Hypertensive status is associated with renoprotection by remote ischemic conditioning for acute myocardial infarction—a meta-regression and trial sequential analysis of randomized clinical trials

The potential modifiable factors for remote ischemic conditioning (RIC) in reducing contrast-associated acute kidney injury (CA-AKI) in patients with acute myocardial infarction (AMI) have not been investigated. The aim of this meta-regression was to address these issues.We searched Pubmed, Embase and the Cochrane Library database for published randomized controlled trials (RCTs) with registration number CRD42020155532. Nine RCTs comprising of 1540 subjects were included in our meta-analysis. Compared with control group, RIC was associated with reduced incidence of CA-AKI [(9 studies, 1540 subjects, relative risk (RR) 0.51, 95% confidence intervals (CI) 0.35 to 0.76, p = 0.000, I = 52%, p for heterogeneity 0.04)] and major adverse cardiovascular events (MACE) (5 studies, 1078 subjects, RR 0.52, 95% CI 0.38 to 0.73, p = 0.000, I = 9%, p for heterogeneity 0.36) for AMI. In addition, both meta-regression and subgroup analyses have shown that RIC was more effective in the hypertensive patients in reducing CA-AKI for AMI (regression coefficient = –0.05, p = 0.021; for subgroup with more hypertensive patients: RR 0.36, 95% CI 0.25 to 0.52 vs the one with less hypertensive patients: RR 0.72, 95% CI (0.40 to 1.30, p for subgroup difference 0.008). Subsequent trial sequential analysis confirmed the effect of RIC in both CA-AKI and MACE. RIC is an effective strategy in reducing CA-AKI and MACE in patients with AMI, especially for patients with hypertension.


1.Introduction
Acute myocardial infarction (AMI) is one of the leading causes of mortality and morbidity globally. Emergency percutaneous coronary intervention (PCI) was recommended as the standard therapy to perfuse the ischemic myocardium, especially for the ST-segment elevated myocardial infarction (STEMI) [1,2]. However, reperfusion seems like as a double-edged sword, acting the role of rescuing but causing injury for the myocardium. The latter is also known as ischemia-reperfusion injury (IRI) [3]. IRI during emergency PCI not only directly causes damage to heart, but also inducing systematic inflammatory response potentiating the impairment of vital organs such as kidney [4,5].
Contrast-associated acute kidney injury (CA-AKI), as defined by the increment of serum creatinine >44.2 mmoL/L or 25% above the baseline value within 48-72 h of contrast media exposure [6], is a common complication during cardiovascular intervention for reduced renal blood of renal medulla [7,8]. What's more, emergency PCI is more frequently associated with hemodynamic instability which could deteriorate the renal blood infusion. Moreover, patients breaking out with AMI are likely to possess other combined risk factors [9]. Many clinical studies have indicated that CA-AKI not only prolongs the hospitalization but also increases the mortality in long term [10]. Till now, many strategies such as use isotonic contrast agents, cysteine, statins have been tried but failed to show the effective results [11]. Remote ischemic conditioning (RIC), including pre-, per-and post-conditioning, a technique to apply the mild nonlethal ischemia and reperfusion in one organ and protect lethal IRI in other organs or tissues [12]. Accumulating experimental and clinical evidence have reported that RIC was effective in reducing infracted size, attenuating left ventricular remodeling and increasing myocardial salvage after AMI [13][14][15][16]. Many studies have reported that RIC is helpful to reduce CA-AKI in patient with stable coronary artery diseases undergoing either elective coronary artery bypass graft (CABG) surgery or PCI [17,18].
However, the effect of RIC on CA-AKI for AMI is still controversial. Some studies have indicated that RIC was helpful in reducing CA-AKI while others did not show the similar results [19][20][21][22][23][24][25][26][27][28][29]. Given the mixed background, we performed a meta-analysis of randomized controlled trials (RCTs) to explore the effect of RIC in reducing CA-AKI as well as to explore the potential factors affecting RIC in renoprotection for AMI.

Literature search
We reported this meta-analysis following the guidance of PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analysis)statement [30]. We searched Pubmed (Medline), Embase and the Cochrane Library database (http://www.cochrane.org) (up to June 2020) with registration number CRD42020155532. We also manually searched reference lists of the retrieved articles. The key words used in search were (ischemic post-conditioning or post-conditioning or ischemic pre-conditioning or preconditioning or remote ischemic conditioning) paired with (myocardial infarction or myocardial injury or percutaneous coronary intervention or acute kidney injury or contrast induced nephropathy).

