O1040: Lung Volume Measurements associated with ECMO use and Mortality in Congenital Diaphragmatic Hernia

Background: Congenital diaphragmatic hernia (CDH) is a common indication for extracorporeal membrane oxygenation (ECMO). The observed/expected lung-to-head ratio (O/E LHR), percent predicted lung volume (PPLV), and observed/expected total lung volume (O/E TLV) are established imaging measures of severity and thus ECMO utilization. We aimed to identify risk factors and lung measurement cutoff values associated with ECMO use and mortality.

Methods: We performed a retrospective cohort study of patients with unilateral CDH at a tertiary referral center from 2012-2023. Categorical data was analyzed with chi-square calculations. Logistic regression was performed for continuous variables associated with ECMO. Multivariate logistic regression was performed to identify significant prenatal risk factors for ECMO. Receiver-operating characteristic (ROC) curve analysis was performed to establish optimal lung measurement cutoff values (Youden’s index) associated with ECMO utilization and mortality.

Results: Of 197 patients, 80 (40.6%) patients required ECMO. On prenatal ultrasound, liver location in the chest was associated with ECMO (p = 0.0065). The presence of a hernia sac was protective against ECMO (p = 0.027).

172 patients (87.3%) had left-sided hernias (LCDH) and 25 (12.7%) had right-sided hernias (RCDH). Right CDH had higher rates of ECMO use (63.6% vs. 40.2% left, p = 0.0382). Lower O/E LHR was associated with ECMO (p < 0.01). Median O/E LHR was 31.1% and 39.4% for LCDH and RCDH, respectively. For O/E LHR ECMO utilization, the optimal cutoffs were 37.5% and 37.2% for LCDH and RCDH, respectively. For mortality, the optimal cutoffs were 33.1% and 36.8% for LCDH and RCDH.

Low PPLV and MRI O/E TLV were also associated with ECMO (p = 0.0026 and p = 0.0002). Median PPLV was 24.4% and 18.9% for LCDH and RCDH, respectively. For PPLV ECMO utilization, the optimal cutoff values were 28% LCDH and 21.6% RCDH. For survival, the optimal PPLV cutoff values were 28.6% LCDH and 19.0% RCDH. Median 32-34 week gestational age MRI O/E TLV was 48.5% and 39.2% for LCDH and RCDH. For ECMO use, the optimal MRI O/E TLV cutoffs were 41.6% and 44.3% for LCDH and RCDH. The optimal cutoffs for O/E TLV mortality were 40.2% and 44.3% for LCDH and RCDH.

Upon multivariate analysis, liver location in the chest (OR 10.05, 95% CI: 1.75–94.39, p = 0.0082), right-sided CDH (OR 11.96, 95% CI: 1.37–214.34, p = 0.023) lower initial O/E LHR (OR 0.92, 95% CI: 0.85-0.97, p=0.016) and absence of a hernia sac (OR 7.65, 95% CI: 1.28–74.26, p = 0.024) remained strong predictors of ECMO utilization.

Conclusion: We reported the optimal cutoff values associated with various lung measurements and ECMO utilization and mortality in CDH patients. The identified risk factors and values may assist in patient selection for fetal intervention.