The hemodynamic effects of sacubitril/valsartan in the first 24 h of initiation in pediatric cardiac intensive care unit patients: insights using high-fidelity data

Abstract Sacubitril/valsartan is increasingly used in pediatric cardiac patients, though its acute physiologic effects in critically ill children remain incompletely characterized. We sought to define the early hemodynamic and tissue-level response to sacubitril/valsartan using high-fidelity physiologic data. This was a retrospective study of children in a cardiac intensive care unit with continuous physiologic monitoring. High-resolution data were temporally aligned to the time of sacubitril/valsartan initiation. Locally weighted regression was used to characterize percent change in physiologic variables relative to baseline. Linear mixed effects models were used to estimate post-initiation trends in absolute units from 2 to 24 h, accounting for repeated measures within patients and adjusting for age, ventricular assist device status, milrinone exposure, and circulatory physiology. Unsupervised clustering using Ward linkage was performed on percent change features to assess heterogeneity in response. Fifty-nine patients were included in the final analyses. Following sacubitril/valsartan initiation, heart rate and systolic blood pressure decreased while diastolic and mean arterial pressure increased. Central venous pressure rose over time. Measures of systemic oxygen delivery improved, including pulse oximetry and renal near-infrared spectroscopy (NIRS), with the largest effect observed in renal NIRS. These changes were consistent across models and remained significant after adjustment. Clustering identified two response patterns, reflecting differences in magnitude rather than distinct physiologic phenotypes. Sacubitril/valsartan initiation in critically ill children is associated with coordinated changes in arterial, venous, and tissue-level physiology consistent with improved systemic oxygen delivery. The response appears continuous, with variability in magnitude across patients rather than discrete phenotypes.