Alveolar Attachments in Emphysema of Human Lungs 1- 3
ATSUSHI NAGAI, ISAO YAMAWAKI, TAKAO TAKIZAWA, and WILLIAM M. THURLBECK
Introduction
On the basis of morphologic studies, Pratt and his colleagues (1) and Anderson and Foraker (2) proposed that the loss of radial traction around bronchioles may be responsible for chronic airflow obstruction (CAO), particularly on expiration due to the loss of radial traction. Wilson and coworkers (3), using post-mortem bronchography and morphology, showed undue narrowing on expiration of focal areas of stenosisofbronchioles and narrowing of bronchioles less than 1 mm in diameter in emphysematous lungs. They attributed this to the loss of elastic recoil. The idea that bronchioles narrowed unduly on expiration was thought unlikely by Hogg and colleagues (4), who found that peripheral airway resistance was not different on inspiration compared with expiration. Their series contained patients with extensiveintraluminal exudate, however,Burrows and coworkers (5) found that expiratory flow resistance was greater in expiration in patients with the emphysematous type of airflow obstruction, whereas inspiratory resistance was increased in patients with dominantly airway disease (chronic bronchitis). Airway narrowing (2) and bronchiolar deformity (6) have been asssociated with loss of alveolar attachment, and we found that bronchiolar deformity and an excess of bronchioles less than 400 urn in diameter were the most important bronchiolar lesions associated with chronic airflow obstruction in patients in the National Institutes of Health (NIH) Intermittent Positive Pressure Breathing (IPPB) trial, most of whom had moderate and severe chronic airflow obstruction (7). The concept of the importance of alveolar attachments has been revived by the Saetta (8) and Petty groups (9, 10). Both these groups of investigators found an association between loss of recoil and diminished radial attachments. Petty and colleagues found a relationship between bronchiolar narrowing and diminished alveolar attachments and also an association between both of these and a re888
SUMMARY We have studied alveolar attachments to membranous bronchioles In 41 patients enrolled In the National Institutes of Health Intermittent Positive Pressure Breathing Trial who died, came to autopsy, and provided adequate tissue for examination. The patients had moderate to severe chronic airflow obstruction and, on the average,severe emphysema. Wemeasured the number of normal attachments per membranous bronchiole per case (N), the number of normal attachments per mm of circumference per bronchiole per case (Nunit), and the ratio of abnormal to all attachments (R). All measurements correlated closely to all measurements of emphysema and to Irregular shape of bronchioles (deformity Index). Measurement of abnormal attachments correlated with abnormal tests of pUlmonary function, Including evidence of airflow obstruction and diffusing eapaelty for carbon monoxide (Oleo). Nunit showed the most frequent correlations, but both Nand Nunlt were only related independently of emphysema for Phase III of the single-breath nitrogen test. R was related to loss of OLeo. We conclude that loss of attachments is not itself a cause of airflow obstruction but rather a correlate of emphysema, which Is the proximate cause of the obstruction. Loss of alveolar attachments has an effect only on the single-breath nitrogen tests Independent of emphysema. AM REV RESPIR DIS 1991; 144:888-891
