High blood eosinophil counts predict sputum eosinophilia in patients with severe asthma

In asthmatic patients sputum eosinophilia predicts steroid response1Green R.H. Brightling C.E. McKenna S. Hargadon B. Parker D. Bradding P. et al.Asthma exacerbations and sputum eosinophil counts: a randomised controlled trial.Lancet. 2002; 360: 1715-1721Abstract Full Text Full Text PDF PubMed Scopus (1514) Google Scholar and identifies an inflammatory phenotype that responds to anti-TH2 biologic agents, such as mepolizumab.2Nair P. Pizzichini M.M. Kjarsgaard M. Inman M.D. Efthimiadis A. Pizzichini E. et al.Mepolizumab for prednisone-dependent asthma with sputum eosinophilia.N Engl J Med. 2009; 360: 985-993Crossref PubMed Scopus (1189) Google Scholar However, sputum induction is not without risk in patients with severe asthma nor is it always successful, and processing requires expertise and is time consuming. A recent study from the Severe Asthma Research Program analyzed same-day blood and sputum eosinophil counts, concluding that the former cannot be reliably used to predict the latter.3Hastie A.T. Moore W.C. Li H. Rector B.M. Ortega V.E. Pascual R.M. et al.Biomarker surrogates do not accurately predict sputum eosinophil and neutrophil percentages in asthmatic subjects.J Allergy Clin Immunol. 2013; 132: 72-80Abstract Full Text Full Text PDF PubMed Scopus (201) Google Scholar Cutoffs to define sputum eosinophilia of 2% or greater and blood eosinophilia of 0.3 × 109 cells/L or greater were chosen. The upper limit of the normal range for sputum eosinophil counts in the healthy (nonasthmatic) population is 1.9% or less,4Reddel H.K. Taylor D.R. Bateman E.D. Boulet L.P. Boushey H.A. Busse W.W. et al.An official American Thoracic Society/European Respiratory Society statement: asthma control and exacerbations: standardizing endpoints for clinical asthma trials and clinical practice.Am J Respir Crit Care Med. 2009; 180: 59-99Crossref PubMed Scopus (1487) Google Scholar and therefore 2% seems appropriate, especially where most patients are already treated with eosinophil-suppressing corticosteroids. For blood eosinophil counts, however, a higher cutoff of normality of greater than 0.45 × 109 cells/L is more commonly used, both in our hospital and in the literature, although this is by no means universal.5Fulkerson P.C. Rothenberg M.E. Targeting eosinophils in allergy, inflammation and beyond.Nat Rev Drug Discov. 2013; 12: 117-129Crossref PubMed Scopus (340) Google Scholar Visual inspection of the graphic data provided in the article by Hastie et al3Hastie A.T. Moore W.C. Li H. Rector B.M. Ortega V.E. Pascual R.M. et al.Biomarker surrogates do not accurately predict sputum eosinophil and neutrophil percentages in asthmatic subjects.J Allergy Clin Immunol. 2013; 132: 72-80Abstract Full Text Full Text PDF PubMed Scopus (201) Google Scholar suggests that this higher cutoff might have a higher positive predictive value for sputum eosinophilia and potentially could be used to classify the airway inflammatory phenotype in a subgroup of patients with severe asthma and hence remove the requirement for sputum induction and processing in some patients. We run a tertiary severe asthma service and routinely analyze sputum cell counts to help characterize the airway status of newly referred patients and to facilitate management decisions in returning patients. Blood is often collected on the same visit, and therefore we investigated whether in our cohort blood eosinophilia could predict sputum eosinophilia and whether a higher cutoff could increase the clinical utility of the test. We collected data from our clinical database of patients attending the Manchester Severe Asthma Service between 2006 and 2013. All patients were receiving at least 1000 μg/d inhaled beclomethasone or equivalent with or without 5 mg of oral prednisone and with or without intramuscular triamcinolone. Sputum was collected either as a spontaneous sample or induced by using hypertonic saline and processed, as previously described.6Pizzichini E. Pizzichini M.M. Efthimiadis A. Hargreave F.E. Dolovich J. Measurement of inflammatory indices in induced sputum: effects of selection of sputum to minimize salivary contamination.Eur Respir J. 1996; 9: 1174-1180Crossref PubMed Scopus (246) Google