Blood eosinophil count as predictor of asthma exacerbation. A meta-analysis

1 Title: Blood Eosinophil Count as Predictor of Asthma Exacerbation. A Meta-analysis Short title: Blood Eosinophil Count and Asthma Exacerbation Narmeen Mallah, MSc1,2; Santiago Rodriguez-Segade, MD3; Francisco-Javier Gonzalez- Barcala, MD3,4,5,6; Bahi Takkouche, PhD1,2,6 1Department of Preventive Medicine, University of Santiago de Compostela, Santiago de Compostela, Spain. 2Centro de Investigación Biomédica en Red de Epidemiología y Salud Pública (CIBER-ESP), Carlos III Health Research Institute, Madrid, Spain. 3Department of Respiratory Medicine, University Hospital of Santiago de Compostela (CHUS), Santiago de Compostela, Spain. 4Department of Medicine, University of Santiago de Compostela, Santiago de Compostela, Spain. 5Spanish Biomedical Research Networking Centre (CIBER-ES), Carlos III Health Research Institute, Madrid, Spain. 6Health Research Institute of Santiago de Compostela (IDIS), Santiago de Compostela, Spain. Corresponding Author: Dr. Francisco-Javier Gonzalez-Barcala, Department of Respiratory Medicine, University Hospital of Santiago de Compostela (CHUS), Santiago de Compostela, 15706, Spain (francisco.javier[email protected]). Authors′ Contributions: NM conducted the literature review, extracted, analyzed, and interpreted the data, and designed and wrote the manuscript. SR-S participated in the studies 2 localization. NM and BT extracted the data. FJG-B conceived the research idea and resolved disagreements about articles՚ selection. BT designed the study and supervised the data analysis and interpretation. All authors reviewed and revised the manuscript and approved it for publication. Acknowledgments: We are grateful to the following authors who answered our queries about their studies: David B. Price (Centre for Academic Primary Care, The Institute of Applied Health Sciences, University of Aberdeen, Aberdeen, UK; Research in Real-Life, Cambridge, UK), Loren C. Denlinger (University of Wisconsin, Madison, Wisconsin) and Kyoung-Hee Sohn (Institute of Allergy and Clinical Immunology, Seoul National University Medical Research Center, Seoul National University College of Medicine, Seoul, Korea). Conflict of interest: none declared Funding sources: no funding was required for this study. Word Count (Abstract): 247 Word Count (Text): 3047 3 ABSTRACT Background: Evidence about the association of high blood eosinophil count with asthma exacerbation is inconsistent and unclear. The objective of this meta-analysis is to determine whether elevated blood eosinophils count predicts asthma exacerbation. Methods: We searched MEDLINE, EMBASE, and additional databases, without any language restriction. We also checked the reference lists of the included studies and of relevant systematic reviews. The main outcome was the occurrence of asthma exacerbation. We calculated global pooled Odds Ratios (ORs) and their 95% Confidence Intervals (CIs) and performed predefined subgroup analyses. We appraised the quality of the studies using Newcastle-Ottawa Scale, examined the heterogeneity between studies, assessed publication bias and carried out sensitivity analyses. Results: Among 1567 retrieved publications, 23 observational studies comprising 155,772 participants met the inclusion criteria. High blood eosinophil count was associated with higher odds of asthma exacerbation [OR: 1.31 (95%CI: 1.16, 1.49)], specifically with asthma-related outpatient visits [OR: 1.46 (95%CI: 1.25, 1.70)] and emergency department visits [OR: 1.63 (95%CI: 1.29, 2.07)]. A significant association was observed starting from an eosinophils’ cut-off value of 200 cells/μl. The association was observed for cohort studies [OR: 1.30 (95%CI: 1.13, 1.49)], North-American studies [OR: 1.43 (95%CI: 1.31, 1.57)], Asian populations [OR: 1.67 (95%CI: 1.34, 2.08)] children [OR: 1.38 (95%CI: 1.22, 1.56)], and studies that adjusted for inhaled corticosteroids therapy [OR: 1.42 (95%CI: 1.28, 1.56)]. Conclusions: Blood eosinophil counts ≥ 200 cells/µl are associated with asthma exacerbation. Blood eosinophil count is a modifiable factor that could be addressed in asthma management strategies. Keywords: asthma exacerbation, blood eosinophils, meta-analysis 4 Abbreviations: CCI: Charlson Comorbidity Index; CI: Confidence Interval; CRP: C-Reactive Protein; FEV1: Forced Expiratory Volume in 1 second; FeNO: Fractional Exhaled Nitric Oxide; ICS: Inhaled Corticosteroids; NOS: Newcastle-Ottawa Scale; OR: Odds Ratio; PEF: Peak Expiratory Flow; RR: Relative Risk 5 INTRODUCTION Asthma is a prominent chronic disease that affects 235 million individuals worldwide and represents the most common chronic disease among children.1,2 When exacerbated, it requires urgent medical assistance to prevent