Crackles in Interstitial Lung Disease

Crackles in Interstitial Lung Disease* Comparison of Sarcoidosis and Fibrosing Alveolitis Robert P. Baughman, M.D., F.C.C.P.; Ralph L Shipley, M.D.; Robert G. Loudon. M.B., Ch.B., F.C.C.P.; and Elyse E. Lower, M.D.

Study Objective: Determine why crackles on chest auscultation are characteristic of most interstitial lung diseases, but may not be heard in sarcoidosis. Design: All patients with sarcoidosis or cryptogenic fibrosing alveolitis seen during a four-week period were studied. In a second study to relate ausculatory findings to anatomy, patients with fibrotic changes on their chest roentgenogram were studied. SeUing: Patients were recruited from outpatient clinics. PatienttJ: In the first part, all patients seen over the course of one month were studied. In the second study, patients with pulmonary fibrosis seen on chest roentgenograms were studied. Interventions: For the first study, two independent observers performed auscultation on five sites for crackles and reviewed four roentgenogram quadrants for changes. For the second study, patients underwent VC measurements, auscultation, and high resolution computer tomography scans. Measurements and Results: For the first study, crackles were noted at >2 sites in all 11 CFA patients, but only one of 17 SARC patients (p2 quadrants in nine of 11 CFA patients and

eight of 17 SARC patients (p=ns). In the second study, the VC was similar in the two groups: SARe: 1.96±.90 L (means±SD), 58±20.4 percent predicted; CFA: 1.81±.33 L, 59 ± 9.2 percent predicted). Only two of 14 SARC patients had crackles in > I area, while all 14 CFA patients had crackles at >2 sites. The HRCT studies were read by a radiologist unaware of the diagnosis. The presence and degree (0 to 3 scale) of subpleural and peribronchial fibrosis were scored. Twelve SARC patients had peribronchial changes (mean score 1.9 ± 1.08), while only eight had subpleural fibrosis (mean score .6 ± .52). There was a significantly different pattern in the CFA patients, where eight had peribronchial fibrosis (mean seore = .9 ± .78, p
(rales) are a common physical finding in C rackles patients with interstitial lung disease. The pres-

not seen with sarcoid. The following study was performed to determine the prevalence of crackles in two types of interstitial lung disease (sarcoidosis and CFA) and the interobserver variability for the presence ofcrackles. In addition, the physical findings were compared to HRCT results to determine if a specific roentgenographic pattern was associated with crackles.

1

ence of fine crackles has been considered an early indicator of certain interstitial lung diseases such as asbestosis and cryptogenic fibrosing alveolitis.i-" In a survey of patients with various interstitial lung diseases, Epler et al4 noted differences in the frequency of crackles in patients with sarcoidosis (18 percent) vs patients with other interstitial lung diseases, especially asbestosis and CFA (60 percent)." High resolution computer tomography scans have recently been used to differentiate various interstitial lung diseases including asbestosis, sarcoidosis, and CFA.~ A relatively specific pattern for disease has been reported for sarcoidosis which is distinctly different from the interstitial pattern of CFA. 5,7 Patients with CFA have been found to have subpleural fibrosis, *From the Departments of Medicine and Radiology University of Cincinnati, Ohio. Supported in part by National Institutes of Health grant RR-00068. Manuscript received Septemher 24, revision accepted November 26. Reprint requests: Dr. Baughman, 2.31 Bethesda Avenul'. RT1l 7511, Cincinnati 45267-0.564 96

=

CFA cryptogenic 6brosing alveolitis; HRCT computer tomography; SARe sarcoidosis

=

=high resolution

METHODS

The study was performed in two parts. First, a standard way of observing and reporting crackles was established and tested. Second, this standard reporting technique was used to characterize crackles and compare the sounds heard to HRCf. Patients were considered to have sarcoidosis if they had a compatible clinical history, chest roentgenogram, and biopsy findings demonstrating noncaseating granulomas," Patients with cryptogenic fibrosing alveolitis had a restrictive pattern on pulmonary function testing and a chest roentgenogram consistent with fibrosing alveolitis. Patients with CFA had at least a transbronchial biopsy showing no evidence of granulomatous disease. Patients with underlying collagen vascular diseases were included in the study. Chest auscultation findings were not used as part of the evaluation process. Patients for the first study were selected from outpatients at a chest clinic. Patients received auscultation simultaneously by two Crackles in Interstitial Lung Disease (Baughman et a/)

