Acute kidney injury in hypoplastic left heart syndrome patients following the comprehensive stage two palliation

Abstract

Background:

An alternative surgical approach for hypoplastic left heart syndrome is the Hybrid pathway, which delays the risk of acute kidney injury outside of the newborn period. We sought to determine the incidence, and associated morbidity, of acute kidney injury after the comprehensive stage 2 and the cumulative incidence after the first two operations in the Hybrid pathway.

Design:

A single centre, retrospective study was conducted of hypoplastic left heart patients completing the second-stage palliation in the Hybrid pathway from 2009 to 2018. Acute kidney injury was defined utilising Kidney Diseases Improving Global Outcomes criteria. Perioperative and post-operative characteristics were analysed.

Results:

Sixty-one patients were included in the study cohort. The incidence of acute kidney injury was 63.9%, with 36.1% developing severe injury. Cumulatively after the Hybrid Stage 1 and comprehensive stage 2 procedures, 69% developed acute kidney injury with 36% developing severe injury. The presence of post-operative acute kidney injury was not associated with an increase in 30-day mortality (acute kidney injury 7.7% versus none 9.1%; p = > 0.9). There was a significantly longer median duration of intubation among those with acute kidney injury (acute kidney injury 32 (8, 155) hours vs. no injury 9 (0, 94) hours; p = 0.018).

Conclusions:

Acute kidney injury after the comprehensive stage two procedure is common and accounts for most of the kidney injury in the first two operations of the Hybrid pathway. No difference in mortality was detected between those with acute kidney injury and those without, although there may be an increase in morbidity.

Acute kidney injury is common after congenital heart surgery, with a reported incidence of 30–60% in operations requiring cardiopulmonary bypass. ^1–6 The presence of post-operative kidney injury is associated with an increased risk of morbidity and mortality. ^1,2,5,7,8 The literature is sparse in examining the incidence of kidney injury and its consequences in infants with single ventricle physiology. ^9–14

Patients with hypoplastic left heart syndrome require multiple interventions in the first year of life. The majority of centres utilise the Norwood procedure as their preferred first palliation, which requires cardiopulmonary bypass and period of circulatory arrest or regional cerebral perfusion. ^15,16 The hybrid pathway is an alternative to the traditional Norwood approach and entails the Hybrid procedure (placement of bilateral pulmonary artery bands by a surgeon and stent placement across the ductus arteriosus followed by a transcatheter atrial septostomy) followed by the comprehensive stage 2 (Comp 2) palliation. ¹⁷ The Comp 2 requires a substantial length of cardiopulmonary bypass time to remove the ductal stent, remove the branch pulmonary artery bands, perform an aorto-pulmonary amalgamation, potentially an atrial septectomy, and the bidirectional Glenn. ¹⁸

Acute kidney injury following palliation for single ventricle physiology is common; with reported post-operative acute kidney injury incidences of 24–75% following the Norwood procedure and 13–29% following the Hybrid Stage 1. ^9–14 However, the Hybrid palliative strategy may result in a greater incidence of acute kidney injury at the time of the Comp 2 palliation, thus making the cumulative acute kidney injury risk negligible between the differing palliative options. No studies have investigated kidney injury after the Comp 2 or analysed the cumulative risk of kidney injury after the first two stages of the Hybrid pathway.

In this study, we investigated the incidence and severity of acute kidney injury attributable to the Comp 2 procedure and the cumulative incidence of acute kidney injury after Hybrid + Comp 2 pathway. We hypothesise that patients undergoing the Comp 2 procedure have a high incidence of post-operative acute kidney injury and that those who develop post-operative acute kidney injury have a higher incidence of morbidity and mortality than those without.

Materials and methods

This is a single centre, retrospective review of patients born with hypoplastic left heart syndrome who underwent the Comp 2 procedure between 2009 and 2018. Patients with a significant abnormal kidney ultrasound, defined as an abnormality likely to impair kidney function, those with incomplete creatinine data, or with incomplete secondary outcomes were excluded. We obtained waiver of informed consent by the Institutional Review Board at Nationwide Children’s Hospital.

Demographic variables were obtained including details of the patient’s hypoplastic left heart syndrome variant and documentation of kidney ultrasound findings. We utilised previously published data from our group for the presence of acute kidney injury after the Hybrid palliation. ¹² Pre-procedural characteristics such as age, height, and weight at time of the Comp 2 were recorded.

