Effects of Tidal Volume on Physiology and Clinical Outcomes of Surgery Patients Undergoing One-lung Ventilation: A Meta-analysis of Randomized Controlled Trials

Objective: It is unclear how tidal volume (Vt) impacts patients during one-lung ventilation (OLV). A meta-analysis was performed to assess the effect of Vt on physiology and clinical outcomes in OLV patients. Methods: PubMed, Cochrane library, and Web of Science were retrieved until February 2020. Randomized controlled trials comparing the application of low and high Vt ventilation in adults with OLV were included. Demographic variables, Vt, physiology, and clinical outcomes were retrieved. Summary odds ratios (ORs) with 95% condence intervals (CIs) and mean difference with standard deviation were calculated using a random-effects model. Results: 12 studies involving a total of 876 participants met inclusion criteria. Signicant difference in blood interleukin-6 (IL-6) was observed between low Vt ventilation and high Vt ventilation (MD -35.51 pg/ml, 95% CI [-66.47, -4.54 pg/ml], p = 0.02). Low Vt ventilation decreased driving pressure (ΔP) (MD -6.02 cmH 2 O, 95% CI [-8.32, -3.72 cmH 2 O], p < 0.0001), Peak pressure (Ppeak) (MD -2.88 cmH 2 O, 95% CI [-4.60, -1.16 cmH 2 O], p = 0.001), and improved PaO 2 /FiO 2 (MD 32.27 mmHg, 95% CI [19.54, 45.01 mmHg], p <0.00001). Risk of atelectasis was the same between the two groups. Furthermore, the study suggested that low Vt ventilation was associated with decreases in the risk of acute lung injury (OR 0.05, 95% CI [0.28, 0.88], p = 0.02). Decreased hospital length of stay in the low Vt group occurred when Vt was set 4-5 ml/kg (MD -0.78 d, 95% CI [-1.45, -0.11 d], p = 0.02). Conclusions: In OLV patients, low Vt ventilation improved PaO 2 /FiO 2 , and it was also associated with decreased blood IL-6, ΔP, Ppeak, and risk of acute lung injury, when the low Vt was set 4-5 ml/kg hospital length of stay was decreased meta-analysis aims to assess the effects of low Vt on the physiology and clinical outcomes of adult surgery patients undergoing OLV. of ARDS patients. In the surgical population, either two lung[32] or one lung[33, 34] ventilation, ∆ P has been identied as a risk factor for the development of postoperative pulmonary complications. Driving pressure to elastance times Vt. it as a surrogate for dynamic alveolar injury. The results of this study showed that low Vt signicantly reduced the driving pressure and peak pressure. it can be speculated that low Vt may be associated with the postoperative pulmonary complications of patients. Further studies are needed to clarify the relationship between Vt, driving pressure, and patient outcomes. other strong attenuated by the of the lower Vt.


Introduction
Lung complications are common after general anesthesia surgery and are associated with signi cantly increased mortality and morbidity [1,2]. Low tidal volume (Vt) mechanical ventilation in anesthetized patients undergoing abdominal surgery can limit postoperative pulmonary complications [3]. Initial evidence of protective ventilation with low Vt ventilation comes from acute respiratory distress syndrome (ARDS) [4]. Recent research has established the applying of low Vt ventilation to surgical populations without ARDS, with consistent results [5]. Patients of thoracic surgery undergoing one-lung ventilation (OLV) are challenged physiologic and clinical circumstances that complicate the application of lung-protective ventilation with low Vt.
OLV causes signi cant physiologic mechanisms changes, thus affects the clinical prognosis of patients. These changed physiologic mechanisms are as follows [6][7][8]: the obligate collapse of the nondependent lung and overdistention of the dependent lung represents in ammatory cascades and is related to high airway pressures; when ventilation is switched from two-lung to OLV, shunt fraction increased, and hypoxemia may occur; because of lower chest wall compliance due to lateral decubitus position compared with two lung ventilation, pulmonary atelectasis occurs more frequently. Therefore, patients of thoracic surgery with OLV are vulnerable to ventilator-induced lung injury and increased length of hospital stay.
The sample sizes of most studies of OLV are small, and these studies primarily focus on reporting physiologic outcomes. Furthermore, due to a paucity of evidence, anesthesiologists implemented different ventilation strategies during OLV [9]. Recently, a study has demonstrated that lung-protective ventilation with recruitment maneuvers and PEEP improved physiologic outcomes in OLV for thoracic surgery [10]. However, it has not been evaluated whether low Vt improved outcomes when it was used during OLV in patients undergoing thoracic surgery. In 2017, El Tahan and his colleagues's study [11] denoted that low Vt of OLV was with no impact on hospital length of stay, while a subsequent study [12] showed that the duration of hospital stay was shorter in low Vt group. The metaanalysis aims to assess the effects of low Vt on the physiology and clinical outcomes of adult surgery patients undergoing OLV.

