Prognostic and diagnotsic biomarkers linked to water and sodium regulation in acute dyspnea

(1)

Prognostic and diagnostic biomarkers linked to water and sodium regulation

in acute dyspnea

Jacob Andreas Winther, M.D.

Division of Medicine

Akershus University Hospital

And

Institute of Clinical Medicine University of Oslo

(2)

© Jacob Andreas Winther, 2019

Series of dissertations submitted to the Faculty of Medicine, University of Oslo

ISBN 978-82-8377-466-5

reproduced or transmitted, in any form or by any means, without permission.

Cover: Hanne Baadsgaard Utigard.

Print production: Reprosentralen, University of Oslo.

(3)

2

ACKNOWLEDGEMENTS

The scientific work presented in this thesis would not have been possible without considerable support from my mentors, colleagues, friends and family that I would like to recognize.

First, I want to convey a deep sense of gratitude to my mentors for providing me with the tools and the skillset required to succeed in modern research. Professor Torbjørn Omland and Helge Røsjø, both well recognized experts in the field of biomarkers, have patiently guided me through all stages of this project, helping me to push forward while maintaining a high standard of research quality. Professor Joseph G. Verbalis, one of the world’s foremost hyponatremia scientists, has also contributed with in- depth knowledge of water and sodium regulation that has enabled comprehensive interpretation of the data collected. In addition, Ivar Følling, a renowned endocrinologist and exemplary teacher, has influenced me substantially, not only as a clinical endocrinologist, but also as a researcher, greatly improving the quality of my scientific work.

My sincere gratitude also goes to the remaining co-authors of the papers presented in this thesis, including Jon Brynildsen, Pål Brekke, Arne Didrik Høiseth, Tor-Arne Hagve, Geir Christensen, Heidi Strand, Ragnhild Røysland, Mohammad Osman Pervez, Magnus Nakrem Lyngbakken, and Ståle Nygård for their work related to the ACE2 Study and insightful comments and suggestion; and to my fellow PhD students in the Cardio Thoracic Research (CRG) group, led by Professor Torbjørn Omland, for their inspiration and support during these past 6 years.

Furthermore, I want to extend my appreciation to the Department of Research at Akershus University Hospital, including Vigdis Bakkelund, Marit Jørgensen, Annika Lorentzen, and all other staff engaged in the ACE 2 study; and to the Secretary of the University of Oslo, Campus Ahus, especially Berit Lund Opheim, Vibeke S. Bjørklund, Ellen E. Westgaard, and André Øien who have helped me with the practical aspects of my academic work.

(8)

6

I am also thankful to Ingrid Nermoen and the rest of my colleagues at the Department of

Endocrinology at Akershus University Hospital, for supporting my academic efforts and at the same time enabling me to preserve and improve my skills as a clinical endocrinologist.

Finally, I want to thank my family. I am deeply grateful to my loving father Jan-Erik Winther and mother May E. Talberg who gave me the gift of curiosity that encouraged academic endeavors.

Unfortunately, my mother passed away before the completion this thesis, but I am convinced that she would have been pleased with the final result. Many family members have helped to maintain a solid foundation at home which has been greatly appreciated, but my deepest gratitude goes to Bente Dalen, the true love of my life, and our three fantastic children Sophie, Julie and Henrik who always have granted me unconditional love and support in every aspect of life.

(9)

7

ABBREVIATIONS

ACE 2 Study Akershus Cardiac Examination 2 Study

AECOPD Acute exacerbation of chronic obstructive pulmonary disease

ANP Atrial natriuretic peptide

AVP Arginine-vasopressin

BMI Body mass index

BNP B-type natriuretic peptide

CI Confidence interval

COPD Chronic obstructive pulmonary disease

CRP C-reactive protein

ED Emergency department

GOLD Global Initiative for Chronic Obstructive Lung Disease

HF Heart failure

HFpEF HF with preserved ejection fraction

HFrEF HF with reduced ejection fraction

HR Hazard ratio

Hs-TnT High-sensitivity troponin T

LVEF Left ventricular ejection fraction

NPs Natriuretic peptides

(10)

8 NIV Non-invasive ventilation

NRI Net reclassification index

NT-proBNP Amino terminal fragment of the pro-hormone of BNP

OR Odds Ratio

MR-proANP Mid-regional fragment of the pro-hormone of ANP

Q Quartile

RAAS Renin-angiotensin-aldosterone system

REC Regional Committees for Medical and Health Research Ethics

ROC Receiver operating characteristic

ROC-AUC Area under the ROC curve SD Standard deviation

SEM Standard error of the mean

(11)

9

LIST OF PAPERS

Paper I

Jacob A. Winther, Jon Brynildsen, Arne Didrik Høiseth, Ivar Følling, Brekke PH, Geir Christensen, Tor-Arne Hagve, Joseph G Verbalis, Torbjørn Omland, Helge Røsjø.

Prevalence and Prognostic Significance of Hyponatremia in Patients with Acute Exacerbation of Chronic Obstructive Pulmonary Disease: Data from the Akershus Cardiac Examination (ACE) 2 Study. PLoS One. 2016;11(8):e0161232.

Paper II

Jacob A. Winther, Jon Brynildsen, Arne Didrik Høiseth, Heidi Strand, Ivar Følling, Geir Christensen, Ståle Nygård, Helge Røsjø, Torbjørn Omland.

Prognostic and diagnostic significance of copeptin in acute exacerbation of chronic obstructive pulmonary disease and acute heart failure: data from the ACE 2 Study. Respir Res. 2017;18(1):

184.

Paper III

Mohammad Osman Pervez, Jacob A. Winther,Jon Brynildsen, Heidi Strand, Geir Christensen, Arne Didrik Høiseth, Peder L. Myhre, Ragnhild Røyslan, Magnus Nakrem Lyngbakken, Torbjørn Omland, Helge Røsjø.

Prognostic and diagnostic significance of mid-regional pro-atrial natriuretic peptide in acute exacerbation of chronic obstructive pulmonary disease and acute heart failure: data from the ACE 2 Study. Biomarkers. 2018:1-31.

(12)

(13)

10

INTRODUCTION

ACUTE DYSPNEA

Dyspnea, or difficulty of breathing, is a major symptom of cardiovascular and pulmonary disease that has been recognized since antiquity. In fact, descriptions of several types of dyspnea including asthma, tachypnea and orthopnea can be found in records by the ancient Greek physicians, collected in the Corpus Hippocrates from around 400 BC [1]. The ancient Greek physicians also understood that dyspnea frequently is related to heart and lung function, however, in those days the exact

pathophysiological mechanisms were left to speculation. For instance, it was suggested that dyspnea could be a result of phlegm from the brain that lodged in the heart and lungs. More recently, the American Thoracic Society have defined dyspnea quite broadly as “a subjective experience of breathing discomfort that consists of qualitatively distinct sensations that vary in intensity,” and that the experience of dyspnea is derived from “interactions among multiple physiological, psychological, social, and environmental factors, and may induce secondary physiological and behavioral

responses"[2]. Thus, dyspnea is a complex symptom that may have many different causes. This thesis will primarily focus on patients with acute dyspnea related to chronic obstructive pulmonary disease (COPD) and the main differential diagnosis to COPD in patients hospitalized for acute dyspnea, which is heart failure (HF).