Scholar A venous blood sample was collected and analyzed during the same episode. For practical reasons, blood samples were sometimes collected up to 2 days on either side of the sputum visit. Data presented are descriptive, with performance characteristics also reported for the ability of blood eosinophil counts of 0.15 × 109 cells/L or greater, 0.3 × 109 cells/L or greater, and greater than 0.45 × 109 cells/L to predict sputum eosinophilia of 2% or more. Data were analyzed with SPSS software (version 20; SPSS, Chicago, Ill). Complete data sets were available from 163 patients providing samples on 202 sample days. Blood samples were collected on the same day as the sputum sample on all but 15 occasions. Demographic details (at the index visits where patients provided multiple samples) were mean age of 46.6 years (SD, 12.2 years), 56% female sex, 63% receiving systemic corticosteroids, 9% receiving second-line systemic immunosuppressants, and mean inhaled beclomethasone dipropionate dose equivalent of 2293 μg/d (SD, 1158 μg/d). Blood and sputum eosinophil counts correlated (Spearman correlation = 0.493, P < .001). Values for each of the 202 blood-sputum pairs are plotted in Fig 1, and a receiver operating characteristic curve with coordinates is presented in Fig E1 and Table E1 in this article's Online Repository at www.jacionline.org. Changing the threshold from 0.3 to 0.45 × 109 cells/L improved the positive predictive value for sputum eosinophilia from 65.6% to 89.2% and the specificity from 84.4% to 97%. There was a marginal loss of sensitivity/negative predictive value (Table I). A blood eosinophil cutoff of 0.15 × 109 cells/L had a positive predictive value of less than 50%.Table IPerformance characteristics (percentages) for prediction of at least 2% sputum eosinophilia in patients with moderate and severe asthma by using blood eosinophil count cutoffs of 0.15, 0.3, and 0.45 × 109 cells/LBlood eosinophil cutoff (× 109 cells/L)≥0.15≥0.30>0.45Sensitivity77.659.749.3Specificity53.384.497.0Positive predictive value45.265.689.2Negative predictive value82.880.979.4 Open table in a new tab In our cohort of patients with severe asthma, the majority receiving very high doses of inhaled or systemic corticosteroids, a blood eosinophil count of greater than 0.45 × 109 cells/L can correctly predict sputum eosinophilia in 9 of 10 cases. We have included only patients with severe asthma in our study, and therefore we believe our results are important because these are the patients who will be considered for novel TH2-targeting biologic agents. The search for biomarkers of airway inflammation has been made more urgent by the advance of novel therapies that act specifically by targeting this inflammation. To this end, recent studies that aimed to investigate the potential benefit of mepolizumab in patients with eosinophilic asthma identified patients by using blood eosinophil cutoffs as low as 0.15 × 109 cells/L.7Pavord I.D. Korn S. Howarth P. Bleecker E.R. Buhl R. Keene O.N. et al.Mepolizumab for severe eosinophilic asthma (DREAM): a multicentre, double-blind, placebo-controlled trial.Lancet. 2012; 380: 651-659Abstract Full Text Full Text PDF PubMed Scopus (1658) Google Scholar, 8Ortega H.G. Liu M.C. Pavord I.D. Brusselle G.G. FitzGerald J.M. Chetta A. et al.Mepolizumab treatment in patients with severe eosinophilic asthma.N Engl J Med. 2014; 371: 1198-1207Crossref PubMed Scopus (1574) Google Scholar As demonstrated both here and in the article from the Severe Asthma Research Program group,4Reddel H.K. Taylor D.R. Bateman E.D. Boulet L.P. Boushey H.A. Busse W.W. et al.An official American Thoracic Society/European Respiratory Society statement: asthma control and exacerbations: standardizing endpoints for clinical asthma trials and clinical practice.Am J Respir Crit Care Med. 2009; 180: 59-99Crossref PubMed Scopus (1487) Google Scholar this cutoff would include many patients who in fact do not have contemporaneous sputum eosinophilia (more than half in this study) and therefore potentially would not be in the group of patients who could benefit from treatment. The consequence of this in a trial of a biologic agent is that the magnitude of benefit could be underestimated or even missed, as happened previously when mepolizumab was trialed in