Our understanding of asthma endotypes and phenotypes has grown substantially in the last decade. Endotype-targeted therapy has become a foundation of management, and classification of patients during initial assessment is extremely important. The use of history, laboratory data, and pulmonary function testing together help to categorize our patients and help guide therapy. One lab test, that of sputum or blood eosinophils, facilitates categorization and has been evaluated for its ability to determine response to medications and predict exacerbations.
In particular, eosinophilia has been extensively studied in severe asthma and is associated with type 2 inflammation. The 2021 GINA guidelines describe type 2 inflammation as characterized by cytokines (especially IL-4, IL-5, and IL-13). “T2-high patients” tend to have elevated blood or sputum eosinophil counts and elevated fractional concentration of exhaled nitric oxide (FENO) and are more likely to respond to biologic therapy. (Global Initiative for Asthma. Global Strategy for Asthma Management and Prevention, 2021).
Courtesy ACCP
Dr. Erin N. Haber
However, what about patients with more mild-to-moderate asthma? Two recent studies have asked this question. In 2020, Pavord and colleagues performed a prespecified secondary subgroup analysis on an open-label randomized control trial comparing prn salbutamol alone to budesonide and as needed salbutamol to as needed budesonide-formoterol. The population was 675 adults with mild asthma receiving only as needed short acting beta-agonists (SABA) at baseline. The primary outcome was annual rate of asthma exacerbation, and whether it was different based on blood eosinophil count, FENO or a composite of both. They had several interesting findings. First, for patients only on an as needed SABA, the proportion having a severe exacerbation increased progressively with increasing blood eosinophil count. Second, inhaled corticosteroids (ICS) plus as needed SABA were more effective than SABA alone in patients with a blood eosinophil count of ≥300 cells/μL, both in terms of total exacerbations and severe exacerbations. The effects of budesonide-formoterol on exacerbations, however, was not associated with blood eosinophil count or FENO. This last point is particularly interesting in light of GINA guidelines that prioritize this combination (Pavord ID et al. Lancet Respir Med. 2020;8[7]:671-80).
Courtesy ACCP
Dr. Daniel B. Jamieson
More recently, a prespecified secondary analysis of the SIENA trial looked at 295 subjects with mild persistent asthma (237 adults aged 18+, and 58 adolescents aged 12-17). The primary outcome was a composite of asthma control (treatment failure, asthma control days, and FEV1). They found that sputum eosinophil levels, blood eosinophil levels, and FENO all predicted response to ICS in adults; however, the area under the receiver operative characteristic curve (AUC) was less than 0.7 for each of these findings, which was below the threshold for acceptability. A blood eosinophil count of ≥100 cells/μL offered 87% sensitivity and 17% specificity for response to ICS (Krishnan JA et al. Ann Am Thorac Soc. 2022;19[3]:372-80).
What does this tell us? Blood eosinophil count may help determine who will respond to ICS, and there remains utility in assessing blood eosinophil count in severe asthma for determining candidacy for biologic therapies. However, the overall utility of blood eosinophils in mild to moderate asthma is not as clear.
But, are we asking the right questions? Many studies look at a single blood eosinophil level, either at a single point in time, a baseline level, or a highest level over a specific time period. But do eosinophil counts vary over time?
A 2018 single-center study initially asked this question. The authors evaluated blood eosinophil levels in 219 adult patients at the NYU/Bellevue Hospital Asthma Clinic over a 5-year period. They found that individual patients had variable eosinophil levels. For example, only 6% (n=13) of patients had levels consistently above 300 cells/μL, but nearly 50% (n=104) had at least one level above 300. The degree of variability was then assessed by K-mean clustering yielding three clusters. Cluster 2 had the largest variability in blood eosinophil counts and a slightly higher absolute eosinophil level. While not significant, there was a suggestion of worse asthma control with more hospitalizations and more prescriptions for multiple controllers in this cluster with more variability. Clearly, this warranted further study (Rakowski E et al. Clin Exp Allergy. 2019;49[2]:163-70).
Variability was re-examined more recently in 2021. A post hoc analysis of two phase III clinical trials from the reslizumab BREATH program looked at eosinophil counts in the 476 patients randomized to receive placebo during the 52-week study. These patients did have eosinophilic asthma by definition and had to have an elevated eosinophil count >400 cells/μL over the 4-week enrollment period to enter the study. However, 124 patients (26.1%) had an eosinophil level <400 cells/μL immediately before the first dose of placebo. The primary outcome was variability in blood eosinophil count. Of patients who started with serum eosinophils <400, 27% to 56% of patients shifted to the ≥400 cells/μL category during the treatment period (this wide range is across three categories of low “baseline” blood eosinophil count; <150, 150 to 300, and 300 to 400). On the contrary, patients who started with eosinophils ≥400 cells/μL tended to stay at that level. The variability is reduced by taking two to three repeat measurements at baseline (Corren et al. J Allergy Clin Immunol Pract. 2021;9[3]:1224-31).
Does this variability have clinical significance? A recent retrospective cohort study looked at 10,059 stable adult patients with asthma from the MAJORICA cohort in Spain, compared with 8,557 control subjects. The primary outcome was total blood eosinophil count and an “eosinophil variability index” (EVI) where EVI=(Eosmax – Eosmin / Eosmax) x 100%. They found that an elevated EVI was associated with hospitalization, more so than maximum eosinophil count or any other eosinophil count variable, with an odds ratio of 3.18 by univariate regression (2.51 by multivariate). They also found that patients with an EVI ≥50% were twice as likely to be hospitalized or visit the ED than those with a lower EVI (Toledo-Pons N et al. Ann Am Thorac Soc. 2022;19[3]:407-14). These results are very interesting and merit further research.
So, what to do with this information? We know that patients with peripheral eosinophilia and severe asthma symptoms are candidates for biologic therapy. They are also more likely to respond to steroids, although the utility of this assessment alone in mild to moderate asthma is less clear. It does seem that more variability in eosinophils over time may be linked to more difficult-to-treat asthma.
Should you check eosinophils in your patients with asthma? GINA 2021 guidelines say to consider it, and list blood eosinophilia as a risk factor for future exacerbation, even if patients have few asthma symptoms. They also say to repeat blood eosinophils in patients with severe asthma, if the level is low at first assessment, based on the studies discussed above. We would agree. We also see the blood eosinophil count as one part of a clinical assessment of a patient’s overall asthma control – even if the patient has mild symptoms. More study on variability is welcome.
Dr. Haber and Dr. Jamieson are with Medstar Georgetown University Hospital, Washington, D.C.