Meta-analysis of the effects of ischemic postconditioning on structural pathology in ST-segment elevation acute myocardial infarction

In this meta-analysis, we assessed cardiac magnetic resonance imaging data to determine the effects of local and remote ischemic postconditioning (LPoC and RPoC, respectively) on structural pathology in ST-segmentel elevation acute myocardial infarction (STEMI). We searched the Pubmed, Embase and Cochrane Library databases up to May 2017 and included 12 randomized controlled trials (10 LPoC and 2 RPoC)containing 1069 study subjects with thrombolysis in myocardial infarction flow grade 0~1. Weighed mean difference (WMD), standardized mean difference (SMD), and odds ratio (OR) were used for the pooled analysis. Random-effect model was used for the potential clinical inconsistency. LPoC and RPoC increased the myocardial salvage index (n = 5; weighted mean difference (WMD) = 5.52; P = 0.005; I2 = 76.0%), and decreased myocardial edema (n = 7; WMD = −3.35; P = 0.0009; I2 = 18.0%). However, LPoC and RPoC did not reduce the final infarct size (n = 10; WMD = −1.01; P > 0.05; I2 = 68.0%), left ventricular volume (n = 10; standardized mean difference = 0.23; P > 0.05; I2 = 93.0%), the incidence of microvascular obstruction (n = 6; OR = 0.99; P > 0.05; I2 = 0.0%) or the extent of microvascular obstruction (n = 3; WMD = −0.09; P > 0.05; I2 = 6.0%). This meta-analysis shows that LPoC and/or RPoC improves myocardial salvage and decreases myocardial edema in STEMI patients without affecting final infarct size, left ventricular volume or microvascular obstruction.


INTRODUCTION
Timely restoration of coronary perfusion is the most effective strategy to limit infarction size (IS) and improve clinical outcomesin patients with ST-segment elevation acute myocardial infarction (STEMI) [1]. However, the progressive changes in structure and morphology of the left ventricle after ischemic myocardial reperfusion is associated with 25% of heart failure (HF) cases [2,3]. Hence, accurate evaluation of the effects of STEMI therapy on cardiac structural pathology is critical [4].
Cardiac magnetic resonance imaging (cMRI) has emerged as the most accurate and reliable tool for the evaluation of cardiac structure. Moreover, contrast-enhanced cMRI has been widely used to measure the infarct size with

Effects of ischemic post conditioning on final IS, MSI, and myocardial edema
As shown in Figure 2

DISCUSSION
In this meta-analysis of 12 randomized trials, weassessed1069 STEMI patients that underwent PCI by cardiac magnetic resonance imaging(cMRI). We observed that both LPoC and/or RPoC reduced the extent of MSI and myocardial edema, thereby offering cardioprotection. However, LPoC and RPoC did not affect final IS, LV volume, and the incidence or the extent of MVO. This meta-analysis is the first comprehensive analysis to evaluate structural effects of ischemic postconditioning in STEMI patients using cMRI.
The protective potential of ischemic postconditioning (PoC) for STEMI patients has been confirmed in clinical trials by assessing cardiac enzyme levels and left ventricular function [11,12] and systematically reviewed previously [10]. Some trials have explored the structural effects of PoC in STEMI by angiography [13], echocardiography [14,15] and SPECT [16]. In order to increase the consistency, we included studies that reported structural effects of PoC in STEMI as assessed by cMRI, which accurately measures the infarct size and LV volumes [17][18][19]. Thus, our meta-analysis provides more solid evidence about the structural effects of PoC in STEMI.
During ischemia/reperfusion (I/R), the hydrostatic pressure within interstitial space increases and results in myocardial edema. This contributes to capillary compression and aggravates the extent of cell damage, which is characteristic of severe I/R injury. Since myocardial edema is central to I/R injury, it is critical to analyze the positive effects of ischemic postconditioning. In a dog model, ex vivo assessment of water content showed that LPoC reduced myocardial edema [31]. Improved detection of in vivo myocardial edema by non-invasive T2-weighted imaging [32,33] has led to evaluation of the efficacy of ischemic postconditioning on attenuating reperfusion injury [21,34]. Thuny et al. showed reduction in the extent of myocardial edema by LPoC in STEMI [26]. However, results of many clinical trials evaluating LPoC [21,22,27,30] and RPoC [23] have been controversial. In our meta-analysis, we combined positive [24,26] and negative [21-23, 27, 30] studies and showed that LPoC and RPoC were associated with reduced myocardial edema after STEMI. However, the mechanisms underlying reduced myocardial edema by ischemic postconditioning need further investigation.
To address the effect of cardioprotective interventions on ischemia injury and myocardial edema, we selected T2 weighted cMRI, which is a water-sensitive  technique that measures myocardial edema in vivo without using radiation or contrast agents and accurately represents the size of area at risk [35]. Our analysis showed that LPoC and RPoC decreased myocardial edema, but did not have any effect on the ischemia. The reasons for these effects are unclear. Myocardial edema includes intracellular and extracellular edema. However, cMRI could not distinguish between the two sources of extracellular edema, namely, intravascular water permeation and water release from necrotic cardiomyocytes into the infarcted area. Reperfusioninduced myocardial edema (extracellular space)increased wall thickness and stiffness favoring collagen deposition and fibrosis, which reduced expansion of the infracted area and left ventricular remodeling, regardless of myocardial salvage [36]. Moreover, recent studies using cMRI have demonstrated a bimodal pattern of myocardial edema after I/R injury, namely, an early phase that is reperfusion induced and occurs within 24 h and a late phase that represents the auto-healing process lasting at least 7days [37,38]. However, the cMRI assessments were mainly performed within 1~7 days after PCI in the included trials. This suggested that ischemic postconditioning enhanced cardiomyocyte healing without affecting the infarction size. Previous studies have shown that LPoC and RPoC decreases inflammation and reactive oxygen species generation, which may prevent extracellular edema by increasing microvascular permeability [1,39]. These studies partly explain the dissociation of the beneficial effects of ischemic postconditioning and structural damage in STEMI.
The main strength of our meta-analysis was that we assessed multiple structural parameterssuch as final IS, MSI, left ventricular volume, MVO, and myocardial edema in two settings of ischemic postconditioning namely, LPoC and RPoC using cMRIin STEMI patients. On the other hand, there were several limitations in our study. First, we included veryfew trials and studies and were unable to access individual patient data. Therefore,we may have underestimated the potential influence of comorbid conditions such as diabetes, dyslipidaemia, multi-vessel disease, and LAD as well as effects of cardiovascular medications such as β-blockers [40], glycoprotein IIb/IIIa inhibitors, and statins [10,41]. Second, the relative small number of the enrolled subjects may have decreased the statistical power of our results. Third, we applied the random effect model based on I 2 ≥ 50% and assumed normalized distribution Oncotarget 8096 www.impactjournals.com/oncotarget [42]. But, we can't rule out heterogeneity influencing the outcomes of our study. Fourth, we excluded non-English language publications .Fifth, more studies are necessary to assess thec MRI data regarding the effect of ischemic postconditioning on cardiac structurein AMI with baseline TIMI flow grade 2~3, especially for RPoC. Finally, the long-term heart failure and cardiac mortality needs to be analyzed and the effect of therapy on cardiac structure needs to be confirmed in future clinical trials.
In conclusion, our meta-analysis of cMRI data showed that both LPoC and/or RPoC reduced the extent of MSI and myocardial edema in STEMI patients. However, there were no improvements in final IS, LV volume, and the incidence or the extent of MVO.

