Prevalence and risk factors for acute renal failure in the postoperative of coronary artery bypass grafting

Objective: To determine the prevalence, risk factors, and the clinical outcome of patients undergone coronary artery bypass grafting who progressed with Acute Renal Failure (ARF). Methods: A prospective study was performed from data of 186 patients undergone surgery from January 2003 through June 2006. The stored data were analyzed using the software STATA 9.0. Results: The prevalence of ARF was of 30.6% (57/186). In 7.0% (4/57) dialysis therapy was needed. The mean age of patients with and without ARF progression was 62.8 % (±9.4) years and 61.3% (±8.8) years respectively (P=NS). CPB time >115 min (p= 0.011) and cross-clamp time >85 min (p=0.044) were related to ARF by the univariate analysis. The need for intra-aortic balloon (P= 0.049), mechanical ventilation >24h (P = 0.006), Intensive Care Unit (ICU) stay > three days (P< 0.0001), bradycardia (P= 0.002), hypotension (P= 0.045), arrhythmia (P=0.005) and inotropic infusion (P= 0.0001) were higher in the ARF group. Only the ICU stay longer > 3 days showed statistical correlation with ARF by the multivariate analysis (P=0.018). The mortality rate with and without ARF was 8.8% (five cases) and 0.8% (one case) respectively (P=0.016), but it reached 50% (2/4) in dialytic patients. Conclusion: ARF was a frequent and severe postoperative complication associated with higher mortality and longer ICU stay, which presented as risk factors: longer CPB and cross-clamp times, mechanical ventilation > 24h and hemodynamic instability. Descriptors: Myocardial revascularization. Risk factors. Renal insufficiency, acute. Resumo Objetivo: Determinar a prevalência, fatores predisponentes e o desfecho clínico dos pacientes submetidos a cirurgia de revascularização do miocárdio que apresentaram insuficiência renal aguda (IRA). Métodos: Estudo do tipo coorte prospectivo, a partir dos prontuários de 186 indivíduos submetidos a cirurgia, no período de janeiro de 2003 a junho de 2006. As informações foram inseridas em um banco de dados e analisadas pelo software STATA 9.0. Resultados: A prevalência de IRA foi 30,6% (57/186), sendo que 7% (4/57) necessitaram de diálise. A idade média dos This study was carried out at University Hospital at Federal University of Maranhão. São Luis, MA, Brazil. Correspondence address: Dyego José de Araújo Brito Rua Matos Carvalho, 28. Olho d’Água. São Luis,MA, Brazil. CEP: 65065270. E-mail: djabrito@hotmail.com Support: Foundation for Research and Scientific and Technology Development Support of the State of Maranhão (FAPEMA) Article received on October 2nd, 2008 Article accepted on June 16th, 2009


INTRODUCTION
Acute renal failure (ARF) is one of the most serious and frequent complications observed in the postoperative period of heart surgery, with prevalence ranging from 1% in patients without prior renal disorder, and 16% to 20% among those with a history of some renal dysfunction [1]. The need for dialysis is observed in up to 30.6% of cases [2,3].
Multifactorial events (pre-, intra-and postoperative) may determine the onset of ARF after heart surgery, even in case of no cardiovascular dysfunction or nephrotoxicity in these patients [4]. In our country, Palomba et al. [5] found that the sum of preoperative factors (age > 65 years, creatinine > 1.2 mg/dL, capillary blood glucose > 140mg/ dL, heart failure of functional class > II and combined surgery), intraoperative (duration of CPB > 120 minutes) and postoperative (low cardiac output, central venous pressure > 14cmH 2 O) may promote the occurrence of ARF.
In patients undergoing heart surgery, it is noted a common volume depletion (because of bleeding during and after surgery) and loss of liquid for the third space (resulting from systemic inflammatory response), which compromises renal perfusion reducing the glomerular filtration rate (GFR) [4]. In the study by De Moraes Lobo et al. [6] it was observed that patients with reduced glomerular filtration rate (GFR) in the postoperative, require a greater infusion of volume and use of sodium nitroprusside and presented lower fractional excretion of sodium (FENa + ), suggesting that changes in renal function observed resulted from an appropriate renal response to low effective blood volume.
The presence of renal failure is an important risk factor for the increase in the number of deaths in the postoperative period of heart surgery [7]. Evidence suggests that small increases in serum creatinine in the postoperative period are associated with significant effects on the risk of death [8]. In general, it seems that mortality among patients undergoing heart surgery may reach 8% of cases, however, when presentation of ARF is established in the postoperative period of these patients, there is an exponential increase in risk of death, surpassing 60% of the cases [9].
Thus, it is necessary to perform the identification of factors associated with increased risk of progression to ARF in the postoperative period, so that preventive measures are taken, reducing the prevalence of the event, through the establishment of clinical (pre-and postoperative) and therapeutic approaches more effective, contributing to reduce morbidity and mortality among patients. This study aimed to determine the prevalence and predisposing factors for acute renal failure among patients undergoing coronary artery bypass grafting.