duced postmortem FEV 1 (9). Increased postmortem closing capacity was also found in patients with reduced alveolar attachments. The series just mentioned used patients whose lobes or lungs were resected for lung cancer with little or no CAO and/or consecutive postmortem lungs. We wished to study alveolar attachments in the well-defined series of patients with moderate and severe CAO. Methods Tissue sections wereobtained from 48 autopsy cases from the National Institutes of Health Intermittent Positive Pressure Breathing clinical trial, and we selected and used 41 cases whose tissue sections were adequate to make morphologic assessments. The criteria for admission to the study were that the subjects had a clinical diagnosis of chronic obstructive pulmonary disease and a prebronchodilator FEV 1 less than 60070 of predicted as well as a prebronchodilator ratio of FEV 1 to FVC less than 0.60. Exclusion criteria from the study included a diagnosis of asthma, total lung capacity lessthan 80% of predicted, radiologic evidenceof other significant complicating lung disease, significant cardiac disease, other life-threatening illness unrelated to chronic airflow obstruction, or arterial oxygen pressure less than 55 rom Hg (11). Pulmonary function data in the present study are those following bronchodilator closest to the time of death and are expressed as percentage of predicted. The morphologic techniques used here have been described in detail elsewhere (12). Relevant to this paper, excised lungs, usually the
left, were inflated for at least 24 h at a transpulmonary pressure of 25 em formalin or glutaraldehyde. The lungs were cut into sagittal slices, and the midsagittal slices were used to produce paper-mounted whole-lung sections. Six stratified random tissue blocks were taken from the medial surface of the slice lateral to the midsagittal slice.Tissue sections 5 11m thick were made and stained with hematoxylin and eosin. Emphysema was measured using the panel grading method on paper-mounted whole-lung sections (Paper) (13),the section grading method using histologic sections (Section) (14), and the mean linear intercept (MLI). Bronchiolar lesions were assessed using the method of Cosio and coworkers (15),and measurement of deformity of bronchioles as described by Nagai's group (12). Briefly, the mean number of normal attachments (N) per membranous bronchiole (nonalveolated, noncartilaginous airways less
(Received in original form May 31, 1990 and in revised form May 3, 1991) 1 From the First Department of Medicine, Tokyo Women's Medical College, Japan, and the Department of Pathology, University of British Columbia, Vancouver, Canada. 2 Supported by Grant-in-Aid for Scientific Research No. 61570379from the Ministry of Education, Science and Culture, Japan, and Grant No. MT-7124 from the Medical Research Council of Canada. 3 Correspondence and requests for reprints should be addressed to Atsushi Nagai, M.D., First Department of Medicine, Tokyo Women's Medical College, 8-1,Kawada-cho, Shinjuku-ku, Tokyo 162, Japan.
ALVEOLAR ATTACHMENTS IN EMPHYSEMA
889 TABLE 1
PATIENT CHARACTERISTICS IN THE STUDY Minimum
5
Paper· Age, yr FEV" 0Al 01 predicted FEV,'FVC, %
53
12 16
Maximum
Mean
SO
90 74 65 58
62.2 63.4 36.7 37.2
20.5 6.4
14.2 9.2
• Assessment of emphysema using the panel grading method.
TABLE 2 CORRELATION COEFFICIENTS BETWEEN ALVEOLAR ATTACHMENTS AND MORPHOLOGIC DATA IN MEMBRANOUS BRONCHIOLES AND LUNG PARENCHYMA· Membranous Bronchioles
Parenchyma Paper Nunit N R
-0.56 -0.46 0.44
MLI
Section
-0.57 -0.59 0.40
-0.58 -0.53 0.43
Pig -0.32 -0.36
-t
Inll
_t _t
_t
Fibr
Def.l.
-t
0.59 0.40 -0.45
0.35
-t
Definition of abbreviations: Nunit = number of normal attachmentslmm circumference; N = number of normal attachments per bronchiole; R = ratio of abnormal to all attachments; Paper = panel grading assessment of emphysema; MLI = mean linear intercept; Section = section grading method of assessing emphysema; Pig = pigmentation; Inll = inflammation; Fibr = fibrosis; Def.l. = deformity index.
