INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Sarcoidosis is a systemic granulomatous disease of unknown origin that is characterized by an accumulation of activated, proliferating T lymphocytes and mononuclear phagocytes in the involved organs, most commonly the lungs. Immunologic studies done with cells obtained by bronchoalveolar lavage (BAL) have considerably improved our knowledge of the immunopathogenesis of sarcoidosis (1, 2). Nowadays, the most widely accepted hypothesis is that a T-helper cell/macrophage alveolitis in pulmonary sarcoidosis is caused by an unknown stimulus activating alveolar macrophages (AM) and lymphocytes to release mediators such as tumor necrosis factor- (TNF-) (3) and interleukin-2 (IL-2) (4), which subsequently recruit and activate other inflammatory cells. A typical feature of pulmonary sarcoidosis is an increase in the percentage of BAL lymphocytes with an accumulation of T-helper cells in the lung, resulting in an increased CD4⁺/CD8⁺ ratio of BAL lymphocytes. Although T-cell and mononuclear-cell activation is restricted to the involved organs (4), sarcoidosis is regarded as a systemic disease. This systemic nature of sarcoidosis is reflected by the fact that it may involve most organ systems. Several serum markers of sarcoidosis (e.g., neopterin [7], angiotensin-converting enzyme [ACE] [8], and soluble IL-2 receptor (sIL-2R) [9, 10]) exist and are used as staging parameters. Although all of these parameters are closely related to the pathogenesis of the disease, insufficient evidence is available about which of them are suitable for predicting clinical deterioration.

The prognosis in sarcoidosis is relatively good, and in approximately 60% of patients a spontaneous remission can be observed (11). However, other patients develop a progressive form of the disease that may result in lung fibrosis and, eventually, death. Despite improved knowledge of the pathogenesis of sarcoidosis, its natural course is still unpredictable in an individual patient. Therefore, prognostic parameters would be very helpful for preventing severe organ damage in the disease.

The present study was designed to investigate BAL (percentage of BAL lymphocytes, CD4⁺/CD8⁺ lymphocyte ratio) and serologic parameters (neopterin, ACE, and sIL-2R) in order to identify those parameters that could be used to predict disease progression in sarcoidosis. Since TNF- is thought to play a key role as a mediator of inflammation in sarcoidosis, we also tested whether the spontaneous release of TNF- from cultured AM obtained from sarcoidosis patients could be a useful parameter for predicting the course of the disease.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

The diagnosis of sarcoidosis was established in 77 patients in accordance with previously defined criteria (1, 2), including noncaseating granulomas identified by transbronchial biopsy. Special stains and cultures for fungus or mycobacteria were negative in BAL fluid (BALF) as well as in tissue specimens. None of the patients received therapy at the time of investigation, nor had they within the previous 2 mo. For the purpose of this prospective study, the patients were allocated to two groups: patients with (Group A) or without (Group B) indications for prednisolone therapy. The two groups did not differ in age, sex, or smoking history (p > 0.2). The characteristics of the study population are summarized in Table 1. In most cases, pulmonary function test results for periods of 3 to 6 mo were available from the referring physicians at the time of BAL. Eight patients who directly sought the service of our referral center were monitored for 3 to 6 mo unless mandatory indications required immediate initiation of therapy. The decision to treat was based on described clinical criteria (11) indicating manifestations of organ damage (i.e., progressive abnormalities in pulmonary function test results or diffusing capacity of carbon dioxide [DL_CO]). In brief, sarcoid patients with evidence of deterioration of pulmonary function during the 3 mo preceding the study, or with an extrapulmonary manifestation requiring therapy (cardiac, neurologic, or ophthalmic involvement or hypercalcemia) were treated with corticosteroids (Group A). The initial dosage was 40 to 50 mg prednisolone in a single daily dose, reduced gradually every 4 wk to a maintenance level of 10 to 15 mg over a period of 6 mo. Patients with no objective evidence of deterioration in the 3 mo preceding the study, and with no extrapulmonary manifestation of sarcoidosis requiring therapy, were not treated but were closely monitored (Group B). Although 31 of 40 patients in Group B had signs of disease activity (e.g., cough or dyspnea, impairment of lung function parameters, radiographic thoracic manifestations of sarcoidosis, or serologic signs of disease activity), they were not treated because mandatory indications for corticosteroid therapy were lacking (e.g., objective evidence of recent deterioration of pulmonary function or extrapulmonary manifestations requiring therapy). This finding supports the concept that functional immune parameters do not necessarily reflect the clinical status in sarcoidosis. The monitoring rather than treatment of patients in Group B is in accord with recommendations of the third meeting of the World Association of Sarcoidosis and Other Granulomatous Disorders (WASOG) (12), that the extent of active disease needs to be distinguished from the total organ damage. Furthermore, disease activity should not be mistaken for a need to start corticosteroid therapy because active, acute sarcoidosis resolves spontaneously in a high percentage of patients. Nevertheless, these patients need to be closely monitored because an ongoing immunologic process may result in disease progression requiring initiation of corticosteroid treatment.

