End-product of fibrinogen is elevated in emphysematous chronic

obstructive pulmonary disease and is predictive of mortality in the

ECLIPSE cohort

Tina Manon-Jensen1,£,#, Lasse L. Langholm1,2,#, Sarah Rank Rønnow1,3, Morten Asser Karsdal1, Ruth

Tal-Singer4*, Jørgen Vestbo5*, Diana Julie Leeming1, Bruce E. Miller4 and Jannie Marie Bülow

Sand1

*The Evaluation of COPD Longitudinally to Identify Surrogate Endpoints (ECLIPSE) study,

investigators. #Contributed equally.

1Nordic Bioscience A/S, Herlev, Denmark, 22University of Copenhagen, Faculty of Health and

Medical Sciences, Department of Biomedical Sciences, Copenhagen, Denmark, 3University of

Southern Denmark, The Faculty of Health Science, Odense, Denmark, 4GSK R&D, Collegeville, PA,

USA, 5Division of Infection Immunity and Respiratory Medicine, The University of Manchester and

Manchester University NHS Foundation Trust, Manchester Academic Health Sciences Centre,

Manchester, England.

£ Corresponding author: Tina Manon-Jensen, Nordic Bioscience A/S, Herlev Hovedgade 205-207

2730 Herlev, Denmark. Tel. +45 4452 5252, Fax: +45 4452 5251, E-mail: TMJ@nordicbio.com

RUNNING TITLE: Fibrinogen processing in patients with COPD

KEYWORDS: fibrinogen, fibrin, ECLIPSE, COPD, fibrosis, D-dimer, fibrinopeptide A, exacerbations,

mortality

Word count: Abstract 310, Text 3561, Figure 4, tables 4

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

ABSTRACT

Background:

Chronic obstructive pulmonary disease (COPD) is characterized by an abnormal epithelial repair

process that may result in intra-airway accumulation of fibrin. Fibrinogen deposition and fibrin-

fibrin cross-links help strengthen the tether formation of the fibrinogen platelets, which are

essential for wound repair processes. Given that plasma fibrinogen is the only FDA qualified

biomarker that predicts mortality and COPD exacerbations, we hypothesized that changes in the

processing of fibrinogen may provide additional characterization of disease endotype and risk of

COPD progression. This was investigated by targeting neo-epitopes of fibrinogen that are released

during tissue repair.

Methods: A subpopulation of participants with COPD, (n=953) smoker (n=205) and non-smoker

controls (n=98) were included from The Evaluation of COPD Longitudinally to Identify Predictive

Surrogate End-points (ECLIPSE) cohort. Two blood biomarkers that specifically target the

thrombin-mediated conversion of fibrinogen into fibrin (FPA), and plasmin-mediated degradation

of cross-linked fibrin (D-dimer) were measured and compared with fibrinogen measurements.

Results: Plasma D-dimer had a predictive value for two-year mortality, with an adjusted hazard

ratio of 1.48 per SD (n=980; 95% Cl 1.18-1.84; p<0.0001). This was comparable to the fibrinogen

hazard ratio of 1.59 per SD (n=983; 95% Cl 1.29-1.96; p=0.0003), whereas FPA was not significantly

associated with mortality. D-dimer (p<0.001), fibrinogen (p<0.0001) and FPA (p<0.05) were

significantly elevated in participants with dyspnea (mMRC ≥ 2) as compared to participants

without dyspnea. By contrast, D-dimer was the only biomarker that was associated with

emphysema (p<0.001), and only plasma fibrinogen (p<0.05) was weakly associated with future

exacerbations.

Conclusion: There is a need for biochemical markers to characterize the heterogeneity of COPD, to

continuously improve clinical trial design and to identify individuals at risk for disease progression

to optimize treatment and early intervention. In the current study, we demonstrate that three

applied fibrinogen biomarkers provide information which reflect the distinct aspects of COPD,

rendering them potential attractive endotype biomarkers in COPD.

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

INTRODUCTION

Chronic obstructive pulmonary disease (COPD) is a heterogeneous and progressive lung disease

with different pathological processes leading to recognition of patient subgroups that may have

their own disease characteristics. Therefore, there is an immediate need for accurate and precise

biomarkers enabling identification of patients in most need of treatment, for optimal use of health

care resources, and to allow for better clinical trial design1. COPD is characterized by an

heightened airway inflammation and an abnormal epithelial repair process2 that may result in

intra-airway accumulation of fibrin causing long-term breathlessness and predisposes patients to

exacerbations, hospitalizations and premature death3,4. Further understanding of the wound

healing cascade which is central to COPD and the processing of fibrinogen may add valuable

information to the heterogeneous nature of COPD and provide better endotype-driven differential

diagnostics tools for patients, drug developers and payers of healthcare.

Wound healing plays a central role in the development of COPD. Cigarette smoke and other

injurious agents disrupt the protective epithelial barrier of the airways and induce injuries that

activate the coagulation cascade and initiate a pro-inflammatory response5,6. Patients with COPD

are repetitively exposed to injurious agents that renew these responses, causing a continued

release of molecules associated with wound healing and repair which results in lasting tissue

alterations such as emphysema and small airways fibrosis7,8. Plasma fibrinogen has been qualified

as a drug development tool by the U.S Food and Drug Administration (FDA)3 and the European

