Imperial College London · Web viewSeven associations (history of tuberculosis, history of RA, exposure to biomass fuels, tobacco smoking, second-hand smoking, serum CRP, and serum

Elucidating the risk factors for chronic obstructive pulmonary

disease: an umbrella review of meta-analyses

Vanesa Bellou1,2, Lazaros Belbasis1, Athanasios K Konstantinidis2, Evangelos Evangelou1,3

1Department of Hygiene and Epidemiology, University of Ioannina Medical School, Ioannina,

Greece

2Department of Respiratory Medicine, University of Ioannina Medical School, Ioannina, Greece

3Department of Epidemiology and Biostatistics, School of Public Health, Imperial College

London, London, UK

Conflict of interest:

None

Running head:

Risk factors for COPD

Word count of the summary: 238 words

Word count of the manuscript: 2779 words

Number of references: 48

Number of tables: 2

Number of figures: 1

Corresponding author:

Dr Evangelos Evangelou

1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

Assistant Professor, [email protected]

Department of Hygiene and Epidemiology, University of Ioannina Medical School, Ioannina,

Greece

Acknowledgements:

Author contributions: VB, LB and EE designed the study. VB and LB performed the

literature search, the study selection and the data extraction. VB and LB performed the

statistical analyses. VB and LB wrote the first draft of the manuscript, whereas AK and

EE critically reviewed the manuscript. All authors approved the final version of the

manuscript.

Guarantor: EE is the guarantor of the content of the manuscript, including the data and

analysis.

Role of sponsors: Not applicable

Conflicts of interest: None

2

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

Summary

Chronic obstructive pulmonary disease (COPD) is commonly attributed to smoking,

ignoring neglecting other potential risk factors. We aim to critically appraise the

epidemiological credibility of the risk factors examined for COPD in published meta-

analyses. We performed a systematic search to capture systematic reviews and meta-

analyses of observational studies on risk factors and biomarkers for COPD. We applied a

set of standardized methodological criteria based on level of statistical significance,

sample size, between-study heterogeneity, and statistical biases. Our search yielded 11

eligible papers including a meta-analysis for 18 risk factors or biomarkers for COPD, and

8 eligible papers performing only a systematic review. Eleven associations achieved

statistical significance at P <0.001 and 6 associations at P <10-6. Thirteen associations

presented an I2 ≥50%, while 6 associations had evidence for small-study effects and/or

excess significance bias. The associations that had high epidemiological credibility for an

increased risk for COPD were history History of tuberculosis and rheumatoid arthritis,

exposure to biomass fuels, tobacco smoking, and second-hand smoking were supported

by high epidemiological credibility for an increased risk for COPD. Furthermore, robust

highly suggestive evidence was found for elevated levels of serum CRP, and serum

fibrinogen in COPD patients compared to healthy controls. Our approach indicates that,

despite thatwhilst a proportion of COPD patients are non-smokers, only a narrow range

of risk factors not related to smoking have been studied for an association with COPD.

There is also a need to decipher possible protective factors in COPD pathogenesis given

that more than a half of ever smokers do not develop COPD.

Keywords: epidemiology; epidemiologic credibility; risk factor; biomarker; bias

3

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

Abbreviation list: CI, confidence interval; COPD, chronic obstructive pulmonary

disease; IQR, interquartile range; MR, Mendelian randomization; OR, odds ratio; RA,

rheumatoid arthritis; RR, risk ratio; SE, standard error

4

60

61

62

Introduction

Chronic obstructive pulmonary disease (COPD) is characterized by progressive

airflow obstruction, due to narrowing of small airways, and destruction of alveolar walls

1,2. The global prevalence of COPD is estimated to be 10.7% in individuals aged 30 years

or more 3. In the USA, COPD is the third leading cause of death 4,5, and it is associated

with a large economic burden 6.

COPD results from an interplay of genetic and environmental risk factors 7. The most

recent meta-analysis of genome-wide association studies found 22 loci associated with

COPD 8. Smoking is considered the main and causal environmental risk factor for COPD

9. However, it is not the only risk factor since about 30% of COPD patients have never

smoked 10. Another fact that should not be neglected is that the attributable fraction for

smoking is lower in developing countries compared with developed countries. 9 This

observation indicates the importance to identify additional risk factors in order to tackle

the prevalence of COPD, especially in low and middle-income countries. For this reason,

a number of other environmental risk factors have been examined for an association with

COPD, including indoor and outdoor air pollution and occupational exposures 9.

Several environmental factors and biomarkers have been considered in systematic

reviews and meta-analyses for risk for COPD. However, there has been no effort to

systematically summarize and critically appraise these associations. We performed an

umbrella review to map the range and validity of current evidence on environmental

factors and biomarkers for affecting the risk for COPD, and to examine their

epidemiological credibility and the presence of statistical biases in these associations.

5

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

Methods

We conducted an umbrella review, which is a systematic collection and evaluation of

multiple systematic reviews and meta-analyses on a specific research topic 11. Identified

risk factors from the umbrella review were assessed using standardized, state-of-the-art

methods.12–14

Search strategy and eligibility criteria

We systematically searched PubMed from inception to December 31, 2017 to

identify systematic reviews and meta-analyses of observational studies examining the

association of environmental factors or serum biomarkers and risk for COPD. We used

the following search algorithm: (“chronic obstructive pulmonary disease” OR “chronic

bronchitis” OR emphysema OR COPD) AND (“systematic review*” OR meta-analys*).

