Stratifying blood eosinophil counts as a biomarker in patients hospitalized for an acute exacerbation of chronic obstructive pulmonary disease (AECOPD) may be beneficial, according to a study published in CHEST.
Previous research has demonstrated that an eosinophil-driven pathway may mediate an AECOPD in up to 30% of cases, with infection rates varying among eosinophil counts.
A derivation cohort of 242 eligible AECOPD admissions was established among 518 confirmed AECOPD admissions in 374 individuals after excluding 132 patients with prehospital oral corticosteroid use. A validation cohort of 115 eligible AECOPD admissions was established among 193 AECOPD admissions in 155 individuals, and was further narrowed to 99 cases after 16 were excluded due to being previously included in the derivation cohort. One-way Kruskal-Wallis statistic (nonparametric data) or analysis of variance (ANOVA, normally distributed data) was used to analyze trends related to clinical outcomes across the low, normal, and high eosinophil groups. Pearson correlation coefficients were used to analyze relationships between different blood biochemistry markers.
Baseline comorbidity and demographic data were similar across all 3 eosinophil groups in both the validation and derivation cohorts. Both cohorts had a similar proportion of exacerbations with low, normal, and high eosinophil counts: 39.3% vs 24.0% vs 36.7% in the derivation cohort and 35.3% vs 25.3% vs 39.4% in the validation cohort. Inhaled corticosteroid prescriptions were high across both cohorts prior to admission (77.8% across both), with no difference observed in median dose or frequency. Infection was detected in 71.4% of the validation cohort and 84.2% in the retrospective cohort, but was far less common across blood eosinophil groups (<50/µL, 90.6%; normal, 66.7%; >150/µL, 51.9%; P =.004 in validation cohort).
In both cohorts, significantly higher eosinophil counts were seen in the AECOPD group not associated with infection (validation cohort: median 185/µL vs 40/µL, respectively; P <.001). Eosinophil counts <50/µL were associated with the highest total white cell counts (WCC), C-reactive protein (CRP), and neutrophil counts, while high eosinophil counts (>150/µL) were associated with the lowest WCC, CRP levels, and neutrophil counts. Blood eosinophil counts correlated negatively with CRP (r = -0.262 and r = -0.244 in the respective cohorts; P <.001). A significant difference in hospital length of stay was identified between eosinophil groups in both the validation (P <.001) and the derivation cohorts (P =.04), which remained after excluding cases with infiltrates on chest X-rays (P <.001). Increasing the cutoff to define eosinophilia from >150/µL to >300/µL to >500/µL reduced the proportion of patients classified with high eosinophil counts from 30.2% (>150/µL) to 22.3% (>300/µL) to 10.9% (>500/µL). The between-group associations with CRP measurements, neurophil counts, and clinical outcomes were unchanged.
Study investigators concluded, “[O]ur data demonstrate for the first time the benefits of stratifying blood eosinophil counts as a biomarker in hospitalised AECOPD. Studies are needed to determine whether low, normal, and high blood eosinophils can be integrated into AECOPD management algorithms to target populations for discriminate therapy with antibiotics and/or corticosteroids.”
Disclosures: Funding support for this study was provided by an unrestricted educational grant from GlaxoSmithKline.
MacDonald MI, Osadnik CR, Bulfin L, et al. Low and high blood eosinophil counts as biomarkers in hospitalised acute exacerbations of chronic obstructive pulmonary disease [published online April 9, 2019]. CHEST. doi:10.1016/j.chest.2019.02.406