Medicines Agency (EMA)9 for use in clinical trials as a prognostic enrichment biomarker of patients

at risk for COPD exacerbation and mortality3,10. Fibrinogen processing (i.e. fibrin deposition and

fibrin-fibrin cross-links) is on the critical path for early stages of wound healing and determines the

outcome of tissue repair11 (Fig. 1). During early wound healing, platelets adhere to and are

activated by collagens at the wound site11 and in parallel, the coagulation cascade is initiated and

results in thrombin-mediated conversion of soluble fibrinogen into insoluble fibrin fibers, leading

to the release of the pro-peptides fibrinopeptides A (FPA) and B (FPB)12. These two processes

combined, form the platelet-fibrin-rich clot which is further strengthened when the fibrin-

stabilizing transglutaminase factor 13 (FXIII) generates fibrin-fibrin cross-links. Once hemostasis

has been achieved, the abundant pro-coagulatory signals that drive coagulation are promptly

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

removed to facilitate fibrinolysis which is required for the wound healing to proceed. Fibrinolysis,

the process that degrades cross-linked fibrin, is initiated when the zymogen plasminogen is

converted to plasmin, resulting in the release of D-dimer, a small fibrinolysis-specific degradation

product of fibrin that can be cleared by the liver13.

In this study, we evaluated plasma fibrinogen in addition to two novel biochemical markers

reflecting neo-epitopes of fibrinogen that are released during tissue repair. We hypothesized that

assessing fibrin deposition and fibrin-fibrin cross-links and not just plasma fibrinogen may provide

additional characterization of disease endotypes and risk of progression in COPD. We examined

serine protease-generated neo-epitopes of fibrin(ogen) with the biomarkers FPA and D-dimer, in a

subpopulation of the Evaluation of COPD Longitudinally to Identify Predictive Surrogate End-points

(ECLIPSE) cohort. These biomarkers reflect a thrombin-generated neo-epitope of fibrinopeptide A

(FPA), that allows quantification of the amount of active wound healing, and the plasmin-

generated neo-epitope of D-dimer that quantifies the absolute level of FXIII-cross-linked fibrin

which is considered accurate wound healing.

MATERIALS AND METHODS

Study population

The analysis was based on the three-year observational longitudinal study ECLIPSE

(ClinicalTrials.gov. number, NCT00292552), described previously14. The full ECLIPSE study included

2163 participants with COPD. The enrollment criteria included a forced expiratory volume in one

second (FEV1) of less than 80% of the predicted value and FEV1/forced vital capacity (FVC) ratio of

0.7 or less, assessed after the use of bronchodilators. Participants with COPD were clinically

evaluated at baseline, month three, six and subsequently every six months in three years. The

current biomarker analysis was performed on heparinized plasma samples from 1286 subjects

consisting of 983 COPD subjects, 205 smoking control participants and 98 non-smoking control

participants. Importantly, fibrinogen was measured in ethylenediaminetetraacetic acid (EDTA)

plasma15. Control participants had to be free of significant comorbidities, determined from

screening investigation, physical examination and medical history and only 95 COPD participants

were age, BMI and gender matched with the control groups. For the analyses, we used clinical

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

data obtained at baseline and available biomarker data obtained at year one. All-cause mortality

was assessed at year three. The study was conducted according to the Declaration of Helsinki and

Good Clinical Practice guidelines and was approved by relevant ethics and review boards.

Participants provided informed consent.

Biomarker measurements

Fibrin(ogen) formation and degradation products were assessed by enzyme-linked

immunosorbent assays (ELISAs) (Nordic Bioscience A/S, Herlev, Denmark). In both ELISAs, mouse

monoclonal antibodies were used to target a neo-epitope of thrombin-mediated degradation of

fibrinogen α-chain (competitive ELISA, FPA, unpublished) and a neo-epitope of plasmin-mediated

degradation of cross-linked fibrin β-chain (Sandwich ELISA, D-dimer, unpublished) (Table 1). The

neo-epitope is defined as a specific amino acid sequence that has been generated by enzymatic

cleavage of fibrin(ogen). The antibodies react only with enzymatically processed, and not with

intact fibrin(ogen).

For FPA competitive ELISA, 96-well streptavidin plate was coated with biotinylated synthetic

peptide dissolved in assay buffer (50mM Tris, 137 mM NaCl, 1% BSA, 0.05% Tween-20, 0.36%

Bronidox L5, pH 8.0) and incubated 30 minutes at 20°C. 20 µL of standard peptide or samples

diluted in assay buffer were added to appropriate wells, followed by 100 µL of monoclonal

antibody 84-23, and incubated 20 hours at 4°C. After plate wash (20 mM Tris, 50 mM NaCl, pH

7.2), 100 µL of Horseradish Peroxidase (HRP) labeled rabbit anti-mouse secondary antibody was

added and incubated at 20°C for one hour. For D-dimer sandwich ELISA, streptavidin plate was

coated with biotinylated antibody dissolved in coating buffer (50mM Tris, 137 mM NaCl, 1% BSA,

0.05% Tween-20, 0.36% Bronidox L5, pH 8.0) and incubated 30 minutes at 20°C. 20µL of standard

dilution or samples followed by 80µL incubation buffer (50mM Tris, 137 mM NaCl, 1% BSA, 0.05%

Tween-20, 0.36% Bronidox L5, 5% liquid II, pH 8.0) were added to appropriate wells and incubated

20 hours at 4°C. After plate wash, 100µL of HRP labeled antibody was added and incubated at 20°C

for one hour. For both assays, the plate was washed and added 100 µl of tetramethylbenzidine

(TMB) and incubated for 15 min at 20oC in the dark, before stopping the HRP reaction by adding

100 µl of stopping solution (1% H2SO4). Plates were read in a SpectraMax M5 (Molecular Devices,

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

CA, USA) at 450nm with 650nm as reference.

