American Journal of Respiratory and Critical Care Medicine

Inhaled beryllium induces specific sensitization and nonspecific effects leading to chronic beryllium disease (CBD). It is not known whether beryllium induces epithelial cell injury and increases alveolar-capillary leak. We hypothesize that lung injury is an early event in this disease and that markers of lung injury reflect severity of CBD. We measured serum and bronchoalveolar lavage fluid (BALF) KL-6 level, a marker of epithelial cell injury, and BALF/serum albumin, a marker of alveolar-capillary permeability, in 26 patients with CBD, 15 beryllium-sensitized subjects without disease (BeS), and 32 control subjects (Ctrl). We examined the association of these markers, BAL cellularity, pulmonary function, gas exchange, serum angiotensin-converting enzyme, chest radiograph, the effects of glucocorticoid therapy, and clinical course. BALF/serum albumin and serum KL-6 increased in CBD and were discriminative markers for CBD. BALF KL-6 and BALF/serum albumin reflected mainly lung cellular and granulomatous inflammation. Serum KL-6, like BALF KL-6, was associated with permeability change and reflected functional and radiologic abnormalities. Serum KL-6 detected early lung injury in BeS. Epithelial injury and permeability changes occur early in CBD, indicating disease severity. Monitoring of these events with serum KL-6 may be useful for management of CBD.

Chronic beryllium disease (CBD) is a granulomatous lung disorder in which inhaled beryllium induces beryllium-specific T-cell-mediated sensitization. In CBD, sensitization precedes granuloma formation (1). In addition, beryllium may trigger lung inflammation by more nonspecific mechanisms of toxicity, beginning at the point of contact with pulmonary epithelium (2). However, the direct effects of this highly reactive bivalent ion have not been examined extensively in humans. Animal models have suggested that beryllium induces epithelial cell injury and increases alveolar-capillary leak (3), but it is not known whether such effects occur in humans or whether they promote chronic inflammation.

In interstitial lung disease, lung injury increases alveolar-capillary permeability (4). Lymphocytes and granulocytes accumulate in the inflammatory lesion, migrating from the blood vessels as a consequence of increased permeability, even though there may be no detectable structural changes in endothelial cells (5, 6). Several methods have been used to evaluate alveolar-capillary permeability in vivo and in vitro, including using bronchoalveolar lavage fluid (BALF) albumin, BALF total protein, BALF/serum albumin ratio, BALF albumin/urea ratio, 99mTc-diethylenetriamine pentaacetate, and 125I-bovine serum albumin. Such markers provide a relative index of the degree to which lung permeability contributes to inflammatory cell and secretory product transit from the blood to sites of pulmonary inflammation.

The mucinlike molecule, KL-6, is expressed on alveolar type II cells and bronchiolar epithelial cells in human lungs. KL-6 is also detected in alveolar macrophages by immunohistochemistry (7). Because of its high expression on injured, regenerating lung epithelium and absence from interstitial cells, KL-6 may be a useful marker of pulmonary epithelial cell injury (8-10). Serum KL-6 level reflects the severity of interstitial lung diseases, such as idiopathic pulmonary fibrosis, collagen vascular disease, hypersensitivity pneumonitis, sarcoidosis, and radiation pneumonitis (7, 9, 11-13). Serum KL-6 also reflects lung injury in patients with pulmonary tuberculosis (14), thus making it a likely candidate marker for monitoring lung injury in granulomatous disorders.

We hypothesized that CBD is the result of an interaction of beryllium-specific immunity and nonspecific lung injury and that direct epithelial cell injury is an early event in disease pathogenesis. Furthermore, we hypothesized that by monitoring two markers in the lung and blood, one that reflects epithelial cell injury (KL-6) and the other changes in permeability (albumin), we could detect and track the degree of lung injury in CBD.

Study Population

The study population included 73 subjects: 26 with CBD, 15 beryllium-sensitized but without lung disease (BeS) (including six patients who progressed to CBD within 2 yr), and 32 control subjects (Ctrl) as described in Table 1. These subjects were enrolled as part of a study supported by a Specialized Center of Research Grant from the National Institutes of Health. Informed consent was obtained from each subject, and the protocol was approved by our Institution's Human Subjects Review Board.


SubjectsCBD (n = 26 )BeS (n = 15 )Ctrl (n = 32)
Sex, M/F20/612/316/16
 Caucasian non-Hispanic191228
 Caucasian Hispanic 4 3 0
 African American 3 0 1
 Asian American 0 0 3
Median age (25%–75%), yr50 (43–60)50 (46–75)39 (31–43)*
 Current smoker 3 5 3
 Former smoker11 7 2
 Never smoker12 327
Corticosteroid therapy 6 1  0

Definition of abbreviations: CBD = subjects with chronic beryllium disease; BeS = beryllium-sensitized subjects without disease; Ctrl = control subjects.