CHRONIC OBSTRUCTIVE PULMONARY DISEASE (COPD)

History

COPD refers to a group of overlapping pulmonary conditions that lead to airflow limitation and dyspnea. These conditions include emphysema, chronic bronchitis, and chronic asthmatic bronchitis.

Descriptions of the anatomical and functional components of COPD can be found in early accounts of research concerning the anatomy of the respiratory system. For instance, the Swiss-born physician

(14)

11

Theophile Bonét described enlarged airspaces and voluminous lungs characteristic for emphysema on post-mortem examinations as early as 1679, and in 1814, the British physician Charles Badham described mucus hypersecretion and chronic cough typical for chronic bronchitis. Later on, in 1821, Dr. Rene Laënnec, the inventor of the stethoscope, also recognized that air-filled lungs frequently were accompanied by increased mucus in trachea, effectively describing the combination of emphysema and chronic bronchitis observed in COPD (Figure 1). However, emphysema and chronic bronchitis were not formally defined until the 1958 CIBA symposium that also formed the foundation for the COPD diagnosis [3].

Figure 1. Normal alveoli and bronchus (top). Bronchitis: inflamed airways filed with mucus (bottom right). Emphysema: damaged alveoli and enlarged airspace (bottom left). Used with permission of Mayo Foundation for Medical Education and Research. All rights reserved.

Definition

From the 1960’s COPD has been defined by the presences of emphysema and chronic bronchitis. With improved spirometry techniques, the obstructive airflow limitation observed in COPD was also characterized (Figure 2) and could be included in the definition [4]. Since 1998, the Global Initiative of Chronic Obstructive Lung Disease (GOLD) have worked continuously to improve the COPD

(15)

12

definition according to published research, and currently, GOLD defines COPD as “a common, preventable and treatable disease that is characterized by persistent respiratory symptoms and airflow limitation that is due to airway and/or alveolar abnormalities usually caused by significant exposure to noxious particles or gases” [5]. Furthermore, “persistent airflow limitation” is defined by

spirometry measures as non-reversible forced expiratory volume in one second (FEV1) less than 70%

of the forced vital capacity (FVC). Asthmatic disease with reversible airway obstruction related to hyper-responsiveness is considered to be a separate condition that can be differentiate from COPD by measuring FEV1 and FVC before and after administration of a bronchodilator. As transient worsening of lung function is common in COPD, it is recommended that spirometry is performed during stable- state disease. The severity of airflow limitation can be classified according to FEV1 (Table 1), however, standardized symptomatic assessment tests, such as the Modified British Medical Research Council (mMRC) questionnaire and the COPD assessment test (CAT) are considered to be more closely related to health status and mortality [6].

Figure 2. Spirometry of (A) normal control and (B) obstructive disease. Reproduced from source [5]

with permission from Global Initiative for Chronic Obstructive Lung Disease.

(16)

13

Table 1. Classification of airflow limitation in chronic obstructive pulmonary disease [5]

GOLD class FEV1 % of predicted FEV1 Severity

1 ≥ 80 Mild

2 50 - 79 Moderate

3 30 - 49 Severe

4 < 30 Very Severe

FEV1, forced expiratory volume in one second; predicted FEV1, normal FEV1 based on gender, age, and height

Etiology

Historically, two main pathophysiological theories emerged concerning the development of COPD, the British hypothesis and the Dutch hypothesis. The British hypothesis proposed that repeated chest infections and exposure to air pollution, in particular tobacco smoke could lead to COPD, while the Dutch hypothesis stated that the main contributor to the pathogenesis of COPD was an underlying genetically determined bronchial hyperactivity [3]. Today, we know that both hypotheses were at least partly correct. The main risk factor for COPD is considered to be tobacco smoking, but other

environmental factors such as biomass fuel exposure and air pollution also contribute. Exposure to these noxious particles and gasses leads to inflammation and destruction of normal lung parenchyma, small airway disease, and acceleration of the natural decline in lung function with increasing age. In addition, a variety of host factors including genetics and airway hyper-responsiveness predispose individuals to develop COPD [5].

Acute exacerbation of COPD (AECOPD)

Most COPD patients periodically experience worsening of lung function termed acute exacerbations of COPD (AECOPD). AECOPD is associated with worsening of dyspnea, cough, increased sputum production, and change in sputum colure, and is defined by GOLD as “an acute worsening of respiratory symptoms that results in additional therapy” [5]. COPD exacerbations are usually triggered by environmental factors or airway infections, and it appears that some patients are more susceptible than others. The severity of exacerbations is variable and not all patients require

(17)

14

hospitalization, however, acute exacerbation is a major cause of hospitalization among COPD patients, and a high rate of exacerbations is also associated with poor prognosis [7].

Epidemiology

COPD is a common condition associated with increased mortality and morbidity. The reported prevalence of COPD varies widely due to differences in populations and diagnostic criteria, but based on a large meta-analysis of population-based studies conducted worldwide, the estimated global prevalence of COPD among people aged 30 years or more was 11.7 % in 2010 [8]. Well-designed studies from Europe and North-America have found the prevalence of COPD to lie between 4-10%

[9]. In Norway, a prevalence of 7% was found using the GOLD criteria in a random sample of the general population in Bergen during 1996 and1997 [10]. Furthermore, the prevalence of COPD is expected to rise over the coming years as a result of aging populations in high-income countries and rising prevalence of smoking in developing countries [11]. With respect to prognosis, COPD was the third most common cause of death worldwide in 2016 according to the World Health Organization (Figure 3). In addition, the impact of COPD on mortality may be underestimated as the condition most likely is underdiagnosed, especially in developing countries. There is also a substantial economic burden associated with COPD. In the European union, the direct annual cost of COPD was estimated to 38.6 Billion Euros or about 6% of the total health budget [5]. In Norway, it was calculated that the annual cost of therapy per COPD patient in 2005 was 2.498 NOK (€284), while the total annual cost was 1.13 billion NOK (€0.13 billion) or 0.7% of the total health expenditure [12]. In addition to the direct costs of COPD, there may be a considerable indirect cost related to loss of working ability and the social burden associated with the disability that comes with increasing severity of COPD.

(18)

15

Figure 3. Top 10 causes of death according to the World Health Organization [13]

Comorbidities

A vast range of co-existing diseases have been described in COPD, including depression, osteoporosis, anorexia, anemia, diabetes mellitus, metabolic syndrome, and cardiovascular disease and heart failure [14]. While some comorbidities arise independently, others may be causally linked by shared risk factors or by one disease increasing the risk of the other. Some authors also view certain comorbidities as systemic manifestations of the inflammation that occurs in COPD [15, 16], however, the

mechanism behind such systemic consequences are currently not well understood. Symptoms of co- existing diseases such as HF can be difficult to differentiate from COPD and may be overlooked, especially in the acute setting. Nevertheless, it is of great importance to identify comorbidities as they significantly impact the course of disease and mortality rates [17].

Therapy

COPD is considered to be a progressive, but treatable disease, and several therapeutic options exist.

