration in exhaled breath is more suitable to reflect the concentration at anesthetic effect sites and for titrating anesthesia, aiming to provide valuable animal data for future clinical applications.

Structure of VUV-TOFMS
The VUV-TOFMS instrument has the size of is 520 mm (L) × 550 mm (W) × 940 m (H), with casters on the bottom for easy movement.The internal structure includes the VUV light, RF ion guide device, electrostatic lenses and TOF-MS.A 1.5 m heated (T: 100 • C) polyetherketone (PEEK) tube (OD: 2.5 mm, ID: 2 mm) is used to continuously feed the gas samples into the ionization chamber, which is maintained at 500 Pa by a molecular pump.The VUV light is a krypton discharge lamp to produce photons with 10.6 eV energies.The wavelengths provided by the light source is 116-117 nm.The resolving power of the home-built TOF-MS is 2500 at mass-to-charge ratio (m/z) of 115.The reflectron-type TOF analyzer features a 265 mm flight tube and the voltage of the MCP detector is 2400 V.In addition, the builtin visual control software can directly set detection parameters such as carrier gas flow rate, ionization voltage, and acquisition frequency.The data analysis system is capable of effectively filtering and denoising the aforementioned data, ultimately enabling precise quantification of propofol levels in gas samples.The self-constructed calibration gas generator was employed to generate propofol calibration gas with varying concentrations, as described in the referenced literature [18,19].Prior to each experiment, the VUV TOF-MS instrument underwent calibration to mitigate inter-day variability.Detailed information on the internal structure, operation process and calibration methods of VUV TOF-MS can be referenced to our previously published studies [18,19].

Experimental animals
Forty-two male Sprague-Dawley rats weighing 300-400 g and aged 8-10 weeks (Chengdu DOSSY Experimental Animals Co., Ltd, Chengdu, Sichuan, China) were maintained in animal rooms (Tianfu Life Science Park of West China Hospital of Sichuan University, Chengdu, Sichuan, China) with temperature at 20 ± 2 • C and humidity at 50 ± 5% and free access to water and food.The experiment was approved by the Animal Ethics Committee of West China Hospital of Sichuan University (Chengdu, Sichuan, China) (20220420001) and in accordance with animal welfare requirements.The experimental procedure is shown in figure 1.

Anesthesia, preparation, and monitoring
All rats were fasted for 12 h before the experiment and allowed to have free access to water.Anesthesia was induced by intraperitoneal injection of 10% pentobarbital sodium (Sinopharm Chemical Reagent Co., Ltd, Shanghai, China) at 35 mg kg −1 .After the loss of righting reflex, the rats were fixed in the supine position on a warmed plate, and the tail vein was punctured and catheterized for drug administration.Tracheotomy was performed for tracheal intubation and the trachea and tube were fixed tightly to prevent air leakage.The parameters of the ventilator (R407, RWD Life Technology Co., Ltd, Shenzhen, China) were set as follows: 100% pure oxygen, tidal volume of 9 ml kg −1 , initial respiratory rate of 50 min −1 , and the inhalation-exhalation ratio of 1:1.End-tidal carbon dioxide partial pressure (P ET CO 2 ) was continuously monitored (BeneView T6/T8, Mindray, Shenzhen, China).Respiratory rate was decreased by 10% when P ET CO 2 was <25 mmHg and increased 10% when P ET CO 2 > 45 mmHg.One of the femoral arteries was catheterized and connected with the monitor (BL-420 F, Chengdu Techman software Co., Ltd, Chengdu, China) for blood pressure monitoring, and the other was catheterized for blood sampling.Throughout the procedure, the body temperature was maintained at 36.5 ± 0.5 • C.During the whole period, the exhaled air was continuously sampled and analyzed.Upon reaching the corresponding measurement time point, the rats were immediately sacrificed by decapitation, and blood and brain tissue samples were collected.

