Eosinophil-guided corticosteroid therapy in hospitalised patients with COPD exacerbation

(CORTICOsteroid Reduction in COPD (CORTICO-COP)):A randomised controlled multicentre investigator-initiated trial

1

1

1

23

4

5

6

7

8

9

1011

2

3

Summary

Background: Systemic corticosteroid exposure in patients with acute exacerbation of COPD

(AECOPD) is associated with debilitating adverse effects. Therefore, systemic corticosteroids

reducing strategies are urgently needed and may be offered by a personalised medicine biomarker-

guided approach. The aim of the study was to determine whether an algorithm based on blood

eosinophils could safely reduce systemic corticosteroid exposure in hospitalised patients with

AECOPD.

Methods: We did a prospective multicentre, investigator-initiated, randomised, controlled open-

label trial in respiratory departments in three different University-affiliated Hospitals in Denmark.

AECOPD patients aged at least 40 years, with known airflow limitation (defined as

postbronchodilator FEV1/FVC ≤ 0·70) and a specialist verified diagnosis of COPD, and who was

decided to start on systemic corticosteroids by the respiratory medicine physician on duty, were

assigned (1:1), using a computer-generated sequence stratified according to the site and age (> 70

years vs. ≤ 70 years) to either the eosinophil-guided group or the control-group within 24-h of

admission. Both investigators and patients were aware of group assignment. All patients received

80 mg of intravenous methylprednisolone on the first day. The eosinophil-guided group were from

the second day given 37·5 mg of prednisolone tablet daily (maximum of 4 days in all) on days when

the blood eosinophil count was ≥ 0·3 x 109 cells/l. On days with eosinophil count < 0·3 x 109 cells/l,

systemic corticosteroid treatment was not administered. If a patient was discharged during the

treatment period, a treatment based on the last measured eosinophil count was prescribed for the

remaining days within the 5-day period (last observation carried forward). The control group

received 37·5 mg of prednisolone tablet daily from the second day and on a daily basis for 4 days.

The primary outcome was days alive and out of hospital within 14 days after recruitment. All

analyses were done by intention to treat. This non-inferiority trial was registered at

ClinicalTrials.gov, number NCT02857842, and was completed in January 2019.

2

4

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

5

6

Findings: Between August 3, 2016, and September 30, 2018, 159 patients in the eosinophil-guided

group and 159 patients in the control-group were included in the analyses. For days alive and out of

hospital within 14 days after recruitment, no significant difference was recorded: mean 8·9 days

[95% CI 8·3–9·6] in the eosinophil-guided group vs. 9·3 days [95% CI 8·7–9·9] in the control-

group (between-group absolute difference -0·4 [95% CI -1·3–0·5], p=0·34) according to the

intention-to-treat analysis, and mean 9·1 [95% CI 8·4 to 9·8] in the eosinophil-guided group vs. 9·4

[95% CI 8·8 to 10·0] in the control group; between-group absolute difference -0·3 [95% CI -1·2 to

0·7]; p = 0·49) in the per-protocol analysis. Treatment failure at 30 days was seen in 42 (26·4%)

and 41 patients (25·8%) (p=0·90) and number of deaths at 30 days were 9 (5·7%) and 6 (3·8%)

(p=0·43) in the two groups, respectively. Median duration of the systemic corticosteroid therapy

was 2 days [IQR 1·0 to 3·0] in the eosinophil-guided group and 5 days [IQR 5·0 to 5·0] in the

control-group, p<0·0001.

Interpretation: Eosinophil-guided therapy substantially reduced the duration of systemic

corticosteroid exposure in hospitalised patients with AECOPD with a similar treatment response.

Eosinophil-guided therapy seems to be an promising treatment option in hospitalised patients with

AECOPD.

Funding: The Danish Regions Medical Fund and The Danish Council for Independent Research.

Keywords: systemic corticosteroids, COPD, exacerbations, eosinophil count, eosinophil-guided

corticosteroid therapy, randomised controlled trial, biomarker.

Word count:

3

7

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

8

9

Introduction

Acute exacerbations of chronic obstructive pulmonary disease (AECOPD) contribute substantially

to high morbidity and mortality and poor quality of life worldwide.1,2 Orally administered systemic

corticosteroids are frequently given to treat AECOPDs to improve recovery from symptoms and

prevent treatment failure. However, systemic corticosteroids have no effect on long-term decline in

lung function, re-exacerbation of COPD after the first month, the length of hospital stay at the

intensive care unit (ICU) or mortality.3 Evidence regarding the optimal dosage and treatment

duration is inadequate, and physicians may be reluctant to shorten the length of the treatment.3-5

Systemic corticosteroid over-use should be avoided as this may cause serious harm to the patient,

including osteoporotic fractures, adrenal insufficiency, sepsis and venous thromboembolism.6-8 The

increasing knowledge of these side effects along with the recognition of COPD as a heterogeneous

disease and the knowledge that systemic corticosteroids are largely ineffective in reducing

neutrophil-associated inflammation has led to increasing interest in a more targeted approach to

systemic corticosteroid treatment.9,10

Eosinophilic inflammation has been demonstrated in 20–40% of AECOPD patients.11-13 There is

now evidence to support the use of peripheral blood eosinophil counts as a diagnostic biomarker to

define an eosinophilic COPD phenotype.13,14 An eosinophilic phenotype based on a peripheral blood

eosinophil count of ≥ 300 cells/μl is associated with an increased risk of AECOPD15,16, and patients

with higher peripheral blood eosinophil count are more likely to benefit from the treatment with

inhaled corticosteroids (ICS) and systemic corticosteroids.17-19 Therefore, this biomarker has been

proposed as a useful tool to guide the systemic corticosteroid treatment in AECOPD.14,17,20 To date,

only one randomised eosinophil-guided trial used blood eosinophils to direct systemic

corticosteroids in moderate exacerbations of COPD and reached non-inferiority in treatment failure

rates in biomarker-directed treatment (blood eosinophil count ≥ 2%).17 Additionally, increased harm

in patients with a low blood eosinophil count (< 2%) who received systemic corticosteroid

treatment was observed.17 Also, a number of studies demonstrated that in-hospital outcomes 4

10

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

8411

12

including hospital stay, mechanical ventilation, ICU stay and ICU mortality following severe

AECOPD have more favorable response to systemic corticosteroids in patients with higher blood

eosinophil count.18,21,22 In daily clinical practice, administration of systemic corticosteroids seems to

lead to a drop in blood eosinophils, and the clinical improvement hereafter indicates that systemic

corticosteroids may be able to reduce eosinophilic inflammation. However, it is not known whether

day to day adjustments of systemic corticosteroids based on eosinophil counts can sufficiently

reduce eosinophil inflammation in patients admitted to hospital with AECOPD.

