169 - The neonatal sequential organ failure assessment score is associated with preoperative mortality in patients with hypoplastic left heart syndrome

Poster Number: 169
Publication Number: 169.3162

Background: Hypoplastic left heart syndrome (HLHS) is universally fatal without surgical intervention soon after birth. Studies of illness severity in the pre-operative period for HLHS patients are rare. The neonatal sequential organ failure assessment (nSOFA) is a validated operationalization of life-threatening organ dysfunction in neonatal intensive care unit patients. The utility of the nSOFA to predict mortality risk among the unique HLHS population is unknown.

Objective: To determine the prognostic utility of the nSOFA for mortality among pre-operative patients with confirmed HLHS.

Design/Methods: Following study approval by the Institutional Review Boards at each center, a retrospective analysis of HLHS cases in the level IV NICUs at Riley Hospital for Children at Indiana University Health and University of Florida School of Medicine was performed. ICD9/ICD10 codes for HLHS (ICD9: 746.7 and ICD10: Q23.4) were used to identify HLHS patients hospitalized between 1/1/2009 and 11/1/2022 (IU) or between 1/1/2012-3/1/2020 (UF). Echocardiogram report review was performed to verify the HLHS diagnosis. HLHS patients were designated as either survived to surgical palliation or died prior to surgical palliation. Groups were pooled between centers. Selected demographic variables were curated by chart review alongside the maximum nSOFA (nSOFAmax) that occurred prior to surgical palliation or death. Data were examined for normality and subsequently compared using Mann-Whitney or Chi square test. Area under receiver operating characteristics curve (AUROC) using nSOFAmax for the outcome of pre-operative mortality was calculated.

Results: We identified 226 HLHS patients including 174 patients from IU and 52 from UF (Table 1). Pre-operative mortality was low (14/226; 6%). The nSOFAmax was greater among those that died compared to those that survived (median 8 (interquartile range (IQR) 6, 12) vs 2 (IQR 0, 4, p< 0.001)) (Fig A). The AUC for the nSOFAmax to discriminate for mortality was 0.93 (95% confidence interval 0.88, 0.98), p< 0.001 (Fig B). Compared to HLHS patients with a nSOFAmax of 0, the likelihood ratio for pre-operative death doubled with a nSOFAmax of 2, tripled with 4, and was 10-fold greater with 7. Center-specific AUCs were IU: 0.94 (0.89, 0.99), p< 0.001 and UF: 0.95 (0.89, 0.99), p< 0.001.

Conclusion(s): The maximum nSOFA score showed excellent discrimination of pre-operative mortality in the unique HLHS population in aggregate and by individual center. The nSOFA may be helpful to stratify risk in future studies in this population.