Study outcomes and selection
The primary endpoint was CA-AKI, which was diagnosed by increment of serum creatinine >44.2 mmoL/L or above the base value of 25% with 48-72 h exposure of contrast without other obvious factors. The secondary outcome was long-term major adverse cardiovascular events (MACE). The definitions of MACE used by each study and included all cause death, cardiac death, heart failure, revascularization or myocardial infarction.The third outcome was the net change in creatinine or estimate glomerular filtration (eGFR) due to RIC.
Inclusion criteria for the retrieved studies were as follows: (1) prospective RCT design; (2) performed in the patients with AMI (including STEMI and non-STEMI (NSTEMI)); (3) inclusion of outcomes of CA-AKI; (4) inclusion of multivariable-adjusted or unadjusted relative risk (RR)/hazard ratio (HR) and their corresponding 95% confidence intervals (CI); or provided the number of events and total population in each group;

Quality assessment
Two authors (Yuehua Li and Ying Lou) assessed the quality of the RCTs by the Cochrane criteria including adequacy of random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, intentionto-treat analysis and other bias [31]. Trials scored one point for each item addressed. If 3 quality criteria were not met, the study was classified as having high risk of bias; others were classified as having moderate (3-5 points) or low (≥5 points) risk of bias.

Data extraction
Data were extracted by two independent authors (Yuehua Li and Ying Lou). Discrepancies were resolved by group discussion. The extracted data included source of study (author, publication year, country), population characteristics [protocol, conditioning, patients number, mean age, male proportion, percentage of smoking, diabetes mellitus (DM), hypertension, hyperlipidemia, multi-vessel disease, left anterior descending (LAD)branch involving, use of angiotensinconverting enzyme inhibitor(ACEI)/angiotension receptor blocker (ARB), beta-blocker (BB) and statin, follow-up period], the clinical endpoints, RRs or HRs and their corresponding 95% CI.

Statistical analysis
We considered the HRs as RRs in the prospective studies. We calculated the RR by the number of events and total population in the RIC and control group. The random-effect model was also used in the pooled analysis for the potential clinical heterogeneity [32]. The heterogeneity was assessed by Q statistic, I 2 and p value (I 2 > 50%, p < 0.05 was considered to be statistically significant). If there is no significant heterogeneity, we would use the fixed-effect model. Univariable meta-regression analyses (including age, percentage of male, smoking, DM, hypertension, hyperlipidemia, multi-vessel disease, LAD inclusion, use of ACEI/ARB, BB and statin, protocol time, study quality, area) were conducted to explore the potential sources of heterogeneity [33]. Subgroup analyses we real so conducted to explore the potential sources of heterogeneity by specified study characteristics including area (China or other countries), study quality [high risk (quality score <3 points) or moderate/low risk (quality score ≥3 points)], protocol (conditioning time ≥15 min or <15 min), conditioning (by preconditioning or post-conditioning), AMI type (STEMI or NSTEMI) and the mean of the mentioned factor such as age, proportion of smoking, DM, hypertension, hyperlipidemia, multi-vessel disease, LAD occlusion, use of ACEI/ARB, BB and statin [33].
Publication bias was assessed by Begg's test and Egger's test [34]. We performed the trial sequential analyses (TSA) of CA-AKI or MACE following AMI based on the data from our pooled analysis (RR and incidence of CA-AKI and MACE) to calculate the required sample size for the statistical power (Two sided: Type-I error = 0.05; λ = 2.0; Power = 80%). Two sided p value < 0.05 was considered to be significant. All data analyses were performed by STATA software (10.0 version, StataCorporation, TX, USA), REVMAN software (version 5.0; Cochrane Collaboration, Oxford, United Kingdom) and Trail Sequential Analysis (Copenhagen Trial Unit, Copenhagen, Denmark).

Search results
We initially identified 9223studies by database and manual searching. After exclusion of duplicates and nonrelevant studies, 29 potential articles were selected for detailed evaluation. We further excluded 20 articles as shown in Fig. 1. Finally, nine studies were included. Among them, all have reported the impact of RIC on CA-AKI and five about the endpoint of long-term MACE [19,[21][22][23]29].

Meta-regression and subgroup analysis
For the main endpoint of CA-AKI, we performed both meta-regression and subgroup analyses by specific study characteristics which was mentioned above. In univariable meta-regression, the percentage of hypertension at baseline was negatively related to risk of CA-AKI (regression coefficient = -0.05, p = 0.021) (Table 2, Fig. 3). This factor was further confirmed by subgroup analysis [for subgroup with hypertensive patients over 58% (mean): RR 0.36, 95% CI 0.25 to 0.52 vsless than 58% (mean): RR 0.72, 95% CI 0.40 to 1.30, p for subgroup difference 0.008] ( Table 3). The subgroup analysis also has indicated that the proportion of age, male, DM, conditioning, quality score might be possible modifiable factors, however these factors were not verified by meta-regression analyses (Tables 3,4).