Methods
The methods that we used to write the meta-analysis comply with the Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines [13].

Search Strategies And Study Screening
The authors searched the PubMed, EMBASE, and the Cochrane Library databases from inception to February 2020 for studies exploring the intraoperative use of OLV with low Vt in patients undergoing thoracic surgery. There were no limits applied for articles published in a different language. The following terms: "protective ventilation or low tidal volume" and "one-lung ventilation or One-Lung Ventilation" were searched. In addition, the authors manually searched the reference lists of relevant studies.
Two reviewers (F.P.X. and Z.H.L.) independently assessed all titles and abstracts and excluded studies that were irrelevant. According to the inclusion and exclusion criteria, the full texts of the remaining articles were then independently reviewed. The priority inclusion criteria were as follows: 1) patients receiving OLV with clearly reported size of Vt; 2) comparison of different Vt during intraoperative ventilation undergoing thoracic surgery; 3) randomized controlled trials (RCT). Studies involving children, lung transplantation, cardiopulmonary bypass, those comparing airway devices, or without de nite time of measurement were excluded. For reaching a consensus inclusion, the discrepancies were resolved by consensus and discussion with a third reviewer. According to the included studies, a low Vt was de ned as 3-6 mL/kg of ideal body weight, and a high Vt was de ned as 8-10 mL/kg of ideal body weight. Included studies were then analyzed to identify which outcome measures were reported in the studies.

Data Extraction And Quality Assessment
Two authors performed data extraction according to standardized author-developed data extraction form in Excel. The following data were extracted from included trials: year of the publication, rst author, type of patients, side of the operation, Vt category, PEEP setting, FiO 2 during OLV, recruitment maneuver, and physiology or clinical outcomes. For the data presented as a median range, we converted it to mean standard deviation [14].
We evaluated included RCTs using the Cochrane Risk of Bias tool, which include the following items: random sequence generation (selection bias), allocation concealment (selection bias), blinding of the participants and personnel (performance bias), blinding of the outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other biases [15]. One author performed data extraction and bias assessment of the included studies, and the extracted data and the risk of bias ware con rmed by the other author. When there was a discrepancy between the two authors, they reached the consensus by discussion.

Statistical analysis
All statistical pooling of the meta-analysis was performed using RevMan (ver5.1). Physiology outcomes for meta-analysis of Vt were blood interleukin-6 (IL-6), driving pressure (ΔP), peak pressure (Ppeak), PaO 2 /FiO 2 , and atelectasis; clinical outcomes for meta-analysis of Vt were hospital length of stay and incidence of acute lung injury. For dichotomous outcomes, the odds ratios (ORs) with corresponding 95% con dence intervals (CIs) we calculated. The mean difference with the standard deviation was calculated for continuous outcomes. For considering clinical heterogeneity, random effects modeling was chosen as a priority [16]. The author quanti ed heterogeneity between studies using the I 2 statistic [17]. By excluding one study one at a time, the author tested the sensitivity and repeated to analyze the robustness of the aggregated results. A two-sided p-value of less than 0.05 was considered statistically signi cant.

Search results
Our research identi ed 2842 studies through database searching, and 25 additional records were identi ed through other sources. 1023 records remained after duplicates were removed by screening the titles and abstracts. Among them, 934 publications were not relevant and were discarded. The full text of the remaining 80 publications was examined for an assessment; According to the exclusion criteria, 69 publications were ineligible and were excluded. Twelve studies were included in the nal meta-analysis ( Fig. 1).

