To the Editor:
Recently, two post hoc analyses of clinical trials in moderate to severe asthma showed that fractional exhaled nitric oxide (FeNO) and the blood eosinophil count provide additive prognostic information on the occurrence of severe asthma attacks (1, 2). The effect is large, with a threefold increased risk in attacks seen in patients with FeNO ⩾50 ppb and blood eosinophils ⩾0.3 × 109/L compared with those with a FeNO <25 ppb and blood eosinophils <0.15 × 109/L (3). Importantly, this risk can be reduced with type 2 cytokine and alarmin-directed biologic agents (4–6). The additive, independent, and differentially modifiable risk associated with these biomarkers suggests that they identify different yet complementary aspects of type 2 airway inflammation.
Although raised FeNO classically identifies corticosteroid responsiveness, the advent of FeNO suppression testing for uncontrolled type 2–high asthma has proved that a third of patients have corticosteroid-resistant elevations in FeNO—and disease burden—despite objective evidence of treatment adherence (7, 8). FeNO nonsuppression provides a convenient model to control for nonadherence and independently study corticosteroid resistance in severe asthma.
We tested the hypothesis that FeNO and blood eosinophils relate differently to inflammation observed in the sputum (reflecting airway) and blood (reflecting systemic) compartments. An important feature of our approach was to study patients in whom we had a high degree of confidence in treatment adherence to high-dose inhaled corticosteroids and/or systemic corticosteroids.
Methods
Induced sputum eosinophils and sputum plus serum mediators were analyzed in a pooled cross-sectional analysis of patients with severe asthma and healthy control subjects.
We included patients with severe asthma who had sputum analyzed after a FeNO suppression test (8) or the RASP-UK (UK Refractory Asthma Stratification Programme) trial (NCT02717689) (9). Adherence was verified using different approaches. The FeNO suppression cohort underwent remotely monitored inhaled corticosteroids via a chipped inhaler and, if FeNO was suppressed by <42% by Day 7, a nurse-administered triamcinolone injection (8). The RASP-UK cohort underwent 8-weekly biomarker or clinically guided treatment advisories for 1 year (9) followed by a range of objective adherence measurements (prescription refills; cortisol and prednisolone blood concentrations if applicable; FeNO suppression testing if FeNO elevated) before being recruited for the associated bronchoscopy study (NCT02883530). Healthy control subjects were nonsmokers, reported no atopy or lung disease, and had normal lung function. All subjects provided written informed consent in ethically approved studies.
Patients and control subjects underwent same-day detailed clinical assessment, sputum induction, and phlebotomy when on maximum intensity treatment; only the FeNO suppression protocol included serum. Twenty-six sputum, serum, and clinical measurements were assessed (Table 1). Inflammatory proteins were measured in duplicates using multiplex electrochemiluminescent assays (Meso Scale Discovery) or single ELISAs (Cayman Chemical). Spearman correlations were computed between FeNO, blood eosinophils, and analytes, controlling for a false discovery rate <0.05. To translate significant correlations, Jonckheere-Terpstra ordinal trend tests were performed across FeNO (<25, 25 to <50, and ⩾50 ppb) and blood eosinophils (<0.15, 0.15 to <0.3, and ⩾0.3 × 109/L) categories. Statistical analyses were performed using SPSS v27 with a two-sided α of 0.05.
Table 1.
