Risk Factors Profile for Acute Kidney Injury after Cardiac Surgery Is Different According to the Level of Baseline Renal Function

Abstract

Acute kidney injury (AKI) occurs frequently after cardiac surgery. Although numerous variables were identified as predictors for AKI, there is a lack of information about possible differences in risk factors according to the level of preoperative renal function. Preoperative, intraoperative, and postoperative data from 4118 adult patients submitted to cardiac surgery from January 1, 2000, to December 31, 2005, were included in the study. AKI was defined by an increase in serum creatinine (S_Cr) ≥ 0.3 mg/dL after surgery. Patients were stratified into two groups: group 1, C_Cr ≥ 60 mL/min/1.73 m² BSA; group 2, C_Cr < 60 mL/min/1.73 m² BSA. Risk factors were assessed using a multiple logistic regression model. In all, 749 patients (42.5%) developed AKI. The mortality rate of the entire population was a 5.2%. In patients of group 1 (n = 2678), the variables independently associated with the onset of AKI were age, diabetes, preoperative use of diuretics, non-scheduled surgery, cardiopulmonary by-pass (CPB) time, CPB mean arterial pressure, hemodilution, and postoperative use of norepinephrine. Baseline S_Cr was associated with AKI only in patients of group 2 (n = 1440). Age, EuroSCORE, non-coronary surgery, hemodilution, and postoperative use of vasoactive drugs were also predictors of AKI in this group of patients. The mortality rate was significantly higher in group 2 than group 1 (8.5% vs. 3.6%, p = 0.000). In conclusion, the present study demonstrated a difference in variables associated with postoperative AKI, according to baseline renal function. The degree of preoperative renal function was a predictor of AKI only in patients with CKD stages 3–4, as well as other risk factors. In addition to other well known risk factors for AKI in this setting, the use of diuretics in patients of group 1 and the level of hemodilution during CPB in both groups should be emphasized, as they are potentially modifiable.

Keywords:

• acute renal failure
• cardiac surgery
• risk factors
• diuretics
• hemodilution

INTRODUCTION

Acute kidney injury (AKI) is a frequent and severe complication after cardiac surgery. Its frequency is variable, ranging from 1–5% for severe forms requiring renal replacement therapy, to up to 30% in less severe forms.[1–3] Moreover, according to some investigators, a transient decline in the renal function is a regular consequence of extracorporeal circulation.[4]

AKI after cardiac surgery is associated with a dramatic rise in mortality rate. Depending on definition of AKI, mortality reaches up to 10–20% in non-dialyzed patients[3],[5] and up to 54–70% in dialyzed patients.[6–9] In addition to mortality, AKI adds morbidity and increases the length of stay and hospital charges.[10]

Numerous variables were identified as risk factors for the development of AKI in this setting, including age, male sex, diabetes, NYHA class, peripheral vascular disease (PVD), previous sternotomy, chronic obstructive pulmonary disease, chronic kidney disease (CKD), hypotension during surgery, use of vasoactive drugs, duration of cardiopulmonary bypass, and aortic cross-clamping, among others.[11],[12] The influence of CKD and particularly the level of previous renal function were extensively studied.[1],[3],[13–15] It has been demonstrated that a modest rise in preoperative serum creatinine is sufficient to increase the risk of developing AKI in this setting.[3],[16]

The hypothesis of the study is the demonstration of different risk factors for post cardiac surgery AKI according to baseline renal function. As a result, preventive strategies could be tailored according to baseline renal function. To test this hypothesis, we stratified patients into two levels of estimated preoperative creatinine clearance in order to analyze the effect of a number of pre-, intra-, and post-operative variables on post-operative renal function.

PATIENTS AND METHODS

Design: Observational, Longitudinal Study

All adult consecutive patients submitted to cardiac surgery in the National Institute for Cardiac Surgery—IMPASA between January 1, 2000, and December 31, 2005, were considered for inclusion in the study. ESRD patients under renal replacement therapy, renal transplant recipients, as well as patients who died in the first 24 hours after surgery were excluded. Cardiopulmonary transplantation is not performed in our center. Data from patients are collected prospectively in the ongoing database of the National Institute for Cardiac Surgery, which includes more than 490 variables; thirty-two of them were selected for the analysis. This registry was approved by the Institutional Review Board to record information in cardiac surgery patients.