*p < 0.05 versus Ctrl and BeS.

For non-beryllium-related skin rash.

Subjects with CBD. We obtained serum from 25 subjects with CBD, and BALF from 25 with CBD, with serum and BAL fluid data missing from one patient for each. The diagnosis of CBD was based on the following criteria (1): (1) history of beryllium exposure; (2) positive BAL beryllium lymphocyte proliferation test; (3) transbronchial lung biopsy yielding pathologic changes consistent with CBD, i.e., noncaseating granulomas and/or mononuclear cell infiltrates; (4) respiratory symptoms, pulmonary function test abnormalities, or chest radiographic small opacities profusion score greater than or equal to 1/0 by the International Labour Organization classification system (15). In addition, none of the subjects had other detectable causes of granulomatous inflammation, and all had negative cultures and special stains for acid-fast bacilli, fungi, and bacteria. Glucocorticoid effect was studied in nine subjects, six with CBD and three BeS who progressed to CBD. The average glucocorticoid oral dose in these nine subjects was 20 to 30 mg /d, with a duration of treatment of 6 to 30 mo.

Beryllium-sensitized subjects without lung disease (BeS). Thirteen serum and 15 BALF samples were obtained from 15 BeS subjects. Subjects included in this category were required to meet the following criteria: (1) a history of beryllium exposure; (2) two or more abnormal blood beryllium lymphocyte proliferation tests; (3) negative BAL beryllium lymphocyte proliferation test; (4) no pathologic evidence of CBD on transbronchial lung biopsy (1). The clinical assessment of all BeS subjects fell within the normal range. Of these subjects, we measured serum (n = 4), BALF (n = 6), KL-6, and albumin levels both at the time of enrollment (Time 1) and at a second sampling time 1 to 2 yr later (Time 2). We compared the baseline (Time 1) and the changes (Time 2 − Time 1) of serum and BALF KL-6 and albumin levels in the subjects with BeS who developed CBD at Time 2 (BeS-CBD, n = 6) and the subjects with BeS who remained sensitized without disease at Time 2 (BeS-BeS, n = 7).

Control subjects. Serum was obtained from 32 Ctrl and BALF from 11 Ctrl. The 32 Ctrl included 21 healthy hospital workers and community volunteers and 11 subjects who underwent bronchoscopy without evidence of either BeS, CBD, or other lung disease on the basis of blood, BAL, and biopsy results.

Sample Collection and Preparation

BAL was performed by previously published methods (1) using four 60-ml aliquots of saline instilled in a subsegment of the right middle lobe. The cellular and BALF components were separated by centrifugation and stored at −70° C until use. There was no significant difference in recovered volume of BAL among the groups (Table 2). We analyzed neat concentration of KL-6 and albumin in recovered fluid in this study, as previously reported (16). Total cell counts and cell differentials were obtained by hemocytometer and Diff-Quik staining, as described previously (17, 18). Serum was obtained from study subjects on the same day as BAL and stored at −70° C until use.


CBD (n = 26)BeS (n = 15)Ctrl (n = 11)
Recovery of BAL, %62  (54–69)59  (48–78)66  (65–70)
WBC (× 105)/ml BALF 5.87(3.95–8.19) 2.63(1.35–7.33) 1.39(1.10–1.69)*
% Lymphocytes56.5  (35–72)10   (5–22)* 15  (11–27)*
% Macrophages43  (28–64)89  (76–92)* 84  (72–88)*
% Neutrophils 0   (0–1) 1   (0–2) 1   (0–2)
% Eosinophils 0   (0–0) 0   (0–1) 0   (0–0)
BALF albumin, μg/ml65.8(32.4–95.3)19.5(14.2–40.4)* 18.3 (9.7–25.3)*
Serum albumin, g/dl 2.79(1.19–5.11) 3.36(2.91–3.92) 3.40(2.37–3.94)

For definition of abbreviations, see Table 1.

*p < 0.05 versus CBD.

Measurement of KL-6 in Serum and BALF

We measured KL-6 levels in serum and BALF by the modified method reported previously (7) using a sandwich-type ELISA kit (ED046; Eisai Co. Ltd., Tokyo, Japan). Samples were evaluated in duplicate.

Alveolar-Capillary Permeability

Alveolar-capillary permeability was estimated by calculating the BALF/serum albumin concentration ratio (19). BALF and serum were diluted 1:10 and 1:10,000 in sterile saline, respectively, and albumin concentrations were measured in duplicate by ELISA (Albuwell®; Exocell, Inc., Philadelphia, PA).