Analytes according to FeNO and Blood Eos–based Stratification Strategies
| Analyte (LLOD)* | FeNO (ppb) |
Blood Eos (×109/L) |
Healthy Control Subjects (n = 10) | ||||||
|---|---|---|---|---|---|---|---|---|---|
| <25 (n = 17) | 25 to <50 (n = 30) | ⩾50 (n = 27) | r (P Value) | <0.15 (n = 21) | 0.15 to <0.30 (n = 22) | ⩾0.30 (n = 31) | r (P Value) | ||
| Biomarker | |||||||||
| FeNO, ppb | 16 (13–20) | 39 (32–42) | 83 (60–123)† | — | 38 (23–55) | 38 (26–74)† | 45 (25–89)† | 0.24 (0.04) | 19 (11–28) |
| Blood Eos, ×109/L | 0.17 (0.1–0.54) | 0.24 (0.1–0.35) | 0.26 (0.19–0.55)† | 0.24 (0.04) | 0.09 (0.05–0.12) | 0.23 (0.19–0.25) | 0.54 (0.36–0.66)† | — | 0.14 (0.09–0.18) |
| Sputum | |||||||||
| Eos, % | 0.8 (0.4–5.3) | 2.7 (1.1–17.8)† | 12.8 (3.3–35.5)† | 0.51 (0.0002) | 2.7 (0.7–6.1) | 5.1 (0.5–30.5)† | 4.3 (1–21)† | NS | 0.3 (0.3–0.4) |
| IL-4 (0.2) | 0.1 (0.1–0.3) | 0.4 (0.1–1.1)† | 0.8 (0.2–1.2)† | 0.48 (<0.0001) | 0.3 (0.1–1)† | 0.4 (0.1–0.9) | 0.3 (0.1–1)† | NS | 0.1 (0.1–0.1) |
| IL-5 (0.5) | 1.2 (0.4–4.6) | 4.6 (1.9–7.8) | 10.9 (2.9–29.8)† | 0.47 (0.0002) | 2.3 (1.1–9.7) | 5.3 (1.5–15.1)† | 4.7 (1.8–10.8)† | NS | 0.3 (0.2–2.7) |
| IL-13 (4.2) | 6.4 (2.1–8.8) | 7 (5.8–14.2)† | 8.4 (6.4–13.9)† | 0.26 (0.04) | 7 (5.1–11.5)† | 8.3 (4–12.5)† | 7.6 (6–12.2)† | NS | 2.1 (2.1–2.1) |
| IL-33 (0.6) | 0.9 (0.3–1.3) | 0.9 (0.3–2.1)† | 1.7 (0.7–2.9)† | 0.35 (0.006) | 0.9 (0.3–1.9) | 1.4 (0.5–2.6)† | 1 (0.3–2.3)† | NS | 0.3 (0.3–0.3) |
| TSLP (0.9) | 2.4 (1–9.3) | 6.4 (2.3–10.7)† | 11.9 (5–20.7)† | 0.41 (0.001) | 4.9 (1.5–16.9) | 9.1 (1.9–2.6) | 7.1 (2.5–15)† | NS | 0.9 (0.5–1.8) |
| Eotaxin-3 (4.2) | 34 (2–71) | 133 (23–369)† | 353 (245–804)† | 0.55 (<0.0001) | 76 (23–264) | 215 (9–418)† | 191 (29–390)† | NS | 2 (2–26) |
| TARC (0.4) | 17 (9–89) | 27 (18–77) | 58 (38–301)† | 0.32 (0.02) | 35 (19–107)† | 41 (9–101) | 36 (17–88) | NS | 6 (2–21) |
| LTE4 (7.8) | 59 (23–114) | 138 (42–465)† | 133 (42–730)† | NS | 64 (23–139) | 94 (48–343) | 163 (49–676)† | NS | 7 (4–70) |
| PGD2 (19.5) | 241 (173–384) | 217 (119–354) | 209 (135–439) | NS | 213 (133–505) | 219 (183–389) | 222 (117–439) | NS | 89 (43–200) |
| IFN-γ (0.3) | 0.3 (0.2–0.5) | 0.4 (0.2–1.8) | 0.6 (0.2–1.5) | NS | 0.5 (0.2–1.7) | 0.4 (0.2–2.6) | 0.3 (0.2–0.8) | NS | 0.2 (0.2–2.1) |
| TNF (0.4) | 1.5 (0.4–10.2) | 2 (0.8–7.5) | 3.3 (1.5–6.7) | NS | 2.5 (1–6.7) | 3.2 (0.5–8.5) | 2 (0.7–8.6) | NS | 2.9 (0.4–16.7) |
| Serum | |||||||||
| IL-4 (0.1) | 0.1 (0.1–0.1) | 0.1 (0.1–0.1) | 0.1 (0.1–0.1) | NS | 0.1 (0.1–0.1) | 0.1 (0.1–0.1) | 0.1 (0.1–0.1) | NS | 0.1 (0.1–0.1) |
| IL-5 (0.4) | 1.1 (1–1.2)† | 0.6 (0.5–0.9)† | 0.6 (0.4–1.6)† | NS | 0.4 (0.4–0.6) | 0.6 (0.6–1.6)† | 0.8 (0.6–1.5)† | 0.41 (0.03) | 0.2 (0.2–0.4) |
| IL-13 (6.7) | 3.3 (3.3–3.3) | 3.3 (3.3–9.9) | 3.3 (3.3–14.1) | NS | 3.3 (3.3–3.3) | 8.8 (3.3–13.3) | 3.3 (3.3–10) | NS | 9.2 (7.8–9.8) |
| IL-33 (0.4) | 0.2 (0.2–0.3) | 0.8 (0.2–0.8) | 0.2 (0.2–0.8) | NS | 0.2 (0.2–0.8) | 0.4 (0.2–0.8) | 0.2 (0.2–0.8) | NS | 0.4 (0.4–0.4) |
| TSLP (0.5) | 1.3 (1–2.1) | 1.8 (0.6–2.5) | 2.7 (1.8–4.5)† | NS | 1.7 (1.2–3.6) | 2.3 (1.9–4.4)† | 1.8 (0.8–2.7) | NS | 1.4 (0.9–1.5) |
