Cardiovascular MRI quantifications in heart failure
Citation for published version (APA):
Saporito, S. (2016). Cardiovascular MRI quantifications in heart failure. Technische Universiteit Eindhoven.
Document status and date: Published: 17/11/2016 Document Version:
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers) Please check the document version of this publication:
• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page numbers.
Link to publication
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal.
If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:
www.tue.nl/taverne Take down policy
If you believe that this document breaches copyright please contact us at: openaccess@tue.nl
providing details and we will investigate your claim.
in heart failure
PROEFSCHRIFT
ter verkrijging van de graad van doctor aan de Technische Universiteit Eindhoven, op gezag van de Rector Magnificus, prof.dr.ir. F. P. T. Baaijens, voor een
commissie aangewezen door het College voor Promoties in het openbaar te verdedigen op donderdag 17 november 2016 om 14.00 uur
door
Salvatore Saporito
Voorzitter: prof.dr.ir. A.B. Smolders 1e promotor: prof.dr.ir. J.W.M. Bergmans 2e promotor: prof.dr. H.H.M. Korsten Copromotoren: dr.ir. M. Mischi
dr.i H.C. van Assen
leden: prof.dr. F.W. Prinzen (Universiteit Maastricht) prof.dr. T. Schaeffter (King’s College London) prof.dr.ir. F.N. van de Vosse
adviseur(s): dr. P. Houthuizen (Catharina Ziekenhuis Eindhoven)
Het onderzoek of ontwerp dat in dit proefschrift wordt beschreven is uitgevoerd in overeenstem-ming met de TU/e Gedragscode Wetenschapsbeoefening.
A catalogue record is available from the Eindhoven University of Technology Library. This research was financially supported by the Dutch Technology Foundation STW.
Printed by: Gildeprint - www.gildeprint.nl Cover design by: Veronica Caprai
c
Copyright 2016, S. Saporito
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission from the copyright owner.
Heart failure (HF) is a chronic and disabling condition that occurs when the heart is unable to maintain organ perfusion at normal filling pressure. HF is a major societal challenge that affects 120,000 people in the Netherlands alone, with prevalence expected to rise as result of an aging population. As one of the first causes of hospitalization, HF places a heavy and growing burden both on patient quality of life and on health-care systems. There is need for quantitative tools to improve the stratification, management, and treatment of HF patients.
Electrical conduction defects, particularly left bundle branch block (LBBB), are known to cause mechanical dyssynchrony, which results in worsening of the cardiac mechanical efficiency and in chronic ventricular remodeling. Cardiac resynchronization therapy (CRT) has been shown to be an effective form of treatment for selected HF patients suffering from conduction disorders. However, 3040 percent of patients do not respond to therapy, therefore, criteria to predict CRT response constitute an active research area. Untreated pathological ventricular remodeling may lead to increased filling pressures and subsequent elevation of trans-pulmonary pressures, which can possibly result in pulmonary congestion with increased thoracic fluid volumes.
In this thesis, novel methods for cardiovascular quantifications in HF are introduced. The pro-posed methods rely on magnetic resonance imaging (MRI), the current gold-standard imaging modality for the assessment of cardiovascular anatomy and function at the global and regional levels. MRI-based quantifications of the dilution occurring in trans-pulmonary circulation, and of the intra-ventricular mechanical dyssynchrony were compared with alternative imaging modalities, as well as with non-invasive techniques. The response of the proposed parameters to physiological maneuvers and to surgical interventions was also investigated.
Nowadays, quantitative assessment of thoracic fluid volumes requires an invasive procedure. In this thesis, a minimally-invasive, model-based method for the characterization of the trans-pulmonary circulation is presented. The method consists of administering a bolus of gadolinium-based MRI contrast agent in a peripheral vein during a dynamic contrast-enhanced MRI
walk kinetic model and is estimated using deconvolution techniques. The dilution system is characterized by the pulmonary transit time (PTT), which is related to the intra-thoracic fluid volumes, and by the skewness measure lambda, a parameter that quantifies the relative impor-tance of diffusion and convection in the dilution process.
Significant differences were observed for the trans-pulmonary dilution parameters obtained by DCE-MRI comparing a population of healthy volunteers and HF patients. The advantages of using a model-based approach were shown with respect to related works in the literature. PTT estimates obtained by DCE-MRI were in agreement with those obtained by contrast-enhanced ultrasound, permitting assessment at the bedside; the proposed dilution parameters correlated with common measures of cardiac function. The repeatability, reproducibility, and reliability of the PTT was assessed, and a method to reduce the inter-observer variability was introduced that made use of spectral clustering for the automatic definition of regions of interest. Also changes in MRI-derived PTT in response to acute fluid displacement achieved by pneumatic leg compression devices were investigated. Thoracic fluid estimates by MRI and relative changes during the fluiddisplacement procedure were in agreement with those obtained non-invasively by surface trans-thoracic bio-impedance spectroscopy.
A novel method for the quantification of intra-ventricular dyssynchrony was validated in LBBB canine models. Regional contraction time estimates were derived from threedimensional cine-MRI short-axis image loops using a cross-correlation approach applied on the radial wall motion curves, derived from endocardial border segmentation. Additionally, a novel method for the quantification of left ventricular mechanical discoordination by analyzing the motion of the center of the left ventricular endocardial border is proposed. Finally, to investigate the role of cardiopulmonary function in the context of CRT, a study was carried out to assess changes in PTT assessed by contrast echocardiography during CRT follow-up, as well as its predictive value for CRT response assessed using clinical and symptomatic criteria.
Mechanical dyssynchrony assessed by cine-MRI was in agreement with dyssynchrony assessed from tagged MRI, the gold standard for myocardial deformation imaging that requires dedicate and time-consuming additional image acquisitions. Dyssynchrony measured by cine-MRI signifi-cantly increased after isolated LBBB induction in animal models. The proposed cross-correlation approach enabled contraction-time estimates that were not limited by the image acquisition frame rate, and also took into account the relaxation phases of the cardiac cycle. Endocardial center motion-derived measures, which took into account both timing and amplitude of the ventricular contraction, were significantly different between healthy subjects and LBBB subjects. With regard to CRT patients, significant changes were observed in PTT during follow-up. Moreover, PTT was significantly longer at the baseline in non-responders to therapy. The presented findings suggest that PTT may provide value for prediction of CRT response.
The results included in this thesis advance the state of the art by providing estimates for intra-thoracic fluid volumes and left ventricular mechanical performance measures. As a result, novel insights on pulmonary circulation in HF and on the different approaches for quantification
protocols.
The presented results motivate an extensive validation of the proposed methods to investigate their potential clinical value. The integration of the proposed imaging techniques with minimally-and non-invasive continuous monitoring solutions through implantable or wearable devices is envisioned. Furthermore, the proposed signal processing methodologies could be further applied to other imaging modalities and to other biomedical domains.
In conclusion, the research presented in this thesis shows that model-based signal processing techniques can be used to extract physiological parameters of clinical interest, which could possibly pave the way to an extended role of MRI in HF management.
