Application of Principal Component Analysis to Heterogenous Fontan Registry Data Identifies Independent Contributing Factors to Decline

Abstract

Single ventricle disease is a serious and deadly illness, and advances in clinical management of individuals with Fontan circulation over the past two decades have yet to yield acceptable survival. Patients remain at risk of developing a diverse assortment of Fontan-associated comorbidities that ultimately require a heart transplant. Our goal in this observational cohort study was to determine if application of principal component analysis (PCA) to heterogeneous data collected from a sizable Fontan cohort (n = 140) would predict functional decline. The data, broadly comprised of blood biomarkers, lymphatic biomarkers, measures of cardiac and vascular function, and exercise (VO₂max), were collected at a single site over 11 years; 16 events occurred over that time that we consider here as a single composite outcome measure. The standardized data was transformed via PCA, and principal components (PCs) characterizing >5% of total variance were thematically labeled based on their constituents and tested for association with the composite outcome. We found that the 6th PC (PC6), which represents 7.1% of the total variance, is superior to ejection fraction (EF) as a measure of proportional hazard, is greatly influenced by blood serum biomarkers and superior vena cava flow, and displays the greatest accuracy (according to area under the curve analysis) for classifying Fontan patients. In bivariate hazard analysis, we determined that models combining lymphatic dysfunction (PC6) and systolic function (EF or PC5) were most accurate, with the former having the highest c-statistic, and the latter having the greatest AIC. Our findings support our hypothesis that improved prognostication in a Fontan population should utilize a multifactorial model.

Keywords

Single ventricle disease / Congenital heart defect / Fontan / Outcomes

Acknowledgments

This research was supported by the Jayden DeLuca Foundation, NIH CTSA Grant UL1 TR002535, and the American Heart Association Children’s Heart Foundation Predoctoral Congenital Heart Defect Research Award 20PRE35260057 to MRF. The authors would like to additionally thank Dunbar Ivy for his support of the dissertation work that led to this manuscript.

Author Contributions

M.R.F. and K.S.H. developed the main ideas for the manuscript; M.S. and M.V.D.M. contributed to revising these ideas over time and maintaining clinical relevance. M.V.D.M. and M.S. performed data extraction and outcomes adjudication. M.R.F. performed all statistical analyses; K.S.H. and M.S. checked these analyses. M.R.F. created the figures and tables. M.R.F. primarily wrote the paper, with assistance from K.S.H. All authors reviewed the manuscript. K.S.H. and M.V.D.M. are both considered senior authors on this manuscript.

Ethics Statement

Approval was granted for the “Fontan at Altitude Registry (FAR)”, protocol #14-1072, by the Colorado Multiple Institutional Review Board (COMIRB) originally on 11 June 2014 and most recently on 8 September 2023. Actions by COMIRB, and in my research practice, are guided by the principles of respect for persons, beneficence, and justice set forth in the Ethical Principles and Guidelines for the Protection of Human Subjects of Research (often referred to as the Belmont Report). Please note that the Belmont Report and the Declaration of Helsinki are both foundational documents that outline ethical principles for conducting research involving human subjects. The Belmont Report emphasizes respect for persons, beneficence, and justice, while the Declaration of Helsinki focuses on the well-being of research participants and the importance of informed consent.

Informed Consent Statement

Because all data used in this study were collected in standard-of-care treatment, data collection was granted a waiver of consent by our institutional review board.

Data Availability Statement

The anonymized data underlying this study are not publicly available due to privacy and ethical considerations. Access may be granted on a case-by-case basis to qualified researchers for non-commercial purposes, subject to approval by the institutional review board and execution of a data use agreement. Requests for data access should be directed to the corresponding author.

Funding

This research was supported by the Jayden DeLuca Foundation, NIH CTSA Grant UL1 TR002535, and the American Heart Association Children’s Heart Foundation Predoctoral Congenital Heart Defect Research Award 20PRE35260057 to MRF.

Declaration of Competing Interest

The authors have no relevant financial or non-financial interests to disclose.