We did not apply any language restrictions. When more than one meta-analysis on the

same research question was available, the meta-analysis with the largest number of

prospective cohort studies was retained for our analysis. We excluded meta-analyses

limited in a specific geographic location. Literature search and selection of eligible

studies were performed independently by two investigators (VB, LB), and in case of

discrepancies the final decision was that of a third investigator (EE).

We additionally searched for Mendelian randomization (MR) studies examining the

potential causal association between risk factors and COPD. In this search, the following

search algorithm was used: (“mendelian randomization” OR “mendelian randomisation”)

AND “obstructive pulmonary disease”. We complemented the search strategy by

6

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

reference screening of eligible MR studies. An MR study is an approach that uses

measured variation in genes of known function to make causal inference of the effect of

an exposure on an outcome. It is an application of the technique of instrumental variables

with genotype acting as an instrument for the exposure of interest. 15

Data extraction

Data extraction was performed independently by two investigators (VB, LB), and in

case of discrepancies the final decision was that of a third investigator (EE). From each

eligible article, we recorded the first author, the year of publication, the examined risk

factors and the number of studies considered. For From each eligible article including a

systematic reviews without a quantitative synthesis, we recorded the justification for not

conducting meta-analysis and the main conclusion.

For From each eligible article including a meta-analysesanalysis, we extracted the

study-specific risk estimates (i.e., standardized mean difference, risk ratio, odds ratio, and

hazard ratio) along with the corresponding 95% confidence interval (CI) and the number

of cases and controls in each study for each risk factor. We screened the component

studies to ensure that none of the eligible meta-analyses included studies with shared

population.

From each MR study, we extracted the first author and year of publication, the risk

factor considered, the level of comparison for exposure, the genetic instrument used, the

applied statistical approach, the sample size, the causal odds ratio and its 95% CI, the P-

value for the association, and whether the authors claimed that a causal relationship

exists.

7

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

Statistical analysis

For each meta-analysis, we estimated the summary effect size estimate and its 95%

CI using fixed-effect and random-effects models. In the case of meta-analyses with

continuous data, the standardized mean difference was transformed to an odds ratio. We

also estimated the 95% prediction interval (PI), which further accounts for between-study

heterogeneity and evaluates the uncertainty for the effect that would be expected in a new

study addressing that the same association. For the largest study of each meta-analysis,

we estimated the SE of the effect size estimate and we examined whether the SE was less

than 0.10.

Between-study heterogeneity was assessed by the I2 metric. I2 ranges between 0%

and 100% and is the ratio of between-study variance over the sum of within-study and

between-study variances. Values exceeding 50% or 75% are usually judged to represent

large or very large heterogeneity, respectively.

We assessed whether there was evidence for small-study effects (i.e., whether

smaller studies tend to give substantially larger estimates of effect size compared with

larger studies) with the Egger’s regression asymmetry test 16. A P <0.10 combined with a

more conservative effect in the largest study than in random-effects meta-analysis was

judged to provide evidence for small-study effects.

We further applied the excess statistical significance test, which evaluates whether

there is a relative excess of formally significant findings in the published literature due to

any reason (e.g., publication bias, selective reporting of outcomes or analyses) 17. This

test assesses whether the observed number of studies with nominally significant results is

8

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

larger than their expected number. Excess statistical significance was claimed at two-

sided P <0.10.

Assessment of epidemiological credibility

We identified associations that had the strongest evidence and no signals of large

between-study heterogeneity or, small-study effects or excess significance bias.

Specifically, we considered as convincing the associations that fulfilled all the following

criteria: statistical significance per random-effects model at P <1 × 10-6, based on >1000

cases, without large between-study heterogeneity (I2 <50%), 95% PI excluding the null

value, and no evidence of small-study effects and excess significance bias. The

associations with >1000 cases, P <1 × 10-6 in random-effects meta-analysis, and largest

study presenting a statistically significant effect (P <0.05) were graded as highly

suggestive. The associations supported by >1000 cases and a statistically significant

effect at P < 1 × 10-3 were considered as suggestive. The remaining nominally significant

associations were considered as having weak evidence. For associations with convincing

and highly suggestive evidence, we performed a sensitivity analysis including only

prospective cohort studies and nested case-control studies, and we examined whether

there was a change in the level of epidemiological credibility.

The statistical analyses and the power calculations were done with STATA, version

12.0.

Ethics approval statement

Our study does not involve human subjects. Thus, ethics approval is not required.

9

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

Results

Our literature search resulted in 1801 papers, and nineteen articles were deemed

eligible (Figure 1). Eight articles performed a systematic review without quantitative

synthesis of the available evidence, and eleven articles performed a meta-analysis as well.

Thirteen articles were excluded after full-text screening, because another more recently

published meta-analysis with more component studies was available. These meta-

analysesarticles pertained to biomass fuels, history of tuberculosis, history of psoriasis,

serum vitamin D, tobacco smoking, second-hand smoking, and occupational exposures.

Five meta-analysespapers were excluded after full-text screening, because they included

meta-analyses due towith fundamental errors in statistical synthesis. Specifically, in these

five papers, we identified errors in the extraction or the calculation of effect size, or

inclusion of multiple component studies with overlapping samples.