Statistics

All biomarker data were not normally distributed (D’agostino Pearson test), thus for comparison,

all data were analyzed as non-parametric. Demographics data was analyzed using Kruskal-Wallis

one-way ANOVA and chi-squared test as appropriate. The Mann-Whitney t-test and Kruskal-Wallis

testing were used for between-group biomarker comparison. Biomarker cutoffs (Youden Index

criterion) estimated using receiver operating curve (ROC) analysis was based on mortality data and

plotted using Kaplan-Meier survival curves. Plasma D-dimer data were dichotomized based on ROC

statistics. Mortality risk for groups below and above the cutoff was compared using Cox

proportional hazards analysis with or without relevant confounders as determined by univariate

Cox proportional hazards analysis. Cox proportional hazard regression analysis used to estimate

hazard ratios (HR) per 1 standard deviation (SD) change for a better comparison between the

biomarkers. HRs were adjusted for confounding factors for mortality: age, smoking history, 6-

minute walking distance (6MWD), prior hospitalizations due to COPD exacerbations and modified

Medical Research Council (mMRC) dyspnea score. Software used was GraphPad Prism version 7.00

for Windows (GraphPad Software, La Jolla California USA) and MedCalc Statistical Software version

14.8.1 (MedCalc Software bvba, Ostend, Belgium). A p-value < 0.05 was considered statistically

significant.

RESULTS

Assessment of fibrinogen turnover in participants with COPD

The basic demographic description of the population and standard clinical parameters of lung

function and disease stage are presented in Table 2. Mean age of the participants was 63.1 years

for the COPD cohort and 55.0 years and 58.9 years for the smoker and non-smoker control group,

respectively.

To investigate the change in fibrinogen turnover in COPD, we compared COPD participants with

smoker and non-smoker controls. Degradation of cross-linked fibrin, represented by the plasmin-

generated fragment D-dimer, increased in participants with COPD compared to smokers and non-

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

smokers (Fig. 2B, p<0.0001, p=0.035), which was similar to the change in fibrinogen (Fig. 2C,

p<0.0001). Notably, in age-, gender- and BMI-matched samples, D-dimer did not show any

significant differences between groups, whereas fibrinogen shows significant difference between

COPD and smoker controls (p<0.05) and non-smoker controls (p<0.0001) (data not shown).

Notably, the matched analysis included only 95 COPD participants and represented mostly less

diseased patients, as compared to the complete cohort (data not shown). The concentration of the

thrombin-generated fibrin fragment FPA, representing wound healing activation, followed the

pattern of fibrinogen, although it did not reach statistical significance (Fig. 2A).

Fibrinogen turnover related to disease phenotype (dyspnea, exacerbation, and emphysema)

To examine the underlying symptoms or endotypes related to COPD, we assessed the relationship

between fibrinogen processing and dyspnea, emphysema and exacerbations. To investigate the

change in fibrinogen turnover related to breathlessness, participants from all groups were divided

into an asymptomatic/mild (n=765) and a symptomatic group (n=475), based on a modified mMRC

dyspnea score cutoff of ≥2, as previously described16. Plasma levels of the three biomarkers of

fibrinogen turnover show significant difference between asymptomatic and symptomatic

participants. Formation of fibrin, represented by the thrombin-generated fragment FPA, increased

significantly in symptomatic participants (p=0.0282; Fig.3A). D-dimer concentration, represented

by the neo-epitope of plasmin-mediated degradation of the cross-linked fibrin beta-chain,

increased substantially corresponding to a 62 % increase in the D-dimer level of symptomatic

participants (p=0.0002; Fig.3B). Among all three biomarkers, the level of fibrinogen shows the

strongest marked difference between asymptomatic and symptomatic participants (p<0.0001,

Fig.3C). None of the biomarkers show significant difference between asymptomatic and

symptomatic participants when assessing the overall health based on a reference cutoff point of

St. George´s Respiratory Questionnaire (SGRQ) ≥25, as suggested by GOLD17 (data not shown).

To investigate the change in fibrinogen turnover related to emphysematous COPD, the COPD

subgroup was separated based on whether or not they had significant emphysema, defined as low

attenuation area at -950 Hounsfield Units (% LAA) ≥ 10% on chest computed tomography (CT)

scans. D-Dimer was the only biomarker that differed significantly between emphysematous

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

(n=613) and non-emphysematous (n=321) COPD with concentrations of 74.00 (95 % CI 46.67–

101.3) ng/mL and 48.48 (95 % CI 38.01-58.94) ng/ mL, respectively (p = 0.0002), corresponding to

a 53 % increase in the levels (Fig. 3H). Biomarkers FPA (Fig.3G) and fibrinogen (Fig.3I) did not show

any significant difference between emphysematous and non-emphysematous COPD.

To investigate change in fibrinogen turnover related to COPD exacerbations, the COPD subgroup

was divided into two groups based on if they had experienced one or more exacerbations (n=547)

or none (n=437) in the two years follow-up period after the blood-sampling. Exacerbations were

characterized as sustained worsening of respiratory symptoms, such as breathlessness or

increased sputum volume, but were self-reported by the patients. Fibrinogen levels were

significantly elevated in patients with future exacerbations (p=0.0164, Fig.3F), while FPA and D-

dimer concentrations were unchanged (FPA, p=0.1897, Fig.3D; D-dimer, p=0.5983, Fig.3E).

Fibrinogen turnover related to mortality

Next, we wanted to investigate how the fibrinogen turnover markers related to mortality, using

data from plasma fibrinogen as a reference to further explore the full fibrinogen turnover profile.

A total of 31 COPD subjects died within the three year study period (952 survivors). The biomarker

D-dimer was significantly elevated 7.7 fold (mean levels 415.7 ng/ml vs. 54.0 ng/ml) in subjects

who died compared to survivors (p=0.0018) (Fig. 4B). Plasma fibrinogen (Fig.4C), but not FPA

(Fig.4A), was also significantly elevated 1.2 fold (mean levels 457.9 ng/ml vs. 390.1 ng/ml) in non-

survivors (p=0.0012).

