Risk Factors and Outcome of Acute Kidney Injury after Congenital Heart Surgery: A Prospective Observational Study.

BACKGROUNDS AND AIMS: Acute kidney injury (AKI) is a frequent event after congenital heart surgery with increased mortality and morbidity. We investigated frequency, risk factors, and associated morbidity and mortality of AKI after pediatric cardiac surgery at a single institution.
METHODS: Children undergoing congenital heart surgery from March 2013 to February 2016 were assessed for development of AKI based on modified pediatric Risk, Injury, Failure, Loss, and End-stage renal disease criteria. They were also investigated for predictive risk factors, associated mortality, and morbidity including duration of mechanical ventilation, Intensive Care Unit (ICU), and hospital length of stay.
RESULTS: Five hundred and nineteen patients were recruited during the study period including 259 (49.9%) males and 260 (50.1%) females. AKI was seen in 150 (28.9%) patients including 101 (67.3%), 42 (28%), and 7 (4.7%) cases with risk, injury, and failure stages, respectively. Patients with AKI had longer ventilation time (P = 0.002), ICU (P = 0.05), and hospital (P = 0.56) stay. Mortality was seen in 31 (2.7%) and 44 (11.9%) patients with and without AKI, respectively (P = 0.01). After multivariable logistic regression, there was an association between AKI and preoperative abnormal levels of creatinine (adjusted odds ratio [aOR] = 0.47, 95% confidence interval [CI] 0.22-1.01; P = 0.05), presence of cyanotic heart disease (aOR = 1.97, 95% CI = 1.15-3.2; P = 0.01), duration of surgery (aOR = 1.05/10 min, 95% CI = 1.01-1.08; P = 0.007), and elevated lactate level (aOR = 1.14, 95% CI = 1.03-1.3; P = 0.01).
CONCLUSION: The presence of cyanotic heart disease, duration of surgery, elevated postoperative lactate level, and likely preoperative creatinine level were independent risk factors for the development of AKI after congenital heart surgery.

INTRODUCTION

Acute kidney injury (AKI) is a significant and common complication after pediatric cardiac surgery. The incidence has been reported from 9.6% to 64.6%[1234] based on the definition of AKI and age group. It is associated with increased morbidity and mortality.[5678] It also increases duration of mechanical ventilation,[19] Intensive Care Unit (ICU) and hospital length of stay,[710] and resource utilization.[5] The increased mortality depends on severity of AKI. Renal replacement therapy (RRT) is needed in 1.6%–7.7% after cardiac surgery that further increases mortality in these patients.[11121314]

A modified version of pediatric Risk, Injury, Failure and Loss, and End-Stage (p-RIFLE) has been proposed for standardization of definition of AKI that uses either a decrease in glomerular filtration rate (GFR) or urine output.

A number of perioperative risk factors have been suggested as predictors of AKI associated with cardiac surgery. These include younger age, prolonged pump time,[110] pump failure, lower weight, sepsis,[310] need for extracorporeal membrane oxygenation (ECMO),[7] and prolonged ventilator requirement.[3]

The objective of this study was to investigate the development of AKI after congenital heart surgery at a single university-affiliated institution. The primary outcome was to determine the incidence of AKI. The secondary outcomes included investigation of risk factors and outcome including duration of mechanical ventilation, ICU and hospital length of stay, and inhospital mortality.

METHODS

In a prospective observational study, we enrolled consecutive children undergoing congenital heart surgery for either palliation or correction with or without cardiopulmonary bypass (CPB) from March 2013 to February 2016 at a university-affiliated Hospital, Mashhad, Iran. Inclusion criteria were patients under 18 years admitted to cardiac ICU after any kind of cardiac surgery. Patients were excluded if they had preexisting end-stage renal disease, a history of renal transplant, or if no documented baseline serum creatinine was available.

Preoperative variables were age at the time of surgery, weight, preoperative hemoglobin, white blood cell (WBC), urea, creatinine, and pulmonary hypertension.

Surgical characteristics including type of surgery, use of CPB, duration of surgery, pump time, cross-clamp time, and intraoperative events including excessive bleeding, life-threatening arrhythmia, and intraoperative transfusion were recorded.

