Prognosis of COPD depends on severity of exacerbation history: A population-based analysis

BACKGROUND: Differences in previous exacerbation history may influence prognosis of chronic obstructive pulmonary disease (COPD). We hypothesized that prognosis differs between individuals with a history of only medically treated exacerbations (moderate exacerbations) and those with a history of hospitalised exacerbations (severe exacerbations).

METHODS: We included 98 614 adults from the Copenhagen General Population Study and assessed risk of moderate and severe exacerbations, pneumonia hospitalisation, and respiratory and all-cause mortality from 2003 until 2013 according to exacerbation history.

RESULTS: Among 6545 individuals with COPD, 6290 had no exacerbations in the preceding year, 109 had one moderate exacerbation, 108 had two or more moderate exacerbations, and 38 had one or more severe exacerbations. During 9.4 years of follow-up, we observed 926 moderate and 244 severe exacerbations, 477 pneumonias, and 707 deaths, including 69 from respiratory disease. Compared to individuals without previous exacerbations, lung function and symptom adjusted hazard ratios (HRs) for future moderate exacerbation were 4.68 (95% confidence interval:3.31-6.62) for individuals with one previous moderate exacerbation, 21 (13-33) for individuals with two or more previous moderate exacerbations, and 5.30 (3.44-8.15) for individuals with one or more previous severe exacerbations. Corresponding HRs were 1.62(0.78-3.34), 1.29(0.57-2.89), and 5.43 (2.56-12) for severe exacerbation, 1.86(1.06-3.27), 1.74(1.01-2.99), and 4.85 (2.94-8.02) for pneumonia, 0.53(0.10-2.99), 1.65(0.53-5.17), and 2.98 (1.14-7.83) for respiratory mortality, and 1.34(0.79-2.29), 1.57(1.00-2.47), and 1.49 (0.85-2.62) for all-cause mortality, respectively.

CONCLUSION: Individuals with COPD and a history of hospitalised exacerbations carried the poorest prognosis compared to those with a history of only medically treated exacerbations, suggesting difference in risk profile.