serious outcomes including death.3 Asthma exacerbation occurs in higher frequency and with more severity when asthma is not controlled; nevertheless, it may also affect patients on asthma treatment.4 The global mortality rate from asthma was 0.19 per 100,000 individuals in 2012, with stalled improvements in lowering mortality over the past decade.5 Indeed, in 2016, asthma was classified as the 23rd leading cause of premature mortality.6 The Global Asthma Report 2018 estimates that 1000 people die every day from asthma worldwide.6 Asthma prevalence is still growing in many countries, imposing major consequences on public health, society and economy.6 These observations emphasize the need to better understand predisposing factors to predict the occurrence and severity of asthma exacerbation. Increased clinical interest in the use of eosinophil count as a predictor of asthma worsening was observed in the recent literature. However, to reach appropriate clinical decisions, a careful interpretation of the available evidence is required. Findings concerning blood eosinophils and asthma exacerbation diverged widely across studies, as the directions and the magnitudes of measures of effect were inconsistent. While some studies did not find a correlation between increased blood eosinophil count and asthma severity,7,8 others reported a borderline effect,9-11 or a positive association which magnitude varied greatly between studies, ranging between 30% and 150% increase in the risk of asthma worsening.12-19 On the other side, some authors reported opposite findings which associated elevated levels of blood eosinophils with a lower risk of asthma exacerbation.20,21 In light of these contradictory observations, and taking into account that blood eosinophil count may be lowered through adequate therapy, we considered that there was a 6 need for a meta-analysis that assesses the association between blood eosinophil count and asthma exacerbation, and explores the threshold of blood eosinophils which potentially induces asthma worsening. METHODS Data sources and searches We performed a systematic review and meta-analysis of blood eosinophil counts as predictors of the risk of asthma exacerbation. Asthma exacerbation is also sometimes referred as asthma attack.22,23 We defined asthma exacerbation (or asthma attack) as a deterioration in asthma requiring any of the following: treatment with an oral corticosteroid, visit to an emergency department, outpatient visit, or hospitalization. We included studies that ascertained asthma exacerbations using self-assessment tools. However we later stratified the analysis based on the mean used for outcome ascertainment. We searched the following databases since their inception until May 2020: MEDLINE, EMBASE, Conference Proceedings Citation Index- Science, the five regional bibliographic databases of the World Health Organization; as well as the Open Access Thesis and Dissertations. We did not apply any restriction on language or publication date. In MEDLINE we applied the search syntax: ("Asthma"[Mesh] AND "Eosinophils"[Mesh] AND (exacerbation OR worsening)), which we then adapted to other databases. We completed our search by using the terms: asthma, blood eosinophil, exacerbation, worsening, attack, as free text words, and by checking manually the reference lists of the eligible studies as well as that of other relevant systematic reviews. 7 Study selection Two researchers (NM and SR-S) independently reviewed the titles and abstracts of the retrieved studies and subsequently selected the potentially eligible ones for a full-text revision. Disagreements were resolved by referring to a third researcher (FJG-B). We included observational studies that reported Odds Ratio (OR) or Relative Risk (RR) of the association between high blood eosinophil count and asthma exacerbation. We excluded studies in which the comparison group was limited to healthy subjects. In the case of duplicated reports, we considered the most complete and updated version. The study protocol is registered in the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42020157743). Data extraction and quality assessment Two epidemiologists (NM and BT) extracted data on: 1) study source: first author’s name and publication year; 2) study characteristics: country, study design, blood eosinophil cut-off limit, sample size, participants’ inclusion and exclusion criteria, effect measure and 95% confidence interval (CI), and adjustment variables; 3) patients’ characteristics: age and sex; and 4) asthma exacerbation definitions: hospital admission, emergency department visit, outpatient visit and/or increased corticosteroids treatment. When relevant data were incomplete or not available, we contacted the corresponding author to request the missing information.8,9,14,15,24-28 To assess the methodological quality of the studies, we used the Newcastle-Ottawa Scale (NOS) for cohort and case-control studies,29 and an adapted NOS version for the evaluation of crosssectional studies.30 NOS is a widely used tool to evaluate the methodological aspects of observational studies such as samples՚ representativeness, participation rate, control for confounding factors, and ascertainment of exposure and outcome. Quality scoring was performed 8 independently by two reviewers and the average score between reviewers was assigned to the studies. A maximum of 8 points was attributed to each study.