Serum Surfactant Proteins A and D as Prognostic ... - ATS Journals

Serum Surfactant Proteins A and D as Prognostic Factors in Idiopathic Pulmonary Fibrosis and Their Relationship to Disease Extent HIROKI TAKAHASHI, TAKUYA FUJISHIMA, HIROYUKI KOBA, SEIJI MURAKAMI, KEIZO KUROKAWA, YOSHIE SHIBUYA, MASANORI SHIRATORI, YOSHIO KUROKI, and SHOSAKU ABE Third Department of Internal Medicine and Department of Biochemistry, Sapporo Medical University School of Medicine, Sapporo, Japan

Idiopathic pulmonary fibrosis (IPF) is a progressive, life-threatening, interstitial lung disease of unknown etiology. For optimal therapeutic management of IPF an accurate tool is required for discrimination between reversible and irreversible types of the disease. However, such noninvasive tools are few, and even with high-resolution computed tomography (HRCT), which is the most trusted method for doing so, the nature of the disease activity in IPF cannot always be accurately predicted. The aims of the present study were to assess the values of surfactant protein (SP)-A and SP-D in semiquantifying the extent of disease in IPF and in predicting deterioration in restrictive pulmonary function and survival over a follow-up period of 3-yr. SP-A and SP-D in sera were measured with enzyme-linked immunosorbent assays as previously described. Fifty-two IPF patients were studied to evaluate the association between serum SP-A and SP-D and disease extent on HRCT, deterioration in pulmonary function, and survival during 3 yr of follow-up. Both SP-A and SP-D concentrations were significantly correlated with the extent of alveolitis (a reversible change), whereas they did not correlate with the progression of fibrosis (an irreversible change). The SP-D concentration, unlike that of SP-A, was also related to the extent of parenchymal collapse and the rate of deterioration per year in pulmonary function. The concentrations of SP-A and SP-D in patients who died within 3 yr were significantly higher than in patients who were still alive after 3 yr. We propose that assays of SP-A and SP-D in sera from IPF patients are useful tools for understanding some pathologic characteristics of the disease, that SP-D may be a good predictive indicator of the rate of decline in pulmonary function, and that a combination of the assays for SP-A and SP-D may be helpful in predicting the outcome of patients with IPF.

Idiopathic pulmonary fibrosis (IPF) is a progressive interstitial lung disease of unknown etiology and with a poor prognosis. Carrington and colleagues (1) reported that the mean duration from diagnosis to death in IPF was nearly 6 yr. Schwarz and colleagues (2) stated that 41 of 74 (55.4%) patients with IPF died in a mean follow-up period of 4 yr after the onset of pulmonary symptoms, and that the median survival was 28.2 mo. However, predicting the progression of IPF and its prognosis remain difficult. The hydrophilic surfactant proteins (SP)-A and SP-D belong to the collectin subgroup of the C-type lectin superfamily, along with mannose-binding glycoproteins and collectin CL43 (3). Two types of nonciliated epithelial cells in the peripheral

(Received in original form October 19, 1999 and in revised form March 28, 2000) Supported by a Grant-In-Aid for Scientific Research from the Ministry of Education, Japan. Correspondence and requests for reprints should be addressed to Hiroki Takahashi, M.D., Ph.D., Third Department of Internal Medicine and Department of Biochemistry, Sapporo Medical University School of Medicine, South-1 West-16, Chuo-ku, Sapporo 060-8543, Japan. E-mail: [email protected] Am J Respir Crit Care Med Vol 162. pp 1109–1114, 2000 Internet address: www.atsjournals.org

airways, Clara cells and alveolar type II cells, produce these lung collectins (4, 5). We previously prepared the monoclonal antihuman SP-A antibodies PC6 and PE10 (6) and developed a sandwich enzyme-linked immunosorbent assay (ELISA) for SP-A by using these antibodies (7). This ELISA has been clinically applied to determine SP-A levels as an indicator of surfactant in amniotic fluids (8). By using the ELISA system and Western blot analysis, we showed that the 35-kD molecule of SP-A is also detected in blood, and that the levels of SP-A in sera from patients with IPF are significantly higher than those in sera of healthy subjects (9, 10). This assay system provides high sensitivity (71%) and specificity (87%) for detection of IPF. We also prepared monoclonal antibodies against human recombinant SP-D and developed an ELISA for the detection of SP-D in amniotic fluid (11). We applied this ELISA to the measurement of SP-D in sera and demonstrated that its concentrations are prominently increased in patients with IPF (12). Moreover, we recently developed a modified ELISA for the detection of human SP-D that is designed to avoid the effect of serum factors, including rheumatoid factor (13). When the modified method was used, the levels of SP-D in sera from IPF patients were also significantly higher than those of healthy subjects. This assay system provides high sensitivity (93%) and high specificity (92%) for detection of IPF. These findings show that measurements of SP-A and SP-D are useful as noninvasive and lung-specific clinical biomarkers. During an acute exacerbation of IPF, SP-A and SP-D concentrations increase in serum, making diagnosis of the disease easier (10, 12). However, it is not clear whether SP-A levels correlate with SP-D levels in chronic and steady-state IPF, and whether the levels of both surfactant proteins are related to deterioration of lung function and to mortality. In this study we evaluated the relationship between serum concentrations of SP-A and SP-D and changes seen with high-resolution computed tomography (HRCT), and whether measurements of SP-A and SP-D can predict either the progression of IPF as measured by pulmonary function tests (PFTs) or survival.