Meta-analyses of risk factors and biomarkers for COPD

The 11 eligible articles included 18 unique meta-analyses. The median number of

studies per meta-analysis was 5 (IQR, 4 – 11) and the median number of cases was 2732

(IQR, 2108 – 3359). Nine of 18 meta-analyses examined exposures to various noxious

particles, 4 meta-analyses focused on medical conditions as risk factors for developing

COPD, and 5 meta-analyses examined the levels of serum inflammatory biomarkers in

COPD patients compared with healthy controls.

Thirteen associations were nominally significant (P <0.05), and eleven of them

remained statistically significant at P <0.001. Only six of them were statistically

significant at P <1 × 10-6 (Table 1). These associations pertained to exposure to biomass

10

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

fuels, history of tuberculosis, history of rheumatoid arthritis (RA), tobacco smoking,

second-hand smoking, serum C-reactive protein (CRP), and serum fibrinogen. Only three

associations had 95% PI that excluded the null value (Table 1). These associations

focused on history of tuberculosis, tobacco smoking, and second-hand smoking. In fifteen

meta-analyses, the result in the largest study was more conservative than the summary

result under in the random-effects model. Also, in 10 meta-analyses, the standard error of

the largest study was less than 0.10 in a log OR scale.

Five associations (biological dust, second-hand smoking, serum TNFα, traffic

intensity on nearest road, and traffic load on major roads within 100 meters) had small or

moderate between-study heterogeneity (I2 <50%). Three associations (gases or fumes,

mineral dust, and serum CRP) had large heterogeneity (I2 ≥50% and I2 ≤75%). Ten

associations (history of tuberculosis, exposure to biomass fuels, tobacco smoking, history

of psoriasis, history of RA, waterpipe smoking, serum vitamin D, vitamin D deficiency,

and serum fibrinogen) presented very large heterogeneity (I2 >75%).

In 5 associations, there was evidence for small-study effects (Table 1). These

associations pertained to history of tuberculosis, exposure to biomass fuels, history of

psoriasis, second-hand smoking, and serum vitamin D. In 4 associations, there was

evidence for excess statistical significance and these associations pertained to exposure to

biomass fuels, second-hand smoking, serum vitamin D, and vitamin D deficiency (Table

1). In four meta-analyses, pertaining to history of RA and psoriasis, traffic intensity on

nearest road, and traffic load on major roads within 100 meters, excess significance test

could not be performed because sample sizes of component studies were not available.

Assessment of epidemiological credibility

11

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

Seven associations (history of tuberculosis, history of RA, exposure to biomass fuels,

tobacco smoking, second-hand smoking, serum CRP, and serum fibrinogen) were graded

as highly suggestive (P <1 × 10-6 in random-effects meta-analysis and largest study with a

statistically significant effect at P <0.05). Five of them (history of tuberculosis, exposure

to biomass fuels, tobacco smoking, serum CRP, and serum fibrinogen) with highly

suggestive evidence also presented a large effect size (i.e., OR >2.00).

Three associations (history of psoriasis, serum vitamin D, and serum leukocytes)

presented suggestive evidence (more than 1000 cases, and P <0.001 but P >1 × 10-6).

Three associations (waterpipe smoking, vitamin D deficiency, and serum TNFα) were

supported by weak evidence. Five associations (exposure to biological dust, gases or

fumes, mineral dust, traffic intensity on nearest road, and traffic load on major roads

within 100 meters) were not statistically significant (P >0.05 under the random-effects

model).

From the associations that were supported by highly suggestive evidence, exposure

to biomass fuels 18, history of tuberculosis 19 and history of RA 20 were examined in case-

control studies, cross sectional studies and retrospective cohort studies. In contrast, the

associations on active smoking and second-hand smoking were mainly examined in

prospective cohort studies. In the sensitivity analysis, the association for active tobacco

smoking remained highly suggestive, but the association for second-hand smoking

presented weak evidence due to small number of COPD cases (Table 2). In the meta-

analyses on serum CRP and serum fibrinogen 21, only one prospective cohort study was

included.

Mendelian randomization studies for COPD

12

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

Overall, we captured 4 MR studies for risk of developing COPD 22–25. Three MR

studies 22,23,25 examined the causal effect of genetically elevated levels of serum CRP on

COPD risk. These MR studies constructed haplotypes of polymorphisms in CRP gene as

instrumental variable. An additional MR study examined the causal effect of genetically

elevated levels of serum interleukin-6 on COPD risk and used 8 single nucleotide

polymorphisms as instrumental variables 24. All MR studies used individual-level data.

The findings from the MR studies indicated that a genetically elevated serum CRP and

serum interleukin-6 was not associated with increased risk for developing COPD.

Systematic reviews without a meta-analysis

The eight eligible systematic reviews that did not perform quantitative synthesis

focused on n-3 fatty acids intake 26,27, exposure to pesticides 28, dietary fiber intake 27,

menopause 29, agricultural work 30, construction work 31, and socioeconomic status 32. The

systematic reviews avoided quantitative synthesis due to large methodological

heterogeneity of observational studies. The authors of the systematic reviews concluded

that low socioeconomic status 32, higher intake of dietary fibers 27, exposure to

agricultural and construction workplaces 30,31, and exposure to pesticides 28 may increase

the risk for COPD, but they argued that the evidence for n-3 fatty acids, and menopause

was inconsistent 26,27,29.

