EG-1 POSITIVE EOSINOPHILS IN ASTHMA

To the Editor :

There is conflicting information about EG1-positive eosinophils contained in two articles published in the November 1999 issue of the American Journal of Respiratory and Critical Care Medicine.

Li and colleagues suggest that salmeterol has an antiinflammatory effect because of its ability to cause significant decreases in the number of EG1-positive eosinophils in the lamina propria of asthmatic patients (1). However, Faul and colleagues show that it is EG2-positive but not EG1-positive eosinophils that are a highly reproducible and reliable parameter of asthmatic airway inflammation (2).

We would like to ask the authors to comment on this apparent controversy.

Lahey Clinic Medical Center, Burlington, Massachusetts

1. Li X, Ward C, Bish FTR, Bamford T, Bao X, Bailey JW, Wilson JW, Walters EH. An antiinflammatory effect of salmeterol, a long acting ₂ agonist, assessed in airway biopsies and bronchoalveolar lavage in asthma. Am J Respir Crit Care Med 1999; 160: 1493-1499 [Abstract/Free Full Text].

2. Faul JL, Demers EA, Burke CM, Poulter LW. The reproductibility of repeat measures of airway inflammation in stable atopic asthma. Am J Respir Crit Care Med 1999; 160: 1457-1461 [Abstract/Free Full Text].

From the Authors :

We recently showed that the clinical benefit of 12 weeks of salmeterol treatment in asthma was accompanied by a significant reduction in airway eosinophilic infiltrate as assessed by immunostaining with EG-1 and a nonsignificant trend for reduced EG-2 counts (1).

This followed prospective evaluation of the reproducibility of airway biopsy (2) and bronchoalveolar lavage (3), which defined sample sizes for subsequent bronchoscopic studies. This novel work was subsequently followed by others, including the paper by Faul and colleagues (4).

Faul and colleagues confirmed our earlier findings that there was considerable variability in immunostaining of biopsy inflammation. In further analysis, intraclass correlation coefficients (ICC) were calculated to set a rather arbitrary cut off (0.6) for "acceptable" reproducibility. T cell, CD45RO and EG2 reproducibility over 2 weeks were deemed acceptable, whereas EG1 and CD68 were not. This paper (4) would seem to have originated from a pharmacological study (5), with a post-hoc analysis of the placebo limb. If we are correct in our assumption, it is unfortunate that only 9 of the 12 subjects in the original study (5) (Table 5, p. 759), were used in the subsequent analysis (4). An n of 9 would seem an inadequate basis for the didactic recommendations of the paper.

Our original reproductibility analyses (2, 3) were influenced by the work of eminent biostatisticians (6), with an increasing body of statistical literature expressing strong reservations about the use of ICCs (6). Our original paper (2) did not, however, include data on all the markers that we have studied. In our recent pharmacological study (1), we found a larger within-patient S.D. for EG-1 counts (15 cells/mm basement membrane) compared with EG-2 (7 cells/mm bm), but with twice the number of EG-1 cells counted (mean 13 versus 5 cells/mm bm). Our data support the view that EG-1 and EG-2 (and other biopsy markers) exhibit considerable variability, but that the use of adequate sample sizes and study design (2, 3) ensures that such biopsies provide valuable data in pathophysiological studies (1). Thus, our finding of decreased EG-1 counts in the lamina propria of asthmatics following salmeterol overcame the inherent biological variation.

We thank Dr. Stone and colleagues for their constructive questioning.

Freeman Hospital and University of Newcastle-upon-Tyne, United Kingdom

E. Haydn Walters

University of Tasmania, Hobart, Australia

1. Li X, Ward C, Thien F, Bish R, Bamford T, Bao X, Bailey M, Wilson JW, Walters EH. An antiiinflammatory effect of salmeterol, a long-acting ₂ agonist, assessed in airway biopsies and bronchoalveolar lavage in asthma. Am J Respir Crit Care Med 1999; 160: 1493-1499 .

2. Richmond I, Booth H, Ward C, Walters EH. Intrasubject variability in airway inflammation in biopsies in mild to moderate stable asthma. Am J Respir Crit Care Med 1996; 153: 899-903 [Abstract].

