Flow profile characteristics in Fontan circulation are associated with the single ventricle dilation and function: principal component analysis study

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

Single ventricle heart disease is a severe and life-threatening illness, and improvements in clinical outcomes of those with Fontan circulation have not yet yielded acceptable survival over the past two decades. Patients are at risk of developing a diverse variety of Fontan-associated comorbidities that ultimately requires heart transplant. Our observational cohort study goal was to determine if principal component analysis (PCA) applied to data collected from a substantial Fontan cohort can predict functional decline (N=140). Heterogeneous data broadly consisting of measures of cardiac and vascular function, exercise (VO2max), lymphatic biomarkers, and blood biomarkers were collected over 11 years at a single site; in that time, 16 events occurred that are considered here in a composite outcome measure. After standardization and PCA, principal components (PCs) representing >5% of total variance were thematically labeled based on their constituents and tested for association with the composite outcome. Our main findings suggest that the 6th PC (PC6), representing 7.1% percent of the total variance in the set, is greatly influenced by blood serum biomarkers and superior vena cava flow, is a superior measure of proportional hazard compared to EF, and displayed the greatest accuracy for classifying Fontan patients as determined by AUC. In bivariate hazard analysis, we found that models combining systolic function (EF or PC5) and lymphatic dysfunction (PC6) were most predictive, with the former having the greatest AIC, and the latter having the highest c-statistic. Our findings support our hypothesis that a multifactorial model must be considered to improve prognosis in the Fontan population.

Extraction and Digitization of ECG Signals from Standard Clinical Portable Document Format Files for the Principal Component Analysis of T-wave Morphology

INTRODUCTION

T-wave analysis from standard electrocardiogram (ECG) remains one of the most available clinical and research methods for evaluating myocardial repolarization. T-wave morphology was recently evaluated to aid with diagnosis and characterization of diastolic dysfunction. Unfortunately, PDF stored ECG datasets limit additional numerical post-processing of ECG waveforms. In this study, we apply a simple custom process pipeline to extract and re-digitize T-wave signals and subject them to principal component analysis (PCA) to define primary T-wave shape variations.

METHODS

We propose simple pre-processing and digitization algorithms programmable as a MATLAB tool using standard thresholding functions without the need for advanced signal analysis. To validate digitized datasets, we compared clinically standard measurements in 20 different ECGs with the original ECG machine interpreted values as a gold standard. Afterwards, we analyzed 212 individual ECGs for T-wave shape analysis using PCA.

RESULTS

The re-digitized signal was shown to preserve the original information as evidenced by excellent agreement between original - machine interpreted and re-digitized clinical variables including heart rate: bias ~ 1 bpm (95% CI: -1.0 to 3.5), QT interval: bias ~ 0.000 ms (95% CI: -0.012 to 0.012), PR interval: bias = -0.015 ms (95% CI: -0.015 to 0.003), and QRS duration: bias = -0.001 ms (95% CI: -0.007 to 0.006). PCA revealed that the first principal component universally modulates the T-wave height or amount of repolarization voltage regardless of the investigated ECG lead. The second and third principal components described variation in the T-wave peak onset and the T-wave peak morphology, respectively.

CONCLUSION

This study presents a straightforward method for re-digitizing ECGs stored in the PDF format utilized in many academic electronic medical record systems. This process can yield re-digitized lead specific signals which can be retrospectively analyzed using advanced custom post-processing numerical analysis independent of commercially available platforms.

Central Venous Waveform Patterns in the Fontan Circulation Independently Contribute to the Prediction of Composite Survival

It is well appreciated that the Fontan circulation perturbs central venous hemodynamics, with elevated pressure being the clearest change associated with Fontan comorbidities, such as Fontan-associated liver disease (FALD) and protein-losing enteropathy (PLE). Our group has better quantity of these venous perturbations through single- and multi-location analyses of flow waveforms obtained from magnetic resonance imaging of Fontan patients. Here, we determine if such analyses, which yield principal components (PC) that describe flow features, are associated with Fontan survival. Patients with a Fontan circulation (N = 140) that underwent free-breathing and mechanically ventilated cardiac MRI were included in this study. Standard volumetric and functional hemodynamics, as well as flow analysis principal components, were subjected to univariate and bivariate Cox regression analyses to determine composite clinical outcome, including plastic bronchitis, PLE, and referral and receipt of transplant. Unsurprisingly, ventricular function measures of ejection fraction (EF; HR = 0.88, p < 0.0001), indexed end-systolic volume (ESVi; HR 1.02, p < 0.0001), and indexed end-diastolic volume (EDVi; HR = 1.02, p = 0.0007) were found as specific predictors of clinical events, with specificities uniformly > 0.75. Additionally a feature of IVC flow (PC2) indicating increased flow in systole was found as a highly sensitive predictor (HR = 0.851, p = 0.027, sensitivity 0.93). In bivariate prediction, combinations of ventricular function (EF, ESVi, EDVi) with this IVC flow feature yielded best overall prediction of composite outcome. This suggests that central venous waveform analysis relays additional information about Fontan patient survival and that coupling sensitive and specific measures in bivariate analysis is a useful approach for obtaining superior prediction of survival.

