The Primary Care Respiratory Society

Inspiring best practice in respiratory care

 Facebook LogoTwitter Logo

This website is for healthcare professionals only

Second opinion - Use of blood eosinophil count as criteria for ICS use

Second Opinion.  Your respiratory questions answered…

This question was posted by a PCRS-UK member and is answered by Dr Iain Small and Dr Vince Mak.


I attended the PCRS conference last year (2016) and there was a presentation by a Physician who was talking about the use of blood eosinophil count as a useful criteria for inhaled corticosteroids use in patients with COPD.

He also talked about its potential value in deciding on whether oral steroids are appropriate for some patients when suffering an exacerbation. This spurred me on to look at the way we practice locally, as we seem to have a large number of patients accessing rescue packs with (sometimes) alarming frequency. As a colleague said, “we dish out rescue packs with antibiotics and steroids like sweeties”.

 So, given the increased risk of Pneumonia and other steroid related side effects, should all patient rescue packs contain Oral Cortico-steroid or should we be looking at each patient individually and possibly using blood eosinophil level as a clinical marker to identify a sub-group of patients who should take Prednisolone, and a group for whom risk would outweigh benefit?

We could potentially reduce unnecessary long-term side effects and also make a prescribing cost saving- a win-win situation. I've just started to look at the evidence as to whether this is a reliable enough marker, and how it can be used (if at all) to make these sort of decisions.

So to my questions;

  • are there currently any pathways and local clinical guidelines out there that already incorporate this?
  • Is there any convincing data to support this strategy?  In my literature review, I identified The Hull & east Riding COPD pathway, but they only talk about using blood eosinophil levels as a marker to decide on withdrawal of inhaled steroids, not oral.

ANSWER- a debate

Dr Iain R Small General Practitioner Aberdeenshire Scotland

In brief, there is little evidence to support the use of OCS in COPD exacerbations (Professor Alyn Morice, the Physician to whose presentation you refer pointed this out with aplomb), but we all do it and patients seem to like them.

Working out which are 'infective' and which are 'inflammatory' exacerbations (should such differences genuinely exist) is extremely difficult in real time, and in primary care, I believe it to be almost impossible.

The hypothesis that we can use blood eosinophil levels to help us in this regard has been generated from various post hoc analyses of the data from the recent large clinical trials, whose Primary End Points were NOT directly related to your question. For the most part, the discussion surrounding blood eosinophilia in COPD relates more to trying to decide which patients are likely to gain benefit (rather than suffer harm) from Inhaled Corticosteroids, as compared to maximal bronchodilatation through different physiological and pharmacological pathways.

I don't think at this stage we can say that the evidence supports a clinical strategy to use this marker as a predictor of when to use Oral Corticosteroids (although many eminent researchers in the field might disagree, particularly those who advocate ‘treatable traits’ as a way of identifying patient sub-groups. See Alvar Agusti’s classic paper;  Agusti A et al European Respiratory Journal 2016 47: 410-419; ).  The cost saving you mention would be indirect, as Oral Corticosteoids are inexpensive. Indirect costs savings around Type 2 Diabetes Mellitus and Osteoporosis, for example, do have potential.

There is a second issue worth considering however. There have been a series of recent publications that suggest that there is no such thing as an ‘exacerbator phenotype’ in COPD. These observational cohort studies suggest that some  patients move in and out of an exacerbating behavior pattern over time. This means you can't really pin down a plan for sub sets of the population and reliably expect them to continue to behave the way they previously have.

My pragmatic, non evidence based common sense approach over the past few years has been- If you have a patient who feels fevered with muco-purulent sputum and deteriorating dyspnoea, start with an antibitotic for the first 48 hours (as well as maximal bronchodilation), and add systemic steroids if the patient isn’t responding as quickly as might have been expected. Conversely, a patient who is less systemically unwell but is wheezy might benefit from an earlier introduction of systemic steroid. We don’t always have to use both, in conjunction, in every case.

The patient’s blood eosinophil count is unlikely to hold sway against other clinical variables, and as such, the role for their measurement is, at present, in my opinion, limited.

Dr Vince Mak Consultant Integrated Care Physician, London

I think that there is a lot of data out there that suggest that there is a subset of COPD patients who are not obviously asthmatic, who do have raised blood and/or sputum eosinophilia during exacerbations and there is some evidence to show that these patients do seem to respond better to oral corticosteroids in terms of speed of recovery and protection from future exacerbations and hospital re-admissions.  Professor Dave Singh did a recent big review in the European Respiratory Journal

I also asked Professor Morice at the PCRS UK Conference referred to in the question what effect Inhaled Corticosteroid has on blood eosinophils and he said that they were reduced. This gives us a problem, as looking for eosinophilia in those already on triple therapy is not very helpful.

At present, I think that the pragmatic approach is much as Iain says, if there is infection, then, despite the poor overall evidence base for their use, treat with antibiotics. Of course, reliably identifying bacterial infection in a timely manner is challenging in clinical practice. If the patient doesn’t get better, or if wheeze and shortness of breath are troublesome despite maximal bronchodilators, use Oral Corticosteroids. 

As a clinical team, we do not proactively measure eosinophilia in every patient, but we do look for blood or sputum eosinophils in those who have recurrent exacerbations.

I court a degree of controversy by being someone who still uses long term low dose oral corticosteroids in patients who are frequent exacerbators as this seems to be clinically effective in s small subgroup. The justification for this approach is that it is better for them to have a low dose (say 10mg per day of Prednisolone) long term than frequent high dose courses which are far more damaging to bones and increase problems in those with established Diabetes Mellitus.

With more and more patients popping Oral Corticosteroids from rescue packs they have at home, by the time we see them, sputum and blood eosinophilia will have disappeared, as eosinophils are very sensitive to corticosteroids, and their levels change within hours of introducing them.

Thus the practicality of using them as a marker to drive consistent decision making, as suggested in the original question, is fraught with difficulties and potentially misleading anomalies.

So pragmatically, I agree with Iain, but I do think that eosinophils play a role in determining those patients more likely to benefit from corticosteroids.


Small, I. Mak, V
Clinical resource or information
PCRS Produced / Collaboration
Clinical Area: 
Listing Status: