One low eosinophil count does not rule out an eosinophilic COPD phenotype

Design

pooled post hoc analysis of three phase 3 randomised, double-blind, placebo-controlled trials

Population

patients with COPD receiving mepolizumab 100 mg or placebo in METREX, METREO and MATINEE, with historical eosinophil counts

Primary outcome

annualised rate of moderate or severe exacerbations, and of exacerbations needing emergency attendance or admission

Effect

21% reduction with ≥300 cells/µL at any timepoint; 12–27% persistently raised; 22–36% with variable counts; no benefit persistently <150 cells/µL

Blood eosinophil count guides biologic and inhaled corticosteroid decisions in COPD, and it is usually treated as a stable trait measured once. This pooled analysis of the phase 3 METREX, METREO and MATINEE trials asked what happens when you look at how the count behaves over time, using historical values from the 12 months before randomisation alongside screening and baseline measurements.

Mepolizumab reduced, or trended toward reducing, moderate or severe exacerbations in every group with type 2 inflammation: a 21% reduction in those reaching 300 cells/µL at any pre-randomisation timepoint, 12 to 27% in those persistently raised, and 22 to 36% in those whose counts varied over time. From 150 cells/µL upwards, mepolizumab reduced exacerbations requiring emergency attendance or admission. Patients persistently below 150 cells/µL got no benefit.

The variable-count group is the finding. Those are the patients a single measurement misclassifies — the count was 180 on the day it was taken and 420 four months earlier, and current practice records the 180 and closes the question. On these data, intermittent elevation identifies a treatable phenotype as reliably as persistent elevation, and possibly more so. What follows in clinic is cheap: look back through the record for previous full blood counts before concluding a patient is not eosinophilic, and repeat the count rather than treating one value as the answer.

This is a post hoc pooled subgroup analysis, so the specific percentages should not be quoted as effect sizes. The direction is consistent enough across differently defined groups to change how the test is used.

• Look back through previous full blood counts before calling a COPD patient non-eosinophilic
• Repeat a borderline count rather than acting on a single value
• Treat persistent counts below 150 cells/µL as the group without benefit
• Record the range of counts, not just the latest, when referring for a biologic decision

Why it matters

A single eosinophil count is being used to close a question the count is too variable to answer.

Don't overread it

Post hoc pooled subgroup analysis — the percentage reductions are descriptive, not effect estimates from a trial designed to test them.

The statistics, in plain English

These are pooled subgroups defined after the trials ended, so the reduction figures — 12 to 36% depending on the group — describe ranges across several definitions rather than one estimated effect. Confidence intervals are not given for most of them in the abstract, and 'trended toward reduction' means some did not reach significance. The finding that persistently low counts showed no benefit is the most robust part, because it is consistent with the trials' original results.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.