Abstract
Introduction With the widely used technique of One Lung Ventilation (OLV) in patients throughout thoracic surgery, it’s unclear whether inhaled or intravenous anesthetics were associated with postoperative complications. The purpose of the current study is to compare the effects of intravenous and inhaled anesthetics on the postoperative complications within the patients suffering OLV.
Methods We searched the related randomized controlled trials in PubMed\EMBASE\Medline and the Cochrane library up to 09\2021.Inclusive criteria were as follows: We included all the randomized controlled trials which compared the effects of intravenous and inhaled anesthetics on the postoperative complications[listed as: (a) major complications; (b)postoperative pulmonary complications (PPCs); (c) postoperative cognitive function (MMSE score); (d) length of hospital stay; (e) 30-days mortality] for the patients undergoing one lung ventilation.
Results Thirteen randomized controlled trials with 2522 patients were included for analysis. Overall, there were no significant differences in the postoperative major complications between inhaled and intravenous anesthetics groups (OR 0.78, 95%CI 0.54 to 1.13, p=0.19; I2=0%). However, more PPCs were detected in intravenous groups when compared to inhaled groups (OR 0.62, 95%CI 0.44 to 0.87, p=0.005; I2=37%). Both the postoperative MMSE scores (SMD -1.94, 95%CI -4.87 to 0.99, p=0.19; I2=100%) and the length of hospital stay (SMD 0.05, 95%CI -0.29 to 0.39, p=0.76; I2=73%) were comparable between two groups. Besides, the 30-day mortality didn’t differ significantly across groups either (OR 0.79, 95%CI 0.03 to 18, p=0.88; I2=63%).
Conclusions In patients undergoing OLV, generous anesthesia with inhaled anesthetics could reduce PPCs compared with intravenous anesthetics, but no evident advantages were provided over other major complications, cognitive function, hospital stay or mortality.
Introduction
According to the literature, approximately 3% surgical patients develop major complications, resulting in 0.4% deaths peri-operatively(1). And lung cancer is the leading reason of cancer-related death in the United States(2). Since one-lung ventilation (OLV) could facilitate the performance of surgery and prevent contamination of the other lung, it has become a necessary technique during thoracic surgery(3). However, OLV could cause ischemia and hypoxia in the nonventilated lung, barotrauma and excessive fluids in the ventilated lung tissue, and alveolar or even systematic inflammatory responses, thus increasing the risk of postoperative complications(4). The incidence of postoperative pulmonary complications (PPCs) is much higher in patients operated with OLV than in those without OLV(5).
Christopher.U et al. found that in cardiac surgery, inhaled anesthesia was associated with major benefits in outcomes and reduced mortality(6). However, Bassi.A(7) and Modolo.NS(3) revealed that few evidence from randomized controlled trials(RCTs) demonstrated significant difference in particular postoperative outcomes between general anesthesia maintained by inhaled and intravenous anesthetics in case of OLV in 2008 and 2013. Later, several RCTs and systematic reviews suggested that inhalation might preserve intraoperative cardiac function, reduce PPCs, attenuate local alveolar inflammatory responses in patients undergoing OLV(8-10).
Since 2013, increasing clinical RCTs have been published to explore the distinct effects of different sedative anesthetics on the major complications in patients with OLV. Therefore, we conducted this meta-analysis to compare the effects of inhaled anesthetics (Sevoflurane or Desflurane) and intravenous anesthetic (Propofol) regarding patients’ postoperative outcomes.
Methods
We followed the recommendation from the Cochrane Handbook for the Systematic Review(11). And the meta-analysis has been registered on https://www.crd.york.ac.uk/prospero/ with No. CRD420202222856.
Retrieval strategy
Two authors (JY, QHH) separately searched the Pubmed, Medline, Embase and Cochrane central register for the relevant RCTs from Jan 1st, 2000 to Sep 31st, 2021. We used various combinations of the key words and MeSH terms to perform the search. The search terms were listed as Table 1 and the search was limited to English.
The specific keywords and MeSH terms during the screening process
Inclusion criteria
Population: Patients (>18 years old) were scheduled for elective thoracic surgery under OLV.
Intervention: Patients receiving anesthesia maintained with inhaled anesthetics when OLV.
Comparison: Patients receiving intravenous anesthetics like propofol to maintain the anesthesia when OLV.
Outcomes:
The primary endpoint was the number of major complications assessed by Clavien-Dindo score (grade III to V) or not (complications needs more intensive treatments including overall cardiac events, myocardial infarction, acute renal failure, hepatic failure, disseminated intravasal coagulation, extrapulmonary infection, gastrointestinal failure, coma).