Fibrinogen has a published defined cutoff of 350 mg/dl for classifying patients as having high

fibrinogen levels and increased mortality and exacerbation events10. Notably, fibrinogen

concentrations in the ECLIPSE study were adjusted with -13.6% to account for the use of EDTA

plasma (the K-assay) instead of citrate plasma (Clauss method) for measuring fibrinogen based on

data provided by the manufacturer10. We estimated similar cutoff for the D-dimer using ROC curve

analysis, which resulted in an AUC of 0.66 (95%CI 0.63:0.69) for D-dimer with an associated

Youden index criterion of 32 ng/ml for predicting mortality. The corresponding AUC of fibrinogen

was 0.67 (95%CI 0.64:0.70). The biomarker FPA was excluded from these analyses as it did not

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

show any significant difference between survivors and non-survivors (Fig.4A). Kaplan-Meier plots

show a difference in survival times when subjects are separated by the dichotomized biomarker

groups for fibrinogen (Fig.4E) and D-dimer (Fig.4D). Subjects with high plasma levels of D-dimer

(≥32 ng/ml) had a significantly increased mortality risk within the two years of follow up, which is

similar to what we see with high fibrinogen (≥350 mg/dl) (Fig.4A, B). We used cox proportional

hazard regression to investigate if D-dimer was an independent predictor of mortality as indicated

by the Kaplan-Meier curve. Univariate cox regression analysis was performed to evaluate

confounders of mortality risk and here age, 6MWD, mMRC, prior hospitalizations due to COPD

exacerbations and current smoking status were identified as confounders (Table 3). Indeed, with

an adjusted hazard ratio of 1.48 per SD (n=980; 95% Cl 1.18-1.84; p<0.0001) (Table 4), D-dimer had

a predictive value for mortality which compares to plasma levels greater than or equal to 350

mg/dl for fibrinogen with an adjusted hazard ratio of 1.59 per SD (n=983; 95% Cl 1.29-1.96;

p=0.0003). We used cox regression analysis to assess the discriminatory power of a D-dimer model

for predicting death. We evaluated D-dimer alone and in an adjusted model using the

dichotomized biomarker cutoff 32 ng/ml and significant confounders (Table 3). The adjusted

model (including D-dimer, age, and current smoking status) showed D-dimer to be independently

associated with mortality, with the high biomarker group having a hazard ratio of 4.31 (95% CI

1.80-10.3) compared to the low biomarker group. No model could be obtained including

fibrinogen (using a cutoff of 350 mg/dL).

DISCUSSION

Given that fibrinogen is the only FDA and EMA qualified as a drug development tool biomarker for

predicting mortality and exacerbations in COPD3, we hypothesized that more refined assays of

fibrinogen, such as neo-epitopes of fibrinogen that are released during wound repair as activity

measures of healing may provide more specificity additional disease characteristics and enable

personalized health care in COPD. Here, we measured plasma levels of fibrinogen processing by

detecting the neo-epitope of thrombin-mediated degradation of fibrinogen α-chain, FPA, and the

neo-epitope of plasmin-mediated degradation of cross-linked fibrin β-chain, D-dimer. FPA and D-

dimer measure a specific processing of fibrinogen and may add additional biological value

compared to measuring fibrinogen alone. We found that D-dimer, a measure of ongoing fibrin

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

formation and degradation, had a predictive value for mortality, similar to fibrinogen, whereas

FPA, a measure of fibrin formation, was not significantly associated with mortality. Plasma D-

dimer, fibrinogen and FPA were significantly elevated in symptomatic participants as compared to

asymptomatic participants. In contrast, D-dimer was the only marker that was associated with

emphysema and only plasma fibrinogen was associated with exacerbations, albeit weakly.

A variety of fibrinogen assays exist18, in particular for D-dimer which utilizes antibodies to detect

different epitopes of degraded fibrinogen18. D-dimer levels in COPD subjects have previously

shown high variations across studies19–24. Our findings suggest an impaired regulation of fibrinogen

processing to be a part of COPD pathogenesis. In line with other studies25, D-dimer and fibrinogen

were increased in COPD compared to smoking and non-smoking controls. Current smoking, as

opposed to never smoking, is accompanied by increased levels of most circulating inflammatory

markers including plasma fibrinogen26–28. Unexpectedly, we found no significant difference

between smoking and non-smoking control subjects regardless of whether the subject groups

were matched for age, gender and BMI. Also, no significant difference between former and

current COPD smokers was found for any of the fibrinogen biomarkers (data not shown). Smoking

cessation may not fully attenuate the inflammatory process once COPD is developed and past

smoking may result in long-term changes in fibrinogen levels26. Notably, not all COPD participants

had a high degree of systemic inflammation defined by C-reactive protein levels despite high levels

of fibrinogen, fully supporting the notion that impaired deposition and possible fibrinolysis are

pathological events in COPD3.

The presented data not only confirm that D-dimer is elevated in COPD, but also that D-dimer could

be a marker of emphysema. Emphysema is a condition in which the alveoli are destroyed as a

result of an aberrant inflammatory reaction to irritating gases like cigarette smoke. Interestingly,

neither FPA nor intact fibrinogen were associated with emphysema indicating that fibrinogen is

not just fibrinogen and that assessing neo-epitope protein fragments of fibrinogen provide

different information to the pathogenesis of COPD, which is not provided by assessing intact

fibrinogen. As expected, we found the D-dimer concentration significantly highest in symptomatic

participants compared to asymptomatic participants, supporting D-dimer concentrations to be

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

highest during active disease, where regenerative capacity and activity of tissue turnover are

expected to be at their highest. Data derived from multiple large clinical trials indicate that the

average lung function decline appears to be higher in an earlier state of COPD compared to later

stage where large parts of the lung tissue have already been damaged29. However, FEV1 is not the

optimal tool for staging patients with COPD, as a low FEV1 can be obtained with both normal and

accelerated decline, depending the patient’s starting point30. Assessment of active wound healing

as well as tissue damage and repair may provide a more accurate picture of the current

pathological state31–33.