1 General introduction 1
1.1 Heart failure . . . 1
1.1.1 Definition . . . 1
1.1.2 Heart failure figures . . . 2
1.2 Congestion in heart failure . . . 2
1.2.1 Invasive pressure measurements . . . 4
1.2.2 Alternative techniques . . . 4
1.2.3 Volume measurements . . . 5
1.3 Dyssynchrony in heart failure . . . 8
1.3.1 Conduction disorders . . . 8
1.3.2 Cardiac resynchronization therapy . . . 9
1.3.3 Dyssynchrony measures and patient selection . . . 10
1.4 Magnetic resonance imaging . . . 12
1.5 Scope of the thesis . . . 13
1.6 Objective . . . 15
1.7 Outline of the thesis . . . 16
2 Reliability of pulmonary transit time assessment 29 2.1 Introduction . . . 31 2.2 Methods . . . 31 2.2.1 Patients . . . 31 2.2.2 Measurement protocol . . . 32 2.2.3 Statistical analysis . . . 33 ix
2.3 Results . . . 35
2.4 Discussion . . . 37
2.4.1 Limitations . . . 40
2.5 Conclusions . . . 42
3 Automatic regions of interest definition 45 3.1 Introduction . . . 47 3.2 Method . . . 49 3.2.1 Image acquisition . . . 49 3.2.2 Image analysis . . . 50 3.2.3 Validation . . . 53 3.3 Results . . . 54
3.3.1 Results on simulated data . . . 54
3.3.2 Results on DCE-MRI . . . 55
3.3.3 Results on CE-US . . . 56
3.4 Discussion . . . 58
3.5 Conclusions . . . 59
4 Characterization of the trans-pulmonary circulation by DCE-MRI 63 4.1 Introduction . . . 65
4.2 Methods . . . 66
4.2.1 Population . . . 66
4.2.2 Magnetic resonance imaging . . . 66
4.2.3 DCE-MRI . . . 67
4.2.4 Dilution system identification . . . 68
4.2.5 Statistical methods . . . 68
4.3 Results . . . 69
4.3.1 Comparison between HF and HV . . . 71
4.3.2 Comparison with standard cardiac function measures . . . 73
4.3.3 Comparison with time-to-peak measures . . . 73
4.4 Discussion . . . 73
4.4.1 Feasibility and contrast agent usage . . . 73
4.4.2 Comparison of HF patients versus healthy subjects, reproducibility, and repeatability . . . 73
4.4.3 Comparison with difference in time-to-peak approach . . . 74
4.4.4 Quantitative comparison with PTT values reported in the literature . . . 74
4.4.5 Comparison of the method with the literature . . . 75
4.4.6 PTT as congestion measure . . . 76
4.4.8 Limitations . . . 77
4.5 Conclusions . . . 77
5 Comparison between CE-US and DCE-MRI for PTT assessment 81 5.1 Introduction . . . 83
5.2 Methods . . . 83
5.2.1 Study population . . . 83
5.2.2 Pulmonary transit time estimation . . . 84
5.2.3 Dynamic contrast-enhanced MRI . . . 84
5.2.4 CE-US and Doppler echocardiography . . . 85
5.2.5 N-terminal pro-B-type natriuretic peptide sampling . . . 86
5.2.6 Statistical analysis . . . 86
5.3 Results . . . 87
5.3.1 PTT as a measure for cardiac function . . . 89
5.3.2 (n)PTT by CE-US in relation to echocardiographic parameters . . . 90
5.4 Discussion . . . 90
5.4.1 Limitations . . . 93
6 Comparison between MRI and thoracic impedance 97 6.1 Introduction . . . 99
6.2 Methods . . . 101
6.2.1 Study population . . . 101
6.2.2 Leg compression procedure . . . 101
6.2.3 Data acquisition . . . 102
6.2.4 Data analysis . . . 103
6.2.5 Statistical methods . . . 106
6.3 Results . . . 107
6.3.1 Cardiovascular magnetic resonance . . . 107
6.3.2 Bioimpedance spectroscopy . . . 108
6.3.3 Comparison of BIS and CMR results . . . 109
6.4 Discussion . . . 111
6.4.1 Limitations . . . 113
6.5 Conclusions . . . 113
7 Assessment of left ventricular mechanical dyssynchrony 119 7.1 Introduction . . . 121
7.2.1 Data acquisition . . . 122 7.2.2 Data processing . . . 123 7.2.3 Measures of dyssynchrony . . . 124 7.2.4 Statistical methods . . . 125 7.3 Results . . . 125 7.4 Discussion . . . 127
8 Endocardial center motion characterization 135 8.1 Introduction . . . 137
8.2 Materials and Methods . . . 138
8.2.1 Magnetic resonance imaging protocol . . . 138
8.2.2 Image analysis . . . 138 8.2.3 Statistical methods . . . 140 8.3 Results . . . 141 8.3.1 Patient characteristics . . . 141 8.3.2 ECM measures . . . 143 8.4 Discussion . . . 144
8.4.1 ECM measures as discoordination measures . . . 145
8.4.2 Comparison with other discoordination measures . . . 145
8.4.3 Discoordination measures and lead optimization . . . 146
8.4.4 Limitations . . . 146
8.4.5 Conclusions . . . 146
9 Pulmonary transit time and CRT 149 9.1 Introduction . . . 151
9.2 Methods . . . 151
9.2.1 Trans–thoracic echocardiography . . . 153
9.2.2 Estimation of the pulmonary transit time . . . 153
9.2.3 Evaluation of clinical response to CRT . . . 154
9.2.4 Evaluation of echocardiographic and biomarker response to CRT . . . . 154
9.2.5 Statistical analysis . . . 154 9.3 Results . . . 156 9.3.1 Response to CRT . . . 158 9.3.2 Prediction of response to CRT . . . 158 9.4 Discussion . . . 160 9.4.1 Primary results . . . 160 9.4.2 Decrease in PTT/nPTT due to CRT . . . 160 9.4.3 PTT as a predictor of response to CRT . . . 160
9.4.4 Potential clinical implications . . . 161
9.5 Conclusion . . . 162
10 General discussion and future prospects 165 10.1 Application context . . . 165
10.2 Fluid shifts and accumulation . . . 166
10.3 Left ventricular dyssynchrony and resynchronization . . . 169
10.4 Future prospect and conclusions . . . 171
Curriculum Vitae 185
General introduction
1.1
Heart failure
1.1.1
Definition
Heart failure (HF) is a chronic, disabling condition where the heart can not supply enough blood to meet the requirements of the rest of the body. Cardiovascular conditions, such as acute myocardial infarction (MI) or hypertension, but also non-cardiac conditions, such as diabetes, renal failure, and metabolic syndromes, are known triggers for HF development [1]. HF typically presents with a heterogeneous pathophysiology, often complicated by the presence of several chronic co-morbidities [2]. HF diagnosis carries risks of frequent hospital admissions, and the 5-years survival-rate is close to 50%, comparable to the most common cancer diagnoses [3]. HF severity is commonly assessed by measuring the left ventricular (LV) ejection fraction (EF), typically by echocardiography [4]. EF is the fraction of blood pumped from the heart during each heartbeat, and it is computed as the ratio between the stroke volume (SV) and maximum LV volume. Although recently the outcomes for HF patients have improved with the development of multiple drug and device therapies [5], HF is still associated with symptoms worsening over time, declining functional capacity, diminished quality of life, and premature death [4].
1.1.2
Heart failure figures
Cardiovascular diseases (CVDs) are still among the main causes of mortality in the European Union, despite a strong decline in mortality for CVDs in western countries over last decades [6]. Among CVDs, HF imposes a significant socio-economic burden on health-care systems and societies in general. HF affects 2 to 3% of the population in industrialized countries [1, 7], with a marked rise over age 70, where it can reach 10% prevalence [6, 8]. Prevalence is expected to increase in the near future, mainly due to the demographic trend of population aging and to the improved treatment of several cardiac and non-cardiac conditions, such as the prolonged survival in patients suffering from CAD [9, 10], and the better prevention of sudden cardiac death [11]. It has been reported that some 26 million worldwide [5], 15 million Europeans and 6 million US Americans suffer from HF [1, 12]. In the Netherlands alone, approximately 120,000 persons are currently diagnosed with HF, with an expected increase in the coming decade to an approximately 200,000 [13, 14]. The estimated annual cost for HF in the US increased from 24 billion dollars in 2003 to 40 billion dollars in 2010 (around 2% of the total health-care budget) [15], and is projected to increase to around 70 billion dollars by 2030[16]. Similar figures and trends are expected in EU countries[17].
