Prevalence of acute kidney injury after cardiac surgery:A systematic review and meta-analysis on risk factors and different diagnostic criterias(AKIN,RIFLE,KDIGO)

Objective: This systematic review and meta-analysis aimed to determine the incidence and some of risk factors of AKI after cardiac surgery using all three diagnostic criteria (AKIN,RIFLE, and KIDGO). Method: We searched for published literature in the English language in MEDLINE via PubMed, EMBASETM via Ovid, The Cochrane Library, and Trip database. For literature published in other languages, we searched national databases (Magiran and SID) , KoreaMed and LILACS, and we searched OpenGrey (www.opengrey.eu/) and the World Health Organization Clinical Trials Registry (who.int/ictrp) for unpublished literature and ongoing studies.To ensure the literature saturation, the list of the included research references or the relevant reviews found by searching was studied(MS). The keywords used in the search strategy were Acute kidney injury,acute renal failure, creatinine,cardiac surgery,heart surgury, Coronary artery bypass grafting (CABG),valve replacement,RIFLE (risk, injury, failure, loss, end-stage renal disease) ,Acute Kidney Injury Network (AKIN),KDIGO (Kidney Disease: Improving Global Outcomes) , which were combined using the AND, OR, and NOT operators. Results: A total of 33298 patients who had undergone the cardiac surgery were studied. Based on the random effect model the total prevalence of AKI in 33298 patients undergone the heart surgery was 26.3% (95% confidence interval[CI]:26.1%,26.6%, I 2 =99.5%). the total prevalence of AKI in patients undergone cardiac surgery based on AKIN criteria was 21.6% (95% CI –21.2%, 22.1%,I 2 =98.5%) of whom 21.6% (95% CI-21.2%,22.1%,I 2 =98.5%) were classified as AKIN stage 1, whereas 3% (95% CI-2.7%%,3.4%%,I 2 =90.6%) were classified as AKIN stage 2, and 3.2% (95% CI-3.0%%,3.4%%,I 2 =97.2%) were classified as AKIN stage 3. the total prevalence

Conclusion: AKI after cardiac surgery is a common symptom, although most often more severe in elderly patients. The prevalence of AKI after cardiac surgery based on KDIGO criteria was found to be higher than RIFLE and AKIN. The prevalence of AKI regardless of the definition used showed a decreasing trend from 2009 to 2019. Our findings pointed to the superiority of the KDIGO criterion over RIFLE and AKIN for diagnosing and evaluating AKI after cardiac surgery. However, the widespread acceptance of consensus definitions ( RIFLE and AKIN criteria) for AKI is still reflected in the studies. In order to progress further, establishment of a uniform definition for AKI seems necessary.

Background
Acute renal injury (AKI) as a severe complication occurs in 3.5-31.0% of the patients who undergo cardiac surgery, and was found to be one of the most common complications affecting these patients (1). Acute kidney injury (AKI) after cardiac surgery increases the mortality rate (2)(3)(4)(5), prevalence of complications, ICU and hospital stay, and health care costs(6-8)In a study conducted by Kochi et al.( 9), the postoperative AKI prevalence in patients undergone the heart surgery was 34.0%. In contrast, some studies including Santos et al.(10), Chertow et al.(11)and Conlon et al.(12),reported prevalence rates of 16.1%, 2.4%, and 1.1%, respectively. The considerable inconsistency in the post-cardiac surgery AKI prevalence can be mainly attributed to the variation of the diagnostic criteria used (13).To develop a standardized definition of acute renal failure, the Acute Dialysis Quality Initiative Workgroup released the RIFLE classification which covers the three stages of acute renal dysfunction,:risk (R), injury (I), and failure (F) based on baseline changes in sCr, the estimated glomerular filtration rate, and urine output, as well as the two clinical outcomes of loss (L) and end-stage kidney disease (E), based on the length of required renal replacement therapy (14). In light of the fact that little changes in sCr compared to the changes proposed by the RIFLE definition can also be associated with adverse outcomes, AKIN definition and classification were proposed by the Acute Kidney Injury Network (AKIN, 2006). These criteria cover the three stages corresponding to the RIFLE risk, injury, and failure stages, respectively (15). The main difference between the two classifications lies in the fact that the AKIN criteria requires a period of 48 hours for detecting changes in renal function and a smaller change in sCr than RIFLE criteria, aiming to improve the sensitivity and early detection of AKI. Moreover, the AKIN criteria do not entail the knowledge of a baseline sCr. In contrast, they regard the kidney injury as any acute increase in sCr establishing the criteria. Some risk factors affecting acute kidney injury are older age, male gender, diabetes mellitus (16), and the pre-existing chronic kidney disease as the most important one (17).. It also serves as a predictor of acute kidney injury found in the postoperative time (18). Despite several attempts indetermining the etiology and prophylactic measures, there is limited data on its incidence and predictors in the post-cardiac operative stage based on all three different definitions and diagnostic criteria. This systematic review and meta-analysis aimed to determine the incidence and some of risk factors of AKI after cardiac surgery using all three diagnostic criteria (AKIN,RIFLE, and KIDGO).