Plasma fibrinogen is a useful biomarker to stratify individuals with COPD into those with a high or

low risk of future exacerbations, and may be used to identify those with a higher risk of

mortality34. Exacerbations occur when COPD subjects experience that their symptoms are much

worse than usual causing a sudden accelerated decline in respiratory functions35, often leading to

hospitalizations. COPD associated with frequent exacerbations is regarded as a distinct and

persistent phenotype of COPD that is largely independent of lung function36. Contrary to intact

fibrinogen, neither FPA nor D-dimer concentrations were increased in patients experiencing

exacerbations. Our results may reflect exhaustion of COPD or a feedback mechanism

downregulating activation of fibrin formation to counteract a sudden burst in inflammation.

Contrary to our findings, elevated concentration of D-dimer has been associated with the risk of

developing acute exacerbations37. Other papers have established D-dimer as an independent risk

factor for in-hospital mortality38 and for poor long-term prognosis in patients admitted for acute

exacerbations39. Also, D-dimer has been suggested to be used as an evaluation index for the

severity of COPD with acute exacerbations40. Notably, the diagnostic efficiency of D-dimer tests is

continuously debated as D-dimer levels are conflicting. Multiple D-dimer assays exist using

antibodies which detect different epitopes of degraded fibrinogen. Contrary to the D-dimer

assessed in this study, some of the commercial kits have obvious cross-reaction to fibrinogen18,

which will affect the accuracy of the D-dimer test18.

In the present study, D-dimer was independently associated with increased risk of death, with a

hazard ratio similar to fibrinogen, while FPA was not. Furthermore, a model including

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

dichotomized D-dimer showed that high D-dimer (≥32 ng/ml) was associated with a significantly

increased risk of mortality within the two years of follow-up. Surprisingly, we did not find any

association between fibrinogen and increased mortality using the established 350 mg/dl cutoff.

This may be explained by the smaller sample size used in our study. Furthermore, for comparison,

we looked at a clinical model consisting of BODE, age and previous hospitalizations as previously

described41. The AUC of predicting mortality for this model were 0.76, which could be increased to

0.83 by including fibrinogen. This was comparable to the D-Dimer cutoff model with an AUC of

0.81 (data not shown). This indicates that D-Dimer show comparably predictive power to

fibrinogen.

A major limitation of our study is the lack of information about when blood samples were taken in

relation to time since last exacerbation. Polosa et al. showed elevated levels of D-dimer in COPD

subjects during acute exacerbation, which declined substantially when the patients reached

clinical stability42. Furthermore, the relatively small mortality rate would need a larger validation

cohort to properly evaluate the prognostic value. Likewise, the current D-dimer cutoff used for

assessing mortality needs further validation in large independent cohorts before it can be

compared to the well-defined cutoff of fibrinogen, but it does indicate that biomarkers of

fibrinogen turnover could add prognostic value.

CONCLUSION

There is an immediate need for accurate and precise biochemical markers to characterize the

heterogeneity of COPD, to identify progressors of disease, for efficient use of health care

resources, and to allow for better clinical trial design. We demonstrate that the three applied

fibrinogen markers each provide information that mark distinct aspects of COPD, making them

potential attractive endotype biomarkers. D-dimer, neo-epitope detection of plasmin-degraded

cross-linked fibrin, may serve as a biomarker for predicting disease phenotype by identifying

patients with emphysematous COPD and may show value as a prognostic marker for all-cause

mortality in COPD. Given the heterogeneity of COPD, and to enable personalized health care, it is

most likely that multiple prognostic or enrichment tools alongside plasma fibrinogen are needed

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

to facilitate drug development and to provide add-on information to the pathological wound

healing process central to COPD.

CONFLICT-OF-INTEREST DISCLOSURE: TMJ, LLL, SRR, JMBS, DJL, TMJ and MAK are employed by

Nordic Bioscience, Biomarkers and Research. MAK, DJL and TMJ hold stocks in Nordic Bioscience.

Nordic Bioscience is a privately owned small-medium sized enterprise, partly focused on the

development of biomarkers for connective tissue disorders and rheumatic diseases. None of the

authors from Nordic Bioscience received any kind of financial benefits or other bonuses for the

work described in this manuscript. RTS and BEM are employees and shareholders of GSK. JV has

received honoraria for presenting and advising from Astra Zeneca, Boehringer-Ingelheim, Chiesi,

GlaxoSmithKline and Novartis, outside the submitted work.

AUTHOR CONTRIBUTIONS.

Two representatives of GlaxoSmithKline (RTS, BEM) and one academic (JV), together representing

the ECLIPSE study investigators, developed the current study design and concept in collaboration

with representatives of Nordic Bioscience (DJL, MAK, TMJ, JMBS). They approved the plan for the

current analyses, had full access to the data, and were responsible for the decision to publish. The

study sponsor did not place any restrictions with regard to statements made in the final paper.

The biomarkers were measured by LLL and SRR. The data analysis and interpretation of data was

done by LLL, SRR and TMJ. Drafting the manuscript was done by TMJ and LLL. All authors read and

approved the final version of the manuscript, and provided interpretation of data, and revising

critically for important intellectual content.