tnvesngators (R.G.L. and R.P.B.) with at least one investigator not aware of the patient's diagnosis. For the second part of the study, patients with known fibrotic changes on chest roentgenogram and a clinical diagnosis of sarcoidosis or fibrosin~ a1veolitis were recruited to undergo HRCf. Patients who consented to a protocol approved by our Institutional Review Board received auscultation by two or three investigators (R.G .L.. R.P.B., or E .E.L.) within a week of the HRCf. For all patients, chest roentgenogram and pulmonary function studies were obtained within one month of the time of auscultation. No patient was studied during a time of rapid change in lung function . The pulmonary function studies were performed with either a water seal spirometer (Collins) or a pneumotachometer system (Spirotech 300). The forced vital capacity was measured at least three times and the best value recorded . The percent predicted was calculated using Morris's predicted values.!"

Evaluation a/Crackles Chest auscultation was performed in a standard fashion by two or more observers on all patients. Observers were trained to auscultate crackles by the use of a lung sounds tape and by comparing observations on a group of patients prior to initiating the study. The patient was instructed to breathe normal tidal volume breaths; auscultation was performed at five sites : both bases. both apices, and at the mouth . The presence or absence of crackles was noted. If present. crackles were recorded as either fine or coarse, timing was noted, and loudness and profusion were graded on four point scales .

High Resolution Computerized Tomography The HRCf protocol consisted of 1.5-mm thick slices using a Cf scanner with high spatial resolution algorithm and tangential to each lung separately. Twenty-eight patients underwent imaging at six levels each in prone and supine positions. The levels were determined on each patient in relation to the carina. diaphragm , and apex of the lung. Figure I shows the levels examined. Three patients had imaging in the supine position only because of inability to lie prone; two of these patients had imaging at only three levels (top, middle, and bottom of the lungs). The HRCf was scored as follows: subpleural opacities were scored from 0 (none) to 3 (marked) if irregular densities were present within the lung adjacent to the visceral pleura. Peribronchial disease was similarly scored if abnormal opacities were present

FIGURE 1. A posterior-anterior chest roentgenogram with the levels at which the HRCf was performed .

Table I-Comparison of Score Uring a Four-Point Scale for Fibrosis or Crackles Chest Roentgenogram

Total number of observations Amount of disagreement None One Two Three

Crackles

Sarcoid

CFA

Sarcoid

CFA

48

36

85

45

34 Il 2 1

27 9 0 0

80 4 1 0

26 13 3 2

adjacent to bronchovascular bundles. Bronchiectasis was judged present if airways were thickened and dilated or beaded. Honeycombing was judged present if focal lucencies of less than 1 ern diameter with visible walls were present in clusters. To further differentiate bronchiectasis from honeycombing, additional criteria were used . If the lucencies appeared to branch and follow the expected course of bronchovascular bundles, they were judged to be bronchiectasis. If they were predominantly subpleural and nonsegmental, they were judged to be honeycombing. Ground glass opacity was scored if poorly defined confluent airspace opacity was present.

Statistics For comparisons between observers. kappa analys is was used." RESULTS

For the first study, 17 sarcoid patients and 11 patients with cryptogenic fibrosing alveolitis were encountered over the course of one month. No patient refused participation in the study. In 21 patients, chest roentgenograms within one month of exam were available (12 sarcoids, nine CFA). The roentgenograms were reviewed by two chest physicians who graded the four quadrants of the chest roentgenogram for the degree of fibrosis on a four-point scale . Table 1 demonstrates the agreement between the two observers. Each quadrant was scored and a comparison was made between the two observers' scores for each quadrant. It was found that in all but three quadrants (4 percent), there was either perfect agreement (73 percent) or agreement within one point (23 percent). Kappa analysis was done for both chest roentgenogram and auscultation, and there was good agreement between the two observers (roentgenogram: kappa = 0.45; auscultation: kappa = 0.46). If one allowed for one point disagreement between the observers, the kappa analysis showed near perfect agreement between observers (roentgenogram: kappa =0.93; auscultation: kappa = 0.86). A comparison was also made between the two observers on the profusion and loudness of the crackles heard. Table 1 displays the good interobserver agreement on these two scales of crackles, with 82 percent agreement, 13 perent with one point disagreement, and only 5 percent disagreement for two or more points. The amount of interobserver variation was CHEST I 100 I 1 I JULY, 1991