Type of study
We performed a prospective study with data collected retrospectively from the medical records of 186 patients who had undergone a retrospective cohort surgery from data collected from medical records of patients who had undergone CABG surgery in our institution.
Data of postoperative monitoring • Mean Arterial Pressure: obtained through the arithmetic mean of the highest and lowest blood pressure measurements recorded each day.
• Diuresis: presented in the form of mL/kg/day, by defining oliguria when the value is less than 1 mL/kg/day.

Criteria for definition of ARF
We considered patients with acute renal failure postoperatively, those patients who presented [4]: 1) 50% increase in creatinine in the postoperative period, with baseline > 1.3 mg/dL preoperatively; 2) increase of 0.5 mg/dL in creatinine among those with preoperative value d" 1.3 mg/dL, or 3) need for dialysis postoperatively.
To calculate the prevalence of dialytic ARF, only patients who had undergone dialysis during the ICU stay were considered, excluding those who required dialysis after discharge from this unit.

Ethical aspects
The study was approved by the Research Ethics Committee of the Institution (Protocol No. 33104-0372/2007) as the Resolution 196/96 of the National Health Council for research involving humans.

Statistical analysis
First, descriptive analysis of the variables studied was performed. Qualitative variables are presented by frequencies and percentages, and quantitative variables as mean and standard deviation.
For comparison of quantitative variables we used the bilateral student's t test for independent samples and for the qualitative variables we used the Chi-square, where applicable, and Fisher's exact test. Normality of quantitative variables was analyzed by Shapiro Wilk test.
To identify factors associated with acute renal failure postoperatively, we performed univariate analysis using the model of logistic regression.
To establish the overall association of the factors

Data collection
From the list of patients who had undergone surgery, records from patients who met the criteria for inclusion in the study were collected. Subsequently, we collected information on the identification data, laboratory tests, clinical presentation, monitoring and treatment applied.
The data were related to the preoperative period and ICU postoperative follow-up, characterized in six times, namely: • T0: corresponded to the preoperative period, before induction of anesthesia; • T1: from the admission of the patient to the ICU up to 6h59min of the next morning; • T2: starting at 7am on the 1st postoperative day up to 6h59min of the next day or time of patient's death (if it had occurred 24 hours after the end of the procedure); • T3: starting at 7am on the 2nd postoperative day up to 6h59min of the next day or time of discharge from ICU or patient's death; • T4: starting at 7 am on the 3rd postoperative day up to 6h59min of the next day or time of discharge from ICU or patient's death; • Discharge from ICU

Variables
We approached the following variables in this study: Clinical data • Age: in years, presented as mean ± standard deviation (SD) and age; • Gender: male and female; • Weight: in kg to calculate estimated Glomerular Filtration Rate (GFR) and daily diuresis; • Personal comorbidities, as described in the medical records: diabetes mellitus, systemic arterial hypertension, GFR between 15-59 mL/min, previous heart surgery, acute myocardial infarction or angina pectoris, stroke, among others.; • Underlying pathologies.