Study Characteristics
We included 12 studies that compared low Vt with high Vt in thoracic surgery patients undergoing OLV. The characteristics of each study are reported in Table 1. The included studies were published between 2005 and 2019. The analysis involved 876 participants, individual study with different samples size ranging from 26 to 343. Low Vt varied from 3 mL/kg to 6 mL/kg of ideal body weigh, while high Vt was from 8 mL/kg to 10 mL/kg of ideal body weigh. Ten studies applied PEEP varying from 3 cmH 2 O to 8 cmH 2 O in the low Vt groups. Ten studies did not apply or did not mention applying PEEP in the high Vt groups. The fraction of inspiration O 2 (FiO 2 ) applied in eight of the studies during the surgery was adjusted according to oxygen saturation or protocol. Three studies administered recruitment maneuvers (RM) in the low Vt group, two studies in the high Vt group. 12 [12,[18][19][20][21][22][23][24][25][26][27][28] of these studies reported data on physiology outcomes (IL-6, ΔP, Ppeak, PaO 2 /FiO 2 , atelectasis), and 7 studies [12, 18, 22-24, 26, 27] included data on clinical outcomes (hospital length of stay, the incidence of acute lung injury). The details of the risk of bias assessment are outlined in Fig. 2.  Figure S2) with low Vt ventilation.
When IL-6 was analyzed in terms of low Vt for less than 6 mL/kg or 6 mL/kg, among those who received a low Vt of less than 6 mL/kg, there was a signi cant decrease in the low Vt group than in the high Vt group (MD -74.62 pg/ml, 95% CI [-110.73, -38.51 pg/ml], p < 0.0001; Fig. 3), with possible moderate heterogeneity (I 2 = 32%). However, among those who received a low Vt of 6 mL/kg, there was no signi cant difference in IL-6 between low Vt and high Vt groups (MD -4.08 pg/ml, 95% CI [-11.08, 2.93 pg/ml], p = 0.25; Fig. 3). There was also a possibility of substantial heterogeneity (I 2 = 84.7%) between the subgroups (Fig. 3).

Clinical Outcomes
Thirteen studies reported acute lung injury and hospital length of stay as clinical outcomes. Overall, there was a signi cant decrease in the risk of acute lung injury with low Vt ventilation (OR 0.05, 95% CI [0.28, 0.88], p = 0.02; Fig. 6). There was a possible low heterogeneity (I 2 = 0%) between the low Vt group and the high Vt group (Fig. 6).  Fig. 7). There was a possible low heterogeneity between the two subgroups (I 2 = 45.4%, Fig. 7).

