Effect of remote ischemic preconditioning on lung function after surgery under general anesthesia: a systematic review and meta-analysis

Remote ischemic preconditioning (RIPC) protects organs from ischemia–reperfusion injury. Recent trials showed that RIPC improved gas exchange in patients undergoing lung or cardiac surgery. We performed a systematic search to identify randomized controlled trials involving RIPC in surgery under general anesthesia. The primary outcome was the PaO2/FIO2 (P/F) ratio at 24 h after surgery. Secondary outcomes were A-a DO2, the respiratory index, duration of postoperative mechanical ventilation (MV), incidence of acute respiratory distress syndrome (ARDS), and serum cytokine levels. The analyses included 71 trials comprising 7854 patients. Patients with RIPC showed higher P/F ratio than controls (mean difference [MD] 36.6, 95% confidence interval (CI) 12.8 to 60.4, I2 = 69%). The cause of heterogeneity was not identified by the subgroup analysis. Similarly, A-a DO2 (MD 15.2, 95% CI − 29.7 to − 0.6, I2 = 87%) and respiratory index (MD − 0.17, 95% CI − 0.34 to − 0.01, I2 = 94%) were lower in the RIPC group. Additionally, the RIPC group was weaned from MV earlier (MD − 0.9 h, 95% CI − 1.4 to − 0.4, I2 = 78%). Furthermore, the incidence of ARDS was lower in the RIPC group (relative risk 0.73, 95% CI 0.60 to 0.89, I2 = 0%). Serum TNFα was lower in the RIPC group (SMD − 0.6, 95%CI − 1.0 to − 0.3 I2 = 87%). No significant difference was observed in interleukin-6, 8 and 10. Our meta-analysis suggested that RIPC improved oxygenation after surgery under general anesthesia. Clinical trial number: This study protocol was registered in the University Hospital Medical Information Network (registration number: UMIN000030918), https://center6.umin.ac.jp/cgi-open-bin/ctr_e/ctr_view.cgi?recptno=R000035305.


Risk of bias
The quality of the 71 RCTs was assessed independently by two authors (SK and one of CY, AY, DN, or TM) using the Cochrane Collaboration tool.Detailed quality assessments are shown in Supplemental Table S2.Thirty-two of the 71 the included trials showed a 'low' risk of bias, whereas the rest were judged as 'unclear' as the methods used to protect against bias were not sufficiently reported.

Effect of RIPC on oxygenation
Seven RCTs with 735 patients were analysed to investigate the effect of RIPC on the P/F ratio at 24 h after surgery [7][8][9]12,30,41,67 . The combine results are shown in Fig. 2a.In comparison to patients without ischemic preconditioning, those with RIPC showed a significantly higher P/F ratio at 24 h after surgery, with substantial heterogeneity (MD 36.6 Torr, 95% CI: 12.8 to 60.4,I 2 = 69%) (Fig. 2a).When the analysis was repeated using trials with a low risk of bias (4 RCTs) 8,30,41,67 , we confirmed similar results (MD 50.4 Torr, 95% CI 4.9 to 96.0,I 2 = 65%) (Supplemental Fig. S1).The evidence quality of the P/F ratio was graded as 'low' as there were limitations in terms of the presence of substantial heterogeneity and potential publication bias.
Our pre-specified subgroup analysis was performed from three perspectives: age (over 18 years old or not), anesthetic agent (total intravenous anesthesia or volatile anesthesia), and type of surgery (lung lobectomy or not).All seven of the RCTs targeted adult patients.Three RCTs were performed using total intravenous anesthesia 8,9,30 , and two RCTs were conducted in patients who underwent lung lobectomy 7,9 .No differences were found between the sub-groups (Supplemental Fig. S2).We additionally performed a post-hoc sub-group analysis according to the place of inflation cuff (arm or thigh) and total inflation time (over 15 min or not).The differences between subgroups were non-significant (Supplemental Fig. S2).