Acute HF syndromes are the direct cause of approximately one million hospitalizations and an additional factor in other 2.4 million hospitalizations each year in the US [18]. HF has been identified as the largest expenditure for a single diagnosis-related group [19], and as result of the aforementioned demographic and health care trends, costs attributed to HF are expected to steadily increase.
Management of HF patients by remote monitoring has been proposed as a cost saving strategy [20, 21] as a reduction in health-care resource utilization may be obtained by limiting in the number of HF hospitalizations [22]. However, randomized controlled clinical trials performed using telemonitoring with non-implanted [22], and implanted devices [23, 24] have given inconsistent results [25]. At present moment, monitoring strategies based on indirect measures of intra-thoracic fluids have not shown to improve patient outcomes [26]. Clinical value for these monitoring strategies may possibly be improved by tailoring them in a patient-specific manner, by the combination with physiological sensors in other modalities, static baseline characteristics, and additional biomarkers [23].
1.2
Congestion in heart failure
Acute heart failure (AHF) is commonly defined as a rapid onset, or worsening, of signs and symptoms of HF. It is a life-threatening condition that requires immediate medical attention and usually leads to an unplanned hospitalization or emergency room visit. Recurrent hospitalizations, occurring with a rate of 25% in the 60 to 90 days after discharge [27, 28], are an important source of costs, and account for a large fraction in annual HF expenditures [29–31]. Also short-term re-hospitalizations, within 30 days, remain frequent [19, 32].
In the IMPACT-HF trial at 60 days after discharge, approximately 45% of patients experienced worsening HF, and 25% required re-hospitalization [28]. Traditionally, episodes of AHF were considered to be characterized by increased pulmonary capillary wedge pressure (PCWP) [33] and decreased cardiac output (CO) [34]. However, data from the ADHERE, OPTIMIZE-HF trials shown that most hospitalizations for AHF occur because of congestion rather than a low CO [18, 27, 33–36]. Congestion, regarded as a nearly universal finding in AHF [34, 37], it is emerging as a main contributor to the progression of HF [38], and as an important predictor of mortality [18, 39].
Little weight change before or during an AHF event suggests that in many cases congestion is caused by other mechanisms than pure fluid accumulation, such as fluid redistribution and neurohormonal or inflammatory activation, and complicated by renal dysfunction [40]. Fluid loss in itself is not related to improvements in symptoms in most patients.
Fluid redistribution may be the result of a combined vascular and cardiac process reducing capacitance in the large veins and increasing arterial stiffness and resistance. The heightened preload causes the increased intracardiac pressures to be transmitted back to the pulmonary circulation [41], eventually redistributing fluid to the lungs and promoting the clinically observed pulmonary congestion [42]. The increase in LV filling pressures may cause augmented LV wall stress; this often results in changes of the ventricular shape (geometric remodeling), and in secondary mitral valve insufficiency [34, 43, 44], eventually inducing ischemia, which leads to further myocardial cell death. Moreover, the increased PCWP can lead to redistribution of fluid within the lungs, resulting in edema, depending on several other factors such as permeability and integrity of the alveolar-capillary membrane, and efficiency of lymphatic drainage [18].
The development of signs and symptoms represents the main reason for hospitalization in HF patients. However, elevated PCWP and LV filling pressure may be present days or weeks before the need for hospitalization [18]. Gheorghiade et al. distinguish the elevation of the LV filling pressures, the hemodynamic congestion, as opposed to clinical congestion, which occurs later and is evidenced by clinical signs as dyspnea, pulmonary rales, edema, and jugular venous distension It has been shown that an increase in intra-thoracic fluids characterizing hemodynamic congestion could be detected as early as 18 days before the resulting clinical congestion requiring hospitalization [45].
Currently, congestion is often not adequately addressed in AHF as approximately 50% of patients are discharged with persistent symptoms [2, 18, 28, 32], despite the fact that congestion was the main reason for admission. Therefore, there is an urgent need for techniques for the assessment and management of congestion, in order to improve quality of life and to reduce future hospitalizations [33, 46].
1.2.1
Invasive pressure measurements
During pulmonary artery catheterization (PAC), a catheter is inserted into the right side of the heart, and subsequently passed into the pulmonary artery. This allows estimation of pressures in the right heart, as well as PCWP by the inflation of a balloon on the catheter tip. PAC is the current gold standard [19] to measure PCWP and cardiac output (CO). Elevated PCWPs, associated with volume overload, correlate with symptoms and survival rates in HF [33, 36, 47, 48]. The Evaluation Study of Congestive HF and PAC Effectiveness (ESCAPE) demonstrated that PCWP was one of the most important predictors of post-discharge survival, although routine use of PAC to guide therapy does not result in improved outcomes [49].
However, PAC is an invasive procedure associated with complications that may compromise patient safety [50]. Therefore, at present moment there are no indications for routine use of PACs to adjust therapy during hospitalization for AHF [49] given the cost, potential complications, and the lack of demonstrable benefits [4].
1.2.2
Alternative techniques
Clinical signs of congestion include abnormal heart sounds, pulmonary rales, and atypical jugular venous pulse [51]. While physical signs and hemodynamic status are associated in AHF, the chronic conditions are characterized by the presence of compensatory mechanisms that may frequently cause signs to be absent despite elevations of filling pressures [51]. In previous studies, the combination of clinical signs showed a good specificity but only a 58% sensitivity in detecting patients with elevated PCWP, therefore providing insufficient accuracy [51, 52]. Chest X-Ray is a standard clinical method for assessing pulmonary edema, allowing estimation of the presence of alveolar and interstitial lung water [53]. It is commonly used to identify a pulmonary explanation for a patient’s symptoms [4]; however, the absence of chest X-ray find-ings does not exclude the presence of high PCWP [54] and in general only a semi–quantitative score for congestion can be derived. Moreover it requires the use of ionizing energy and specific operator expertise [53].
B-lines are artifacts observed during ultrasound imaging of the lungs; in presence of edema, dis-crete, vertical reverberation lines from the pleural line arise. Assessment of lung fluids by B-lines provides a technique for thoracic fluid assessment, [55, 56], potentially expanding the already established role of trans-thoracic echocardiography in HF [53]. They are a chest ultrasound sign of extravascular lung water that can be obtained in a simple, non time-consuming way at the bedside, and they are not restricted by traditional cardiac acoustic window limitations [56, 57]. B-lines strongly predicted re-hospitalizations over 6-months [58]. However, the anatomical and physical basis of the phenomenon, which is still considered an artifact in the traditional ultrasound physics, is not completely known at present moment [59] and further research would be necessary to reach consensus guidelines for its interpretation [59, 60]. Finally, at present moment, only semi-quantitative scores such as line–count can be provided [57, 60].
B-type Natriuretic peptides (BNPs) are a group of hormones secreted in increased amount when the load on any cardiac chamber is increased [61, 62]. BNP levels provide prognostic information in HF, and their decrease during hospitalization is associated with lower rates of readmissions [34, 63, 64]. While it has been shown that BNP levels at the baseline and at discharge in AHF patients provide complementary and independent prognostic information, changes in concentrations of the BNP during hospitalization were less informative [65]. It has been suggested that BNP concentration cannot be used to follow dynamic changes in congestion since BNP pattern of release is slower respect to hemodynamic variations therefore a reliable fluid status assessment can not be performed based on BNP levels alone [18].
