Oxygen delivery in pediatric cardiac surgery and its association with acute kidney injury using machine learning

Development and Validation of a Nomogram for Predicting Acute Kidney Injury in Pediatric Patients Undergoing Cardiac Surgery

Acute kidney injury (AKI) is a common complication after cardiac surgery and associated with adverse outcomes. The purpose of this study is to construct a nomogram to predict the probability of postoperative AKI in pediatric patients undergoing cardiac surgery. We conducted a single-center retrospective cohort study of 1137 children having cardiac surgery under cardiopulmonary bypass. We randomly divided the included patients into development and validation cohorts at a ratio of 7:3. The least absolute shrinkage and selection operator regression model was used for feature selection. We constructed a multivariable logistic regression model to select predictors and develop a nomogram to predict AKI risk. Discrimination, calibration and clinical benefit of the final prediction model were evaluated in the development and validation cohorts. A simple nomogram was developed to predict risk of postoperative AKI using six predictors including age at operation, cyanosis, CPB duration longer than 120 min, cross-clamp time, baseline albumin and baseline creatinine levels. The area under the receiver operator characteristic curve of the nomogram was 0.739 (95% CI 0.693-0.786) and 0.755 (95% CI 0.694-0.816) for the development and validation cohort, respectively. The calibration curve showed a good correlation between predicted and observed risk of postoperative AKI. Decision curve analysis presented great clinical benefit of the nomogram. This novel nomogram for predicting AKI after pediatric cardiac surgery showed good discrimination, calibration and clinical practicability.

Artificial intelligence in early detection and prediction of pediatric/neonatal acute kidney injury: current status and future directions

Acute kidney injury (AKI) has a significant impact on the short-term and long-term clinical outcomes of pediatric and neonatal patients, and it is imperative in these populations to mitigate the pathways leading to AKI and be prepared for early diagnosis and treatment intervention of established AKI. Recently, artificial intelligence (AI) has provided more advent predictive models for early detection/prediction of AKI utilizing machine learning (ML). By providing strong detail and evidence from risk scores and electronic alerts, this review outlines a comprehensive and holistic insight into the current state of AI in AKI in pediatric/neonatal patients. In the pediatric population, AI models including XGBoost, logistic regression, support vector machines, decision trees, naïve Bayes, and risk stratification scores (Renal Angina Index (RAI), Nephrotoxic Injury Negated by Just-in-time Action (NINJA)) have shown success in predicting AKI using variables like serum creatinine, urine output, and electronic health record (EHR) alerts. Similarly, in the neonatal population, using the "Baby NINJA" model showed a decrease in nephrotoxic medication exposure by 42%, the rate of AKI by 78%, and the number of days with AKI by 68%. Furthermore, the "STARZ" risk stratification AI model showed a predictive ability of AKI within 7 days of NICU admission of AUC 0.93 and AUC of 0.96 in the validation and derivation cohorts, respectively. Many studies have reported the superiority of using biomarkers to predict AKI in pediatric patients and neonates as well. Future directions include the application of AI along with biomarkers (NGAL, CysC, OPN, IL-18, B2M, etc.) in a Labelbox configuration to create a more robust and accurate model for predicting and detecting pediatric/neonatal AKI.

Indexed oxygen delivery during pediatric cardiopulmonary bypass is a modifiable risk factor for postoperative acute kidney injury

BACKGROUND

Acute kidney injury after pediatric cardiac surgery is a common complication with few established modifiable risk factors. We sought to characterize whether indexed oxygen delivery during cardiopulmonary bypass was associated with postoperative acute kidney injury in a large pediatric cohort.

METHODS

This was a retrospective analysis of patients under 1 year old undergoing cardiac surgery with cardiopulmonary bypass between January 1, 2013, and January 1, 2020. Receiver operating characteristic curves across values ranging from 260 to 400 mL/min/m2 were used to identify the indexed oxygen delivery most significantly associated with acute kidney injury risk.

RESULTS

We included 980 patients with acute kidney injury occurring in 212 (21.2%). After adjusting for covariates associated with acute kidney injury, an indexed oxygen delivery threshold of 340 mL/min/m2 predicted acute kidney injury in STAT 4 and 5 neonates (area under the curve = 0.66, 95% CI = 0.60 - 0.72, sensitivity = 56.1%, specificity = 69.4%). An indexed oxygen delivery threshold of 400 mL/min/m2 predicted acute kidney injury in STAT 1-3 infants (area under the curve = 0.65, 95% CI = 0.58 - 0.72, sensitivity = 52.6%, specificity = 74.6%).

CONCLUSION

Indexed oxygen delivery during cardiopulmonary bypass is a modifiable variable independently associated with postoperative acute kidney injury in specific pediatric populations. Strategies aimed at maintaining oxygen delivery greater than 340 mL/min/m2 in complex neonates and greater than 400 mL/min/m2 in infants may reduce the occurrence of postoperative acute kidney injury in the pediatric population.