According to the outlined inclusion criteria, 4118 patients were studied. All patients have at least two measurements of serum creatinine (S_Cr): at baseline and in the post-operative period during hospitalization. When patients had more than one post-operative S_Cr, the highest value was selected.

AKI is defined by an increase in S_Cr equal to or greater than 0.3 mg/dL after cardiac surgery during hospitalization.[17] Baseline creatinine clearance (C_Cr) was estimated by the Cockroft-Gault equation.[18] Patients were stratified for the analysis into two groups according to estimated C_Cr: group 1, C_Cr ≥ 60 mL/min/1.73 m² BSA; group 2, C_Cr < 60 mL/min/1.73 m² BSA. This cutoff point was selected considering the National Kidney Foundation definition of chronic kidney disease.[19]

Surgical and anaesthetic management were described elsewhere.[16] Furosemide was excluded from the anaesthetic protocol because we demonstrated that it has a harmful effect on renal function in the aforementioned study.

Variables described in previous trials as risk factors for AKI in this setting were selected to be tested in the present series of patients. These include age, gender, body surface area, EuroSCORE (see appendix), comorbidities (diabetes, hypertension, COPD, peripheral vascular disease, cerebrovascular disease, smoking, hypertension), left ventricular ejection fraction, prior cardiac surgery, prior miocardial infarction, and drug exposure (radiocontrast, diuretics, digitalis, non-steroids anti-inflammatory drugs, β-blockers, calcium-channel antagonists, angiotensin-converting enzyme inhibitors). Intraoperative variables were also considered, including on-pump and off-pump surgery and inotropic and vasoactive agents. In patients with on-pump surgery, variables analyzed were cardiopulmonary bypass time (CPBT), aortic cross-clamping time (ACT), arterial pressure under on-pump perfusion and aortic cross-clamping time, dieresis, and hematocrit. The use of vasoactive agents and the hemodynamic parameters in post-operative period were also included for the analysis.

Statistical Analysis

Univariate analysis was performed using χ² test for categorical variables and the Student t- test, or Mann- Whitney test, for continuous variables. A separate multiple logistic regression model using the forward stepwise elimination method was performed in each group, with postoperative AKI being the dependant variable (yes/no). The large number of patients included in the analysis allows exploring for confounding factors and multivariate adjust for confounding, including age and serum creatinine, within each strata. This stratified analysis made it possible to explore for the residual contribution of the baseline renal function to the variance of AKI incidence within each group. The variables that reached a 10% probability of random association to onset of post-operative AKI in the univariate analysis were included in the model. The exit criterion was set at level 0.05. The statistical package SPSS 10.0 (Chicago, Illinois, USA) was used for statistical analysis.

RESULTS

In all, 1749 patients (42.5%) developed acute kidney injury according to the definition adopted by authors. In patients of group 1, which corresponds to eC_Cr ≥60 mL/kg/min, the frequency of AKI was 41.9% (1125/2680). In patients of group 2, with preoperative renal insufficiency corresponding to CKD stage 3–4, the frequency of AKI was 43.4% (624/1438). The frequency of AKI that needed renal replacement therapy was 0.6% for group 1 and 3.7% for group 2 (p = 0.001). Mortality rate of the entire population was 5.2% (216/4118). Data are shown in Table 1. Patients with AKI had higher mortality than patients without AKI (8.8% vs. 2.6, p = 0.000). Moreover, mortality rate within groups was also different: all-cause 30-day mortality was higher in group 2 than group 1: 8.5% vs. 3.6% (p = 0.000).