Propofol administration method and dosage
Forty-two rats were randomly divided into seven groups (n = 6) according to the measurement time points using the random number generator (Excel 2019, Microsoft, Redmond, WA): 1 min, 3 min, 5 min, 7 min, 10 min, 20 min, 30 min.The rats were administered propofol (10 mg ml −1 ) at a dosage of 12 mg kg −1 , which corresponds to the 2ED 50 dosage resulting in the loss of the righting reflex [20].The administration was performed using a micro-infusion pump (BeneFusion n series, Shenzhen Mindray Bio-Medical Electronics Co., Ltd, Shenzhen, China), with a steady injection rate of 0.4 ml/30 s.

Breath samples collecting and propofol measurement
A three-way medical valve (made of polycarbonate) connected the tracheal cannula (made of polyurethane) of the breathing circuit (made of polytetrafluoroethylene) was directly coupled to the VUV TOF-MS instrument.The exhaled air of the rats was sampled every 20 s without sample pre-separation and analyzed directly by the VUV TOF-MS instrument.The detection conditions were set as follows: voltage was adjusted to 77 V; the flow rate of carrier gas was maintained at 70 ml min −1 ; the ion source temperature was controlled at 100 • C; the sampling tube was kept constant at 100 • C and the sampling time was 20 s.The selection of m/z 177.6 as the mass peak of propofol was based on rigorous testing of propofol standards using VUV TOF-MS.Furthermore, the absence of this peak in the mass spectrogram of blank exhaled breath further supports its specificity for propofol detection.Additionally, our investigation revealed a consistent correlation between changes in m/z 177.6 and drug concentration in vivo following intravenous administration of propofol to rats.C E was calculated from the calibration curve of intensity (peak area) versus propofol concentration (parts per billion by volume, ppbv) measured by the VUV TOF-MS instrument.

Blood and brain tissue sampling and measurement
Arterial blood samples (200 µl each) were obtained from the femoral artery at 0, 1, 3, 5, 7, 10, 20 or 30 min after propofol injection.Brains tissues were immediately removed after decapitation, frozen in liquid nitrogen and stored at −80 • C in a refrigerator.For plasma samples, 50 µl blood supernatant was transferred to EP tubes after centrifugation (3500 rpm for 10 min at 4 • C) to which 150 µl of thymol was added.After the second centrifugation (20 000 rpm for 10 min at 4 • C), 100 µl of supernatant from each sample was transferred to vials for chromatographic separation and measurement.For brain tissues samples, 0.1 g of deep-frozen brain tissue diluted in 1 ml 0.9% of saline was homogenized and then centrifuged at 3500 rpm for 10 min at 4 • C. 50 µl of supernatant from each sample was transferred to an EP-tube containing 150 µl of thymol and then centrifuged at 20 000 rpm for 10 min at 4 • C. Finally, 100 µl of supernatant from each sample was transferred to vials for chromatographic separation and measurement.
The instrument employed in this study for quantifying propofol concentration in plasma or brain tissues was a Waters Model 2695 high-performance liquid chromatography mass spectrometry (HPLC-MS, Waters, USA) coupled with a Waters model 2475 fluorescence detector.Experimental conditions were as follows: the column used was SWELL C18 (150 mm × 4.6 mm, 5 µm); maintained at a temperature of 30 • C; the mobile phase consisted of pure water and acetonitrile (38:62, v/v); fluorescence detection occurred at an excitation wavelength (Ex) of 276 nm and an emission wavelength (Em) of 310 nm [21,22]; the flow rate was set to be 1.0 ml min −1 ; injection volume amounted to 10 µl; retention times were observed at thymol internal standard peak (3.9 min) and propofol peak (7.4 min).