ACKNOWLEDGEMENTS: The study was sponsored by GlaxoSmithKline; the Danish Agency for

Science, Technology and Innovation; and the Danish Research Foundation. JV is supported by the

National Institute of Health Research Manchester Biomedical Research Centre (NIHR Manchester

BRC). The authors acknowledge all participants, medical, nursing, and technical staff involved in

the ECLIPSE study.

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

REFERENCE

1. Agusti, A. et al. Treatable traits: toward precision medicine of chronic airway diseases. Eur.

Respir. J. 47, 410–9 (2016).

2. Vestbo, J. COPD: definition and phenotypes. Clin. Chest Med. 35, 1–6 (2014).

3. Miller, B. E. et al. Plasma fibrinogen qualification as a drug development tool in chronic

obstructive pulmonary disease: Perspective of the chronic obstructive pulmonary disease

biomarker qualification consortium. American Journal of Respiratory and Critical Care

Medicine (2016). doi:10.1164/rccm.201509-1722PP

4. Patel, I. S., Roberts, N. J., Lloyd-Owen, S. J., Sapsford, R. J. & Wedzicha, J. A. Airway epithelial

inflammatory responses and clinical parameters in COPD. Eur. Respir. J. 22, 94–9 (2003).

5. Petecchia, L. et al. Bronchial airway epithelial cell damage following exposure to cigarette

smoke includes disassembly of tight junction components mediated by the extracellular

signal-regulated kinase 1/2 pathway. Chest 135, 1502–1512 (2009).

6. Patel, I. S. et al. Airway epithelial inflammatory responses and clinical parameters in COPD.

(2003). doi:10.1183/09031936.03.00093703

7. Hogg, J. C. et al. The nature of small-airway obstruction in chronic obstructive pulmonary

disease. N. Engl. J. Med. 350, 2645–53 (2004).

8. Cosio, M. et al. The relations between structural changes in small airways and pulmonary-

function tests. N. Engl. J. Med. 298, 1277–81 (1978).

9. Medicines Agency, E. Product Development Scientific Support Department Qualification

opinion on Plasma Fibrinogen as a Prognostic Biomarker (Drug Development Tool) for All-

Cause Mortality and COPD Exacerbations in COPD subjects. (2018).

10. Mannino, D. M. et al. Plasma Fibrinogen as a Biomarker for Mortality and Hospitalized

Exacerbations in People with COPD. Chronic Obstr. Pulm. Dis. (Miami, Fla.) 2, 23–34 (2015).

11. Manon-Jensen, T., Kjeld, N. G. & Karsdal, M. A. Collagen-mediated hemostasis. J. Thromb.

Haemost. n/a-n/a (2016). doi:10.1111/jth.13249

12. Stalker, T. J. et al. Hierarchical organization in the hemostatic response and its relationship

to the platelet-signaling network. Blood 121, 1875–85 (2013).

13. Danese, S. et al. Inflammation and coagulation in inflammatory bowel disease: The clot

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

thickens - CME. American Journal of Gastroenterology 102, 174–186 (2007).

14. Vestbo, J. et al. Evaluation of COPD Longitudinally to Identify Predictive Surrogate End-

points (ECLIPSE). Eur. Respir. J. 31, 869–73 (2008).

15. Miller, J. et al. Comorbidity, systemic inflammation and outcomes in the ECLIPSE cohort.

Respir. Med. (2013). doi:10.1016/j.rmed.2013.05.001

16. Vogelmeier, C. F. et al. Global Strategy for the Diagnosis, Management, and Prevention of

Chronic Obstructive Lung Disease 2017 Report. GOLD Executive Summary. Am. J. Respir.

Crit. Care Med. 195, 557–582 (2017).

17. Rabe, K. F. et al. Global strategy for the diagnosis, management, and prevention of chronic

obstructive pulmonary disease: GOLD executive summary. Am. J. Respir. Crit. Care Med.

176, 532–55 (2007).

18. Dempfle, C. E., Zips, S., Ergül, H., Heene, D. L. & Fibrin Assay Comparative Trial study group.

The Fibrin Assay Comparison Trial (FACT): evaluation of 23 quantitative D-dimer assays as

basis for the development of D-dimer calibrators. FACT study group. Thromb. Haemost. 85,

671–8 (2001).

19. Alessandri, C. et al. Hypercoagulability state in patients with chronic obstructive pulmonary

disease. Chronic Obstructive Bronchitis and Haemostasis Group. Thromb. Haemost. 72,

343–6 (1994).

20. Arregui, M. A. A., Ezquerra, K. L., López, F. C. & Lacasa, R. C. [Hypercoagulability state and

endotelial injury in stable chronic obstructive pulmonary disease patients]. An. Sist. Sanit.

Navar. 33, 43–50

21. Hartmann, I. J., Hagen, P. J., Melissant, C. F., Postmus, P. E. & Prins, M. H. Diagnosing acute

pulmonary embolism: effect of chronic obstructive pulmonary disease on the performance

of D-dimer testing, ventilation/perfusion scintigraphy, spiral computed tomographic

angiography, and conventional angiography. ANTELOPE Study Group. Ad. Am. J. Respir. Crit.

Care Med. 162, 2232–7 (2000).

22. Silva, D. R., Coelho, A. C., Gazzana, M. B., Menna Barreto, S. S. & Knorst, M. M. D-dimer

levels in stable COPD patients: a case-control study. COPD 9, 426–31 (2012).

23. Akgun, M. et al. Comparison of clinical characteristics and outcomes of patients with COPD

exacerbation with or without venous thromboembolism. Respiration. 73, 428–33 (2006).

