Histopathological Comparison of Endobronchial Biopsies from Different Pulmonary Lobes of Severe Asthmatic Patients

Abstract

Fiber-optic bronchoscopy with endobronchial biopsy (EBB) has become an important tool for accurate diagnosis and treatment selection in airway diseases. In a recent study, histopathological EBB findings revealed an overlapping COPD-asthma phenotype; however, the challenge of selecting the proper lung region for biopsy was raised because dissimilarities in the histology of different lobes could affect the homogeneity of samples.1Papakonstantinou E. Savic S. Siebeneichler A. et al.A pilot study to test the feasibility of histological characterisation of asthma-COPD overlap.Eur Respir J. 2019; 53: 1801941Crossref PubMed Scopus (7) Google Scholar In the present study, we compared histopathological features among different lung lobes of severe asthmatics. We included 30 patients with severe asthma undergoing evaluation for bronchial thermoplasty. The study was approved by the local institutional review board (EKNZ 2016-01057). The patient characteristics are shown in Table 1. Diagnosis of asthma was based on European Respiratory Society/American Thoracic Society guidelines and according to Global Initiative for Asthma 2019 criteria. Bronchoscopy procedures were performed by two senior pneumologists, transnasally or transorally, with the patients in the semirecumbent position. Nasal anesthesia was achieved by 2% lidocaine gel and the patients received propofol in repeated IV boluses, as previously described.2Grendelmeier P. Tamm M. Jahn K. Pflimlin E. Stolz D. Flexible bronchoscopy with moderate sedation in COPD: a case-control study.Int J Chron Obstruct Pulmon Dis. 2017; 12: 177-187Crossref PubMed Scopus (15) Google Scholar, 3Grendelmeier P. Tamm M. Pflimlin E. Stolz D. Propofol sedation for flexible bronchoscopy: a randomised, noninferiority trial.Eur Respir J. 2014; 43: 591-601Crossref PubMed Scopus (34) Google Scholar, 4Schlatter L. Pflimlin E. Fehrke B. Meyer A. Tamm M. Stolz D. Propofol versus propofol plus hydrocodone for flexible bronchoscopy: a randomised study.Eur Respir J. 2011; 38: 529-537Crossref PubMed Scopus (42) Google Scholar During the procedure, diagnostic EBB specimens were obtained from first- and second-generation bronchi using 2.2-mm-wide single use biopsy forceps with Endo-Glide Sheath (Radial Jaw, Boston Scientific). We chose to assemble samples from proximal airways because airway smooth muscle cell (ASMC) mass does not increase significantly in peripheral small airways of asthmatic patients.5Ijpma G. Panariti A. Lauzon A.M. Martin J.G. Directional preference of airway smooth muscle mass increase in human asthmatic airways.Am J Physiol Lung Cell Mol Physiol. 2017; 312: L845-L854Crossref PubMed Scopus (24) Google Scholar,6Kuwano K. Bosken C.H. Pare P.D. Bai T.R. Wiggs B.R. Hogg J.C. Small airways dimensions in asthma and in chronic obstructive pulmonary disease.Am Rev Respir Dis. 1993; 148: 1220-1225Crossref PubMed Scopus (517) Google Scholar From each patient, we obtained 10 EBBs from five locations: right upper lobe (one biopsy), middle lobe (two biopsies), right lower lobe (three biopsies), lingula (one biopsy), and left lower lobe (three biopsies). Specimens were washed in phosphate-buffered saline, fixed in formalin, and transferred to pathology. For each specimen, five sequential sections were stained with hematoxylin and eosin and Elastica van Gieson, adding to 5 to 15 analytic units per lobe and up to 50 analytic units per patient that were evaluated blindly by two senior pathologists for: (1) inflammation in the stroma, tissue lymphocyte/eosinophil/granulocyte infiltration and thickening of reticular basement membrane (BM), using a 0 to 3 scale: 0 = absence/ normal, 1 = mild-moderate, 2-3 = severe; (2) ASMC mass (total percentage of the submucosal area occupied by ASMC); and (3) distance between BM and ASMC in micrometers, measured from the parenchymal site of BM toward ASMC, without including the thickness of the BM and at the part of the biopsy where