• r = 0.31, P < 0.05; r = 0.40, P < 0,01.
t
Not significant.
than 2 mm in internal diameter) per case was calculated, and the internal diameter and circumference of bronchioles were measured using a computer-assisted digitizer. Normal and abnormal attachments were determined as described by Saetta's group (8). The ratio of abnormal attachments (R) was calculated. Saetta and colleagues also measured the distance between normal attachments by dividing the circumference by the number of attachments. We used the reciprocal of this, thus expressing the data as the number of normal attachments per mm circumference (Nunit), since this may obviate the problem that N may reflect the size of airways. Statistical analysis included correlation coefficients and multiple comparison using one-wayanalysisof variance and the StudentNewman-Keuls test for three independent samples. Since we found significant correlations between emphysema and certain mea100 90 80
i .:
8-
«
surements of alveolar attachments, we performed partial regression and partial correlation tests between emphysema, alveolar attachments, and pulmonary function tests. A levelof p < 0.05wasaccepted as significant. Results
The characteristics of the patients in the study are shown in table 1. Most patients had moderate or severe emphysema as assessed by a paper-mounted gradingsystem. These patients had moderate and severe airflow limitation. The correlations between alveolar attachments and parenchymal and bron-
chiolar lesions are shown in table 2. As measurements of alveolar attachments worsened, so did measurements of emphysema, and this applied to all measurements of emphysema and all measurements of attachments. Correlations were reasonably high between emphysema and Nand Nunit, but considerable scatter was presented in the data (figure 1). When bronchioles were analyzed by size of internal diameter and severity of emphysema (figure 2), the attachments wereparticularly reduced in bronchioles up to 800 J.1m in diameter, and in all groups moderate to severe emphysema had fewer attachments than mild emphysema. The differences between moderate-severe and severe emphysema were not significant. All measurements of attachments showed correlations with the deformity index: as they got worse, there was more bronchiolar deformity. The highest correlation was between the Nunit and the deformity index (r = 0.59), and once again there was scatter (figure 3). As Nunit and N worsened, there was more pigment. N was also related to fibrosis, there being less fibrosis with decreased attachments. Inflammation was not associated with loss of attachments. Bronchiolar muscle and goblet cell metaplasia were not related to attachments. Correlation coefficients between alveolar attachments and emphysema and tests of pulmonary function are shown in table 3. Nunit showed the highest number of significant correlations, and R the fewest. In general, the correlations are poor and considerable scatter was present, an example (FEV 1/ FVC) being shown in figure 4. Correlation coefficients were also consistently
NS ,.--,
P
ATSUSHI NAGAI, ISAO YAMAWAKI, TAKAO TAKIZAWA, and WILLIAM M. THURLBECK
Introduction
On the basis of morphologic studies, Pratt and his colleagues (1) and Anderson and Foraker (2) proposed that the loss of radial traction around bronchioles may be responsible for chronic airflow obstruction (CAO), particularly on expiration due to the loss of radial traction. Wilson and coworkers (3), using post-mortem bronchography and morphology, showed undue narrowing on expiration of focal areas of stenosisofbronchioles and narrowing of bronchioles less than 1 mm in diameter in emphysematous lungs. They attributed this to the loss of elastic recoil. The idea that bronchioles narrowed unduly on expiration was thought unlikely by Hogg and colleagues (4), who found that peripheral airway resistance was not different on inspiration compared with expiration. Their series contained patients with extensiveintraluminal exudate, however,Burrows and coworkers (5) found that expiratory flow resistance was greater in expiration in patients with the emphysematous type of airflow obstruction, whereas inspiratory resistance was increased in patients with dominantly airway disease (chronic bronchitis). Airway narrowing (2) and bronchiolar deformity (6) have been asssociated with loss of alveolar attachment, and we found that bronchiolar deformity and an excess of bronchioles less than 400 urn in diameter were the most important bronchiolar lesions associated with chronic airflow obstruction in patients in the National Institutes of Health (NIH) Intermittent Positive Pressure Breathing (IPPB) trial, most of whom had moderate and severe chronic airflow obstruction (7). The concept of the importance of alveolar attachments has been revived by the Saetta (8) and Petty groups (9, 10). Both these groups of investigators found an association between loss of recoil and diminished radial attachments. Petty and colleagues found a relationship between bronchiolar narrowing and diminished alveolar attachments and also an association between both of these and a re888
SUMMARY We have studied alveolar attachments to membranous bronchioles In 41 patients enrolled In the National Institutes of Health Intermittent Positive Pressure Breathing Trial who died, came to autopsy, and provided adequate tissue for examination. The patients had moderate to severe chronic airflow obstruction and, on the average,severe emphysema. Wemeasured the number of normal attachments per membranous bronchiole per case (N), the number of normal attachments per mm of circumference per bronchiole per case (Nunit), and the ratio of abnormal to all attachments (R). All measurements correlated closely to all measurements of emphysema and to Irregular shape of bronchioles (deformity Index). Measurement of abnormal attachments correlated with abnormal tests of pUlmonary function, Including evidence of airflow obstruction and diffusing eapaelty for carbon monoxide (Oleo). Nunit showed the most frequent correlations, but both Nand Nunlt were only related independently of emphysema for Phase III of the single-breath nitrogen test. R was related to loss of OLeo. We conclude that loss of attachments is not itself a cause of airflow obstruction but rather a correlate of emphysema, which Is the proximate cause of the obstruction. Loss of alveolar attachments has an effect only on the single-breath nitrogen tests Independent of emphysema. AM REV RESPIR DIS 1991; 144:888-891
duced postmortem FEV 1 (9). Increased postmortem closing capacity was also found in patients with reduced alveolar attachments. The series just mentioned used patients whose lobes or lungs were resected for lung cancer with little or no CAO and/or consecutive postmortem lungs. We wished to study alveolar attachments in the well-defined series of patients with moderate and severe CAO. Methods Tissue sections wereobtained from 48 autopsy cases from the National Institutes of Health Intermittent Positive Pressure Breathing clinical trial, and we selected and used 41 cases whose tissue sections were adequate to make morphologic assessments. The criteria for admission to the study were that the subjects had a clinical diagnosis of chronic obstructive pulmonary disease and a prebronchodilator FEV 1 less than 60070 of predicted as well as a prebronchodilator ratio of FEV 1 to FVC less than 0.60. Exclusion criteria from the study included a diagnosis of asthma, total lung capacity lessthan 80% of predicted, radiologic evidenceof other significant complicating lung disease, significant cardiac disease, other life-threatening illness unrelated to chronic airflow obstruction, or arterial oxygen pressure less than 55 rom Hg (11). Pulmonary function data in the present study are those following bronchodilator closest to the time of death and are expressed as percentage of predicted. The morphologic techniques used here have been described in detail elsewhere (12). Relevant to this paper, excised lungs, usually the
left, were inflated for at least 24 h at a transpulmonary pressure of 25 em formalin or glutaraldehyde. The lungs were cut into sagittal slices, and the midsagittal slices were used to produce paper-mounted whole-lung sections. Six stratified random tissue blocks were taken from the medial surface of the slice lateral to the midsagittal slice.Tissue sections 5 11m thick were made and stained with hematoxylin and eosin. Emphysema was measured using the panel grading method on paper-mounted whole-lung sections (Paper) (13),the section grading method using histologic sections (Section) (14), and the mean linear intercept (MLI). Bronchiolar lesions were assessed using the method of Cosio and coworkers (15),and measurement of deformity of bronchioles as described by Nagai's group (12). Briefly, the mean number of normal attachments (N) per membranous bronchiole (nonalveolated, noncartilaginous airways less
(Received in original form May 31, 1990 and in revised form May 3, 1991) 1 From the First Department of Medicine, Tokyo Women's Medical College, Japan, and the Department of Pathology, University of British Columbia, Vancouver, Canada. 2 Supported by Grant-in-Aid for Scientific Research No. 61570379from the Ministry of Education, Science and Culture, Japan, and Grant No. MT-7124 from the Medical Research Council of Canada. 3 Correspondence and requests for reprints should be addressed to Atsushi Nagai, M.D., First Department of Medicine, Tokyo Women's Medical College, 8-1,Kawada-cho, Shinjuku-ku, Tokyo 162, Japan.