The secondary endpoints were the number of PPCs (hypoxemia, acute respiratory distress syndrome, pulmonary infiltrates, pneumonia, pleural effusions, atelectasis, pneumothorax, bronchospasm, cardiopulmonary edema, aspiration pneumonitis); scores of Mini-mental State Examination (MMSE), length of hospital stay and 30-day mortality after OLV.
(5) Study: Randomized controlled studies.
Any trials reported the familiar Population, Intervention, Comparison, Studies, and at least one outcome as listed before were included.
Exclusion criteria
Duplicated studies, nonhuman or pediatric studies, conference abstracts, studies published before the 2000s, and the studies which data could not be extracted.
Data extraction
Based on the criteria above, two authors (JY, QHH) sequentially enrolled the trials and extracted the data independently: publication information (first author’s name, publication year), characteristics of participants (sample size, type of surgery, anesthesia induction schemes, duration of OLV and surgery, OLV strategies) and outcomes information. Disagreements over eligibility between the two researchers were resolved by discussion. If necessary, the third researcher (RC) was involved and adjudicated. The data were extracted or calculated from figures using the program Engauge Digitizer 5.1 if necessary (M. Mitchell, Engauge Digitizer, http://digitizer.sourceforge.net). All the extracted data were collected in the standardized Excel file by the two authors, and YC double-checked the accuracy.
Risk of Bias Assessment and Strength of Evidence
Two reviewers (JY, QHH) independently employed the method recommended by Cochrane Collaboration to assess the methodological quality of included trials. For each trial, the criteria used for quality assessment were random sequence generation, allocation concealment, performance bias, detection bias, attribution bias, reporting bias, and others. Each criterion is classified as “yes”, “no” or “unclear”, and the brief assessment for each trial were classified as three levels (low risk of bias, unclear risk of bias and high risk of bias). The Grading of Recommendations, Assessment, Development and Evaluations approach (GRADEpro; gdt.gradepro.org) was approved to rating the overall quality of evidence for each outcome. In the approach, each outcome began as high-quality evidence, but may be rated down by one or more of five categories of limitations: risk of bias, inconsistency, indirectness, imprecision, and reporting bias. Finally, the approach drew an evident quality of each outcome as low, moderate or high.
Statistical Analysis
According to DerSimonian and Laird method performed by Review Manager 5.3 (RevMan, The Cochrane Collaboration, Oxford, UK), differences were expressed as risk ratio (RR) with 95% confidence intervals (CI) for dichotomous data, and the differences between continuous data were expressed as mean differences (MDs) or standardized mean differences (SMD) with 95% CI. Due to the limited number and heterogeneity between included studies, the Hartung-Knapp-Sidik-Jonkman (HKSJ) method were further conducted for random effects to the data pooled by less than 5 trials, or the value of heterogeneity was more than 50%. Since Joanna(12) had proved that HKSJ method were outperformed than DerSimonian-Laird method for the meta-analysis among limited number studies or high heterogeneity.
The heterogeneity between pooled studies was represented by I2 value, and the criteria to identify whether the combined data being high or low heterogeneity was 50%. On account of the inconsistency of the surgery process, method of anesthesia, time of OLV and the factors which increased the heterogeneity, random effects model was conducted for significant heterogeneity (I2>50%, p ≤ 0.1). And sensitivity analyses were implied to explore the possible explanation for the high heterogeneity too.
Results
Study Identification
The search yielded 1945 publications by the original screening. Based on the inclusion and exclusion criteria, 1319 potentially eligible trials were excluded based on the title or abstract. Within the full text screening we excluded 46 studies (10 articles were not RCT, 19 studies did not match the population criteria, 9 compared intravenous anesthetics to regional anesthesia or other drugs, and 9 studies did not report any outcome as listed before). Finally, we included 13 studies and meta-analyzed the data(13-25). The flow chart was presented in Fig.1.
Flow diagram of selecting process
Study Characteristics and Quality
The main characteristics of the included trials were depicted in Table 2. The 13studies(13-25) included 2522 patients, which published from 2000 to 2021. As showed in Figure.2, 8(13, 15-20) out of 14 studies presented a low risk of random sequence generation and allocation concealment by describing the randomized method in detail. Although 6(14, 17, 21-24) out of 14 failed to report the details of blinding to participants or outcome evaluators, the outcomes were little influenced by the lack of blinding. According to the GRADEpro system, the quality of every outcome were showed in Table 3. The evidence quality of PPCs was high, and the evidence quality of major complications, 30-days mortality and hospital stay were moderate. However, the evidence of MMSE score was low.