Measurement of Serum Angiotensin-converting Enzyme Activity

Serum angiotensin-converting enzyme (ACE) activity was measured by colorimetric assay as described previously (20).

Clinical Assessment

Pulmonary physiology. Lung function was assessed in all CBD and BeS subjects. We measured FVC and FEV1 with a recording spirometer or a pneumotachograph, reporting maximal values obtained from three satisfactory maneuvers. Normal predicted values were derived from the work of Morris and coworkers (21). We measured TLC in a constant-pressure body plethysmograph using the predicted normal values of Goldman and Becklake (22). We used the single-breath method of Ogilvie and coworkers (23) to determine the diffusing capacity for carbon monoxide (Dl CO). Dl CO data were presented as the percentage of predicted normal values of Crapo and Morris (24). We evaluated these subjects' maximal exercise capacity and gas exchange during exercise using a Siemens Elema 380B cycle ergometer (Siemens Elema, Solna, Sweden), with continuous monitoring of cardiac rhythm and oxygen saturation (Hewlett-Packard, Waltham, MA). A mass spectrometer (1100 medical gas analyzer; Perkin-Elmer Medical Instruments, Pomona, CA) was used to measure inspired air and expired oxygen and carbon dioxide concentrations. Using an ABL-2 blood gas analyzer (Radiometer, Copenhagen, Denmark), we measured arterial blood gases at rest and after each minute of graded exercise, obtained through an indwelling arterial catheter in the radial or brachial artery.

Radiology. We scored the radiographic severity of interstitial lung disease using standard posteroanterior chest radiographs according to the International Labour Organization classification system (15). All films were scored by a certified B-reader who was blind to diagnosis. Normal profusion of small opacities was defined as a profusion score less than or equal to 0/1. Data were analyzed on a 0- to 10-point ordinal scale in which profusion ranks 0 /- and 0/0 were combined.

Statistical Analysis

We used Dunn's non-parametric multiple-comparisons procedure after Kruskal-Wallis test or Wilcoxon's non-parametric rank sum procedure for ranked data to compare differences between all or between two groups (25). We assessed associations using Spearman's correlation coefficient (rho). All data are expressed as medians with 25%– 75% interquartile ranges reported in parentheses. We required a p value < 0.05 for statistical significance.

Serum and BALF KL-6 and Albumin Concentrations

As shown in Figures 1 and 2, serum and BALF KL-6 levels in Ctrl were 281.5 (169.5–330.5) and 172 (102–258) U/ml, respectively. In the Ctrl, we observed no significant differences based on sex, age, smoking status, or race. The median serum KL-6 level in CBD [1,002 (436.5–1,741.5) U/ml] was significantly elevated compared with Ctrl or BeS [402 (335–699.5) U/ml] (p < 0.05) (Figure 1). We observed no differences in serum KL-6 between BeS and Ctrl. There were no significant differences in BALF KL-6 between CBD [360 (143–764.5) U/ml], Ctrl [172 (102–258) U/ml], and BeS [210 (159–404) U/ml] (p = 0.159) (Figure 2). There was a statistically significant correlation between serum KL-6 and BALF KL-6 (rho = 0.45, p  < 0.001, n = 52).

As shown in Table 2, the BALF albumin concentration was significantly higher in subjects with CBD than in Ctrl or in BeS (p < 0.05). There were no significant differences in serum albumin concentration between the groups. The BALF/serum albumin ratio was significantly higher in subjects with CBD than in either Ctrl or BeS (p < 0.05) (Figure 3).

The BALF/serum KL-6 ratio was significantly lower in subjects with CBD than in Ctrl (p < 0.05) (Figure 4). However, when we corrected BALF KL-6 by BALF albumin, we found no significant difference between the groups: Ctrl [11.3 (5.6– 17.7)], BeS [7.1 (5.6–15.3)], and CBD [6.4 (3.7–8.8)] U/μg albumin.

To assess the relationship between albumin leak from the blood compartment to the alveolar space and conversely the KL-6 leak from alveolar space to blood, we calculated BALF/ serum albumin, serum KL-6, BALF KL-6, and BALF/serum KL-6 ratios (Table 3). We observed significant correlation between serum KL-6 concentration and BALF/serum albumin ratio in CBD alone and in a combined (CBD + BeS + Ctrl) group, as well as between BALF KL-6 and BALF serum albumin in this combined group. We noted a trend toward negative correlation between BALF/serum albumin ratio and BALF/ serum KL-6 ratio in the combined group.