Tissue Doppler imaging indexes, such as the E/Ea given by the ratio between the early trans– mitral velocity (E peak, influenced by filling pressures) measured by pulsed–wave Doppler to early diastolic velocities recorded at the lateral corner of the mitral annulus (Ea, a preload– independent index of LV relaxation influenced by longitudinal myocardial shortening) by tissue Doppler, is relatively simple to obtain and relates well to mean PCWP. The E/Ea ratio has the potential to estimate LV filling pressures by correcting the trans–mitral velocity for the influence of relaxation, allowing the discrimination between different degrees of diastolic dysfunction [66, 67]. E/Ea ratio has been shown to predict adverse outcomes in relatively small HF popula-tions [68]. However, this parameter may be difficult to assess in the presence of confounders such as alterations in myocardial structure, ischemia or infarction [66, 69], mitral regurgitation, or the presence of synchronized pacing, which reduce its predictive value [69].
Similarly, the ratio between the peak early mitral inflow velocity and the color M-mode Doppler flow propagation velocity (E/Vp) [70] has been considered for the estimation of LV filling pressures; however, its relationship with PCWP observed in large population was not confirmed within individual subjects [71]. Both non-invasive indices do not reliably track changes in left-sided filling pressures [69, 71, 72]; their relationship with the actual PCWP is so variable that it may not be accurate enough for adjusting therapy [50, 71].
1.2.3
Volume measurements
Body weight (BW) is often used as a marker of congestion in both inpatient and outpatient settings. Although a change in BW observed over long time (such as weeks or months) may be associated with other factors such as diet, daily variations most likely reflects changes in volume status. It has been shown that important increases in BW start occurring at least 1 week before hospitalization for HF [73, 74]. However, recent studies have suggested minimal BW changes before, during and after an AHF episode [42], and that the degree of BW loss during an AHF admission was not associated with a substantial and sustained improvement in patients’ symptoms or outcome [75], suggesting that the relationship among BW, congestion, and clinical outcomes could be complex [76]. A decrease in BW cannot be used as an indiscriminate target for reducing hospitalization events, since BW fluctuations do not always reflect changes in intravascular volumes [75].
Thoracic Impedance (TI) measurement provided by an external or implanted device has been investigated in recent years for assessing thoracic fluid status. The measurement principle is based on the idea that accumulation of intra-thoracic fluid increase the conductance (or decrease the impedance) to electrical currents passing across the lungs [45]. TI can be measured non-invasively with surface electrodes, using a portable or even wearable system [77–79]. TI provides an indirect measurement of thoracic fluid content [80, 81] which correlates with PCWP and fluid balance [82] and predicts hospitalizations for HF [34, 45]. In particular, a reduc-tion in TI has been observed starting as early as 18 days before an AHF event [45] with higher sensitivity than acute BW changes [16, 83]. However, TI monitoring is hampered by significant inter–subject variability, as measurements are affected by adiposity, muscularity, height, intrinsic lung characteristics, and in the case of wearable devices, by electrode placement [45], subject posture [79, 84], activity, and skin preparation for electrodes [85].
Due to the variability in measurements related to these factors, the clinical utility of impedance measurements is limited to observation of relative changes over time within subjects [80] as show in Figure 1.1. This may partially explain the results from trials such as DEFEAT-PE considering the variations in intra-thoracic impedance measured by implanted devices, where the threshold to obtain a practically acceptable false-positive rate resulted in reduced sensitivity for the AHF prediction [86].
Figure 1.1: Fluid accumulation indexes can be derived from the difference between the daily and reference thoracic impedance. Alerts are triggered by short–term fluctuations of fluid indexes measuring relative variations in thoracic fluids.
Studies suggested that the integration of other parameters (such as heart rate variability [87], physical activity [88]), which are currently also recorded by implantable electrical devices, may provide a more sensitive tool to predict HF worsening than TI monitoring alone. Others stressed the importance of patient-specific adjustment of the threshold for TI changes, which would require normalization or baseline calibration [16]. Although at present moment TI monitoring represents only an adjunctive clinical tool for managing HF in patients with implanted devices, its use as part of home telemonitoring solutions could be extended, beyond anticipation of AHF episodes, to therapy guidance, medication adjustment or assessment of medication compliance. Indicator dilution methods provide a measurement technique for the assessment of cardio-vascular volumes [89]. In these methods, a suitable indicator is injected in the circulation as a bolus [89]. Subsequent sampling of the indicator concentration allows the derivation of indicator dilution curves (IDCs). Model-based interpretation of the IDCs, often performed using pharmaco-kinetic models [90], allows the derivation of physiological parameters of interest, such as cardiac output [91, 92], ejection fraction [93–95], pulmonary blood volume [96, 97], and extravascular lung water [91, 98]. Specifically, the central volume theorem [99, 100] pos-tulates that the amount of fluid between two IDC measurement sites is equal to the difference in the indicator mean transit time multiplied by the flow rate in the system. An important distinction has to be made between purely intravascular indicators, such as indocyanine green and microbubbles, and extravasating indicators, such as temperature and most of the clinically approved gadolinium-based magnetic resonance imaging (MRI) contrast agents, which can assess extravascular extracellular spaces; based on the characteristics of the indicator, different diffusion volumes can be quantified [101].
Recently, Cao et al. [102] have shown that transit times provide an attractive minimally invasive technique for the assessment of LV filling pressures, as shown in Figure 1.2. Moreover, as shown by Bogaard et al. [103] and Giuntini et al. [97], the skewness of the transit time distribution of an extra-vascular indicator is influenced by the relative contributions of diffusion and convec-tion, which might be altered in a two-compartment dilution system, such as in the presence of pulmonary edema and extra-vascular lung fluids. Originally, blood volume analysis techniques used red blood cells tagged by radioisotopes, or tagged albumin to directly measure blood volume [104]. Blood volumes correlated with invasive measurements of cardiac filling pres-sures in patients presenting with decompensated HF [68]. Alternatively, cold-saline [105, 106], dyes [107, 108], and lithium [109] have been used as indicators. In particular, trans-pulmonary thermodilution (TPTD) is currently regarded as the clinical standard for pulmonary blood volume (PBV) and extravascular lung water (EVLW) measurement [98, 110], and it is used to calibrate commercial devices used in clinical practice [111]. As the invasive detection of these indicators requires central vessel catheterization, these methods carry risks for patient safety and are not suitable for monitoring or routine screening; therefore, they cannot be used to measure volume variations over a short period of time.
Recently, imaging contrast agents have been proposed for minimally invasive application of indi-cator dilution methFods using contrast–enhanced ultrasound (CE–US) [90, 105, 112], CT [113]
and dynamic contrast-enhanced MRI (DCE-MRI) [114–117], as sampling of the indicator con-centration can be performed outside the body using dynamic imaging techniques. An important requirement for applying indicator dilution methods is the monotonic relationship the contrast agent concentration vs. signal intensity, which allows to approximate IDCs using the image time-enhancement curves. Accurate volume estimates of variable volume physical phantoms have been reported using CE-US and DCE-MRI [115, 118]. Moreover, blood volume mea-surements based on CE-US have been shown to have good correlation with invasive TPTD in patients undergoing cardiac surgery [119]. However, also when applied in a minimally–invasive way, indicator dilution methods remain intrinsically intermittent measurement techniques.
Figure 1.2: Comparison between left atrial normalized transit time and left ventricular end-diastolic pressure. r=0.91, p<0.001 Adapted from: Cao et al. Circ Cardiovasc Imaging 2011, 4(2) 130-8
1.3
Dyssynchrony in heart failure
1.3.1
Conduction disorders
Under normal conditions, the LV electrical activation can be considered synchronous, resulting in a relatively fast and synchronous mechanical contraction along the myocardium. Altered pathways for the electrical depolarization propagation, which induces the muscle contraction, result in a prolonged QRS complex [19]. Impaired depolarization leads to regional delays in ventricular mechanical contraction timing, resulting in a globally inefficient and discoordinate contraction, and producing progressive LV dilation and anatomical remodeling [120].
