Published ahead of print on November 1, 2007, doi:10.1164/rccm.200705-742OC
© 2008 American Thoracic Society doi: 10.1164/rccm.200705-742OC
Phenotyping Sarcoidosis from a Pulmonary Perspective1 Department of Pneumology and 2 Department of Laboratory Chemistry, University Hospital Freiburg, Freiburg, Germany Correspondence and requests for reprints should be addressed to Antje Prasse, M.D., Abteilung für Pneumologie, Universitätsklinik Freiburg, Killianstr. 5, 79106 Freiburg, Germany. E-mail: antje.prasse{at}uniklinik-freiburg.de
Rationale: Sarcoidosis is known as a disease with high heterogeneity in clinical severity and inflammatory activity. On the basis of radiologic criteria, John Guyette Scadding developed a classification system, which is widely used, but is insufficient for clinical decision making. Therefore, biomarkers and genetic markers that predict outcome are urgently needed. Objectives: To obtain a classification system based on clinical criteria to evaluate biomarker and genetic markers. Methods: We developed a protocol for classification of various disease courses (sarcoid clinical activity classification [SCAC]) based on clinical criteria with three categories: (1) whether the disease was acute or nonacute in onset, (2) whether treatment was required, and (3) whether there was need for long-term treatment. Measurements and Main Results: In total, we evaluated 225 patients with sarcoidosis, applying both classification protocols retrospectively. The described SCAC protocol based on clinical criteria was retrospectively able to stratify patients according to disease outcome. The classes of patients with chronic disease differed significantly regarding pulmonary function test parameters and bronchoalveolar lavage fluid cell composition. Most interestingly, concentrations of soluble IL-2 receptor and neopterin were particularly increased in patients with acute disease who required long-term treatment. A moderate but significant increase in soluble IL-2 receptor and neopterin was also observed in patients with nonacute disease who needed long-term treatment. In contrast to the clinical classification system, the system based on radiologic criteria separated the patients with the need for long-term therapy insufficiently, but identified patients with advanced fibrotic remodeling. Conclusions: The described SCAC protocol is practicable and gives additional information not yet acquired by radiologic typing and seems suitable for studies evaluating genetic influence and biomarkers.
Key Words: sarcoidosis • chest X-ray type • treatment • phenotyping • classification
Sarcoidosis is a chronic multisystem, granulomatous disease of unknown etiology. The variety of the disease is well known (1–6). As far back as 1953, Löfgren and coworkers (7, 8) had already shown that there is a benign form of sarcoidosis with an acute onset, bihilar lymphadenopathy, and bilateral arthritis, which was named after its discoverer. Because of the possibilities opened up by genetic sequencing methods, the interest in classification protocols based on clinical criteria for sarcoidosis has regained considerable focus (9–12). More recently, such approaches have been successfully used in systemic lupus erythematodes and bronchial asthma, where specific phenotypes have been shown to be associated with distinct genotypes (13, 14). New techniques in human genetics and molecular biology open up the possibility to identify genetic risk profiles for certain sarcoid phenotypes. To this end, clinical phenotypes need to be defined that segregate in clinical course and therapeutic consequences. For example, it is well described that patients with Löfgren's syndrome have a better prognosis than do patients with other manifestations of the disease (15). However, there is no uniform definition of Löfgren's syndrome (16). Some authors tend to classify all patients with an acute onset of disease, weight loss, and malaise as having Löfgren's syndrome, whereas others include only patients with the full picture of the disease described by Löfgren, including bihilar lymphadenopathy, arthritis, as well as erythema nodosum. A strong association between Löfgren's syndrome and variants of genes on chromosome 3 in the vicinity of CCR2 has been described (17, 18). These data demonstrate the need to classify patients according to criteria of clinical relevance. In this context, we have developed a classification protocol that classifies patients as having either acute or nonacute disease. In addition, we further subdivided patients according to their need for treatment with steroids and response to treatment. In this study, we evaluated this clinical classification protocol in comparison with Scadding's well-known system based on chest X-ray types (19) by a retrospective analysis of 225 patients with sarcoidosis who attended our clinic. Some of the results of these studies have been previously reported in the form of an abstract (20).