The hypothesis of this trial is that an eosinophil-guided reduction of the dose of systemic

corticosteroids for hospitalised patients with AECOPD, compared to” standard care” does not lead

to inferior treatment effect in regard to “days alive and out of hospital within 14 days” after

recruitment.

Methods

Study design

The CORTICOsteroid reduction in COPD (CORTICO-COP) study was a multicentre investigator-

initiated randomised controlled open-label non-inferiority trial involving respiratory departments at

three University Hospitals in Denmark. Patients were recruited and followed up between August

2016 and January 2019 (last patient was recruited on September 30, 2018, and followed up to

January 8, 2019).

The trial was carried out in accordance with the published trial protocol,23 the ICH-GCP guideline24,

applicable government requirements, the Helsinki Declaration25 and the requirements of the

CONSORT guideline and check list.26 The study was registered at clinicaltrials.gov using the

identifier NCT02857842, and the full protocol is included in the appendix23; both the protocol and

the statistical analysis plan were made available on our website (www.coptrin.dk) before the

recruitment stopped. This study has been approved by the Ethics Committees of all participating

sides (H-15012207), the Danish Medicines Agency (EudraCT no: 2015-003441-26) and the Danish

5

13

85

86

87

88

89

90

91

92

93

94

95

9697

98

99

100

101

102

103

104

105

106

107

108

109

110

14

15

Data Protection Agency (HGH-2015-038 and I-Suite number 04014). It was monitored according to

Good Clinical Practice (GCP) by the GCP unit of the Capital Region of Denmark. No financial

incentive was provided to the investigators or participants.

Participants

All consecutive patients admitted to the participating sites were eligible if included within 24 h of

admission; were aged ≥ 40 years; had spirometry-verified airflow limitation (defined as

postbronchodilator FEV1/FVC ≤ 0·70) and a specialist verified diagnosis of COPD based on data

from stable disease state. In accordance with our National guidelines, all AECOPD patients had

systemic corticosteroids administered, upon decision of the respiratory medicine physician on duty,

shortly after admission, and thus the patients who were recruited for the trial had all received the

first dose. Thus eosinophil-guided strategy was solely applied on days 2 through 5. Exacerbations

were defined according to the consensus definition, stated by the Global Initiative for Chronic

Obstructive Lung Disease committee: An acute worsening of respiratory symptoms that results in

additional therapy.27 Patients for whom the primary reason for admittance was not AECOPD,

including cardiovascular disease, were not included in the study. The baseline date which is also

the calendar date of recruitment in the trial was named “day 1”.

Patients who met the inclusion criteria were invited to participate in the trial. Exclusion criteria

included the following: (1) Self-reported or physician-diagnosed asthma, (2) Life expectancy of less

than 30 days, (3) Severe COPD exacerbation requiring invasive ventilation or admission to the ICU,

(4) Allergy to systemic corticosteroids, (5) Severe mental illness that cannot be controlled by

medication, (6) People detained under the act on the use of coercion in psychiatry, (7) Severe

language difficulties or the inability to provide a written informed consent, (8) Pregnancy and

lactation, (9) Systemic fungal infections, or (10) Patients receiving above 10 mg of chronic systemic

corticosteroids daily. The written informed consent was obtained from patients prior to

randomisation. Patients could only participate in the trial once.

6

16

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

17

18

Randomisation

Patients were randomly assigned in a 1:1 ratio to receive either treatment according to eosinophil

guidance based on the daily peripheral blood eosinophil count (intervention group) or the standard

of care (control group). The random sequence was generated using the Sealed Envelop sequence

generator stratified according to the site (respiratory departments in three different University-

affiliated Hospitals) and age (> 70 years vs. ≤ 70 years). Online inclusion of patients according to

the concealed sequence was performed with an independent, centralised 24h-available, web-based

system at https://www.sealedenvelope.com. The block sizes varied, and nobody has access to the

sequence. Both the investigators and patients were aware of the treatment allocation. After

randomisation, the principal investigator was informed by a system generated email about the

treatment assignment. Then, the investigator informed the research nurses when, and with whom,

they had to do follow-up.

Procedures

Patients were randomised to one of the two treatment arms: a) Intervention group: 80 mg of

intravenous methylprednisolone on the first day followed by 37·5 mg of prednisolone tablet daily

(maximum of 4 days in all) on days when the blood eosinophil count was ≥ 0·3 x 109 cells/l. On

days with eosinophil count < 0·3 x 109 cells/l, systemic corticosteroid treatment was not

administered. If a patient was discharged during the treatment period, a treatment based on the last

measured eosinophil count was prescribed for the remaining days within the 5-day period (last

observation carried forward) OR b) Control group: 80 mg of intravenous methylprednisolone on the

first day followed by 37·5 mg of prednisolone tablets on a daily basis for 4 days.

The study did not interfere with the decision of the physician to commence systemic corticosteroids,

since this decision was made before the patients could be recruited for the trial. Treating physicians

were told always to contact the investigator if she or he wished to deviate from the protocol

algorithm. This approach was used in order to reduce non-adherence to the protocol. The decision 7

19

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

20

21

to discharge was made by the treating physician and sites were encouraged to not involve

investigators in the decision to discharge a patient included in the trial. Baseline measurements were

obtained on the calendar date of recruitment. For patients randomly assigned to the eosinophil-

guided group, measurements of blood eosinophil count were carried out once a day and made

available to the attending physicians. Blood eosinophil count was only measured in the control

group on the day of inclusion in order to avoid spill-over treatment switching.

Increased dyspnoea, increased sputum volume, increased sputum purulence and cough were

registered at recruitment. Blood glucose measurements were done in a standardised manner, every

day for five days, at the first blood sampling round in the morning, and was thus not influenced by

the investigator.

The COPD Assesment Test (CAT) score, spirometry (forced expiratory volume in 1 second (FEV1),

forced vital capacity (FVC), and FEV1/ FVC), Body Mass Index (BMI) and the level of dyspnoea

using the Medical Research Council (MRC) Dyspnoea Scale were assessed at baseline (also day 3

for spirometry), and on 30 and 90 days after discharge.

Outcomes and follow-up

The primary efficacy outcome was DAOH within 14 days after recruitment (non-inferiority

approach). The primary outcome measure was chosen to summarize “good outcome” (alive and out

of hospital) during the first 14 days in opposition to “bad outcomes” (dead or still admitted to

hospital).28 Among other advantages, lead-time bias due to death will be avoided using this outcome

measure (i.e. patients who die early will not be counted as short length of stay).

