The Effects of Protective Ventilation on the Production of Endogenous Melatonin and Prognosis in Patients Undergoing Esophageal Cancer Surgery: A Prospective Randomized Double-Blind Controlled Study


 Background: Exogenous melatonin exerts a similar effect to protective ventilation on attenuating ventilator-induced lung injury (VILI) by inhibiting NLRP3 inflammasome activation in mouse model. However, the effect of protective ventilation on the production of endogenous melatonin and prognosis in patients undergoing esophageal cancer surgery remains unknown. In this study, we aimed to reveal the effects of protective ventilation on the production of endogenous melatonin, interleukin (IL)-1β, IL-18 and major complications in patients undergoing esophageal cancer surgery. Methods: Eight-eight patients were randomized to receive “conventional” ventilation (Vt=10 mL/kg) or lung protective ventilation [Vt=5 mL/kg along with 5 cm of H2O positive end-expiratory pressure (PEEP)]. IL-1β, IL-18 and melatonin levels in bronchoalveolar lavage fluid (BALF) and serum were measured. Respiratory variables and outcomes were evaluated.Results: Lung protective ventilation decreased the peak airway pressure (Ppeak), plateau airway pressure (Pplat) and driving pressure (ΔP) compared with the “conventional” ventilation group. Lung protective ventilation inhibited polymorphonuclear (PMN) cells invasion into the BALF (P=0.000). Likewise, lung protective ventilation suppressed alveolar and serum IL-1β and IL-18 secretion after mechanical ventilation. Furthermore, lung protective ventilation resulted in a decrease in the inhibition of endogenous MT production compared to “conventional” ventilation (P=0.000). In addition, lung protective ventilation reduced the incidence of postoperative pulmonary complications (P=0.04) and the rate of major postoperative complications (P=0.023).Conclusions: Taken together, lung protective ventilation for esophageal cancer surgery suppressed the secretion of IL-1β, IL-18 and restored the endogenous melatonin level. Meanwhile, lung protective ventilation improved postoperative outcomes after esophageal cancer surgery.Trial registration: The Chinese Clinical Trial Registry, ChiCTR1900026190. Registered 25 September 2019, http://www.chictr.org.cn/edit.aspx?pid=34677&htm=4


Background
One-lung ventilation (OLV) is required for esophageal cancer and can contribute to the surgical eld [1].
However, inappropriate ventilation modes may cause or augment acute lung injury, which is known as ventilator-induced lung injury (VILI) [2]. Lung protective ventilation [low tidal volume + positive endexpiratory pressure (PEEP)] was shown to achieve good clinical effects and protect against VILI [3,4].
Furthermore, clinical studies have demonstrated that lung protective ventilation induced an immune response with lower concentrations of in ammatory mediators than that of "conventional" ventilation [5].
Therefore, further studies of the effect of lung protective ventilation on the pulmonary immune response are essential to prevent VILI.
Increasing studies have shown that OLV may lead to proin ammatory cytokine release and in ammatory signaling pathway activation [5][6][7][8]. Overdistension in ventilated lungs followed by compression of alveolar vessels initiates a robust release of proin ammatory cytokines, such as interleukin (IL)-6, IL-8 and tumor necrosis factor (TNF)-a, in bronchoalveolar lavage uid (BALF) [5,9]. These proin ammatory cytokines are important chemotactic factors for polymorphonuclear (PMN) cells [10]. Excessive PMN cell aggregation will amplify the in ammatory cascade. Furthermore, a recent study showed that in mouse alveolar macrophages, Nucleotide-binding domain and leucine-rich repeat protein 3 (NLRP3) in ammasome activation contributes to the development of VILI [11].
Melatonin (N-acetyl-5-methoxytryptamine, MT), which is mainly secreted in the pineal gland, has welldocumented anti-in ammatory and immunomodulatory functions [12,13]. Early preliminary studies have shown that exogenous MT ameliorates VILI by increasing the anti-in ammatory response [14]. Recently, Zhang et al. demonstrated that exogenous MT inhibited NLRP3 in ammasome activation in mice with acute lung injury [15]. However, researchers have not determined whether lung protective ventilation affects NLRP3 in ammasome-related in ammatory cytokine and endogenous melatonin production in patients.
Our study aimed to investigate the effects of lung protective ventilation on NLRP3 in ammasome-related in ammatory cytokine and endogenous MT secretion in patients undergoing video-assisted thoracoscopic esophagectomy (VATS). In addition, the effect of lung protective ventilation on postoperative complications was also investigated.