Trial Characteristics
A summary of assessment of risk bias of each RCTs
The details of GRADE evidence among each outcome
1. The primary outcome: major postoperative complications
Five studies(13, 15, 17, 18, 24) assessed the major complications post-surgery including 1083 patients, among which fixed effects model showed that there have no significantly difference between intravenous or inhaled anesthetics groups (OR 0.78, 95%CI 0.54 to 1.13, p=0.19; I2=0, Fig.3.).
Forest plot for the number of postoperative major complications between inhaled and intravenous groups
2. The second outcomes
2.1 PPCs
Seven RCTs(14, 15, 17-20, 24) compared the effect of intravenous and inhaled anesthetics on PPCs in 763 patients with OLV. The fixed effects model depicted that compared to intravenous anesthetics, inhaled anesthetics significantly reduced the number of patients who had PPCs with low heterogeneity (OR 0.62, 95%CI 0.44 to 0.87, p=0.005; I2=37%, Fig.4.).
Forest plot for the number of postoperative pulmonary complications between inhaled and intravenous groups
2.2 Postoperative MMSE scores
As Fig.5. showed, five RCTs(16, 21-23, 25) estimated the postoperative cognitive function after OLV through MMSE scores in 1324 patients. The pooled data concluded that anesthetics barely have effect on MMSE scores (SMD -1.94, 95%CI -4.87 to 0.99, p=0.19; I2=100%). In terms of the extremely high heterogeneity, sensitivity analysis and HKSJ method were further conducted to strength the outcome. Nonetheless, we failed to obtain the original heterogeneity study by removing any individual trials and HKSJ method reached the same conclusion as before (SMD -1.94, 95%CI -5.11 to 1.23, p=0.16).
Forest plot for the postoperative MMSE scores between inhaled and intravenous groups
2.3 Length of hospital stay
Data for the length of hospital stay were extracted from five trials(13, 15, 17, 19, 24) with 772 patients. Fig.6. revealed that the types of anesthetics were not associated with significant differences in the length of hospital stay at all (SMD 0.05, 95%CI -0.29 to 0.39, p=0.76; I2=73%, Fig.6.). Sensitivity analysis detected that Mahmoud 2011(19) contributed to the total heterogeneity. After removing this study, the pooled data outlined that the patients in the intravenous group spent significantly less time in hospital (SMD 0.19; 95 % CI 0.05 to 0.34, p= 0.01; I2=0) compared with patients in the inhaled group. However, HKSJ method strength the conclusion that anesthetics were not related to the length of hospital stay after excluded the Mahmoud 2011 and depicted the instability of result above (SMD 0.19, 95%CI -0.04 to 0.42, p=0.07).
Forest plot for the length of hospital stay between inhaled and intravenous groups
2.4 30-days mortality
Two out of five studies(15, 17-19, 24) which intended to assess the mortality within 30 days after surgery and found that there has no difference in 30-days mortality between two groups (SMD 0.79, 95%CI 0.03 to 18, p=0.88; I2=63%, Fig.7.).
Forest plot for the 30-day mortality between inhaled and intravenous groups
Discussion
In the present analysis, we included 13 eligible trials(13-25) with 2522 patients who underwent OLV to illustrate that compared to intravenous anesthetics, inhaled anesthetics were associated with fewer risk of PPCs with the high evidence evaluated by GRADEpro system. However, different types of anesthetics had no significantly different effect on the major complications, postoperative MMSE scores, length of hospital stay or 30-days mortality.
According to the meta-analysis accomplished by Uhlig et.al(6), general anesthesia with inhaled anesthetics was associated with the reduction of major complications and mortality in cardiac surgery, which might possibly due to the cardioprotective properties of volatile anesthetics through coronary vasodilation and reduced stress responses(26).Similarly, Uhlig(6) concluded that total inhaled anesthetics seemed to be associated with less benefits in major complications, mortality, length of hospital stay compared to intravenous anesthetics in noncardiac surgery. Even though, studies had demonstrated that the anti-inflammatory effects of volatile anesthetics affected other organs like lungs, brain, kidneys and liver in virto(27-29). In terms of patients receiving noncardiac surgery (including thoracic, vascular and abdominal surgery), major complications, length of hospital stay and mortality might be more relevant to types of surgery, characteristic of patients, different standardized operating procedures rather than types of anesthetics. Thus, the systematic organ protection of inhaled anesthetics was diluted for the patients undergoing OLV that we discussed.