3. Ward C, Gardiner PV, Booth H, Walters EH. Intrasubject variability in airway inflammation sampled by bronchoalveolar lavage in stable asthmatics. Eur Respir J 1995; 8: 1866-1871 [Abstract].

4. Faul JL, Demers EA, Burke CM, Poulter LW. The reproducibility of repeat measures of airway inflammation in stable atopic asthma. Am J Respir Crit Care Med 1999; 160: 1457-1461 .

5. Faul JL, Leonard CT, Burke CM, Tormey VJ, Poulter LW. Fluticasone propionate induced alterations to lung function and the immunopathology of asthma over time. Thorax 1998; 53: 753-761 [Abstract/Free Full Text].

6. Bland JM, Altman DG. Comparing two methods of clinical measurement: a personal history. Int J Epidemiol 1995; 24: S7-S14 [Abstract].

From the Authors :

We thank Dr. Stone and colleagues for highlighting important issues that relate to the interpretation and design of biopsy studies in asthma (1). Li and colleagues demonstrate significant reductions in EG1 cell numbers on bronchial biopsy during salmeterol therapy for asthma (2). Although these changes might be due to a novel antiinflammatory effect of salmeterol therapy, we believe that significant alterations in EG1 might also occur due to biologic variability. The intraclass correlation coefficient (ICC) is a dimensionless statistic bound by 0 and 1 that describes the reliability of repeat measures in the same populations (3). In a clinically stable population, repeated measures with an ICC greater than 0.6 are thought to be clinically reliable, whereas repeat measures with an ICC less than 0.6 are generally not clinically useful (3). When we examine the reproducibility of repeat measures of airway inflammation in bronchial biopsies from clinically stable, steroid-naive, subjects with atopic asthma (who take beta agonists as required and a placebo inhaler twice daily), we find that some parameters of airway inflammation (e.g. T cells, EG2) are more reproducible than others (e.g., CD68 and EG1) (4). In particular, the ICC for repeat measures of EG1 is 0.4 after a 2-wk interval and 0.26 after an 8-wk interval (4). When we examine the reproducibility of repeat measures of EG1 in subjects who take high-dose inhaled fluticasone propionate (2000 µg/d) (5), we calculate an ICC of 0.4 after a 6-wk interval (unpublished data). Therefore, significant within-subject changes in EG1 numbers, quantified in bronchial biopsies, might occur due to natural variability. Such alterations in measures of EG1 are seen in the absence of a change in asthma therapy (4, 5), and in the absence of a significant change in asthma control (4, 5). Our interpretation of the current data is that intra-individual, between-biopsy, repeat measures of EG1 are highly variable and any observed change should be interpreted as due to a combination of biologic variability and the effect of an intervention.

Stanford University Medical Center, Stanford, California

Conor M. Burke

JCMH, Dublin, Ireland

Leonard W. Poulter

Royal Free Hospital, London, United Kingdom

1. Poulter LW, Burke CM, Jarjour NN, Pyke SD. Designing bronchial biopsy studies. Respir Med 2000;94(Suppl F)S3-S8.

2. Li X, Ward C, Thein F, Bish R, Bamford T, Bao X, Bailey M, Wilson JW, Walters EH. An anti-inflammatory effect of salmeterol, a long acting Beta 2-agonist, assessed in bronchial biopsies and bronchoalveolar lavage in asthma. Am J Respir Crit Care Med 1999; 160: 1493-1499 .

3. Armitage P, Berry G. Measurement error and correlation coefficients. Br Med J 1996; 313: 41-42 [Free Full Text].

4. Faul JL, Demers EA, Burke CM, Poulter LW. The reproducibility of repeat measures of airway inflammation in stable atopic asthma. Am J Respir Crit Care Med 1999; 160: 1457-1461 .

5. Faul JL, Leonard CT, Burke CM, Tormey VJ, Poulter LW. Fluticasone propionate induced alterations to lung function and the immunopathology of asthma over time. Thorax 1998; 53: 753-761 .