Discussion
In order to better elaborate the physiology and clinical impact of the Vt in thoracic surgery patients of OLV, we have performed the meta-analysis. The results demonstrate that low Vt ventilation during OLV signi cantly improved PaO 2 /FiO 2 , and it was associated with decreased blood IL-6, ΔP, Ppeak, and risk of acute lung injury. Furthermore, hospital length of stay was decreased in low Vt ventilation when the Vt was set 4-5 ml/kg. Meanwhile, low Vt ventilation had no impact on the risk of atelectasis.
Low Vt ventilation strategy aimed to limit lung overdistension, leading to a reduction in the incidence of ALI along with a shorter hospital stay. Conventional Vt of two lung ventilation may lead to overdistension of the aerated one lung and increase the shear forces generated due to repetitive open and collapse of atelectatic areas. When compared with conventional Vt, a low Vt ventilation strategy is proved to be bene cial for both injured lungs [4] and anesthetized patients [3]. However, there is a lack of large sample randomized controlled studies to evaluate the effect of low Vt on ALI. Our results denoted that the incidence of ALI was low when patients were with one-lung ventilation during thoracic surgery, which was consisted of the ndings of Hu's study [29]. In Hu's study, they did not explore hospital stay length, while our study denoted that hospital stay length was decreased when the Vt was set 4-5 ml/kg. However, hospital stay length did not differ between the ventilatory strategies in Lee' s [30] study, for actual body weight was used to set Vt in their research, which may affect the results. Ahn's [18] research showed that low Vt had no advantage on hospital length of stay. The negative results in this study were likely to be explained by the fact that although it was ventilated with low Vt, the platform pressure was not high, less than 20 cmH 2 O. This result had implied that pressure during ventilation were needed to be taken into account.
In a retrospective study, Amato and his colleges [31] had identi ed high driving pressure as the risk of outcomes of ARDS patients. In the surgical population, either two lung [32] or one lung [33,34] ventilation, ∆P has been identi ed as a risk factor for the development of postoperative pulmonary complications. Driving pressure equals to elastance times Vt. Thus, it may serve as a surrogate for dynamic alveolar injury. The results of this study showed that low Vt signi cantly reduced both the driving pressure and peak pressure. Therefore, it can be speculated that low Vt ventilation may be associated with the postoperative pulmonary complications of patients. Further studies are needed to clarify the relationship between Vt, driving pressure, and patient outcomes. In addition, our results denoted other strong indications that lung injury was attenuated by the application of the lower Vt.
High Vt associated with deformation of the alveolar epithelium and cyclic opening of collapsed areas during OLV is thought to local production and release of in ammatory mediators leading to ALI. In ammatory biomarkers were a direct assessment of lung damage. A decreased in ammatory response was observed in healthy lungs after low Vt ventilation compared with conventional Vt [35]. Previous meta-analysis [11] had envaluated impact of low Vt on OLV patients without in ammatory biomarkers. In our study, the patients who received low Vt ventilation showed decreased serum IL-6, which was indicated as a useful marker of induced injury [36]. Our result was consistent with the previous nding [23] described in patients undergoing esophagectomy, when combined Vt of 5 ml/kg with positive end-expiratory pressure (PEEP) 5 cm H 2 O during one-lung ventilation, resulted in lower levels of IL-6 being released into the serum after the surgery. However, Kim and his colleges [20] did not observe a difference in plasma IL-6. This may due to the calculation of the sample size for this study was not based on postoperative outcomes, making the sample size small. Therefore, the clinical impact on pulmonary in ammation remains further to be investigated.
Hypoxemia during OLV may be prevented by applying a ventilation strategy that avoids alveolar collapse while minimally impairing perfusion of the dependent lung. Strategies to improve ventilation/perfusion ratio and thereby to maintain arterial oxygen tension during OLV [8] in thoracic surgery include using low Vt and positive end-expiratory pressure (PEEP) to the ventilated lung and titrating inspired FiO 2 to maintain a SpO 2 . In our study, PaO 2 /FiO 2 was improved under low Vt ventilation. Consistent with Lee' s [30] study that low Vt ventilation was associated with better oxygenation compared with conventional ventilation requiring OLV. While in Liu's study [37], there was no difference in PaO 2 /FiO 2 between groups, for they compared different modes rather than different Vt of ventilation. Therefore, large Vt, which was potentially injurious to the lung, do not translate into better oxygenation as compared with low Vt. Low Vt ventilation that keeps the lung open without impeding perfusion improved oxygenation during OLV.
We also found that postoperative atelectasis was not evident in the low Vt group compared with the conventional Vt group. During intraoperative ventilation, atelectasis may be caused by [38,39] ventilator-associated lung injuries, and the reduction of functional residual capacity consequent to OLV and muscle paralysis. The previous study [30] had shown that lung atelectasis could be reduced by low Vt ventilation when assessed using lung ultrasound. Apart from Vt, the application of PEEP [40] in the low Vt group contributed to the prevention of atelectasis. Considering that PEEP was applied only in the low Vt group among six of the seven included studies, the importance of proper PEEP in the low Vt group may contribute to the prevention of atelectasis in our study.
This meta-analysis also has some limitations. First, the magnitude of hypoxemia generally peaks about 20 minutes after beginning OLV. While, in our study, data were collected at 15 minutes to 2 hours among different studies, which would overestimate the effect of low Vt on oxygenation. Second, there was heterogeneity in the use of PEEP and RM between the two groups among the included studies, for the use of PEEP and RM may associate with improvements in oxygenation and development of ALI. Therefore, analyses were needed to be interpreted with caution due to the heterogeneity.

Conclusions
This study assessed the physiology and clinical impact of low Vt ventilation during OLV. In OLV patients, low Vt was associated with improvement in PaO 2 /FiO 2 , and with decreased blood IL-6, ΔP, Ppeak, risk of acute lung injury. Furthermore, low Vt would decrease hospital length of stay of the patients when the Vt was set 4-5 ml/kg, which implied that lower Vt should be applied in patients of OLV, and further research may require to be con rmed.

Abbreviations
Vt tidal volume; OLV = one-lung ventilation; ORs = odds ratios; CIs = con dence intervals; IL-6 = interleukin-6; Ppeak = Peak pressure; ARDS = acute respiratory distress syndrome; PEEP = positive endexpiratory pressure; ΔP = driving pressure Declarations Figure 2 Risk of bias summary of the included studies. The reviews' judgements about seven risk of bias item for each study. Red indicates high risk; yellow indicates unclear risk; green indicates low risk.

Figure 3
The association of low tidal volume ventilation and IL-6 in surgery patients undergoing one-lung ventilation. The association of low tidal volume ventilation and driving pressure in surgery patients undergoing one-lung ventilation.

Figure 5
The association of low tidal volume ventilation and PaO2/FiO2 in surgery patients undergoing one-lung ventilation.

Figure 6
The association of low tidal volume ventilation and the risk of acute lung injury in surgery patients undergoing one-lung ventilation.