BALF/Serum Albumin
CBD (n = 20)BeS (n = 12)Ctrl (n = 10)CBD + BeS + Ctrl (n = 42)
rhop Valuerhop Valuerhop Valuerhop Value
Serum KL-6    0.460.041* 0.500.0950.390.2600.60< 0.001*
BALF KL-60.380.0950.280.379−0.190.6030.320.036*
BALF/Serum KL-6−0.080.7400.060.829−0.320.366−0.290.064

For definition of abbreviations, see Table 1.

*p < 0.05.

Correlation between KL-6 and Clinical Measures

To examine the relationship between KL-6 levels and the severity of lung injury, we correlated serum and BALF KL-6 with the clinical variables shown in Table 4. BALF variables provide an estimate of the degree of lung inflammation (26). As shown in Table 2, the white blood cell count (WBC) /ml BALF was significantly higher in CBD than in Ctrl, with higher percent lymphocytes and lower percent macrophages in CBD than in Ctrl or BeS, consistent with previous reports (26, 27). In subjects with CBD, we observed significant positive correlations between BALF KL-6 levels and WBC number/ml BALF, as well as with percent lymphocytes. We observed negative correlations between BALF KL-6 levels and percent macrophages in BAL. Measurements of lung volumes, air flow, and gas exchange are indicative of CBD severity (28). Serum KL-6 levels were inversely associated with TLC percent predicted (%TLC), %FVC, %FEV1, %Dl CO percent predicted, PaO2 at rest, and PaO2 at maximal exercise. Serum KL-6 was associated positively with aaPo 2 at rest, aaPo 2 at maximal exercise, and radiograph profusion score. Similarly, BALF KL-6 was negatively associated with %FEV1 and PaO2 at maximal exercise, and positively associated with serum ACE, aaPo 2 at maximal exercise, and radiograph profusion score. In the BeS group, serum KL-6 level was negatively associated with %FVC, %FEV1, and PaO2 at maximal exercise, even though the actual values in this group of subjects were within the normal range.


Serum KL-6 (n = 25)BALF KL-6 (n = 25)Serum KL-6 (n = 13)BALF KL-6 (n = 15)
rhop Valuerhop Valuerhop Valuerhop Value
Bronchoalveolar lavage
 WBC (× 105)/ml0.39  0.0520.64< 0.001* 0.460.1120.280.315
 % Macrophage−0.25  0.220−0.67< 0.001* −0.400.180−0.170.556
 % Lymphocyte0.33  0.129    0.48  0.023* 0.410.2090.220.475
Serum ACE0.33  0.129    0.48  0.023* 0.410.2090.220.475
Pulmonary physiology, %
 TLC−0.63< 0.001* −0.27   0.188−0.400.176−0.350.198
 FVC−0.53  0.007* −0.31   0.136−0.660.014* −0.440.098
 FEV1 −0.62  0.001* −0.48  0.015* −0.550.050* −0.370.182
 Dl CO −0.69  0.001* −0.22   0.321−0.320.288−0.350.238
Rest and exercise gas exchange
 PaO2 at rest−0.51  0.011* −0.15   0.492−0.500.086−0.140.639
 PaO2 at max exercise−0.65  0.001* −0.50  0.017* −0.630.028* −0.310.283
aaPo 2 at rest0.42  0.042* 0.21   0.3240.300.3170.140.639
aaPo 2 at max exercise0.64  0.001* 0.58  0.005* 0.560.0600.370.188
Radiograph profusion score0.60  0.002* 0.57   0.003* 0.240.4380.380.167

*p < 0.05.

Correlation between BALF/Serum Albumin and Clinical Measures

As shown in Table 5, in CBD, BALF/serum albumin was significantly associated only with WBC number/ml BALF. In the combined CBD + BeS group, BALF/serum albumin was associated positively with WBC number/ml BALF, percent lymphocytes in BALF, serum ACE activity, aaPo 2 at maximal exercise, and negatively with percent macrophage in BALF and with %Dl CO.