(e-Appendix 1) Disagreement between raters was measured by the Bland-Altman limits of agreement method.31 Data synthesis and statistical analysis We weighted the study-specific adjusted measures of effect by the inverse of their variance to obtain a pooled OR. When various thresholds of blood eosinophil count were reported, we used 400 cells/μl as a cut-off limit.32 When effect measures were reported for different subgroups within the same study, we pooled the different estimates in order to obtain one global OR per study. We computed fixed and random effect models but considered only the latter when heterogeneity was detected. We checked for heterogeneity using DerSimonian and Laird’s Q test and quantified it using Ri, the proportion of total variance due to between-study variance.33 We interpreted Ri values as follows: low heterogeneity (Ri<0.4), moderate heterogeneity (0.4≤ Ri≤0.75), and high heterogeneity (Ri>0.75). Finally, we stratified the analysis according to study design, type of cohort study, quality score, age category, geographical location, blood eosinophils’ cut-off limit, asthma exacerbation definitions (indicators), methods of asthma exacerbation ascertainment, and frequency of asthma attacks. We planned a priori all these subgroup analyses. We explored the presence of publication bias, first visually using a funnel plot and then, more formally, using Egger’s regression test. We also carried out a sensitivity analysis assuming that cross-sectional studies are the least likely to be published when they show a null association. We recalculated the pooled OR assuming that: 1) the cross-sectional studies included in our meta- 9 analysis represent only half of the studies that have ever been conducted on this topic, 2) the unpublished cross-sectional studies found a null association of OR = 1 and 3) the prevalence of asthma exacerbation in the unpublished studies was equal to the average prevalence of the published ones. In addition, to further assess publication bias, we performed a “trim and fill” analysis.34 We carried out the analyses using the software HEpiMA version 2.1.3,35 and STATA version 12 (Stata Corp, College Station, TX, USA). RESULTS Literature search and study characteristics We identified 1567 studies in our search (Figure 1). Twenty-three studies with a total population of 155772 patients met our inclusion criteria (Table 1 and Figure 2). Four studies were not considered eligible for this meta-analysis because they did not use any specific cut-off value for blood eosinophil count.9,24,26,27 One of the four studies measured the association using percentage of blood eosinophils instead of counts,9 and 3 other studies provided continuous effect measures.24,26,27 We merged data from different studies carried out in the same population as explained in eAppendix 2. We classified one cohort study as cross-sectional given that the blood eosinophil test was performed at admission to the hospital, and hence no follow-up stage was carried out.20 16 Additional analyses of our data have shown that misclassification of the outcome due to selfascertainment of asthma exacerbation and misclassification of blood eosinophil count are unlikely to explain our results (data not shown). We also found a large amount of heterogeneity between studies that remained present in many subgroup analyses. This heterogeneity was to a certain point expected, due to the wide range of cut-off limits and the differences in the definitions of asthma worsening used in the studies. We, therefore, based our interpretation on random effects estimates as recommended.58,59 Metaanalysis experts emphasize that no degree of heterogeneity can be deemed unacceptable, provided the eligibility criteria are clear and the data are correct,58 and that heterogeneity in metaanalysis, due to the differences in methods and populations, should be viewed as the “expectation, rather than the exception”.60 Finally, due to the lack of data, it was not possible to assess whether the increase in the odds of asthma exacerbation was associated in a dose-response fashion to the increase in the count of blood eosinophils. Asthma is a frequent life-long inflammatory disorder that causes a global substantial burden of disease. Eosinophilic inflammation in asthma patients is a well-characterized feature of the disease. We showed that the occurrence of asthma exacerbation is related to increased levels of blood eosinophils, a modifiable exposure factor. Future research is needed to investigate this association in understudied vulnerable populations including elderly patients. The findings of this meta-analysis could prove useful in the management of asthma. 