A Novel Prediction Model of Acute Kidney Injury Based on Combined Blood Variables in STEMI

Background Development of acute kidney injury (AKI) is associated with poor prognosis in patients with ST-segment elevation myocardial infarction (STEMI). Objective This study sought to investigate whether a combination of pre-procedural blood tests could predict the incidence of AKI in patients with STEMI. Methods A total of 908 consecutive Japanese patients with STEMI who underwent primary percutaneous coronary intervention within 48 hours of symptom onset were recruited and divided into derivation (n = 617) and validation (n = 291) cohorts. A risk score model was created based on a combination of parameters assessed on routine blood tests on admission. Results In the derivation cohort, multivariate analysis showed that the following 4 variables were significantly associated with AKI: blood sugar ≥200 mg/dL (odds ratio [OR]: 2.07), high-sensitivity troponin I >1.6 ng/mL (upper limit of normal ×50) (OR: 2.43), albumin ≤3.5 mg/dL (OR: 2.85), and estimated glomerular filtration rate <45 mL/min/1.73 m2 (OR: 2.64). Zero to 4 points were given according to the number of those factors. Incremental risk scores were significantly associated with a higher incidence of AKI in both cohorts (P < 0.001). Receiver-operating characteristic curve analysis of risk models showed adequate discrimination between patients with and without AKI (derivation cohort, area under the curve: 0.754; 95% confidence interval: 0.733-0.846; validation cohort, area under the curve: 0.754; 95% confidence interval: 0.644-0.839). Conclusions Our novel laboratory-based model might be useful for early prediction of the post-procedural risk of AKI in patients with STEMI.

long-term outcomes (5,6). The mechanisms underlying AKI in the clinical setting of STEMI and subsequent primary percutaneous coronary intervention (PCI) are multifactorial (7). Contrast agents are known to have direct cytotoxic effects on renal tubular cells, as well as indirect cytotoxic effects through changes in renal blood flow causing regional hypoxia (8,9).
In contrast, key pathogenetic mechanisms of AKI result from systemic and renal hemodynamic changes secondary to impaired cardiac output and increased venous congestion (4). Moreover, an imbalance of endogenous vasodilatory and vasoconstrictive factors is involved (10). A burst of excess immunological and inflammatory activation is also likely the potential cause of further renal injury (5). Thus, these pathophysiological responses that occur during the postacute myocardial infarction (AMI) phase have a strong association with the occurrence of AKI.
Several risk score models have been developed to assess the incidence of AKI, including the Mehran score (11)(12)(13)(14). Mehran's model requires 8 variables to assess the risk of AKI, as follows: age >75 years, hypotension, congestive heart failure, hemoglobin, estimated glomerular filtration rate (eGFR), diabetes, contrast volume, and need for intra-aortic balloon pumping (IABP). However, the volume of contrast material administered and use of a hemodynamic support device are unknown before the procedure.
To date, several blood biomarkers have been reported to be potential tools for predicting AKI in patients with AMI (13,(15)(16)(17)(18). Given that these biomarkers may reflect different aspects of pathophysiological responses that occur during the post-AMI phase, we hypothesized that a combination of biomarkers might provide more accurate and useful information for risk stratification than the information provided by any individual biomarker. In this study, we sought to develop a risk score prediction model, based on a combination of parameters obtained on routine blood tests, for in-hospital AKI in patients with STEMI who underwent primary PCI and to compare the predictive utility of that model with that of the conventional Mehran model.  (20). Then, the patients received optimal medications (20).