Inclusion Criteria
The methods used in this systematic review were developed based on the PRISMA guidelines. Cross-sectional studies, case-control, and cohort studies were included in this study, while case studies, letters to editors, case reports, clinical trials, study protocols, systematic reviews, and review articles were excluded.
Participants: All studies on the prevalence of acute renal failure in patients undergoing cardiac surgery were reviewed.

Results:
The main purpose of this study was to determine the prevalence of acute renal failure in patients undergoing cardiac surgery and the data were collected according to the released reports.

Sampling Methods and Sample Size:
All observational studies were included in the systematic review regardless of their design. The minimum sample size was 25 patients.

Searching Strategy
The searches were conducted by two independent researchers to find studies published from 1/1/2009 to 30/5/2019. Studies published in MEDLINE were searched through PROSPERO was also searched to find recent or ongoing reviews .the keywords used in the search strategy were: Acute kidney injury, acute renal failure, creatinine, cardiac surgery, coronary artery bypass grafting (CABG), valve replacement, RIFLE (risk, injury, failure, loss, end-stage renal disease), Acute Kidney Injury Network (AKIN), KDIGO (Kidney Disease: Improving Global Outcomes) which were combined using the AND, OR, and NOT operators.

Study Selection and Data Extraction
Two researchers independently reviewed the research titles and abstracts based on the inclusion criteria. After excluding repeated studies, the full manuscripts of the studies were reviewed according to the inclusion criteria and the data and information about the authors were collected if necessary. The data extracted included general information (the author(s), province, and year of publication), study information (sampling technique, diagnostic criteria, the method of data collection, research conditions, the sample size, and risk of bias) and output scale (prevalence of acute kidney injury).

Quality Assessment
To assess the quality of the methodology and the bias risk, each observational study was evaluated using the scale developed by Hoy et al. This 10-item scale assessed the quality of the study in terms of the external validity (items 1 to 4 used to assess the target population, the sampling frame, the sampling method, and the minimum participation bias) and the internal validity (items 5 to 9 used to assess the methods of data collection, the problem statement, and the instruments and scale, and item 10 evaluates the analysis bias). The risk of bias was assessed independently by the two researchers and any inconsistency in the obtained values was resolved by consensus.

Data Collection Procedure
All studies that met the inclusion criteria were used in the data collection process after systematic review and data were mixed using the funnel plot. The random-effects model was evaluated based on the overall prevalence of the disease among the participants. The heterogeneity of the initial studies was assessed using the I 2 test. Besides, the subgroups were analyzed to determine heterogeneity by the participants' age, year of publication, and country. Finally, a meta-analysis of the collected data was performed using STATA14 software.

Research Selection
A total of 1413 articles were extracted through preliminary searches in various databases.
Of the 1294 studies identified by the analysis of titles and abstracts, 387 were eliminated because of irrelevant titles. Out of 907 existing articles, 657 articles were excluded for some reasons. Of the 657 excluded studies, 377 did not have a complete manuscript, 25 were review articles, 7 were letters to the editor-in-chief, and 248 did not meet the inclusion criteria. Of the remaining 250 studies, 22 met the inclusion criteria ( Fig.1).