Left bundle-branch block (LBBB) is a conduction defect that results in of altered electrical activation of the LV as shown in Figure 1.3. LBBB causes inter- and intra-ventricular me-chanical dyssynchrony [121, 122]. LBBB is a predictor of mortality [123, 124] and causes ventricular remodeling [125]. An early and abrupt contraction of the inter-ventricular septum, known as septal flash (SF) [126, 127], is often presented by LBBB patients. SF occurs during the pre-ejection period, before the contraction of the delayed posterior wall[128, 129]. It has
been noticed the relative right and LV size and pressures are also determinants of the abnormal septal motion [130, 131], resulting in a variety of contraction patterns [132]. Additionally, asynchronous relaxation occurs in presence of LBBB; the prolonged isovolumic relaxation time [133, 134] results in abnormalities also in diastolic filling patterns [135–137].
Figure 1.3: Typical examples of three-dimensional left ventricular electrical activation time in canine hearts during normal conduction (left) and after induction of left bundle branch block (right), obtained by a combined endocardial and epicardial electrical mapping
Adapted from: Strik et al. J Cardiovasc Transl Res. 2012 Apr; 5(2): 135145.
Figure 1.4: Examples of septal longitudinal strain patterns in 3 left bundle branch block patients derived from ultrasound speckle tracking.
Adapted from: Leenders et al. Circulation: Heart Failure. 2012, 5(1):87-96.
1.3.2
Cardiac resynchronization therapy
Cardiovascular implantable electronic devices (CIEDs) include cardiac pacemakers, implantable cardioverter-defibrillators, and cardiac resynchronization therapy (CRT) devices. CRT devices deliver electrical impulses to the different chambers of the heart to promote synchronized electrical activity and improve the heart’s pump function. CRT is currently an established therapy for selected HF patients with conduction defects [138, 139]. Large randomized controlled clinical trials have clearly shown that CRT improves symptoms and quality of life [140], reduces mortality [141], and induces reverse ventricular remodeling [142].
The number of device implantations has markedly increased over the past decade and so have the associated costs [1]. The majority of the costs for these devices occurs at the time of implantation while the maintenance costs are relatively low, especially compared other therapeutic options, such as drugs that present cumulative costs increasing over time [1]. Optimal selection of patients referred for device implantation offers an opportunity to reduce HF related costs.
1.3.3
Dyssynchrony measures and patient selection
Electrical dyssynchrony
Currently, patient selection for CRT referral is based on severity of disease and of LV systolic function, and on QRS complex duration, measured by electrocardiogram [143]. However, pa-tient selection remains challenging, since with current criteria, up to 30% do not respond to therapy [144–146] while LV reverse remodeling is achieved in only 50% of the patients [147]. Therefore, extensive research towards the identification of better predictors of response is still ongoing [148–150]. The limited response rate mentioned above can be ascribed to the limited predictive value of current patient selection criteria [143, 151]. The criteria for response is commonly defined in terms of reverse remodelling, confirmed by a decrease in LVESV which translate in improved mechanical pump function.
Several studies have shown a discrepancy between electrical activation and mechanical contrac-tion. [152–155]. It has been reported that mechanical dyssynchrony is common in HF patients with impaired LV function, even without prolonged QRS duration [156–159]. Moreover, QRS duration reflects interventricular dyssynchrony as well [160], which has shown to be unrelated to improvements during CRT [161, 162].
Mechanical dyssynchrony
Mechanical dyssynchrony is associated with impairment of LV pump function, resulting in increased mortality and morbidity [123, 163]. Research suggested that mechanical dyssyn-chrony may provide additional prognostic value with respect to electrocardiographic mea-sures [164, 165]. Several imaging–based approaches have been proposed to assess mechanical dyssynchrony in a non invasive way, allowing the evaluation of the effects of the conduction disorders on the LV mechanical function, and the accurate localization of contraction abnor-malities. Imaging techniques provide a tool for visualizing paradoxical septal motion patterns, whose presence seems to be associated with a more favorable long-term survival and remodeling after CRT [126, 166]. Echocardiography is a cost-effective, routinely used imaging test for the assessment of cardiovascular function in HF patients [4]. Echocardiography has been firstly used to perform quantitative analysis of LV dyssynchrony [161], suggesting that intraventricular dyssynchrony may be a better predictor for CRT response and reverse LV remodeling than
ECG markers [167, 168]. Tissue Doppler imaging (TDI) can be used to measure the velocity of myocardial motion. Although it has been applied for CRT optimization and patient selection with promising results [162, 169, 170], its applicability is restricted by ultrasound probe position and orientation dependency [171, 172]. Strain imaging has also been proposed [173, 174], yielding indicators with higher discriminatory value between different dyssynchronopathies [175].
Limitations
Different trials have shown promising results for prediction of CRT response based on echocar-diographic assessment of mechanical dyssynchrony [162, 168, 170]; however, the results were not confirmed in larger multi-center studies [158]. It has been suggested that this may have been due to intrinsic limitations of echocardiography such as operator dependency, the requirement of suitable acoustic windows which affect the image quality, an intrinsic anisotropic sensitivity in longitudinal and lateral direction. All the factors might lead to a generally limited reproducibility [175–177]. Recent literature suggested that other factors beside mechanical dyssynchrony may play a role in CRT response prediction. Ischemic aetiology was negatively associated with CRT outcome [178, 179] since the infarcted myocardium forms scar, which is less compliant than viable tissue [180]. Other investigators underlined also the lack of agreement on the criteria for CRT response [181].
A common problem to all imaging-based approaches to dyssynchrony quantification is that currently there is no agreement of which could be the optimal dyssynchrony metrics for prospec-tive CRT response prediction [182]. Although several indexes have been proposed [182], it is still unclear how to derive a reliable and sensitive measure of dyssynchrony. As a result, dyssynchrony detected by one index is often not confirmed when a different index is used [183]. Most authors considered indices of dyssynchrony given by measures of spread of events in the cardiac cycle, such as the standard deviation of the time to onset of contraction, time to peak radial motion, or time to peak circumferential strain [173]. However, LBBB subjects may exhibit significantly different spatial contraction patterns [132]; therefore, the definition of the time to peak is challenging as LV motion is characterized by multiple peaks as shown in Figure 1.4. Only a limited number of studies considered hemodynamic improvements induced by CRT, while most research focused on the improvement of intra-operative measures of cardiac performance alone [184, 185]. Part of the hemodynamic benefit induced by CRT may also be explained by optimization of cardiac time intervals, improving the diastolic filling [186]. The effect of CRT on diastolic function have not been extensively described and are still debated [187]; however, studies suggested that the improvement in LV systolic performance is associated with improvements also in LV relaxation and LV filling pressures [188]. Moreover, several studies reported weak agreement between response assessed by reverse remodeling observed by echocardiography, and symptomatic and functional improvement, referred as clinical response [181]. These findings suggest that further research is needed to investigate the role of diastolic function and of hemodynamic changes in context of response to CRT.
1.4
Magnetic resonance imaging
MRI sometimes referred to as cardiovascular magnetic resonance (CMR) in this context, allows comprehensive assessment of HF patients and it is now the gold standard imaging technique to assess myocardial anatomy, regional and global function, and viability [189]. CMR overcomes many of the limitations of echocardiography since it has the ability to image in any desired plane and with a nearly unrestricted field of view; the list of proposed indications for the use of CMR in HF patients is likely to further expand in the near future. CMR is regarded the gold standard with respect to accuracy and reproducibility of volumes, mass, and wall motion [189]. Because CMR yields good image quality in most patients, it is a valuable alternative modality in patients with poor echocardiographic acoustic windows.