Acute kidney injury was defined utilising the Kidney Diseases Improving Global Outcomes criteria; with stage 1 being an increase in creatinine of ≥50% from baseline or increase in creatinine of 0.3 mg/dL, stage 2 an increase in creatinine of ≥100%, and stage 3 an increase in creatinine ≥200% or estimated glomerular filtration rate (eGFR) < 35 mL/min per 1.73 m. ^2,7,19 Serum creatinine was recorded through post-operative day 3. The urine output definition of acute kidney injury was excluded from the criteria to be consistent with the limited studies available for comparison. ^1,9–11

Acute kidney injury was defined as stage 1 and greater and severe acute kidney injury stage 2 and 3. Utilising data previously published, we compiled the cumulative incidence of acute kidney injury after the Hybrid + Comp 2 pathway. ¹²

Data regarding urine output, fluid overload percentage, utilisation of blood products, and diuretic requirements were obtained at 8, 24, and 48 hours post-operatively. Diuretic requirements were described as utilisation (yes/no) and as furosemide equivalents in which furosemide was converted to intravenous dosage (i.e., oral furosemide is 0.5 the equivalency of the intravenous form). Fluid overload was calculated utilising net fluid balance at the above-mentioned time periods divided by the weight prior to the Comp 2. Daily weight change from pre-operative through post-operative day 3 was recorded. Vasoactive inotrope score, as previously described in the literature, was recorded through post-operative day 2. ^20,21 The utilisation of extracorporeal membrane oxygenation and renal replacement therapy was documented through the patient’s hospital course.

The primary outcome of the study was the effect of post-operative acute kidney injury following the Comp 2 on 30-day mortality. Secondary outcomes investigated included the impact of acute kidney injury on need/length of respiratory support (including intubation and length of positive pressure ventilation), length of ICU stay, and length of hospital stay.

Data were summarised using frequency (percentage) for categorical variables, mean (standard deviation) for normally distributed continuous variables, and median (interquartile range) for skewed continuous variables. Comparisons were made between acute kidney injury (≥ Stage 1) and no injury and, separately, severe acute kidney injury (≥ Stage 2) compared to no severe acute kidney injury (No Injury and Stage 1). Two-sample t or nonparametric Wilcoxon rank-sum tests were used for continuous variables and Pearson’s chi-square or Fisher’s exact tests were used for categorical variables. Two-sided p-values < 0.05 were considered statistically significant.

Logistic regression analysis was used to assess if the duration of cardiopulmonary bypass time was a risk factor for developing acute kidney injury and, separately, severe acute kidney injury. The logistic regression coefficients were estimated using iteratively reweighted least squares and were transformed to odds ratios for ease of interpretation. Ninety-five percent Wald confidence intervals were estimated and provided. Multivariate significance tests were calculated using Wald tests and evaluated at the 0.05 level of significance.

The association between the presence of post-operative acute kidney injury prevalence, and, separately, severe acute kidney injury prevalence with an increased the length of stay in the hospital was estimated using Poisson regression with the log link function. The model adjusts for vasoactive inotrope score at 24 hours post-operation, positive pressure ventilation, and fluid overload on post-operative day 2. The model also included the interaction between acute kidney injury and fluid overload on post-operative day 2. This same modelling was conducted utilising the severe acute kidney injury criteria with the only modification being fluid overload on post-operative day 1 was utilised. The coefficients were estimated using iteratively reweighted least squares and were transformed to risk ratios. The 95% confidence intervals were constructed using robust standard errors estimated by means of heteroscedasticity-consistent covariance matrix estimation. ^22,23 All models were constructed using R statistical software.

Results

Sixty-nine patients were identified as having undergone the Comp 2 during the study period. Eight patients were excluded from the study; six due to incomplete post-operative creatinine data, one due to inability to obtain secondary outcome data, and one patient death on post-operative day 1. The excluded mortality patient underwent the Comp 2 after the family declined transplantation. The study cohort was 61 patients.

The cohort was predominantly male (66%) and Caucasian in ethnicity (75%). The subtypes of HLHS were mitral atresia/aortic atresia (38%), followed by mitral stenosis/aortic stenosis (31%), and mitral stenosis/aortic atresia (31%). The median age at Comp 2 was 155 (143, 171) days, median weight at Comp 2 was 6.15 (5.6, 6.96) kg. There were no significant differences in pre-operative patient characteristics for those with and without acute kidney injury or severe kidney injury, as outlined in Table 1.

Table 1. Patient characteristics comparing the presence of post-comprehensive stage 2 procedure acute kidney injury.

The incidence of acute kidney injury after the Comp 2 was 63.9% (39/61), with 36.1% (22/61) developing severe acute kidney injury (Fig. 1). Of the patients who developed acute kidney injury, 63% met their maximum stage of injury on post-operative day 2 (Supplemental Figure 1). Based on the cumulative incidence of acute kidney injury after the Hybrid Stage 1 and the Comp 2, the majority of the cohort developed acute kidney after the Comp 2, 64.9%, compared to 13.4 % after the initial Hybrid palliation (Fig. 2). ¹²

Figure 1. Acute kidney injury incidence after the comprehensive stage 2 procedure. AKI distributions after the comprehensive stage 2 procedure are shown as frequency (Percentage %) and classified by kidney disease improving global outcomes criteria. AKI = acute kidney injury.