Abbreviations

SVD, single ventricle disease; PLE, protein losing enteropathy; PB, plastic bronchitis; FALD, Fontan-associated liver disease; PC MRI, phase contrast magnetic resonance imaging; VVCR, ventricular vascular coupling ration; VO₂, rate of oxygen consumption; AAo, ascending aorta; SVC, superior vena cava; IVC, inferior vena cava; LPA, left pulmonary artery; cMRI, cardiac MRI; TCPC, total cavopulmonary connection; PCA, principal component analysis; PCs, principal components; EF, ejection fraction; EDVi, end diastolic volume index; ESVi, end systolic volume index; SVi, stroke volume index; CO, cardiac output; CI, cardiac index; BNP, B-type natriuretic peptide; GGT, gamma-glutamyl transferase; AST, aspartate aminotransferase; SaO₂, arterial oxygen saturation; FEV1, forced expiratory volume in one second; mSVCP, mean SVC pressure; mPAP, mean pulmonary artery pressure; HLHS, hypoplastic left heart syndrome; TA, tricuspid atresia; double outlet right ventricle, DORV; double inlet left ventricle, DILV; hypoplastic right heart syndrome, HRHS; TCPC, total cavopulmonary circuit; FVC, forced vital capacity; RV, residual volume; TLC, total lung capacity; Alk phos, alkaline phosphatase; BUN, blood urea nitrogen.

References

[1] Kritzmire SM, Cossu AE. Hypoplastic Left Heart Syndrome. In StatPearls; StatPearls Publishing: Treasure Island, FL, USA, 2020.

[2] Rychik J. Forty years of the Fontan operation: a failed strategy. Semin. Thorac. Cardiovasc. Surg. Pediatr. Card. Surg. Annu. 2010, 13, 96-100.

[3] Lee YK, Lee ER, Park BU. PRINCIPAL COMPONENT ANALYSIS IN VERY HIGH-DIMENSIONAL SPACES. Stat. Sin. 2012, 22, 933-956.

[4] Witteveen E, Wieske L, van der Poll T, van der Schaaf M, van Schaik IN, Schultz MJ, et al. Increased Early Systemic Inflammation in ICU-Acquired Weakness; A Prospective Observational Cohort Study. Crit. Care Med. 2017, 45, 972-979.

[5] Qi Y, Jia JA, Li H, Wan N, Zhang S, Ma X. Lymphocyte-monocyte-neutrophil index: a predictor of severity of coronavirus disease 2019 patients produced by sparse principal component analysis. Virol. J. 2021, 18, 115.

[6] Kristono GA, Holley AS, Hally KE, Brunton-O’Sullivan MM, Shi B, Harding SA, et al. An IL-6-IL-8 score derived from principal component analysis is predictive of adverse outcome in acute myocardial infarction. Cytokine X 2020, 2, 100037.

[7] Federolf PA, Boyer KA, Andriacchi TP. Application of principal component analysis in clinical gait research: identification of systematic differences between healthy and medial knee-osteoarthritic gait. J. Biomech. 2013, 46, 2173-2178.

[8] Schäfer M, Frank BS, Humphries SM, Hunter KS, Carmody KL, Jacobsen R, et al. Flow profile characteristics in Fontan circulation are associated with the single ventricle dilation and function: principal component analysis study. Am. J. Physiol. Heart Circ. Physiol. 2020, 318, H1032-H1040.

[9] Ferrari MR, Schäfer M, Hunter KS, Di Maria MV. Coupled waveform patterns in the arterial and venous fontan circulation are related to parameters of pulmonary, lymphatic and cardiac function. Int. J. Cardiol. Congenit. Heart Dis. 2023, 11, 100429.

[10] Ferrari MR, Schäfer M, Hunter KS, Di Maria MV. Central Venous Waveform Patterns in the Fontan Circulation Independently Contribute to the Prediction of Composite Survival. Pediatr. Cardiol. 2023, 45, 1617-1626.