Table 1 Demographic data in the general population

Variable
Number of patients	4118
Age (years)	64.8 ± 10.8
Male gender	2737 (66.5%)
EuroSCORE (median, range)	6 (0–22)
Diabetes (%)	1037 (25.2%)
Baseline SCr (mg/dL)	1.11 ± 0.39
eGFR (mL/min/1.73 m^2)	74.2 ± 32.7
AKI	1749 (42.5%)
AKI-D	69 (1.6%)
Thirty-day mortality rate	216 (5.2%)

S_Cr and eGFR data are expressed as mean ± SD.

Abbreviations: S_Cr: = serum creatinine, eGFR = estimated glomerular filtration rate, AKI = acute kidney injury, AKI-D = acute kidney injury that needed dialysis.

Twenty-five risk factors for AKI in the general population were identified in the univariate analysis (see Table 2). When patients were stratified according to baseline eC_Cr, the risk factors profile was different between groups. In patients of group 1 (n = 2678), advanced age, diabetes, preoperative use of diuretics, non-scheduled surgery, CPB time, the lowest MAP during CPB, the lowest hematocrit during CPB, and post-operative use of norepinephrine were independently associated with the onset of AKI (see Table 3). In patients of group 2 (n = 1440), advanced age, EuroSCORE, non-coronary surgery, baseline S_Cr, the lowest hematocrit during CBP, and post-operative use of vasoactive drugs (dopamine, epinephrine and norepinephrine) were predictors of AKI (see Table 4).

Table 2 Univariate analysis for AKI risk factors in the general population

Variable	AKI (n = 1749)	Non AKI (n = 2369)	p
Age (years old)	66.9 ± 9.7	63.2 ± 11.2	0.000
Diabetes (%)	28.1	23.0	0.000
Obesity (%)	33.4	27.2	0.000
EuroSCORE (median, range)	6 (0–21)	5 (0–22)	0.000
Baseline seric creatinine (mg/dL)	1.10 ± 0.45	1.12 ± 0.35	0.098
Peripheral vascular disease (%)	11.5	7.1	0.000
Preoperative use of ACEI (%)	41.6	38.8	0.078
Preoperative use of digitalics (%)	10.7	8.9	0.055
Preoperative use of diuretics (%)	23.0	16.8	0.000
Preoperative use of nitrates (%)	22.1	25	0.032
Acute myocardial infarction (%)	72.0	68.7	0.025
Preoperative LVEF (%)	52.5 ± 11.6	54.3 ± 24.3	0.008
Non-coronary surgery (%)	44.1	33.8	0.000
Emergency surgery (%)	1.9	0.8	0.030
On-pump surgery (%)	82.8	68.7	0.000
Pre-CBP Hct	36.6 ± 5.4	37.7 ± 8.3	0.000
Post-CBP Hct	25.9 ± 6.2	26.5 ± 6.2	0.048
Lowest intraoperative Hct	23.4 ± 3.8	24.1 ± 3.9	0.000
Intraoperative MAP (mmHg)	50.7 ± 15.1	51.2 ± 16.1	0.004
CBP time (min)	82.7 ± 49.0	63.0 ± 45.4	0.000
Aortic cross clamping time (min)	51.9 ± 34.1	40.3 ± 32.0	0.000
PO use of dopamine (%)	59.0	47.1	0.000
PO use of dobutamine (%)	73.2	65.5	0.000
PO use of epinephrine (%)	3.4	0.8	0.000
PO use of norepinephrine (%)	32.2	17.6	0.000

Variables that reach a random probability lower than 10% in the univariate analysis were included in the multivariate model for each strata of preoperative eGFR. Quantitative data are expressed as mean ± SD.

Abbreviations: eGFR = estimated glomerular filtration rate, AKI = acute kidney injury, ACEI = angiotensin-converting enzime inhibitor, LVEF = left ventricular excretion fraction, Hct = hematocrit, MAP = mean arterial pressure, CBP = cardiopulmonary bypass, PO = postoperative.

