During clinical practice of anesthesia, patients are under various blood partial pressures of CO₂ by artificial ventilations or by external CO₂ loading such as intra-abdominal insufflation during laparoscopic surgery or video assisted endoscopic surgery to ensure visibility in the subcutaneous organs, such as robot-assisted thyroidectomy. One of the adaptations of human body to maintain homeostasis is to maintain a normal range of pH between 7.35-7.45 [1]. pH maintenance within this range is mainly governed by a buffer system based on bicarbonate ions, with the involvement of the renal system, and carbon dioxide regulation by the respiratory system [1,2]. This mechanism allows for partial or even complete compensation for certain disturbances, whereby the pH remains within the physiological range despite the presence of disruptions. A pathological condition in which the pH falls below 7.35 is called acidosis, and a condition in which it rises above 7.45 is called alkalosis [2,3]. These acid-base changes not only affect the efficacy of non-depolarizing neuromuscular blockers such as vecuronium or pancuronium, but also indirectly induce muscle contractions [4]. Respiratory alkalosis is known to delay the action of rocuronium, and it is also considered to delay the effects of rocuronium during hyperventilation [5]. On the contrary, in respiratory acidosis induced by hypoventilation, it was also discovered that the duration of neuromuscular blockade induced by neuromuscular blocking agents (NMBAs), such as rocuronium could be prolonged [6]. Therefore, we hypothesized that changes in pH induced by CO₂ loading or washout during anesthesia could influence neuromuscular blockade and recovery.

Sugammadex is a chelating agent that reverses neuromuscular blockade induced by NMBA, which have a steroidal backbone [7]. Sugammadex has a well-defined lipophilic internal cavity and can encapsulate lipophilic guest molecules, such as steroids, forming a host-guest inclusion complex also known as chemical encapsulation [7]. It has a modified γ-cyclodextin which encapsulates charged molecules of steroidal NMBAs forming a stable complex which prevents the pharmacological action of the NMBA [7,8]. It was speculated that if the charged form of steroidal NMBAs is affected by environmental pH, the interaction of this host-guest reaction may be affected, resulting in change in the sugammadex-induced recovery profile. This study investigated the effect of acid-base imbalance induced by different mixed gases containing different percentages of CO₂ on rocuronium-induced neuromuscular blockade and its reversal by sugammadex, focusing on both first twitch tension (T1) of the train-of-four (TOF) stimulation and TOF ratio (TOFR) and in an ex vivo rat phrenic nerve-hemidiaphragm model, as well as the relationship between these parameters.

After receiving approval from the Asan Ethics Committee of the Laboratory of Animal Research on May 19, 2023 (Protocol No. 2023-13-083), phrenic nerve-hemidiaphragm specimens were obtained from 30 adult Sprague-Dawley rats (male; mean weight 315.2 g, range 278.3-355.0 g).

Each rat was anesthetized with an intraperitoneal injection of urethane (0.5g/kg). The thoracic cage was immediately isolated and phrenic nerve-hemidiaphragm tissue specimens were obtained. The sampled tissues were immersed in Krebs buffer solution (120 mM NaCl, 2.5 mM CaCl₂, 4.7 mM KCl, 1.2 mM MgSO₄, 1.2 mM KH₂PO₄, 25 mM NaHCO₃, and 11 mM α-D-glucose) maintained 95% O₂ and 5% CO₂ with continuous bubbling to ensure tissue viability throughout the experimental session; after group assignment, mixed gas was changed at respiratory acidosis (97.5% O₂ + 9% CO₂) and respiratory alkalosis (91% O₂ + 2.5% CO₂) groups. And the temperature of the bath was maintained at 35°C with a warm water circulator. The tissue was fixed to a frame with platinum electrodes and immersed in a 100 ml organ bath containing 75 ml of oxygenated Krebs buffer solution. Subsequently, the tendinous portion of the diaphragm of each sample was hooked to a Grass FT03 Force Transducer (Grass Technologies), and 40 mN of resting tension was applied to create preload tension. The phrenic nerve was fixed to a platinum bipolar electrode and stimulated using a Grass S88 Stimulator (Grass Technologies). Supramaximal stimulation with 0.2 ms of a square wave pulse was administered every 20 s in trains of 2 Hz for 2 s (TOF stimulation). All waveforms were acquired using the PowerLab 4/26 Data Acquisition System (AD Instruments) and displayed, monitored. and stored by LabChart 7 Software (AD Instruments).