Research characteristics
A total of 33,298 patients who underwent heart surgery were evaluated. Of the 21 studies, 9 were retrospective, 11 were prospective, and the research design was not mentioned in the other study. A total of 21 studies from 6 countries that met the inclusion criteria were  Table 1).

The meta-analysis of the prevalence of acute renal failure after cardiac surgery according to the RIFLE criteria
According to the random-effects model and the RIFLE criteria, the overall prevalence of acute renal failure in patients undergoing cardiac surgery was 26.0% (25.6% -26.5%, CI = 95%, I 2 = 99.4%) ( Table 4, Fig. 4).

The meta-analysis of the prevalence of acute renal failure after cardiac surgery according to KDIGO criteria
According to the random-effects model and KDIGO criteria, the overall prevalence of acute renal failure in patients undergoing cardiac surgery was 34.7% (33.8%-35.7%, CI = 95%, I 2 = 98.%) (  (Fig. 7).

A meta-analysis of the prevalence of acute renal failure after cardiac surgery in children
According to the random-effects model, the overall prevalence of acute renal failure after heart surgery in children was 37.5% (40.9% -34.5%, CI = 95%, I 2 = 94%) ( Table 6, Fig. 6).

Meta-regression between the year of publication and the prevalence of acute renal failure after cardiac surgery
The meta-regression of the studies was evaluated according to the relationship between the prevalence of acute renal failure, the year of publication of the study, and the overall rate of acute renal failure. There were no significant linear trends in the univariate metaregression to explain the change in the effect size of the year the study was published (Fig. 10).

The results of meta-regression between participants' age and frequency of acute renal failure in patients undergoing cardiac surgery
The regression test was performed to explore the relationship between the prevalence of acute renal failure, the participants' age, and the total rate of acute renal failure. There was no significant linear trend in univariate meta-regression to explain the effect size of participants' age ( Fig. 9).

Publication bias
The funnel plot displayed in Fig. 8 does not show the publication bias and the funnel plot is symmetric. The circle size indicates the size of the studies (larger circles indicate a greater number of samples and smaller circles indicate fewer samples) (Fig. 8).
[1] Confidence interval Discussion Nearly, 1 million persons are annually admitted for coronary artery bypass grafting in the world (43). 30% of them are being at the risk of developing AKI (44).Despite all advances made in the AKI pathogenesis and treatment, it is still found as a relatively common complication and an independent risk factor associated with mortality in critically ill patients (45-47). Furthermore, there are still some disagreements about AKI and, in particular, its definition (48). The inconsistency in data on the actual incidence of this complication is the result of the lack of consensus supported by a considerable variation (1% -30%) in the reported incidence of post cardiac-surgery AKI (44

Limitations
The required data were collected from the existing information from various studies. Some studies were clinical trials while others were observational investigations. The plans and objectives of the studies, the input and output indexes and variables, and the data collected were significantly different in these studies. In addition, there were a few studies on the KDIGO criteria as the diagnostic inputs.

Strength
This study is the first systematic review and meta-analysis that compared the prevalence of AKI after cardiac surgery based on three diagnostic criterias. It also compared simple diagnostic instruments based on easily accessible information, even for non-nephrologists; and thus, the present study, despite the limitations described, makes an important contribution to the medical community. This study is also significant for its comprehensive review of all relevant studies, accurate assessment of the bias risk of selected studies using the QUADAS-2 tool, extraction of duplicate data and subgroup analysis, and its sensitivity to investigate differences in the prevalence of AKI in studies with high, low, and     prevalence of AKI after cardiac surgery based on RIFLE criteria Figure 5 prevalence of AKI after cardiac surgery based on the KDIGO criteria   Meta-regression between age of the participants and prevalence of AKI after Cardiac surgery Figure 10 Meta-regression between publication year of study and prevalence of AKI after Cardiac surgery