Table 3 Adjusted risk factors for AKI in patients with the baseline preoperative eGFR ≥60 mL/min/1.73 m²

Variable	OR	CI 95%	p
Advanced age	1.085	1.067–1.103	0.000
Diabetes	1.848	1.309–2.513	0.000
Diuretics	1.642	1.167–2.310	0.004
Emergency surgery	10.000	1.135–88.09	0.038
CBP time	1.007	1.003–1.011	0.002
Lowest Hct	0.936	0.900–0.972	0.001
IO MAP (mmHg)	1.011	1.005–1.018	0.001
PO-norepinephrine	1.885	1.350–2.632	0.000

Abbreviations: CBP = cardiopulmonary bypass, Hct = hematocrit, PO = postoperative, IO = intraoperative, MAP = mean arterial pressure.

Table 4 Adjusted risk factors for AKI in patients with the baseline preoperative eGFR <60 mL/min/1.73 m²

Variable	OR	CI 95%	p
Advanced age	1.027	1.002–1.052	0.034
EuroSCORE	1.087	1.010–1.170	0.027
Non-coronary surgery	1.780	1.258–2.517	0.001
Baseline SCr (mg/dL)	2.029	1.208–3.406	0.007
Lowest Hct	0.924	0.882–0.968	0.001
PO-dopamine	1.645	1.464–2.325	0.005
PO-epinephrine	3.330	1.189–9.320	0.022
PO-norepinephrine	1.532	1.048–2.238	0.028

Abbreviations: PVD = peripheral vascular disease, CBP = cardiopulmonary bypass, PO = postoperative.

DISCUSSION

There are a growing number of patients submitted to cardiac surgery worldwide as a result of the rise in the incidence of coronary artery disease and the inclusion of older and sicker patients for surgery. As an example, comparing data of our own center, mean age of patients submitted to surgery in a series published in 1997[20] was 58 ± 0.9 years, while in the current series the mean age of patients is 64.8 ± 10.8 years (p = 0.001).

Several risk factors for renal failure have been identified,[2],[3],[11],[12] and algorithms to accurately predict the occurrence of ARF in this setting have been designed with the aim of preventing this complication, or at least for advising patients on a more accurate decision-making at the time of surgery. In a large cohort of 42,773 patients, Chertow and colleagues[1] developed a risk stratification algorithm that, according to the authors, allows them to estimate the risk and design interventions directed to improve the outcome. Thakar et al.[14] also developed a clinical score to predict AKI after open-heart surgery. More recently, Rajendra Mehta and co-workers,[21] using the largest series of patients to date, designed a bedside tool for predicting the risk of postoperative AKI. Unfortunately, most of the variables associated with acute kidney injury are barely modifiable, if they are modifiable at all. CKD is one of the most important predictors of AKI, and it was demonstrated that even modest changes in serum creatinine are strongly associated with AKI and mortality after cardiac surgery.[7],[22]. However, there is a lack of information on the eventual differences in risk factors for postoperative AKI, according to preoperative level of renal function.

There was no difference in the incidence of AKI between groups according to the adopted definition. However, the present study demonstrated that baseline S_Cr was associated with the onset of AKI only in patients with a C_Cr lower than 60 mL/min/1.73 m² BSA (CKD stages 3 and 4). This finding is in accordance with other studies that demonstrated a higher risk for postoperative AKI when C_Cr is lower than 60 mL/min/1.73 m² BSA.[1],[14] In the aforementioned study from Chertow et al.,[1] a higher risk for AKI was also found in patients with a C_Cr <60 mL/min/1.73 m² BSA, but it should be mentioned that a more restrictive definition of AKI (need of postoperative dialysis) was used in this study, as well as in the study from Thakar and co-workers.[14] Moreover, other studies demonstrated that patients with CKD and C_Cr <60 mL/min/1.73 m² BSA are also prone for non-renal complications, mainly cardiovascular events.[23–25] This fact is very significant, considering that 1449 out of 4118 of the patients in present series were in these stages of CKD. Interestingly, although the incidence of AKI in both groups was similar, variables associated with the onset of AKI were different between the two groups.