Specimens were randomly allocated to control (95% O₂ + 5% CO₂), respiratory acidosis (91% O₂ and 9% CO₂) respiratory alkalosis (97.5% O₂ and 2.5% CO₂). After a 30-min stabilization period, rocuronium initial loading (500 μg) and booster (250 μg) doses were serially administered at 10-min intervals in each sample until > 95% depression of T1 was achieved. Ten minutes of rest time passed, confirming that T1 responses were completely diminished. Recovery progression was followed to produce > 95% recovery of T1 tension using the sugammadex concentration, which was set to a 1:1 equimolar dose of rocuronium. Recovery from neuromuscular blockade was simultaneously monitored using T1 and TOFR until the recovery of T1 to > 95% and TOFR to > 0.9. The buffer solution was sampled at the end of study, and the pH of the solution was measured by portable pH tester (Caleo International Ltd). Study protocol is summarized at Fig. 1.

There were no statistically significant intergroup differences in body weight of rats, wet weight or size of the phrenic nerve-hemidiaphragm tissue specimens (Table 1). The amount of rocuronium used for > 95% blockade was statistically different in the acidosis group compared to the control or alkalosis groups (Table 2). As sugammadex amount at a 1:1 equimolar dose was used, sugammadex also statistically different in acidosis group compared to other groups. The total dose of rocuronium and sugammadex differed between the respiratory acidosis and alkalosis groups (P = 0.001, and P = 0.001 respectively; Table 2).

Different CO₂ loading resulted in intergroup differences in the pH of the buffer solution (Table 1). However, no intergroup differences were observed in the time to > 95% T1 twitch tension, RI, or time to > 0.9 TOFR (Table 2).

The rocuronium dose-response curve of T1% in the respiratory acidosis group was shifted to the left (P = 0.001) compared to that in the respiratory alkalosis group. Mean regression coefficients were -3.0 × 10^-5, -0.99 × 10^-5, and -2.0 × 10^-5 in the control, the respiratory acidosis, and the respiratory alkalosis groups, respectively, and showed a statistical difference between the respiratory acidosis and the respiratory alkalosis groups and between the acidosis and alkalosis groups (P < 0.001, Fig. 2A). However, there was no difference between those in control and the respiratory alkalosis groups (P = 0.026).

The rocuronium dose-response curves of TOFR in each group were also compared. Mean regression coefficients (a) were -2.6 × 10^-7, -2.3 × 10^-7, and -2.1 × 10^-7 in control, the respiratory acidosis, and respiratory alkalosis groups, and there were intergroup differences in all groups (P < 0.001, Fig. 2B).

The recovery progression pattern of the sugammadex-induced reversal of T1 was quite different among the three groups. Mean regression coefficients of T1 in each group are -1.038 × 10^-5, -1.205 × 10^-5 and -1.000 × 10^-5 in the control, the respiratory acidosis, and alkalosis groups, respectively, and showed intergroup differences (P < 0.001, Fig. 3). However, the regression curves obtained from TOFR did not fit well (R² < 0.4) with the nonlinear mixed model. The recovery progressions of TOFR in each group were not plotted well, and their coefficients could not be set. Therefore, intergroup comparisons of the recovery coefficients of TOFR were not performed.

All groups recovered to > 95% of T1 within 1 h and intergroup comparisons of the time from 5% to 95% showed no statistically significant differences. The time from 25% to 75% T1 recovery (RI) was measured and no intergroup differences were observed (Table 2). The time from sugammadex administration to the reappearance of the T1 response tended to be delayed in the respiratory alkalosis group, although the difference was not statistically significant.