Subjects Retrospectively, on the basis of a digital data system of our hospital, 654 patients were identified who were diagnosed as having sarcoidosis and for further evaluation their records were obtained. Only patients for whom there was a record of clinical data of histologically proven sarcoid disease for at least the previous 2 years were recruited in the study, after obtaining their informed consent. Only patients who had acute Löfgren's syndrome defined as bihilar lymphadenopathy, erythema nodosum, and arthritis were included despite the absence of histologic proof. Finally, 225 patients with sarcoidosis fulfilled the inclusion criteria. Our study is limited by the bias that our patient cohort is highly selected. Our sarcoidosis center is situated at a university hospital and therefore has the status of a referral center. Our patient cohort is not able to represent normal distribution among patients with sarcoidosis. We have developed a classification system that classifies patients as having sarcoidosis either with acute or nonacute disease onset. Acute disease was defined as a sudden onset of the disease (defined as malaise) accompanied by fatigue, subfebrile temperature, weight loss, and night sweat. Both sudden onset and at least two of these four symptoms had to be present. The time frame in which the symptoms had to be present was 1 month. Patients who were initially diagnosed with an acute onset but showed later in the course a relapse of the disease with no further signs of acute disease were classified as nonacute, whereas patients who were initially diagnosed with an acute onset and later showed remitting symptoms of acute disease (fatigue, subfebrile temperature, weight loss, and night sweat) were classified as having acute sarcoidosis. In addition, we further subdivided patients according to their need for treatment with steroids and their response to treatment. Patients who needed only one period of treatment, lasting for at most 1 year, were separated from patients who needed treatment with steroids for more than one period or long-term treatment lasting for more than 1 year (Table 1). It is worth noting that this system requires a standardization of treatment, which is done in our center according to the guidelines proposed by the American Thoracic Society/European Respiratory Society (6). This means only patients with life-threatening organ involvement, progressive loss of pulmonary function, or other situations defined in the guidelines have been treated. According to these criteria, all patients could be classified as belonging to one of the six distinct groups.
In this study, we compare our classification system based on clinical data with the well-known Scadding system (20, 21) based on chest X-ray types. Each patient's chest X-ray was analyzed independently by a radiologist and a specialist in respiratory care medicine: type 0, no radiographic abnormalities; type 1, bilateral hilar adenopathy without parenchymal abnormalities; type 2, bilateral hilar adenopathy with interstitial parenchymal infiltrates; type 3, interstitial parenchymal infiltrates without hilar adenopathy; and type 4, cicatricial changes. In the case of discordant typing computed tomographic scans were obtained and a unanimous decision was reached.
Bronchoalveolar Lavage and Processing of Bronchoalveolar Lavage Fluid Cells
Pulmonary Function Tests
Analysis of Serum Parameters The determination of ACE was performed by spectrophotometric assay, using the synthetic tripeptide substrate N-[3-(2-furyl)acryloyl]-L-phenylalanylglycylglycine (FAPGG), which is catalyzed by ACE to FAP and glycylglycine (Trinity Biotech, Bray, Ireland). This hydrolysis results in a decrease in absorbance at 340 nm. C-reactive protein (CRP) was analyzed by particle-enhanced immunoturbidimetric assay (Roche Diagnostics, Darmstadt, Germany). Anti-CRP antibodies coupled to Latex microparticles react with antigen in the sample to form an antigen–antibody complex. The subsequent agglutination is measured turbidimetrically. sIL-2R was measured by means of a solid-phase two-site chemiluminescence immunometric assay (DPC Biermann, Bad Nauheim, Germany). sIL-2R in the sample binds to the solid-phase antibody and reagent antibody (coupled with alkaline phosphatase), forming a sandwich complex. A chemiluminescent signal is generated in proportion to the bound enzyme. A solid-phase ELISA was used to determine neopterin (IBL, Hamburg, Germany). Only data from patients who did not receive any immunosuppressive therapy were considered. For ACE, data from 155 patients were available. In our department routine analysis of neopterin and sIL-2R is performed for 4 years, and sIL-2R data were available for 101 patients and neopterin data were available for 87 patients.