As far as we are concerned, inhaled anesthetics interferes with hypoxic pulmonary vasoconstriction (HPV), and causes hypoxemia when used at a minimum alveolar concentration greater than one when OLV(30). However, Prakash(31) observed that volatile agents had a direct effect on bronchial smooth muscle and contributed to bronchodilation, which acted with lower pressures and greater Cdyn in cases of OLV when compared with propofol. Thus, there are both advantages and disadvantages of inhaled anesthetics on the pulmonary function when OLV.
Regarding to inflammatory responses in vitro(32) or in vivo(33), inhaled anesthetics significantly reduced inflammatory responses to lung injury, contributing to immunomodulatory and organ protective effects. In case of the clinical surgery proceeded with OLV, inhaled anesthetics were found to play anti-inflammatory role by acting on the cytokine responses, ischaemia-reperfusion and oxidative stress(15, 34). Besides, meta-analysis concluded that compared with total intravenous anesthesia, total inhaled anesthesia could reduce the alveolar inflammatory responses but had no significantly effect on the systematic inflammatory responses in the meantime(10), which may contribute to the results that inhaled agents have a benefit effect on the incidence of PPCs rather than the other systematically complications.
Given the report by International Perioperative Neurotoxicity Working Group, few evidence is detected on which particular anesthetic is preferable to postoperative cognitive function in general anesthesia(35). Trials have proved that cereal oxygenation saturation is associated with postoperative cognitive dysfunction(36), and OLV is relevant to a definite reduction in partial pressure of oxygen when compared to baseline as well(37). Moreover, trials revealed that oxygenation index were higher in intravenous anesthetics groups compared to inhaled group within the first 30 minutes after OLV(8). However, in consistent with the Working Group’s Recommendation, we found postoperative cognitive function screening (MMSE scores) were still comparable between two groups after OLV. The reasons may lay in that MMSE screening tool is insufficient to measure cognitive function, and the following time of included trials may be not long enough due to postoperative cognitive dysfunction may last for weeks to months. More importantly, desaturation was rare in all participants, which may offset the different effect of anesthetics on cognitive function.
We came across several limitations through the analysis. First, not all included trials applied Clavien-Dindo score to assess major complications systematically in which complications were analyzed with 0 to 5 different severity grades. Due to limited published articles, to reduce the risk of bias as far as possible, the postoperative events are defined as the events that need more intensive treatments. Second, only two included trials reported the rate of mortality. The mortality is relatively low and affected more by multi-factors than anesthetics alone. Thus, the conclusion may be referenced with consideration. Next, parts of data we acquired was transferred from median/range or graph. Although the method we chose was used commonly, the data was not completely original and may add to the risk of error rate. Finally, we limited the language to English, which may increase the risk of publication bias. Thus, as two Cochrane Meta-analysis recommended(3, 7), if researchers doubted the prognosis of different anesthetics when OLV, they should design and carry out more high-quality and large-scaled trials to assess the standardized outcomes in the future.
Conclusion
For the patients suffering OLV, compared to intravenous anesthetics, inhaled anesthetics showed significantly protective effects on the PPCs, but not on the postoperative major complications, cognitive function, hospital stay or 30-days mortality either. Further studies are required to test and verify the conclusion.
Key messages
Comparing to propofol, inhaled anesthetics provided protective effects over postoperative pulmonary complications for the patients undergoing OLV.
No significantly advantages were observed on the other major complications, cognitive function, hospital stay or 30-days mortality between different anesthetics when OLV.
Conflicts of interest
The authors certify that there is no conflict of interest with any financial organization regarding the material discussed in the manuscript.
Funding
This research was funded by a grant from Chengdu Municipal Health Commission (No.2020095).
Authors’ contributions
Conceptualization: Jing Yang, Qing-Hua Huang, Yu Cui. Data retrieval: Jing Yang, Qing-Hua Huang.
Project administration: Yu Cui. Supervision: Rong Cao.
Validation: Rong Cao.
Writing – original draft: Jing Yang, Qing-Hua Huang. Writing – review and editing: Rong Cao, Yu Cui.
All authors read and approved the final version of the manuscript.
Abbreviations
- OLV
- one lung ventilation
- PPCs
- postoperative pulmonary complications
- MMSE
- Mini-mental State Examination.