The present study found that a change in CO₂ could change the pH of the buffer solution and affect neuromuscular blockade of rocuronium and the recovery progression of sugammadex. Although eye-catching differences could not be displayed in the regression lines of blockade and recovery progression, statistical differences could be obtained in some variables among the groups.

In Sprague-Dawley rats, apneic threshold under urethane anesthesia was evaluated, and the mean PaCO₂ level was 32.8 ± 0.4 mmHg [9]. Volatile inhalation anesthetics cause a rightward shift in the apneic threshold [10], which is the minimum PaCO₂ required to initiate spontaneous respiration. Thus, spontaneous breathing efforts do not occur if mechanical or assisted ventilation efforts drive PaCO₂ levels below the CO₂ threshold during general anesthesia [11]. To overcome these problems, hypoventilation and CO₂ retention are often induced at the end of surgery, to stimulate the respiratory drive center, and initiate spontaneous respiration. To promote rapid recovery from general inhalation anesthesia, artificial hyperventilation is also used to wash out volatile inhalation anesthetics such as desflurane or sevoflurane. However, based on these present results and the report of Ono et al. [4], delayed recovery or deterioration of the blockade might occur, resulting in respiratory distress, although the respiratory drive is regained. Although the number of ionized/unionized molecules changed during hypoventilation, this alone does not explain the altered potency of NMBA during pH changes. NMBA are relatively hydrophilic drugs compared with other anesthetic drugs and therefore usually remains in completely ionized at pH 7.4 [12], pH changes might be minimal within the range used during anesthesia. As such, it is speculated that homeostasis occurs because of pH changes at the neuromuscular junction (NMJ). It has been suggested that protons, which are co-released with neurotransmitters at the NMJ of mice may act as a trigger for the up-regulation of neurotransmitter output (measured by quantal content, QC), which is caused by retrograde signal through the activation of acid-sensing ion channels (ASICs) [13,14]. Presynaptic regulation (presynaptic homeostatic potentiation, PHP) is the increase in presynaptic neurotransmitter release that compensates for a decrease in postsynaptic receptivity [15]. In the NMJ of vertebrate animals, including humans, this homeostatic regulation is very rapid, occurring within seconds following the partial blockade of nicotinic acetylcholine receptors (nAChRs). Recently, Zhu et al. [16] implicated extracellular protons as carriers of this retrograde signal and ASICs as mediators. Most ASICs are activated by acidic pH and are permeable to Na⁺, although some ASICs also conduct Ca²⁺ [17-19]. Zhu et al. [16] found that the pharmacological inhibition of ASICs prevents the upregulation of QC in response to the partial blockade of nAChRs by d-tubocurarine in the NMJ of mice, demonstrating the necessity of ASICs for PHP. Accordingly, the neuronal presynaptic areas of the NMJ can compensate for repeated neuronal stimulation. However, Imomnazarov et al. [13] found that QC increased when the pH of perfusion saline decreased to 7.20, 7.15, or 7.10, but failed to trigger QC upregulation in more acidic (pH 7.00) or basic (pH 7.25) environments. If QC upregulation occurs in accordance with a decrease in pH, sensitivity to NMBA might not be increased or may even be resistant to neuromuscular blockade. In contrast, if QC upregulation fails at a given pH, it is more sensitive to NMBA. In the present study, the acidosis group became more sensitive to rocuronium, although the mean pH of this group was 7.1. This result is significantly different from that obtained by Imomnazarov. According to their results, increased sensitivity to rocuronium in the acidosis group should not be observed because the pH of this group was within the range of Imomnazarov’s, which can activate ASICs and PHP occurred [13]. A possible explanation is that the pH of the NMJ might be more acidic because of the protons that are contained in the QC and are released and accumulated by repeated indirect stimulations.