| Eotaxin-3 (4.2) | 9 (6–30) | 18 (7–32) | 16 (13–29) | NS | 18 (14–31) | 29 (12–34)† | 13 (7–18) | NS | 8 (4–13) |
| TARC (0.2) | 427 (108–571) | 252 (147–463) | 226 (89–430 | NS | 314 (195–664) | 190 (92–252) | 344 (146–480) | NS | 206 (124–285) |
| IFN-γ (0.3) | 0.4 (0.4–0.5) | 0.5 (0.2–0.7) | 0.3 (0.2–1.2) | NS | 0.3 (0.2–0.6) | 0.7 (0.2–2.3) | 0.2 (0.2–0.6) | NS | 0.4 (0.2–0.7) |
| TNF (0.4) | 1.8 (0.9–3.8)† | 0.6 (0.2–2) | 1.2 (0.2–4.2)† | NS | 1.7 (0.2–2.3)† | 0.6 (0.2–2) | 0.9 (0.3–1.9)† | NS | 0.2 (0.2–0.3) |
| Clinical | |||||||||
| ACQ-5 score | 1.2 (0.5–1.8) | 1.6 (0.2–2.2) | 2 (0.8–3) | NS | 1.6 (0.5–2.1) | 1.7 (0.7–2.9) | 1.2 (0.6–2.2) | NS | — |
| FEV1, % predicted | 88 (78–103) | 85 (75–98) | 81 (72–96) | NS | 81 (77–94) | 83 (74–97) | 85 (76–99) | NS | — |
| FEV1/FVC, % | 72 (64–82) | 68 (61–79) | 72 (60–77) | NS | 71 (62–82) | 68 (61–77) | 72 (61–80) | NS | — |
| Asthma attacks in the past year | 1 (0–3) | 1 (0–4) | 3 (0–5) | 0.25 (0.03) | 1 (0–5) | 1.5 (0–4) | 1 (0–4) | NS | — |
Definition of abbreviations: ACQ-5 = five-item Asthma Control Questionnaire; Eos = eosinophils; FeNO = fractional exhaled nitric oxide; LLOD = lower limit of detection; LTE4 = leukotriene E4; NS = not significant; PGD2 = prostaglandin D2; TARC = thymus activation–regulated cytokine (CCL17); TNF = tumor necrosis factor; TSLP = thymic stromal lymphopoietin.
P values ⩾0.05 are not significant.
Data are presented as median (interquartile range) in pg/ml unless stated otherwise.
Spearman correlation coefficients (r) and associated P values are in bold if retained after controlling for a false discovery rate <0.05 across the 52 computed correlations.
Cytokine levels that were not quantified were assigned the arbitrary value of 0.5 × the lower limit of detection to allow analysis.
Adjusted P value <0.05 compared with healthy control subjects on Kruskal-Wallis test adjusted for six comparisons.
Results
We included 74 patients with severe asthma and 10 healthy control subjects. Patients included from the FeNO suppression cohort (n = 34) and RASP-UK cohort (n = 40) were similar. Patients with asthma were 55% male, 74% atopic, and 85% nonsmokers. The mean (±SD) age was 53 ± 15 years; the mean Asthma Control Questionnaire score was 1.6 ± 1.2; the mean beclomethasone dipropionate-equivalent dose was 2,391 ± 1,084 μg/d; the mean post-bronchodilator FEV1 was 85 ± 19% predicted; the mean FEV1/FVC ratio was 70 ± 11%; and 53% were assessed on systemic corticosteroids. There were 60 sputum supernatants and 30 serum samples available for analysis in asthma.
We observed significant correlations between FeNO and sputum eosinophils, IL-4, IL-5, and IL-33, TSLP (thymic stromal lymphopoietin), eotaxin-3, TARC (thymus activation–regulated cytokine), and asthma attacks in the past year. Blood eosinophils correlated with serum IL-5 (Table 1). We observed no correlation between the Asthma Control Questionnaire score and the 26 analytes. Sputum eosinophils inversely correlated with lung function and closely mirrored the correlations observed with FeNO (Figure 1).
Figure 1.