Pharmacological therapy, including bronchodilators (beta2-agonists, anticholinergics,

0 1 2 3 4 5 6 7 8 9 10

Ischaemic heart disease Stroke Chronic obstructive pulmonary disease Lower respiratory infections Alzheimer disease and other dementias Trachea, bronchus, lung cancers Diabetes mellitus Road injury Diarrhoeal diseases Tuberculosis

Deaths (millions)

Top 10 global causes of death in 2016

(19)

16

methylxanthines) and anti-inflammatory therapy (corticosteroids and phophodiesterase-4 inhibitors), alleviate chronic symptoms and reduce the severity of exacerbations, but no conclusive clinical trial evidence support that existing medication reduce long-term decline in lung function [5]. Pulmonary rehabilitation, including physical exercise and enhanced self-management, is also an effective therapeutic strategy to improve symptoms and quality of life [18]. Non-invasive ventilation (NIV) is standard care for patients with acute respiratory failure during COPD exacerbations, but the effect of NIV in chronic respiratory failure remains undetermined [19]. The only therapies that clearly enhance survival in COPD are smoking cessation [20] and long-term oxygen therapy for those with severe resting hypoxemia [21]. In addition, improved COPD prognosis can be achieved with pneumococcal and influenza vaccinations that decrease the incidence of lower respiratory tract infections and thus reduce the frequency and severity of exacerbations [22, 23]. Finally, management of COPD includes identification and treatment of comorbidities that also impact on prognosis [17].

HEART FAILURE (HF)

The primary differential diagnosis to AECOPD among patients hospitalized with acute dyspnea is acute heart failure. Heart failure is a clinical syndrome resulting from a structural or functional cardiac disorder that impairs the ability of the heart to fill or eject blood. Symptoms and signs of heart failure fall into one of two classes, those due to reduced cardiac output and reduced peripheral circulation (i.e.

forward failure) such as hypotension and fatigue, and those due to congestion (i.e. backward failure) leading to peripheral edema and dyspnea. The pathophysiology of congestion is complex and involves activation of several neuro-hormonal pathways and endothelial dysfunction leading to fluid retention and redistribution [24]. Importantly, both right and left heart failure contribute congestion. The main causes of HF are coronary heart disease and arterial hypertension, but the syndrome can result from a wide range of disorders of the pericardium, myocardium, heart valves, great vessels or certain

metabolic abnormalities. Notably, pulmonary disease such as COPD can also lead to heart failure as a result of pulmonary hypertension [25]. Several classification systems exist for HF that provide

complimentary information. The severity of symptoms in HF are usually classified according the New

(20)

17

York Heart Association (NYHA) Functional Classification, while the American College of Cardiology Foundation (ACCF)/American Heart Association (AHA) stages categorizes HF according to the development of structural heart disease (Table 2). Based on the fraction of blood ejected from the left ventricle (LVEF) HF can be subdivided into patients with preserved LVEF (HFpEF) or reduced LVEF (HFrEF). HF classification according to LVEF is considered to be important as the two groups have different patient demographics, comorbid conditions, prognosis, and response to therapies [26]. The ACCF/AHA defines LVEF ≤ 40% as HFrEF, LVEF 41-49% as borderline HFrEF, and LVEF ≥ 50%

as HFpEF [27]. Recently, the European Society of Cardiology (ESC) guidelines have implemented a similar classification according to LVEF that also includes additional criteria concerning the

concentration of natriuretic peptides and the presence of structural heart disease or diastolic function (Table 3).

(21)

18

Table 2. Comparison of ACCF/AHA Stages of HF [27] and NYHA Functional Classification [28]

ACCF/AHA Stages of HF NYHA Functional Classification A At high risk for HF but without structural heart

disease or symptoms of HF

None

B Structural heart disease but without signs or symptoms of HF

I Cardiac disease but without resulting limitation of physical activity. Ordinary physical activity does not cause undue fatigue, palpitation, dyspnea, or anginal pain.

C Structural heart disease with prior or current symptoms of HF

II Cardiac disease resulting in slight limitation of physical activity. They are comfortable at rest. Ordinary physical activity results in fatigue, palpitation, dyspnea, or anginal pain.

III Cardiac disease resulting in marked limitation of physical activity. They are comfortable at rest. Less than ordinary activity causes fatigue, palpitation, dyspnea, or anginal pain.

IV Cardiac disease resulting in inability to carry on any physical activity without discomfort. Symptoms of heart failure or the anginal syndrome may be present even at rest. If any physical activity is undertaken, discomfort is increased.

D Refractory HF requiring specialized interventions

ACCF indicates American College of Cardiology Foundation; AHA, American Heart Association; HF, heart failure; and NYHA, New York Heart Association.

(22)

19

Table 3. Definition of heart failure with preserved, mid-range, and reduced ejection fraction according to ESC guidelines [29]

HFrEF HFmrEF HFpEF

Criteria 1 Symptoms ± Signsâ Symptoms ± Signsâ Symptoms ± Signsâ

2 LVEF < 40 % LVEF 40-49 % LVEF ≥ 50 %

3 1. Elevated levels of natriuretic peptides ^b

2. At least one additional criterion:

a. Relevant structural heart disease (LVH and/or LAE) b. Diastolic dysfunction

1. Elevated levels of natriuretic peptides ^b

2. At least one additional criterion:

a. Relevant structural heart disease (LVH and/or LAE) b. Diastolic dysfunction

BNP, B-type natriuretic peptide; HF, heart failure; HFmrEF, heart failure with mid-range ejection fraction; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; LAE, left atrial enlargement; LVEF, left ventricular ejection fraction;

LVH, left ventricular hypertrophy; NT-proBNP, N-terminal pro-B type natriuretic peptide. ^aSigns may not be present in the early stages of HF (especially in HFpEF) and in patients treated with diuretics. ^b BNP.35 pg/ml and/or NT-proBNP.125 pg/mL.

BIOMARKERS

Definition

The term biomarker (biological marker) was defined in 2001 by The National Institute of Health Consortium as a “characteristic that is objectively measured and evaluated as an indicator of normal biological processes, pathogenic processes, or pharmacologic responses to a therapeutic

intervention” [30]. A biomarker can be recorded directly from the patient (e.g. temperature, pulse and blood pressure) or measured in a bio-sample (e.g. blood, urine or tissue), electrophysiological

recording (e.g. electrocardiogram or electromyography), or imaging technique. Thus, a wide array of measurements that indicate health or disease characteristics, can be referred to as biomarkers. In this thesis, we will explore several biomarkers measured in blood.

Applications

There are several applications of biomarkers. In the clinical setting, biomarkers can be used as a screening or diagnostic tool, to evaluate disease stage, progression, and prognosis, or to predict and monitor the effect of intervention. In research, biomarkers that closely relate to study outcomes are

(23)

20

also increasingly popular as surrogate endpoints. Finally, most biomarkers are integral parts of different biological systems, and thus, the study of biomarkers can improve our understanding of biological and pathological processes. The evaluation of any candidate clinical biomarker should involve the following: (1) demonstration of independent association with the outcome of interest, (2) prospective validation in independent cohorts representative for the target population, (3) finding incremental value to biomarkers already in clinical use, (4) evaluation of the effect on patient management and outcome, (5) cost vs. benefit, (6) ease of use and interpretation. The ideal clinical biomarker should help physicians to optimally manage patients in a timely manner. Good examples of useful biomarkers in clinical practice are circulating cardiac biomarkers, such as troponins and B-type natriuretic peptide (BNP), that have evolved into essential tools for the diagnosis and management of acute coronary syndromes and heart failure over the last decades [31]. The application of biomarkers in COPD is however, currently not clear due to lack of validation of potential candidates [32]. In this thesis, we analyze diagnostic and prognostic properties of biomarkers linked to sodium and water regulation in patients with acute dyspnea, with particular focus on AECOPD.