ALVEOLAR ATTACHMENTS IN EMPHYSEMA
889 TABLE 1
PATIENT CHARACTERISTICS IN THE STUDY Minimum
5
Paper· Age, yr FEV" 0Al 01 predicted FEV,'FVC, %
53
12 16
Maximum
Mean
SO
90 74 65 58
62.2 63.4 36.7 37.2
20.5 6.4
14.2 9.2
• Assessment of emphysema using the panel grading method.
TABLE 2 CORRELATION COEFFICIENTS BETWEEN ALVEOLAR ATTACHMENTS AND MORPHOLOGIC DATA IN MEMBRANOUS BRONCHIOLES AND LUNG PARENCHYMA· Membranous Bronchioles
Parenchyma Paper Nunit N R
-0.56 -0.46 0.44
MLI
Section
-0.57 -0.59 0.40
-0.58 -0.53 0.43
Pig -0.32 -0.36
-t
Inll
_t _t
_t
Fibr
Def.l.
-t
0.59 0.40 -0.45
0.35
-t
Definition of abbreviations: Nunit = number of normal attachmentslmm circumference; N = number of normal attachments per bronchiole; R = ratio of abnormal to all attachments; Paper = panel grading assessment of emphysema; MLI = mean linear intercept; Section = section grading method of assessing emphysema; Pig = pigmentation; Inll = inflammation; Fibr = fibrosis; Def.l. = deformity index.
• r = 0.31, P < 0.05; r = 0.40, P < 0,01.
t
Not significant.
than 2 mm in internal diameter) per case was calculated, and the internal diameter and circumference of bronchioles were measured using a computer-assisted digitizer. Normal and abnormal attachments were determined as described by Saetta's group (8). The ratio of abnormal attachments (R) was calculated. Saetta and colleagues also measured the distance between normal attachments by dividing the circumference by the number of attachments. We used the reciprocal of this, thus expressing the data as the number of normal attachments per mm circumference (Nunit), since this may obviate the problem that N may reflect the size of airways. Statistical analysis included correlation coefficients and multiple comparison using one-wayanalysisof variance and the StudentNewman-Keuls test for three independent samples. Since we found significant correlations between emphysema and certain mea100 90 80
i .:
8-
«
surements of alveolar attachments, we performed partial regression and partial correlation tests between emphysema, alveolar attachments, and pulmonary function tests. A levelof p < 0.05wasaccepted as significant. Results
The characteristics of the patients in the study are shown in table 1. Most patients had moderate or severe emphysema as assessed by a paper-mounted gradingsystem. These patients had moderate and severe airflow limitation. The correlations between alveolar attachments and parenchymal and bron-
chiolar lesions are shown in table 2. As measurements of alveolar attachments worsened, so did measurements of emphysema, and this applied to all measurements of emphysema and all measurements of attachments. Correlations were reasonably high between emphysema and Nand Nunit, but considerable scatter was presented in the data (figure 1). When bronchioles were analyzed by size of internal diameter and severity of emphysema (figure 2), the attachments wereparticularly reduced in bronchioles up to 800 J.1m in diameter, and in all groups moderate to severe emphysema had fewer attachments than mild emphysema. The differences between moderate-severe and severe emphysema were not significant. All measurements of attachments showed correlations with the deformity index: as they got worse, there was more bronchiolar deformity. The highest correlation was between the Nunit and the deformity index (r = 0.59), and once again there was scatter (figure 3). As Nunit and N worsened, there was more pigment. N was also related to fibrosis, there being less fibrosis with decreased attachments. Inflammation was not associated with loss of attachments. Bronchiolar muscle and goblet cell metaplasia were not related to attachments. Correlation coefficients between alveolar attachments and emphysema and tests of pulmonary function are shown in table 3. Nunit showed the highest number of significant correlations, and R the fewest. In general, the correlations are poor and considerable scatter was present, an example (FEV 1/ FVC) being shown in figure 4. Correlation coefficients were also consistently
NS ,.--,
P