Correlation matrix for FeNO, blood Eos, and selected analytes in severe asthma. Bold Spearman coefficient of correlations (r) and P values were those retained after controlling for a false discovery rate <0.05 in the primary analysis (first two columns and rows); the rest of the matrix is exploratory. Asthma attacks are defined as acute events requiring ⩾3 days of systemic corticosteroids in the past year. ACQ-5 = five-item Asthma Control Questionnaire; Eos = eosinophils; FeNO = fractional exhaled nitric oxide; ns = not significant (P ⩾ 0.05); TARC = thymus activation–regulated cytokine (CCL17); TSLP = thymic stromal lymphopoietin.
FeNO nonsuppression was associated with higher sputum eosinophils (fold difference in median values, FeNO <25 to ⩾50 ppb: 17-fold, P for trend = 0.001), IL-4 (7.6-fold, P = 0.0006), IL-5 (8.9-fold, P = 0.006), IL-33 (1.8-fold, P = 0.02), TSLP (5-fold, P = 0.002), eotaxin-3 (10-fold, P = 0.00003), TARC (3.5-fold, P = 0.005), and asthma attacks in the past year (3-fold, P = 0.03). Greater blood eosinophils (<0.15 to ⩾0.3 × 109/L) was associated with higher serum IL-5 (1.9-fold, P = 0.04) (Table 1).
The highest FeNO and blood eosinophil categories generally had greater sputum eosinophils, sputum/serum type 2 cytokine, and chemokine and alarmin levels than healthy control subjects (Table 1).
The directions of trends were consistent when removing systemic corticosteroid–treated patients or when separating the RASP-UK and FeNO suppression cohorts. Exploratory multiple regression showed no additive effect for biomarkers to identify inflammation levels.
Discussion
We found that in severe asthma, FeNO nonsuppression identifies increased airway type 2 cytokines (IL-4 and IL-5), chemokines (eotaxin-3 and TARC), alarmins (IL-33 and TSLP), and sputum eosinophilia. In contrast, blood eosinophils correlate with serum IL-5 and not with any assessed measure of airway inflammation. We base these conclusions on our cross-sectional study of patients with extremely high corticosteroid exposure and proven adherence.
Our results are consistent with the cross-sectional bronchial biopsy-based ADEPT study (10) but extend their findings by showing correlations between FeNO and almost all of the assessed components of the airway type 2 immune response for a population with confirmed treatment adherence. The most striking finding of our study was the different relationship between FeNO, blood eosinophils, and markers of airway and systemic type 2 inflammation. Our findings imply that FeNO and blood eosinophils relate to different components and compartments of type 2 inflammation: FeNO reflects airway type 2 activity and the chemotactic pull to the airways, whereas blood eosinophils reflect the systemic pool of available effector cells and circulating IL-5.
Our study has several limitations. Its cross-sectional design assessed correlation, not causality. The analysis of serum analytes was underpowered (β = 0.43 for r = 0.40 with critical P < 0.041), and we pooled two cohorts that used different approaches to confirm treatment adherence, although a sensitivity analysis analyzing both independently was supportive of our results. Unexpectedly, sputum IL-13 did not correlate with FeNO after controlling for multiplicity of testing. This may reflect the complex dimeric receptor system signaling both IL-4/-13, a greater steroid-sensitivity of IL-13, and/or a slightly underpowered analysis.
To conclude, we found that FeNO and blood eosinophils provide different and complementary mechanistic information in severe asthma. How airway signaling (reflected by FeNO) and an increased systemic eosinophil pool (reflected by blood eosinophils) relate to the pathogenesis of asthma attacks and the response to treatment remains an important question.
Acknowledgments
Acknowledgment
The authors thank Catherine Borg, Clare Connolly, Tilly Downs, Beverley Hargadon, Anna Hayman, Gareth Hynes, Karolina Krassowska, Angela Moran, Sophie Morgan, Sarah Poole, Timothy Powell, and Samantha Thulborn for participation in data collection and/or specimen processing; the participants for their time and generosity; and Sanjay Ramakrishnan for manuscript revision. This study includes data from www.clinicaltrials.gov trial number NCT 02883530.
Footnotes
Supported by a nonrestricted research grant from Sanofi Genzyme for investigator-initiated type 2 innovation research; by the National Institute for Health Research (NIHR Leicester Biomedical Research Centre and NIHR Oxford Biomedical Research Centre); by the Medical Research Council (MRC Refractory Asthma Stratification Programme); and in part by the Wellcome Trust (211050/Z/18/Z) and Beit Fellowship (211050/Z/18/A). The views expressed are those of the authors and not necessarily those of the National Health Service, the NIHR, or the Department of Health.