The patients were assessed postoperatively for highest lactate level and inotrope score ([dopamine + dobutamine] + [10] milrinone + [100] [epinephrine + norepinephrine + phenylephrine], μg/kg/min), WBC on postoperative day one, duration of mechanical ventilation more than 24 h, ICU and hospital length of stay, low cardiac output, need for RRT, intraoperative events, and inhospital death. RRT was used in case of refractory acid-base and electrolyte disorders, signs of hypervolemia, or uremic brain (defined as decreased level of consciousness assuming to be related to hyperuremia). RACHS-1 score was used to categorize surgeries according to their complexity.

AKI was defined according to p-RIFLE criteria that are based on postoperative fall in GFR compared to baseline as R:risk (a decrease in GFR by 25%), I:injury (a decrease in GFR by 50%), F: failure (a decrease in GFR by 75% or <35 mL/min/1.73 m2), and L:loss (loss of renal function >4 weeks). GFR was calculated using the following equation: Estimated glomerular filtration rate (mL/min/1.73 m2) = k × height (cm)/plasma creatinine (mg/dL) where k = 0.55; if <1 year, 0.45; if male and older than 13 years, 0.7.

Urine output was not used as a criterion for renal failure in our study. Recovery was defined as return of GFR to a level with <25% drop compared to preoperative level.

Low output syndrome was defined as hepatomegaly, oliguria, tachycardia, a fall in systolic blood pressure under the age-related normal value, a base excess lower than _4 mmol/L, or lactate level above 2 mmol/L in 2 consecutive arterial blood samples.

The patients were under standard monitoring and meticulous attention by critical care team including an intensivist, a critical care fellow, and anesthesiology residents to maintain cardiopulmonary, cerebral, and renal function. Patients were weaned from mechanical ventilation and extubated according to a standardized protocol.

Statistical analysis

Means and standard deviation were used for normal distribution variables and median and interquartile range for otherwise. Frequencies and percentages were used for categorical variables. The Student's t-test or the Mann–Whitney test was used for comparison of continuous variables between patients with and without AKI. Chi-square test or Fisher's exact test was used for categorical variables.

Binary logistic regression model was constructed for assessment of independent effect of each variable on the prediction of occurrence of AKI. Variables with P < 0.1 in univariable analysis consisting ventilation time, duration of ICU stay, preoperative creatinine level, history of cyanotic heart disease, intraoperative bleeding, duration of surgery, level of albumin, and lactate level were included in multivariable analysis. Model goodness of fit was evaluated by the Hosmer–Lemeshow test. Receiver operating characteristic curve was used for prediction of sensitivity and specificity of cutoff point of significant variables. Statistical significance was considered as P < 0.05. The statistical analysis was performed using SPSS® version 16 (IBM SPSS, Chicago, IL, USA).

RESULTS

Five hundred and nineteen patients were included during the study period. There were 259 (49.9%) males and 260 (50.1%) females with mean age of 40.3 ± 43.7 and 40.07 ± 45.9 and median 24 and 22 months for males and females, respectively. Demographic data, types of surgery, and surgical characteristics of patients are presented in Tables 1–3, respectively.

Table 1

Demographic data and patients’ characteristics

Table 3

Surgical characteristics of patients

AKI was seen in 150 (28.9%) patients. Of these, 101 (67.3%), 42 (28%), and 7 (4.7%) cases developed risk, injury, and failure, respectively. Three patients needed RRT. AKI occurred in 72 (50.3%), 44 (30.8%), and 12 (8.4%) patients on day 1, 2, and 3, respectively. Recovery was seen in 47 (63.5%), 13 (17.6%), and 7 (9.5%) patients on the 1st, 2nd, and 3rd day, respectively. The incidence of AKI based on surgical complexity using RACHS-1 score is presented in Table 4.

Table 4

Frequency of AKI according to RACHS-1 score

Thirty-one (2.7%) and forty-four (11.9%) patients with and without AKI died (P = 0.01). Mortality was seen in 13 (13%), 12 (28.6%), and 6 (85.7%) cases in those with risk, injury, and failure, respectively (P < 0.001). All three patients who needed dialysis died.

Ventilation time was 1164 ± 2402 min in the AKI group and 496 ± 1300 min in the non-AKI group (P = 0.002). Duration of ICU stay was 4.45 and 3.39 days in patients with and without AKI, respectively (P = 0.05). Hospital length of stay was 8.82 and 8.04 days in patients with and without AKI, respectively (P = 0.56). Association of RACHS-1 score and AKI is displayed in Table 5.

Table 5

Univariate analysis of RACHS-1 score on predicting AKI

Using multivariable logistic regression, we found the following variables as independent risk factor for the development of AKI.