Trial sequential analysis
To confirmed the pooled effect sizes of CA-AKI or MACE as true estimated effect, the required sample sizes for the CA-AKI is 959 (RR reduction = 50.0%, incidence of Control arm = 22.0%), and MACE (RR reduction = 50.0%, incidence of Control arm = 17.0%) is 561. However, the same size of CA-AKI (1540 vs 959) or MACE (1078 vs 561) is enough for the estimated effect ( Supplementary  Figs. 4,5).

Publication bias
The publication bias of CA-AKI was not observed by Begg's adjusted rank correlation test (p = 0.35) and Egger's test (p = 0.06) (Supplementary Fig. 6).

4.Discussion
In this meta-analysis of nine RCTs including 1540 subjects, we found that RIC was an effective strategy in reducing the incidence of CA-AKI in patients with AMI, with a profound protection associated with a 0.51-fold lower risk. What's more, our meta-regression showed that the effect of RIC seemed to be more beneficial for the hypertensive patients. In addition, RIC was also showed longterm protective effect with a reduce risk of MACE for long term (RR 0.52). To our knowledge, this is the first metaregression analysis focusing on the modifiable factors for renoprotection of RIC in AMI patient.
The effect of RIC from previous studies in reducing CA-AKI in patients with AMI was inconsistent. We combined all available RCTs and found that RIC was an effective strategy in preventing RIC. Our study was in line with previous two meta-analyses which have showed that RIC was beneficial for prevention of acute kidney injury in patients with PCI, coronary artery bypass grafting and other cardiac surgeries [35,36]. However, these two metaanalyses did not discriminate the effect of RIC in different conditions separately. Our study has found that RIC was not only effective in reducing the incidence of CA-AKI in AMI, but also associated with an improved long-term prognosis. The current meta-analysis has extended previous finding that that RIC was effective in reducing CA-AKI in elective PCI [37,38]. In addition, our meta-analysis has indicated that RIC also showed a long-term cardiac protection in reducing MACE. This result was consistent with recent studies which have reported that RIC was an benefit for AMI patents in reducing MACE, heart failure as well as myocardial edema levels, myocardial salvage index [26]. Moreover, the TSA results showed that, assuming future trials record the same event rates as the published trials, they did not need additional participants to provide future meta-analysis with the power to confirm the benefit for CA-AKI and MACE. This meaningful and intriguing finding indicates that RIC is probably an effective renoprotection strategy in the condition of AMI.
The effect of RIC in reducing CA-AKI for AMI may be influenced by some modifiable factors. Our metaregression has showed that hypertension status was negatively associated with the reduced risk of CA-AKI. What's more, our subgroup analyses have showed that the group with more percentage of hypertensive patients was benefit more (RR 0.36 vs 0.72) from RIC treatment under the condition of AMI. Our results were in lined with previous studies which have reported that RIC was effective in reducing systolic blood pressure about 5 mmHg [39,40]. In addition, hypertension is a well-known risk factor for AMI and our results have also suggested that RIC might be more effective for some high-risk patients. Our study was consistent with previous trials which have also suggested that RIC was more effective in kidney protection for high-risk patients undergoing PCI or cardiac surgery [41,42]. Nevertheless, the influential effect of hypertension in RIC-induced renoprotection needs more large sample-size and high-quality clinical trials to verify in future.
Results from our meta-analysis indicated that RIC was an effective strategy in reducing CA-AKI and MACE for AMI. This meaningful and intriguing finding indicates that RIC is probably an effective renoprotection strategy in AMI. Hence, routine performance of RIC would be helpful for renal protection under the condition of acute ischemic events. Future experimental studies are needed to explore the mechanism about the RIC for kidney protection. In addition, large randomized controlled trials are necessary to extend the investigation of the effect of RIC for renal protection in both cardiac and non-cardiac conditions.

Limitation
Our meta-analysis has some limitations. First, we did not use the multivariable adjusted RR for the effect size, resulting inpotential residual confounders. Second, the excluded studies which performed in patients with acute coronary syndrome might influence on the effect size. Third, five included studies were from China. Although our subgroup analyses have indicated that area was not a modifiable factor, it needs further investigation for RIC in reducing CA-AKI for AMI. Finally, our meta-analysis used pooled data, rather than individual data, which restricted detailed analysis for the potential confounding factors.

Conclusions
Our meta-analysis of nine RCTs comprising 1540 patients has demonstrated that RIC remains an effective strategy in reducing CA-AKI for AMI, especially for the hypertensive group. Routine RIC in AMI should be recommended for renal protection.