BALF/Serum Albumin
CBD (n = 20)BeS (n = 12)CBD + BeS (n = 32)
rhop Valuerhop Valuerhop Value
Bronchoalveolar lavage
 WBC (× 105)/ml 0.630.003* −0.170.6020.400.022*
 % Macrophage−0.13   0.593−0.230.467−0.510.003*
 % Lymphocyte0.13   0.5780.020.9400.490.004*
Serum ACE0.38   0.1380.150.6750.410.033*
Pulmonary physiology, %
 TLC −0.15   0.535−0.330.296−0.30   0.097
 FVC−0.02   0.927−0.530.080−0.29   0.113
 FEV1 −0.21   0.363−0.400.208−0.30   0.097
 Dl CO −0.26   0.2680.220.485−0.410.019*
Rest and exercise gas exchange
 PaO2 at rest0.27   0.260−0.180.569−0.09   0.647
 PaO2 at max exercise−0.14   0.5790.140.688−0.31   0.104
aaPo 2 at rest−0.45   0.700
aaPo 2 at max exercise0.23   0.384−0.020.9570.400.040*
Radiograph profusion score0.01   0.950.270.4000.29   0.103

For definiton of abbreviations, see Table 1.

*p < 0.05.

Effect of Glucocorticoid Therapy

In CBD, the median serum KL-6 levels before and during glucocorticoid therapy were 1,437 (426–3,397) and 866.5 (536– 1,349) U/ml, respectively (n = 8) p = NS). The median BALF KL-6 levels before and during therapy were 367 (91–934) and 330 (192–574) U/ml, respectively (n = 9) (p = NS). Similarly, the BALF/serum albumin ratio did not change significantly during the course of therapy.

KL-6 and Albumin in the Progression from BeS to CBD

As shown in Table 6, baseline serum KL-6 levels were significantly higher in BeS-CBD than in BeS-BeS (p < 0.05). There was no significant difference between baseline BALF KL-6 levels in those two groups. Baseline BALF/serum albumin tended to be higher in BeS-CBD than in BeS-BeS (p = 0.07). Serum KL-6 levels did not rise significantly higher during the progression from BeS to CBD compared with BeS-BeS. However, BALF KL-6 levels rose significantly over time in BeS-CBD compared with BeS-BeS (p < 0.05). We observed no significant difference in the change of BALF/serum albumin ratio between the two groups.


(BeS-BeS) Unchanged BeS(BeS-CBD) Progression to Disease
Median(25%–75%)Median(25%–75%)p Value*
Baseline KL-6
 Serum366     (244–445) 726     (433–1094)     0.047
 BALF165    (139–404) 204(140–545)§ 0.610
Baseline BALF/serum albumin (× 10−4)      5.6  (4.2–1.2)§      15.3(11.6–21.9) 0.070
Change of KL-6
 Serum46      (1–135) 150 (−7–409) 0.500
 BALF−84( −173– −34) 150(−69–397)§        0.45
Change of BALF/serum albumin (× 10−4)       3  (6.3–9.8)§        −0.1(−7.6–19.1)        0.850

For definition of abbreviations, see Table 1.

*Wilcoxon's rank-sum test.

p < 0.05.

n = 4.

§n = 6.

  n = 7.

Our data demonstrate that levels of BALF albumin and of serum KL-6 are elevated in CBD and can distinguish patients with the disease from those with beryllium-sensitization alone and from normal control subjects. These data lead us to conclude that both pulmonary epithelial cell injury, as reflected by KL-6, and increased alveolar-capillary permeability, as reflected by the BALF/serum albumin ratio, are important events in the lungs of patients with CBD. Interestingly, both BALF KL-6 and BALF albumin levels reflected the cellularity and degree of granulomatous inflammation in the lungs of patients with CBD, whereas serum KL-6 correlated most strongly with measures of the physiologic and radiologic severity of disease. Serum KL-6 proved superior to BALF KL-6 in distinguishing CBD from BeS and from Ctrl. Taken in aggregate, these findings suggest that measurements of mucinlike molecules in the peripheral blood may reflect not only epithelial cell injury in the lung but also the degree of leakage from the alveolus to the capillary. Thus, as a surrogate measure of two important early events in CBD pathogenesis, serum KL-6 may become a useful prognostic indicator of BeS progression to disease.

There are several possible explanations for the greater sensitivity of serum KL-6 compared with BALF KL-6. KL-6 in BALF may reflect the production and release by injured epithelial cells within the lung compartment. Even in normal epithelial lining fluid, KL-6 levels are usually more than 200 times greater than in serum (16). In order to be able to detect KL-6 in the peripheral blood, leakage of the 2,000 kD mucinlike molecule must occur as a consequence of increased alveolar-capillary permeability. Our data showing increase of BALF/ serum albumin ratio and a correlation between BALF/serum albumin ratio and serum KL-6 indicate that such permeability changes do occur in CBD. The high BALF/serum albumin ratio that we observed in CBD is consistent with studies of other interstitial lung diseases and with animal beryllium inhalation models showing permeability changes (3, 29). The differences between serum and BALF KL-6 may also be explained, in part, by the dilutional effect of lavage relative to serum. Furthermore, baseline KL-6 levels in epithelial lining fluid are so high that a further increase would not be detected easily. Our data do not directly demonstrate whether pulmonary epithelium is actively producing KL-6 in CBD, or if it is simply releasing it because of cell toxicity. However, these data indicate that the net effect is to raise KL-6 concentrations in serum, making this a potentially useful surrogate measure of lung injury and leak in granulomatous inflammatory disorders.