17 DATA AVAILABILITY The data that support the findings of this study are openly available in “FigShare” at: https://figshare.com/s/e7c1358ae298d033fa91 REFERENCES 1. GBD 2015 Mortality and Causes of Death Collaborators. Global, regional, and national life expectancy, all-cause mortality, and cause-specific mortality for 249 causes of death, 1980-2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet. 2016;388(10053):1459-1544. doi: 10.1016/S0140-6736(16)31012-1. 2. WHO. Asthma. https://www.who.int/news-room/q-a-detail/asthma. Accessed October 4, 2020. 3. Reddel HK, Taylor DR, Bateman ED, Boulet L-P, Boushey HA, Busse WW, 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(1):59-99. doi: 10.1164/rccm.200801- 060ST 4. Global Initiative for Asthma (GINA). Asthma management and prevention (for adults and children older than five years old). A pocket guide for health professionals updated 2019. https://ginasthma.org/wp-content/uploads/2019/04/GINA-2019-main-Pocket-Guide- wms.pdf. Accessed October 4, 2020. 5. Ebmeier S, Thayabaran D, Braithwaite I, Bénamara C, Weatherall M, Beasley R. Trends in international asthma mortality: analysis of data from the WHO Mortality Database from 46 countries (1993-2012). Lancet. 2017(10098);390:935-945. doi: 10.1016/S0140- 6736(17)31448-4 18 6. Global Asthma Network. The global asthma report 2018. Auckland, New Zealand: Global Asthma Network, 2018. http://www.globalasthmareport.org/Global%20Asthma%20Report%202018.pdf. Accessed October 4, 2020. 7. Malinovschi A, Janson C, Borres M, Alving K. Simultaneously increased fraction of exhaled nitric oxide levels and blood eosinophil counts relate to increased asthma morbidity. J Allergy Clin Immunol. 2016;138(5):1301-1308:e2. doi: 10.1016/j.jaci.2016.01.044 8. Tran TN, Khatry DB, Ke X, Ward CK, Gossage D. High blood eosinophil count is associated with more frequent asthma attacks in asthma patients. Ann Allergy Asthma Immunol. 2014;113(1):19-24. doi: 10.1016/j.anai.2014.04.011 9. Ban GY, Ye YM, Kim SH, Hur GY, Kim JH, Shim JJ, et al. Plasma LTE4/PGF2alpha ratio and blood eosinophil count are increased in elderly asthmatics with previous asthma exacerbation. Allergy Asthma Immunol Res. 2017;9(4):378-382. doi: 10.4168/aair.2017.9.4.378 10. Kerkhof M, Tran TN, van den Berge M, Brusselle GG, Gopalan G, Jones RCM, et al. Association between blood eosinophil count and risk of readmission for patients with asthma: historical cohort study. PLoS One. 2018;13(7):e0201143. doi: 10.1371/journal.pone.0201143 11. Zeiger RS, Schatz M, Dalal AA, Suruki RY, Kawathar AA, Qian L. Blood eosinophil count and outcomes in severe uncontrolled asthma: a prospective study. J Allergy Clin Immunol Pract. 2017;5(1):144-153:e8. doi: 10.1016/j.jaip.2016.07.015 19 12. Belda J, Giner J, Casan P, Sanchis J. Mild exacerbations and eosinophilic inflammation in patients with stable, well-controlled asthma after 1 year of follow-up. Chest. 2001;119(4): 1011-1017. doi: 10.1378/chest.119.4.1011 13. Casciano J, Krishnan JA, Small MB, Buck PO, Gopalan G, Li C, et al. Burden of asthma with elevated blood eosinophil levels. BMC Pulm Med. 2016;16:100. doi: 10.1186/s12890-016-0263-8 14. Price D, Wilson AM, Chisholm A, Rigazio A, Burden A, Thomas M, et al. Predicting frequent asthma exacerbations using blood eosinophil count and other patient data routinely available in clinical practice. J Asthma Allergy. 2016;9:1-12. doi: 10.2147/JAA.S97973 15. Price DB, Rigazio A, Campbell JD, Bleecker ER, Corrigan CJ, Thomas M, et al. Blood eosinophil count and prospective annual asthma disease burden: a UK cohort study. Lancet Respir Med 2015;3(11):849-858. doi: 10.1016/S2213-2600(15)00367-7 16. Vedel-Krogh S, Fallgaard Nielsen S, Lange P, Vestbo J, Nordestgaard BG. Association of blood eosinophil and blood neutrophil counts with asthma exacerbations in the copenhagen general population study. Clin Chem. 2017;63(4):823-832. doi: 10.1373/clinchem.2016.267450 17. Velthove KJ, Bracke M, Souverein PC, Schweizer RC, Berg MJT, Leufkens HGM, et al. Identification of exacerbations in obstructive lung disease through biomarkers. Biomarkers. 