METHODS Subjects We conducted a retrospective study of 52 patients with IPF (43 male and nine female, consisting of 35 current smokers, 10 ex-smokers, and seven nonsmokers) aged 62.5 ⫾ 7.5 yr (mean ⫾ SD) (range: 45 to 78 yr) who were inpatients or outpatients at Sapporo Medical University Hospital between 1990 and 1996. The diagnosis of IPF was based on accepted criteria (1), which included either evidence of varying degrees of interstitial fibrosis and alveolitis or evidence of diffuse parenchymal infiltrates on chest radiography. The diagnoses were confirmed by open lung biopsies, video-assisted thoracoscopic surgery (VATS), or transbronchial biopsies. All patients enrolled in the study had the typical findings of IPF on chest radiography and pulmonary function testing. Exclusion criteria consisted of a clinically relevant history of environmental or occupational exposure, hypersensitivity

1110

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

VOL 162

2000

pneumonitis, collagen vascular disease, or other interstitial lung diseases such as acute interstitial pneumonia, diffuse alveolar damage, bronchiolitis obliterans organizing pneumonia, or lymphocytic interstitial pneumonia. All patients studied were clinically stable at the time of entry into the study. Forty of the patients were smokers and none had been given steroid therapy for their IPF. For analysis of receiver operating characteristics (ROC) we also studied 108 healthy volunteers (42 male and 66 female, consisting of 34 current smokers and 74 nonsmokers) aged 56.0 ⫾ 10.2 yr (range: 20 to 61 yr).

Collection and Analysis of Blood Samples Peripheral venous blood samples, which were collected from the patients at their initial visits and from healthy subjects at the time of registration for the study, were used for SP-A and SP-D assays. The serum samples had been stored at ⫺80⬚ C and were analyzed in a blinded-fashion with regard to the clinical status of the patients. SP-A assay was done with ELISA kits provided by the Teijin Institute of Bio-medicine (Tokyo, Japan). A method based on that of Shimizu and coworkers (7) was adapted with minor modifications (10). The assay system was able to detect SP-A at concentrations of 2.0 to 250 ng/ ml. The concentration of SP-D was also measured with an ELISA, using recombinant SP-D as a standard and two monoclonal antibodies against human SP-D (13). This assay system was able to detect SP-D at concentrations of 1.56 to 100 ng/ml. All assays were performed in duplicate, and the results were given as the mean value. In immunoblot analyses, specific antibodies used for the SP-A and SP-D assay systems showed no nonspecific crossreactivity, such as with mannose binding protein or other serum proteins.

Chest Computed Tomographic Findings To assess the relationship between findings on chest computed tomography (CT) and levels of SP-A and SP-D, we evaluated 49 IPF patients who were examined through HRCT scanning as a routine clinical test. All of these patients were clinically stable and had not been receiving steroid therapy. The HRCT scan, done with a GE 9800 scanner (General Electric, WI), was performed within 2 wk after collection of the blood sample taken at the patients’ initial visits. All images were obtained at maximal inspiration, using 1.5-mm collimation with 10-mm gaps from the apex of the lung to the diaphragm in the supine position. Images were reconstructed with a high–spatial-frequency (bone) algorithm. Evaluation of the findings on HRCT was made by three clinicians who are experts in radiographic diagnosis and were blinded to the patients’ clinical courses and serum SP-A and SP-D values. Final conclusions were reached by means of consensus. The presence and distribution of CT findings were determined as follows: (1) honeycombing was defined as cystic spaces with thickened walls; and (2) a ground-glass opacity (GGO) was defined as a minor increase in density in lung parenchyma. The extent of honeycombing and that of GGO were scored separately on scales ranging from 1 to 3, based on modified criteria of a scoring system for IPF developed by Kazerooni and associates (14). Observers scored two CT images recorded at the level of the carina and 2 cm above the diaphragm. Scores of 1, 2, and 3 for honeycombing and GGO were defined as 0% to ⬍ 5%, 5% to ⬍ 25%, and ⭓ 25% involvement of the total lung field on the two CT images. On the basis of total characteristics identified on HRCT, the 49 IPF patients were grouped according to two types of findings: a parenchymal collapse–opacity (PCO)-dominant type and a GGO-dominant type (Figure 1). PCO was defined as air bronchiolograms with intense lung attenuation with parenchymal collapse, often accompanied by thickened vessels and bronchial wall and subpleural involvement, on the basis of the observations of Nishimura and coworkers (15).