The secondary endpoints were accounting from baseline: i) Treatment failure at day 30 (recurrence

of AECOPD resulting in emergency room visits, hospitalisation or need to intensify

pharmacological treatment), ii) Readmission with AECOPD or death within 30 days, iii) Time to

readmission with AECOPD or death within 30 days, iv) Cumulative corticosteroid dose during

hospitalisation, on day 30 and 90, v) Mortality rate at day 30, vi) Hyperglycaemia during index 8

22

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

23

24

admission (≥ 7·0 mmol/l), and vii) A new onset of diabetes mellitus at day 30 (HbA1c ≥ 48

mmol/mol), viii) Worsening of diabetes mellitus at day 30 (any increase in HbA1c). The following

endpoints were assessed at day 90 after recruitment : i) all infections requiring antibiotic treatment,

ii) changes in PTH and vitamin D status (appendix p 39) and iii) dyspepsia, ulcer complication or

initiation of new proton-pump inhibitor treatment. Following endpoints were assessed on day 30

and 90 from baseline: Spirometric changes (also at day 3) and change in BMI, CAT and MRC

score. All patients were planned for outpatient visits at day 30 and day 90. If the patients did not

attend the follow-up visits (if they found it too strenuous), our project nurses visited them at home.

If patients were discharged before day 5, we made a phone call on day 5 to record real-time use and

adherence of systemic corticosteroids.

Statistical analysis

The aim of this trial was to establish whether the eosinophil-guided corticosteroid therapy was non-

inferior in terms of clinical outcome assessed as DAOH within 14 days in comparison to the

standard guideline-based systemic corticosteroid therapy and at the same time could reduce the use

of systemic corticosteroids in hospitalised patients with AECOPD.

We estimated that 318 patients would be required for the trial to have 80% power (1-β) and a one-

sided significance level (α) of 0·025.

The maximum decrease of the primary outcome was set at 1·2 days and the standard deviation

based on other reports, was set at 3·8 days (appendix p 38).3,29 This non-inferiority margin was

based on the most recent Cochrane meta-analysis on administration of systemic corticosteroids as

compared to placebo, offered 1·2 days shorter admission.3 We included no loss to follow-up in the

sample size estimate due to a complete follow-up of the primary outcome in central registries. Thus,

we did not accept any loss to follow-up in the case report form. This sample size calculation was

based on a group sequential design based on the approach of O’Brien & Fleming30, normal

distribution of the means, one-sided non-inferiority and two interim analyses on 1/3 and 2/3 patients 9

25

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

26

27

recruited. The primary outcome was presented as mean (95% CI), according to the protocol,

statistical analysis plan and assessment of the distributions.

We compared baseline characteristics and outcomes with a t-test or Mann-Whitney U test for

continuous outcomes and χ² test for nominal outcomes. We calculated a cumulative event estimate

using a hazard ratio (HR) with 95% confidence interval (95% CI). Mixed linear models were used

to test differences between the eosinophil-guided group and the control group for FEV 1 %predicted,

CAT score, MRC score and BMI after controlling for baseline value. Posthoc a multiple regression

model for the primary outcome measure was performed in order to adjust for pre-existing disability

(activity of daily living, previous exacerbation history and comorbidities). Activity of daily living

was divided into the following categories 1) living in a nursing home or own home with nurse help

several times a day or 2) living at home on own conditions or with help maximally once daily.

(supplement). Due to risk of multiple comparisons, Bonferroni correction was conducted for the

five secondary endpoints (supplement). Statistical analyses were performed using the SAS

statistical software 9.4 (SAS Institute Inc., Cary, NC, USA) and the statistical software R (version

3.4.3). The sample size calculation was done using StudySize 3.0, Frölunda, Sweden.

Interim and final analysis

An independent Data and Safety Monitoring Board (DSMB) reviewed the trial’s progress and

evaluated the safety, efficacy and data completeness during the trial. The board consisted of an

intensivist, an anaesthesiologist and a pulmonologist (appendix p. 4). The interim analyses were

conducted as planned after the recruitment of 1/3 and 2/3 of the target sample size, and the

statistician was blinded to the trial-group assignments. Investigators had no access to data from the

interim analyses. For both the interim analyses, we used the following terms regarding efficacy and

harm: if the z value for mortality analysis was larger than the upper boundary value at the specified

interim analysis, the trial could possibly be prematurely stopped (final decision not based on the Z-

value alone).10

28

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231232233

234

235

236

237

238

239

240

241

29

30

The DSMB recommended at both the interim analyses that the trial should be continued. These

analyses consisted of data on recruitment rates, baseline characteristics, follow-up data and the

primary and the following secondary outcomes: i) Mortality rate within 30 days from recruitment,

ii) Readmission with COPD exacerbation or death within 30 days from recruitment, and iii) Time to

readmission with COPD exacerbation or death within 30 days from recruitment. Dr. Eklöf

performed all interim analyses in strict confidentiality, provided the DSMB with these data and was

available for the board.

The study group, including the steering committee, was blinded to the data until they had been

entered into the database for the scheduled primary analyses. When the processes of recruitment

and follow-up were completed, the database was locked, and all data that could not unblind the

principal investigator were made accessible for analysis. Data were analysed as ‘group A’ vs.

‘group B’. The statistical analysis was performed in accordance with the statistical analysis plan,

and the analysis, with the exception below, was performed by the principal investigator (PS) who

was unaware of the trial-group assignments. Analyses of blood eosinophils and corticosteroid use

were conducted in strict confidentiality by Dr. Eklöf (not to unblind the principal investigator (PS)

and the analysis supervisor (JUJ)). The latter data were presented separately at the unblinding

meeting on January 25, 2019, where all investigators were unblinded to the study results. This trial

was registered as an International Standard Randomised Controlled Trial with ClinicalTrials.gov,

number NCT02857842.