97

Table 2-Crackles or Fibrosis in Two or More Quadrants. Fibrosis

Sarcoidosis Fibrosing alveolitis

Crackles

No . of Patients

Yes

No

Yes

No

17 11

8 9

9 2

1 11

16 0

·Chi square = 24.2, p
similar between chest roentgenogram reading for the presence or absence of fibrosis and the auscultation of the chest for the presence or absence of crackles. In Table 2, CFA was compared to sarcoidosis for the presence or absence of crackles and fibrosis in two or more quadrants of the chest. As can be seen, all patients with CFA had crackles at two or more sites, while only one of 17 sarcoid patients had crackles at two or more sites (Chi square = 24.2, p
sarcoidosis. All patients had fibrotic changes noted on their chest roentgenogram . The vital capacity was similar in the two groups: sarcoid : 1.96± .90 L (mean ± SD), 58 ± 20.4 percent predicted; CFA : 1.81 ± .33 L, 59 ± 9.2 percent predicted. Only two sarcoid patients had crackles at more than one site. All 14 CFA patients had crackles at two or more sites. Table 3 summarizes the quality, timing, profusion , and loudness of crackles heard for each patient. These values were the mean determination of the two or more observers of each patient, and the mean value for the bases are shown. Crackles were heard in the upper lobes of only seven patients (one with sarcoid, six with CFA). In all but one sarcoid patient, the crackles were more prominent in the bases than in the upper lobes. In the 17 patients with crackles, all but two had crackles which were considered fine. Both patients with coarse crackles had CFA. The timing of crackles when present was usually mid (nine patients), late (two patients) or paninspiratory (sixpatients). Although crackles were heard more frequently in patients with CFA, there was no difference in terms of timing or quality of the crackles between patients with sarcoid and CFA.

Bases"

Pt. No .

T

Q

1 2 3 4 5 6 7

0 0 0 0 0 0 L 0 0 0 0 M M 0

0 0 0 0 0 0 F 0 0 0 0 F F 0

0 0 0 0 0 0

.5 1 0

.5 1.5 0

M P P M M P L P M P M P M M

F F F C F C F F F F F F F F

3 2.5 1.5 1 1 3 2 3 2 2 3 2 2 2

3 2.5 2 1.5 2 1.5 2 3 1.5 2 3 2.5 2 2

8 9 10 11 12 13 14 CFA Patients 1 2 3 4 5 6 7

8 9 10 11 12 13 14

*T:timing: E, early; M, mid; L, late ; P, pan. Q : quality: F, fine, C, coarse. L: loudness: Mean of both lungs, scale of 0-3 . P: profusion: mean of both lungs, scale of 0-3 .

98

L

P 0 0 0 0 0 0

.5 0 0 0 0

.5 0 0 0 0

FIGURE 2. The presence of subpleural fibrosis is demonstrated in the HRCT scan of a patient with CFA . Crackles in Interstitial lung Disease (Baughman at eI)

FI G U II ~: 3 0 Pe rihm nch ial

Iihrosis is dem on strat ed in this patien t with

sa rco idos is,

Th e H H T studies were read by a radiologist unaware of th • diagnosis. Th e presence and d gre > (0-3 scale) of subpl ' ural and p eribronch ial fibrosis were scor ' d . Exampl es of subpl ' ural fibrosis and peribronchial fibrosis arc shown in Figures 2 and 3, FI GlIlIE 5 0 Diffuse

honeycombing see n in a patient with sarcoidosis,

respectively. The presence of bronchiectasis (Fig 4) or honeycombing (Fig 5) was also noted. Table 4 shows the HRCf results for each patient. Only one patient (sarcoid patient 3) had no abnormality seen on the HRCT scan. Twelve of 14 sarcoid patients had peribronchial changes (mean score 1.9 ± 1.07), while only eight had subpleural fibrosis (mean score .6 ± .51, p
Fine crackles are an important auscultatory finding in interstitial lung diseases. This study demonstrated CHEST I 100 I 1 I JULY, 1991

99

Table 4-HRCT Results Pt. No.

