Background
Acute coronary syndrome (ACS), including ST-elevation myocardial infarction (STEMI), non-ST elevation myocardial infarction (NSTEMI) and unstable angina (UA), is one of the most dangerous presenting conditions. It usually results from the abrupt reduction of coronary blood flow caused by the ruptured coronary atherosclerotic lesion and superimposed acute thrombosis. Primary percutaneous coronary intervention (PCI) is recommended as a first-line treatment of the reperfusion strategy and it contributes to the salvage of the myocardium and improvement of the prognosis [
1‐
3]. However, even after successful PCI, both the short-and long-term morbidity and mortality of ACS patients undergoing PCI are alarmingly high [
4‐
6]. Therefore, it is critical to investigate the risk factors associated with worse outcomes.
Contrast-induced acute kidney injury (CI-AKI), also known as contrast-induced nephropathy, is a frequent complication of PCI due to the use of iodinated contrast agent [
7‐
10]. In particular, patients who underwent primary PCI are at higher risk of CI-AKI than those treated with elective PCI due to hemodynamic instability and lack of adequate renal prophylactic therapy [
8,
11]. The primary manifestation is the mild decline in renal function, occurring 1 to 3 days after contrast media exposure. According to several reports, CI-AKI after PCI for ACS was associated with adverse clinical outcomes, including all-cause mortality [
12‐
14], short-term in-hospital mortality [
15‐
17], major adverse cardiac events (MACE) [
18‐
20], major adverse cardiovascular and cerebrovascular events (MACCE) [
21‐
23] and stent restenosis [
19,
24,
25]. However, several other studies demonstrated the inconsistent results [
26,
27]. In addition, the evaluated values among studies varied widely. For example, the increased risk of all-cause mortality for CI-AKI patients after PCI for ACS varied from 54% [
26] to 799% [
17].
The aim of the present meta-analysis was to evaluate the associations between CI-AKI and adverse clinical consequences, including all-cause mortality, short-term all-cause mortality, MACE, MACCE and stent restenosis in patients undergoing PCI for ACS.
Methods
Our study was designed, conducted and reported based on Preferred.
Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).
guidelines (see Additional file
1) [
28].
Search strategy
EMBASE, PubMed, Web of Science™ and Cochrane Library databases were systematically searched from inception to December 16, 2016 using Google Chrome software with following terms: “acute myocardial infarction”, “AMI”, “myocardial infarction”, “MI”, “non-ST segment elevation myocardial infarction”, “ST-segment elevation myocardial infarction”, “acute coronary syndrome”, “ACS”, “unstable angina”, “acute ST-elevation myocardial infarction”, “acute non-ST segment elevation myocardial infarction”, “complete revascularization”,“revascularization”, “revascularization”, “percutaneous coronary intervention”, “PCI”, “contrast-induced acute kidney injury”, “contrast-induced nephropathy”, “contrast media use”, “CIN”, “CI-AKI”, “mortality”, “cardiovascular”, “events”, “death”, “outcome”, “Hemodialysis”, “Haemodialysis”, “Dialysis”, “peritoneal dialysis”, “Adverse effect”, “prognosis”, “chronic kidney disease”, “CKD”, “end-stage renal disease”, “ESRD” (see Additional file
2). Furthermore, we searched the reference lists of included articles for additional studies. The full text of a record was reviewed if there was any doubt to the eligibility of it. Two of the authors (Y.Y. and K.G.) independently screened titles and abstracts, analyzed full-text articles, and ascertained the final eligible records. Divergences were resolved by discussion or consulting a third author.
Selection criteria
We selected articles that (1) were cohort studies published as original studies, (2) enrolled the populations of general PCI-treated ACS patients, including acute myocardial infarction, non ST-segment elevation myocardial infarction, ST-segment elevation myocardial infarction, unstable angina pectoris or ACS (i.e., without any additional pathologies, neither be given other special operation or interventional therapy), (3) observed the exposure of CI-AKI with clear definition, (4) evaluated the associations between CI-AKI and risk of any clinical outcomes during follow-up, and quantitative effect size could be extracted, calculated or obtained from original authors, including relative risks, hazard ratios or odds ratios (RRs, HRs or ORs) and associated 95% confidence intervals (CIs), and (6) had no restriction of language.
Exclusion criteria
Studies should be excluded by the following criteria: (1) meta-analyses, reviews, editorials, meeting abstracts, comments, letters, notes, animal studies, case-control and cross-sectional studies were excluded; (2) we could not extract available data from the published result, and the authors did not reply our request for more data; (3) only one or two studies evaluated the association between CI-AKI and risk of the same outcome.
If two articles were from the same cohort but reported the different outcomes, we included both; otherwise, we included the most informative study.
We extracted the information of interest from each study including study characteristics (study group name, publication year, geographical location, study design, sample size, clinical scenario, age at baseline, male percentage, follow-up duration, prevalence of CI-AKI, CI-AKI definition, smoker percentage, body mass index, Killip class, left ventricular ejection fraction, baseline serum creatinine, baseline eGFR, Family history for coronary artery disease), complication (prevalence of hypertension, diabetes, hyperlipemia), diseases’ history (myocardial infarction, previous coronary bypass surgery, PCI), medications for ACS (Aspirin, ACEI/ARB, β-blocker, Statins), use of contrast volume and confounder-adjusted and unadjusted risk estimates (including RRs, HRs and ORs) and 95% CIs for any adverse outcome. All outcomes in our eligible articles included all-cause mortality, in-hospital mortality, 30-day mortality, cardiovascular mortality, MACE, MACCE, stent restenosis, reinfarction and major bleeding. We conducted meta-analyses if more than two studies had available data for an outcome. Two authors independently extracted data using standardized tables and resolved discrepancies through discussion and consensus. We made a request to the corresponding author of original studies for further information when necessary and included it if responses were obtained.