In addition to baseline renal reserve, other variables, such as need of vasoactive drugs, hemodilution, age, EuroSCORE and the type of surgery, were also associated with AKI in this group of patients.

In patients of group 1 (CKD stages 1 and 2), AKI was independently associated with some underlying conditions (older age, diabetes), preoperative administration of diuretics, urgency surgery, CPB time, hypotension during CBP, hemodilution, and the need of inotropic and vasoactive drugs after surgery, which is a marker of low cardiac output.

The preoperative use of diuretics was strongly associated with AKI in group 1 patients and increased the odds for renal dysfunction by 64%. A detrimental effect of the use of diuretics and mainly furosemide during cardiac surgery has been demonstrated by us and others.[16],[26] The present study also demonstrated the adverse effect of diuretics when they are administrated before surgery, in accordance with data from Chertow et al.[1] A potential nephrotoxic effect of the drug associated with volume depletion and/or cardiac failure could be an explanation for this finding.

On-pump lowest hematocrit was associated with an increase in the risk of AKI in both groups of patients. In patients of group 1, every 1% descent in on-pump nadir hematocrit increases the risk of AKI by 6.4%. In patients of group 2, the risk increases by 7.6%. Hemodilution during CBP has been identified recently as a risk factor for AKI.[27–29] In a comprehensive and elegant study, Habib and colleagues[23] established a detrimental threshold of hemodilution when the nadir hematocrit was <24%. They also demonstrated that prolonged CBP, intraoperative transfusion of red cells, and preoperative impaired renal function exacerbated the probability of AKI.

The mortality rate of the entire population is in the usual figures.

CONCLUSIONS

According to the present study, the risk factor for AKI varies according to the estimated preoperative renal function. Only in patients with CKD stages 3–4, the degree of preoperative renal dysfunction and the need of vasoactive drugs were strong predictors of AKI. This fact highlights the weight of the reduction of renal reserve as a risk factor in this group of patients with advanced stages of chronic kidney disease. In addition to other well-known risk factors for AKI in this setting, the use of diuretics in patients of group 1 and the level of hemodilution during CBP in both groups appear as very important risk factors, as they are potentially modifiable.

ACKNOWLEDGMENTS

The authors thank the medical staff of the National Institute of Cardiac Surgery and the Department of Critical Care Medicine for contributions to the database.

APPENDIX: EUROSCORE

Table

		Score
Patient-related factors
Age	Per 5 years or part thereof over 60 years	1
Sex	Female	1
Chronic pulmonary disease	Long-term use of bronchodilators or steroids for lung disease	1
Extracardiac arteriopathy	Any one or more of the following: claudication, carotid occlusion or >50% stenosis, previous or planned intervention on the abdominal aorta, limb arteries or carotids	2
Neurological dysfunction disease	Severely affecting ambulation or day-to-day functioning	2
Previous cardiac surgery	Requiring opening of the pericardium	3
Serum creatinine	>200m micromol/L preoperatively	2
Active endocarditis	Patient still under antibiotic treatment for endocarditis at the time of surgery	3
Critical preoperative state	Any one or more of the following: ventricular tachycardia or fibrillation or aborted sudden death, preoperative cardiac massage, preoperative ventilation before arrival in the anaesthetic room, preoperative inotropic support, intraaortic balloon counterpulsation, or preoperative acute renal failure (anuria or oliguria<10 mL/hour)	3
Cardiac-related factors
Unstable angina	Rest angina requiring IV nitrates until arrival in the anaesthetic room	2
LV dysfunction	Moderate or LVEF 30–50%	1
	Poor or LVEF <30	3
Recent myocardial infarct	<90 days	2
Pulmonary hypertension	Systolic PA pressure > 60 mmHg	2
Operation-related factors
Emergency	Carried out on referral before the beginning of the next working day	2
Other than isolated CABG	Major cardiac procedure other than or in addition to CABG	2
Surgery on thoracic aorta	for disorder of ascending, arch, or descending aorta	3
Postinfarct septal rupture		4