Figure 2. Cumulative incidence of acute kidney injury after the Hybrid Stage 1 palliation and the comprehensive stage 2 procedure. Describes the cumulative incidence of AKI after the first two palliative operations of the Hybrid palliation pathway, Hybrid Stage 1 palliation and the comprehensive stage 2 procedure. AKI was defined as greater than or equal to stage 1 AKI. Severe AKI was defined as meeting Stage 2 injury or greater. This table was created utilising data previously published by our group 1. ¹² AKI = acute kidney injury; post-Hybrid = AKI incidence after the Hybrid palliation; post-comp 2= cumulative AKI incidence after the Hybrid palliation and comprehensive stage 2 procedure.

Table 2 describes the post-operative findings comparing those who developed post-operative acute kidney injury to those who did not and, separately, severe acute kidney injury to those who did not. There was no statistical difference in 30-day mortality between the groups (Acute kidney injury; 3 (7.7%) versus No Acute Kidney Injury 2 (9.1%), p-value=> 0.9) Those with severe acute kidney injury had a longer cardiopulmonary bypass times (severe acute kidney injury 272.68 ± 72.27 minutes versus not-severe: 238.18 ± 38.19 minutes, p = 0.017). Subjects with acute kidney injury and severe acute kidney injury had a greater percent fluid overload through post-operative day 1 (acute kidney injury 6.91 ± 7.33% versus no injury 3.21 ± 4.38%, p = 0.017) and required a greater amount of furosemide equivalent diuretics on post-operative day 2. There was no significant difference in weight change from pre-op in either group (Supplemental Table 1). Patients who developed acute kidney injury and severe acute kidney injury had a longer duration of invasive mechanical ventilation and non-invasive positive pressure (intubated; acute kidney injury 32 (8, 155) hours vs. no injury 9 (0, 19) hours, p = 0.018/positive pressure ventilation; acute kidney injury 0 (0,1) days vs. no acute kidney injury 2 (0, 6) days, p-value = 0.006). Those with kidney injury and severe kidney injury required more vasoactive support on post-operative day 1 and 2 (Table 2). Patients with post-operative kidney injury and severe kidney injury required a longer intensive care stay and those with severe injury had a longer hospital length of stay (hospital length of stay; severe acute kidney injury 23 (13, 52) days versus not severe acute kidney injury: 10 (9, 17) days p = 0.002). No patients required renal replacement therapy, and one patient required post-operative extracorporeal membrane oxygenation (Supplemental Table 1).

Table 2. Post- comprehensive stage 2 procedure variables comparing the presence of post-op acute kidney injury.

Utilising logistic regression, no association was found between cardiopulmonary bypass time and acute kidney injury (odds ratio: 1.003, 95% confidence interval (0.991, 1.016), p = 0.2268, Supplemental Table 2). Utilising Poisson regression, neither acute kidney injury nor severe acute kidney injury increased the risk of a longer hospital stay (hospital length of stay: acute kidney injury relative risk 0.71; 95% confidence intervals (0.484, 1.04), Supplemental Table 3). Acute kidney injury and fluid overload on post-operative day two were not associated with an increased risk of a longer hospital stay (relative risk: 0.977; 95% confidence intervals (0.92, 1.04)). There was also no increased risk of severe acute kidney injury and fluid overload on post-operative day one with an increased length of stay (Supplemental Table 3). A higher vasoactive inotrope score at 24 hours and a longer duration of positive pressure ventilation increased the risk of a longer hospital stay (hospital length of stay: vasoactive inotrope score 24 hours post-operation relative risk: 1.069; 95% confidence intervals (1.04, 1.1), Supplemental Table 1).

Discussion

Acute kidney injury is a common complication after the Comp 2 palliation. There was no association between acute kidney injury and mortality in our cohort. Patients with acute kidney injury after the Comp 2 may require more advanced medical therapies during their stay. Providers caring for these patients should be aware of acute kidney injury in the immediate post-operative period.

The Comp 2 has been referred to as a combined Norwood and Glenn procedure. ¹⁸ Therefore, we compared our data to the available literature for the Norwood and Glenn palliations. Garcia et al. investigated post-operative acute kidney injury incidence after the Norwood procedure, defining kidney injury as paediatric risk, injury, failure, loss, end-stage kidney disease ≥ injury, and found an incidence of 53% compared to 36% we found post Comp 2. ¹⁰ In another post-operative Norwood cohort, Goldstein et al., defining acute kidney injury as Kidney Diseases Improving Global Outcomes ≥ stage 1, found an incidence of 67%, which is similar to the 64% we found after the Comp 2. ¹¹ Two studies, both utilising Kidney Diseases Improving Global Outcomes ≥ stage 1 injury, found vastly discrepant incidences of acute kidney injury post Glenn procedure: Huggins et al. at 30% and Wong et al. at 68%. ^9,24 Wong et al. is the only study to look at acute kidney injury incidence at each stage of the Norwood palliative strategy for hypoplastic left heart syndrome. While they did not discuss cumulative incidence of acute kidney injury, they did publish an incidence of 75%, with 21% meeting stage 3 injury criteria, after the Norwood and an incidence of 68%, with 12% meeting stage 3, injury after the Glenn procedure.