ASMCs were closest to BM (the shortest distance in all biopsies). We evaluated only specimens with tangential sections (n = 96) so that the orientation of the sections would not affect measurements. We assigned the median score of all analytic units per lobe when evaluating qualitative values such as inflammation, tissue lymphocyte/eosinophil/granulocyte infiltration and thickening of reticular BM, and the mean score of all analytic units per lobe for the quantitative values such as ASMC mass and BM-ASMC distance. The scores of both pathologists agreed in more than 95% in the evaluation of each of seven parameters for individual analytic units.Table 1Characteristics of the PatientsParameterSevere Asthma Patients (n = 30)Age (y, mean ± SD)57.8 ± 2.8Sex (female %)56.7%Pack-years ± SD14.0 ± 22.3BMI ± SD27.8 ± 5.0Therapy, No. (%) LABA+ICS30 (100) Oral corticosteroids20 (66.7)Skin prick test positivity, No. (%)16 (53.3)Blood eosinophils (n × 109/L, mean ± SD)0.3 ± 0.3Blood leukocytes (n × 109/L, mean ± SD)8.8 ± 2.4Blood neutrophils (n × 109/L, mean ± SD)5.9 ± 2.6CRP (mg/L)2.6 ± 2.4Current smoking, No. (%)4 (13.3)Post-BD FEV1 % pred (mean ± SD)64.8 ± 19.3Post-BD FVC % pred (mean ± SD)92.7 ± 19.4Post-BD FEV1/FVC (mean ± SD)0.58 ± 0.15FEV1 reversibility % (mean ± SD)11.4 ± 8Post-BD Dlco % pred (mean ± SD)92.7 ± 19.6Blood cell numbers represent absolute numbers. BD = bronchodilator; CRP = C-reactive protein; Dlco = diffusing capacity of the lung for carbon monoxide; ICS = inhaled corticosteroid; LABA = long-acting beta-adrenoceptor agonists; pred = prednisone. Open table in a new tab Blood cell numbers represent absolute numbers. BD = bronchodilator; CRP = C-reactive protein; Dlco = diffusing capacity of the lung for carbon monoxide; ICS = inhaled corticosteroid; LABA = long-acting beta-adrenoceptor agonists; pred = prednisone. Associations between lung lobes and the three different levels of severity of inflammation and existence of infiltration and thickening of BM (absence/normal, mild/moderate, and severe) were examined using mixed ordinal logistic regression models. The effect of the patient was included as a random effect in the model and the step-down Bonferroni method was used to adjust for multiple comparisons. Associations between lung lobes and ASMC mass and BM-ASMC distance were examined using ranked transformed mixed linear regression models. In these models, the effect of the patient was included as a random effect, but the Dunnett-Hsu method was used to adjust for multiple comparisons. Associations between upper and lower lung lobes and ASMC mass and BM-ASMC distance were examined using rank transformed mixed linear regression models, where the factor patient was included as a random effect and the position in the lung as repeated effect (two measurements for the upper position and three measurements for the lower position). There were no significant differences among the different lobes (Table 2) concerning inflammation in the stroma, tissue lymphocyte, eosinophil and granulocyte infiltration, and thickening of the BM. Significant difference between the upper lobes and the lower lobes was observed only for tissue lymphocyte infiltration (OR, 2.12; 95% CI, 1.05-4.29; P = .036) (Table 2). Conversely, we observed a significant difference in the ASMC mass and the BM-ASMC distance among lobes (P < .001 and P = .017 accordingly; Table 2, Fig 1). The right upper lobe presented less ASMC than the other lobes and the left upper lobe depicted the smallest BM-ASMC distance (Fig 1). To our knowledge, the significantly less ASMC in the right upper lobe, compared with other lobes, cannot so far be explained by any embryological developmental step or biologic rational. The discrepancies in the BM-ASMC distance were not significantly associated with the level of inflammation (adjusted central value [95% CI], 58.24 [37.60-73.48] for absence, 59.44 [45.50-71.50] for mild-moderate, and 71.53 [56.11-89.49] for severe, P = .315). Indeed, Elliot et al7Elliot J.G. Noble P.B. Mauad T. et al.Inflammation-dependent and independent airway remodelling in asthma.Respirology. 