Late gadolinium-enhanced (LGE) MRI is a well-established tool for the assessment of presence and extent of myocardial fibrosis caused by MI [190] which is another known important factor, independent from LV dyssynchrony, for CRT response prediction [191].
Using phase-contrast MRI sequences, velocity-encoded images can be generated in arbitrary orientation with flexible spatial and temporal resolution. Velocity and flow can be measured in regions of interest within blood vessels or in cardiac regions [192], similarly to Doppler imaging [193] without the limitations imposed by poor acoustic windows [194, 195].
DCE–MRI has been originally introduced for (cardiac) perfusion imaging [196]: changes in tissue (myocardial) signal intensities over time are recorded during the first passage of a suitable contrast media; ischemic areas are identified by abnormal contrast enhancement with respect to the healthy tissue, reflecting differences in contrast kinetics. While T1–weighted MRI acquisition sequences can be used to derive absolute contrast agent concentration values, in order to fully characterize the relationship between signal intensities and contrast concentration, additional calibration acquisitions are required, e.g. native T1 tissue mapping. Nevertheless, the technique provides clinical value also when semi–quantitative approaches are applied, e.g. when using of small dose of contrast media, the signal enhancement can be considered approximately linear with respect to the contrast concentration [197].
Tagged MRI consists in imposing a series of dark bands on the image, modulating tissue magnetization [198, 199]. These bands will then move together with the underlying tissue. Tagged MRI allows tracking myocardial deformation resulting in accurate strain maps on a global and regional level [200]. Although different strategies for shortening the data processing time have been proposed, such as phase demodulation methods [201, 202], the large scale clinical applicability of this technique is still limited by the necessary post–processing. Additionally, its accuracy may be affected by wall Cine MRI is widely used technique providing high myocardial tissue contrast resulting in reproducible and accurate image loops [203] where the LV wall motion can be accurately visualized and traced [204] with relatively short acquisition time. Examples of tagged and cine MRI images in short-axis view are shown in Figure 1.5. Other techniques, such as phase-contrast MRI [205], have been proposed to this end; however, none of them is part of a regular cardiac MRI examination limiting their applicability as they would require additional scan time.
Figure 1.5: Examples of cardiac MRI views from a canine model. Four-chamber view cine MRI with a green line indicating the corresponding short-axis view (A); sample frame from a short-axis cine MRI loop (B); sample frames from a short-axis tagged MRI with vertical (C) and horizontal (D) tag pattern
Drawbacks of MRI include limited availability, high costs, and low tolerance to the procedure in some patients, often because of claustrophobia. Since evidence has been reported that also patients with cardiac implanted devices can (under certain conditions) undergo MRI scans [206, 207], the same technique may be valuable also during CIED implantation follow-up.
1.5
Scope of the thesis
The presented epidemiological and economical figures shown in section 1.1.2 emphasize the clinical and societal need for research in the diagnosis, management, and treatment of HF. It is widely recognized that future strategies to contain the costs related on HF should have as primary goal the reduction of hospitalization that represents the largest part of treatment costs, and the identification of which patients are most likely to benefit from the range of therapies available [88].
Congestion is an universal finding in hospitalized AHF and it is emerging an as important prognostic factor, predictive for recurrent hospitalization [208]. Clinical evidence suggests that redistribution of fluid volumes (rather than accumulation alone) is a relevant pathophysiological process leading to AHF [42], and that it occurs before the actual acute symptomatic events [45]. Therefore, there is a clinical need for techniques enabling the assessment of the pulmonary circulation status. Existing techniques are either non-specific, invasive, or semi-quantitative, while indicator dilution methods allow direct volume quantifications.
Figure 1.6: Left to right: example of four-chamber view by DCE-MRI with overlaid regions of interest before contrast agent injection (A), during contrast agent first passage in the right ventricle (B), during first passage in left ventricle (C), corresponding indicator dilution curves for right (blue) and left ventricle (red) (D).
DCE-MRI allows minimally invasive estimation of indicator dilution parameters to provide a characterization of the trans-pulmonary hemodynamics. Example images from a DCE-MRI acquisition and corresponding IDCs is shown in Figure 1.6. The injection of a small bolus of contrast agent ensures a linear relationship between image enhancement and contrast concentra-tion, allowing a description of the contrast agent kinetics using indicator transit time distribution models. The local density random walk model, introduced by Sheppard and Savage [209] is a solution of the drift-diffusion equation. The indicator transport is interpreted as a combination of dispersion Brownian process and indicator drift. One feature of this model is that it allows the interpretation of the skewness factor of the distribution as proportional to the P´eclet number, a dimensionless measure of the relative contributions of convection and diffusion in the indicator transport. An increased contribution of diffusion has been previously interpreted as result of different mechanisms, such as increased trans-capillary exchange in edema or heterogeneous resistances to flow in different parts of the vascular bed [210].
To overcome the intrinsic limitations of MRI–based approaches, CE-US can be used for minimally-invasive measurement at the bedside, while TI would allow continuous monitor-ing in outpatient settmonitor-ings. A combination of these techniques may have the potential to provide a quantitative, noninvasive thoracic fluid content assessment. Moreover, diastolic function and cardiopulmonary hemodynamics have received little attention in the context of CRT response, despite the fact that hemodynamic improvement is used intra-operatively to optimize CRT delivery [211].
Although the relationship between mechanical dyssynchrony and CRT remains still contro-versial [212, 213], there is evidence that mechanical dyssynchrony could still play a role in the optimization of CRT delivery [156, 182], together with scar imaging to guide lead place-ment [214, 215]. Recently, researchers hypothesized that the presence of LBBB [216–218] could predict LV reverse modeling induced by CRT [150]. The 2012/2013 guidelines for patient selection included LBBB QRS morphology as ECG criterion, and long QRS duration >150 ms when LBBB morphology is absent [219]. Patients with conduction disturbances such as right bundle branch block (RBBB) [220] or interventricular dyssynchrony have been identified as non-responders to CRT [217, 221, 222]; or may even experience harm in the long-term [223]. RBBB
presents motion abnormalities between inferior and anterior LV walls, similar to abnormalities between septal and free LV walls found in LBBB [224]. However, identifying true LBBB on the ECG can be challenging as different criteria have been proposed [148]. Currently it is not clear which electrophysiogical parameter should be preferred as primary marker for selection of CRT patients [148].
The direct observation of mechanical features associated with LBBB enabled by imaging may provide additional information to electrical ECG criteria alone [225]. Most of the dyssynchrony measures proposed in the literature are positive scalar numbers that lack the ability to localize the origin of dyssynchrony [183]. As an example, the standard deviation of the time to peak myocardial tissue velocities [226] will have the same value regardless whether the most delayed segment is in the basal lateral or apical septal. Moreover, indicators taking into account mul-tiple phases of the cardiac cycle have shown higher diagnostic and predictive power for CRT response [180, 218, 227]. Finally, indices based on timing alone do not necessarily reflect the impact of dyssynchrony on the global coordination [173].
A method to analyze mechanical contraction pattern is proposed. Cine MRI is used to analyze the LV wall motion exploiting its excellent depiction of the myocardial borders [228]. The method is based on generalized cross-correlation, and it can be applied on different imaging acquisitions, e.g. radial motion and circumferential strain, to derive contraction maps and global indicators of dyssynchrony. An alternative approach based on tracking the center of the endocardial border on cine MRI image loops is also proposed; the center motion is influenced by local contraction time, but also by the local contraction amplitude and by the relative positioning of the myocardial segments.