Author Contributions: S.C. collated the data, analyzed specimens, and drafted and approved the final manuscript. R.S. participated in data collection and specimen analysis and approved the final manuscript. R.C., A.H.M., L.P.M., L.G.H., S.J.F., P.B., T.S.C.H., and I.D.P. participated in data collection and approved the final manuscript. T.S.C.H. participated in manuscript preparation, approved the final publication, and is the guarantor of this publication.
Originally Published in Press as DOI: 10.1164/rccm.202104-1040LE on June 15, 2021
Author disclosures are available with the text of this letter at www.atsjournals.org.
Contributor Information
Collaborators: in collaboration with the MRC UK Refractory Asthma Stratification Programme (RASP-UK)
References
- 1. Shrimanker R, Keene O, Hynes G, Wenzel S, Yancey S, Pavord ID. Prognostic and predictive value of blood eosinophil count, fractional exhaled nitric oxide, and their combination in severe asthma: a post hoc analysis. Am J Respir Crit Care Med. 2019;200:1308–1312. doi: 10.1164/rccm.201903-0599LE. [DOI] [PubMed] [Google Scholar]
- 2. Busse W, Wenzel S, Bateman E, Casale T, FitzGerald J, Rice M, et al. Baseline FeNO as a prognostic biomarker for subsequent severe asthma exacerbations in patients with uncontrolled, moderate-to-severe asthma receiving placebo in the LIBERTY ASTHMA QUEST study: a post hoc analysis. Lancet Respir Med. doi: 10.1016/S2213-2600(21)00124-7. [DOI] [PubMed] [Google Scholar]
- 3. Couillard S, Laugerud A, Jabeen M, Ramakrishnan S, Melhorn J, Hinks T, et al. Derivation of a prototype asthma attack risk scale centred on blood eosinophils and exhaled nitric oxide. Thorax. doi: 10.1136/thoraxjnl-2021-217325. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Castro M, Corren J, Pavord ID, Maspero J, Wenzel S, Rabe KF, et al. Dupilumab efficacy and safety in moderate-to-severe uncontrolled asthma. N Engl J Med. 2018;378:2486–2496. doi: 10.1056/NEJMoa1804092. [DOI] [PubMed] [Google Scholar]
- 5. Pavord ID, Korn S, Howarth P, Bleecker ER, Buhl R, Keene ON, et al. Mepolizumab for severe eosinophilic asthma (DREAM): a multicentre, double-blind, placebo-controlled trial. Lancet. 2012;380:651–659. doi: 10.1016/S0140-6736(12)60988-X. [DOI] [PubMed] [Google Scholar]
- 6. Corren J, Parnes JR, Wang L, Mo M, Roseti SL, Griffiths JM, et al. Tezepelumab in adults with uncontrolled asthma. N Engl J Med. 2017;377:936–946. doi: 10.1056/NEJMoa1704064. [DOI] [PubMed] [Google Scholar]
- 7. Heaney LG, Busby J, Bradding P, Chaudhuri R, Mansur AH, Niven R, et al. Medical Research Council UK Refractory Asthma Stratification Programme (RASP-UK) Remotely monitored therapy and nitric oxide suppression identifies nonadherence in severe asthma. Am J Respir Crit Care Med. 2019;199:454–464. doi: 10.1164/rccm.201806-1182OC. [DOI] [PubMed] [Google Scholar]
- 8. McNicholl DM, Stevenson M, McGarvey LP, Heaney LG. The utility of fractional exhaled nitric oxide suppression in the identification of nonadherence in difficult asthma. Am J Respir Crit Care Med. 2012;186:1102–1108. doi: 10.1164/rccm.201204-0587OC. [DOI] [PubMed] [Google Scholar]
- 9. Heaney LG, Busby J, Hanratty CE, Djukanovic R, Woodcock A, Walker SM, et al. investigators for the MRC Refractory Asthma Stratification Programme. Composite type-2 biomarker strategy versus a symptom-risk-based algorithm to adjust corticosteroid dose in patients with severe asthma: a multicentre, single-blind, parallel group, randomised controlled trial. Lancet Respir Med. 2021;9:57–68. doi: 10.1016/S2213-2600(20)30397-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Silkoff PE, Laviolette M, Singh D, FitzGerald JM, Kelsen S, Backer V, et al. Airways Disease Endotyping for Personalized Therapeutics (ADEPT) study investigators. Identification of airway mucosal type 2 inflammation by using clinical biomarkers in asthmatic patients. J Allergy Clin Immunol. 2017;140:710–719. doi: 10.1016/j.jaci.2016.11.038. [DOI] [PubMed] [Google Scholar]