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

24. Zhang, M. et al. D-dimer as a potential biomarker for the progression of COPD. Clin. Chim.

Acta. 455, 55–9 (2016).

25. Zhang, M. et al. D-dimer as a potential biomarker for the progression of COPD. Clin. Chim.

Acta. 455, 55–9 (2016).

26. Gan, W. Q., Man, S. F. P., Senthilselvan, A. & Sin, D. D. Association between chronic

obstructive pulmonary disease and systemic inflammation: a systematic review and a meta-

analysis. Thorax 59, 574–80 (2004).

27. Levitzky, Y. S. et al. Relation of smoking status to a panel of inflammatory markers: the

framingham offspring. Atherosclerosis 201, 217–24 (2008).

28. Gan, W. Q., Man, S. F. P. & Sin, D. D. The interactions between cigarette smoking and

reduced lung function on systemic inflammation. Chest 127, 558–64 (2005).

29. Bihlet, A. R. et al. Biomarkers of extracellular matrix turnover are associated with

emphysema and eosinophilic-bronchitis in COPD. Respir. Res. (2017). doi:10.1186/s12931-

017-0509-x

30. Lange, P. et al. Lung-Function Trajectories Leading to Chronic Obstructive Pulmonary

Disease. N. Engl. J. Med. 373, 111–122 (2015).

31. Stolz, D. et al. Systemic Biomarkers of Collagen and Elastin Turnover Are Associated With

Clinically Relevant Outcomes in COPD. Chest 151, 47–59 (2017).

32. Sand, J. M. B. et al. Accelerated extracellular matrix turnover during exacerbations of COPD.

Respir. Res. (2015). doi:10.1186/s12931-015-0225-3

33. Groenewegen, K. H. et al. Increased Systemic Inflammation Is a Risk Factor for COPD

Exacerbations. Chest 133, 350–357 (2008).

34. Duvoix, A. et al. Blood fibrinogen as a biomarker of chronic obstructive pulmonary disease.

Thorax 68, 670–6 (2013).

35. Patil, S. P., Krishnan, J. A., Lechtzin, N. & Diette, G. B. In-Hospital Mortality Following Acute

Exacerbations of Chronic Obstructive Pulmonary Disease. Arch. Intern. Med. 163, 1180

(2003).

36. Hurst, J. R. et al. Susceptibility to exacerbation in chronic obstructive pulmonary disease. N.

Engl. J. Med. 363, 1128–38 (2010).

37. Ishikawa, G., Acquah, S. O., Salvatore, M. & Padilla, M. L. Elevated serum D-dimer level is

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

associated with an increased risk of acute exacerbation in interstitial lung disease. Respir.

Med. 128, 78–84 (2017).

38. Hu, G. et al. Prognostic role of D-dimer for in-hospital and 1-year mortality in exacerbations

of COPD. Int. J. Chron. Obstruct. Pulmon. Dis. 11, 2729–2736 (2016).

39. Fruchter, O., Yigla, M. & Kramer, M. R. D-dimer as a prognostic biomarker for mortality in

chronic obstructive pulmonary disease exacerbation. Am. J. Med. Sci. 349, 29–35 (2015).

40. Liu, B. H. et al. Detection and study of plasma D-dimer change in patients with acute

exacerbation of chronic obstructive pulmonary disease. J. Biol. Regul. Homeost. Agents 30,

839–845

41. Celli, B. R. et al. Inflammatory biomarkers improve clinical prediction of mortality in chronic

obstructive pulmonary disease. Am. J. Respir. Crit. Care Med. 185, 1065–1072 (2012).

42. Polosa, R. et al. Effect of acute exacerbations on circulating endothelial, clotting and

fibrinolytic markers in COPD patients. Intern. Emerg. Med. 8, 567–74 (2013).

478

479

480

481

482

483

484

485

486

487

488

489

490

491492

493

494

TABLES

Table 1. Biomarker description

Structural biomarker

Specification Measure

FPA Neo-epitope of thrombin mediated degradation of fibrinogen (competitive ELISA)

Fibrin formation. Fibrinopeptide-A released from conversion of fibrinogen into fibrin

Quantify the amount of active wound healing

D-dimer Neo-epitope of plasmin-mediated degradation of cross-linked fibrin (sandwich ELISA)

Fibrinolysis - Fibrin cross-linking

Quantify the absolute level of factor 13 transglutaminase (FXIII) generated cross-links

Fibrinogen Previously measured in the ECLIPSE study by the Kamiya (K) assay

495

496

497

498

Table 2. Main clinical characteristics of the study population by subgroup at baseline

COPD(n=983)

Smoker Controls(n=205)

Non-smoker controls (n=98)

P-value

Male sex, n (%) 627 (63.8) 85 (41.5) 50 (51.0) P<0.0001#

Age, years 64 (63-65) 53 (52-56) 59 (57-61) P < 0.001

BMI, kg/m2 26.8 (5.9) 26.6 (4.3) 28.4 (4.6) P < 0.001

FEV1, post-bronchodilator, baseline, L

1.36 (1.31-1.39) 3.07 (2.99-3.17) 3.11 (2.96-3.46) P < 0.001

FEV1, post-bronchodilator, year 1, L

1.33 (1.28-1.38) 2.95 (2.88-3.04) 3.07 (2.91-3.38) P < 0.001

FEV1, % of predicted year 1 45.8 (44.2-46.9) 101.9 (100.4-

104.5) 111.3 (107.3-115.0) P < 0.001

%LAA, year 1 14.4 (13.5-15.7) 1.5 (1.3-2.0) 4.1 (-)* P < 0.001

Current smokers, n (%) 363 (36.9) 205 (100) 0 (0) P = 0.79#

Smoking history (pack years) 43 (42-45) 26 (23-28) 0 (0) P < 0.001

mMRC 1 (1-2) 0 (0) 0 (0) P < 0.001

6MWD, mean meters (SD) 385 (378-395) - -

BODE 3 (2-3) - -

GOLD grade n (%)

