Interdiscip Cardiovasc Thorac Surg. 2026 Jul 30:ivag213. doi: 10.1093/icvts/ivag213. Online ahead of print.
ABSTRACT
OBJECTIVES: Failing Fontan is an increasingly recognised complication after total cavopulmonary connection, yet longitudinal biomarker data remain scarce. We aimed to characterise the incidence, biomarker trajectories, and prognostic determinants of failing Fontan.
METHODS: All patients who underwent primary total cavopulmonary connection (n = 650) or conversion to total cavopulmonary connection (n = 19) at our centre between 1994 and 2022 were reviewed. Lymphocyte count, N-terminal pro-brain natriuretic peptide and its zlog value, and Fibrosis-4 index were assessed at 1 year before, 6 months before, at onset, and at last follow-up. Patients were classified into protein-losing enteropathy/plastic bronchitis and heart failure phenotypes.
RESULTS: Failing Fontan developed in 78 primary total cavopulmonary connection patients (12.0%) and 12 conversion patients (63.2%). Conversion patients developed failing Fontan earlier (P < 0.001), but survival after onset survival was comparable. Lymphocyte counts declined before onset (2.49 to 0.89 ×10³/µL, P < 0.001) and Fibrosis-4 index increased (0.060 to 0.330, P < 0.001). Phenotypes showed divergent profiles: zlog-N-terminal pro-brain natriuretic peptide at onset was 0.71 in protein-losing enteropathy/plastic bronchitis versus 5.34 in heart failure (P < 0.001). Lymphocytopenia predicted mortality early (hazard ratio 6.77, P = 0.013) but appeared protective at last follow-up (hazard ratio 0.18, P = 0.049), reflecting a phenotypic shift. On multivariable analysis, lower lymphocyte count independently predicted mortality (hazard ratio 2.44, P = 0.041), while stent implantation was independently associated with lower mortality (hazard ratio 0.32, P = 0.040).
CONCLUSIONS: Lymphocyte counts and Fibrosis-4 index change progressively before clinical onset and may serve as early warning biomarkers. The prognostic interpretation of lymphocytopenia depends on the underlying phenotype, underscoring the need for phenotype-aware monitoring.
PMID:42530872 | DOI:10.1093/icvts/ivag213