Pulmonary Function Tests Inspiratory VC and TLC were measured with a Chestac-55V system (Chest, Tokyo, Japan), and all values were expressed as percentages of reference values (16). Twenty-three of the 52 IPF patients in the study had serial PFTs over a period of more than 24 mo. The PFTs were repeated several times while the patients were followed for various periods (24 to 44 mo). The times and time intervals of PFTs were not the same for all of the patients because the study was done retrospectively. The minimum and maximum intervals between a PFT and

Figure 1. Typical images of two types of IPF distinguished by characteristics on chest HRCT. (A) GGO-dominant type. (B) PCO-dominant type. See METHODS for detailed description of HRCT findings in the two types of disease.

the next PFT were 3 mo and 12 mo, respectively. The follow-up interval was 35 ⫾ 7 mo. PFTs performed during exacerbations of IPF were excluded from analysis. To simplify the study design, we analyzed only two values (the initial and final PFT values for each patient). The initial values were obtained within 2 wk after collection of the blood sample at the initial visit. The final values were the final PFT values in the follow-up period. The relation of SP-A or SP-D to the percentage change in each parameter (⌬%VC and ⌬%TLC) was studied. For example, the percent change in VC was calculated according to the formula: ⌬%VC ⫽ (%VCfinal ⫺ %VCinitial)/%VCinitial ⫻ 100. Moreover, the percentage change per year in each functional parameter was calculated as ⌬%VC/yr ⫽ ⌬%VC/number of follow-up months ⫻ 12.

Analysis of Survival To evaluate whether initial values of SP-A and SP-D can predict survival for IPF patients, we compared 10 patients who died of acute exacerbations of IPF within 3 yr from the initial measurement of SP-A and SP-D with 42 patients who were still alive 3 yr later. The average survival time of the 10 patients who died was 11.4 mo (range: 3 to 28 mo).

Statistical Analysis Data are expressed as mean ⫾ SD. Differences between two variables were assessed with the Mann-Whitney U test. Differences between data from the three study groups that were created on the basis of the scoring on HRCT were analyzed with one-way analysis of variance. Significant differences between means were analyzed with the post hoc test using Scheffe’s F. Spearman’s rank correlation coefficient was used to analyze correlations between the score on HRCT and concentrations of SP-A or SP-D. Analysis of correlations between SP-A and SP-D was done with Pearson’s product moment coefficient of correlation. The concentrations of SP-A and SP-D were further analyzed by using ROC curves in order to find the cutoff levels indicating the best sensitivity and specificity of these two measures (17). Significance was defined as p ⬍ 0.05.

RESULTS Determination of Cutoff Levels of SP-A and SP-D for the Detection of IPF

The ROC curves drawn for determination of the cutoff levels of SP-A and SP-D are shown in Figure 2. When tentative cutoff levels for SP-A measured in the 52 IPF patients and the 108 healthy volunteers were evaluated at intervals of 2.5 ng/ml from 0 ng/ml to 80.0 ng/ml, 45.0 ng/ml was found to be the best value in the relationship between sensitivity (78.8%) and specificity (94.4%). When tentative cutoff levels for SP-D measured in the same subjects were evaluated at intervals of 10 ng/ml from 0 ng/ml to 250 ng/ml, 110 ng/ml was found to be the best value in the relationship between sensitivity (84.6%) and specificity (95.4%). Therefore, the definitive cutoff levels of SP-A and SP-D were set at 45.0 ng/ml and 110 ng/ml, respectively.

1111

Takahashi, Fujishima, Koba, et al.: Serum SP-A and SP-D in IPF

Figure 2. ROC curves showing sensitivity and specificity of SP-A (left) and SP-D (right) for the detection of IPF. The ranges of concentrations of SP-A and SP-D used for this analysis were 0 to 80 ng/ml and 0 to 250 ng/ml, respectively. Each interval for plotting was 2.5 ng/ml for SP-A and 10 ng/ml for SP-D. The ROC curve identified 45 ng/ml and 110 ng/ml, respectively, as the best cutoff values for SP-A and SP-D. See RESULTS for description of setting of the cutoff levels.