2009;14(7):523-528. doi: 10.3109/13547500903150763 18. Yii ACA, Tay TR, Puah SH, Lim H-F, Li A, Lau P, et al. Blood eosinophil count correlates with severity of respiratory failure in life-threatening asthma and predicts risk of subsequent exacerbations. Clin Exp Allergy. 2019;49(12):1578-1586. doi: 10.1111/cea.13465 20 19. Zeiger RS, Schatz M, Li Q, Chen W, Khatry DP, Gossage D, et al. High blood eosinophil count is a risk factor for future asthma exacerbations in adult persistent asthma. J Allergy Clin Immunol Pract. 2014;2(6):741-750. doi: 10.1016/j.jaip.2014.06.005. 20. Gonzalez-Barcala FJ, San-Jose ME, Nieto-Fontarigo JJ, Carreira J-M, Calvo-Alvarez U, Cruz M-J, et al. Association between blood eosinophil count with asthma hospital readmissions. Eur J Intern Med. 2018;53:34-39. doi: 10.1016/j.ejim.2018.02.034 21. Pola-Bibian B, Dominguez-Ortega J, Vila-Nadal G, Entrala A, González-Cavero L, Barranco P, et al. Asthma exacerbations in a tertiary hospital: clinical features, triggers, and risk factors for hospitalization. J Investig Allergol Clin Immunol. 2017;27(4):238- 245. doi: 10.18176/jiaci.0128 22. Fuhlbrigge A, Peden D, Apter AJ, Boushey HA, Camargo Jr CA, Gern J, et al. Asthma outcomes: exacerbations. J Allergy Clin Immunol. 2012;129(3 Suppl):S34-48. doi: 10.1016/j.jaci.2011.12.983 23. Global Initiative for Asthma (GINA). Global strategy for asthma management and prevention updated 2019. https://ginasthma.org/wp-content/uploads/2019/06/GINA-2019- main-report-June-2019-wms.pdf. Accessed October 4, 2020. 24. Denlinger LC, Phillips BR, Ramratnam S, Ross K, Bhakta NR, Cardet JC, et al. Inflammatory and comorbid features of patients with severe asthma and frequent exacerbations. Am J Respir Crit Care Med. 2017;195(3):302-313. doi: 10.1164/rccm.201602-0419OC 25. Price DB, Bosnic-Anticevich S, Pavord ID, Roche N, Halpin DMG, Bjermer L, et al. Association of elevated fractional exhaled nitric oxide concentration and blood eosinophil count with severe asthma exacerbations. Clin Transl Allergy. 2019;9:41. doi: 10.1186/s13601-019-0282-7 21 26. Semprini R, Williams M, Semprini A, McDaouall A, Fingleton J, Holweh C, et al. Type 2 biomarkers and prediction of future exacerbations and lung function decline in adult asthma. J Allergy Clin Immunol Pract. 2018;6(6):1982-1988. doi: 10.1016/j.jaip.2018.03.004 27. Wu AC, Tantisira K, Li L, Schuemann B, Weiss ST, Fuhlbrigge AL. Predictors of symptoms are different from predictors of severe exacerbations from asthma in children. Chest. 2011;140(1):100-107. doi: 10.1378/chest.10-2794 28. Sohn KH, Song WJ, Park JS, Park HW, Kim TB, Park CS, et al. Risk factors for acute exacerbations in elderly asthma: what makes asthma in older adults distinctive? Allergy Asthma Immunol Res. 2020;12(3):443-453. doi: 10.4168/aair.2020.12.3.443 29. Wells G, Shea B, O'Connell D, Peterson J, Welch V, Losos M, et al. The Newcastle- Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp. Accessed October 4, 2020. 30. Modesti PA, Reboldi G, Cappuccio FP, Agyemang C, Remuzzi G, Rapi S, et al. Panethnic differences in blood pressure in Europe: a systematic review and meta-analysis. PLoS One. 2016;11(1):e0147601. doi: 10.1371/journal.pone.0147601 31. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1(8476):307-310. 32. Aleman F, Lim HF, Nair P. Eosinophilic endotype of asthma. Immunol Allergy Clin N Am. 2016; 36(3):559-568. doi: 10.1016/j.iac.2016.03.006 33. Takkouche B, Cadarso-Suarez C, Spiegelman D. Evaluation of old and new tests of heterogeneity in epidemiologic meta-analysis. Am J Epidemiol. 1999;150(2):206-215. doi: 10.1093/oxfordjournals.aje.a009981 22 34. Duval S, Tweedie R. Trim and fill: A simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics. 2000;56(2):455-463. doi: 10.1111/j.0006-341x.2000.00455.x 35. Costa-Bouzas J, Takkouche B, Cadarso-Suarez C, Spigelman. HEpiMA: software for the identification of heterogeneity in meta-analysis. Comput Methods Programs Biomed. 2001;64(2):101-107. doi: 10.1016/s0169-2607(00)00087-0 36. Shah SP, Grunwell J, Shih J, Stephenson S, Fitzpatrick AM. Exploring the utility of noninvasive type 2 inflammatory markers for prediction of severe asthma exacerbations in children and adolescents. J Allergy Clin Immunol Pract. 2019;7(8):2624-2633. doi: 10.1016/j.jaip.2019.04.043 37. Zeiger RS, Schatz M, Li Q, et al. The association of blood eosinophil counts to future asthma exacerbations in children with persistent asthma. J Allergy Clin Immunol Pract. 