BALF KL-6 reflects both BAL cellularity and serum ACE activity, whereas serum KL-6 does not. Analogously, the BALF/serum albumin ratio was associated with BAL WBC, macrophage and lymphocyte percent, and serum ACE, although the correlations were weaker than for the KL-6 measure. Serum ACE is considered a marker of the total body burden of granulomatous inflammation in sarcoidosis and in CBD (20, 30-33). In CBD, granulomatous inflammation is confined predominantly to the lungs. Thus, it is logical to expect BALF KL-6 to offer a more direct reflection of intrapulmonary events such as inflammatory cell accumulation compared with serum KL-6. Our data are consistent with previous studies in which BALF KL-6 correlated with both WBC and lymphocyte number in BAL from patients with idiopathic pulmonary fibrosis, sarcoidosis, and hypersensitivity pneumonitis (16). Kobayashi and colleagues reported that serum KL-6 was not associated with BAL lymphocytosis, although they found an association with CD4+/CD8+ cell ratios in BAL in sarcoidosis. Our study did not examine CD4+/CD8+ ratios; however, previous work suggests that the majority (> 80%) of BAL lymphocytes in CBD are CD4+ (26, 27, 34). Our observed correlations between KL-6 and lymphocytes in CBD does not necessarily mean that these cells produce KL-6, but simply reflects the degree of pulmonary inflammation. Previous studies have not found lymphocytes to stain with anti-KL-6 antibodies (7, 9).

Although BALF KL-6 provides the better estimate of lung inflammation, serum KL-6 levels reflect better the degree of gas exchange abnormality and lung restriction in CBD. Serum KL-6 correlated inversely with the degree of gas exchange derangement at rest and at maximal exercise, with Dl CO, and with FVC and TLC. We found much weaker associations using BALF KL-6. This set of results is similar to our previous observation that serum KL-6 is correlated inversely with lung restriction in pulmonary tuberculosis (14). Our data suggest that the physiologic severity of CBD may be linked more directly to the degree of epithelial cell injury and alveolar-capillary leak than to the amount of granulomatous involvement in the lung interstitium. As in previous studies of sarcoidosis and tuberculosis (13, 14), we observed that serum and BALF KL-6 correlate significantly with the profusion of small opacities on chest radiographs in CBD.

Interestingly, in BeS subjects who, by definition, have had negative lung biopsies, no demonstrable BAL cellular infiltration, and normal pulmonary function, we found significant inverse correlations between serum KL-6 and FVC and FEV1, and PaO2 at maximal exercise. Although we carefully classified the study subjects, we cannot exclude the possibility that some BeS subjects were at an early subclinical stage of CBD. In fact, these findings suggest that KL-6 detects subclinical lung injury caused by beryllium earlier than do conventional clinical tests such as BAL beryllium lymphocyte proliferation test, BAL cellularity, chest radiograph, pulmonary function tests, or transbronchial biopsy. We speculate that this is because of beryllium's direct toxic effects on the alveolar epithelium. Future studies should address whether persons who have been exposed to beryllium but who are not sensitized have elevated KL-6 levels. The implication of such a finding would be that beryllium nonspecifically induces lung injury and inflammation that in some instances may be followed by epithelial repair but in other circumstances may help promote chronic lung injury and the cell-mediated immune responses to beryllium that lead to granulomas and CBD.

Although the subject number was small, we found that the subjects with BeS in whom the serum KL-6 was higher at Time 1 tended to progress into CBD (BeS-CBD) and to have increases in BALF/serum albumin compared with those who did not progress from sensitization to disease (BeS-BeS). Prognostically, this suggests that in patients who progress from BeS to CBD, measurement of KL-6 and/or permeability may help to detect disease better than the more conventional tests. Future studies employing larger numbers of subjects will be needed to determine the merits of KL-6 as a prognostic marker and measure of disease progression.

We found no significant changes in serum or BALF KL-6 levels or in the BALF/serum albumin ratio before and after glucocorticoid therapy. However, the number of subjects available for this aspect of the study was small and the degree of clinical response to the steroids was highly variable. These limitations preclude us from drawing firm conclusions concerning the effect of corticosteroids on KL-6, albumin, epithelial cell repair, or alveolar-capillary leak.