Statistical Analysis
Patient Characteristics In total, 225 patients were included (n = 104 female, n = 121 male). The mean age was 45.5 ± 13.4 years at initial diagnosis. Of these 225 patients, 9 patients (4%) were classified as chest X-ray type 0, 63 (28%) as chest X-ray type I, 103 (46%) as chest X-ray type II, 46 (20%) as chest X-ray type III, and 4 (2%) as chest X-ray type IV. According to the sarcoid clinical activity classification (SCAC) system proposed by us, 40% (n = 89) of the study cohort had acute sarcoidosis, whereas 136 patients (60%) had nonacute sarcoidosis. Thirty-one patients with Löfgren's syndrome were included in the acute sarcoidosis class (14%). None were included in the nonacute disease class. Forty-four patients (20%) were classified as having class 1 disease, 26 (12%) had class 2 disease, 19 (8%) had class 3 disease, 52 (23%) had class 4 disease, 32 (14%) had class 5 disease, and 52 (23%) had class 6 disease (see Table 1). Further patient characteristics according to the distinct classes are listed in the online supplement. Only three patients fulfilled the criteria of acute disease but presented later in the disease course with a relapse without fulfilling the criteria. These patients were classified as having nonacute disease and all have experienced progressive pulmonary disease. Comparison of both classification systems showed that all patients with sarcoidosis chest X-ray type IV were classified to class 6 in the classification system based on clinical data (Table 2). Class 6 consisted of patients with nonacute disease requiring long-term therapy (more than 12 mo) (see Table 2; and Figure E1 in the online supplement). It is worth noting that the majority of patients who were classified as belonging to class 6 had type II chest X-rays. Chest X-ray type II was also found in all other classes; however, in class 1 predominantly patients with chest X-ray type I were found. Patients with chest X-ray type 0, III, and IV were classified mostly as having nonacute disease (classes 4 through 6), whereas patients with chest X-ray type I were classified mostly as having acute disease (classes 1 through 3; see Table 2). Half of the patients with chest X-ray type II were classified as having acute disease and the other half as having nonacute disease. Thus, chest X-ray type II exhibits the widest distribution over the six SCAC classes.
Patients with Acute Disease More Often Have Extrapulmonary Involvement Than Do Patients with Nonacute Disease Patients with acute disease more often had extrapulmonary involvement (71.3%) than did patients with nonacute disease (26.3%) (P < 0.0001; statistically significant). Only half of the patients with acute disease had Löfgren's syndrome. Patients with nonacute disease and extrapulmonary involvement tended to have skin involvement, whereas patients with acute disease were affected more frequently with extensive adenopathy, arthritis, and renal or liver involvement (see Table E2). The percentage of patients with an indication for steroid treatment was lower in the acute disease group than in the nonacute disease group (P < 0.0001).
Analysis of Pulmonary Function Data
Analysis of BAL Fluid Cell Composition Patients with sarcoidosis chest X-ray type IV had a statistically significant increase in neutrophil and macrophage BAL fluid cell counts, whereas BAL fluid CD4+ cell counts tended to be decreased compared with those of patients with all other chest X-ray types (Figure 2 and Figure E3). Patients with chest X-ray type III also had lower BAL fluid CD4+ cell counts compared with patients with chest X-ray type I. If data were analyzed by applying the classification system based on clinical criteria, patients with nonacute disease and long-term treatment (class 6) had higher neutrophil cell counts than other patient classes, but these differences did not reach statistical significance (see Figure 2). BAL fluid CD4+ cell counts tended to be higher in patients with acute disease compared with patients with nonacute disease, but the difference did not reach statistical significance (see Figure E3). BAL fluid macrophage cell counts from nonacute patients with the need for treatment (classes 5 and 6) were significantly lower than those from patients with no need for treatment (classes 1 and 4).
Analysis of Serologic Markers In the analysis of serologic markers the differences between patient subgroups according to clinical classes were more clearly detectable than in the patient subgroups according to chest X-ray types. sIL-2R and neopterin concentrations were dramatically increased in patients with acute disease and the need for long-term treatment (Figure 3). Patients with nonacute disease and a need for treatment also had a statistically significant increase in serum sIL-2R and neopterin concentrations, but the levels were less than those from patients belonging to class 3. In contrast to sIL-2R and neopterin, we could not obtain any statistically significant differences in ACE concentrations regarding the different classes (Figure E4 in the online supplement). Only patients with sarcoidosis chest X-ray type 0 had significantly lower serum ACE concentrations compared with patients with chest X-ray types I and II (see Figure E4). CRP levels were also significantly different between the SCAC classes (see Figure E4). Patients with acute disease had higher CRP concentrations than did patients with nonacute disease.