2015;3(2):283-287. doi: 10.1016/j.jaip.2014.10.009 38. Turner SW, Murray C, Thomas M, Burden A, Price DB. Applying UK real-world primary care data to predict asthma attacks in 3776 well-characterised children: a retrospective cohort study. NPJ Prim Care Respir Med. 2018;28:28. doi: 10.1038/s41533-018-0095-5 39. Nagasaki T, Sato K, Kume N, Oguma T, Sunadome H, Ito I, et al. The prevalence and disease burden of severe eosinophilic asthma in Japan. J Asthma 2019;56(11):1147-1158. doi: 10.1080/02770903.2018.1534967 40. Hakansson KEJ, Rasmussen LJH, Godtfredsen NS, Tupper OD, Eugen-Oslen J, Kallemose T, et al. The biomarkers suPAR and blood eosinophils are associated with hospital readmissions and mortality in asthma - a retrospective cohort study. Respir Res. 2019;20(1):258. doi: 10.1186/s12931-019-1234-4 23 41. Westerhof GA, de Groot JC, Amelink M, de Nijs SB, Brinke AT, Weersink EJ, et al. Predictors of frequent exacerbations in (ex)smoking and never smoking adults with severe asthma. Respir Med. 2016;118:122-127. doi: 10.1016/j.rmed.2016.08.006 42. Mogensen I, Alving K, Jacinto T, Fonseca J, Janson C, Malinovschi A. Simultaneously elevated FeNO and blood eosinophils relate to asthma morbidity in asthmatics from NHANES 2007-12. Clin Exp Allergy. 2018;48(8):935-943. doi: 10.1111/cea.13137 43. Zhang XY, Simpson JL, Powell H, Yang IA, Upham JW, Reynolds PN, et al. Full blood count parameters for the detection of asthma inflammatory phenotypes. Clin Exp Allergy. 2014;44(9):1137-1145. doi: 10.1111/cea.12345 44. Johansson MW. Activation states of blood eosinophils in asthma. Clin Exp Allergy. 2014;44(4):482-498. doi: 10.1111/cea.12292 45. Peters MC, Wenzel SE. Intersection of biology and therapeutics: type 2 targeted therapeutics for adult asthma. Lancet 2020; 395: 371–83. doi: 10.1016/S0140- 6736(19)33005-3 46. Yancey SW, Keene ON, Albers FC, Ortega H, Bates S, Bleecker ER, et al. Biomarkers for severe eosinophilic asthma. J Allergy Clin Immunol. 2017;140(6):1509-1518. doi: 10.1016/j.jaci.2017.10.005 47. Aleman F, Lim HF, Nair P. Eosinophilic endotype of asthma. Immunol Allergy Clin North Am. 2016;36(3):559-568. doi: 10.1016/j.iac.2016.03.006 48. Roche N, Chapman KR, Vogelmeier CF, Herth FJ, Thach C, Fogel R, et al. Blood eosinophils and response to maintenance chronic obstructive pulmonary disease treatment: data from the FLAME trial. Am J Respir Crit Care Med 2017;195:1189–1197. doi: 10.1164/rccm.201701-0193OC 24 49. Rydman RJ, Isola ML, Roberts RR, Zalenski RJ, McDermott MF, Murphy DG, et al. Emergency department observation unit versus hospital inpatient care for a chronic asthmatic population: a randomized trial of health status outcome and cost. Med Care. 1998;36(4):599-609. doi: 10.1097/00005650-199804000-00015 50. Zhu J, Message SD, Qiu Y, Mallia P, Kebadze T, Contoli M, et al. Airway inflammation and illness severity in response to experimental rhinovirus infection in asthma. Chest. 2014;145(6):1219-1229. doi: 10.1378/chest.13-1567 51. Jackson DJ, Sykes A, Mallia P, Johnston SL. Asthma exacerbations: origin, effect, and prevention. J Allergy Clin Immunol. 2011;128(6):1165-1174. doi: 10.1016/j.jaci.2011.10.024 52. Mehta NK, Lee H, Ylitalo KR. Child health in the United States: recent trends in racial/ethnic disparities. Soc Sci Med. 2013;95:6-15. doi: 10.1016/j.socscimed.2012.09.011 53. Koo S, Gupta A, Fainardi V, Bossley C, Bush A, Saglani S, et al. Ethnic variation in response to IM Triamcinolone in children with severe therapy-resistant asthma. Chest. 2016;149(1):98-105. doi: 10.1378/chest.14-3241 54. Park HW, Tantisira KG. Genetic signatures of asthma exacerbation. Allergy Asthma Immunol Res. 2017;9(3):191-199. doi: 10.4168/aair.2017.9.3.191 55. Hernandez-Pacheco N, Pino-Yanes M, Flores C. Genomic predictors of asthma phenotypes and treatment response. Front Pediatr. 2019;7:6. doi: 10.3389/fped.2019.00006 56. Sinharoy A, Mitra S, Mondal P. Socioeconomic and environmental predictors of asthmarelated mortality. J Environ Public Health. 2018;2018:9389570. doi: 10.1155/2018/9389570 25 57. Kaplan A, Hardjojo A, Yu S, Price D. Asthma across age: insights from primary care. Front Pediatr. 2019;7:162. doi: 10.3389/fped.2019.00162 58. Higgins JP. Commentary: Heterogeneity in meta-analysis should be expected and appropriately quantified. Int J Epidemiol. 2008;37(5):1158-1160. doi: 10.1093/ije/dyn204 59. Council N. Combining information: statistical issues and opportunities for research. Washington (DC): National Academy Press 1992:52. 60. Berlin JA. Invited commentary: benefits of heterogeneity in meta-analysis of data from epidemiologic studies. Am J Epidemiol. 