Discussion

In this umbrella review, we summarized and critically appraised the evidence on 18

risk factors and serum biomarkers for COPD. Overall, three toxic exposures (i.e., active

13

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

and passive tobacco smoking, and biomass fuels exposure), and two medical conditions

(i.e., history of tuberculosis, and RA) presented robust evidencehigh epidemiological

credibility for an elevated COPD risk. Additionally, highly suggestive evidence indicated

that serum levels of two biomarkers (i.e. CRP, fibrinogen) were higher in COPD patients

than healthy controls.

Smoking is the foremost risk factor for developing COPD. There is a vast amount of

observational studies supporting an association with COPD. Despite the very large

between-study heterogeneity, the 95% PI excluded the null value, indicating a robust

credible association between smoking and risk for COPD. Smokers presented a 4-fold

increased risk compared to never smokers. Additionally, passive smoking was also

strongly associated with risk for COPD, increasing the lung’s total burden of inhaled

particles and gases 33. Higher smoking intensity, commonly measured by the number of

pack-years, is related to COPD severity 7. Only 40% to 50% of lifetime smokers develop

COPD 34, indicating that unknown protective factors and potential gene-environment

interactions attenuate the risk in smokers 33.

About 30% of COPD patients are never-smokers, indicating that there are additional

factors modifying the risk for COPD. Our umbrella review indicated that exposure to

biomass fuel was also associated with risk for developing COPD. In developing

countries, exposure to indoor air pollution is considered the main risk factor for COPD

and it contributes to the pathogenesis of COPD through deposition of noxious particles

on the airways 33. Although the largest exposure burden is reported in low-income

countries, biomass fuels, primarily wood, are also used in developed countries, mainly for

14

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

heating 9. Combustion of biomass fuel produces a variety of air pollutants especially

particulate matter, which is inhaled into the small airways and alveolar spaces. 35

Moreover, our results depicted a highly suggestive relationship between history of

tuberculosis and COPD, indicating that COPD could be a long-term complication of

tuberculosis 36. The 95% PI for this association excluded the null value, but evidence for

small-study effects was found, indicating that the effect size could be inflated. Also, this

association was examined only in case-control studies and cross-sectional studies. The

underlying mechanism explaining this association could be fibrosis and inflammation of

the airways induced directly by the mycobacterium or by the immune response of the

host. 2 In developing countries, epidemiological studies indicated that history of

tuberculosis exerted a much larger risk for COPD compared with exposure to biomass

fuels or tobacco smoking. 2,37 It is disputed that smoking could be the underlying link

between tuberculosis and COPD, because A meta-analysis of observational studies

showed that smokers have an elevated risk of tuberculosis, and this observation could

indicate that the association between tuberculosis and risk for COPD could be partly

explained through smoking 34. However, there are studies showing that the association

remains statistically significant even after adjusting for tobacco smoking and biomass

fuels. This evidence supports an association of history of tuberculosis with COPD,

independent of smoking. is independent of smoking 2,37 Tobacco smoking and biomass

fuels may also contribute in the airflow obstruction caused by tuberculosis, resulting in

COPD

Furthermore, our study showed that highly suggestive evidence exists for an

increased risk of developing COPD in RA patients., but tIn the respective meta-analysis,

15

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

this association was examined exclusively in retrospective cohort studies, and a temporal

association could not be established. Smoking is a risk factor for both RA and COPD;

hence, this association is commonly attributed to the effect of smoking.38–40 The first

prospective cohort study was recently published and was based on the population of

Nurses’ Health Study.41 In this study, a statistically significant association was found

between RA and risk for COPD, whereas the association remained significant after

adjusting for confounding and mediation by smoking and other exposures through

marginal structural models. Under the light of this observation, RA could be considered

to increase the risk for COPD independently of smoking. This association could be

explained by the fact that smoking is a shared risk factor for both conditions 38–40. We

should also note that current evidence from genome-wide association studies do not

support the presence of shared genetic polymorphisms between RA and COPD.8,42 A

previous study showed that RA, besides increasing the risk for COPD, also facilitates a

shortening of time course for developing COPD.

COPD is associated with systemic oxidative stress, activation of circulating

inflammatory cells and increased plasma levels of proinflammatory cytokines, which

include CRP, interleukin-6, fibrinogen, white blood cells and TNF-α 21,43. The findings of

our umbrella review indicated that there is highly suggestive evidence for increased

levels of two markers of systemic inflammation in COPD patients, serum levels of CRP

and fibrinogen in COPD patients. Only suggestive and weak evidence was found for

elevated levels of serum leukocytes and serum TNF-α in COPD patients, respectively.

However, genetically elevated serum CRP and serum interleukin-6 were not associated

with risk for COPD based on the findings from MR studies 22–25. The findings of MR

16

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

studies indicate that either clinical manifestations of COPD lead to an elevation of serum

CRP or the observation association is caused by confounders. 23 For the association of

fibrinogen and risk for COPD, we did not identify any MR study. This observation

indicates that inflammatory markers do not modify the risk for COPD, and probably the

elevated levels of these markers in COPD patients areis attributed to the physical course

of disease. The absence of a causal association between serum inflammatory markers and

risk for COPD does not preclude their clinical significance. These markers are useful as

prognostic indicators, as well as surrogate markers to monitor response to treatment,

rather than diagnostic tools. 23

Our umbrella review has some strengths and introduces some innovations in the field

of COPD epidemiology. First, we summarized all previously published meta-analyses of

observational studies on risk environmental factors and biomarkers affecting the risk for

COPD. Second, we applied a set of additional statistical tools (i.e., level of significance,

95% PI, and excess significance test) and methodological criteria to further examine the

robustness credibility of the associations. However, our umbrella review also has some

limitations. First, we considered only risk factors that were examined in published

systematic reviews and meta-analyses. There might be additional risk factors with

adequate number of observational studies that have not yet been considered in a meta-

analysis. Also, Egger’s test should be interpreted with caution in presence of large

between-study heterogeneity. However, it is unlikely that the credibility assessment

would be different if we did not consider the hints for small-study effects.