Subpleural

Sarcoidosis patients 1 0 1 2 3 0 4 1 5 0 6 0 7 1 1 8 9 1 10 1 11 0 12 1 13 1 0 14 .6 Mean CFA Patients 1 1 2 3 2 3 4 1 5 2 1 6 7 1 8 2 1 9 10 2 11 2 1 12 13 1 14 2 Mean 1.6t

Bronchiectasis

Honeycombing

Pattern

3 2 0 0 2 2 3 3 2 2 2 3 3 1 1.9

3 1 0 1 2 2 3 2 1 2 0 3 2 1 1.6

No No No No No No Yes No Yes No No Yes Yes No

Equal Equal Equal Lower Lower Upper Upper Equal Upper Upper Lower Upper Equal Upper

2 1 2 2 0 0 0 1 2 2 0 0 3 0

2 2 2 1 0 0 0 1 1 2 1 0 2 1 1.0

Yes Yes Yes No Yes No No Yes No Yes No Yes Yes Yes

Equal Lower Lower Upper Lower Equal Lower Equal Upper Equal Equal Equal Lower Lower

Peribronchial

.9t

-Differs from subpleural, p
that crackles were usually heard in patients with CFA, but rarely heard in patients with sarcoidosis. Disagreement on physical findings is well established. 12•13 Since there is an interpretive element to lung auscultation, there has always been some difficulty in obtaining consistent information for lung sound research. Murphy et al3 have shown that with training, reasonable agreement can be obtained on the presence and quality of crackles heard on lung auscultation. In our study, using a lung sound tape and interaction on individual cases prior to this study, we were quickly able to arrive at good agreement on lung sound findings. The level of agreement was similar to that obtained upon reading chest roentgenograms. 13 In order to prove the reliability of recording crackles, patients underwent auscultation by two observers. We found agreement on the presence or absence of crackles in 75 percent of the cases. In addition, there was no case in which one observer felt crackles were present in two or more quadrants and another did not agree. This compares with previous studies in which training has been shown to improve interobserver agreement.s-' 100

FIG UIl E 6. Th e marked d ilfer en ce in lum; Involvemen t in th e uppe r lobe and a relatively norm al uppcaruuce of the lower lobe of a patient with sarcoidosi s.

The comparison between chest auscultation and HRCT was undertaken in order to explain the difference in lung sound findings between sarcoidosis and CFA. We specifically chose to study patients with similar degrees of lung disease so that auscultatory differences could not be attributed to degree of disease only. The HRCT has proved a useful technique to characterize parenchymal involvement due to interstitial lung dtsease.>" With this technique, it is possible to differentiate changes in the interstitium from the effects of fluid overload and low lung volumes. The major drawback to the technique is its limited sampling since only a small percentage of the chest is examined. However, both sarcoidosis and fibrosing alveolitis are diffuse diseases, and therefore, the limited views appear to be a good reflection of a wider portion of the chest.F' It should be noted that in patients with fibrosing alveolitis, the disease was in the bases of the chest, as assessed by the HRCT, although two patients had upper lobe predominance. In patients with sarcoidosis, the disease by HRCT was often in the upper part of the chest. This difference in pattern of involvement has been previously observed when viewing the plain chest roentgenogram . 14 Crackles in Inlerstitiallung Disease (Baughman at al)