Statistical analysis
Statistical analysis was performed using Graphpad Prism software (Graphpad Prism, USA).All data were presented as mean (standard deviation, SD).To directly compare C P and C B , the specific density of 1.03 g ml −1 for plasma was assumed in our study [23].The t-test was performed to compare C p and C B .Statistical significance was set at P < 0.05.
Based on published studies, the correlation of C E and C P was fitted with the linear regression model and the degree of fit (R 2 ) was calculated.The higher the R 2 , the better the correlation.According to our previous studies, the correlations of C E or C P and C B were evaluated with the one-phase exponential association model.The least squares method was used to obtain the best model parameters.Y 0 , Plateau and K are the model parameters The linear regression equations of the predicted propofol concentration in brain tissues (Pre C B_E , C B predicted by C E ; Pre C B_P , C B predicted by C P ) calculated with the above formula and the actual C B (Act C B ) were constructed, and the degree of fit (R 2 ) was compared.Meanwhile, the Bland-Altman analysis was performed to evaluate the agreement of Pre C B_E or Pre C B_P with C B .The difference of the measured results was taken as the vertical axis and the mean of the measured results was taken as the horizontal axis.The scatter plot is drawn and the 95% confidence interval (CI) is marked.If the difference between the two measurement results is within the 95% CI, it can be assumed that the measurement results of the two methods are in good agreement.In addition, when comparing Pre C B_E , Pre C B_P with Act C B , the closer the bias is to 0, the better the agreement.Comparisons were made between propofol concentrations in plasma and brain tissues, assuming a specific density of 1.03 g ml −1 for plasma.

C E , C P and C B
The calibration information of VUV TOF-MS and HPLC-MS is provided as supplementary material in appendix 1.A total of 42 male rats from the same batch were involved in this experiment, the average weight of 348.27 ± 20.24 g.After a bolus injection of 6 mg −1 , the dose of propofol administrated was 3.91 ± 0.28 mg.The values of C E ranged from 0.63 (0.11) ppbv to 11.43 (5.49) ppbv within 30 min (figure 2).The values of T max of C E , C P and C B were all at 1 min.The values of C max were 9.36 (2.37) ppbv of C E ; 4.69 (0.95) µg ml −1 of C P ; 8.33 (0.93) µg g −1 of C B , respectively.The ratio of C E /C P ranged from 0.05% to 0.24%.Meanwhile, the values of C B were higher than that of C P at all measurement points (P < 0.05).