Correlation between Scoring of HRCT Findings and Serum Levels of SP-A and SP-D

HRCT findings from 49 IPF patients were evaluated for correlation between scoring of HRCT findings and serum levels of SP-A and SP-D. The extent of GGO correlated significantly with serum levels of SP-A and SP-D (SP-A: p ⫽ 0.791, p ⬍ 0.0001; SP-D: p ⫽ 0.446, p ⫽ 0.0034), whereas the extent of honeycombing did not correlate with levels of either surfactant protein (Figure 3). There was also a significant correlation between the levels of SP-A and those of SP-D (r ⫽ 0.491, p ⫽ 0.0003). The relatively low value of r suggested that some other findings on HRCT might also be involved in the elevation of serum levels of SP-A and/or SP-D. One possible candidates for this was PCO, which mainly reflects a collapse of peripheral airway spaces in which alveolar type II cells and Clara cells exist and secrete SP-A and SP-D. Since it was difficult to quantify the extent of PCO, we grouped the IPF subjects simply into a PCO-dominant type (n ⫽ 23) and a GGO-dominant type (n ⫽ 26). SP-A levels (98.3 ⫾ 55.8 ng/ml) in the GGOdominant type were significantly (p ⫽ 0.0003) higher than those (51.3 ⫾ 33.3 ng/ml) in the PCO-dominant type, whereas SP-D levels were not significantly different for the two types

(GGO-dominant type: 243.1 ⫾ 142.4 ng/ml; PCO-dominant type: 266.6 ⫾ 161.1 ng/ml) (Figure 4). The sensitivities of the SP-A and SP-D assays were estimated and compared between the two types by using the cutoff levels of SP-A and SP-D (45.0 ng/ml and 110 ng/ml, respectively). The sensitivity of SP-A in the GGO-dominant–type group (21 of 26 patients, 81%) was much higher than that in the PCO-dominant–type group (12 of 23 patients, 52%). In contrast, the sensitivity of SP-D in the two type groups was not different (GGO-dominant type: 22 of 26 patients, 85%, PCO-dominant type: 19 of 23 patients, 83%). Collectively, these results indicate that PCO may be related more strongly to the elevation of SP-D than to that of SP-A levels. Relationship between Deterioration of PFT Values and Serum Level of SP-A or SP-D

Twenty-three IPF patients were studied for the relationship between deterioration of PFT values and serum level of SP-A or SP-D. At the start of follow-up, levels of neither SP-A nor SP-D correlated with %VC or %TLC. However, SP-D levels at the initial time of study correlated significantly with ⌬%VC (r ⫽ ⫺0.560, p ⫽ 0.0047), ⌬%VC/yr (r ⫽ ⫺0.520, p ⫽ 0.0099), ⌬%TLC (r ⫽ ⫺0.476, p ⫽ 0.0205), and ⌬%TLC/yr (r ⫽ ⫺0.429, p ⫽ 0.0405) (Figure 5). In contrast, SP-A levels did not correlate significantly with any parameters. These results indicate that SP-D, but not SP-A, may reflect the rate of worsening of restrictive pulmonary disturbances. Relationship between Survival and SP-A or SP-D

Ten patients who died of acute exacerbations of IPF before the end of the 3-yr-follow-up period (nonsurvivors) were compared with 42 patients who were still alive (survivors) for study of the relationship between survival and SP-A or SP-D (Figure 6A). Overall mortality was 19%. No significant differ-

Figure 3. Relationship between concentrations of SP-A and SP-D and extent of GGO (A and B) or honeycombing (C and D) on HRCT. Horizontal lines are cutoff levels of SP-A and SP-D. Grades 1, 2, and 3 correspond to 0 to ⬍ 5%, 5 to 25%, and ⭓ 25% extent of involvement, respectively. The correlation was analyzed with Spearman’s rank correlation test. The statistical comparison was done with a post hoc test using Scheffe’s F. *p ⬍ 0.01.

Figure 4. Comparison of serum concentrations of SP-A and SP-D in GGO-dominant type (n ⫽ 26) and PCO-dominant type (n ⫽ 23) IPF on HRCT. Each bar represents the mean ⫾ SD. Horizontal lines indicate cutoff levels for SP-A and SP-D.