Study Population
Patients with esophageal cancer who were treated at our hospital were considered for enrollment. The inclusion criteria were as follows: American Society of Anesthesiologists (ASA) physical status I -, requirement for OLV during operation, and aged 45-77 years. Exclusion criteria were preexisting hypoxemia, diagnosed major obstructive or restrictive pulmonary disease [preoperative forced expiratory volume in 1 second (FEV1) and forced vital capacity (FVC) <70% of the predicted value], pulmonary infection before surgery, body mass index (BMI) of less than 20 or more than 35, and use of immune modulators.

Randomization and Blinding
The randomized numbers were generated by a research coordinator using block sizes on a 1:1 ratio. This ensured that each group had an equal number of subjects. Then, the research coordinator sealed the numbers in opaque envelopes. Before mechanical ventilation, the anesthesia assistant opened the envelopes, set the breathing parameters and covered the breathing parameters using opaque paper. The anesthesia assistant did not participate in the next study. One anesthesiologist collected the specimens, and another anesthesiologist recorded the breathing parameters. Both physicians were blinded to the allocation.

Study Protocol
Standard monitoring devices were applied after admission to the operating room. Before induction of anesthesia, an artery catheter was inserted into the left radial artery. Anesthesia induction was performed with 2.0-2.5 mg/kg propofol, 0.02-0.06 mg/kg midazolam, 0.4-0.6 µg/kg sufentanil and 0.6-0.9 mg/kg rocuronium. 3 min after assisted breathing, a double-lumen endotracheal tube (Broncho-Cath ® 35 F or 37 F; Covidien, Ireland) was inserted into the left main bronchus. Anesthesia was maintained with 50-100 µg/kg/min of propofol, 0.1-1 µg/kg/min of remifentanil and 5.0-10.0 µg/kg/min of rocuronium to maintain the proper depth of anesthesia (BIS 40-60). A forced-air warming system (3M Company, Shanghai, China) was used to keep the patients warm.
After intubation, the patients were randomly divided into 2 groups. In the control group (group A), patients received volume controlled mechanical ventilation with a tidal volume of 10 mL/kg of ideal body weight (IBW). In the lung protective ventilation group (group B), lung protective ventilation with a low tidal volume (Vt=5 mL/kg IBW) and 5 cm H 2 O PEEP was chosen. After 15 minutes, all patients were turned to the left lateral position and OLV was initiated. During OLV, the ventilation mode was not changed except the plateau airway pressure (Pplat) exceeded 30 cmH 2 O. If the Pplat exceeded 30 cmH 2 O, the tidal volume was decreased and this patient discontinued the experiment. With both two-lung ventilation (TLV) and OLV, mechanical ventilation was performed with an inspiratory to expiratory ratio of 1:2, and an appropriate respiratory rate to maintain an end-tidal CO 2 (ETCO 2 ) below 45 mmHg. All surgeries were always performed at 8:30 in the morning.

Observational Indexes
Peak airway pressure (Ppeak), Pplat, respiratory rate and blood gas analyses were evaluated at two stages: during TLV before surgery and 30 minutes after OLV. Furthermore, the driving pressure (ΔP) was recorded. ΔP was de ned and calculated as follows: ΔP=Pplat-PEEP [16].
The major postoperative complications were pulmonary complications and nonpulmonary complications. The pulmonary complications included pulmonary infection, acute lung injury or acute respiratory distress syndrome and reintubation or invasive mechanical ventilation. Nonpulmonary complications included anastomotic stula, incision infection, ICU stay and death before hospital discharge.
Bronchoalveolar lavage was performed after induction of general anesthesia (baseline) and at the end of the surgical procedures. BALF was aspirated from the lung after instillation of 20 mL of sterile isotonic saline. Then, the recovered BALF was centrifuged at 700 g for 10 minutes at 4°C, and the supernatant was stored at -80°C. The cell pellets were resuspended in ice-cold sterile isotonic saline for staining and counting.
Blood samples were obtained during TLV after induction of general anesthesia (baseline) and at the end of the surgical procedures. Five milliliters of arterial blood samples was centrifuged at 800 g for 5 minutes. The upper serum phase was separated and stored at -80°C.
MT, IL-18 and IL-1β concentrations in the serum and BALF were determined using commercial ELISA kits (Cusabio, Wuhan, China). We performed the assays according to the manufacturer's instructions. The limitations for MT, IL-18 and IL-1β were 0.1 pg/mL, 7.8 pg/mL and 2.2 pg/mL, respectively.

