Article of the Month - Nov 2021

Epidemiology of Acute Kidney Injury After Neonatal Cardiac Surgery: A Report From the Multicenter Neonatal and Pediatric Heart and Renal Outcomes Network

Crit Care Med. 2021;49(10):e941-e951. doi:10.1097/CCM.0000000000005165. PMID: 34166288.

Alten JA, Cooper DS, Blinder JJ, et al.

Reviewed by Austin Rutledge

Background:

Neonatal AKI is common after cardiac surgery and is associated with significant morbidity and mortality; however, the reported rates of cardiac surgery-associated AKI (CS-AKI) have varied widely (15-75%) across prior single-center studies. Although the neonatal modified Kidney Disease Improving Global Outcomes (KDIGO) criteria for diagnosing AKI is now widely accepted, it has yet to be evaluated within a multicenter neonatal cardiac surgery cohort. Multicenter studies have the benefit of larger sample sizes to enhance generalizability of results, though risk confoundment by variation in clinical practice among centers if not adherent to strict protocols.

What was the purpose of the studies?

Similar to the single-center study by Beken et al., the multicenter study was designed to 1) describe the rate of neonatal CS-AKI using the modified neonatal KDIGO AKI definition, 2) evaluate perioperative predictors of neonatal CS-AKI, and 3) investigate the impact of CS-AKI on outcomes.

What were study design and methods?

This was a multicenter retrospective cohort study from September 2015 to January 2018 conducted among 22 hospitals in North America using the NEPHRON registry within the Pediatric Critical Care Consortium (PC4). The previously reviewed single-center study was performed in a level III center in Turkey, which may have differing available resources and baseline rates of comorbidities or overall mortality.

AKI was defined using either serum creatinine (SCr) or urine output (UOP) criteria of the neonatal modified KDIGO definition. Beken et al. used only SCr criteria as UOP measurements were not readily available. The highest KDIGO UOP or SCr staging criteria was used for analysis. Patients requiring kidney replacement therapy (KRT) after 24hrs following surgery were automatically assigned stage 3 AKI.

Enrollment criteria included neonates undergoing cardiac surgery from 22 PC4 hospitals. Whereas the single-center study enrolled only patients >37 weeks gestation, premature neonates were included in the multicenter cohort. The following exclusion criteria were unique to the multicenter study: 1) cardiopulmonary bypass (CPB) reoperation within 7 days, 2) preoperative KRT, 3) preoperative SCr elevation, and 4) extracorporeal membrane oxygenation during the first 24hrs after surgery.

What are the results?

2,240 patients met enrollment criteria for the multicenter study, of which 291 (13%) were born premature.

1,207 patients (53.8%) developed CS-AKI and of those, 58.2% had stage 1, 25.0% had stage 2, and 16.8% had stage 3. 12 of the 203 patients with stage 3 AKI were categorized based on the need for KRT.

The percentage of patients who underwent CPB were similar between the multicenter (74%) and single-center (71%) studies as were the distribution of AKI stages among those requiring CPB.

The 35.6% AKI rate in the single-center cohort fell within the wide range of 27-86% across the participating institutions within the multicenter study. However, the single-center stage 3 AKI rate of 46.5% was higher than the multicenter range of 0-31%. This difference is even more substantial given 68% of those with stage 3 AKI in the multicenter study were diagnosed as a result of oligo/anuria, a criterion not utilized by Beken et al.

Among the 1,195 CS-AKI patients (excluding 12 requiring KRT), 462 (38.6%) had oligo/anuria with normal SCr. Only 312 (26.1%) met any stage criteria for both SCr and UOP. There was also considerable discordance between SCr and UOP criteria with 352 (29.5%) having a 2 or more AKI stage difference.

In multivariable regression analysis, prophylactic peritoneal dialysis (PD) (OR, 1.67; 95% CI, 1.07-2.6), longer duration of CPB (OR, 1.53; 95% CI, 1.01-2.32), and use of modified ultrafiltration (MUF) (OR, 1.54; 95% CI, 1.01-2.34) were associated with CS-AKI whereas preoperative feeding (OR, 0.68; 95% CI, 0.52-0.9) and delayed sternal closure (OR, 0.76; 95% CI, 0.61-0.96) were associated with lower rates of CS-AKI.
Compared to the single-center cohort, the multicenter mortality rates were much lower at 4.5% in those with CS-AKI vs. 3.5% in those without; however, only stage 3 AKI was independently associated with mortality (OR, 2.44; 95% CI, 1.30-4.61).

The multicenter study did not find an association between CS-AKI and other morbidities such as longer duration of mechanical ventilation as shown by Beken et al. and other previous studies.

What was the conclusion?

The multicenter study confirmed the common occurrence of AKI after cardiac surgery in neonates using the modified KDIGO criteria although with substantial variability across participating institutions. CS-AKI with both oliguria and elevated SCr occurred infrequently with nearly 40% of the CS-AKI cohort having oligo/anuria, but normal SCr. Prophylactic PD, longer duration of CPB, and use of MUF independently predicted CS-AKI. Unlike the single-center study, the multicenter study included premature neonates and although a known risk factor for neonatal AKI, prematurity was not associated with CS-AKI in this cohort. The overall mortality rates were considerably lower in the multicenter cohort both with and without CS-AKI possibly reflecting baseline differences and a more homogenous patient population (i.e., stricter exclusion criteria including need for reoperation in the first 7-day study period, pre-operative renal replacement therapy, pre-operative serum creatinine elevation, and post-operative ECMO within the first 24 hours). In contrast to the single-center study, only stage 3 AKI was associated with mortality. Previously reported associated adverse outcomes of CS-AKI were not reproduced in the multicenter cohort.

Why is this important?

The NEPHRON report is the first multicenter study to describe the epidemiology of neonatal CS-AKI. Importantly, the criteria and biomarkers used for CS-AKI diagnosis impact not only the estimates of AKI rates but also the identification of events that may have varying clinical relevance in terms of impact on outcomes. The discordance in assigned CS-AKI stages between UOP and SCr criteria further support that not all SCr elevation and/or oliguria are of equal clinical importance in different circumstances. While standardized definitions are helpful for establishing a common language, the definition may require refinement in different clinical situations such as in the first 24 hours following cardiac surgery.  Both these studies show that additional research is warranted to better discern predictive factors, diagnose clinically significant CS-AKI using more precise biomarkers, and identify the patients most at risk for associated morbidities.