1112


VOL 162

2000

Figure 5. Relationship between serum concentrations of SP-A and SP-D and percentage change per year in parameters of restrictive pulmonary disturbance.

ence was found between the ages of the nonsurvivors and the survivors (63.2 ⫾ 8.7 yr and 62.3 ⫾ 7.3 yr, respectively). The percentage of smokers at the start of the follow-up period was greater among the survivors (34 of 42 patients, 81%) than among the nonsurvivors (six of 10 patients, 60%). Initial SP-A levels in the nonsurvivors (117.7 ⫾ 66.8 ng/ml) were significantly higher than those in the survivors (68.8 ⫾ 40.4 ng/ml) (p ⫽ 0.0125). Initial SP-D levels in the nonsurvivors (453.7 ⫾ 290.3 ng/ml) were also significantly higher than those in the survivors (248.0 ⫾ 176.4 ng/ml) (p ⫽ 0.0032). Forty-two (80.8%) and 44 (84.6%) of the 52 IPF patients studied showed higher levels of SP-A and SP-D, respectively, than the cutoff level for each protein. Only 27 (64.3%) of the 42 patients still alive showed SP-A and SP-D levels above the cutoff level. In contrast, all of the 10 nonsurvivors exhibited levels of SP-A and SP-D that were higher than the cutoff level (Figure 6B). Thus, the study suggests that finding normal levels of SP-A and/or SP-D at the times of initial elevation in serum levels of these proteins is a better prognostic indicator, since none of the patients who had normal levels died.

DISCUSSION Alveolar type II cells are the major sources of the SP secreted in alveoli. On the other hand, it has been reported that messenger RNAs (mRNAs) for SPA and SP-D are expressed in other organs such as rat gastric mucosa, small intestine, and colon (18, 19), and human gastric mucosa (20), whereas secretion of these two proteins by gastrointestinal mucous cells has not been demonstrated. Expression of SP-A and SP-D in the gastrointestinal tract appears very weak as compared with that in the lung. A previous study revealed that serum protein binding to the mannose-affinity matrix exhibits SP-A with an apparent molecular mass identical to that of SP-A isolated from bronchoalvelor lavage fluid (BALF), and which is recognized by a monoclonal anti-SP-A antibody, indicating the presence of SP-A in serum (21). Thus, most SP-A and SP-D in the serum is probably derived from lung. We have reported that serum SP-A and SP-D levels were increased in IPF patients (9, 10, 12), and we hypothesize that such increases reflect pathologic changes in the lung. Because HRCT can precisely image pathologic changes, it is now commonly used as a noninvasive method of assessment of patients with diffuse infiltrative lung diseases (22, 23). Areas of GGO

Figure 6. Relationship between serum concentrations of SP-A and SP-D at the initial time of the study and survival during 3 yr follow-up. (A) mean ⫾ SD and individual values of SP-A and SP-D. Horizontal lines indicate cutoff levels of SP-A and SP-D. Asterisk indicates p ⬍ 0.05 compared with values for survivors. (B) (Top panel) patients showing the cutoff level or higher; (bottom panel) patients showing levels below the cutoff level. See METHODS section for description of setting of cutoff levels.

on HRCT most commonly reflect increased cellularity in alveolar interstitium in patients with IPF, whereas areas of honeycombing, regarded as synonymous with reticular opacity, correspond to fibrotic histologic change (15, 24, 25). Therefore, as a first step in clarifying the significance of the increases in serum levels of SP-A and SP-D in IPF, we evaluated the relationship between the extent of disease on HRCT and serum levels of SP-A and SP-D. Our study showed that the extent of GGO correlated significantly with serum levels of SP-A and SP-D, whereas the extent of honeycombing did not correlate with levels of either protein. GGO is related in part to alveolitis, which is a potentially reversible and treatable parenchymal abnormality (26, 27), and may in part represent very fine intralobular fibrosis (28). In contrast, honeycombing is a finding made at the end stage of fibrotic change, and seldom regresses, even after medication. Our study indicates that increased SP-A and SP-D levels may reflect the extent of potentially reversible moieties in diverse abnormalities in the lungs of IPF patients. Since the possibility of optimal therapeutic management of IPF requires an accurate tool for discriminating between the reversible and irreversible forms of the disease, assays for SP-A and SP-D could help in selecting therapies for it. Among all the subjects of the present study, SP-A significantly correlated with SP-D, but the correlation coefficient was not large (r ⫽ 0.491). This suggests that the mechanisms of the increases in the two proteins could be partly different from each other, and that some factors other than the extent of GGO are independently involved in the increase in SP-A or SP-D in IPF. The collapsed and fibrotic changes in the peripheral air spaces of alveoli and bronchioles, called parenchymal collapse-opacity (PCO), also occur in IPF, and as frequently as do GGO and honeycombing (as stated by Nakamura and associates [15]). We hypothesized that these changes might be the “other factors” in the increase in levels of SP-A or SP-D. This change is also commonly seen in interstitial lung diseases