1.6
Objective
The general objective of this thesis is to introduce novel methods for cardiovascular quantifica-tions in HF. The thesis focuses on MRI-based evaluation of: (1) the dilution process occurring in trans-pulmonary circulation, and of (2) intra-ventricular mechanical dyssynchrony. The proposed techniques were compared with alternative imaging modalities, as well as with other non-invasive techniques; the response of the proposed parameters to physiological maneuvers and to surgical interventions was also investigated.
1.7
Outline of the thesis
In this section, an overview of how the thesis contributions are organized is provided.
The repeatability, reproducibility, and reliability of the pulmonary transit time (PTT) estimation by CE-US is assessed in Chapter 2 [JP1]; the inter- and intra-observer variability introduced by regions of interest (ROIs) drawing is quantified together with the test–retest repeatability in three repeated injections.
In Chapter 3 [JP2] a method based on the use of spectral clustering for the automatic definition of ROIs to reduce the inter-observer variability is introduced. The method exploits the correla-tion between pixel-based indicator dilucorrela-tion curves belonging to the same cardiac chamber to automatically define ROIs in CE-US and DCE-MRI, after a modality dependent preprocessing. In Chapter 4 [JP3] a model-based method for the characterization of the trans-pulmonary circu-lation is presented. The method consists of a bolus injection of MRI contrast agent followed by a dynamic contrast-enhanced MRI (DCE-MRI) acquisition. Indicator dilution curves are derived in the different heart chambers; the dilution impulse response of the trans-pulmonary circulation is modeled by the local density random walk model and is estimated using system identification techniques. The dilution system is characterized by the PTT, related to the pulmonary fluid volumes, and by the skewness measure λ , a novel parameter quantifying the ratio between convection and diffusion in the dilution process.
In Chapter 5 [JP4], PTT estimates obtained by CE-US, permitting assessment at the bedside, are compared to those by DCE-MRI; PTT by CE-US is also compared with BNP levels and common cardiac performance measures.
In Chapter 6 [JP5], changes in MRI-derived thoracic fluid measures in response to acute fluid displacement achieved by a pneumatic leg compression device are investigated; thoracic fluid estimates by MRI and changes during the fluid displacement procedure are compared with those obtained non-invasively by surface trans-thoracic bio-impedance spectroscopy.
A novel method for the quantification of intra-ventricular mechanical dyssynchrony is evalu-ated in LBBB animal models and it is presented in Chapter 7 [JP6]. The method is based on three-dimensional cine-MRI image loops; regional contraction time estimates are derived from the radial endocardial motion using a cross-correlation approach. Mechanical dyssynchrony assessed by cine-MRI is compared with dyssynchrony assessed from tagged MRI, gold standard for myocardial deformation imaging, in an animal model after isolated LBBB induction. In Chapter 8 [JP7], a novel method for the quantification of LV mechanical discoordination by tracking the motion of the center of the LV endocardial border is presented. Endocardial center motion derived measures are compared in healthy subjects and LBBB patients.
In Chapter 9 [JP8], the role of cardiopulmonary function, quantified by PTT measured by CE-US, in the context of CRT is investigated. Predictive value for CRT response, based on clinical and symptomatic criteria, as well as changes during CRT follow-up are assessed.
[1] Braunschweig F, Cowie M. R, and Auricchio A. What are the costs of heart failure? Europace, 2011, 13(Suppl 2):13–17.
[2] Gheorghiade M. Reassessing treatment of acute heart failure syndromes: the ADHERE Registry. Eur. Hear. J. Suppl., 2005, 7(Suppl B):B13–B19.
[3] Stewart S, MacIntyre K, Hole D. J et al. More ’malignant’ than cancer? Five-year survival following a first admission for heart failure. Eur. J. Heart Fail., 2001, 3(3):315–322.
[4] McMurray J. J, Adamopoulos S, Anker S. D et al. Esc guidelines for the diagnosis and treatment of acute and chronic heart failure 2012. Eur. J. Heart Fail., 2012, 14(8):803–869.
[5] Ambrosy a. P, Fonarow G. C, Butler J et al. The global health and economic burden of hospitalizations for heart failure: Lessons learned from hospitalized heart failure registries. J. Am. Coll. Cardiol., 2014, 63(12):1123–1133.
[6] Laribi S, Aouba A, Nikolaou M et al. Trends in death attributed to heart failure over the past two decades in Europe. Eur. J. Heart Fail., 2012, 14(3):234–239.
[7] Dickstein K, Cohen-Solal A, Filippatos G et al. ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure 2008. Eur. Heart J., 2008, 29(19):2388–442.
[8] Mosterd A and Hoes A. W. Clinical epidemiology of heart failure. Heart, 2007, 93(9):1137–46.
[9] Gottdiener J. S, Arnold A. M, Aurigemma G. P et al. Predictors of congestive heart failure in the elderly: The cardiovascular health study. J. Am. Coll. Cardiol., 2000, 35(6):1628–1637.
[10] Senni M, Tribouilloy C. M, Rodeheffer R. J et al. Congestive heart failure in the community: trends in incidence and survival in a 10-year period. Archives of Internal Medicine, 1999, 159(1):29–34.
[11] Fang J, Mensah G. A, Croft J. B et al. Heart Failure-Related Hospitalization in the U.S., 1979 to 2004. J. Am. Coll. Cardiol., 2008, 52(6):428–434.
[12] Lloyd-Jones D, Adams R. J, Brown T. M et al. Heart disease and stroke statistics2010 update a report from the american heart association. Circulation, 2010, 121(7):e46–e215.
[13] van der Velden J, van der Wall E, and Paulus W. Heart failure with preserved ejection fraction: current status and challenges for the future. Netherlands Heart Journal, 2016, 24(4):225–226.
[14] Leening M, Siregar S, Vaartjes I et al. Heart disease in the netherlands: a quantitative update. Netherlands Heart Journal, 2014, 22(1):3–10.
[15] Lloyd-Jones D, Adams R. J, Brown T. M et al. Executive summary: Heart disease and stroke statistics-2010 update: A report from the american heart association. Circulation, 2010, 121(7).
[16] Abraham W. T, Compton S, Haas G et al. Intrathoracic Impedance vs Daily Weight Monitoring for Predicting Worsening Heart Failure Events: Results of the Fluid Accumulation Status Trial (FAST). Congest. Hear. Fail., 2011, 17(2):51–55.
[17] Liao L, Allen L. A, and Whellan D. J. Economic burden of heart failure in the elderly. Pharmacoeconomics, 2008, 26(6):447–462.
[18] Gheorghiade M, Filippatos G, De Luca L et al. Congestion in acute heart failure syndromes: an essential target of evaluation and treatment. The American journal of medicine, 2006, 119(12):3–S10.
[19] Gheorghiade M and Pang P. S. Acute Heart Failure Syndromes. JAC, 2009, 53(7):557–573.
[20] Ladapo J. A, Turakhia M. P, Ryan M. P et al. Health Care Utilization and Expenditures Associated With Remote Monitoring in Patients With Implantable Cardiac Devices. Am. J. Cardiol., 2016, 117(9):1455–1462. [21] Klersy C, De Silvestri A, Gabutti G et al. Economic impact of remote patient monitoring: An integrated
economic model derived from a meta-analysis of randomized controlled trials in heart failure. Eur. J. Heart Fail., 2011, 13(4):450–459.
[22] Anker S. D, Koehler F, and Abraham W. T. Telemedicine and remote management of patients with heart failure. Lancet, 2011, 378(9792):731–739.
[23] Hawkins N. M, Virani S. A, Sperrin M et al. Predicting heart failure decompensation using cardiac implantable electronic devices: a review of practices and challenges. Eur. J. Heart Fail., 2015.