Myocardial strain-curve deformation patterns after Fontan operation

Myocardial deformation analysis by cardiac MRI (CMR) yielding global circumferential and longitudinal strain (GCS and GLS) is an increasingly utilized method to accurately quantify systolic function and predict clinical events in patients with Fontan circulation. The purpose of this study was to use principal component analysis (PCA) to investigate myocardial temporal deformation patterns derived from strain-time curves to learn about latent strain features beyond peak values. We conducted the study with specific attention to dominant single left or right ventricle (SLV and SRV) morphologies. Methods and Results: Patients remote from Fontan operation who underwent follow-up CMR were analyzed for standard volumetric and function hemodynamics including myocardial deformation parameters including GCS and GLS. We applied PCA to investigate in an unbiased fashion the strain-time curve morphology and to calculate patient specific shape scores. All variables were subjected to single variable Cox regression analysis to detect composite clinical outcome including death, heart transplant, protein losing enteropathy and plastic bronchitis. A total of 122 patients, (SLV = 67, SRV = 55) with a mean age of 12.7 years underwent comprehensive CMR analysis. The PCA revealed 3 primary modes of strain-curve variation regardless of single ventricle morphology and type of strain investigated. Principle components (PCs) described changes in (1) strain-time curve amplitude, (2) time-to-peak strain, and (3) post-systolic slope of the strain-time curve. Considering only SLV patients, GCS was only CMR variable predictive of clinical events (HR 1.46, p = 0.020). In the SRV group, significant CMR predictors of clinical events were derived indexed end-diastolic (HR 1.02, p = 0.023) and end-systolic (HR 1.03, p = 0.022) volumes, GCS (HR 1.91, p = 0.003) and its related first component score (HR 1.20, p = 0.005), GLS (HR 1.32, p = 0.029) and its third component score (HR 1.58, p = 0.017). CMR derived global strain measures are sensitive markers of clinical outcomes in patients with Fontan circulation, particularly in patients with the SRV morphology. Myocardial strain-time curve morphology specific to SLV and SRV patients inspired by unbiased PCA technique can further aid with predicting clinical outcomes.

The Influence of Respiration on Blood Flow in the Fontan Circulation: Insights for Imaging-Based Clinical Evaluation of the Total Cavopulmonary Connection

Congenital heart disease is the most common birth defect and functionally univentricular heart defects represent the most severe end of this spectrum. The Fontan circulation provides an unique solution for single ventricle patients, by connecting both caval veins directly to the pulmonary arteries. As a result, the pulmonary circulation in Fontan palliated patients is characterized by a passive, low-energy circulation that depends on increased systemic venous pressure to drive blood toward the lungs. The absence of a subpulmonary ventricle led to the widely believed concept that respiration, by sucking blood to the pulmonary circulation during inspiration, is of great importance as a driving force for antegrade blood flow in Fontan patients. However, recent studies show that respiration influences pulsatility, but has a limited effect on net forward flow in the Fontan circulation. Importantly, since MRI examination is recommended every 2 years in Fontan patients, clinicians should be aware that most conventional MRI flow sequences do not capture the pulsatility of the blood flow as a result of the respiration. In this review, the unique flow dynamics influenced by the cardiac and respiratory cycle at multiple locations within the Fontan circulation is discussed. The impact of (not) incorporating respiration in different MRI flow sequences on the interpretation of clinical flow parameters will be covered. Finally, the influence of incorporating respiration in advanced computational fluid dynamic modeling will be outlined.

Prediction of Fluid Responsiveness by Stroke Volume Variation in Children Undergoing Fontan Operation

Background

Fontan operation is a palliative medical procedure performed on children with single-ventricle defects. As postoperative success of the procedure largely depends on the preload volume, it is necessary to maintain an appropriate pressure gradient between the systemic vein and the left atrium to ensure the effective volume of systemic circulation. However, there is a lack of effective indexes to evaluate fluid responsiveness in Fontan patients. Stroke volume variation (SVV) is a dynamic hemodynamic parameter based on cardiopulmonary interaction in mechanical ventilation. This study is aimed at validating the sensitivity and specificity of SVV and central venous pressure (CVP) in assessing the fluid responsiveness of Fontan patients.

Method

Sixty-four children with single ventricle who underwent modified Fontan operation between May 2018 and January 2020 were included in this study. Patients were administered 10 ml·kg⁻¹ albumin for fluid challenge within 10 min after cardiopulmonary bypass. Before and after fluid challenge, the invasive arterial pressure module was connected to MostCare™ equipment to collect the cardiac index (CI) and SVV dynamically in a time window of 30 s at a frequency of 1000 Hz. According to the range of CI change, patients with ΔCI ≥ 15% were classified into the responder (R) group and those with ΔCI < 15% into the nonresponder (NR) group. Using SVV and CVP as indicators, the receiver operating characteristic (ROC) curve of the patients was established, and the area under curve (AUC), diagnostic threshold, sensitivity, and specificity were calculated.

Results

The SVV values were 16.28% (25th and 75th percentiles 14.17%-19.24%) and 13.68% (25th and 75th percentiles 12.90%-15.89%) before and after fluid challenge treatment in responders, respectively, and the values were 18.60 ± 1.83 mmHg before and 20.20 ± 2.39 mmHg for CVP after treatment. The AUC of SVV was 0.74 (95% confidence interval (CI) 0.54-0.94, P < 0.05), and the cutoff value was 16%, offering a sensitivity of 50% and a specificity of 91.7%. Meanwhile, the AUC of CVP was 0.70 (95% CI 0.50-0.92, P > 0.05), and the cutoff value was 19.5 mmHg, offering a sensitivity of 58% and a specificity of 76%.

Conclusion

SVV exhibited a good predictive value for fluid responsiveness in pediatric Fontan patients. Appropriate fluid therapy according to SVV could improve the cardiac function of such patients. Trial registration. This study was registered in Chinese Clinical Trail Registry on Jan 26, 2018. Registration number is ChiCTR1800014654. Registry URL is http://www.chictr.org.cn/showproj.aspx?proj=25019. This observational prospective study was approved by the Local Ethics Committee of Shanghai Children's Medical Center affiliated to Shanghai Jiao Tong University (SCMCIRB-K2017035).