METHODS
DEFINITION OF AKI. The initial serum creatinine level at admission was used to determine baseline renal function. The definition of AKI was based on change in serum creatinine using the Kidney Disease: Improving Global Outcomes criteria, as follows (21): stage 1, a rise in serum creatinine by $0.3 mg/dL  Values are mean AE SD, n (%), median (interquartile range), or n.
AKI ¼ acute kidney injury; BNP ¼ B-type natriuretic peptide; BS ¼ blood sugar; CABG ¼ coronary artery bypass grafting; CK ¼ creatine kinase; CRP ¼ C-reactive protein; ECMO ¼ extracorporeal membrane oxygenation; eGFR ¼ estimated glomerular filtration rate; HbA1c ¼ glycosylated hemoglobin; HDL-CHO ¼ high-density lipoprotein cholesterol; HR ¼ heart rate; hsTnI ¼ high-sensitivity troponin I; IABP ¼ intra-aortic balloon pumping; LAD ¼ left anterior descending; LCX ¼ left circumflex; LMT ¼ left main trunk; LDL-CHO ¼ low-density lipoprotein cholesterol; LVEF ¼ left ventricular ejection fraction; MI ¼ myocardial infarction; PCI ¼ percutaneous coronary intervention; RCA ¼ right coronary artery; TIMI ¼ Thrombolysis In Myocardial Infarction; WBC ¼ white blood cell. within 48 hours, or an increase in serum creatinine of at least 1.5 times the baseline that is known or presumed to have occurred within the prior 7 days; stage 2, more than a 2-to 2.9-fold increase in serum creatinine; and stage 3, more than a 3-fold increase in serum creatinine from baseline or initiation of renal replacement therapy or an increase in serum creatinine to >4.0 mg/dL.   Values are mean AE SD or median (interquartile range).
Abbreviations as in Table 1.   Table 3), and 4 variables were ultimately proved to be significantly associated with AKI, as follows:   Tables 1 and 2).
LABORATORY-BASED PREDICTION OF AKI. In the derivation cohort, an increased total risk score was significantly associated with an elevated incidence of AKI (P for trend < 0.001) (Figure 2A). The risk score also showed a significant trend for the incidence of AKI in the validation cohort (P for trend <0.001) ( Figure 2B).
The risk model displayed adequate discrimination between patients with or without AKI in the validation ( Figure 3A) and derivation ( Figure 3B Figures 4A and 4B).   Several biomarkers related to cardiac, metabolic, hematologic, and inflammatory responses have been reported to be independent predictors of AKI in patients with AMI (13,(15)(16)(17)(18). Pre-existing chronic kidney disease is known as the strongest risk factor, with a lower level of kidney function associated with a higher degree of risk (11)(12)(13)(14). Our present study added the following 3 blood variables as potential predictors of AKI: hsTnI, albumin, and blood sugar.
Our study demonstrated a relationship between hsTnI and the risk of AKI in treated patients with STEMI. A previous report showed that troponin kinetics could be used to estimate the onset of ischemia in patients with STEMI, and this "biochemical ischemic time" correlates well with infarct size and long-term mortality (23). Previous AKI risk prediction models have focused on ischemia time or left ventricular ejection fraction (12,24). Thus, it is suggested that the incidence of AKI is associated with the degree of myocardial injury as assessed by hsTnI.
Inflammation and oxidative stress have been reported to be mechanisms underlying AKI (25).
Acute coronary syndrome is an inflammatory state, and the serum albumin level often decreases in this situation (26). Therefore, the serum albumin level might have a role as an inflammatory marker in patients with STEMI. Moreover, albumin has antioxidant activities (27). In fact, several studies have suggested that albumin itself can protect the kidney from injury (28). Nonetheless, it is still unclear whether the unfavorable impact of hypoalbuminemia in the early phase of acute coronary syndrome reflects the state of inflammation or an independent effect of albumin itself.
Numerous studies have described an association between hyperglycemia on admission and AKI (16).
The mechanisms underlying this relationship are explained by the fact that an acute increase in blood sugar suppresses flow-mediated vasodilatation, likely through increased production of oxygen-derived free Goriki et al We focused on routine pre-procedural blood tests to develop a risk score model for predicting in-hospital acute kidney injury in Japanese patients with ST-segment elevation myocardial infarction who underwent primary percutaneous coronary intervention. The predictive value of our laboratory-based model composed of 4 laboratory parameters was comparable to that of the conventional Mehran's model composed of laboratory and nonlaboratory information. Our results suggest that this laboratory-based model is useful for early prediction of the postprocedural risk of acute kidney injury in that patient population. AUC ¼ area under the curve; CI ¼ confidence interval; eGFR ¼ estimated glomerular filtration rate; hsTnI ¼ high-sensitivity troponin I; IABP ¼ intra-aortic balloon pumping. radicals, and increases oxidative stress that may exacerbate the deleterious effects of contrast agents on the kidney (29,30). Moreover, acute hyperglycemia may induce a transient osmotic diuresis, resulting in volume depletion and dehydration, which are associated with an increased risk and severity of AKI.
Recently, Tsai et al (14). developed a new risk score model from data based on large numbers of patients undergoing PCI. That model was composed of 11 preprocedural variables and showed good discrimination between STEMI patients with and without AKI.

CONCLUSIONS
Our novel laboratory-based model might be helpful for early prediction of the risk of AKI in patients with STEMI who undergo primary PCI within 48 hours after onset.
ACKNOWLEDGMENT The authors thank Ms. Aya Yamada (Saga University) for her excellent support.