unselected patients.9Flood-Page P. Swenson C. Faiferman I. Matthews J. Williams M. Brannick L. et al.A study to evaluate safety and efficacy of mepolizumab in patients with moderate persistent asthma.Am J Respir Crit Care Med. 2007; 176: 1062-1071Crossref PubMed Scopus (633) Google Scholar The other potential consequence of using blood eosinophil counts as a surrogate for sputum is the very high risk of false-negative results; from our data, approximately 1 in 5 patients without blood eosinophilia had airway eosinophilia and hence could be missed by using this selection method. Choosing a simple and universally available biomarker (blood eosinophil counts) is attractive, but in this case does not lead to a personalized medicine approach. In conclusion, we have shown that using a cutoff of 0.45 × 109 cells/L for blood eosinophilia can usefully predict airway eosinophilia in patients with severe asthma receiving high levels of treatment. However, we agree with Hastie et al3Hastie A.T. Moore W.C. Li H. Rector B.M. Ortega V.E. Pascual R.M. et al.Biomarker surrogates do not accurately predict sputum eosinophil and neutrophil percentages in asthmatic subjects.J Allergy Clin Immunol. 2013; 132: 72-80Abstract Full Text Full Text PDF PubMed Scopus (201) Google Scholar that lower cutoffs are not useful. We would propose that study design for novel antieosinophil therapies could use this higher cutoff to investigate the effect of these therapies but that patients should also be included with airway but not blood eosinophilia, as previously described.8Ortega H.G. Liu M.C. Pavord I.D. Brusselle G.G. FitzGerald J.M. Chetta A. et al.Mepolizumab treatment in patients with severe eosinophilic asthma.N Engl J Med. 2014; 371: 1198-1207Crossref PubMed Scopus (1574) Google Scholar In the meantime, more work needs to be done to develop useful minimally invasive and accurate surrogate markers for airway inflammation, including composite biomarkers comprising, for example, exhaled nitric oxide and serum periostin levels together with blood eosinophil counts. Table E1Coordinates of the receiver operating characteristics curve in Fig E1 showing sensitivity and 1 − specificity for blood eosinophil counts to predict sputum eosinophils of at least 2% (eg, blood eosinophil counts of at least 0.46 × 109 cells/L has a sensitivity of 49.3% and a specificity of 97.0% for sputum eosinophilia)Blood eosinophil cutoff (× 109 cells/L)Sensitivity1 − Specificity0.001.0001.0000.010.9700.9330.020.9550.9110.030.9550.8810.040.9400.7930.050.9400.7560.060.9250.6960.070.9100.6670.080.8660.6440.090.8360.6220.100.8360.6000.110.8210.5410.120.8060.5110.130.7910.5040.140.7760.4890.150.7760.4670.160.7760.4300.170.7760.3630.180.7610.3410.190.7460.2960.200.7460.2670.210.7160.2670.220.7160.2440.230.6720.2300.240.6720.2150.250.6420.2000.260.6270.1780.270.6120.1780.290.5970.1630.310.5970.1560.330.5820.1330.340.5820.1110.350.5820.0890.360.5670.0810.380.5520.0740.390.5520.0590.400.5220.0590.420.5220.0520.430.5070.0440.440.4930.0440.460.4930.0300.470.4780.0300.490.4630.0300.510.4480.0300.520.4330.0300.530.4330.0220.540.4030.0220.560.3730.0220.590.3580.0150.620.3430.0150.640.3130.0150.660.2990.0070.670.2840.0070.700.2540.0070.740.2390.0070.780.2240.0070.810.2090.0070.860.2090.0000.890.1940.0000.920.1640.0000.960.1490.0000.990.1340.0001.010.1190.0001.040.1040.0001.080.0900.0001.150.0600.0001.210.0450.0001.310.0300.0001.800.0150.0003.220.0000.000 Open table in a new tab Blood eosinophil counts predict treatment response in patients with severe eosinophilic asthmaJournal of Allergy and Clinical ImmunologyVol. 136Issue 3PreviewIn a letter to the editor published in the March 2015 issue of the Journal, Fowler et al1 found that a blood eosinophil cutoff of 150 cells/μL had a positive predictive value of less than 50%. They conclude that using a blood eosinophil cutoff of 150 cells/μL would include many patients who do not have contemporaneous sputum eosinophilia and translate this into a potential lack of response or underestimation of response to new biologic medications, such as mepolizumab. For the mepolizumab program, the goal of using blood eosinophils as a biomarker is not to accurately predict the sputum eosinophil percentage but rather to serve as a marker of treatment response. Full-Text PDF