(78) From Theory to Practice: Initial Hemodynamic Phenotypes in Congenital Diaphragmatic Hernia

Background: Congenital diaphragmatic hernia (CDH) is associated with significant morbidity and mortality, in part from heterogenous hemodynamic complexity in the first 24 hours. Three echocardiographically-defined hemodynamic phenotypes in CDH have been proposed but have not yet been intensively investigated: (P1) no/mild pulmonary hypertension (PH), (P2) pre-capillary PH ± right ventricular (RV) dysfunction, and (P3) post-capillary PH with left ventricular (LV) dysfunction. The P3 phenotype may have an adverse response to inhaled nitric oxide (iNO) and thus timely identification is key.

Objectives: Herein, we (1) systematically define the “as-yet hypothetical” initial hemodynamic phenotypes in CDH using treatment-naïve and pre-treated targeted neonatal echocardiographs (TNE), (2) assess for key outcome association, and (3) identify predictors of the P3 phenotype.

Design/Methods: All inborn neonates with CDH from 2018-2024 with a TNE in the first 24 hours were included. TNE were classified as treatment-naïve if no cardiovascular therapies were administered prior (including iNO). Phenotypes were classified according to the algorithm in Figure 1. The primary outcome was mortality/ extracorporeal life support (ECLS) in addition to select secondary outcomes. Variables found to be significant (p < 0.05) on univariate analysis were included in logistic regression to predict P3. Statistical analysis via Mann-Whitney, Chi square, Kruskal Willis, and logistic regression.

Results: Of 28 included neonates, 16 (57%) had a treatment-naïve TNE. The hemodynamic phenotype distribution was similar for treatment-naïve vs pre-treated TNE: P1 12% vs 0 %, P2 63% vs 67%, and P3 25% vs 33%, respectively. There was no significant association with the primary outcome; however, the P3 phenotype was associated with an earlier day of death, decreased survival to repair (Table 1), and mortality with iNO (Table 2). Multiparametric logistic regression including LV s’, LV e’, and RV fractional area change was predictive of P3 with an AUROC of 0.971 (Figure 2).

Conclusions: An algorithmic approach using early TNE can be utilized to identify hemodynamic phenotypes in CDH. Reassuringly, this was still feasible even if cardiovascular therapies were already started. While not directly associated with mortality, the association of iNO and mortality in the P3 phenotype underscores the importance of early and accurate phenotype delineation. The identification of a novel three-part echo-based predictor may help to identify this iNO-vulnerable P3 phenotype.