Quality assessment
We evaluated the quality of studies using Newcastle–Ottawa Quality Assessment Scale (NOS) of cohort studies [
29] which assessed the following: (1) exposed cohort truly or somewhat representative, (2) non-exposed cohort drawn from the same community as the exposed cohort, (3) ascertainment of exposure, (4) outcome of interest not present at start, (5) study controls for age, (6) study controls for≥3 additional risk factors, (7) assessment of outcome (independent blind assessment or record linkage), (8) follow-up≥24 m, (9) complete accounting for cohorts or subjects lost to follow-up unlikely to introduce bias. This scale allocates a maximum of 9 points for quality of study participants. Overall study quality was graded as good (score, 7–9), fair (score, 4–6), or poor (score, 0–3) [
30]. Two authors performed the quality assessment independently and divergence was resolved by discussion or consensus.
Statistical analysis
The studies included in our analysis reported different effect measures (HRs, RRs or ORs). The RR was used as a common measure of association across studies. We combined HRs as RRs in our article. OR was transformed into RR according to the formula RR = OR/[(1-P0) + (P0 × OR)] where P0 stands for the incidence of CI-AKI whenever possible [
31]. Otherwise, ORs was considered as RRs because of the low incidence of CI-AKI. Heterogeneity of RRs across included studies was examined using χ
2 based on Q-statistical test and quantified by I
2 index. Roughly, heterogeneity was considered significant if
P < 0.10 and Higgins I
2 values of 25, 50, and 75% were respectively considered as low, moderate, and high inconsistencies. We used random effects models when there was obvious heterogeneity among studies (I
2 ≥ 50%); otherwise fixed effects models were used. Publication bias was assessed through Begg’s and Egger’s testing (
P < 0.10 was significant) as well as funnel plot. Jackknife sensitivity analyses were conducted by recalculating pooled RRs with the removal of a single study once from the baseline group. To explore the source of heterogeneity, subgroup meta-regressions were conducted. All analyses and graphs were made using Stata 10.0 (College Station, TX, USA) and GraphPad Prism 6.
P < 0.05 by 2 tailed were set to be significant.
Discussion
To our knowledge, this study was the first meta-analysis to evaluate the association between CI-AKI and risk of adverse outcomes in ACS patients treated with PCI. Twenty two studies were eligible. We confirmed that CI-AKI might increase 216, 455, 49, 86 and 50% relative risk for all-cause mortality, short-term all-cause mortality, MACE, MACCE, and stent restenosis in ACS patients treated with PCI.
There was moderate heterogeneity among the studies included for all-cause mortality and short-term all-cause mortality, but no heterogeneity among the studies included for MACE, MACCE and stent restenosis.
CI-AKI was associated with an increased risk of all-cause mortality, short-term all-cause mortality, MACE, MACCE and stent restenosis of ACS patients after PCI in our study. Especially for short-term all-cause mortality, the increased risk was up to 455% compared to none CI-AKI patients. According to other reports, in ACS patients treated with PCI, CI-AKI was also associated with other adverse outcomes, including net adverse clinical events (HR = 1.53, 95% CI: 1.23–1.9 [
19]), major bleeding (HR = 2.69, 95% CI: 2.26–3.2 [
24] and HR = 2.07, 95% CI: 1.57–2.73 [
19]) and in-hospital ischaemic stroke (HR = 2.91, 95% CI: 1.03–8.24 [
38]).
The reasons for taking CI-AKI as a strong predictor of clinical adverse outcomes were complex. It was demonstrated that the renal function of 54.6% of CI-AKI patients did not return to baseline level after 2 weeks [
39]. CI-AKI patients were at increased risk of progressive decline in kidney function [
40,
41]. Over 45% of patients had persistent renal dysfunction leading to chronic kidney disease (CKD) [
42]. Patients with CKD have more severe and diffuse coronary artery disease, higher rates of traditional coronary risk factors and complications such as diabetes mellitus, hypertension, left ventricular hypertrophy dyslipidemia, atrial fibrillation, and congestive heart failure [
43,
44], and more restrictions to therapies for other diseases [
43,
45].
For patients undergoing coronary angiography, CI-AKI was also reported to be associated with an increased risk of adverse clinical outcomes in a meta-analysis [
46]. It included 39 studies and showed that CI-AKI was associated with increased risk of mortality (adjusted pooled RR = 2.39, 95% CI: 1.98–2.90), MACE (adjusted pooled RR = 1.98, 95% CI: 1.52–2.59) and end-stage renal disease (pooled crude RR, 15.26, 95% CI: 1.86–125.01) for patients treated with coronary angiography. Our meta-analysis focused on ACS patients treated with PCI. Noteworthy, only one same study [
12] was included in both our study and the above study. The occurrence of ACS was more of an emergency and increased risk of short-term mortality associated with CI-AKI was higher (pooled RR 5.55; 95% CI 3.53–8.73). Therefore, CI-AKI should be paid more attention to for ACS patients treated with PCI.
There were some limitations of our meta-analysis. Firstly, the diagnostic criteria for CI-AKI, MACE and MACCE were not completely uniform in all included studies. Secondly, studies included in this meta-analysis were not adjusted for consistent factors when assessing the association between CI-AKI and outcomes. Finally, the differences of demographic and clinical characteristics regarding race, genetics, geographical locations and PCI procedure across the different populations pooled were not thoroughly studied. All the limitations might have effect on the risk evaluation of prognosis.