This study is the first to examine KL-6 in the U.S. population. Previous analyses of serum and BALF KL-6 have been conducted in Japanese subjects only (7, 16). Our study of 66 Caucasians, four African Americans, and three Asian Americans demonstrates that the levels are similar across races. We observed no significant differences in serum or in BALF KL-6 by sex, age, or smoking status.

In conclusion, significant epithelial cell injury and increases in alveolar-capillary permeability form part of the spectrum of physiologic alterations in CBD. Future research should focus on how these nonspecific alterations at the alveolar level interact with beryllium-specific cell-mediated immunity in promoting granulomatous disease. Serum KL-6 appears to be a promising marker for distinguishing between disease and beryllium-sensitization without disease, and may help identify those sensitized patients at highest risk of progression to CBD.

The writers thank Ronald Balkissoon, M.D., D.I.H., for his assistance in clinical management of the patients, Sally Tinkle, Ph.D., for helpful discussions, Mary Solida, R.N., for patient care, Gerald S. Davis, M.D., for helpful discussion regarding albumin measurement, and Nina Rice for secretarial assistance.

Supported by U.S. Public Health Service Grant R29 ES-04843, Specialized Center of Research Grant HL-27353, and General Clinical Research Center Grant M01 RR-00051 from the National Institutes of Health.