1113

Takahashi, Fujishima, Koba, et al.: Serum SP-A and SP-D in IPF

associated with collagen vascular diseases such as polymyositis/dermatomyositis (29). In our results, the concentrations of SP-A in patients with GGO-dominant type lung damage were significantly higher than those in patients with PCO-dominant type damage. On the other hand, SP-D increased not only in patients with GGO-dominant type but also in those with PCOdominant damage. Katzenstein and colleagues (30) stated that alveolar collapse and reorganization appears to be related in part to the pathogenesis of honeycombing in usual interestital pneumonia. Therefore, SP-D rather than SP-A might predict the progress from alveolar collapse to fibrosis. The difference in their concentrations in IPF patients with PCO-dominant as patients compared with GGO-dominant type lung damage, which in certain instances reflects some pathologic difference, may be brought about by a difference in the mechanism of leakage of the two proteins from lung into the systemic circulation. We attempted to assess whether deterioration in PFT results can be predicted by measuring the concentrations of SP-A and/or SP-D in sera from patients with IPF. Interestingly, we found that the levels of SP-D at the initial time of study, unlike those of SP-A, correlated significantly with the velocity of decline in VC and TLC. Most subjects studied, even many who showed high levels of SP-D, did not have dyspnea at the initial time of study. Nevertheless, our results clearly indicate that high levels of SP-D are involved in subsequent declines in %VC and %TLC. Thus, patients exhibiting higher serum levels of SP-D may have a greater chance of falling into restrictive pulmonary dysfunction, and more rapidly, than patients with low serum levels of SP-D. Several reports (31, 32) suggest that it may be more effective to start treatment for IPF before the manifestations of severe pulmonary fibrosis occur. Our results raise the possibility that the assay of SP-D can help guide therapy with corticosteroids and/or immunosuppressive agents. In the future, a prospective study including serial measurements of SP-A and SP-D is needed to determine whether their values rise as lung function declines. McCormack and coworkers (33) have reported a decrease in the ratio of SP-A to phospholipid in BALF from patients with IPF. The SP-A-to-phospholipid ratio predicted 5-yr survival better than any other lavage factors measured. On the basis of the findings of McCormack and coworkers, we evaluated the utility of assays of serum SP-A and SP-D in establishing the prognosis of patients with IPF. We found that patients who died within 3 yr of the observation period showed significantly higher levels of both SP-A and SP-D than those who were still alive beyond these 3 yr. Moreover, in the 3 yr interval, none of the patients showing SP-A and/or SP-D levels below the respective cutoff levels died. Although the number of patients in our study was small, these findings suggest that a combination of SP-A and SP-D assays is useful to identify patients with the best prognosis in IPF. Concentrations of SP-A in BALF from patients with IPF are significantly lower than those in BALF from healthy subjects (33), whereas differences in concentrations of SP-D in the two populations are not significant (12). A difference between levels of SP-A and SP-D was also observed in patients with acute respiratory distress syndrome (ARDS) (34). To explain the difference in patients with ARDS, Greene and coworkers (34) hypothesized that the decline in SP-A concentration mainly reflects a regulatory abnormality in SP-A metabolism, whereas decreases in SP-D levels in alveoli are consistent with the destruction of alveolar type II cells. This hypothetical interpretation could also be relevant for IPF, another disease marked by chronic lung injury. In addition, SP-A is more soluble than SP-D, since most SP-A tightly binds to

surfactant lipid aggregates, whereas most SP-D appears to be lipid-free. These differences in their response to inflammatory products and solubility could also affect concentrations of SP-A and SP-D in serum as well as BALF. Moreover, our data suggest that the pathologic difference in types of IPF, which is presumed to exist on the basis of HRCT findings, alters the concentrations of both surfactant proteins in serum. More evidence is needed to clarify the mechanism of the leakage of SP-A and SP-D into the bloodstream and the difference in concentrations of these surfactant proteins. It is also necessary to clarify how SP-A and SP-D are removed from the bloodstream, and whether a rational basis exists for a difference in regulation of these proteins. In conclusion, assays of SP-A and SP-D may help to understand some of the pathologic characteristics of IPF seen on HRCT. The SP-D assay may indicate the rate of decline in pulmonary function in this disease. Assays of SP-A and SP-D may also assist in making clinical choices for therapeutic management of patients with IPF. A careful prospective study is needed to verify the hypothesis of this study. Acknowledgment : The writers wish to thank Dr. T. Akino of the Sapporo Medical University for valuable suggestions and encouragement.