2018; 23: 1138-1145Crossref PubMed Scopus (37) Google Scholar have confirmed that the thickening of the ASMC layer and the BM may be present even with the absence of inflammation. Significant differences were found in ASMC mass and BM-ASMC distance when we compared the upper and lower lobes because the upper lobes present a smaller ASMC mass and a smaller BM-ASMC distance (for ASMC mass, adjusted central value [95% CI], 14.32 [12.46-19.99] for lower and 9.06 [6.57-12.17] for upper, P = .003) and for BM-ASMC distance adjusted central value [95% CI], 68.40 [57.18-74.60] for lower and 48.08 [35.66-64.37] for upper, P = .018; Table 2, Fig 1). No significant differences were found in ASMC mass and BM-ASMC distance when we compared the left to the right lobes clustered together. The significance of the results remained intact when we omitted the middle lobe from the cluster of the upper lobes (data not shown).Table 2Descriptive Statistics of the Histological Outcomes From Endobronchial Biopsies Obtained From Different Locations of Pulmonary Lobes in Patients With Severe AsthmaParameterRight Upper LobeMiddle LobeRight Lower LobeLeft Upper LobeLeft Lower LobeP Value for the Effect of the Factor“Lung Lobe” on the OutcomeUpper LobesLower LobesP Value for the Effect of the Factor“Lung Position” (Lower vs Upper) on the OutcomeInflammation in the stromaaQualitative evaluation, 0-3 scale: 0 = absence/normal, 1 = mild-moderate, 2-3 = severe.n = 28n = 28n = 29n = 28n = 27.192bCalculated using mixed multinomial (ordinal) logistic regression models, where the factor patient was included as a random effect and the step-down Bonferroni method was used to adjust for multiple comparisons.n = 56n = 84.646bCalculated using mixed multinomial (ordinal) logistic regression models, where the factor patient was included as a random effect and the step-down Bonferroni method was used to adjust for multiple comparisons. Absence, No. (%)8 (28.6)9 (32.1)3 (10.3)8 (28.6)4 (14.8)16 (28.6)16 (19.1) Mild-moderate, No. (%)11 (39.3)15 (53.6)19 (65.5)16 (57.1)19 (70.4)27 (48.2)53 (63.1) Severe, No. (%)9 (32.1)4 (14.3)7 (24.1)4 (14.3)4 (14.8)13 (23.2)15 (17.9)Tissue lymphocyte infiltrationaQualitative evaluation, 0-3 scale: 0 = absence/normal, 1 = mild-moderate, 2-3 = severe.n = 28n = 27n = 29n = 27n = 27.051bCalculated using mixed multinomial (ordinal) logistic regression models, where the factor patient was included as a random effect and the step-down Bonferroni method was used to adjust for multiple comparisons.n = 55n = 83.036bCalculated using mixed multinomial (ordinal) logistic regression models, where the factor patient was included as a random effect and the step-down Bonferroni method was used to adjust for multiple comparisons. Absence, No. (%)3 (10.7)5 (18.5)6 (20.7)6 (22.2)6 (22.2)9 (16.4)17 (20.5) Mild-moderate, No. (%)12 (42.9)18 (66.7)15 (51.7)14 (51.9)17 (63.0)26 (47.3)50 (60.2) Severe, No. (%)13 (46.4)4 (14.8)8 (27.6)7 (25.9)4 (14.8)20 (36.4)16 (19.3)Tissue eosinophil infiltrationaQualitative evaluation, 0-3 scale: 0 = absence/normal, 1 = mild-moderate, 2-3 = severe.n = 28n = 28n = 29n = 28n = 27.407bCalculated using mixed multinomial (ordinal) logistic regression models, where the factor patient was included as a random effect and the step-down Bonferroni method was used to adjust for multiple comparisons.n = 56n = 84.610bCalculated using mixed multinomial (ordinal) logistic regression models, where the factor patient was included as a random effect and the step-down Bonferroni method was used to adjust for multiple comparisons. Absence, No. (%)18 (64.3)19 (67.9)15 (51.7)15 (53.6)14 (51.9)33 (58.9)48 (57.1) Mild-moderate, No. (%)8 (28.6)5 (17.9)8 (27.6)8 (28.6)10 (37.0)16 (28.6)23 (27.4) Severe, No. (%)2 (7.1)4 (14.3)6 (20.7)5 (17.9)3 (11.1)7 (12.7)13 (15.5)Granulocytes in the stromaaQualitative evaluation, 0-3 scale: 0 = absence/normal, 1 = mild-moderate, 2-3 = severe.n = 28n = 28n = 29n = 28n = 27.594bCalculated using mixed multinomial (ordinal) logistic regression models, where