WATER AND SODIUM REGULATION

Water and sodium homeostasis in humans are maintained by complex neurohormonal systems that regulate thirst, renal sodium and water reabsorption, and vascular tone (Figure 4). The major hormones involved in this regulation are arginine-vasopressin (AVP), natriuretic peptides (NPs) including BNP and atrial natriuretic peptide (ANP), and the renin-angiotensin-aldosterone system (RAAS). These are vasoactive hormones that also have a significant impact on renal function.

Specifically, AVP promotes reabsorption of electrolyte free water, NPs induce sodium excretion, and the RAAS stimulate sodium reabsorption. Thus, the concentration of these hormones are key

determinants of extracellular fluid volume and sodium concentration. As pathologic changes in blood pressure, blood volume or cardiovascular load induce compensatory hormonal release of AVP, NPs, and RAAS, they also represent potential biomarkers. However, direct hormone measurements have been associated with preanalytical and analytical challenges that have limited their clinical application,

(24)

21

especially in the acute setting. On the other hand, assays of more stable metabolites have proven very useful in clinical practice. One example is N-terminal pro-B-type natriuretic peptide (NT-proBNP) that have been integrated into guidelines for the diagnosis and management of patients with suspected acute and chronic HF [27]. Interestingly, COPD is associated with water and sodium retention that cannot be explained by cardiac comorbidity [33]. Thus, changes in sodium concentration and

regulating hormones could provide useful diagnostic and prognostic information in COPD also in the absence of HF. This thesis will focus on the diagnostic and prognostic value of low sodium

concentrations, also termed hyponatremia, and increased concentrations of two emerging biomarkers that are metabolites of AVP and ANP hormone production, specifically copeptin and mid-regional pro-hormone atrial natriuretic peptide (MR-proANP).

(25)

22

Figure 4. Major neurohumoral communication systems involved in water and sodium regulation.ANP, A-type natriuretic peptide; AVP, arginine vasopressin; BNP, B-type natriuretic peptide. Reproduced with permission from source [34]. Copyright Massachusetts Medical Society.

Hyponatremia

The concentration of sodium in plasma is closely related to water homeostasis. While retention of sodium results in increased extracellular fluid volume, the concentration of sodium in plasma is primarily determined by the degree of water retention. Thus, hyponatremia indicates water retention, that in most cases is stimulated by increased levels of AVP. Hyponatremia is the most frequent electrolyte disturbance encountered in clinical practice and it is observed in a wide variety of medical disorders [35]. While the normal sodium concentration ranges between 137 and 145 mmol/L in

healthy adults [36], the cut-off defining hyponatremia varies between studies from 130 to 137 mmol/L.

In one large study of unselected patients admitted to hospital, hyponatremia defined as [Na⁺] below

(26)

23

136 mmol/L was found in 28% of subjects [37]; and several studies have reported high prevalence of hyponatremia in HF, both decompensated (20-27%) [38] and stable disease (17%) [39]. Interestingly, hyponatremia correlate with prognosis in several populations. In a cross-sectional study of adults, subjects with hyponatremia had increased risk of mortality compared to subjects with normal sodium concentration after adjustment for other risk indices [40]. A large study of unselected hospitalized patients also found low admission sodium concentration to be a strong predictor of mortality, including sodium levels within the lower reference range (137-138 mmol/L) [41]. HF patients with hyponatremia have increased short- [42] and long-term mortality [43] compared to normonatremic HF patients regardless of LVEF [38, 44, 45]. Similar to unselected hospitalized patients, it appears that sodium levels in the lower reference range as compared to the upper reference range are associated with worse prognosis in HF [44]. Strong associations between hyponatremia and mortality have also been demonstrated in other conditions, such as liver cirrhosis [46-48], pneumonia [49], and acquired immunodeficiency syndrome [50]. However, very scarce information regarding the prevalence and prognostic significance of hyponatremia in patients with COPD was available before the publication of the first paper included in this thesis.

Copeptin

Copeptin is a novel biomarker that reflects AVP secretion. While AVP is responsible for regulation of renal water reabsorption, the function of copeptin is currently unknown. The copeptin molecule is derived from the 39-amino acid carboxyl-terminal fragment of the AVP precursor molecule, pre-provasopressin (Figure 5) that is synthesized in the hypothalamus and then released from the posterior pituitary gland in equimolar amounts to AVP [51]. The plasma levels of copeptin and AVP are also well corelated to plasma osmolality and to each other [52]. Thus, it is likely that secretion of copeptin is regulated by the same mechanisms that have been established for AVP. Under normal conditions AVP secretion is regulated according to plasma osmolality by osmoreceptors in the hypothalamus, but several strong non-osmotic pathways also exist [53]. In particular, arterial under-filling, as observed during heart failure, stimulates AVP secretion via baroreceptors in the carotid sinus and the aortic arch

(27)

24

[54]. Pulmonary disorders, including COPD are also associated with elevated AVP levels, possibly due to impaired gas exchange or activation of baroreceptors, but the mechanism is not clear [55, 56].

While analytical challenges and stability issues makes reliable measurements of circulating AVP difficult to achieve [57], copeptin is easily measured and stable in plasma for at least 7 days in room temperature and over several freeze and thaw cycles [51, 58].

Copeptin concentrations have been found to be associated with prognosis in several conditions including sepsis and hemorrhagic shock [59, 60], myocardial infarction [61], chronic HF [62, 63] and lower respiratory tract infections [64]. Among patients admitted to hospital with acute dyspnea of various etiologies, copeptin was found to be a strong prognostic marker with superior accuracy compared to BNP and NT-proBNP [65], and in a study of patients with stable-state COPD copeptin predicted two-year mortality independently of selected pulmonary risk factors [66]. On the other hand, a recent multicenter study of AECOPD patients did not find any association between copeptin and a short-term (30 days) composite outcome that included mortality, transfer to the intensive care unit, or a new visit to the emergency room [67]. Thus, sparse and conflicting evidence concerning the

prognostic potential of copeptin in AECOPD existed before we published the second paper included in this thesis.

(28)

25

Figure 5. Polypeptide sequence of pre-provasopressin and single-letter amino acid sequence of copeptin. The immunoassay principle for copeptin is also shown. Numbers indicate amino acid count;

Signal, signal peptide; Tracer, labeled antibody; solid phase, antibody coated on tubes. Republished with permission of P.B. HOEBER, from source [58]; permission conveyed through Copyright Clearance Center, Inc.

Mid-regional pro-hormone atrial natriuretic peptide (MR-proANP)

The natriuretic peptide family includes several genetically distinct peptide hormones that share structural features and physiologic functions, such as stimulating natriuresis, diuresis, and vasodilatation [68]. ANP and BNP are released from myocardial cells in response to myocardial stretch [69], and both hormones are derived from inactive prohormones (proANP and proBNP) that are enzymatically cleaved into the active hormone and a N-terminal fragment (NT-proANP and NT- proBNP) [70, 71]. Subtle differences in induction, storage, and metabolization of ANP and BNP indicate that the two hormones could provide useful complementary information in selected patient groups. For instance, BNP secretion is primarily stimulated by enhanced ventricular wall stress, while ANP secretion is mainly triggered by atrial stretch and enhanced preload [72, 73]. However, as BNP and ANP are released from both atrial and ventricular cells, the source of circulating hormones in patients with cardiac dysfunction probably depends on the severity and etiology of heart disease [74].