2 487 (49.5) - -

3 396 (40.3) - -

4 100 (10.2) - -

Data presented as median (95% Confidence interval), unless stated otherwise. BMI, Body Mass Index; FEV1, Forced Expiratory Volume in one second; FVC, Forced Vital Capacity; LAA%, Low attenuation area; 6MWD, 6 minute walking distance; BODE, Body mass index, airflow Obstruction, Dyspnea, and Exercise capacity; GOLD, the Global Initiative for Obstructive Lung Disease classification; mMRC, modified Medical Research Council dyspnea score. *, sample size too small (n=5). One-way ANOVA analysis used to determine P-values between groups, # Chi-squared test. The P-values were considered statistically significant if P < 0.05.

499

500

Table 3. Confounding factors of mortality risk assessed by Cox regression analysis

Parameter Hazard ratio

(95% confidence interval)

P-value

Age 1.13 (1.06;1.21) 0.002

Sex 0.96 (0.46;1.99) 0.910

BMI 1.00 (0.94;1.06) 0.960

BODE 1.41 (0.97;2.05) 0.070

6MWD 1.00 (0.99;1.00) 0.021

mMRC 1.51 (1.08;2.10) 0.015

Prior exacerbations 0.98 (0.72;1.33) 0.886

Hospitalizations 1.51 (1.01;2.24) 0.044

Current smoker 0.18 (0.06;0.59) 0.005

Smoking pack years 1.01 (1.00;1.02) 0.096

BMI, body mass index; BODE, Body mass index, airflow Obstruction, Dyspnea and Exercise capacity; 6MWD, 6 minute walking distance; mMRC, modified Medical Research Council dyspnea score; Prior exacerbations, exacerbations in year prior to screening; Hospitalizations, due to exacerbations.

Table 4. Hazard ratios for biomarkers

Covariate Hazard Ratio P-ValueD-Dimer (adjusted) 1.48 per 1 SD (95% Cl 1.18-1.84) <0.0001Fibrinogen (adjusted) 1.59 per 1 SD (95% Cl 1.29-1.96) 0.0003FPA (adjusted) 0.94 per 1 SD (95% Cl 0.65-1.35) 0.728D-Dimer model (≥ 32 ng/ml):

D-Dimer AgeFormer smoker

4.31 (95%CI 1.80:10.3)1.07 (95%CI 1.00:1.15)5.56 (95%CI 1.70:18.2)

0.0010.0490.005

Fibrinogen model (≥350 mg/dl):AgemMRCFormer smoker

1.09 (95%CI 1.02:1.17)1.41 (95%CI 1.01:1.98)5.15 (95%CI 1.21:21.8)

0.0120.0440.027

95% Cl, 95% confidence interval.

501

502

503504

505

FIGURES

Figure 1: Schematic processing of fibrinogen. During early wound healing, platelets are activated and recruited at the wound site where clot formation and fibrinolysis occurs. The coagulation cascade is initiated and results in thrombin-mediated conversion of soluble fibrinogen into insoluble fibrin fibers, leading to the release of the pro-peptides fibrinopeptides A (FPA) and B (FPB). Subsequently, the insoluble fibrin is crosslinked by transglutaminase factor 13 (FXIII) to strengthen the fibrin clot formation, and once hemostasis has been achieved, the fibrinolytic system, which includes a serial of enzymes and inhibitors, degrades fibrin in order to dissolve the clot, while D-dimer is plasmin-cleaved from fibrin and can be cleared by the liver.

Figure 2. Biomarker levels in COPD and control groups. A: FPA B: D-dimer C: Fibrinogen, assessed in plasma from year 1. All graphs present median ± 95% confidence interval. Differences were considered statistically significant if p<0.05 and significance levels are displayed as: *, p≤ 0.05. ****, p≤ 0.0001.

506

507

508

509510511512513514515516517518519520521522

523524525526527528529

530

Figure 3. Symptoms and endotypes related to COPD. Symptomatic (n=475) vs asymptomatic (n=765), future exacerbations (n=547) vs no future exacerbations (n=437), and emphysema (n=613) vs no emphysema (n=321) are represented for biomarkers FPA (A,D,G), D-dimer (B,E,H), and fibrinogen (C,F,I). Symptomatic participants defined as mMRC dyspnea score ≥ 2. Exacerbations defined as 1 ≥ exacerbations recorded in the year prior to biomarker measurement. Significant emphysema defined as LAA% ≥ 10%. All graphs present median ± 95% confidence interval. Differences between groups were considered statistically significant if p<0.05 and significance levels are displayed as: *, p≤ 0.05. **, p≤ 0.01. ***, p≤ 0.001, ****, p≤ 0.0001.

531532533534535536537538539

Figure 4. Mortality. Biomarker level comparison between survivors (n=953) and non-survivor (n=30) COPD participants. A: FPA; B: D-dimer; C: Fibrinogen. Kaplan-Meier Survival curve of D-dimer (D) and fibrinogen (E) with defined cutoffs. Log rank test for comparison of survival curves; D-dimer, 32 ng/ml cutoff, P<0.0001, E: Fibrinogen. 350 mg/dl cutoff, P=0.037. All graphs present median ± 95% confidence interval. Differences between survivors and non-survivors are considered statistically significant if p<0.05 and significance levels are displayed as: **, p≤ 0.01.

540

541542543544545546547548