Table 1 Summary of evidence of reviewed biomarkers in relation to questions raised by Sin & Vestbo [4]

Is there a strong biological plausibility in terms of its role in pathogenesis of disease?

Evidence from animal studies and gene-association studies [10]-[19]

Suggested from in vitro, animal study and gene-association studies [18],[52],[53]

N.A.

Experimental data suggest role in systemic effects and comorbidity [81]

Experimental data suggest role in systemic effects and comorbidity [108],[109]

Is there a strong, consistent and independent association between the biomarker and COPD?

Level IIb[16],[20]-[35]

Level IIb [33],[45],[63]-[66]

Level III [77]-[79]

Conflicting results from large population studies [78],[85]-[99]

Level IIb [32],[87],[112]-[122],[125]

Is there a strong, independent association between the biomarker and hard clinical outcomes such as mortality and hospitalisations?

Level IIa [36],[37]

No evidence [37]

Level IIb [79]

Level IIa; on all cause mortality [84],[86],[99]

Level IIa [37],[112]-[114],[126]

Is there evidence from randomised controlled trials that the biomarker is modifiable by interventions?

Evidence from 3 cohort studies of prednisolone treatment [25],[38],[39]

Evidence from one RCT of TNF-R antibody treatment [67] and one RCT in salmeterol/fluticasone propionate-arm [68]

Evidence from one RCT with prednisolone treatment [79]

Evidence from one RCT with inhaled glucocorticoid, prednisolone or placebo [100]

Evidence from one RCT with p38 MAPK inhibitor[68],[127]

Is there evidence from randomised controlled trials that changes in the biomarker status results in changes in an important (and accepted) clinical outcome (e.g. mortality, exacerbations, rate of decline in FEV₁ and health status)?

N.A.

N.A.

N.A.

N.A.

N.A.