Furthermore, some studies suggest that the production and secretion of ANP is greater in the right atrium relative to the left atrium [75-77]. In contrast to BNP, ANP is stored in granules that provides a

(29)

26

source for rapid release into circulation [78]. In addition, ANP and NT-proANP have a shorter in-vivo half-life than BNP and NT-proBNP, respectively [79, 80]. Among COPD patients, BNP

concentrations are increased during exacerbations and consistently corelates with pulmonary artery pressure and left ventricular ejection fraction [81], while ANP concentrations correlate to right ventricular end-diastolic volume [82, 83]. Interestingly, one study also found that ANP levels correlated inversely to lung function as measured by FVC among patients with COPD [84].

High-performance liquid chromatography analysis suggest that the NT-proANP molecule is subject to truncation leading to molecular heterogenicity [85]. Thus, assays based on monoclonal antibodies directed against the mid-regional part of NT-proANP (MR-proANP, Figure 6) could provide more accurate measurements compared to ANP and NT-proANP assays [86-88]. In addition, MR-proANP concentrations have been shown to be more closely correlated to left ventricular filling pressures than NT-proBNP concentrations [89], thus increments in circulating MR-proANP concentrations could identify patients with elevated ventricular filling pressures and subsequent atrial enlargements, which previously has been identified as an index of poor prognosis [90]. Accordingly, in the third paper included in this thesis, we hypothesized that MR-proANP concentrations would provide relevant diagnostic and prognostic information complementary to NT-proBNP in patients with acute dyspnea.

(30)

27

Figure 6. Polypeptide sequence of pro-atrial natriuretic peptide (proANP) and single-letter amino acid sequence of mid-regional proANP (MR-proANP). The immunoassay principle for MR-proANP is also shown. Numbers indicate amino acid count; Tracer, labeled antibody; solid phase, antibody coated on tubes. Reprinted from source [91] with permission from Elsevier.

Tracer

Solid phase

(31)

28

RATIONALE AND AIMS

RATIONALE

AECOPD and acute HF are common causes of acute dyspnea with overlapping symptoms and poor prognosis. Timely diagnosis and risk stratification are essential in order to provide adequate therapy and thereby improve survival. In this respect, biochemical biomarkers can be of great clinical use.

According to previous studies, water and sodium handling is abnormal in COPD and HF. Thus, analyzing the concentrations of sodium and stable metabolites of the hormones involved in the regulation water sodium could provide clinically useful diagnostic and prognostic information among patients with acute dyspnea. In addition, such analysis could lead to enhanced understanding of the pathophysiological mechanisms behind the abnormal sodium and water regulation in these patients.

GENERAL AIM

The general aim of the work presented in this thesis was to establish the prevalence and prognostic properties of hyponatremia in patients with AECOPD compared to patients with acute HF, and to assess the diagnostic and prognostic significance of increased concentrations of copeptin and MR- proANP compared to NT-proBNP in patients with AECOPD and acute HF.

SPECIFIC AIMS OF THE PAPERS

Paper I

To establish the prevalence and prognostic significance of hyponatremia among patients with AECOPD and acute HF.

Paper II

To establish the prognostic and diagnostic significance of copeptin compared to NT-proBNP in patients with AECOPD and acute HF.

(32)

29 Paper III

To establish the prognostic and diagnostic significance of MR-proANP compared to NT-proBNP in patients with acute AECOPD and acute HF

(33)

30

MATERIAL AND METHODS

POPULATION

The Akershus Cardiac Examination (ACE) 2 Study cohort

The Akershus Cardiac Examination (ACE) 2 Study is a single-center prospective study of biomarkers in patients admitted to hospital with acute dyspnea. It was conducted at Akershus University Hospital, a Norwegian teaching hospital that covered a catchment area with 460 000 inhabitants during the study period. Patients were eligible for study inclusion if dyspnea was considered to be the primary cause for hospitalization by the emergency department (ED) physician, they were at least 18 years of age, and less than 24 hours had passed from hospital admission to study inclusion. Exclusion criteria were disseminated malignant disease, acute myocardial infarction or coronary intervention, major surgery within the last 2 weeks, incomplete study blood sampling, hemoglobin <10 g/dL and dementia or other causes precluding informed patient consent. Dedicated study personnel, who attended all briefings in between sets, identified eligible patients in cooperation with medical interns and residents working in the ED. Patient inclusion occurred between 8 am and 2 pm Monday to Thursday from June 2009 until November 2010. The ACE 2 study was approved by the Akershus University Hospital data protection officer and the Norwegian Regional Committees for Medical and Health Research Ethics (REC) South East and conducted in agreement with the Declaration of Helsinki. All participants provided written informed consent prior to study enrolment. Previously published work based on the ACE 2 Study includes analysis of the predictive properties of secretoneurin [92] and osteoprotegerin [93], and the influence of glycosylation on the predictive properties of NT-proBNP [94] and

chromogranin A [95].

(34)

31 DATA COLLECTION

Clinical information was obtained from the ED physicians, directly from the patients, and from hospital records by dedicated study personnel. First, the ED physicians were approached at the briefings in between sets. Data was registered by the use of a standardized questionnaire that included questions on the duration of dyspnea prior to admission, clinical findings such as respiratory

frequency, peripheral edema, cyanosis, raised jugular venous pressure, and ECG pathology. The ED physicians were also asked to classify the functional capacity of all patients according to the New York Heart Association (NYHA) functional scale. Next, using a more comprehensive questionnaire, clinical information was retrieved directly from the patients. This second questionnaire included questions pertaining to clinical status prior to admission, medical history, co-morbidities, and smoking status. Previous medical history and medication were checked against medical records. Blood

pressure, heart rate, and body temperature on hospital admission were collected from patient records.

Height and weight were either collected directly from the patient or from the patient records. Body mass index (BMI) was calculated as weight (Kg)/[height (m)]². Data on left ventricular ejection fraction (LVEF) were recorded from routine echocardiography reports, and spirometry data, such as forced expiratory volume of the first second (FEV1) and forced vital capacity (FVC) were registered from spirometry reports if available. Pre-hospital capillary oxygen saturation (SpO2) was collected retrospectively from hospital or ambulance records where available. Finally, survival status was recorded from electronic hospital records, which are synchronized with Statistics Norway, until the end of follow-up November 1^st, 2012.

LABORATORY ANALYSIS

Biochemical work-up including serum sodium, potassium, glucose, and creatinine were determined on hospital admission by standard biochemical methods and the results were retrieved from the hospital’s electronic records. Hyponatremia was defined as [Na⁺]<137 mmol/L according to the local laboratory reference range based on the Nordic reference interval project [36]. To avoid registration of hypertonic

(35)

32

hyponatremia related to hyperglycemia, sodium concentrations were corrected by the Hillier formula [96]. The Hillier formula increases sodium by 2.4 mmol/L for every 5.6 mmol/L (100 mg/dL) increase in glucose above 5.6 mmol/L, which appears to provide the best overall correction. Study blood sampling for biomarkers were performed by venipuncture within 24 hours of admission.

Repeated measurements were also performed, for a subset of patients, on day 2 and before discharge.