Role of the funding source

Funding was provided by the Danish Regions Medical Fund and the Danish Council for

Independent Research. The funders of the study had no role in the study design, data collection,

data analysis, data interpretation, or in the writing of the report. The principal investigator had full

access to all of the data after the unblinding meeting and the final responsibility of the decision to

submit for publication.11

31

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

32

33

Results

Between August 3, 2016, and September 30, 2018, we screened 1363 patients admitted with COPD

exacerbations in three respiratory medicine departments in Denmark. A total of 318 patients (23%)

of 1363 were randomly assigned to either the eosinophil-guided group (n=159) or the control group

(n=159) within 24 h of admission (figure 1). Thirty patients, after being randomly assigned to a

group, were not fully adherent to treatment algorithm (appendix p 37), resulting in the per-protocol

population of 288 patients, 140 in the eosinophil-guided group and 148 in the control group (figure

1), and thus an adherence to the eosinophil-guided algorithm of the 88·1% and 93·1% in the control

group (supplement). In the eosinophil-guided group, 137 (86%) of 159 patients were discharged

without involving the investigator. In the control group, this number was 143 (90%) of 159

patients. Median time was xx minutes [interquartile range (IQR) xx to yy ] from administration of

systemic corticosteroids until blood samplings for eosinophils was taken on day 1. All 318

randomised patients entered the intention-to-treat analysis. The treatment groups were well

balanced at baseline with respect to demographic and disease characteristics (table 1).

The follow-up for the primary endpoint, re-exacerbation, death, infections requiring antibiotic

treatment and corticosteroid use during admission was complete (100%) for all patients.

Primary endpoint

For the primary outcome (intention-to-treat analysis), DAOH within 14 days after recruitment, no

difference was recorded (mean DAOH of 8·9 [95% CI 8·3 to 9.6] in the eosinophil-guided group

vs. 9·3 [95% CI 8·7 to 9·9] in the control group; between-group absolute difference -0·4 [95% CI -

1·3 to 0·5]; p=0·34).

Results for the per-protocol population were similar (mean DAOH of 9·1 [95% CI 8·4 to 9·8] in the

eosinophil-guided group vs. 9·4 [95% CI 8·8 to 10·0] in the control group; between-group absolute

difference -0·3 [95% CI -1·2 to 0·7]; p = 0·49). 12

34

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

35

36

Secondary endpoints

Within the first 30 days, there was no increase in treatment failure (recurrence of AECOPD

resulting in emergency room visits, hospitalisation or the need to intensify pharmacological

treatment): 42 (26·4%) of 159 patients in the eosinophil-guided group compared to 41 (25·8%) of

159 patients in the control group (difference -0·6% [95% CI -10·0% to 9·0%]; p = 0·90).

Readmission with AECOPD or death at 30 days was found to be 39 (24·5%) of 159 patients and 27

(17·0%) of 159 patients, respectively, in the two groups (figure 2) (difference of 7·5% [95% CI -

1·3% to 16·4%]; p=0·10). After 30 days, 9 (5·7%) of 159 patients died in the eosinophil-guided

group compared to 6 (3·8%) of 159 patients in the control group (difference 1·9% [95% CI -2·8%

to 6·5%]; p = 0·43). The daily usage and the median duration of the systemic corticosteroid therapy

was substantially lower in the eosinophil-guided group compared to the control group (2 days [ IQR

1·0 to 3·0] corresponding to 117·5 mg and 5 days [IQR 5·0 to 5·0] corresponding to 230 mg,

respectively, p < 0·0001 (figure 3). The difference in the mean cumulative corticosteroid dose

continued throughout days 30 and 90 (figure 3, table 2).

During the 90-day follow-up, there was no difference in infections requiring antibiotic treatment in

the eosinophil-guided group compared to the control group (55 (34·6%) of 159 patients vs. 68

(42·8%) of 159 patients; difference of -8·2% [95% CI -18·8% to 2·5%]; p = 0·13, figure 2).

Furthermore, we did not observe differences in dyspepsia, ulcer complication or initiation of new

proton-pump inhibitor treatment at 90 days in the two groups, respectively (table 2). During the 30-

day follow-up, there was no difference in new onset of diabetes mellitus according to HbA1c

definition for patients without pre-existing diabetes (table 2). However, worsening of diabetes

(defined as increasing HbA1c from baseline to day 30) in patients with pre-existing diabetes was

significantly higher in the control group (table 2). Blood glucose levels during admission were

either equal or lower in the eosinophil-guided group from days 3 to 5, but only significantly

different between the groups on day 5 (table 2). At all the time points measured, the mean change 13

37

294

295296297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

38

39

from baseline in FEV1, CAT score, BMI and MRC did not differ between the two groups (table 2

and figure 4). Furthermore, there was no difference in the change in mean of PTH and D-vitamin at

90 days (appendix p. 39) in the groups.

Sensitivity analysis

As post hoc analyses we repeated the analyses for the primary outcome while adjusting for i)

activity of daily living, ii) number of severe AECOPD in the previous year and iii) for all baseline

comorbidities (table 1). This did not affect the signal (supplement). In another post hoc analyse, we

stratified the primary outcome analysis for baseline evidence of pneumonia (table 2). This did not

alter the signal. Additional exploratory analyses were performed as described in the supplement.

Discussion

In this multicentre randomised trial, we assessed patients with severe acute exacerbation in chronic

obstructive pulmonary disease requiring hospital admission. An eosinophil-guided algorithm where

systemic corticosteroids were withheld whenever the daily eosinophil count was below 0·3 109/l did

not result in fewer days alive or out of hospital within 14 days than standard guideline based

treatment. This strategy, however, reduced the duration of systemic corticosteroids to less than half.

Although small numerial differences were observed, in term favouring the control group or the

eosinophil-guided group, no stastistical differences were noted on any secondary endpoints, except

regarding worsening in diabetes mellitus, which was more frequent in the control group. Although

we did not find statistical difference, real differences could happen, but not be identified, especially

for infrequent endpoints, i.e. power issues.

In the Steering Committee, we wanted to make sure, when investigating our hypothesis in the

current trial, that the tested interventional strategy of eosinophil-guided systemic corticosteroids in

AECOPD, would not result in worse outcome regarding length of stay, than not at all using

systemic corticosteroids. Thus, we chose a maximum worsening in the primary outcome measure

equal to the effect of systemic corticosteroids versus placebo, as the non-inferiority boundary.3

14

40

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

41

42

We decided to include the component of all-cause mortality in the primary endpoint in order not to

introduce lead-time bias. Although the tested strategy was successfull in reducing systemic

corticosteroids, we cannot exclude the possibility that a more aggressive algorithm, i.e. one single

dose, may have been more effective.

To the best of our knowledge, only one randomised controlled trial (n = 109) assessed the use of

eosinophil-guided treatment for AECOPD; however, mostly in moderate non-hospitalised

exacerbations.17 In that trial, patients were randomised to either standard oral prednisolone for 14

days to treat AECOPD, while patients in the biomarker-directed arm received prednisolone or

placebo according to a single exacerbation eosinophil count (≥ 2 % vs. < 2%). The investigators

did not not observe any differences in any outcome measures, but they did find a pronounced

reduction of systemic corticosteroids for AECOPD in the biomarker arm. 17 However, unlike our

trial , only ten patients were hospitalised and suffered from a severe exacerbation, and thus the

study population differed markedly from ours where all were hospitalised.17 In line with this,

another randomised controlled trial showed non-inferiority in clinical outcome and reduction of

systemic corticosteroid therapy from 14 to 5-day treatment for AECOPD.5 That study, like ours, had

a spectrum of patients with severe COPD, used a large sample size (n=314) and was a multicentre

trial. Since that trial was published, more knowledge has been acquired on eosinophilic

inflammation in COPD and additionally, the utmost importance of reduction in exposure to

corticosteroids has been increasingly acknowledged.6 Regarding corticosteroid side effects, we only

noted differences regarding worsening of diabetes after one month, which was more frequent in

control patients than in eosinophil-guided patients. The lack of apparent differences regarding other

side-effect measures may, however, also have been caused by insufficient power for these

endpoints, and a relatively short follow-up, which may not capture side-effects that take long time

to develop; e.g. osteoporotic fractures. Specifically, we did not observe any reduction in the number

of infections requiring antibiotic treatment in the eosinophil-guided group. In this context, a recent

study from our group with a large cohort of COPD patients showed that systemic corticosteroids 15

43

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

44

45

increased the risk of infections up to 12 months after exposure.31 The blood glucose levels on day 5

was slightly higher (not hyperglycemia) among patients in the control group than in the eosinophil-

guided group. Although, these results could be interesting from a physiological point of view, the

clinical relevance of this is disputed.

In spite of some strengths of our trial, a number of limitations deserve to be mentioned: First,

although our trial was multicentre and randomised, the design was open-label. The investigators

were blinded to the database, and the authors performing the interim analyses and final analyses

were unaware of the trial-group assignments. However, the physicians, nurses and patients were

aware of the study group assignments and this may have resulted in some non-protocolized co-

interventions. However, the investigators were instructed not to manage the treatment of the

patients in the trial. In some cases, this was not possible for logistic reasons and may have had some

influence. Second, our study did not have sufficient power to detect differences in mortality rate;

this would have demanded several thousands of patients. However, the 30-day mortality in the

current trial was relatively low in both treatment arms (5·7% vs. 3·8%) compared to U.S. data from

>150,000 admissions at 4,500 U.S. hospitals32 in which 30-day mortality was 8·6%, ranging from

5·9% to 13·5%. Additionally, a recent Cochrane meta-analysis of RCT´s (n=1,319) exploring

systemic corticosteroids against placebo, did not find any altered 30-day mortality [OR 1·0, 95% CI

0·6 to 1·7] and thus we consider it unlikely that our algorithm could cause a worse outcome

regarding mortality than placebo.3 Nevertheless, we do acknowledge that our trial is not powered to

detect small differences in mortality. Third, the adherence to our eosinophil-guided algorithm was

not complete for all intervention days (day 2 through 5): only in 88·1% the adherence was

complete. The main reasons for non-adherence was that the treating physician refused (n=10).

However, this fraction was reasonable compared to other trials using biomarker-guided

algorithms.33 For the standard group, the adherence was higher, 93·1%, but still not complete for all

days. It is a limitation that this decision could be made by the treating physician. In this context, in

a few cases, the investigator was involved in conference decisions to discharge patients, which 16

46

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

47

48

could potentially have influenced some of the endpoints. Fourth, it was not possible to measure the

eosinophil count after discharging the patients of the intervention group from hospital. Optimally,

we could have visited the discharged patients at home and made blood samples and continued real-

time eosinophil guidance of systemic corticosteroids. We decided not to do this, since we wanted to

test an intervention that could be implemented directly; this would not be the case, if the

intervention included post-hospitalisation home visits in all patients. We do, however, acknowledge

that these measurements may have been informative. Fifth, we did not have long-term follow-up on

our patients. For some endpoints like infections, adrenal insufficiency, venous thromboembolisms

and osteoporotic fratures34,35, this may be very important, but awaiting these events to happen,

would have delayed the reporting of the trial for several years, and we prioritized to report the

primary, protocolized results immediately. Sixth, only about one in four of the screened patients

were recruited and we cannot be sure that our strategy can be implemented with the same result

among patients more similar to those screened but not eventually recruited. Seven, systemic

corticosteroid administration before blood sampling for esoinophil on day 1 could potentially

influence the blood eosinophil count. However, systemic corticosteorid effect on eosinophils

normally take hours after administration36. Therefore, this effect is probably not so pronounced,

since blood sampling was taken shortly after administration. Eight, all patients received initially

high dose of methylprednisolone according to national guidelines, and we can not exclude that this

could have influenced the outcome. Finally, approximately 2% of the patients were never-smokers.

Although this is a low frequency, and smoking history is not a formal criterion for COPD

diagnosis27, we acknowledge that this is disputed.

In conclusion, our eosinophil-guided strategy resulted in a substantial reduction in systemic

corticosteroid duration in patients suffering from hospitalisation-demanding acute exacerbation of

COPD and was non-inferior regarding days spent alive out of the hospital. Potentially inappropriate

systemic corticosteroid use was stopped in the hospital setting in this highly vulnerable group of

COPD patients with severe COPD without apparent increased risk of death and readmission. 17

49

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

50

51

However, larger studies would help in evaluating the full safety profile of the treatment and control

group, including assesing more accurately which clinical events may be prevented with the new

treatment. Despite several limitations including the open-label design of the trial, eosinophil-guided

corticosteroid treatment is a promising option in hospitalised patients with AECOPD.

Contributors

The study director JUJ and the principal investigator PS of the CORTICO-COP trial take

responsibility for the integrity and the accuracy of their analysis. JUJ, JV, NS, TW and PS

generated the hypothesis for the study. All authors designed the study, collected and interpreted

data, and critically revised the manuscript. PS and JUJ wrote the first protocol, obtained funding for

the study, and initiated the study. PS, TL, RRL, FSH, CS, PSI, KA, CM, JE, JJ, MM, TP, EB and

AKM recruited and followed up patients. PS, JE and JUJ performed the statistical analyses. PS, TL,

RL and JUJ coordinated the trial and were responsible for monitoring. All authors read and

approved the final manuscript.

Declaration of interests

The authors declare that they have no competing interests. The CORTICO-COP study is an

investigator-initiated and driven study. No commercial sponsor had or will have any involvement in

the design and conduct of the studies. The participating investigators in the Steering Committee

have the right to the results.