The heterogeneity among patients with sarcoidosis is well known. Several authors have therefore suggested that there might be more than one causal agent involved in the pathogenesis of sarcoidosis (23). Advances in genomic research have opened up the possibility to evaluate gene variants possibly defining the susceptibility to sarcoidosis (10–12, 17, 18, 24, 25). Because of the well-known heterogeneity of sarcoidosis, a system that is able to categorize patients according to clinical phenotype is urgently needed for genotype–phenotype correlation. Furthermore, because the majority of patients with sarcoidosis will have a spontaneous resolution of the disease, there is a need for early identification of patients with chronic ongoing disease, who will require steroid treatment. Genotyping might be promising in the identification of these patients. Predicting the clinical course on the basis of genotyping is in its infancy (10). Therefore, to establish clinically practical genotype-based risk profiles a method of phenotyping is required that separates disease courses with different clinical consequences. The presented SCAC classes are a first approach. Because several authors have described a clinical difference as well as a genomic difference between patients who have Löfgren's syndrome (acute diseases of different types) and patients who do not have Löfgren's syndrome, we also share the assumption that this category is fundamental (17, 18). Thus, we applied a broad definition of Löfgren's syndrome, which we termed acute disease. In line with this, we first differentiated between patients with an acute onset of the disease with fatigue, subfebrile temperature, and weight loss and patients with a subacute onset of the disease, which normally cannot be exactly dated by the patients. In addition, to identify patients with chronic, recalcitrant disease, we classified patients further according to the necessity for or response to treatment with steroids. In total, we obtained six different patient classes: three classes describe patients with acute onset and three describe patients with nonacute onset. Using this method, it was possible to classify every patient included in this study. From this we conclude that our proposed classification system is clinically practicable in a routine setting. Interestingly, the patients in the classes of acute disease experienced nonpulmonary organ involvement significantly more often than did patients with nonacute onset. Furthermore, these patients with acute onset had less often a need for long-term treatment than did patients with nonacute onset, and the indication to treat was due predominantly to extrapulmonary organ involvement. In summary, we believe that our data show that there are great differences between both of these patient groups, which justifies our classification protocol. Our study shows that the subtypes divided by chest X-ray according to the system proposed by Scadding (20) properly describe the various subcohorts of patients with sarcoidosis. In particular, patients with fibrotic remodeling were clearly identified, which is correlated with fibrogenic mechanisms as shown in a publication by our group (27) and by others (26, 27). Patients with sarcoidosis type IV disclosed a statistically significant decrease in pulmonary function data. The change in PO2 showed a decrease with the radiologic type of the disease, in line with published data from Medinger and coworkers (28) and Keogh and Crystal (29). Furthermore, BAL fluid cell composition showed a statistically significant decrease in CD4+ lymphocyte counts and an increase in BAL fluid neutrophil cell count in sarcoidosis type IV. However, the number of patients with sarcoidosis type IV was extraordinarily low, comprising only 2% of our study cohort and further analysis did not show any statistically significant differences between patients with sarcoidosis type II and type III, either regarding pulmonary function data or BAL fluid cell composition. We clearly demonstrate that radiologic typing does not correlate with clinical phenotypes. Patients with nonacute recalcitrant disease (class 6) were considerably more numerous than the patient cohort with chest X-ray type IV. All patients with chest X-ray type IV were properly classified as having chronic recalcitrant disease by our classification system. However, class 6 consisted, in addition, of patients with chest X-ray types 0, I, II, and III. Patients with advanced fibrotic disease were clearly the minority in this class. Therefore, we conclude that the majority of the patients with complex disease courses were not identified by chest X-ray typing. In the analysis of our patient cohort according to the classification protocol proposed by this study, we obtained several statistically significant differences between the patient classes. We observed that the distribution of pulmonary function data was different among classes. Patients with nonacute onset and a need for long-term treatment (class 6) had the most severe pulmonary function data. Although all patients classified as belonging to class 3 (acute onset and chronic disease course) had extrapulmonary indications for immunosuppressive treatment, pulmonary function data for this group indicated worse functioning compared with the data for patients classified as belonging to class 1, 2, or 4. However, the fact that clinical decisions are partially based on pulmonary function data might limit these findings, but with the current treatment strategy this is inevitable. Our data suggest that patients with a complex disease course and symptoms of acute disease might also have a risk for pulmonary disease progression and are in line with data by Baughman and coworkers (4), who showed that patients who need treatment (independently of the indication to start therapy) had lower pulmonary function data compared with patients with an indication for treatment. Drent and coworkers (30, 31) and we (32) have shown that BAL fluid neutrophil cell counts are increased in patients with progressive and advanced disease. Our data show that especially patients with chest X-ray type IV had increased neutrophil cell counts. BAL fluid CD4+ lymphocyte counts tended to be increased in patients with acute sarcoidosis compared with patients with nonacute disease, but the difference did not reach statistical significance. This finding is in line with previous observations that patients with Löfgren's syndrome have higher BAL fluid CD4+ lymphocyte counts compared with other patients (30–34). Furthermore, our data are in line with reports suggesting that neither CD4+ BAL fluid cell counts nor the CD4+ cell/CD8+ cell ratio is able to predict the course of the disease (30–34). Therefore, these data again indicate that the clinical classification system worked properly.