Study blood samples were stored on ice immediately after collection and centrifuged within 1 hour.

Serum aliquots of 500 μL were frozen and stored at -80 ºC prior to biomarker analysis. NT-proBNP and high-sensitivity cardiac troponin T (hs-TnT) were measured in 2012 on the Elecsys platform by Roche Diagnostics (Basel, Switzerland). Copeptin and MR-proANP concentrations were measured in 2015 using a commercially available sandwich immunoassay based on the Kryptor system (Thermo Fisher Scientific Inc., Clinical Diagnostics, BRAHMS GmbH, 16761 Hennigsdorf, Germany). The copeptin assay had a detection limit of 0.9 pmol/L, a functional sensitivity (inter-analysis variation <

20%) above 2.0 pmol/L, and a normal reference range (2.5-97.5 percentile) of 0.9-14.9 pmol/L for healthy adults. The MR-proANP assay had a detection interval of 2.1-10000 pmol/L and a total intra- assay coefficient of variation below 4%.

ADJUDICATION OF DIAGNOSIS

The cause of dyspnea leading to the index hospitalization was established by an adjudication committee of senior physicians in a similar fashion to what has been applied in previous studies [97, 98]. The two members of the adjudication committee worked independently and reviewed all medical records, including follow-up data, and they were blinded to each other’s results. They had no

knowledge of study biomarker levels but had access to cardiac biomarker measurements that were ordered by the treating physician, including NT-proBNP and troponin T. Discordant diagnoses were resolved by consensus. Patients were classified into 3 groups according to the final diagnoses: (1) Acute HF, (2) AECOPD, and (3) Non-HF, non-COPD. AECOPD was defined according to the GOLD criteria as acute worsening of the patient’s symptoms (dyspnea, cough and/or sputum production) that was beyond day-to-day variation and that lead to a change in medication [5]. The acute heart failure

(36)

33

diagnosis was based on the European Society of Cardiology criteria, requiring typical signs and symptoms of HF and objective evidence of structural or functional myocardial abnormality [99]. In the event that a patient fulfilled the criteria of both acute HF and AECOPD, the most probable cause of the index hospitalization was diagnosed. Acute HF patients were also stratified into patients with reduced LVEF (HFrEF) or preserved LVEF (HFpEF). Patients categorized to the HFpEF group had LVEF ≥ 50%, but evidence of structural heart disease or diastolic dysfunction on echocardiography or other modalities [100].

VALIDATION

In paper I, a second cohort was analyzed to validate the prognostic effect of hyponatremia found among AECOPD patients in the ACE 2 Study cohort. The validation cohort included 99 AECOPD patients admitted to Akershus University Hospital in a previous study [101]. In this previous study all patients with assumed AECOPD by the attending physicians in the ED were eligible for inclusion.

Eligible patients that did not meet any exclusion criteria were included if the AECOPD diagnosis was confirmed. Exclusion criteria were age < 50, metastatic cancer and Eastern Cooperative Oncology Group (ECOG) performance status grade ≥ 2, neuromuscular disease with respiratory failure, and non- cooperability. The AECOPD diagnosis was confirmed by two study doctors that worked

independently of each other, if (a) the COPD diagnosis was established by a chest physician, (b) symptoms of COPD exacerbation were present at inclusion, and (c) no other condition were more likely than AECOPD to cause the presenting symptoms. Disagreements were settled by consensus.

The study was approved by REC South East, and all included patients provided written informed consent to participate. Patients were recruited from January 2005 until November 2006, i.e. four years earlier than in the ACE 2 Study, however, standard biochemical data including serum sodium were routinely measured at admission and analyzed by the same laboratory and the duration of observation was similar. Thus, admission sodiummeasurements and survival data were considered comparable between the two cohorts.

(37)

34 STATISTICAL ANALYSIS

Statistical analyses were performed using SPSS (SPSS, Armonk, NY), STATA (Stata Corp LP, TX, USA), MedCalc Statistical Software (Broekstraat, Mariakerke, Belgium), and R (R Foundation for Statistical Computing, Vienna, Austria). We considered p < 0.05 to be statistically significant.

Baseline analysis

For all three papers presented in this thesis continuous variables with normal distribution are reported as mean ± standard deviation [SD] or standard error of mean [SEM], while variables with non-normal distribution (as assessed by histograms and the Kolmogorov-Smirnov test) are presented as median with quartile [Q] 1-3. Differences between groups were compared by Student´s t test or Mann- Whitney U (Wilcoxon rank-sum) tests as appropriate. Binary data were compared by the chi-square test and are presented as absolute numbers and percentages. The concentration of sodium was analyzed as a continuous variable and as binary variable for the presences of hyponatremia or not.

Correlations between variables were explored by Spearman's rank correlation. Patient survival stratified according to biomarker concentrations at admission was illustrated by Kaplan-Meier plots and compared by the log-rank test. Positively skewed variables, including copeptin, MR-proANP, and NT-proBNP, were log transformed by the natural logarithm to approach normal distribution before regression analysis.

Regression analysis

Regression analysis were used to assess the relationship between outcome (dependent variable) and predictors including biomarkers (independent variables). Different types of regression analysis were applied depending on the outcome variable. Linear regression was used for continuous outcome variables, while logistic regression was used for binary outcome variables. In addition, Cox regression analysis, a particular type of logistic regression that also takes into account the time from inclusion to event (binary outcome), was used to analyze the association between predictors and mortality. The

(38)

35

results from regression analysis are presented as effect measures, beta coefficients for linear

regression, odds ratio (OR) for logistic regression and hazard ration (HR) for Cox regression. These measures indicate the predictive effect on the outcome for each unit increment of the predictor. For continuous variables such as sodium concentrations this means that predictive effect is calculated per mmol/L, whereas for categorical variables, like the presence of hyponatremia or not, the effect is calculated between groups. As several biomarkers were log transformed by the natural logarithm it is important to keep in mind that OR and HR for these biomarkers reflects the predictive effect on the logarithmic scale, thus, one log unit increase represents a 2.7-fold increase in the biomarker

concentration. OR and HR are presented with 95% confidence intervals (CI) that constitute the range of predictive effects that we would have obtained 95% of the times if we had repeated the same study infinitely. In other words, it is 95% certain that the “true” OR/HR of a biomarker lie within the

corresponding 95% CI for the current population. One advantage of regression analysis is the ability to adjust for confounding effects and thereby identifying independent associations by multivariable model analysis. In this respect, different methods exist for building the multivariable models. We first identified all variables associated with outcome by univariable regression analysis and then

constructed a multivariable model of independent predictors by automatic stepwise forward selection using the likelihood ratio criterion. This method was employed in all types of multivariable regression analysis included in this thesis.

Receiver operating characteristic curve analysis

The receiver operating characteristic (ROC) curve illustrates the predictive abilities of a continuous variable with respect to a binary outcome. In fact, the ROC curve is a two-way plot of the true positive rate vs. the false positive rate for increasing cut-off levels of a continuous predictor. We calculated the area under ROC curves (ROC-AUC) of different biomarkers to compare their diagnostic and

prognostic accuracies. The ROC-AUC of a biomarker represents the probability that the biomarker concentration is higher in patients with the outcome of interest (“case”) compared to patients that does not experience the outcome (“control”). Randomly guessing the outcome is equivalent to a ROC-AUC

(39)

36

of 0.5, while a perfect predictor variable that always assigns a higher value to the case vs. control would produce a value of 1.0 (Figure 7). Thus, variables with ROC-AUC closer to 1.0 are superior predictors of the outcome. It should be noted that comparing ROC-AUCs of different predictors by statistical tests produce very conservative test sizes with low statistical power (i.e. very low probability of falsely rejecting the null hypothesis), and thus, significant results can be difficult to achieve with this method [102].