There are several possible mechanisms .for the difference in the amount of crackles detected in sarcoidosis and CFA, including the following: patients with sarcoidosis have less pulmonary fibrosis' than those with CFA; crackles are associated with honeycombing, which is seen more frequently in CFA than sarcoidosis; and crackles are associated with subpleural fibrosis, which is seen more frequently with CFA. In the second part of this study, we specifically chose sarcoidosis and CFA patients with equally severe disease as assessed by the vital capacity so that the discrepancies in crackles could not be explained by severity of the fibrosis. In addition, patients were selected for study based on the finding of fibrosis on their plain chest roentgenogram. Epler et al," in a previous survey of lung sounds in pulmonary fibrosis, found crackles more frequently in patients with honeycombing. Although honeycombing appeared more frequently in CFA than in sarcoidosis in the current study, it did not correlate with crackles in the CFA group. It has also been shown that crackles may appear several years before any other evidence of pulmonary fibrosis in patients with asbestosis and other pulmonary fibrotic diseases.v' In patients with asbestosis and normal chest roentgenograms, honeycombing was not found on HRCT. 8 Thus, honeycombing alone does not appear to explain the crackles heard. The presence of bronchiectasis, a condition associated with coarse crackles, did not correlate with the fine crackles heard in either group studied. Although several patients had bronchiectasis on HRCT, few appeared to have coarse crackles. In this study, we could not prove that subpleural fibrosis was necessary for crackle production. All patients with crackles showed subpleural fibrosis on HRCT, but some patients with sarcoidosis had subpleural fibrosis but no crackles. Peribronchial fibrosis was seen in both groups, although more frequently in the sarcoid group. There was no association between peribronchial fibrosis and crackles. In addition to the possibilities raised above, there may be changes in the ventilatory pattern of sarcoid vs CFA patients leading to differences in the ability to auscultate crackles. Ventilation is an important element of crackle formation.P'!" Also, the fibrotic changes due to sarcoid may be different from those of CFA due to granulomas found with sarcoid. The

granulomas may add to the roentgenographic changes seen without causing additional fibrosis. Neither of these possibilities was investigated in this study In summary, we found that crackles were always heard in patients with significant cryptogenic fibrosing alveolitis but rarely heard in patients with sarcoidosis. The basis for this difference in auscultation was studied using HRCf. The HRCT scans showed that crackles were associated with lower lobe, subpleural fibrotic changes. REFERENCES

1 Loudon R, Murphy RLH. Lung sounds. Am Rev Respir Dis 1984; 130:663-73 2 Shirai F, Kudoh S, Shibuya A, Sada K, Mikami R. Crackles in asbestos workers: auscultation and lung sound analysis. Br J Dis Chest 1981; 75:386-96 3 Murphy RLH, Gaensler EA, Holford SK, Del Bono EA, Epler G. Crackles in the early detection of asbestosis. Am Rev Hespir Dis 1984; 129:375-79 4 Epler GR, Gaensler EA, Carrington CB. Crackles (rales) in the interstitial pulmonary diseases. Chest 1978; 73:333-39 5 Lynch DA, Webb WR, Gamsu G, Stulbarg M, Golden J. Computed tomography in sarcoidosis. J Comput Assist Tomogr 1989; 13:405-10 6 Muller NL, Staples CA, Miller RR, Veda! S, Thurlbeck WM, Ostrow DN. Disease activity in idiopathic pulmonary fibrosis: CT and pathologic correlation. Radiology 1987; 165:731-34 7 Zerhouni EA, Naidich De Stitik Khouri NF, Siegelman SSe Computed tomography of the pulmonary parenchyma: Part 2. Interstitial disease. J Thorac Imag 1985; 1:54-64 8 Staples CA, Gamsu G, Hau CS, Webb WR. High resolution computed tomography and lung function in asbestos-exposed workers with normal chest radiographs. Am Rev Respir Dis 1989; 139:1502-08 9 Mitchell DN, Scadding JG. Sarcoidosis. Am Rev Respir Dis 1974; 110:774-802 10 Morris JF, Koski A, Johnson Le. Spirometric standards for healthy non-smoking adults. Am Rev Respir Dis 1971; 103:57 11 Fleiss JL. Statistical methods for rates and proportions. 2nd ed. New York: John Wiley and Sons, 1981:213-20 12 Smylie HC, Blendis LM, Armitage I! Observer disagreement in physical signs of respiratory system. Lancet 1965; 2:412 13 Koran LM. The reliability of clincal methods, data, and judgments. N Engl J Med 1975; 293:642 14 Schwan MI. Radiologic recognition of chronic diffuse lung disease. In: Schwan MI, KingTE, eds. Interstitial lung disease. Philadelphia: BC Decker Inc, 1988:27-36 15 Fredberg JJ, Holford SK. Discrete lung sounds: crackles (rales) as stress-relaxation quadrupoles. J Acoust Soc Am 1983; 73:103646 16 Ploysongsang Y, Schonfeld SA. Mechanism of production of crackles after atelectasis during low volume breathing. Am Rev Respir Dis 1982; 126:413-15

n

CHEST I 100 I 1 I JUL~ 1991

101