the factor patient was included as a random effect and the step-down Bonferroni method was used to adjust for multiple comparisons.n = 56n = 84.286bCalculated using mixed multinomial (ordinal) logistic regression models, where the factor patient was included as a random effect and the step-down Bonferroni method was used to adjust for multiple comparisons. Absence, No. (%)23 (82.1)23 (82.1)22 (75.9)26 (92.9)23 (85.2)49 (87.5)68 (80.9) A few, No. (%)5 (17.9)4 (14.3)7 (24.1)2 (7.1)3 (11.1)7 (12.5)14 (16.7) Many, No. (%)0 (0.0)1 (3.6)001 (3.7)0 (0.0)2 (2.4)BM thickeningaQualitative evaluation, 0-3 scale: 0 = absence/normal, 1 = mild-moderate, 2-3 = severe.n = 28n = 28n = 29n = 28n = 27.206bCalculated using mixed multinomial (ordinal) logistic regression models, where the factor patient was included as a random effect and the step-down Bonferroni method was used to adjust for multiple comparisons.n = 56n = 84.258bCalculated using mixed multinomial (ordinal) logistic regression models, where the factor patient was included as a random effect and the step-down Bonferroni method was used to adjust for multiple comparisons. Normal, No. (%)5 (17.9)6 (21.4)2 (6.9)3 (10.7)5 (18.5)8 (14.3)13 (15.5) Mild-moderate, No. (%)16 (57.1)15 (53.6)20 (69.0)12 (42.9)14 (51.9)28 (50.0)49 (58.3) Severe, No. (%)7 (25.0)7 (25.0)7 (24.1)13 (46.4)8 (29.6)20 (35.7)22 (26.2)ASMC mass (%)n = 27n = 28n = 29n = 28n = 26<.001cCalculated using rank-transformed mixed linear regression models, where the factor patient was included as a random effect and the Dunnett-Hsu method was used to adjust for multiple comparisons.n = 55n = 83.003dCalculated using rank-transformed mixed linear regression models, where the factor patient was included as a random effect and position in the lung as repeated effect (two measurements for the upper position and three measurements for the lower position). Mean ± SD9.4 ± 20.119.8 ± 16.919.4 ± 17.722.5 ± 24.217.5 ± 13.316.1 ± 23.018.9 ± 16.0 Median (IQR)1 (0-10.0)16.3 (6.3-30.0)13.3 (5.0-30.0)10.3 (5.0-40.0)14.2 (8.8-20.0)5 (0-20.0)15.0 (7.5-25.0)Distance BM-ASMC, (μm)n = 11n = 25n = 23n = 18n = 22.017cCalculated using rank-transformed mixed linear regression models, where the factor patient was included as a random effect and the Dunnett-Hsu method was used to adjust for multiple comparisons.n = 29n = 70.018dCalculated using rank-transformed mixed linear regression models, where the factor patient was included as a random effect and position in the lung as repeated effect (two measurements for the upper position and three measurements for the lower position). Mean ± SD74.8 ± 56.958.8 ± 32.884.8 ± 57.242.6 ± 34.680.9 ± 43.054.8 ± 46.274.2 ± 46.0 Median (IQR)80 (30.0-112.0)53.3 (37.6-70.0)73.5 (33.9-124.5)29 (17.0-75.0)71.3 (45.5-123.0)33 (17.0-84.5)67.3 (37.6-96.0)For the “Upper lobes” cluster, we combined the data from the right and left upper lobes and for the “Lower lobes” cluster we combined the data from the right lower, left lower and middle lobes. ASMC = airway smooth muscle cell; BM = basement membrane; IQR = interquartile range.a Qualitative evaluation, 0-3 scale: 0 = absence/normal, 1 = mild-moderate, 2-3 = severe.b Calculated using mixed multinomial (ordinal) logistic regression models, where the factor patient was included as a random effect and the step-down Bonferroni method was used to adjust for multiple comparisons.c Calculated using rank-transformed mixed linear regression models, where the factor patient was included as a random effect and the Dunnett-Hsu method was used to adjust for multiple comparisons.d Calculated using rank-transformed mixed linear regression models, where the factor patient was included as a random effect and position in the lung as repeated effect (two measurements for the upper position and three measurements for the lower position). Open table in a new tab For the “Upper lobes” cluster, we combined the data from the right and left upper lobes and for the “Lower lobes” cluster we combined the data from the right lower, left lower and middle lobes. ASMC = airway