COPD is associated with excess mortality. The biomarker D-dimer has been associated with all-cause mortality in both healthy individuals and COPD patients. We aimed to determine the i) distribution of D-dimer in patients with moderate to severe COPD, ii) association between D-dimer and all-cause mortality, and iii) impact of anticoagulant treatment on this relationship. Based on the Danish registries, we formed a cohort of stable COPD patients with an out-patient D-dimer sample. D-dimer was dichotomized at the upper quartile (0.76 mg Fibrin-Equivalent-Units/L) and was used to evaluate time to death using an adjusted Cox proportional hazards regression. Unadjusted Cox proportional hazards regression showed association between high D-dimer and all-cause mortality, with a hazard ratio (HR) of 2.29 (95% Confidence Interval (CI) 1.08?4.83). In the fully adjusted Cox regression, the HR was 1.73 (CI 0.78?3.84). No significant interactions between D-dimer and anticoagulant or antiplatelet therapy were identified for mortality. Findings for the secondary outcome, days alive and out of hospital were of similar magnitude with similar attenuation throughout adjustment. D-dimer levels did not differ with smoking status but were affected by both anticoagulant and antiplatelet therapy.
Thus, in patients with moderate to severe COPD, patients with high level of D-dimer were more likely to die; however this did not reach significance for the adjusted analyses. Explanations include a true ?no-signal? and type II error.