[11] Schäfer M, Ivy DD, Barker AJ, Kheyfets V, Shandas R, Abman SH, et al. Characterization of CMR-derived haemodynamic data in children with pulmonary arterial hypertension. Eur. HeartJ. —Cardiovasc. Imaging 2016, 18, 424-431.

[12] Schäfer M, Kheyfets VO, Schroeder JD, Dunning J, Shandas R, Buckner JK, et al. Main pulmonary arterial wall shear stress correlates with invasive hemodynamics and stiffness in pulmonary hypertension. Pulm. Circ. 2016, 6, 37-45.

[13] Ghelani SJ, Lu M, Sleeper LA, Prakash A, Castellanos DA, Clair NS, et al. Longitudinal changes in ventricular size and function are associated with death and transplantation late after the Fontan operation. J. Cardiovasc. Magn. Reson. 2022, 24, 56.

[14] Rathod RH, Prakash A, Kim YY, Germanakis IE, Powell AJ, Gauvreau K, et al. Cardiac magnetic resonance parameters predict transplantation-free survival in patients with fontan circulation. Circ. Cardiovasc. Imaging 2014, 7, 502-509.

[15] Ishizaki U, Nagao M, Shiina Y, Inai K, Mori H, Takahashi T, et al. Global strain and dyssynchrony of the single ventricle predict adverse cardiac events after the Fontan procedure: Analysis using feature-tracking cine magnetic resonance imaging. J. Cardiol. 2019, 73, 163-170.

[16] Saiki H, Eidem BW, Ohtani T, Grogan MA, Redfield MM. Ventricular-Arterial Function and Coupling in the Adult Fontan Circulation. J. Am. Heart Assoc. 2016, 5, e003887.

[17] Diller GP, Dimopoulos K, Okonko D, Li W, Babu-Narayan SV, Broberg CS, et al. Exercise intolerance in adult congenital heart disease: comparative severity, correlates, and prognostic implication. Circulation 2005, 112, 828-835.

[18] Mori M, Hebson C, Shioda K, Elder RW, Kogon BE, Rodriguez FH, et al. Catheter-measured Hemodynamics of Adult Fontan Circulation: Associations with Adverse Event and End-organ Dysfunctions. Congenit. Heart Dis. 2016, 11, 589-597.

[19] Loomba RS, Frommelt PC, Anderson RH, Flores S, Tweddell JS. Timing of Fontan Completion in Children with Functionally Univentricular Hearts and Isomerism: The Impact of Age, Weight, and Pre-Fontan Arterial Oxygen Saturation. Pediatr. Cardiol. 2019, 40, 753-761.

[20] Tomkiewicz-Pajak L, Plazak W, Kolcz J, Pajak J, Kopec G, Dluzniewska N, et al. Iron deficiency and hematological changes in adult patients after Fontan operation. J. Cardiol. 2014, 64, 384-389.

[21] Cheng AL, Takao CM, Wenby RB, Meiselman HJ, Wood JC, Detterich JA.Elevated Low-Shear Blood Viscosity is Associated with Decreased Pulmonary Blood Flow in Children with Univentricular Heart Defects. Pediatr. Cardiol. 2016, 37, 789-801.

[22] Schafer JL.Multiple imputation: a primer. Stat. Methods Med. Res. 1999, 8, 3-15.

[23] Rychik J, Atz AM, Celermajer DS, Deal BJ, Gatzoulis MA, Gewillig MH, et al. Evaluation and Management of the Child and Adult With Fontan Circulation: A Scientific Statement from the American Heart Association. Circulation 2019, 140, e234-e284.

[24] Papadopoulou-Legbelou K, Kavga M, Fotoulaki M. Protein-losing enteropathy after Fontan operation: enteric capsule findings and management with atrial pacing. Hippokratia 2017, 21, 208.

[25] Chin AJ, Stephens P, Goldmuntz E, Leonard MB. Serum alkaline phosphatase reflects post-Fontan hemodynamics in children. Pediatr. Cardiol. 2009, 30, 138-145.