Study search strategy and inclusion criteria
We performed this meta-analysis in accordance with PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) [43]. We searched PubMed, EMBase, and Cochrane Library databases up to May 2017 with the following keywords:ischemic postconditioning,remote ischemic conditioning, ischemic postconditioning, acute myocardial infarction and percutanenous coronary intervention. Only prospective RCTs that were published in English and that reported STEMI undergoing percutanenous coronary intervention were included in this meta-analysis. Studies those (1) reported only cardiac enzyme levels and/or left ventricular ejection fraction; (2) did not use cMRI for structural assessment and (3) used pre-procedural TIMI flow grade ≥ 2 for potential spontaneous reperfusion [44] were excluded.

Study selection, quality assessment and data extraction
Two investigators, Yadong Cui and Haiyang Gao, independently reviewed all abstracts and the full text according to the described search strategy and criteria. In case of disagreements, consensus was achieved by discussion. Quality assessment was performed according to the Jadad scoring system: randomization; blinding; withdrawals and dropouts. Each study obtained a score between 0 and 5 based on withdrawals and dropouts and those with a score ≥ 3 were considered high-quality [45].
Data extraction included trial design parameters such asyear, country,protocol algorithm, conditioning delay, symptom-to-balloon time, and follow up, and demographic data of patients such as age, gender, and presence or absence of diabetes mellitus, hypertension, smoking, dyslipidemia, stenting technique, multi-vessel, left anterior descending artery disease(LAD), and treatment with glycoprotein IIb/IIIa inhibitor, β-blockers, and statins.

Evaluation of left ventricular structure by cMRI
We extracted endpoints assessed by structural cMRI imaging after PCI, which included primary endpoints such as final IS, MSI, and myocardial edema as well as additional endpoints such as LV volume and MVO. The final IS was assessed by late gadolinium enhancement of the cMRI images and expressed as percentage of LV mass [46]. The MSI was defined as the AAR minus IS; AAR was assessed by cMRI [47]. The myocardial edema was expressed as percentage of LV mass and evaluated by the T2 weighted method [33]. The LV volume was recorded as LV end-diastolic volume(LVEDV), or LVEDV index (LVEDVI), which is defined as LVEDV divided by body surface area [20]. The incidence and extent of MVO was expressed as percentage of LV mass and detected by late gadolinium enhancement of the cMRI images [20].

Statistical analysis
Data was expressed as mean ± standard deviation or median ± interquartile range for continuous variables. We calculated WMD or SMD for LVR to obtain the pooled estimates with 95% confidence intervals (CIs). For dichotomous ones (reported with incidence), we calculated odds ratio (OR) with 95% CIs. We set I 2 ≥ 50.0% as significant heterogeneity and used random-effects model for analyzing such parameters [48]. P < 0.05 (2-sided) was considered statistically significant. All statistical analysis was performed by Stata version 9.0(Stata Corporation, College Station, TX) and RevManversion 5.0(Cochrane Collaboration, Oxford, UK) softwares.

ACKNOWLEDGMENTS AND FUNDING
This work was supported by the Health Industry Special Scientific Research Project (NO.201402019). We are grateful to Dr. Yadong Cui and Dr. Haiyang Gao from Beijing hospital.

CONFLICTS OF INTEREST
The authors declare that there are no conflicts of interest.