To estimate the sugammadex-induced TOFR recovery, no suitable regression model could be derived from the present study data. During typical TOFR recovery (spontaneous or using acetylcholinesterase inhibitors), T1 twitch tension reappears followed by T2, T3, and T4, and TOFR recovery pattern is linear or sigmoid. However, the TOFR recovery pattern is quite different from the classic recovery when sugammadex is used. In the present study, all responses to TOF stimulation reappeared simultaneously from the starting point, with an initial TOFR > 0.7, even though recovery T1 twitch tensions were only 5% of their initial tensions. As T1 increased, the TOFR gradually decreased to below 0.5, and then recovery continuously progressed to > 0.9. This ‘notch’ formation during recovery made it hard to set a fittable linear or nonlinear mixed model. Eleveld et al. [20] reported a temporary decrease in TOFR response and T1 twitch height during sugammadex-induced recovery from rocuronium-induced neuromuscular blockade. They described this phenomenon as “muscle relaxation rebound” and hypothesized that it may be due to redistribution of unbound molecules of neuromuscular blocking drugs from peripheral compartments back into central and effect compartments. The present study was performed in an ex-vivo environment using the phrenic nerve-hemidiaphragm harvested from a Sprague-Dawley rat thoracic cage, and pharmacokinetic components were eliminated from our data. The recovery of TOFR was similar to that reported by Eleveld et al. [20]. Eleveld et al. [20] speculated that this phenomenon was related to the amount of sugammadex administered. The sugammadex dose was set as the equimolar dose of rocuronium because the molecules were bonded to each other in a 1:1 manner. According to Eleveld et al. [20], this recovery notch may appear when insufficient sugammadex is used, even though complete recovery of T1 is achieved. It was suspect that sugammadex arrives at the effector site only through diffusion from the surface of the organ bath. As such, although an equimolar dose of sugammadex was administered, only small amounts of sugammadex could reached at NMJ in the early recovery period in this ex-vivo study.

The present study has several limitations. First, this was an ex-vivo study performed using tissue specimen immersed in a Krebs buffer solution. Although this buffer solution contains a buffer system similar to the in-vivo environment, there are more buffering systems in in-vivo environment such as protein buffer system. Second, the pH of the buffer solution was measured only at the end of the study. CO₂ in the mixed gas is bubbled through and dissolved in the buffer solution in the ex-vivo environment. The gas may be more dissolved if the gas is more exposed in the solution. The gas-supplying catheters were placed freely in organ baths, and the speed or amount of dissolved gas would possibly be different if the depth of the catheter tip were changed. We attempted to match the time and amount of gas exposure in the buffer solution by placing the catheter tip at the base of the organ bath, just above the orifice of the drain, and setting the gas supply pressure to 1psi. Furthermore, the pH of the NMJ could not be measured, where the co-released protons might create a more acidic environment. In the acidosis group, the mean pH was about 7.15 (Table 1), which was measured at the end of each session by sampling the buffer solution. This range might be at the margin of the pH at which QC upregulation could occur by ASIC, as described in an article by Imomnazarov et al. [13]. This may be one of the clues that the data of each group were so intermingled, as shown in Fig. 2A. If the pH at the NMJ was out of the range in which QC up-regulation occurred, QC at the presynaptic neuronal side would be relatively insufficient for the repeated indirect stimulation applied in this study. Although there are other homeostatic mechanisms for compensating for repeated stimulation in the NMJ, such as muscarinic receptors, adenosine receptors, and other minor receptor systems [21,22], we only tried to change the pH to make ASIC the main homeostatic system in the NMJ in this study. The last limitation in this study is that we performed this experiment in ex-vivo environment, which lacks pharmacokinetic properties. Therefore, these results cannot be directly applied in clinical practice.

In conclusion, respiratory acid-base alterations may influence rocuronium-induced neuromuscular blockade and its reversal by sugammadex. However, despite some intergroup differences, we did not obtain statistically significant results across the experimental groups. Therefore, within the pH range examined in this ex-vivo study, the effect of respiratory acid-base imbalance on neuromuscular blockade and recovery appears to be limited or minimal. Further investigations are warranted to elucidate the precise effects of acid-base disturbances on both neuromuscular blockade and reversal.