Influence of Reliance on Historical Blood Eosinophil Counts on ICS Prescribing by GOLD 2019 Thresholds in COPD

In patients with COPD, higher blood eosinophil count (BEC) predicts a greater reduction in moderate and severe exacerbations in response to inhaled corticosteroid (ICS) therapy. The Global Initiative for Chronic Obstructive Lung Disease strategy (GOLD 2019) recommends the use of BEC to guide ICS therapy and states that eosinophil levels above 300 cells/μL can help identify responders, guiding initial dual therapy, with "little or no effect at a BEC < 100 cells/μL". The National Institute for Health and Care Excellence (NICE) COPD 2018 guideline states that a higher BEC is associated with ICS response, but does not specify a threshold. An earlier post hoc risk-benefit analysis suggested that at lower levels of BEC the harm of ICS due to pneumonia is greater than the benefit of severe exacerbation reduction.

The association between BEC and reduction in exacerbation frequency is based on BEC measured when the patient is clinically stable. Transient eosinopenia (BEC < 50 cells/μL) during severe exacerbation is extremely common. In the Dyspnoea, Eosinopenia, Consolidation, Acidaemia and atrial Fibrillation (DECAF) score derivation and validation studies combined, eosinopenia was present on admission in 1,340 of 2,645 severe exacerbations of COPD (ECOPD) and is associated with longer length of stay (DECAF combined cohorts: median difference 1 day, p<0.001 unpublished data), and higher in-hospital and one year mortality. Although eosinophilic COPD exacerbations occur, overall BEC during moderate or severe exacerbation is lower than stable state. In ECOPD managed in critical care low BEC is associated with higher Acute Physiology, Age, Chronic Health Evaluation II (APACHE II) scores, rate of septic shock and mortality. BEC are also suppressed during other acute illnesses, notably sepsis. Failure to recognise that BEC are often suppressed during acute illness compared to stable state may lead to ICS therapy being inappropriately withheld. Regrettably, whilst the effect of eosinophils on exacerbation is well recognised, the effect of exacerbation and other acute illnesses on eosinophils is under- appreciated. Both NICE and GOLD fail to mention whether BEC should be prospectively measured when patients are stable (reflecting RCT evidence), or if reliance on historical values is acceptable.

In routine practice some clinicians rely on previous BECs to avoid a delay in treatment decisions. A number of these historical counts will have been taken during illness, underestimating the patients' stable-state BEC. Conversely, COPD is associated with multimorbidity, and BEC may be requested for reasons other than acute illness. Using the highest BEC from multiple measures in the previous 24 months may therefore better agree with stable state counts.

The primary aim of this trial is to assess the reliability of using BEC over the preceding 24 months to assess COPD eosinophil phenotype at both GOLD thresholds. The primary outcome will be based on using the highest of at least three BEC. Secondary outcomes include a) the level of agreement between baseline stable state BEC and both mean and the highest BEC over the preceding 24 months, b) the influence of the number of BEC measures available and c) the effect of limiting the time frame from 24 months to the previous 12 months.

Whilst BEC is relatively stable over weeks and months, agreement wanes over longer timeframes. BEC is also associated with disease severity, providing further evidence that COPD eosinophil phenotype may change over time. As an exploratory analysis, periods of sustained change in eosinophil phenotype will be sought, and the relationship between eosinophil phenotype and patient characteristics and medication (including prophylactic azithromycin) will be assessed. Recent work suggests that a low Eosinophil to Basophil ratio predicts exacerbation risk and mortality, whilst a high neutrophil to lymphocyte ratio is related to disease severity, exacerbations and mortality.

The investigators will also assess the relations between the dependent variables stable state absolute eosinophil and basophil counts and both eosinophil to basophil and neutrophil to lymphocyte ratios and the following clinical outcomes: a) moderate and severe exacerbations and b) mortality.

Importance to the NHS

1. This study will assess the variation in BEC measured during both stable state and acute illness in patients with COPD. The investigators anticipate that the results will help prevent inappropriate denial of ICS therapy due to reliance on BEC sampled when a patient is unwell. Both NICE COPD guidelines and the GOLD strategy statement appropriately recommend that BEC should help direct ICS therapy in COPD, but fail to emphasise that the evidence supporting this advice is based solely on BEC measured when the patient is stable. BEC are suppressed during many acute illnesses, including severe exacerbations of COPD, and it is during these episodes that blood tests are most likely to taken. Reliance on BEC sampled during such episodes may lead to inappropriate denial of ICS therapy and increased risk of moderate and severe ECOPD.
2. This study will determine whether the highest of at least three BEC measured over the preceding 24 months provides an acceptable level of agreement with stable state BEC in terms of ICS treatment choice. In a sensitivity analysis, the investigators will assess the influence of using mean rather than highest BEC, the number of BEC available, and timeframe on this outcome.

In the world of realistic medicine, ensuring a patient attends their GP practice or hospital clinic for measurement of BEC when they are stable, and then returns when the result is available before a treatment decision is made is unlikely to be achieved in most cases. However multiple previous BEC are often available, and some are likely to have been measured during recovery or taken at other times when the patient was stable. The clinical systems used in primary and secondary care allow all previous BEC available for a patient to be collated and displayed on screen simply by clicking on a single eosinophil measurement. The outcome of this study will inform clinicians of the level of certainty several historical BEC provide in regard to ICS treatment decisions.