Conclusions

17

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

In this umbrella review, we showed that active and passive smoking, exposure to

biomass fuels, history of tuberculosis and history of RA are associated with an increased

risk for developing COPD and these associations are supported by highly suggestive

evidence. Smokers had a 4-fold increase in the risk of developing COPD, whereas

individuals exposed to the rest of the risk factors had an at least 2-fold increase in the risk

for COPD. Also, we found that serum CRP and serum fibrinogen are increased in patients

with COPD. From the perspective of public health, the prevention of exposure to noxious

particles through tobacco control, and avoidance of indoor air pollution through exposure

to biomass fuels is the most appropriate strategy for the prevention of COPD. Also, in

developing countries, population-wide public health interventions to control tuberculosis

could be an additional approach towards the prevention of COPD. There is a need for

more research on protective factors that might prevent or delay the ons et of COPD in

high-risk populations. Given the chronic course and the debilitating property of COPD,

the identification of risk factors and biomarkers could help define a high-risk population

to target for screening.

18

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

References

1. Barnes PJ, Shapiro SD, Pauwels RA. Chronic obstructive pulmonary disease:

molecular and cellular mechanisms. Eur Respir J [Internet]. 2003 Oct;22(4):672–

88. Available from: http://www.ncbi.nlm.nih.gov/pubmed/14582923

2. Salvi SS, Barnes PJ. Chronic obstructive pulmonary disease in non-smokers.

Lancet (London, England) [Internet]. 2009 Aug 29;374(9691):733–43. Available

from: http://www.ncbi.nlm.nih.gov/pubmed/19716966

3. Adeloye D, Chua S, Lee C, Basquill C, Papana A, Theodoratou E, et al. Global

and regional estimates of COPD prevalence: Systematic review and meta-analysis.

J Glob Health [Internet]. 2015 Dec;5(2):20415. Available from:

http://www.ncbi.nlm.nih.gov/pubmed/26755942

4. Ford ES, Croft JB, Mannino DM, Wheaton AG, Zhang X, Giles WH. COPD

surveillance--United States, 1999-2011. Chest [Internet]. 2013 Jul;144(1):284–


5. Kochanek KD, Murphy SL, Xu J, Arias E. Mortality in the United States 2013.

Hyattsville, MD, MD; 2014.

6. Ford ES, Murphy LB, Khavjou O, Giles WH, Holt JB, Croft JB. Total and state-

specific medical and absenteeism costs of COPD among adults aged ≥ 18 years in

the United States for 2010 and projections through 2020. Chest [Internet]. 2015

Jan;147(1):31–45. Available from:


19

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

7. Decramer M, Janssens W, Miravitlles M. Chronic obstructive pulmonary disease.

Lancet (London, England) [Internet]. 2012 Apr 7;379(9823):1341–51. Available


8. Hobbs BD, de Jong K, Lamontagne M, Bossé Y, Shrine N, Artigas MS, et al.

Genetic loci associated with chronic obstructive pulmonary disease overlap with

loci for lung function and pulmonary fibrosis. Nat Genet [Internet]. 2017

Mar;49(3):426–32. Available from:


9. Eisner MD, Anthonisen N, Coultas D, Kuenzli N, Perez-Padilla R, Postma D, et al.

An official American Thoracic Society public policy statement: Novel risk factors

and the global burden of chronic obstructive pulmonary disease. Am J Respir Crit

Care Med [Internet]. 2010 Sep 1;182(5):693–718. Available from:


10. Postma DS, Bush A, van den Berge M. Risk factors and early origins of chronic

obstructive pulmonary disease. Lancet (London, England) [Internet]. 2015 Mar

7;385(9971):899–909. Available from:


11. Ioannidis JPA. Integration of evidence from multiple meta-analyses: a primer on

umbrella reviews, treatment networks and multiple treatments meta-analyses.

CMAJ. 2009 Oct;181(8):488–93.

12. Bellou V, Belbasis L, Tzoulaki I, Evangelou E, Ioannidis JPA. Environmental risk

factors and Parkinson’s disease: An umbrella review of meta-analyses.

20

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

Parkinsonism Relat Disord [Internet]. 2016 Feb;23:1–9. Available from:


13. Bellou V, Belbasis L, Tzoulaki I, Middleton LT, Ioannidis JPA, Evangelou E.

Systematic evaluation of the associations between environmental risk factors and

dementia: An umbrella review of systematic reviews and meta-analyses.

