Background Acute exacerbations of chronic obstructive pulmonary disease (AECOPD) are a common cause of hospitalization. Recent studies have shown that eosinopenia, neutrophil-lymphocyte ratio (NLR) and c-reactive protein (CRP) can be predictors of short-term mortality, with no consensual cut-offs. A recent prospective study has shown that eosinopenia could be helpful predicting ICU admission.

Aims and objectives To evaluate the usefulness of inflammatory markers as predictors of poor outcomes in our population.

Methods Retrospective analysis of AECOPD in a pulmonology ward of a secondary hospital, between 2017-2021. Patients without FEV1/FVC <0.7 were excluded. We collected demographic data, eosinophils, neutrophil, lymphocyte and CRP at admission, mechanical ventilation (MV) and mortality. Eosinopenia was defined as <40 cel/uL and a cut-off of 10.2 was used for NLR.

Results 137 patients were enrolled (mean age 66 ± 9.6 years; 78.1% males). 3.6% patients were submitted to MV and 8.8% died during hospitalization. 31.4% had eosinopenia at admission, the median NLR and CRP was 6.0 and 2.6 mg/dL, respectively. MV rates were significantly higher in patients with eosinopenia (p 0.017) and in patients with higher NLR (9.4% vs 1.9%, p 0.049), with no significant difference in CRP level (median 13.3 vs 6.0, p 0.748). Regarding in-hospital mortality, there was no significant difference between the groups with or without eosinopenia, higher NLR and CRP.

Conclusions Eosinopenia and high NLR at admission might be predictors of MV need in these patients. These tools are fast, easy and cost-effective and can alert us which patients may have a poor evolution needing MV.