Discussion
In our study, the concentration of propofol in exhaled air was measured on-line using the VUV TOF-MS instrument.In addition, propofol concentration in plasma was analyzed simultaneously with that in brain tissues using HPLC-MS.We found that both propofol concentration in plasma and exhaled air had good correlations with that in brain tissues, while propofol concentration exhaled air had a better correlation with that in brain tissues.
The VUV TOF-MS we developed has the characteristics of compact size, low noise and easy to move, rendering it suitable for implementation in operating rooms.The VUV light ion source emits photons with an energy of 10.6 eV to softly ionize volatile organic compounds in gas samples rapidly.The study by Chawaguta et al showed that soft chemical ionization mass spectrometry can detect propofol and its major metabolites in real time with high levels of confidence by monitoring specific different product ions [3].The photoionization limits the formation of fragment ions, facilitating the facile separation of propofol from other volatiles present in exhaled breath.Moreover, only volatiles with ionization energies lower than 10.6 eV can be ionized while major constituents of exhaled air (i.e.N 2 , O 2 , and CO 2 ) remain unaltered.The m/z of propofol measured in our research was 177.6 instead of 178.1 due to the propofol mass exceeding our mass calibration range.The mass calibration method in our study was introducing benzene, oxygen, and tetrachloroethylene into the VUV TOF-MS.The m/z of the three characteristic peaks and their flight time in the TOF-MS were used to fit the curve for calibration, leading to a deviation at high m/z.The actual m/z for propofol is 178.1, which refers to the singly ionized parent.The mass calibration is out by −0.5 amu, which is not significant given the resolving power of this instrument.However, the m/z of 177.6 in this experiment indeed corresponds to the characteristic peak specific to propofol, as evidenced by its presence in both the propofol standard gas and the exhaled breath samples.This VUV TOF-MS achieved a limit of quantification at 0.12 ppbv, satisfying the requirements for measuring exhaled propofol concentration in humans.Compared with the device of Edmon, our apparatus experiences less interference from water resulting in more stable output results [9].Our pilot study demonstrated its capability to simultaneously determine the concentrations of various anesthetics, including propofol, ciprofol [18,19], fos-propofol, sevoflurane, and desflurane.This finding highlights its enhanced clinical practicality.
In our study, the human equivalent dose of propofol was injected, and the exhaled propofol concentration in rats was 0.63 (0.11) to 11.43 (5.49) ppbv, which is similar to that in humans [5,13,14].However, the exhaled propofol concentration demonstrated a significant decrease for the same plasma concentration when compared to Müller-Wirtz's continuous infusion data in rats.Because no other exhalation analysis studies have been published in rats, and propofol in exhaled breath varies widely between species [24], we cannot determine whether the differences are due to instrumental performance.However, inconsistencies in testing standards between studies may be a potentially important reason, which is also a critical issue for breath research.Breath composition and compound concentrations can be strongly influenced by external factors such as recent activities, dietary intake, medications, and environmental exposures.Bruderer suggested that online studies require the same instrumentation at each site, with highly standardized procedures and trained operators [25].
Propofol can easily cross the blood-brain barrier and cause rapid loss of consciousness.Our results showed that propofol concentrations in plasma or brain tissue peaked at 1 min after a single intravenous injection.The higher propofol concentration in brain tissue than in plasma may be related to its lipid solubility.Repeatedly measured exhaled and plasma propofol concentrations showed a strong correlation as assessed by the established linear regression model, which is higher than that in rats using IMS (R 2 = 0.71) [16], and similar to results reported in humans using the headspace solid-phase microextraction coupled to a GC-MS instrument (R 2 = 0.85) [26], compared to previous observations.Although the volume of exhaled air in the rat experiments is small, the VUV TOF-MS instrument still demonstrated good performance.In fact, the VUV TOF-MS instrument can achieve on-line analysis in 1 s.However, sampling was extended to 20 s intervals to reduce the influence of ventilation, which is also considered the cut-off value for real-time monitoring [27], thus avoiding the loss of important propofol concentration information.
Accurate assessment of the sedative effects of propofol to avoid too deep or light anesthesia has always been the pursuit of anesthesiologists.Propofol concentration in the central nervous system is crucial in influencing the sedative effects and serves as a fundamental basis for anesthesiologists' decisionmaking.However, direct detection of propofol concentration in clinical anesthesia through brain tissue or cerebrospinal fluid sampling is impractical.As a result, it is often indirectly reflected by invasive and discontinuous plasma propofol concentration, which is time-consuming and unsuitable in clinical practice.In contrast, exhaled breath sampling offers noninvasive and continuous sampling characteristics [25].Additionally, although direct detection of cerebral propofol concentration during clinical work remains unattainable, sedative effects are often quantified through EEG analysis and conversion.Wellestablished indicators such as BIS or Narcotrend index have proven effective [28,29].Several studies have attempted to establish correlations between exhaled propofol concentration and EEG indicators.For instance, Hornuss used IMR-MS to measure the time course of exhaled propofol concentration in 21 patients and monitored BIS as the cerebral propofol effect.They found that after an IV bolus dose, exhaled propofol concentration measured by IMR-MS had similar time to detection and peak concentration compared with BIS, suggesting that exhaled propofol concentration may be useful for titrating intravenous anesthesia [14].Similarly, Liu et al found that the exhaled propofol concentration measured by IMS was consistent with the propofol effect represented by the BIS and concluded that monitoring exhaled propofol may improve the safety of anesthesia [15].But the correlations were not good in all studies.In pediatric patients, Heiderich et al used MCC-IMS to demonstrate moderate correlations between exhaled propofol concentration and the Narcotrend Index [30].The study of Braathen et al found that the modest correlation whether in normal-weight or obese surgical patients makes the bedside online IMS device's everyday clinical usefulness questionable [31].In further exploration, Müller-Wirtz et al employed MCC-IMS to directly compare the concentrations of propofol in exhaled breath and brain tissues, revealing a significant correlation between them from a pharmacokinetic perspective [16].Our study demonstrated that propofol concentration in exhaled breath exhibits a stronger correlation with that in brain tissues compared to that in plasma in rats, providing valuable animal data for its potential clinical practicality in the field of anesthesia.
The observed ceiling effect on propofol brain concentrations compared to plasma and exhaled concentrations in the one-phase association fitting models may be caused by the single injection methodology in this study.After intravenous injection, most of propofol accumulates in blood vessels, leading to a sudden increase in blood concentration.Simultaneously, rapid gas exchange takes place once propofol reaches the lungs, allowing free propofol present in blood vessels to appear in exhaled breath as it attains saturated vapor pressure (0.142 mmHg at 20 • C) [32].Subsequently, propofol is transferred from blood vessels to other tissues, including the effect-site brain, resulting in a rapid decline in the concentration of propofol in blood vessels.However, the concentration of propofol in brain tissues decreased relatively slowly after reaching the peak.
Certainly, our study has some limitations.Firstly, the sampling method only allows for collection of brain tissue samples at a single time point, which results in relatively high animal and economic costs.In future research, continuous sampling techniques similar to microdialysis technology should be adopted to further optimize the experimental design.Secondly, several factors can easily influence exhalation analysis in rats, including the small exhalation volume, unstable positioning of the tracheal catheter against the lung, and a relatively large blood sample volume.Frequent blood sampling may decrease the hemoglobin content, erythrocyte count, and plasma protein levels, potentially affecting propofol distribution and reducing free propofol volatilization from blood into alveolar air [33].However, single blood sampling in this experiment did not impact actual blood volume in rats.Thirdly, exhaled propofol concentrations are influenced by pulmonary blood flow and cardiac output [34].During basal anesthesia with varying volumes of pentobarbital sodium administration in rats, non-uniform effects on pulmonary blood flow and cardiac output may occur; thus affecting the ratio of pulmonary ventilation to blood flow which could subsequently impact exhaled gas concentration measurements.Further research is required to confirm these effects on final results.Additionally, in our previous study we found that this instrument can measure multiple anesthetics simultaneously; however, this current study only explored propofol measurement, and future studies should also investigate simultaneous measurement of multiple anesthetics.Finally, due to lack of EEG monitoring in rats, the proximity between propofol concentration in exhaled air and its cerebral site of action needs further elucidation through future studies.