[24] Brachmann J, B¨ohm M, Rybak K et al. Fluid status monitoring with a wireless network to reduce cardiovascular-related hospitalizations and mortality in heart failure: Rationale and design of the Op-tiLink HF Study (Optimization of Heart Failure Management using OptiVol Fluid Status Monitoring). Eur. J. Heart Fail., 2011, 13(7):796–804.
[25] B¨ohm M, Drexler H, Oswald H et al. Fluid status telemedicine alerts for heart failure: a randomized controlled trial. Eur. Heart J., 2016, 24(5):442–463.
[26] Van Veldhuisen D. J, Braunschweig F, Conraads V et al. Intrathoracic impedance monitoring, audible patient alerts, and outcome in patients with heart failure. Circulation, 2011, 124(16):1719–1726.
[27] Cleland J. G. F, Swedberg K, Follath F et al. The EuroHeart Failure survey programme - A survey on the quality of care among patients with heart failure in Europe. Part 1: Patient characteristics and diagnosis. Eur. Heart J., 2003, 24(5):442–463.
[28] Gattis W. A, O’Connor C. M, Gallup D. S et al. Predischarge initiation of carvedilol in patients hospitalized for decompensated heart failure: results of the initiation management predischarge: Process for assessment of carvedilol therapy in heart failure (impact-hf) trial. J. Am. Coll. Cardiol., 2004, 43(9):1534–1541. [29] O’Connell J. B. The economic burden of heart failure. Clin. Cardiol., 2000, 23(3 Suppl):III6–I10.
[30] Stewart S, Jenkins a, Buchan S et al. The current cost of heart failure to the National Health Service in the UK. Eur. J. Heart Fail., 2002, 4(3):361–371.
[31] Stewart S. Financial aspects of heart failure programs of care. Eur. J. Heart Fail., 2005, 7(3):423–428. [32] Yancy C. W, Lopatin M, Stevenson L. W et al. Clinical presentation, management, and in-hospital outcomes
of patients admitted with acute decompensated heart failure with preserved systolic function: A report from the Acute Decompensated Heart Failure National Registry (ADHERE) database. J. Am. Coll. Cardiol., 2006, 47(1):76–84.
[33] Fonarow G. C, Abraham W. T, Albert N. M et al. Organized program to initiate lifesaving treatment in hospitalized patients with heart failure (OPTIMIZE-HF): Rationale and design. Am. Heart J., 2004, 148(1):43–51.
[34] Gheorghiade M, De Luca L, Fonarow G. C et al. Pathophysiologic Targets in the Early Phase of Acute Heart Failure Syndromes. Am. J. Cardiol., 2005, 96(6):11–17.
[35] Adams K. F, Fonarow G. C, Emerman C. L et al. Characteristics and outcomes of patients hospitalized for heart failure in the United States: Rationale, design, and preliminary observations from the first 100,000 cases in the Acute Decompensated Heart Failure National Registry (ADHERE). Am. Heart J., 2005, 149(2):209–216.
[36] Lucas C, Johnson W, Hamilton M. A et al. Freedom from congestion predicts good survival despite previous class IV symptoms of heart failure. Am. Heart J., 2000, 140(6):840–847.
[37] Tavazzi L, Maggioni A. P, Lucci D et al. Nationwide survey on acute heart failure in cardiology ward services in italy. European heart journal, 2006, 27(10):1207–1215.
[38] Filippatos G, Leche C, Sunga R et al. Expression of fas adjacent to fibrotic foci in the failing human heart is not associated with increased apoptosis. American Journal of Physiology-Heart and Circulatory Physiology, 1999, 277(2):H445–H451.
[39] Drazner M. H, Rame J. E, Stevenson L. W et al. Prognostic importance of elevated jugular venous pressure and a third heart sound in patients with heart failure. New England Journal of Medicine, 2001, 345(8):574– 581.
[40] Cotter G, Moshkovitz Y, Milovanov O et al. Acute heart failure: A novel approach to its pathogenesis and treatment. Eur. J. Heart Fail., 2002, 4(3):227–234.
[41] Cotter G, Felker G. M, Adams K. F et al. The pathophysiology of acute heart failure-Is it all about fluid accumulation? Am. Heart J., 2008, 155(1):9–18.
[42] Cotter G, Metra M, Milo-Cotter O et al. Fluid overload in acute heart failure - Re-distribution and other mechanisms beyond fluid accumulation. Eur. J. Heart Fail., 2008, 10(2):165–169.
[43] Sabbah H. N, Rosman H, Kono T et al. On the mechanism of functional mitral regurgitation. Am. J. Cardiol., 1993, 72(14):1074–1076.
[44] Kerzner R, Gage B. F, Freedland K. E et al. Predictors of mortality in younger and older patients with heart failure and preserved or reduced left ventricular ejection fraction. Am. Heart J., 2003, 146(2):286–290. [45] Yu C. M, Wang L, Chau E et al. Intrathoracic impedance monitoring in patients with heart failure: Correlation
with fluid status and feasibility of early warning preceding hospitalization. Circulation, 2005, 112(6):841– 848.
[46] Young J. B, Abraham W, Stevenson L et al. Intravenous nesiritide vs nitroglycerin for treatment of decompensated congestive heart failure-a randomized controlled trial. Jama-Journal of the American Medical Association, 2002, 287(12):1531–1540.
[47] Shah M. R, Hasselblad V, Stinnett S. S et al. Hemodynamic profiles of advanced heart failure: association with clinical characteristics and long-term outcomes. Journal of cardiac failure, 2001, 7(2):105–113. [48] Shah M. R, Hasselblad V, Stinnett S. S et al. Dissociation between hemodynamic changes and symptom
improvement in patients with advanced congestive heart failure. Eur. J. Heart Fail., 2002, 4(3):297–304. [49] Binanay C, Califf R, Hasselblad V et al. Evaluation study of congestive heart failure and pulmonary artery
catheterization effectiveness: the escape trial. Jama, 2005, 294(13):1625–1633.
[50] Bouhemad B, Nicolas-Robin A, Benois A et al. Echocardiographic Doppler assessment of pulmonary capillary wedge pressure in surgical patients with postoperative circulatory shock and acute lung injury. Anesthesiology, 2003, 98(5):1091–100.
[51] Stevenson L. W and Perloff J. K. The limited reliability of physical signs for estimating hemodynamics in chronic heart failure. JAMA, 1989, 261(6):884–888.
[52] Remes J, MlEttinen H, Reunanen A et al. Validity of clinical diagnosis of heart failure in primary health care. European heart journal, 1991, 12(3):315–321.
[53] Picano E, Frassi F, Agricola E et al. Ultrasound lung comets: A clinically useful sign of extravascular lung water. J. Am. Soc. Echocardiogr., 2006, 19(3):356–363.
[54] Chakko S, Woska D, Martinez H et al. Clinical, radiographic, and hemodynamic correlations in chronic congestive heart failure: Conflicting results may lead to inappropriate care. Am. J. Med., 1991, 90(1):353– 359.
[55] Lichtenstein D, Meziere G, Biderman P et al. The comet-tail artifact: an ultrasound sign of alveolar-interstitial syndrome. American journal of respiratory and critical care medicine, 1997, 156(5):1640–1646.
[56] Jambrik Z, Monti S, Coppola V et al. Usefulness of ultrasound lung comets as a nonradiologic sign of extravascular lung water. Am. J. Cardiol., 2004, 93(10):1265–70.
[57] Picano E and Pellikka P. A. Ultrasound of extravascular lung water: a new standard for pulmonary congestion. European heart journal, 2016, page ehw164.
[58] Gargani L, Pang P. S, Frassi F et al. Persistent pulmonary congestion before discharge predicts rehospitaliza-tion in heart failure: a lung ultrasound study. Cardiovasc. Ultrasound, 2015, 13(1):40.