1995;142(4):383-387. doi: 10.1093/oxfordjournals.aje.a117645 32 Table 2. Pooled OR and 95% CI of increased blood eosinophil count and asthma exacerbation Number of studies OR (95% CI) Fixed effects OR (95% CI) Random effects Ri‡ Q test p-value All studies 22† 1.39 (1.35, 1.44) 1.31 (1.16, 1.49) 0.90 <0.0001 Study design Cohort 15 1.40 (1.36, 1.45) 1.30 (1.13, 1.49) 0.92 <0.0001 Cross-Sectional 6 1.31 (1.13, 1.51) 1.27 (0.94, 1.72) 0.73 0.003 Cohort study type Prospective 7 1.42 (1.37, 1.46) 1.35 (1.17, 1.56) 0.92 <0.0001 Retrospective 8 1.26 (1.13, 1.41) 1.29 (0.95, 1.76) 0.86 <0.0001 Age category Only children 4 1.38 (1.22, 1.56) 1.38 (1.22, 1.56) 0.00 0.64 Exacerbation ascertainment Medical Record 18 1.40 (1.36, 1.44) 1.30 (1.13, 1.50) 0.93 <0.0001 Self-Reporting 4 1.40 (1.17, 1.67) 1.40 (1.17, 1.67) 0.00 0.42 Adjustment for ICS treatment Not adjusted for ICS or unspecified 7 1.08 (0.94, 1.24) 1.13 (0.73, 1.74) 0.90 <0.0001 Adjusted for ICS 15 1.42 (1.38, 1.46) 1.42 (1.28, 1.56) 0.79 0.001 Quality Score < 4 point 11 1.39 (1.35, 1.44) 1.21 (0.93, 1.59) 0.98 <0.0001 ≥ 4 points 11 1.44 (1.33, 1.55) 1.44 (1.33, 1.55) 0.00 0.86 Geographical location Europe 12 1.39 (1.35, 1.44) 1.19 (0.95, 1.50) 0.97 <0.0001 North America 7 1.43 (1.31, 1.57) 1.43 (1.31, 1.57) 0.00 0.54 Asia 3 1.67 (1.34, 2.08) 1.67 (1.34, 2.08) 0.00 0.70 Blood eosinophils cut-off limit 150 4 1.21 (1.08, 1.34) 1.16 (0.77, 1.74) 0.93 <0.0001 200 5 1.16 (1.11, 1.22) 1.22 (1.02, 1.46) 0.92 <0.0001 250 3 1.24 (1.18, 1.29) 1.42 (1.09, 1.83) 0.95 0.005 300 11 1.29 (1.24, 1.35) 1.23 (1.08, 1.41) 0.83 <0.0001 350 2 1.34 (1.28, 1.41) 1.34 (1.28, 1.41) 0.00 0.56 400 12 1.41 (1.34, 1.48) 1.41 (1.23, 1.62) 0.78 <0.0001 >400 3 1.54 (1.46, 1.61) 1.54 (1.46, 1.61) 0.00 0.70 Definition of asthma exacerbation Asthma related hospitalization 8 0.99 (0.86, 1.15) 1.40 (0.82, 2.40) 0.92 <0.0001 Asthma related emergency room visit 3 1.63 (1.29, 2.07) 1.63 (1.29, 2.07) 0.00 0.58 Asthma related outpatient acute visits 2 1.46 (1.25, 1.70) 1.25 (0.71, 2.20) 0.90 0.14 Frequency of asthma attack Frequent (≥2) 5 1.47 (1.38, 1.56) 1.28 (0.94, 1.74) 0.92 0.05 Less frequent 4 1.04 (0.90, 1.20) 0.95 (0.70, 1.29) 0.71 0.04 †: The total number of studies is 22, as 2 cohort studies were analyzed together (e-Appendix 2). ‡: Proportion of total variance due to between-study variance 33 Figure Legends Figure 1. Flow diagram of studies՚ localization and selection Figure 2. Study-specific and pooled Odds Ratios of blood eosinophil count and asthma exacerbation Figure 3. Funnel plot of Odds Ratios of blood eosinophil count and asthma exacerbation 34 Appendices e-Appendix 1: Quality Assessment of included studies We assessed the following aspects that were common between the three study designs: 1) if the method used to count blood eosinophils was described (1 point), else (0 point); 2) if asthma exacerbation was determined based on medical records (1 point), else (0 point). For cohort studies we also determined: 1) if the exposed cohort was representative of the population (1 point), else (0 point); 2) if non-exposed patients were recruited from the same population as the exposed patients (1 point), else (0 point); 3) if included patients were free from asthma exacerbation events when recruited (1 point), else (0 point); 4) if the measure of effect was adjusted for sex, age, weight, asthma treatment and smoking habits (1 point), else (0 point); 5) if patients were followed up for a minimum of 1 year (1point), else (0 point); 6) if the proportion of patients who left the study or were lost to follow up was ≤ 20% (1 point), else (0 point). For case-control studies, we determined: 1) if cases were representative of patients with asthma exacerbation in the population (1 point), else (0 point); 2) if controls were selected from the same population as the cases (1 point), else (0 point); 3) if blood eosinophil count in controls was ascertained using the same method as that applied for cases (1 point), else (0 point); 4) if participation rate was ≥ 80% (1 point), else (0 point). For cross-sectional studies we checked: 1) if the sample was representative of the whole population (1 point), else (0 point); 2) if the response rate was reported (1 point), else (0 point); 3) if the statistical analysis was well described and appropriate (1 point), else (0 point); 4) if the sample size was justified (1 point), else (0 point). The comparability in case-control and cross-sectional studies was rated as follows: 1) if the measurement of effect was adjusted for asthma treatment (1 point), else (0 point); 2) if the measurement of effect was adjusted for sex, age, weight, and smoking habits (1 point), else (0 point). When details on a specific item were not provided, we graded this item with 0 point. Finally, we summed the points across the items in order to obtain a global quality score of a maximum 8 points. e-Appendix 2: Further details on different publications using the same population Two different publications using the same population were considered for our meta-analysis as they measured a different outcome. The first publication measured the risk of “any asthma exacerbation” while the second assessed the risk of “frequent asthma exacerbation”.1,2 The results of these studies were never pooled in the same group. We used the first one for the overall pooled estimation and all the subgroup analyses,1 