COPD exacerbation phenotypes in a real-world five year hospitalisation cohort

COPD exacerbation phenotypes have been defined in research populations by predominantly infective or inflammatory aetiology. We sought to characterise this in patients admitted to our centre.

Case-notes of consecutive patients discharged alive after treatment for acute COPD exacerbations between December 2012 and January 2017 were analysed. Data were collected on treatment, length of stay, C-reactive protein (CRP), eosinophil count and bacterial sputum culture positivity for potentially pathogenic microorganisms (PPM).

1029 exacerbations were included. There was an inverse correlation between CRP and eosinophil count (rho = -0.277, p < 0.01). The proportion of eosinophilic exacerbations (eosinophils ≥0.3 × 10⁹/L) was low (157, 15%). Median length of stay was longer in patients with a CRP >100 mg/L (4d [3,8] vs 4d [2,7], p < 0.01) or when given antibiotics (4d [2,8] vs 3d [1,6], p < 0.001) and shorter if receiving corticosteroids (4d [2,6] vs 6d [3,7], p < 0.001). Being sputum culture positive on first exacerbation was associated with sputum culture positivity in subsequent exacerbations. Patients with PPM in sputum culture had a significantly higher median CRP than culture negative patients (38 mg/L [18.75, 57] v 18 mg/L [8.5,45.5] p < 0.05). Length of stay, eosinophil count and CRP were significantly correlated between exacerbation pairs.

This real-world population found eosinophilic and high CRP exacerbations to be distinct and significantly stereotyped within individual patients across recurrent exacerbations. High CRP exacerbations are associated with greater healthcare utilisation and chance of sputum positivity with PPM. Eosinophilic exacerbations were associated with lower rate of readmission. Phenotype-driven treatment warrants further investigation in this population.