Alzheimers Dement [Internet]. 2017 Apr 4;13(4):406–18. Available from:

http://linkinghub.elsevier.com/retrieve/pii/S1552526016328539

14. Bellou V, Belbasis L, Tzoulaki I, Evangelou E. Risk factors for type 2 diabetes

mellitus: An exposure-wide umbrella review of meta-analyses. PLoS One

[Internet]. 2018;13(3):e0194127. Available from:


15. Lawlor DA, Tilling K, Davey Smith G. Triangulation in aetiological

epidemiology. Int J Epidemiol [Internet]. 2017 Jan 20;dyw314. Available from:


16. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected

by a simple, graphical test. BMJ. 1997 Sep;315(7109):629–34.

17. Ioannidis JPA, Trikalinos TA. An exploratory test for an excess of significant

findings. Clin Trials. 2007 Jan;4(3):245–53.

18. Hu G, Zhou Y, Tian J, Yao W, Li J, Li B, et al. Risk of COPD from exposure to

biomass smoke: a metaanalysis. Chest [Internet]. 2010 Jul;138(1):20–31.

Available from: http://www.ncbi.nlm.nih.gov/pubmed/20139228

21

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

19. Byrne AL, Marais BJ, Mitnick CD, Lecca L, Marks GB. Tuberculosis and chronic

respiratory disease: a systematic review. Int J Infect Dis [Internet]. 2015

Mar;32:138–46. Available from: http://www.ncbi.nlm.nih.gov/pubmed/25809770

20. Ungprasert P, Srivali N, Cheungpasitporn W, Davis Iii JM. Risk of incident

chronic obstructive pulmonary disease in patients with rheumatoid arthritis: A

systematic review and meta-analysis. Joint Bone Spine [Internet]. 2016

May;83(3):290–4. Available from:


21. Gan WQ, Man SFP, Senthilselvan A, Sin DD. Association between chronic

obstructive pulmonary disease and systemic inflammation: a systematic review

and a meta-analysis. Thorax [Internet]. 2004 Jul;59(7):574–80. Available from:


22. van Durme YMTA, Verhamme KMC, Aarnoudse A-JLHJ, Van Pottelberge GR,

Hofman A, Witteman JCM, et al. C-reactive protein levels, haplotypes, and the

risk of incident chronic obstructive pulmonary disease. Am J Respir Crit Care Med

[Internet]. 2009 Mar 1;179(5):375–82. Available from:


23. Dahl M, Vestbo J, Zacho J, Lange P, Tybjærg-Hansen A, Nordestgaard BG. C

reactive protein and chronic obstructive pulmonary disease: a Mendelian

randomisation approach. Thorax [Internet]. 2011 Mar;66(3):197–204. Available


24. van Durme YMTA, Lahousse L, Verhamme KMC, Stolk L, Eijgelsheim M, Loth

22

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

DW, et al. Mendelian randomization study of interleukin-6 in chronic obstructive

pulmonary disease. Respiration [Internet]. 2011;82(6):530–8. Available from:


25. Wium-Andersen MK, Orsted DD, Nordestgaard BG. Elevated C-reactive protein,

depression, somatic diseases, and all-cause mortality: a mendelian randomization

study. Biol Psychiatry [Internet]. 2014 Aug 1;76(3):249–57. Available from:


26. Fulton AS, Hill AM, Williams MT, Howe PRC, Coates AM. Paucity of evidence

for a relationship between long-chain omega-3 fatty acid intake and chronic

obstructive pulmonary disease: a systematic review. Nutr Rev [Internet]. 2015

Sep;73(9):612–23. Available from:


27. Fonseca Wald ELA, van den Borst B, Gosker HR, Schols AMWJ. Dietary fibre

and fatty acids in chronic obstructive pulmonary disease risk and progression: a

systematic review. Respirology [Internet]. 2014 Feb;19(2):176–84. Available


28. Doust E, Ayres JG, Devereux G, Dick F, Crawford JO, Cowie H, et al. Is pesticide

exposure a cause of obstructive airways disease? Eur Respir Rev [Internet]. 2014

Jun;23(132):180–92. Available from:


29. Campbell B, Davis SR, Abramson MJ, Mishra G, Handelsman DJ, Perret JL, et al.

Menopause, lung function and obstructive lung disease outcomes: a systematic

23

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

review. Climacteric [Internet]. 2018 Feb;21(1):3–12. Available from:


30. Fontana L, Lee S-J, Capitanelli I, Re A, Maniscalco M, Mauriello MC, et al.

Chronic Obstructive Pulmonary Disease in Farmers: A Systematic Review. J

Occup Environ Med [Internet]. 2017 Aug;59(8):775–88. Available from:


31. Borup H, Kirkeskov L, Hanskov DJA, Brauer C. Systematic review: chronic

obstructive pulmonary disease and construction workers. Occup Med (Lond)

[Internet]. 2017 Apr 1;67(3):199–204. Available from:


32. Gershon AS, Dolmage TE, Stephenson A, Jackson B. Chronic obstructive

pulmonary disease and socioeconomic status: a systematic review. COPD

[Internet]. 2012 Jun;9(3):216–26. Available from:


33. From the Global Strategy for the Diagnosis, Management and Prevention of

COPD, Global Initiative for Chronic Obstructive Lung Disease (GOLD) 2017

[Internet]. 2017. Available from: http://goldcopd.org

34. Jayes L, Haslam PL, Gratziou CG, Powell P, Britton J, Vardavas C, et al.

SmokeHaz: Systematic Reviews and Meta-analyses of the Effects of Smoking on

Respiratory Health. Chest [Internet]. 2016 Jul;150(1):164–79. Available from:


35. Assad NA, Kapoor V, Sood A. Biomass smoke exposure and chronic lung disease.

24

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

Curr Opin Pulm Med [Internet]. 2016 Mar;22(2):150–7. Available from:


36. Chakrabarti B, Calverley PMA, Davies PDO. Tuberculosis and its incidence,

special nature, and relationship with chronic obstructive pulmonary disease. Int J

Chron Obstruct Pulmon Dis [Internet]. 2007;2(3):263–72. Available from:


37. Menezes AMB, Hallal PC, Perez-Padilla R, Jardim JRB, Muiño A, Lopez M V, et

al. Tuberculosis and airflow obstruction: evidence from the PLATINO study in

Latin America. Eur Respir J [Internet]. 2007 Dec;30(6):1180–5. Available from:


38. Shen T-C, Lin C-L, Chen C-H, Tu C-Y, Hsia T-C, Shih C-M, et al. Increased risk

of chronic obstructive pulmonary disease in patients with rheumatoid arthritis: a

population-based cohort study. QJM [Internet]. 2014 Jul;107(7):537–43. Available


39. Di Giuseppe D, Discacciati A, Orsini N, Wolk A. Cigarette smoking and risk of

rheumatoid arthritis: a dose-response meta-analysis. Arthritis Res Ther [Internet].

2014 Mar 5;16(2):R61. Available from:


40. Sugiyama D, Nishimura K, Tamaki K, Tsuji G, Nakazawa T, Morinobu A, et al.

Impact of smoking as a risk factor for developing rheumatoid arthritis: a meta-

analysis of observational studies. Ann Rheum Dis [Internet]. 2010 Jan;69(1):70–


25

499

500

501

502

503

504

505

506

507

508

509

510

511

512

513

514

515

516

517

518

519

520

41. Sparks JA, Lin T-C, Camargo CA, Barbhaiya M, Tedeschi SK, Costenbader KH,

et al. Rheumatoid arthritis and risk of chronic obstructive pulmonary disease or

asthma among women: A marginal structural model analysis in the Nurses’ Health

Study. Semin Arthritis Rheum [Internet]. 2018 Apr;47(5):639–48. Available from:

http://linkinghub.elsevier.com/retrieve/pii/S0049017217303827

42. Okada Y, Wu D, Trynka G, Raj T, Terao C, Ikari K, et al. Genetics of rheumatoid

arthritis contributes to biology and drug discovery. Nature [Internet]. 2014 Feb

20;506(7488):376–81. Available from:


43. Sinden NJ, Stockley RA. Systemic inflammation and comorbidity in COPD: a

result of “overspill” of inflammatory mediators from the lungs? Review of the

evidence. Thorax [Internet]. 2010 Oct;65(10):930–6. Available from:


44. Alif SM, Dharmage SC, Bowatte G, Karahalios A, Benke G, Dennekamp M, et al.

Occupational exposure and risk of chronic obstructive pulmonary disease: a

systematic review and meta-analysis. Expert Rev Respir Med [Internet]. 2016

Aug;10(8):861–72. Available from:


45. Schikowski T, Adam M, Marcon A, Cai Y, Vierkötter A, Carsin AE, et al.

Association of ambient air pollution with the prevalence and incidence of COPD.

Eur Respir J [Internet]. 2014 Sep;44(3):614–26. Available from:


26

521

522

523

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

46. Zhu B, Zhu B, Xiao C, Zheng Z. Vitamin D deficiency is associated with the

severity of COPD: a systematic review and meta-analysis. Int J Chron Obstruct

Pulmon Dis [Internet]. 2015;10:1907–16. Available from:


47. Waziry R, Jawad M, Ballout RA, Al Akel M, Akl EA. The effects of waterpipe

tobacco smoking on health outcomes: an updated systematic review and meta-

analysis. Int J Epidemiol [Internet]. 2016 Apr 13;dyw021. Available from:


48. Zhu M, Wang T, Wang C, Ji Y. The association between vitamin D and COPD

risk, severity, and exacerbation: an updated systematic review and meta-analysis.

Int J Chron Obstruct Pulmon Dis [Internet]. 2016;11:2597–607. Available from:


27

543

544

545

546

547

548

549

550

551

552

553

554

555

556

Figure. Flow chart of literature search

28

20 articles were excluded

13 were not the largest meta-analysis5 were meta-analyses with fundamental statistical errors1 was meta-analysis focused in limited geographic location1 was meta-analysis not adequately presenting the effect estimates


22 were diagnostic, prognostic or screening studies 122 had outcomes other than risk for COPD15 were treatment studies 6 were incidence or prevalence studies8 were observational studies


866 were treatment studies 312 had outcomes other than risk for COPD 138 were editorials or narrative reviews 130 were articles about genetic epidemiology 40 were articles about health economics and quality of life47 were diagnostic, prognostic or screening studies31 were methodological papers22 were incidence or prevalence studies3 were observational studies

39 articles reviewed by full text screening

19 eligible articles published until December 31, 2017

212 articles reviewed by abstract screening

1801 articles reviewed by title screening

557

558

559

560

561

Table 1. Characteristics of 18 associations of risk factors for chronic obstructive pulmonary disease.