except for the subgroup “Frequency of asthma attacks” in which we used the second publication in the category “frequent” (≥ 2 attacks).2 A third publication by the same group was included in this meta-analysis after confirming with the author that there was no overlap between the population of this study,3 and that of the previous two studies.1,2 Other 3 studies, published by the same first author (Zeiger RS), were unrelated and encompassed different populations.4-6 One of these studies involved adult patients recruited between 2009 and 2010,4 whereas the second publication included patients ≥ 12 years of age after 2012.5 The third study included a population of children aged 5 to 11 years old.6 Online-Only References 1. Price DB, Rigazio A, Campbell JD, Bleecker ER, Corrigan CJ, Thomas M, et al. Blood eosinophil count and prospective annual asthma disease burden: a UK cohort study. Lancet Respir Med. 2015;3(11):849-858. doi: 10.1016/S2213-2600(15)00367-7 2. Price D, Wilson AM, Chisholm A, Rigazio A, Burden A, Thomas A, et al. Predicting frequent asthma exacerbations using blood eosinophil count and other patient data routinely available in clinical practice. J Asthma Allergy. 2016;9:1-12. doi: 10.2147/JAA.S97973 3. Price DB, Bosnic-Anticevich S, Pavord ID, Roche N, Halpin DMG, Bjermer L, et al. Association of elevated fractional exhaled nitric oxide concentration and blood eosinophil count with severe asthma exacerbations. Clin Transl Allergy. 2019;9:41. doi: 10.1186/s13601-019-0282-7 4. Zeiger RS, Schatz M, Li Q, Chen W, Khatry DB, Gossage D, et al. High blood eosinophil count is a risk factor for future asthma exacerbations in adult persistent asthma. J Allergy Clin Immunol Pract. 2014;2(6):741-750. doi: 10.1016/j.jaip.2014.06.005 5. Zeiger RS, Schatz M, Dalal AA, Chen W, Sadikova E, Suruki RY, et al. Blood eosinophil count and outcomes in severe uncontrolled asthma: a prospective study. J Allergy Clin Immunol Pract. 2017;5(1):144-153:e8. doi: 10.1016/j.jaip.2016.07.015 6. Zeiger RS, Schatz M, Li Q, Chen W, Khatry DB, Gossage D, Tran TN. The association of blood eosinophil counts to future asthma exacerbations in children with persistent asthma. J Allergy Clin Immunol Pract. 2015; 3(2):283-287. doi: 10.1016/j.jaip.2014.10.009 . Records identified through database searching (n = 1567) EMBASE: 492 Medline: 695 WOS Proceedings: 293 WHO databases: 70 OATD: 17 Screening Included Eligibility Identification Additional records identified through other sources (n = 75) Records after duplicates removed (n = 1132) Records screened (n = 1132) Records excluded (n = 995) Full-text articles assessed for eligibility (n = 137) Studies included in qualitative synthesis (n = 23) Studies included in quantitative synthesis (meta-analysis) (n = 23) Full-text articles excluded, with reasons (n = 114) Do not measure the direct association between blood eosinophils count and asthma exacerbation (41) Insufficient data for the metaanalysis (49) OR cannot be calculated from dichotomous variables (4) Duplicated data about the same population (1) Conference proceedings that were published later as full text articles (13) Review articles (6) Overall Zeiger et al, 2015 Sohn et al, 2020 Mogensen et al, 2018 Nagasaki et al, 2019 Tran et al, 2014 Malinovschi et al, 2016 Velthove et al, 2009 Price et al, 2015 & 2016 Shah et al, 2019 Pola-Bibian et al, 2017 Price et al, 2019 Yii et al, 2019 Zeiger et al, 2017 Casciano et al, 2016 Turner et al, 2018 Kerkhof et al, 2018 Vedel-Krogh et al, 2017 Zeiger et al, 2014 Gonzalez-Barcala et al, 2018 Study Westerhof et al, 2016 Belda et al, 2001 Hakansson et al, 2019 1.31 (1.16, 1.49) 1.40 (1.02, 1.94) OR (95% CI) 1.69 (1.29, 2.20) 1.56 (1.15, 2.01) 1.25 (0.58, 2.67) 1.23 (0.92, 1.64) 1.25 (0.76, 2.07) 15.30 (3.90, 60.00) 1.42 (1.36, 1.47) 1.52 (1.30, 1.77) 0.46 (0.33, 0.64) 1.53 (1.09, 2.15) 1.80 (1.10, 2.90) 1.40 (1.02, 1.90) 3.06 (1.14, 8.22) 1.46 (1.20, 1.78) 1.49 (1.04, 2.13) 1.37 (1.18, 1.60) 1.31 (1.07, 1.60) 0.59 (0.37, 0.92) 2.94 (1.10, 8.10) 4.50 (1.80, 38.00) 0.64 (0.50, 0.90) 100.00 5.20 Weight 5.79 5.66 2.06 5.55 3.54 0.79 7.69 6.96 5.10 5.01 3.67 5.31 1.38 6.54 4.82 6.98 6.50 3.93 1.35 0.64 5.50 1.31 (1.16, 1.49) 1.40 (1.02, 1.94) OR (95% CI) 1.69 (1.29, 2.20) 1.56 (1.15, 2.01) 1.25 (0.58, 2.67) 1.23 (0.92, 1.64) 1.25 (0.76, 2.07) 15.30 (3.90, 60.00) 1.42 (1.36, 1.47) 1.52 (1.30, 1.77) 0.46 (0.33, 0.64) 1.53 (1.09, 2.15) 1.80 (1.10, 2.90) 1.40 (1.02, 1.90) 3.06 (1.14, 8.22) 1.46 (1.20, 1.78) 1.49 (1.04, 2.13) 1.37 (1.18, 1.60) 1.31 (1.07, 1.60) 0.59 (0.37, 0.92) 2.94 (1.10, 8.10) 4.50 (1.80, 38.00) 0.64 (0.50, 0.90) 100.00 5.20 Weight 5.79 5.66 2.06 5.55 3.54 0.79 7.69 6.96 5.10 5.01 3.67 5.31 1.38 6.54 4.82 6.98 6.50 3.93 1.35 0.64 5.50 Favor lower odds Favor higher odds 1.0167 1 60 OR (95%CI)