Preoperative abnormal creatinine level (adjusted odds ratio [aOR] = 0.47, P = 0.05) and cyanotic heart disease (aOR = 1.97, P = 0.01) were independent predictors of AKI occurrence [Table 6]. There was an association between intraoperative bleeding and duration of surgery and AKI. However, after multivariable logistic regression, only duration of surgery aOR = 1.05/10 min (P = 0.007) was found to be an independent predictor for AKI [Table 6]. Of postoperative variables, only a high lactate level (aOR = 1.13, P = 0.01) was an independent predictor for development of AKI [Table 6].

Table 6

Perioperative risk factors associated with AKI after congenital heart surgery

Operation time longer than 137 min had a sensitivity of 82% and specificity of 76% for prediction of AKI (area under curve [AUC] = 59%, P = 0.02). Lactate level of higher than 1.35 had a sensitivity of 81% and specificity of 74% for prediction of AKI (AUC = 59%, P = 0.02) [Figure 1]. Young age, weight <10 kg, preoperative albumin level, preoperative anemia, pulmonary hypertension, use of high dose inotropes, and reoperation were not significantly associated with AKI.

Figure 1

Receiver operating characteristic curve for prediction of sensitivity and specificity of cutoff point for creatinine

DISCUSSION

Our study revealed that AKI was seen in 28.9% of patients following congenital heart surgery and was associated with increased morbidity and mortality. We also found that preoperative creatinine level, presence of cyanotic heart disease, duration of surgery, and postoperative lactate level were independent risk factors for AKI.

A number of mechanisms have been proposed as the etiology for development of AKI after congenital heart surgery including altered renal blood flow, hypoperfusion, inflammation and loss of pulsatile flow, ischemia, decreased autoregulation, and nephrotoxic drugs including antibiotics.

Similar to Cardoso et al.,[9] we found that higher postoperative lactate level, a sign of hypoperfusion, is an independent risk factor for AKI confirming hypoperfusion as a proposed mechanism of AKI.

In line with findings of other reports,[19] patients who developed AKI in our study had a higher baseline preoperative creatinine level. However, the association was not very strong and is not clinically significant (aOR = 0.47, P = 0.05). This discrepancy can be explained by different age group population and use of CPB in these studies.

Unlike other studies,[110] we did not find an association between younger age than 12 months and weight <10 kg and development of AKI. Aydin et al.[1] suggested intrinsic immaturity of neonatal renal tubules and its inability to adapt to inflammatory and ischemic changes induced by CPB as the possible mechanism. In contrast, our patients’population was not limited to infancy and use of CPB.

Similar to other studies, we found that AKI is associated with increased mortality and morbidity.[11516] Even those with mild drop in GFR (risk score) had higher mortality than those without AKI. The mortality increased to 18.8%, 50%, and 85.7% in those with risk, injury, and failure scores, respectively, indicating an increase in mortality as the severity of AKI worsens.

Our finding is in keeping with other studies showing that AKI increases duration of mechanical ventilation, ICU, and hospital length of stay.[91718] Nevertheless, the difference for ICU and hospital length of stay was not statistically significant. It seems that factors other than AKI might affect these morbidities.

Unlike Sethi et al.[10] and Aydin et al.,[1] we did not find a strong association between duration of CPB and development of AKI. This can be explained by shorter pump time in our series compared to these reports.

Our study had a few limitations. First, we did not know the baseline preoperative blood pressure of our patients and did not investigate the changes in comparison to intraoperative ones. This might affect the incidence of AKI as is reported in adults.[1920] However, to our best of knowledge, there is no such report in pediatric population. Second, this was a single center study with limited resources that can affect patients’management and hence the findings cannot be generalized. Third, there was a vast range of different cardiac surgeries in our survey that makes it difficult to correlate type of surgery to development of AKI because of small sample size. Fifth, we did not use urine output as a definition for AKI since many patients received diuretics and this might affect accuracy of urine output as a variable for the definition of AKI.

CONCLUSION

We found a significant association between development of AKI and presence of cyanotic heart disease, duration of surgery, elevated postoperative lactate level, and likely preoperative creatinine level after congenital heart surgery.

Financial support and sponsorship

This study was supported financially by deputy research of Mashhad University of Medical Sciences.

Conflicts of interest

There are no conflicts of interest.

Table 2

Type of surgery