Given the low incidence of acute kidney injury post-Hybrid procedure, the Hybrid palliative strategy may mitigate an occurrence of kidney injury in the neonatal period. This mitigation may be beneficial in reducing the overall risk of chronic kidney disease in the single ventricle population, which is a known susceptible population for the development of chronic kidney disease. ^12,25 However, patients are high risk for developing acute kidney injury after the Comp 2. Unfortunately, a limitation of our dataset was the inability to determine if an injury event at the Hybrid led to increased risk at the Comp 2. This was due to the overall low incidence of acute kidney injury post Hybrid palliation. ¹² Multi-centre longitudinal studies would clarify the true incidence and impact of kidney injury in this population.

Those with acute kidney injury and severe acute kidney injury also had a higher percentage of fluid overload, which has been demonstrated in other studies. ^26,27 While there is an association with fluid overload and acute kidney injury, this does not necessarily mean causation and it has been difficult to derive which is the main driver of this derangement to the patient’s physiology. ²⁸ In the regression model for risk of length of hospital stay, we found no statistical difference in acute kidney injury, fluid overload, or combined acute kidney injury and fluid overload. The study may be underpowered to detect these differences.

Those who developed post-operative acute kidney injury may require more advanced medical therapies. Patients who developed acute kidney injury and severe acute kidney had a higher vasoactive inotrope score at 24 hours post-operation. Vasoactive inotrope score has been associated, in previous studies, with an increase in morbidity and mortality after paediatric cardiac bypass. ^20,21 Those with acute kidney injury after the Comp 2 required a longer duration of invasive mechanical ventilation and non-invasive positive pressure. This increased duration of respiratory support for patients with post-operative acute kidney injury has been documented previously in the paediatric cardiac literature. ^10,26,29 The respiratory morbidity may be secondary to fluid overload leading to increased pulmonary oedema and reduced lung compliance.

Contrary to our study’s main hypothesis and that seen in previous post-operative cardiac acute kidney injury studies, the presence of post-operative acute kidney injury was not associated with an increased incidence of 30-day mortality. ^1,3,8 Severe acute kidney injury was also not associated with an increased incidence of 30-day mortality. This may be due to the low number of mortality events in the cohort. In a recent multi-centre investigation of neonates post congenital heart surgery, the author’s only found an increased risk of mortality with stage 3 acute kidney injury. ⁶ This group also found that in non-bypass cardiac surgery, acute kidney injury was not associated with an increase in mortality. ³⁰

There are several limitations to this study which should be acknowledged. First, the Comp 2 is a unique procedure that is performed regularly in only a few centres. Our centre utilises the Hybrid as the preferred approach to single ventricle palliation. From 2009 to 2018, our centre performed 99 Hybrid palliations and 4 Norwood procedures. Given our experience with the Comp 2, our centre’s results may not be applicable to those institutions who reserve this strategy for their high-risk patients (i.e., low-birth weight infants or those unable to undergo cardiopulmonary bypass). This is a retrospective chart review and is subject to the limitations inherent in this methodology. The use of urine output criteria was not utilised in our criteria for acute kidney injury. This decision was made to make apt comparisons to the available literature investigating acute kidney injury after single ventricle population. ^9–11 A study in neonates has shown that urine output and response to diuretics may be an important indicator of kidney injury. ²⁷ We did not investigate the time to resolution of acute kidney injury in this paper. ¹⁴ Of the patients excluded from the study cohort with incomplete creatinine or secondary outcome data, no patients developed kidney injury based on the available data. We were unable to capture the amount of intra-operative contrast utilised in each case. However, in the available data, there was no statistical difference between these groups (Supplemental Table 1).

Acute kidney injury is a common complication after the Comp 2 procedure, with 63.9% of patients developing acute kidney injury and 36.1% developing severe acute kidney injury. In the Hybrid + Comp 2 pathway, the cumulative incidence of acute kidney injury is 69% with 36% developing severe acute kidney injury. The cumulative risk of the Hybrid + Comp 2 pathway may be less than the traditional Norwood + Glenn pathway but further studies are needed to fully answer this question. Post-operative kidney injury after the Comp 2 may be associated with an increased need for post-operative ventilatory support but not 30-day mortality.