Conclusions
The VUV TOF-MS instrument enables online detection of exhaled propofol in rats, with the exhaled concentration showing a stronger correlation to that in brain tissues compared to plasma concentration.This suggests that exhaled propofol may be a more suitable indicator for titration of anesthesia.The rat data obtained from this study provides valuable support for future clinical human studies.

Figure 1 .
Figure 1.Schematic representation of the comparison of propofol concentrations in exhaled breath, plasma, and brain tissue in rats.VUV TOF-MS, vacuum ultraviolet time-of-flight mass spectrometry; HPLC-MS, high-performance liquid chromatography mass spectrometry.Propofol concentrations in plasma and brain tissue samples were collected at 1, 3, 5, 7, 10, 20 and 30 min.Exhaled breath was measured by VUV TOF-MS continuously, plasma and brain samples were measured by HPLC-MS, and the correlations between them were established.

Figure 2 .
Figure 2. The fitted time-course curves of CE, CP and CB.CB: propofol concentration in brain tissues; Cp: propofol concentration in plasma; CE: propofol concentration in exhaled breath; ppbv: parts per billion by volume.Comparisons were made between propofol concentrations in plasma and brain tissues, assuming a specific density of 1.03 g ml −1 for plasma.

Figure 3 .
Figure 3.The fitting models of CE, CP and CB.CB: propofol concentration in brain tissues; Cp: propofol concentration in plasma;CE: propofol concentration in exhaled breath; ppbv: parts per billion by volume.Pre CB_P, the predictive propofol concentration in brain tissues according to the blood concentration of propofol; Pre CB_E, the predictive propofol concentration in brain tissues according to the exhaled concentration of propofol; Act CB, the actual propofol concentration measured in brain tissues.The correlation of propofol concentration in exhaled breath and plasma can be well fitted with the linear regression model, while the correlations of propofol concentration in exhaled air or plasma with that in brain tissues can be fitted with the one-phase exponential association model.

Figure 4 .
Figure 4.The Bland-Altman analysis of CE, CP and CB.CB, propofol concentration in brain tissues; Cp: propofol concentration in plasma; CE: propofol concentration in exhaled breath; ppbv: parts per billion by volume.Pre CB_P, the predictive propofol concentration in brain tissues according to the blood concentration of propofol; Pre CB_E, the predictive propofol concentration in brain tissues according to the exhaled concentration of propofol; Act CB, the actual propofol concentration measured in brain tissues.