Statistical Analysis
According to previous studies, the cell numbers in the BALF increased by more than 30% after OLV [5], which required 12 patients per group with α=0.05 and β=0.02; thus, we aimed to enroll 88 patients to allow for dropouts. The sample size was calculated using PASS 11.0 software.
Data are presented as the mean ± SD or number of patients (proportion, %). The independent-samples ttest or paired-samples t-test were used to analyze normally distributed data. Non-normally distributed data were analyzed by chi-square tests or Fisher's exact test. All statistical analyses were performed with SPSS 19, and a P value of < 0.05 was considered signi cant.

Results
Baseline Parameters of Patients 88 patients were included and assessed. Four patients did not meet the criteria, and 84 were included in this study. However, two patients withdrew for technical reasons, and the other patient were excluded for higher Pplat. Finally, 81 patients completed the study (Fig. 1). The patient characteristics and preoperative details showed no signi cant differences between the two groups (Table 1).

Changes In Respiratory Parameters
The respiratory and gas exchange variables are presented in Table 2. The Ppeak, Pplat and ΔP were signi cantly decreased in the protective ventilation group. While the respiratory rate increased substantially compared with that in the control group. Additionally, the oxygenation index in the protective ventilation group was higher than that in the control group at 30 minutes after OLV (P = 0.006). Data were presented as the mean and SD. Group A: the patients chose volume controlled mechanical ventilation with a routine tidal volume (Vt = 10 mL/kg) as control; Group B: the patients chose lung protective ventilation with a low tidal volume (Vt = 5 mL/kg) and 5 cm H 2 O PEEP. OLV: one-lung ventilation; TLV: two-lung ventilation. a Compared Group A with Group B, P < 0.05

Changes In The Number Of Cells In The Balf
The cells in the BALF were counted after Wright-Giemsa staining. The number of total cells (in groups A and B, P = 0.000) and PMN cells (in groups A and B, P = 0.000) in the BALF were substantially increased after mechanical ventilation (Fig. 2). However, in the group treated with the lung-protective strategy, the total cells (P = 0.000) and PMN cells (P = 0.000) in the BALF were signi cantly reduced compared to the control group (Fig. 2).

Changes in IL-1β and IL-18 Levels in the BALF and Serum
Commercial ELISA kits were used to detect the levels of both IL-1β and IL-18 in the BALF and serum. The IL-1β and IL-18 levels in the BALF and serum showed an increasing trend after mechanical ventilation ( Fig. 3A, B, D and E). However, lung protective ventilation resulted in a signi cant decrease in the BALF and serum IL-1β and IL-18 concentrations compared to the control group (Fig. 3A, B, D and E).

Changes In Mt Levels In The Balf And Serum
Endogenous MT levels in both the BALF and serum were also detected. In contrast to the IL-18 and IL-1β levels, the BALF and serum MT levels were signi cantly decreased in both groups after mechanical ventilation (Fig. 3C, F). Additionally, lower BALF and serum MT concentrations were observed in the control group than in the lung protective ventilation group (Fig. 3C and F).

The Incidence Of Complications
Pulmonary complications occurred in 2/41(4.88%) patients in the protective ventilation group and 8/40 (20%) patients in the control group (P = 0.04). 2 (4.88%) patient in the protective ventilation group developed a nonpulmonary complication compared with 4 (10%) patients in the control group (P = 0.382). The rate of major postoperative complications was 9.76% and 30% in the protective ventilation group and control group, respectively (P = 0.023). The incidence of major postoperative complications was lower in the lung protection group than in the control group (Table 3).