Complementary exams
• Serum creatinine (Cr s ): assessing prior renal function and daily postoperative follow-up up to discharge from the ICU. Reference value: 0.5 to 1.3 mg/dL.
• Urea: preoperative values and daily postoperative follow-up up to T4. Reference value: 30 to 40 mg/dL.
• Creatinine clearance (CrCl) was estimated by the Cockcroft-Gault method [10] to classify preoperative renal function, considering patients with moderate or severe renal failure those with GFR between 15-59mL/min [11].
Surgical data • Number of surgical procedures performed, length of studied with ARF, we used the model of logistic regression. Variables with a p-value less than 0.20, in the univariate analysis, were considered in the model of logistic regression analysis. The selection of variables was performed using the stepwise method by elimination. Only variables with pvalue less than 0.05 remained in the final model. We also estimated the odds ratios (OR) and their respective confidence intervals of 95%. The data were analyzed using 9.0 STATA statistical software.

RESULTS
The mean age of patients was 61.7 (± 8.9) years, and 61.8% were aged less than 60 years. Most of them were males (109 -58.6%).
The mean age of patients who developed ARF was 62.8 (± 9.4) years, comparing with 61.3 (± 8.8) years in the group without ARF, with P-value = 0.293.
Univariate analysis of preoperative variables studied (age, gender, serum creatinine, history of diabetes, hypertension, myocardial infarction/angina and heart surgery), by comparing the groups of patients with and without ARF, there was no association of these informations collected with the event of renal failure in the postoperative period.
The time of cardiopulmonary bypass and aortic clamping were risk factors for postoperative complications associated with ARF, as shown in Table 1.
The evolution of serum creatinine in the postoperative period showed that values were higher in the group that developed ARF and statistically significant difference in the T1 to T4 interval was noted, when compared to the group with no ARF ( Table 2).
In 149 medical records assessed there was information on the value of serum creatinine in the patient's discharge from the ICU, and 1.61 (±1.10) mg/dL on average for the ARF group and 1.05 (±0.41) mg/dL in the group without ARF, with P-value = 0.0004.
Regarding postoperative monitoring, we found that the average of the mean arterial pressure in the T1-T4 interval was almost equal in both groups, being 93.5 (± 10.9) mmHg in the ARF group and 93.9 (±10.2) mmHg in patients without ARF, with P-value=0.792. The average diuresis was also similar between patients with and without ARF, respectively, as follows, 1.98 (± 0.97) mL/kg/h and 2.06 (±0.81) mL/kg/h and P-value=0.561.
Qualitative variables related to postoperative monitoring are described in Table 3. The main complications presented in the postoperative period are described in Table 4. We highlight the fact that hypotension, bradycardia and arrhythmia have presented any statistical relationship with the occurrence of Acute Renal Failure.
After univariate analysis, a multivariate logistic model  Table 5.