smooth muscle cell; BM = basement membrane; IQR = interquartile range. We further explored the association of blood eosinophils with histopathological measurements including ASMC mass, airway tissue eosinophilic infiltration, and BM thickness. Blood eosinophilia was significantly associated with ASMC mass in the right lower lobe (linear regression model: beta coefficient, 0.008, SE, 0.003, R2 %, 24.122, P = .009). Blood eosinophilia was not significantly related to tissue eosinophilic infiltration or BM thickening at any lobe or clustered lobes. Similarly, among different levels of BM thickening, there was no significant difference in tissue eosinophilic inflammation at any lobe or clustered lobes. To our knowledge, no other histopathological study has been performed in such a large group of severe asthmatics (n = 30). With asthma severity, ASMC mass increases, BM-ASMC distance decreases,8Benayoun L. Druilhe A. Dombret M.C. Aubier M. Pretolani M. Airway structural alterations selectively associated with severe asthma.Am J Respir Crit Care Med. 2003; 167: 1360-1368Crossref PubMed Scopus (672) Google Scholar and BM thickness increases.9Bourdin A. Neveu D. Vachier I. Paganin F. Godard P. Chanez P. Specificity of basement membrane thickening in severe asthma.J Allergy Clin Immunol. 2007; 119: 1367-1374Abstract Full Text Full Text PDF PubMed Scopus (115) Google Scholar The main finding of our study is that ASMC mass and BM-ASMC distance differ between the lobes of the same patient, indicating that the lung location of EBB affects significantly the pathological results obtained. However, those qualitative differences in the BM-ASMC distance could partially be attributed to artifacts produced mechanically during EBB because the forceps squash the area under the BM and not uniformly dissect it. In another study of 10 severe asthmatic patients, Pretolani et al10Pretolani M. Dombret M.C. Knap D. et al.Reduction of airway smooth muscle mass by bronchial thermoplasty in patients with severe asthma.Am J Respir Crit Care Med. 2014; 190: 1452-1454Crossref PubMed Scopus (130) Google Scholar also demonstrated that before bronchial thermoplasty, the different lung lobes present with different percentages of ASMC area. However, it was not stated if this difference across the lung lobes was statistically significant and it was indicated that the right upper lobe has the highest percentage of ASMC area.10Pretolani M. Dombret M.C. Knap D. et al.Reduction of airway smooth muscle mass by bronchial thermoplasty in patients with severe asthma.Am J Respir Crit Care Med. 2014; 190: 1452-1454Crossref PubMed Scopus (130) Google Scholar This apparent discrepancy to our results may be attributed to the small sample size of the Pretolani study.10Pretolani M. Dombret M.C. Knap D. et al.Reduction of airway smooth muscle mass by bronchial thermoplasty in patients with severe asthma.Am J Respir Crit Care Med. 2014; 190: 1452-1454Crossref PubMed Scopus (130) Google Scholar To date, no other study has evaluated histological differences across the different pulmonary lobes. A limitation of our study is, however, the lack of biopsy samples from control subjects that would help us to clarify if the observed differences are solely related to asthma or are also similar in healthy subjects. Unfortunately, because ethical issues are raised for obtaining multiple endobronchial biopsies from healthy subjects, we were not able to include such samples in our study. The present study indicates that the structure of the mesenchymal and epithelial airway components assessed by EBB may differ according to the anatomical location within the bronchial tree. Thus, the provenance of biopsies should always be considered to obtain homogenous and accurate results. Novel studies investigating histopathological alterations in different pulmonary lobes of healthy individuals and COPD patients are additionally anticipated. Role of sponsors: The sponsor had no role in the design of the study, the collection and analysis of the data, or the preparation of the manuscript.