# Early Prediction of Postoperative Peritoneal Dialysis Using Lung Ultrasound Scoring in Neonates After Cardiopulmonary Bypass

## TL;DR

This study shows that lung ultrasound scores can help predict if neonates will need dialysis after heart surgery, enabling early and timely management.

## Contribution

The study introduces lung ultrasound scoring as a novel predictor for peritoneal dialysis needs in neonates post-cardiac surgery.

## Key findings

- Lung ultrasound scores at 24–48 hours post-surgery best predicted dialysis need with high sensitivity and specificity.
- Higher lung ultrasound scores at 0–2 hours and 24–48 hours independently predicted dialysis requirement.
- Lung ultrasound scores increased early post-surgery and gradually declined by days 5–7.

## Abstract

Background: Neonates and young infants undergoing cardiac surgery with cardiopulmonary bypass (CPB) are highly vulnerable to pulmonary dysfunction, systemic inflammation, capillary leak, and fluid overload, which may lead to acute kidney injury (AKI) and the need for peritoneal dialysis (PD). Lung ultrasound (LUS) is a bedside, radiation-free tool that allows real-time assessment of lung aeration and pulmonary congestion. However, its role in predicting postoperative renal support remains limited. This study aimed to evaluate whether early postoperative LUS scores could predict the need for PD in neonates after congenital heart surgery with CPB. Methods: In this prospective single-center study, 53 neonates undergoing cardiac surgery with CPB between June 2025 and January 2026 were included. LUS was performed preoperatively and at 0–2 h, 24–48 h, 72 h, 120 h, and 168 h postoperatively using a standardized six-zone scoring system (0–18). The primary outcome was postoperative PD requirement. ROC analysis assessed predictive performance, and multivariable logistic regression identified independent predictors. Results: Total LUS scores significantly increased in the early postoperative period, remained elevated for 24–72 h, and gradually declined by days 5–7. Infants requiring PD (n = 16) had significantly higher LUS scores at 0–2 h, 24–48 h, and 72 h (p < 0.05). The 24–48 h (AUC = 0.784; sensitivity 87%, specificity 62% at cut-off ≥ 11.5) LUS score showed the best predictive value for PD (AUC = 0.831; sensitivity 86%, specificity 74% at cut-off ≥ 13). In multivariable analysis, higher LUS scores at 0–2 h (OR 1.625, p = 0.048) and 24–48 h (OR 1.621, p = 0.048) independently predicted PD. Conclusion: Postoperative LUS is a reliable, noninvasive tool that can aid in predicting the need for PD in neonates undergoing cardiac surgery with CPB, supporting timely fluid and renal management.