1. Newman L. S., Kreiss K., King T. E., Seay S., Campbell P. A.Pathologic and immunologic alterations in early stages of beryllium disease: reexamination of disease definition and natural history. Am. Rev. Respir. Dis.139198914791486
2. Freiman D. G., Hardy H. L.Beryllium disease: the relation of pulmonary pathology to clinical course and prognosis based on a study of 130 cases from the U.S. Beryllium Case Registry. Hum. Pathol.119702544
3. Sendelbach L. E., Witschi H. P.Bronchoalveolar lavage in rats and mice following beryllium sulfate inhalation. Toxicol. Appl. Pharmacol.901987322329
4. Roberts C. M., Cairns D., Bryant D. H., Burke W. M., Yeates M., Blake H., Penny R., Shelley L., Zaunders J. J., Breit S. N.Changes in epithelial lining fluid albumin associated with smoking and interstitial lung disease. Eur. Respir. J.61993110115
5. Meyrick B., Perkett E. A., Harris T. R., Brigham K. L.Correlation of permeability with the structure of the endothelial layer of pulmonary artery intimal explants. Fed. Proc.46198725162520
6. Damle N. K., Doyle L. V.Ability of human T lymphocytes to adhere to vascular endothelial cells and to augment endothelial permeability to macromolecules is linked to their state of post-thymic maturation. J. Immunol.144199012331240
7. Kohno N., Kyoizumi S., Awaya Y., Fukuhara H., Yamakido M., Akiyama M.New serum indicator of interstitial pneumonitis activity: sialylated carbohydrate antigen KL-6. Chest9619896873
8. Kohno N., Akiyama M., Kyoizumi S., Hakoda M., Kobuke K., Yamakido M.Detection of soluble tumor-associated antigens in sera and effusions using novel monoclonal antibodies, KL-3, and KL-6, against lung adenocarcinoma. Jpn. J. Clin. Oncol.181988203216
9. Hamada H., Kohno N., Akiyama M., Hiwada K.Monitoring of serum KL-6 antigen in a patient with radiation pneumonia. Chest1011992858860
10. Kohno N., Inoue Y., Hamada H., Fujioka S., Fujino S., Yokoyama A., Hiwada K., Ueda N., Akiyama M.Difference in sero-diagnostic values among KL-6-associated mucins classified as cluster 9. Int. J. Cancer819948183
11. Kohno N., Hamada H., Fujioka S., Hiwada K., Yamakido M., Akiyama M.Circulating antigen KL-6 and lactate dehydrogenase for monitoring irradiated patients with lung cancer. Chest1021992117122
12. Kobayashi J., Kitamura S.KL-6: a serum marker for interstitial pneumonia. Chest1081995311315
13. Kobayashi J., Kitamura S.Serum KL-6 for the evaluation of active pneumonitis in pulmonary sarcoidosis. Chest109199612761282
14. Inoue Y., Nishimura K., Shiode M., Akutsu H., Hamada H., Fujioka S., Fjuino S., Yokoyama A., Kohno N., Hiwada K.Evaluation of serum KL-6 levels in patients with pulmonary tuberculosis. Tuber. Lung Dis.761995230233
15. International Labour Organization. 1980. Guidelines for the Use of ILO International Classification of Radiographs of Pneumoconioses. In Occupational Safety and Health Series, No. 22 REV. International Labour Office, Geneva, Switzerland.
16. Kohno N., Awaya Y., Oyama T., Yamakido M., Akiyama M., Inoue Y., Yokoyama A., Hamada H., Fujioka S., Hiwada K.KL-6, a mucin-like glycoprotein in bronchoalveolar lavage fluid from patients with interstitial lung disease. Am. Rev. Respir. Dis.1481993637642
17. Willcox M., Kervitsky A., Watters L. C., King T. E.Quantification of cells recovered by bronchoalveolar lavage: comparison of cytocentrifuge preparations with the filter method. Am. Rev. Respir. Dis.13819887480
18. King, T. E., Jr. 1992. The handling and analysis of bronchoalveolar lavage specimen. In R. P. Baughman, editor. Bronchoalveolar Lavage. Mosby-Yearbook, St. Louis. 3–25.
19. Dusser D., Mordelet-Dambrine M., Collignon M. A., Barritault L., Chretien J., Huchon G.Respiratory permeability studies by clearance of an aerosol solute and bronchoalveolar lavage in interstitial pneumopathies. Bull. Eur. Physiopathol. Respir.201984223227
20. Newman L. S., Orton R., Kreiss K.Serum angiotensin-converting enzyme activity in chronic beryllium disease. Am. Rev. Respir. Dis.14619923942
21. Morris, A. H., R. E. Kanner, R. O. Crapo, and R. M. Gardner, editors. 1984. Clinical Pulmonary Function Testing: A Manual of Uniform Laboratory Procedures, 2nd ed. Intermountain Thoracic Society, Salt Lake City, UT. 154–155.
22. Goldman H. I., Becklake M. R.Respiratory function tests: normal values at medium altitudes and the prediction of normal results. Am. Rev. Tuberc.791959457
23. Ogilvie C. M., Forster R. E., Blakemore W. S., Morton J. W.A standardized breathholding technique for the clinical measurement of the diffusing capacity of the lung for carbon monoxide. J. Clin. Invest.361956117
24. Crapo R. O., Morris A. H.Standardized single breath normal values for carbon monoxide diffusing capacity. Am. Rev. Respir. Dis.1231981185189
25. Daniel, W. W. 1990. Applied Non-parametric Statistics, 2nd ed. PWS-Kent Publishing Company, Boston. 240–242.
26. Newman L. S., Bobka C., Schumacher B., Daniloff E., Zhen B., Mroz M. M., King T. E.Compartmentalized immune response reflects clinical severity of beryllium disease. Am. J. Respir. Crit. Care Med.1501994135142
27. Rossman M. D., Kern J. A., Elias J. A., Cullen M. R., Epstein P. E., Preuss O. P., Markham T. N., Daniele R. P.Proliferative response of bronchoalveolar lymphocytes to beryllium: a test for chronic beryllium disease. Ann. Intern. Med.1081988687693
28. Pappas G. P., Newman L. S.Early pulmonary physiologic abnormalities in beryllium disease. Am. Rev. Respir. Dis.1481993661666
29. Sendelbach L. E., Tryka A. F., Witschi H.Progressive lung injury over one-year period after a single inhalation exposure to beryllium sulfate. Am. Rev. Respir. Dis.139198910031009
30. Lieberman J.Elevation of serum angiotensin-converting-enzyme (ACE) level in sarcoidosis. Am. J. Med.591975365372
31. Thompson P. J., Kemp M. W., McAllister W. A. C., Al S. M. H., Habet, Turner-Warwick M. E. H.Angiotensin-converting enzyme. Am. Rev. Respir. Dis.134198610751077
32. Hiwada K., Inoue Y., Takada V., Hashimoto A., Akutsu H., Kitatani F., Kokubu T.Direct radioimmunoassay of angiotensin-converting enzyme in sera from patients with pulmonary diseases. Lung16519872735
33. Inoue Y.Direct radioimmunoassay of angiotensin-converting enzyme in serum and bronchoalveolar lavage fluid from patients with pulmonary diseases. Nippon Kyobu Shikkan Gakkai Zasshi261988129138
34. Saltini C., Winestock K., Kirby M., Pinkston P., Crystal R. G.Maintenance of alveolitis in patients with chronic beryllium disease by beryllium-specific helper T cells. N. Engl. J. Med.320198911031109
Correspondence and requests for reprints should be addressed to Lee S. Newman, M.D., M.A., F.C.C.P., National Jewish Medical and Research Center, 1400 Jackson Street, Room G010, Denver, CO 80206.


No related items
American Journal of Respiratory and Critical Care Medicine

Click to see any corrections or updates and to confirm this is the authentic version of record