Elucidating the risk factors for chronic obstructive pulmonary disease: an umbrella review of meta-analyses

Chronic obstructive pulmonary disease (COPD) is commonly attributed to smoking, and other potential risk factors are ignored. We aimed to critically appraise the epidemiological credibility of the risk factors for COPD that have been examined in published meta-analyses. We performed a systematic search to capture systematic reviews and meta-analyses of observational studies on environmental factors and biomarkers for risk of COPD. We applied a set of standardised methodological criteria based on the level of statistical significance, sample size, between-study heterogeneity and statistical biases. Our search yielded 11 eligible papers, including 18 meta-analyses on environmental factors or biomarkers for COPD risk, and eight eligible papers with systematic reviews only. Eleven associations achieved statistical significance at P < 0.001 and six associations at P < 1 x 10(-6). Thirteen associations presented an I-2 >= 50%, while six associations had evidence of small-study effects and/or excess significance bias. History of tuberculosis or rheumatoid arthritis, exposure to biomass fuels, tobacco smoking and second hand smoking were supported by high epidemiological credibility for an increased risk of COPD. Furthermore, highly suggestive evidence was found for increased levels of serum C-reactive protein, and serum fibrinogen in COPD patients compared with healthy controls. To summarise, our approach suggests that, while a proportion of COPD patients are nonsmokers, 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.