Discussion
The present meta-analysis showed a significant benefit of RIPC on oxygenation in patients undergoing various surgical procedures under general anesthesia.Furthermore, we showed that RIPC improved other respiratory outcomes, including A-a DO 2 and the respiratory index.Finally, we found that RIPC had beneficial effects with respect to the reduced duration of MV after surgery and the lower incidence of postoperative ARDS.The lung protective effect of RIPC has recently been evaluated in patients undergoing major surgery including aortic aneurysm repair, pulmonary resection and cardiac surgery 8,9,29 .Although the effect after 24 h postoperatively was not analyzed in our study, othe results are in accordance with these studies.However, the effect of RIPC with respect to the improvement of the P/F ratio showed substantial heterogeneity.Past studies showed that RIPC failed to show a beneficial effect in paediatric patients undergoing cardiac surgery 73 .Additionally, the interaction between RIPC and anesthetic agents is controversial [73][74][75][76] .Although these factors were examined in a sub-group analysis, we failed to find the cause of heterogeneity, as the differences between the groups were nonsignificant.With these results, we further performed a post-hoc sub-group analysis.Based on the past reports of which cardioprotective effect of RIPC depends on the duration and volume of tissue exposed to ischemia 77,78 , the analysis was according to the place of cuff and total inflation time, which also showed non-significant differences between the groups.
The P/F ratio is a widely used clinical indicator of oxygenation, although it is affected by F I O 2 , ventilation perfusion mismatch, and pulmonary shunting.Therefore, we also investigated other oxygenation indices: A-a DO 2 and the respiratory index.Unlike with the pooled data of the P/F ratio, several trials that included those   indices were targeted at pediatric patients 4,18,21,24 .Both A-a DO 2 and the respiratory index were improved in the RIPC group, which reinforces the effectiveness of RIPC on oxygenation in surgical patients.Most eligible trials evaluated cardiovascular surgery, which often requires postoperative MV.We further evaluated the effect of RIPC on the duration of MV.Forty-seven trials recorded the duration of postoperative MV.Each of the trials reported no significance difference in the duration of MV between the RIPC and control groups.However, a substantial proportion of studies showed a trend toward shorter MV duration in the RIPC group.Likewise, Min et al. reported that RIPC reduced the incidence of MV > 48 h after cardiac surgery 8 .Similarly to past studies 79,80 , the present analysis showed that RIPC was associated with a slightly shorter duration of MV (0.9 h).Our analysis also showed high heterogeneity (I 2 = 78%), which was not explained by a sub-group analysis stratified by age, anesthetic regimen, place of inflation cuff or total inflation time.We speculated that the timing of extubation and the duration of MV may be affected by various factors, including recovery from anesthesia or sedatives, haemodynamic/haemostatic stability, and even human resources in the ICU.
We also investigated whether RIPC reduced the incidence of ARDS.Although the management of ARDSincluding lung protective ventilation, the avoidance of excess fluid administration, prone positioning, and steroid therapy-have become widely known, ARDS is still a common cause of postoperative lung injury and worsens mortality and morbidity in surgical patients 81,82 .The results of integrated studies reported that RIPC is associated with a lower incidence of ARDS with a risk ratio of 0.73 and low heterogeneity (Fig. 3b).In past studies, the reported incidence of postoperative ARDS ranged from 3.4 to 20% 82,83 , whereas the studies included in our analysis showed a higher incidence of 5.5-54% 9,12,14 .One possible reason for the discrepancy is that the included studies contained patients treated with highly invasive procedures (e.g., cardiac surgery and lung lobectomy).Another possible explanation is that the definition of ARDS has changed with the times as what was once defined as acute lung injury is now included in mild ARDS whose P/F ratio ranges 200-300 Torr 84,85 .Most of included trials focused on ALI defined according to the diagnostic criteria of American-European Consensus Conference 1994, therefore the effect of RIPC on moderate to severe ARDS requires further investigation 12,14,84 .
Although past studies on RIPC have mostly focused on the heart and kidneys, the mechanism of RIPC in lung tissue has also been investigated 86,87 .We further investigated the correlation of inflammatory cytokines and RIPC.In accordance with another recent analysis 11 , our analysis observed that RIPC was associated with a reduced level of serum TNFα with high heterogeneity, which may be due to the difference of anesthetic method or surgery type (Supplemental Fig. S4).It is speculated that volatile anesthetics have an anti-inflammatory effect, which may mask the effect of RIPC 88 .However, the interpretation of the results should be noted because the number of trials with total intravenous anesthesia was small.
The present study was associated with several limitations.First, most of our analyses showed high heterogeneity.Although we conducted a subgroup analysis based on the findings in past studies, the cause of heterogeneity was not found, especially in regard to oxygenation indices.Second, the possibility of publication bias in relation to our primary outcome, the P/F ratio at 24 h, cannot be excluded due to the small number of trials.These are reasons for downgrading the quality level of a body of evidence.Third, some of the included trials showed a high or unclear risk of bias.Although our main result was not changed even when limited to trials with a low risk of bias (Supplemental Fig. S2), well-designed RCTs are needed to verify the conclusions.Fourth, our analyses have several secondary outcomes.Therefore, the risk of type I error inflates due to multiple comparison.Fifth, 58 of the 71 included trials involved cardiac surgery, although we imposed no restriction on surgery type.Therefore, it is unclear whether the reported effects of RIPC are applicable to other types of surgery (e.g., major abdominal surgery).Finally, the lung-protective effect of RIPC was not evaluated by biomarkers that directly reflect lung injury.Although several studies assessed lung oxidative stress markers (e.g., serum malondialdehyde or 8-isoprostane), a large number of the included studies did not measure specific markers 7,9 .Further studies investigating the efficacy of RIPC on lung protective effects, such as the measurement of lung-specific biomarkers or collecting samples from bronchoalveolar lavage fluid, should be conducted.Likewise, studies examining other protective effect of RIPC on non-lung organ such as kidney or brain and its impact on mortality or hospital stay are needed because past studies including systematic review presented inconsistent results or indicated inconclusiveness [89][90][91] .
In conclusion, the present meta-analysis showed that RIPC improved oxygenation after surgery under general anesthesia.Additionally, RIPC shortened the duration of postoperative MV and reduced the incidence of postoperative ARDS.These findings warrant further investigations focusing on the lung protective effects of RIPC.