Patients with an acute form of sarcoidosis had significantly increased concentrations of serum CRP, neopterin, and sIL-2R. These findings suggest that patients with acute disease have an elevated acute-phase response and signs of systemic disease. In particular, sIL-2R concentrations were highly increased in patients with an acute form of sarcoidosis and a requirement for treatment with steroids (classes 2 and 3). Serum sIL-2R concentrations were extraordinarily high in patients from class 3 and were 10-fold higher than in the other classes. The magnitude of this difference in sIL-2R concentrations was unexpected and indicates a possible role for sIL-2R as a marker of disease activity and the need for treatment in patients with acute disease. However, this finding might be limited by the fact that in class 3 only five patients could be included. Patients with nonacute sarcoidosis who required treatment also had increased sIL-2R serum concentrations compared with patients with nonacute disease and no need for treatment (class 4), but the difference was less compared with patients from class 3. The differences in sIL-2R concentrations between the classes of patients with nonacute disease were significant but less impressive compared with the differences among patients with acute disease. sIL-2R is described as a marker of disease activity in sarcoidosis (32, 35). Grutters and coworkers (36) have shown that sIL-2R concentrations decreased with increasing chest X-ray type. The authors hypothesized from their findings that patients with extrapulmonary involvement may have increased sIL-2R. Our data confirm this hypothesis, for the first time, insofar as sIL-2R concentrations were particularly increased in patients with active sarcoidosis and extrapulmonary organ involvement. Furthermore, we demonstrate that patients with acute disease had a higher acute-phase response than did patients with nonacute disease. Our data suggest that sIL-2R concentrations need to be adjusted and interpreted in the context of the clinical phenotype of the individual patient. By this approach, sIL-2R serum concentrations might gain relevance in everyday practice. Similar to sIL-2R, neopterin was increased in both classes of patients, either with acute or nonacute disease, who had a need for long-term treatment (classes 3 and 6). Neopterin is produced mainly by macrophages after IFN- In conclusion, our study introduces a new classification protocol based on clinical criteria as a tool for further studies evaluating disease mechanisms to be correlated with the clinical outcome. Analysis of 225 patients demonstrated its practicability by confirming several previous findings regarding pulmonary function data and BAL fluid cell composition. Although Scadding's classification system based on chest X-ray findings is able to identify the small group of patients with sarcoidosis with severe fibrotic lung remodeling, the classification protocol based on clinical criteria identified clinically complex patients with the need for long-term treatment due to high disease activity either from granulomatous disease or fibrotic remodeling. The correlation of clinical phenotypes with genetic markers will result in genetic markers, which will be able to define patients at risk for relapse and outcome. With this information clinicians might be able to adjust their treatment protocols, such as by choosing an immunosuppressive treatment combination, for patients identified as having a poor prognosis. On the other hand, patients identified as having a low risk of progressive disease will benefit if treatment implementation is delayed. Our findings demonstrate that the described SCAC protocol gives additional information not yet acquired by chest X-ray typing, which makes it to a promising tool for further studies evaluating genetic influence, pathomechanisms, and biomarkers.
The authors thank Dr. Claudia Kurz and coworkers from the Department of Radiology of the University Hospital Freiburg (Freiburg, Germany) for chest X-ray typing.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200705-742OC on November 1, 2007 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form May 18, 2007; accepted in final form October 26, 2007
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
PO2) was calculated as the difference in arterial oxygen tension between peak exercise and resting measurements. 
stimulation, whereas sIL-2R is produced mainly by activated lymphocytes (