Discussion
As shown in the present study, lung protective ventilation improved respiratory variables, including Ppeak, Pplat and ΔP. Lung protective ventilation not only inhibited PMN cell invasion but also suppressed IL-1β and IL-18 secretion. Lung protective ventilation resulted in a decrease in the inhibition of endogenous MT production compared to "conventional" ventilation. In addition, lung protective ventilation decreased the incidence of pulmonary complications and major postoperative complications.
OLV is an established procedure performed during VATS. However, clinical studies have shown that the extended use of OLV is an independent risk factor for postoperative pulmonary dysfunction [17]. Excessive stretching or repeated opening of lung tissues is an important cause of VILI during OLV [18]. A lung-protective strategy using low Vt along with PEEP during OLV was con rmed to improve postoperative pulmonary dysfunction [6]. In our study, the lung-protective strategy notably decreased Ppeak and Pplat, indicating that the shear force was reduced by the lung-protective strategy. Meanwhile, we also observed a substantial decrease in ΔP with the lung-protective strategy, which suggested that the lung-protective strategy was associated with a reduced incidence of postoperative pulmonary complications [16]. Indeed, postoperative pulmonary complications occurred less frequently in the lung protective ventilation group in our study.
Increased mechanical strain further activating the in ammatory response is a key event during the development of VILI [5]. The results from previous and recent studies have shown that IL-1β is a special proin ammatory cytokine that promotes VILI in animal models and patients [19][20][21][22]. Regulation and inhibition of IL-1β can nally achieve organ protection because blockade of the IL-1 receptor has been demonstrated to inhibit neutrophil sequestration and edema formation in VILI [23]. In our study, mechanical ventilation clearly increased the alveolar and serum concentration of IL-1β and the alveolar PMN cell counts in the BALF. However, lung protective ventilation blocked the elevated IL-1β level and PMN cell in ltration. Most interestingly, we observed a dramatic increase in both the alveolar and serum concentrations of IL-18 after OLV, while lung protective ventilation resulted in a profound reduction in IL-18 levels. IL-1β and IL-18 were con rmed to be products of NLRP3 in ammasome activation [24].
Furthermore, current studies have demonstrated that NLRP3 in ammasome activation plays a key role in the pathogenesis of VILI in a mouse model [25,26]. Therefore, lung protective ventilation may inhibit in ammatory responses by inhibiting the activation of the NLRP3 in ammasome. For the rst time, we showed that mechanical ventilation may activate the NLRP3 in ammasome, and lung protective ventilation seems to inhibit the NLRP3 in ammasome activation in patients.
In recent years, the anti-in ammatory effects of both exogenous and endogenous MT have been observed in many conditions [27,28]. Paula et al. demonstrated that the exogenous addition of MT protected against VILI through decreasing the levels of in ammatory cytokines in a mouse model [14]. Further research con rmed that exogenous replenishment of MT alleviated lipopolysaccharide-induced acute lung injury by inhibiting NLRP3 in ammasome activation [15]. However, researchers have not determined whether VILI affects the production of endogenous MT. Therefore, we hypothesized that endogenous MT may play a pivotal role in the pathogenesis of VILI. As expected, mechanical ventilation substantially reduced the levels of endogenous MT in patient serum and BALF. Surprisingly, pulmonary protective ventilation signi cantly inhibited the reduction of endogenous MT. Accordingly, our results suggested that endogenous MT may be involved in the pathogenesis of VILI, and pulmonary protective ventilation may attenuate VILI by restoring the level of endogenous MT in patients.
As described above, lung protective ventilation not only improved respiratory parameters but also suppressed NLRP3 in ammasome-related in ammatory cytokine secretion and restored the level of endogenous MT: which are likely to be required to improve outcomes during esophageal surgery. Indeed, lung protective ventilation not only reduced the incidence of pulmonary complications but also decreased the rate of major postoperative complications in our study, consistent with the results reported by Marret [29].
This study has some limitations. First, the sizes of the samples were small, which may lead to bias.
Second, based on our data, we were unable to conclusively determine the relationship between in ammasome-related in ammatory cytokines and endogenous MT. Therefore, the crosstalk between endogenous MT and the NLRP3 in ammasome in VILI requires further animal experiments.

Conclusions
In conclusion, pulmonary protective ventilation improved outcomes by decreasing the rate of pulmonary complications and major postoperative complications. These effects may be attributed to the ability of pulmonary protective ventilation to suppress NLRP3 in ammasome-related in ammatory cytokine secretion and restore the level of endogenous MT in patients undergoing VATS.  Figure 1 Consort ow chart that outline patients assignment and treatment protocols. Group A: Volume controlled mechanical ventilation with a tidal volume of 10 mL/kg was used; Group B: lung protective ventilation with a low tidal volume (Vt=5 mL/kg IBW) and 5 cm H2O PEEP was chosen.