Reference Risk factor Level of comparison

N cases/ N controls

N studies

Effect size

metric

RE summary effect size (95%

CI)P random

95% prediction

intervalI2

Small-study effects/Excess significance

bias

Credibility

Alif, 2016 44 Biological dust High vs. low exposure 2732/19,367 5 OR 0.99 (0.75 – 1.31) 0.935 0.48 – 2.04 31 No/No NS

Alif, 2016 44 Gases or fumes

High vs. low exposure 2732/19,367 5 OR 1.03 (0.73 – 1.45) 0.886 0.36 – 2.95 59.

6 No/No NS

Alif, 2016 44 Mineral dust High vs. low exposure 2732/19,367 5 OR 0.97 (0.68 – 1.39) 0.874 0.32 – 2.96 61.

1 No/No NS

Byrne, 2015 19 History of tuberculosis Yes vs. no 3682/52,140 8 OR 3.10 (2.24 – 4.31) 1.22 × 10-11 1.10 – 8.75 76 Yes/No Highly

suggestive

Eisner, 2010 9 Second-hand smoking Yes vs. no 3206/88,952 12 OR 1.56 (1.40 – 1.74) 3.48 × 10-16 1.23 – 1.99 27.

2 Yes/Yes Highly suggestive

Gan, 2004 21 Serum CRP High vs. low levels 2877/8885 5 OR 2.33 (1.73 – 3.14) 2.17 × 10-8 0.96 – 5.67 63.

4 No/No Highly suggestive

Gan, 2004 21 Serum leukocytes

High vs. low levels 2715/8795 3 OR 2.23 (1.44 – 3.44) 3.01 × 10-4 0.02 – 322.54 83 No/No Suggestive

Gan, 2004 21 Serum TNFα High vs. low levels 116/71 4 OR 2.96 (1.71 – 5.12) 1.00 × 10-4 0.89 – 9.85 0 No/No Weak

Gan, 2004 21 Serum fibrinogen

High vs. low levels 4279/4752 4 OR 2.42 (1.78 – 3.28) 1.33 × 10-8 0.64 – 9.21 89.

1 No/No Highly suggestive

Hu, 2010 18 Biomass fuels smoking Yes vs. no 3501/36,295 15 OR 2.37 (1.72 – 3.26) 1.25 × 10-7 0.72 – 7.75 83.

6 Yes/Yes Highly suggestive

Jayes, 2016 34 Tobacco smoking

Ever vs. never smokers

6238/105,517 22 OR 3.90 (3.08 – 4.94) 1.52 × 10-29 1.40 – 10.86 89.5 No/No Highly

suggestive

Schikowski, 2014 45

Traffic intensity on nearest road

Per 5,000 vehicles/day increase

100/3378 4 OR 1.30 (0.92 – 1.82) 0.135 0.61 – 2.73 0 No/NA NS

Schikowski, 2014 45

Traffic load on major roads within 100 m

Per 500,000 vehicles/day increase

100/3378 4 OR 1.26 (0.95 – 1.70) 0.103 0.68 – 2.33 0 No/NA NS

Ungprasert, 2016 46

History of psoriasis Yes vs. no NA/NA 7 OR 1.45 (1.21 – 1.73) 4.40 × 10-5 0.80 – 2.61 91 Yes/NA Suggestive

29

562

563


N cases/ N controls

N studies

Effect size

metric


CI)P random

95% prediction

intervalI2

Small-study effects/Excess significance

bias

Credibility

Ungprasert, 2016 20

History of rheumatoid arthritis

Yes vs. no NA/NA 4 OR 1.99 (1.61 – 2.45) 1.67 × 10-10 0.80 – 4.92 80.5 No/NA Highly

suggestive

Waziry, 2016 47 Waterpipe smoking

Ever vs. never 2150/63,628 4 OR 3.18 (1.25 – 8.09) 0.015 0.04 – 241.36 95.

2 No/No Weak

Zhu, 2015 48 Serum vitamin D

High vs. low levels 1981/1283 13 OR 0.29 (0.16 – 0.51) 1.65 × 10-5 0.03 – 2.80 94 Yes/Yes Suggestive

Zhu, 2015 48 Vitamin D deficiency Yes vs. no 3312/6918 12 OR 1.77 (1.18 – 2.64) 0.006 0.47 – 6.65 83.

1 No/Yes Weak

CI: confidence interval, CRP: C-reactive protein, OR: odds ratio, NS: not significant, P random: P value under random-effects model, TNF: Tumor necrosis factor

30

564565

566

567

Table 2: Sensitivity analysis of prospective cohort studies for associations with convincing or highly suggestive evidence


N cases/ N controls

N studies

Effect size

metric


CI)P random

95% prediction

intervalI2 Credibility

Jayes, 2016 34 Tobacco smoking

Ever vs. never smokers

6238/105,517 18 OR 4.28 (3.09 – 5.92) 1.77 × 10-18 1.11 – 16.57 90.1 Highly suggestive

Eisner, 2010 9 Second-hand smoking Yes vs. no 732/5363 3 OR 1.43 (1.24 – 1.65) 1.01 × 10-6 0.53 – 3.87 4 Weak

CI: confidence interval, OR: odds ratio, P random: P value under random-effects model

31

568

569

570

571

572

Documents

Imperial College London · Web viewSeven associations (history of tuberculosis, history of RA, exposure to biomass fuels, tobacco smoking, second-hand smoking, serum CRP, and serum