Figure 7. Illustrations of different receiver operating characteristic (ROC) curves and the

corresponding area under the curve (AUC). AUC greater than 0.5 is better than randomly guessing the outcome, and a perfect predictor would have and AUC of 1.0. Reproduced with permission from source [103].

(40)

37 Net reclassification improvement

The value of adding a new biomarker to a basic clinical risk model was investigated by calculating the category free net reclassification index (NRI). This method quantifies the reclassification that is brought about by adding a new biomarker to an existing risk prediction model. A positive NRI means that the new model (with the new biomarker) classifies patients more appropriately compared to the old model with respect to the outcome.

(41)

38

SUMMARY OF RESULTS

POPULATION DATA

In total, 314 of 468 eligible patients were included in the ACE 2 Study (Figure 8). Acute HF was determined to be the primary cause of dyspnea in 143 patients, while 84 patients were diagnosed with AECOPD. Among patients with dyspnea not related to acute HF or AECOPD (n=87) the most frequent diagnoses were pneumonia (27/87), asthma (16/87), and pulmonary embolism (10/87). In the acute HF group, 64% (91/143) of the patients were classified as HFrEF, while 36% (52/143) had HFpEF. The two members of the adjudication committee reached the same diagnosis in 95%

(298/314) of the cases, while the remaining 5% (16/314) were resolved by consensus. Median time from hospitalization to adjudication of diagnosis was 464 days (Q 1-3 304-705). Overall, the baseline characteristics of acute HF and AECOPD patients were consistent with the respective diagnosis. While age, male to female ratio, and body mass index (BMI) were higher in acute HF relative to AECOPD, mortality, follow-up time, and disease severity, as reflected by NYHA functional class, were

comparable between the two groups (Table 4). After a median follow-up of 2.2 years (813 [356-996]

days), 46% (66/143) of HF patients and 42% (35/84) of AECOPD patients had died. Sodiumand MR- proANP measurements at admission were missing for one patient, thus only 313 patients were included in the respective sub-studies reported in paper I and III.

(42)

39

Figure 8. Akershus Cardiac Examination (ACE) 2 Study flowchart. COPD, chronic pulmonary obstructive disease; AECOPD, acute exacerbation of COPD; HF, heart failure. Reproduced from paper II with permission.

(43)

40

Table 4. Akershus Cardiac Examination (ACE) 2 Study baseline characteristics

AECOPD (n=84)

Acute HF (n=143)

Non-HF, non-COPD (n=87)

P^*

Age (years) 69±9 75±11 73±18 <0.001

Male sex 35 (42%) 90 (63%) 39(45%) 0.002

BMI (kg/m²) 24±6 27±6 29±9 0.005

Heart rate (beats/minute) 97±18 92±26 91±22 0.107

MAP (mm Hg) 102±18 104±21 99±16 0.546

Peripheral edema 31 (37%) 77 (54%) 16(18%) 0.014

NYHA class IV vs. II-III 47 (56%) 65 (46%) 24 (28%) 0.127

LVEF (%) 60 (50-60)^† 40 (30-55) 60 (54-60) ^† <0.001

FEV1 % of predicted 39±17 n.a ^‡ n.a ^‡

FEV1/FVC (%) 47±15 ^† n.a^‡ n.a^‡

Current smoker 28 (33%) 30 (21%) 27 (31%) 0.039

Diabetes 9 (11%) 43 (30%) 16 (18%) 0.001

Chronic heart failure 9 (11%) 87 (61%) 5 (6%) <0.001

Coronary artery disease 23 (27%) 78 (55%) 10 (12%) <0.001

Hypertension 26 (31%) 69 (48%) 25 (29%) 0.011

COPD 84 (100%) 61(43%) 10 (12%) <0.001

Beta-blocker 31 (37%) 89 (62%) 19 (22%) <0.001

ACEi/ARB 27 (32%) 87 (61%) 25 (29%) <0.001

Diuretic therapy 33 (39%) 104 (73%) 23 (27%) <0.001

Continuous variables are presented as mean ± standard deviation or median (quartile 1-3). Binary variables are presented as absolute numbers and percentages.

* P for difference between AECOPD and acute HF.

† Missing data >10%; ‡ Missing data > 85%

Abbreviations: ACEi, angiotensin-converting-enzyme inhibitor; AECOPD, Acute exacerbation of chronic obstructive pulmonary disease; ARB, angiotensin II receptor blocker; BMI, Body mass index; FEV1, forced expiratory volume in one second; FVC, forced vital capacity; HF, heart failure; LVEF, left ventricular ejection fraction; n.a., not applicable; NYHA, New York Heart Association; MAP, Mean arterial pressure.

(44)

41 PAPER I

The aim of this paper was to compare the prevalence and prognostic value of hyponatremia between AECOPD and acute HF patients. In the ACE 2 Study cohort, sodium concentrations were lower among AECOPD patients compared to acute HF patients (138.5 [135.9–140.5] vs. 139.2 [136.7–

141.3] mmol/L, p = 0.02), but the prevalence of hyponatremia was comparable between the two groups (27% [22/83] vs. 20% [29/143], p = 0.28). However, in contrast to acute HF patients, hyponatremia was not associated with mortality in AECOPD patients (Figure 9). The same pattern was evident by univariable Cox regression analysis where hyponatremia was associated with an 85%

increased risk of mortality in acute HF (HR 1.85 [95% CI 1.08, 3.16], p = 0.02), while no significant risk of mortality was found for hyponatremia in AECOPD (HR 1.00 [0.47, 2.15], p = 1.00). To

validate the negative results from the AECOPD group in the ACE 2 Study, we included a separate, but comparable cohort of AECOPD patients. In this validation cohort the prevalence of hyponatremia was comparable to that of AECOPD patients in the ACE 2 Study (25% [25/99] vs. 27% [22/83], p = 0.85), but again, hyponatremia was not associated with mortality as analyzed by survival curves or Cox regression analysis (HR 0.79 [0.42, 1.49], p = 0.46). To increase statistical power, we also analyzed the two cohorts in combination, but hyponatremia was still not associated with mortality in AECOPD by univariable Cox regression (HR 0.87 [0.53, 1.40], p = 0.53). Finally, interaction analysis of the combined cohort demonstrated that the diverging effect of hyponatremia on outcome between AECOPD and acute HF patients was statistically significant (p = 0.04).

(45)

42

Figure 9. Kaplan-Meier survival plots according to sodium concentration in (A) total ACE 2 cohort, (B) ACE 2 acute HF patients, (C) ACE 2 AECOPD patients, and (D) AECOPD validation cohort.

Reproduced from paper I with permission.

PAPER II

The main purpose of this paper was to compare the prognostic and diagnostic value of copeptin and NT-proBNP concentrations at admission among patients with AECOPD and acute HF.