References 1. Carrington, C. B., E. A. Gaensler, R. E. Coutu, M. X. Fitzgerald, and R. G. Gupta. 1978. Natural history and treated course of usual and desquamative interstitial pneumonia. N. Engl. J. Med. 298:801–809. 2. Schwartz, D. A., R. A. Helmers, J. R. Galvin, D. S. Van Fossen, K. L. Frees, C. S. Dayton, L. F. Burmeister, and G. W. Hunninghake. 1994. Determinants of survival in idiopathic pulmonary fibrosis. Am. J. Respir. Crit. Care Med. 149:450–454. 3. Day, A. J. 1994. The C-type carbohydrate recognition domein (CRD) superfamily. Biochem. Soc. Trans. 22:83–88. 4. Walker, S. R., M. C. Williams, and B. Benson. 1986. Immunocytochemical localization of the major surfactant apoprotein in type II cells, Clara cells and alveolar macrophages of rat lung. J. Histochem. Cytochem. 34:1137–1148. 5. Voorhout, W. F., T. Veenendaal, Y. Kuroki, Y. Ogasawara, L. M. G. van Golde, and H. J. Geuze. 1992. Immunocytochemical localization of surfactant protein (SP-D) in type II cells, Clara cells, and alveolar macrophages of rat lungs. J. Histochem. Cytochem. 40:1589–1597. 6. Kuroki, Y., Y. Fukada, H. Takahashi, and T. Akino. 1985. Monoclonal antibodies against human pulmonary surfactant apoproteins: specificity and application in immunoassay. Biochim. Biophys. Acta 836:201–209. 7. Shimizu, H., K. Hosoda, M. Mizumoto, Y. Kuroki, H. Sato, K. Kataoka, M. Hagisawa, S. Fujimoto, and T. Akino. 1989. Improved immunoassay for the determination of surfactant protein A (SP-A). Tohoku J. Exp. Med. 157:269–278. 8. Kuroki, Y., H. Takahashi, Y. Fukada, M. Mikawa, A. Inagawa, S. Fujimoto, and T. Akino. 1985. Two-site “simultaneous” immunoassay with monoclonal antibodies for the determination of the surfactant apoproteins in human amniotic fluid. Pediatr. Res. 19:1017–1020. 9. Kuroki, Y., S. Tsutahara, N. Shijubo, H. Takahashi, M. Shiratori, A. Hattori, Y. Honda, S. Abe, and T. Akino. 1993. Elevated levels of lung surfactant protein A in sera from patients with idiopathic pulmonary fibrosis and pulmonary alveolar proteinosis. Am. Rev. Respir. Dis. 147:723–729. 10. Honda, Y., Y. Kuroki, N. Shijubo, T. Fujishima, H. Takahashi, K. Hosoda, T. Akino, and S. Abe. 1995. Aberrant appearance of lung surfactant protein-A in sera of patients with idiopathic pulmonary fibrosis and its clinical significance. Respiration 62:64–69. 11. Inoue, T., E. Matsura, A. Nagata, Y. Ogasawara, A. Hattori, Y. Kuroki, S. Fujimoto, and T. Akino. 1994. Enzyme-linked immunosorbent assay for human pulmonary surfactant D. J. Immunol. Methods 173: 157–164. 12. Honda, Y., Y. Kuroki, E. Matuura, H. Nagae, H. Takahashi, T. Akino, and S. Abe. 1995. Pulmonary surfactant protein-D in sera and bronchoalveolar lavage fluids. Am. J. Respir. Crit. Care Med. 152:1860–1866. 13. Nagae, H., H. Takahashi, Y. Kuroki, Y. Honda, A. Nagata, Y. Ogasawara, S. Abe, and T. Akino. 1997. Enzyme-linked immunosorbent assay using F(ab⬘)2 fragment for the detection of human pulmonary surfactant D in sera. Clin. Chim. Acta 266:157–171.