Data sharing

No further data are available.

Acknowledgment

18

52

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

53

54

We wish to thank patients, relatives, clinical staff and research staff at all trial sites. Without their

support the CORTICO-COP trial would never have been completed. JV is supported by the NIHR

Manchester Biomedical Research Centre.

List of abbreviations

AECOPD: Acute Exacerbation of Chronic obstructive pulmonary disease; COPD: CAT: COPD

Assessment Test; Chronic Obstructive Pulmonary Disease; CRF: Case Report Form; DSMB: Data

and Safety Monitoring Board; FEV1: forced expiratory volume in 1 second; FVC: forced vital

capacity; FEV1/FVC ratio, also called Tiffeneau-index; GCP: Good Clinical Practice; GOLD:

Global Initiative for Chronic Obstructive Lung Disease; ICS: Inhaled corticosteroids; MRC:

Medical Research Council; QoL:Quality-of-life.

References

1.Wedzicha JA, Singh R, Mackay AJ. Acute COPD exacerbations. Clin Chest Med 2014; 35(1): 157-63.2.Seemungal TA, Donaldson GC, Paul EA, Bestall JC, Jeffries DJ, Wedzicha JA. Effect of exacerbation on quality of life in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1998; 157(5 Pt 1): 1418-22.3.Walters JA, Tan DJ, White CJ, Gibson PG, Wood-Baker R, Walters EH. Systemic corticosteroids for acute exacerbations of chronic obstructive pulmonary disease. Cochrane Database Syst Rev 2014; (9): CD001288.4.Laue J, Reierth E, Melbye H. When should acute exacerbations of COPD be treated with systemic corticosteroids and antibiotics in primary care: a systematic review of current COPD guidelines. NPJ Prim Care Respir Med 2015; 25: 15002.5.Leuppi JD, Schuetz P, Bingisser R, et al. Short-term vs conventional glucocorticoid therapy in acute exacerbations of chronic obstructive pulmonary disease: the REDUCE randomized clinical trial. JAMA 2013; 309(21): 2223-31.6.Waljee AK, Rogers MA, Lin P, et al. Short term use of oral corticosteroids and related harms among adults in the United States: population based cohort study. Bmj 2017; 357: j1415.7.Broersen LH, Pereira AM, Jorgensen JO, Dekkers OM. Adrenal Insufficiency in Corticosteroids Use: Systematic Review and Meta-Analysis. J Clin Endocrinol Metab 2015; 100(6): 2171-80.8.Walsh LJ, Wong CA, Oborne J, et al. Adverse effects of oral corticosteroids in relation to dose in patients with lung disease. Thorax 2001; 56(4): 279-84.9.Agusti A, Bel E, Thomas M, et al. Treatable traits: toward precision medicine of chronic airway diseases. Eur Respir J 2016; 47(2): 410-9.10.Brusselle G, Pavord ID, Landis S, et al. Blood eosinophil levels as a biomarker in COPD. Respir Med 2018; 138: 21-31.

19

55

448

449

450

451

452

453

454

455

456

457

458

459

460461462463464465466467468469470471472473474475476477478479480481482483

56

57

11.Bafadhel M, McKenna S, Terry S, et al. Acute exacerbations of chronic obstructive pulmonary disease: identification of biologic clusters and their biomarkers. Am J Respir Crit Care Med 2011; 184(6): 662-71.12.Brightling CE, Monteiro W, Ward R, et al. Sputum eosinophilia and short-term response to prednisolone in chronic obstructive pulmonary disease: a randomised controlled trial. Lancet 2000; 356(9240): 1480-5.13.Singh D, Kolsum U, Brightling CE, et al. Eosinophilic inflammation in COPD: prevalence and clinical characteristics. Eur Respir J 2014; 44(6): 1697-700.14.Bafadhel M, Pavord ID, Russell REK. Eosinophils in COPD: just another biomarker? Lancet Respir Med 2017; 5(9): 747-59.15.Vedel-Krogh S, Nielsen SF, Lange P, Vestbo J, Nordestgaard BG. Blood Eosinophils and Exacerbations in Chronic Obstructive Pulmonary Disease. The Copenhagen General Population Study. Am J Respir Crit Care Med 2016; 193(9): 965-74.16.Yun JH, Lamb A, Chase R, et al. Blood eosinophil count thresholds and exacerbations in patients with chronic obstructive pulmonary disease. J Allergy Clin Immunol 2018; 141(6): 2037-47 e10.17.Bafadhel M, McKenna S, Terry S, et al. Blood eosinophils to direct corticosteroid treatment of exacerbations of chronic obstructive pulmonary disease: a randomized placebo-controlled trial. Am J Respir Crit Care Med 2012; 186(1): 48-55.18.Serafino-Agrusa L, Scichilone N, Spatafora M, Battaglia S. Blood eosinophils and treatment response in hospitalized exacerbations of chronic obstructive pulmonary disease: A case-control study. Pulm Pharmacol Ther 2016; 37: 89-94.19.Bafadhel M, Davies L, Calverley PM, Aaron SD, Brightling CE, Pavord ID. Blood eosinophil guided prednisolone therapy for exacerbations of COPD: a further analysis. Eur Respir J 2014; 44(3): 789-91.20.Camp J, Cane JL, Bafadhel M. Shall We Focus on the Eosinophil to Guide Treatment with Systemic Corticosteroids during Acute Exacerbations of COPD?: PRO. Med Sci (Basel) 2018; 6(3).21.Duman D, Aksoy E, Agca MC, et al. The utility of inflammatory markers to predict readmissions and mortality in COPD cases with or without eosinophilia. Int J Chron Obstruct Pulmon Dis 2015; 10: 2469-78.22.Bafadhel M, Greening NJ, Harvey-Dunstan TC, et al. Blood Eosinophils and Outcomes in Severe Hospitalized Exacerbations of COPD. Chest 2016; 150(2): 320-8.23.Sivapalan P, Moberg M, Eklof J, et al. A multi-center randomized, controlled, open-label trial evaluating the effects of eosinophil-guided corticosteroid-sparing therapy in hospitalised patients with COPD exacerbations - The CORTICO steroid reduction in COPD (CORTICO-COP) study protocol. BMC Pulm Med 2017; 17(1): 114.24.ICH harmonized tripartite guideline: Guideline for Good Clinical Practice. J Postgrad Med 2001; 47(1): 45-50.25.World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. Jama 2013; 310(20): 2191-4.26.Moher D, Hopewell S, Schulz KF, et al. CONSORT 2010 explanation and elaboration: updated guidelines for reporting parallel group randomised trials. BMJ 2010; 340: c869.27.Singh D, Agusti A, Anzueto A, et al. Global Strategy for the Diagnosis, Management, and Prevention of Chronic Obstructive Lung Disease: The GOLD Science Committee Report 2019. Eur Respir J 2019.28.Ariti CA, Cleland JG, Pocock SJ, et al. Days alive and out of hospital and the patient journey in patients with heart failure: Insights from the candesartan in heart failure: assessment of reduction in mortality and morbidity (CHARM) program. Am Heart J 2011; 162(5): 900-6.29.Walters JA, Tan DJ, White CJ, Wood-Baker R. Different durations of corticosteroid therapy for exacerbations of chronic obstructive pulmonary disease. Cochrane Database Syst Rev 2014; (12): CD006897.