DISCUSSION
Studies such as those developed by Yeboah et al. [12] and Abel et al. [13], approaching the relationship between heart surgery and ARF have been published for over 30 years. Interest in this subject is due to the fact that the ARF is a frequent complication related to surgical procedures, especially cardiovascular surgeries, and is related to increased morbidity and mortality in the postoperative [14]. The prevalence of ARF found in our study was 30.6%, similar to that described by Kochi et al. [15], with 34.0%, in a study with 150 patients. In our case, we collected information from medical records of 186 individuals. Other studies, such as by Tuttle et al. [16] and Yehia et al. [3] presented higher prevalence than ours (41.3% and 42.0%, respectively). In the study by Santos et al. [2], Chertow et al. [9], Stallwood et al. [17] and Conlon et al. [18], the rates of ARF were 16.1%, 7.9%, 2.4% and 1.1% respectively, well below those described in our group. Such differences can be explained by the different criteria for the definition of postoperative ARF used by the authors.
The use of dialysis is one of the criteria used to define ARF in the postoperative of heart surgery [2,18,19]. The need for dialysis is found in 0.7% to 30.6% of patients who develop acute renal failure in different publications [2,3]. In rare cases, as in the study published by Tuttle et al. [16], no patient developed ARF with the need for dialysis. Our study identified 7.0% of dialytic ARF, and risk factors for the event were not observed.
Most studies indicate the presence of previous impaired renal function as one of the factors associated with acute renal failure in the postoperative period of heart surgery [18]. The study by Walter et al. [20] pointed out the creatinine clearance estimated by Cockcroft-Gault method, as a more reliable indicator when compared to the value of serum creatinine, to determine the risk of unfavorable outcomes in the postoperative of heart surgery. In our study, as well as developed by Lobo [21], there was no relationship between preoperative creatinine clearance and poor outcome of renal function postoperatively.
Other preoperative variables such as advanced age, diabetes mellitus, previous heart surgery and arterial hypertension, which classically are described as risk factors for ARF in other publications, were not correlated with worsening of renal function among patients approached in this study perhaps by the number of patients studied herein was lower than that of other studies [2,17,18,22].
Among the surgical variables investigated, two were statistically related to the risk of development of ARF. The mean time of cardiopulmonary bypass (CPB) and aortic clamping were higher in the group with ARF (P-value <0.05). In the study by Regner et al. [19], which also assessed these two parameters, only the mean of the CPB showed significant relationship with the abrupt drop of renal function, unlike the study of Tuttle et al. [16], on which the two variables were statistically significant for evolution ARF in the postoperative period, with P-value <0.0001.
The changes caused by the use of cardiopulmonary bypass include contact with nonendothelial surfaces, the exclusion of pulmonary flow, the presence of continuous minimally pulsatile or nonpulsatile flow and the onset of endocrine-metabolic response is proportional to a major surgery, all this culminating in an inflammatory response not susceptible to external control [23]. The positive impact of the absence of CPB on renal function in postoperative of heart surgery can be observed in the study by Milani et al. [24], in which only 0.2% of patients developed ARF, a value less than that found in studies in which CPB was used [25,26]. In another study, developed by Lima et al. [27], assessing octogenarian patients undergoing CABG surgery, the presence of ARF was 19.2% in the on-pump group and none in patients without CPB.
With respect to the follow-up it was observed that in the group with ARF the mean serum creatinine was higher in the range T1 to T4 and at the time of discharge. In the study by Kochi et al. [15] the assessment of the mean value of serum creatinine, 24 and 48 hours after surgery, also showed higher values in the group with ARF (P-value <0.0001). According to the study by Ryckwaert et al. [28], minimum changes in the values of creatinine in the postoperative period are not rare events and have a significant impact on follow-up of patients with decreased survival.
Several studies indicate a relationship between ARF and length of hospital stay. This study assessed only the period of ICU stay, and staying for more than three days -that was the only variable that, after multivariate analysis, showed an association with ARF in the postoperative period. The study of Ryckwaert et al. [28] also pointed out the relationship between prolonged ICU stay and renal failure postoperatively.
The mean time of mechanical ventilation over 24 hours was higher in the ARF when compared to those without ARF (31.6% vs 13.2%, respectively, with p-value = 0, 006). Our result was consistent with that found by the study of Anderson et al. [29], in which the need for mechanical ventilation for more than 24 hours was higher among patients with poor evolution of renal function. Vieira et al. [30] assessing 140 patients admitted to the ICU at an oncology hospital, observed that the occurrence of ARF determined not only a longer duration on mechanical ventilation, but also it took a longer period of time to promote the disconnection from mechanical ventilation.
Two parameters that, according to Santos et al. [2], can indirectly assessing cardiac performance, were associated in our study to the development of ARF in the postoperative period. The need for intra-aortic balloon and inotropic drugs (dopamine or dobutamine), from the period T3, were higher among patients with ARF.
Among the postoperative complications investigated: bradycardia, hypotension and arrhythmia, such complications were related to worsening renal function of patients. The damaged cardiac function, creating a situation of low cardiac output, and therefore a presentation of hypoperfusion, may precipitate pre-renal ARF, which if not promptly corrected, can evolve into the intrinsic ARF and even cortical necrosis, a situation that results in an irreversible presentation of loss of renal function [14].
Acute renal failure is considered an independent risk factor for death in the postoperative period of heart surgery. The mortality reported in the literature ranges from 14.5% to 63.7% [2,9,31,32]. In our study, mortality among patients with ARF was only 8.8%, lower than that found by other authors. This can be explained, in part, because the surgical procedures included in this study were all elected, with patients in better clinical condition before surgery. CONCLUSIONS 1 -The cardiopulmonary bypass time (>115min), aortic clamping time (>85 min), need for intra-aortic balloon, mechanical ventilation for more than 24 hours, bradycardia, hypotension, arrhythmia and the need for inotropic drugs were considered risk factors for ARF; 2 -The presence of renal failure was correlated with a worse postoperative outcome, characterized by longer stay in Intensive Care Unit and higher mortality.