1114


14. Kazerooni, E. A., F. J. Martinez, A. Flint, D. A. Jamadar, B. H. Gross, D. L. Spizarny, P. N. Cascade, R.I. Whyte, J. P. Lynch 3rd, and G. Toews. 1997. Thin-section CT obtained at 10-mm increments versus limited three-level thin-section CT for idiopathic pulmonary fibrosis: correlation with pathologic scoring. Am. J. Roentgenol. 169:977–983. 15. Nishimura, K., M. Kitaichi, T. Izumi, S. Nagai, M. Kanaoka, and H. Itoh. 1992. Usual interstitial pneumonia: histologic correlation with high resolution CT. Radiology 182:337–342. 16. Quanjer, P. H. 1983. Standardized lung function testing. Bull. Eur. Physiolpathol. Respir. 19:7–44. 17. Zweig, M. H., and G. Campbell. 1993. Receiver-operating characteristic (ROC) plot: a fundamental evaluation tool in clinical medicine. Clin. Chem. 39:561–557. 18. Fisher, J. H., and R. J. Mason. 1995. Expression of pulmonary surfactant protein D in rat gastric mucosa. Am. J. Respir. Cell Mol. Biol. 12:13–18. 19. Rubio, S., T. Lacaze-Masmonteil, B. Chailley-Heu, A. Kahn, R. J. Bourbon, and R. Ducroc. 1995. Pulmonary surfactant A (SP-A) is expressed by epithelial cells of small and large intestine. J. Biol. Chem. 270:12162–12169. 20. Motwani, M., R. A. White, N. Guo, L. L. Dowler, A. I. Tauber, and K. N. Sastry. 1995. Mouse surfactant protein-D: cDNA cloning, characterization, and gene localization to chromosome 14. J. Immunol. 155: 5671–5677. 21. Kuroki, Y., H. Takahashi, H. Chiba, and T. Akino. 1998. Surfactant protein A and D: disease markers. Biochim. Biophys. Acta 1408:334–345. 22. Hansell, D. M., and A. U. Wells. 1996. CT evaluation of fibrosing alveolitis—applications and insights. J. Thorac. Imaging 11:231–249. 23. Muller, N. L., and R. R. Miller. 1990. Computed tomography of chronic diffuse infiltrative lung disease: part 1. Am. Rev. Respir. Dis. 142:1206–1215. 24. Engeler, C. E., J. H. Tashjian, S. W. Trenkner, and J. W. Walsh. 1993. Ground-glass opacity of the lung parenchyma: a guide to analysis with high-resolution CT. Am. J. Roentgenol. 160:249–251. 25. Muller, N. L., C. A. Staples, R. R. Miller, S. Vedal, W. M. Thurlbeck,

26.

27.

28.

29.

30. 31.

32.

33.

34.

VOL 162

2000

and D. N. Ostrow. 1987. Decreased activity in idiopathic pulmonary fibrosis: CT and pathologic correlation. Radiology 165:731–734. Remy-Jardin, M., F. Giraud, J. Remy, M. C. Copin, B. Gosselin, and A. Duhamel. 1993. Importance of ground-glass attenuation in chronic diffuse infiltrative lung disease: pathologic-CT correlation. Radiology 189:693–698. Leung, A. N., R. R. Miller, and N. L. Muller. 1993. Parenchymal opacification in chronic infiltrative lung diseases: CT-pathologic correlation. Radiology 188:209–214. Akira, M., S. Yamamoto, K. Yokoyama, N. Kita, K. Morinaga, T. Higashihara, and T. Kozuka. 1990. Asbestosis: high-resolution CT-pathologic correlation. Radiology 176:389–394. Ikezoe, J., T. Johkoh, N. Kohno, N. Takeuchi, K. Ichikado, and H. Nakamura. 1996. High-resolution CT findings of lung disease in patients with polymyositis and dermatomyositis. J. Thorac. Imaging 11:250–259. Katzenstein, A. L. 1985. Pathogenesis of “fibrosis” in interstitial pneumonia: an electron microscopic study. Hum. Pathol. 16:1015–1024. Schwartz, D. A., D. S. Van Fossen, C. S. Davis, R. A. Helmers, C. S. Dayton, L. F. Burmeister, and G. W. Hunninghake. 1994. Determinants of progression in idiopathic pulmonary fibrosis. Am. J. Respir. Crit. Care Med. 149:444–449. Turner-Warwick, M., B. Burrows, and A. Johnson. 1980. Cryptogenic fibrosing alveolitis: clinical features and their influence on survival. Thorax 35:171–180. McCormack, F. X., T. E. King Jr., D. R. Voelker, P. C. Robinson, and R. J. Mason. 1991. Idiopathic pulmonary fibrosis: abnormalities in the bronchoalveolar lavage content of surfactant protein A. Am. Rev. Respir. Dis. 144:160–166. Greene, K. E., J. R. Wright, K. P. Steinberg, J. T. Ruzinski, E. Caldwell, W. B. Wong, W. Hull, J. A. Whitsett, T. Akino, Y. Kuroki, H. Nagae, L. D. Hudson, and T. R. Martin. 1999. Serial changes in surfactantassociated proteins in lung and serum before and after onset of ARDS. Am. J. Respir. Crit. Care Med. 160:1843–1850.