20

58

484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534

59

60

30.O'Brien PC, Fleming TR. A multiple testing procedure for clinical trials. Biometrics 1979; 35(3): 549-56.31.Sivapalan PI, T; Rasmussen,DB, et al. COPD exacerbations: The impact of long versus short courses of oral corticosteroids on mortality and pneumonia: nationwide data on 67 000 patients with COPD followed for 12 months. BMJ Open Resp Res 2019; 6(e000407).32.Lindenauer PK, Grosso LM, Wang C, et al. Development, validation, and results of a risk-standardized measure of hospital 30-day mortality for patients with exacerbation of chronic obstructive pulmonary disease. J Hosp Med 2013; 8(8): 428-35.33.de Jong E, van Oers JA, Beishuizen A, et al. Efficacy and safety of procalcitonin guidance in reducing the duration of antibiotic treatment in critically ill patients: a randomised, controlled, open-label trial. Lancet Infect Dis 2016; 16(7): 819-27.34.Walters JA, Tan DJ, White CJ, Wood-Baker R. Different durations of corticosteroid therapy for exacerbations of chronic obstructive pulmonary disease. Cochrane Database Syst Rev 2018; 3: CD006897.35.Vestergaard P, Rejnmark L, Mosekilde L. Fracture risk in patients with chronic lung diseases treated with bronchodilator drugs and inhaled and oral corticosteroids. Chest 2007; 132(5): 1599-607.36.Thorn GW, Renold AE, Wilson DL, et al. Clinical studies on the activity of orally administered cortisone. N Engl J Med 1951; 245(15): 549-55.

21

61

535536537538539540541542543544545546547548549550551552553554555556

62

63

Research in context

Blood eosinophil counts have been suggested as a biomarker to guide systemic corticosteroid

treatment in patients with exacerbation of COPD. It has been shown that higher blood eosinophil

counts are associated with a better response to oral corticosteroids in patients with an acute

exacerbation of COPD. We searched PubMed for papers using the following seach item: “COPD”

AND “eosinophil count” AND “corticosteroids” with no language or date limitations. We limited

the search to studies in human being. We excluded studies in children and studies in which the

association with inhaled corticosteroids and eosinophil count was assessed. Most of the papers

were stpost hoc analyses of clinical trials and observational studies. Only one randomised controlled

trial was identified. This study concluded that eosinophil-guided corticosteroid treatment could be

used safely in moderate COPD exacerbations. In that trial, patients treated with prednisolone for 2

weeks were compared with a biomarker-guided group (receiving prednisolone or placebo based on

a single exacerbation blood eosinophil count). The results of this study showed no difference in

health status or treatment failure rates between the groups.

22

64

557558

559

560

561

562

563

564

565

566

567

568

569

570

571

572

65

66

Added value of this study

To our knowledge, CORTICO-COP is the first randomised trial to show non-inferiority in treating

hospitalised patients with severe COPD exacerbation through biomarker-guided corticosteroid

therapy and simultaneously reducing the overall exposure to systemic corticosteroids by

approximately half. Implementation of this strategy is most likely to reduce, both on individual and

societal levels, the burden of corticosteroid adverse effects substantially in a large and vulnerable

patient group. Additionally, since the differential count of white blood cells is inexpensive and

readily available in most hospital settings, this strategy can easily be implemented, without

establishing new hospital infrastructure, even in resource-constrained environments.

Implications of all available evidence

Our findings may improve patient care and clinical practice in patients with COPD exacerbations.

Reducing ineffective systemic corticosteroid treatment is important due to the well-known longterm

adverse effects of this treatment.

23

67

573

574

575

576

577

578

579

580

581

582

583

584

585

586

68

69

  Eosinophil-guided group Control group  (n=159) (n=159) Demographics

Age (years) 75 (69-81) 75 (68-82)Male 73 (45·9) 70 (44·0)BMI (kg/m2) 24·2 (20·8-26·6) 23·6 (20·3-27·9)MRC 4 (3-5) 4 (3-5)FEV1 (L) 0·7 (0·5-0·9) 0·7 (0·5-0·9)FEV1 %predicted 32 (23-38·5) 30 (23-40·5)FVC (L) 1·6 (1·2-2·1) 1·6 (1·2-2·1)FVC %predicted 56 (42-72) 56·5 (44-70)FEV1/FVC ratio % 0·44 (0·35-0·54) 0·45 (0·37 - 0·55)

SmokingCurrent 54 (34·0%) 50 (31·4%) Past 103 (64·8%) 105 (66·0%)Pack-years smoked, y 45 (30-57) 48 (35-56)CAT score 21 (17-26) 21 (15-26)NIV 4 (2·5%) 5 (3·1%)

SymptomsIncreased dyspnea 146 (91·8%) 151 (95·0%)Increased sputum volume 33 (20·8%) 34 (21·4%)Increased sputum purulence and cough 45 (28·3%) 47 (29·6%)

Comorbidities    Diabetes mellitus 24 (15·1%) 15 (9·4%)Ischemic heart disease 22 (13·8%) 15 (9·4%)Essential hypertension 64 (40·3%) 61 (38.4%)Hypercholesterolemia 19 (11·9%) 19 (11·9%)Chronic renal failure 12 (7·5%) 10 (6·3%)Heart failure 17 (10·7%) 13 (8·2%)Osteoporosis 33 (20·8%) 26 (16·4%)

Clinical findings    Systolic blood pressure (mm Hg) 130 (117-142) 127 (116 - 139) Diastoic blood pressure (mm Hg) 70 (62-78) 70 (62-80)Heart rate, beats/min 89 (79 - 99) 89 (81 - 99)Oxygen saturation with nasal oxygen 95 (93 - 96) 95 (93-96)Oxygen supply, l/min 2 (1-2) 2 (1-2)Respiratory rate – breaths/min 20 (18-22) 20 (18-22)Temperature (°C) 36·6 (36·3 -36·9) 36·6 (36·4 – 36·9)