Methods
This meta-analysis was performed according to the PRISMA statement as shown in Fig. 1.The study protocol was registered in the University Hospital Medical Information Network (registration number: UMIN000030918).

Eligibility criteria
The inclusion criteria for the analysis required that each study be a prospective randomized trial that evaluated the effectiveness of RIPC in patients undergoing surgery with general anesthesia.There were no restrictions on the type of surgery or anesthetic technique.Animal studies, review papers, and non-randomized controlled trials were excluded.

Search strategy
A comprehensive literature search was performed using MEDLINE, Embase, CENTRAL, Web of Science, Clinical Trials.gov,EU Clinical Trials Registry, and UMIN for randomized controlled trials (RCTs).The following key words were used: remote ischemic preconditioning, anesthesia, and surgical procedures.The full-text of the search formula is shown in Supplemental Document 1.There were no language restrictions.The most recent search was performed in December 2021.

Risk of bias assessment
We estimated the risk of bias in the following methodological domains: sequence generation, allocation concealment, blinding of participants, blinding of healthcare providers, blinding of data collectors, blinding of outcome assessors, incomplete outcome data, selective outcome reporting, and other potential threats to validity.The risk of bias was assessed as low, high, or unclear in each domain.

Quality of evidence assessment
We applied the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach to assess the quality of evidence of the main outcomes as follows: very low, low, moderate, or high.

Data synthesis and analysis
We used the random-effects model of DerSimonian and Laird and calculated the risk ratio (RR) with 95% confidence interval (CI) for dichotomous data 92 .Pooled effects of RIPC on continuous data were estimated using the mean difference (MD) or standardized mean difference (SMD) with 95% CI.We conducted a sensitivity analysis according to the risk of bias.Heterogeneity across studies was quantified using the I 2 statistic, which was considered as low, moderate or high if the value exceeded 25%, 50% or 75%.We conducted a predefined subgroup analysis when there is high heterogeneity: patient age (adult or child), anesthetic method (volatile or intravenous anesthesia), and surgery type (lung surgery or not).A forest plot was used to graphically represent the effect of treatment.The small study effect was evaluated using Egger's regression asymmetry test, with p-values of < 0.1 considered statistically significant.Statistical analyses were performed using the R statistical software package, version 4.0.5 (R Foundation for Statistical computing, Vienna, Austria).The "meta" and "metafor" packages were used to perform the meta-analysis.

Figure 1 .
Figure 1.PRISMA flow diagram for study selection.

Figure 2 .
Figure 2. Forest plot with 95% confidence interval (CI) of P a O 2 /F I O 2 (P/F) ratio (a), A-a DO 2 (b), and respiratory index (c).SD standard deviation, MD mean difference.

Figure 3 .
Figure 3. Forest plot with 95% confidence interval (CI) of the duration of mechanical ventilation (a) and the incidence of acute respiratory distress syndrome (b).SD standard deviation, MD mean difference, RR relative risk.