Copeptin and NT-proBNP concentrations and prognosis

In the ACE 2 Study cohort, copeptin concentrations were associated with increased mortality in AECOPD and acute HF. In contrast, NT-proBNP concentrations were only associated with increased mortality in acute HF (Figure 10). After adjustment for basic clinical risk factors by multivariable Cox regression analysis, the risk of dying increased by 72% in AECOPD (HR 1.72 [1.21-2.45], p=0.003)

(46)

43

and 61% in acute HF (1.61 [1.25-2.09], p<0.001) per log (ln) unit increment of copeptin, i.e. per 2.7- fold increase in the copeptin concentration. In comparison, each log unit increase of NT-proBNP concentration increased the risk of mortality by 62% in acute HF (1.62 [1.27-2.06], p<0.001), while no significant predictive effect for mortality was found in AECOPD (1.12 [0.88-1.42], p=0.373). Among patients with dyspnea not related to COPD or HF, neither copeptin nor NT-proBNP were

independently associated with mortality. When copeptin and NT-proBNP were included in the same model, the predictive effect of copeptin was significant in patients with AECOPD (HR 1.79 [1.20- 2.66], p=0.004), but not in patients with acute HF (1.30 [0.96-1.76], p=0.091). When copeptin was added to the basic clinical risk model, the category free net reclassification index (NRI) was positive in AECOPD (NRI 0.60 [0.19-1.02], p = 0.004) and acute HF (0.39 [0.06-0.71], p=0.020). In AECOPD, the predicted risk of mortality decreased in 67% of survivors and increased in 63% of non-survivors when copeptin was included in the risk model (Figure 11). By ROC-AUC analysis, we could not find any statistical difference in the prognostic accuracy between copeptin and NT-proBNP in AECOPD (ROC-AUC 0.67 [0.55-0.79] vs. 0.56 [0.44-0.69], p=0.111) or acute HF (0.66 [0.57-0.75] vs. 0.67 [0.58-0.76], p=0.695).

(47)

44

Figure 10. Kaplan-Meier survival plots stratified by biomarker concentration quartiles for (A)

copeptin in acute HF, (B) Copeptin in AECOPD, (C) NT-proBNP in acute HF, and (D) NT-proBNP in AECOPD. Reproduced from paper II with permission.

Logrank:

Overall p < 0.0001 Trend p = 0.0002

Range (pmol/L):

Q1 2.3 − 10.2 Q2 10.3 − 22.2 Q3 23.0 − 47.9 Q4 48.7 − 270.5

0.250.500.751.00

35 16 12 3 0

Q4 36 29 25 14 1

Q3

36 30 26 16 1

Q2

36 32 27 18 0

Q1 Number at risk

0 300 600 900 1200

Days

A Copeptin in acute HF

Logrank:

Overall p < 0.006 Trend p < 0.003

Range (pmol/mL):

Q1 1.3 − 5.2 Q2 6.5 − 8.4 Q3 9.1 − 19.3 Q4 19.8 − 131.5

0.250.500.751.00

21 13 11 5 0

Q4 21 18 16 7 2

Q3

21 17 17 15 0

Q2

21 20 17 12 0

Q1 Number at risk

0 300 600 900 1200

Days

B Copeptin in AECOPD

Logrank:

Overall p < 0.0001 Trend p < 0.0001

Range (pg/mL):

Q1 95 − 1601 Q2 1686 − 3600 Q3 3772 − 8396 Q4 8428 − 80733

0.250.500.751.00

35 18 14 7 1

Q4 36 26 20 8 0

Q3

36 33 31 16 1

Q2

36 30 25 20 0

Q1 Number at risk

0 300 600 900 1200

Days

C NT−proBNP in acute HF

Logrank:

Overall p = 0.305 Trend p = 0.110

Range (pg/mL):

Q1 40 − 171 Q2 172 − 367 Q3 391 − 1002 Q4 1013 − 16635

0.250.500.751.00

21 15 15 10 0

Q4 21 17 13 5 0

Q3

21 17 15 10 1

Q2

21 19 18 14 1

Q1 Number at risk

0 300 600 900 1200

Days

D NT−proBNP in AECOPD

Quartiles (Q): 1 2 3 4

(48)

45

Figure 11. Risk reclassification among AECOPD patients. Estimated risk of death for each patient by the basic model (x-axis) and the enhanced model that also included copeptin (y-axis). "Reclassified up" or "reclassified down" represents an increased or decreased risk of death predicted by the enhanced model as compared to the basic model. The prediction model is improved when survivors are reclassified down, and non-survivors are reclassified up. Reproduced from paper II with

permission.

Copeptin and NT-proBNP concentrations and diagnosis

The median copeptin concentration was significantly higher among acute HF patients compared to patients with AECOPD and patients with other causes of dyspnea (22.2 [10.2-47.9] vs. 8.8 [5.2-19.7]

and 8.3 [4.3-18.2] pmol/L, respectively). However, NT-proBNP discriminated acute HF from non-HF related dyspnea more accurately than copeptin (AUC 0.85 [95% CI 0.81-0.89] vs. 0.71 [0.66-0.77], p<0.0001). We did not find any significant difference in the concentration of copeptin or NT-proBNP between patients with AECOPD and patients with other non-HF causes of dyspnea.

0.0 0.2 0.4 0.6 0.8 1.0

0.00.20.40.60.81.0

Survivors

Basic risk model

Basic risk model + lnCopeptin

Reclassified down (67%) Reclassified up (33%)

0.0 0.2 0.4 0.6 0.8 1.0

0.00.20.40.60.81.0

Non−survivors

Basic risk model

Basic risk model + lnCopeptin

Reclassified up (63%) Reclassified down (37%)

(49)

46 PAPER III

In this paper we hypothesized that MR-proANP concentrations would provide relevant prognostic and diagnostic information beyond NT-proBNP in patients with acute dyspnea.

MR-proANP concentrations and prognosis

Across the total ACE 2 study cohort, MR-proANP concentrations at admission were closely related to NT-proBNP concentration (r=0.89, p<0.001), LVEF (r=-0.40, p<0.001), and mortality (Figure 12).

After adjusting for other variables associated with outcome in a multivariable Cox regression model, the risk of dying in the acute HF group increased by 88% for MR-proANP and 46% for NT-proBNP per log unit increment (HR 1.88 [95% CI 1.18-3.01], p=0.01 and 1.46[1.14-1.88], p=0.003;

respectively). No significant difference was found between MR-proANP and NT-proBNP in the accuracy of predicting mortality in acute HF (ROC-AUC 0.66 [0.56-0.75] vs 0.67 [0.58-0.76]).

Among AECOPD patients, the concentration of MR-proANP was associated with increased mortality, while NT-proBNP was not (Figure 13). Every log unit increment of the MR-proANP concentration increased the risk of dying by 98% in the adjusted model (HR 1.98 [1.17-3.34], p=0.01). The optimal MR-proANP concentration cut-off for predicting mortality according to the Youden index was 86 pmol/L.

Prognostic and diagnotsic biomarkers linked to water and sodium regulation in acute dyspnea

Prognostic and diagnostic biomarkers linked to water and sodium regulation

in acute dyspnea

TABLE OF CONTENTS

ACKNOWLEDGEMENTS

ABBREVIATIONS

LIST OF PAPERS

INTRODUCTION

Top 10 global causes of death in 2016

RATIONALE AND AIMS

MATERIAL AND METHODS

SUMMARY OF RESULTS