Laboratory findings    Infiltrate on Chest X-ray 45 (28·3%) 56 (35·2%)Leukocytes (109 cells per liter) 9·8 (7·5-13·5) 9·9 (7·9-13·1)Eosinophils (109 cells per liter) 0·10 (0·01-0·30) 0·06 (0·01-0·20)pH 7·42 (7·38-7·46) 7·42 (7·39 – 7·45)PaCO2 (kPa) 5·4 (4·7 – 6·2) 5·3 (4·7-6·4)PaO2 (kPa) 9·1 (8·2-10·2) 8·8 (7·9-10·1)Ion-Ca (mmol/l) 1·18 (1·15-1·22) 1.19 (1·15-1·22)

Table 1. Baseline Characteristics of Study ParticipantsData are number (%) or median (IQR), unless otherwise specified. BMI = Body Mass Index. FEV1 = forced expiratory volume in 1 second. CAT = COPD Asessment Test. NIV = Non invasiv ventilation.

24

70

587

588

71

72

Primary endpoint Eosinophil-guided group Control groupBetween-group

absolute difference p valueIntention-to-treat (n = 318):Days alive and out of hospital within 14 days after recruitment, mean (95% CI)

8·9 (8·3 to 9·6) 9·3 (8·7 to 9·9) 0·4 (-0·5 to 1·3) 0·34

Per-protocol (n =296):Days alive and out of hospital within 14 days after recruitment, mean (95% CI)

9·1 (8·5 to 9·8) 9·4 (8·8 to 10·0) 0·3 (-0·7 to 1·2) 0·59

Secondary endpointsTreatment failure 30-day 42 (26·4%) 41 (25·8%) 0·6% (-0·09 to 0·10) 0·90

AECOPD readmission or death 30-day 39 (24·5%) 27 (17·0%) 7·5% (-0·01 to 0·16) 0·10

Mortality 30-day 9 (5·7%) 6 (3·8%) 1·9% (-0·03 to 0·07) 0·43Time to readmission with AECOPD or death within 30 days - HR (95% CI)

1·5 (0·9 to 2·5) 1·0 0·09

Cumulative corticosteroid dosageDay 5 2·0 (1·0-3·0) 5·0 (5·0 - 5·0) 3·0 <0·0001Day 30 (mg) mean (95% CI) 173·8 (151·1 to 196·6) 292·7 (272·7 to 312·7) 118·9 (88·8 to 148·9) <0·0001Day 90 (mg) mean (95% CI) 260·8 (216·1 to 305·5) 420·7 (353·1 to 488·3) 159·9 (77·3 to 242·5) 0·0002

FEV1 (ml, change from baseline) Day 3 17·6 (-16·4 to 51·7) 11·0 (-23·2 to 45·2) 6·6 (-41·5 to 54·8) 0·79Day 30 188·3 (130·0 to 246·7) 170·0 (115·1 to 224·9) 18·3 (-61·9 to 98·5) 0·65Day 90 183·9 (122·5 to 245·4) 184·4 (126·4 to 242·3) -0·5 (-85·1 to 84.2) 0·99

CAT Score (change from baseline) Day 30 -2·2 (-3·4 to -1·0) -1·9 (-3·0 to -0·7) -0·3 (-2·0 to 1·3) 0·67Day 90 -3·6 (-4·9 to -2·3) -3·3 (-4·5 to -2·1) -0·3 (-1·9 to 1·4) 0·76

BMI (change from baseline)Day 30 0·006 (-0·3 to 0·3) -0·1 (-0·4 to 0·1) 0·1 (-0·2 to 0·5) 0·48Day 90 -0·06 (-0·3 to 0.2) -0·1 (-0·4 to 0.1) 0·08 (-0·3 to 0·5) 0·71

MRC (change from baseline)Day 30 0·1 (-0·1 to 0·3) -0·1 (-0·3 to 0·1) 0·2 (-0·1 to 0·4) 0·15Day 90 -0·1 (-0·3 to 0·1) -0·2 (-0·4 to -0·1) 0·1 (-0·1 to 0·4) 0·37Adverse events at 90 daysInfection requiring antibiotic treatment 54 (39·6%) 68 (42·8%) 8·8% (-0·2 to 0·0) 0·11Dyspepsia, ulcer complication or new PPI treatment

11 (6·9%) 12 (7·5%) 0·6% (-0·05 to 0·06) 0·83

New onset or worsening of diabetes mellitus at 30 days

12 (7·5%)  10 (6·3%) 1·3% (-0·07 to 0·04) 0·66

Blood glucose - mean (95% CI) Day 1 8·1 (7·6 to 8·6) 8·0 (7·6 to 8·4) 0·1 (-0·7 to 0·5) 0·68 Day 2 6·9 (6·5 to 7·2) 6·9 (6·5 to 7·2) 0·0 (-0·5 to 0·5) 0·97 Day 3 6·1 (5·8 to 6·4) 6·6 (6·1 to 7·1) 0·5 (-0·1 to 1·0) 0·13 Day 4 6·2 (5·9 to 6·6) 6·2 (5·9 to 6·5) 0·0 (-0·5 to 0·5) 0·93 Day 5 6·1 (5·8 to 6·5) 6·7 (6·2 to 7·3) 0·6 (0·0 to 1·2) 0.04

Table 2: Primary and secondary outcome measuresData are number(%), difference (95% CI), or median (IQR), unless otherwise specified. AECOPD = Acute exacerbation of COPD. HR = Hazard Ratio. CI = Confidens Interval.

25

73

589

590

591

74

75

Figure 1: Trial profile

Figure 2: Systemic corticosteroid (SC) use. (A) Proportions of patients on SC. (B) Median duration of SC use during the first 5 days.

26

76

592

593594595596597

77

78

Figure 3: Mean cumulative systemic corticosteroid (SC) dose (mg) during follow-up at: (A) day 30 and (B) day 90.

27

79

598

599

600601602603

80

81

Figure 4: (A) AECOPD readmission during follow-up. (B) infections requiring antibiotic treatment during follow-up

28

82

604605

606

607

608

609610

83

84

Figure 5a: FEV1 change from randomisation to end of follow-up Figure 5b: CAT score change from randomisation to end of follow-up Figure 5c: BMI change from randomisation to end of follow-up Figure 5d: MRC score change from randomisation to end of follow-up

29

85

611

612613614615616617618619620

86

87

30

88

621

622

89

90