Follow-up Study

In all, 77 patients included in the study had their lung-function parameters (i.e., VC, TLC, FEV₁, DL_CO, and Pa_O2) reevaluated after 5.7 ± 0.4 mo mean ± SEM. For TLC and DL_CO, a change of 10% from the baseline value and for VC and FEV₁ a change of 15% from the baseline value was regarded as significant. For Pa_O2, a significant change was defined as ± 15 mm Hg if the baseline values was > 80 mm Hg; ± 10 mm Hg if the baseline value was 55 to 80 mm Hg; and ± 5 mm Hg if the baseline value was < 55 mm Hg. Progressive disease was defined as a significant decrease in two of these parameters or a significant change in one parameter accompanied by a compatible worsening in the level of dyspnea. Stable disease was defined as disease with no change in pulmonary function or level of dyspnea. A change in one parameter without a compatible change in dyspnea was also defined as stable disease. Regressing disease was defined as either no change in the level of dyspnea, accompanied by a significant improvement in two or more parameters, or a significant improvement in one lung-function parameter accompanied by a compatible improvement in the dyspnea level (11). To detect these differences, only accurately performed pulmonary function tests, fulfilling American Thoracic Society (ATS) criteria (13, 14), were used to monitor the changes in lung function. None of the patients exhibited signs of obstructive lung disease. In Group A a relapse was defined as deterioration during the tapering of corticosteroids or within 3 mo after cessation of therapy. In Group B a deterioration that occurred within 6 mo after BAL was regarded as progressive disease.

Serologic Parameters

Neopterin was evaluated with a commercially available radioimmunoassay (RIA) kit (Henning, Berlin, Germany). The 95% normal range in our control group of 25 healthy volunteers was 2.3 to 9.8 nmol/L (median: 5.2 nmol/L), which was in accordance with the data given by the manufacturer. Values of neopterin greater than 9.8 nmol/L were regarded as high. Serum ACE was determined photometrically with a commercially available kit (ACE Colour Test; Fuijrebio, Tokyo). The 95% normal range in our control group was 8 to 23.5 U/L; values exceeding 23.5 U/L were regarded as high. sIL-2R was measured with a sandwich enzyme-linked immunosorbent assay (ELISA) employing two monoclonal antibodies directed against different epitopes of IL-2R (DPC; Biermann, Bad Nauheim, Germany). The 95% normal range in the serum of 50 healthy volunteers was 99 to 919 U/ml (median: 583 U/ml), which was in accordance with the data supplied by the manufacturer. Values greater than 1,000 U/L were regarded as high.

BAL and Preparation of Lung Mononuclear Cells

Bronchoscopy and BAL were performed as previously described (1). In brief, 200 to 300 ml of sterile saline (0.9% NaCl) were instilled into a lingula or middle-lobe segment in 25-ml aliquots. Each aliquot was immediately aspirated. Recovery was 63.1 ± 1.9% in Group A (therapy) and 64.2 ± 2.0% in Group B (no therapy). Cell differential counts were done on Cytospin preparations of at least 200 cells (Cyto-spin II; Shandon Instruments, Sewickley, PA). For immunoperoxidase staining, cells were fixed on poly-L-lysine-coated slides (Bio-Rad, Munich, Germany) and developed with a peroxidase-antiperoxidase (PAP) technique (15), using monoclonal antibodies against CD4, CD8, IL-2R (Ortho Diagnostic Systems; Neckargemünd, Germany) and human leukocyte antigen-DR (HLA-DR) (Becton Dickinson, Heidelberg, Germany). In accordance with the results of the BAL Cooperative Group Steering Committee (16), BAL lymphocyte counts of more than 15% and a CD4⁺/CD8⁺ lymphocyte ratio of more than 3.5 were regarded as elevated.

Cell Culture and TNF- Assay

Freshly isolated BAL immune cells were cultured without any stimuli at a density of 10⁶ cells/ml over a period of 24 h in endotoxin-free RPMI 1640 medium (Seromed, Berlin, Germany) supplemented with 2% human AB serum (kindly provided by the Blutbank Lütjensee, Lütjensee, Germany), 2 mM L-glutamine, 100 U/ml penicillin, and 100 µg streptomycin (all from GIBCO, Wiesbaden, Germany) in a humidified 5% CO₂ atmosphere at 37° C. At the end of the culture period the supernatants were harvested and stored at 70° C until assayed for TNF-. The viability of the cells after culture always exceeded 95% as determined by trypan blue exclusion.

TNF- in cell culture supernatants was measured with a previously described ELISA (17), with slight modifications (3). Antibodies to recombinant human TNF- (-rhTNF, clone 195) were kindly provided by Dr. E. Schlick of Knoll AG, Ludwigshafen, Germany. It was demonstrated for a wide range of TNF- concentrations released by AM either spontaneously or after in vitro stimulation that the TNF- content of the supernatants as measured by ELISA correlated with the bioactivity measured with the L-cell assay (r_s = 0.85; p < 0.001; n = 23; range: 20 to 5,000 pg/ml; data not shown). The TNF- level of our control group of 20 patients who underwent bronchoscopy for diagnostic reasons and who were retrospectively free of any inflammatory or malignant lung disease was 220 ± 37 pg/ml (range: 0 to 589 pg/ml, median: 218 pg/ml; Figure 1). For the purpose of this study, TNF- values of more than 600 pg/ml were regarded as elevated.

View larger version (24K): [in this window] [in a new window]   Figure 1.   Spontaneous release of TNF- by cultured BAL cells of controls and sarcoid patients. The mean is indicated by a horizontal bar and the normal range of TNF- release defined by the control group by the shaded area. Both study groups differ significantly (p < 0.005) from the control group, as indicated by asterisks.

Statistical Analysis

Data are expressed as mean ± SEM. Comparisons were made with the Mann-Whitney U-test. Differences between the groups were evaluated with the chi-square test. Correlations between different parameters were determined with Spearman's rank correlation coefficient. Values of p < 0.05 were regarded as significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Comparison of Clinical Features of the Two Study Groups

The 77 patients included in this study were divided into two groups: Group A (n = 37) with indications for prednisolone therapy at the time of BAL, and Group B (n = 40) with no indications for therapy. The groups did not differ significantly in age, gender, smoking history, or lung function (Table 1). Nevertheless, the VC, FEV₁, DL_CO, and Pa_O2 were lower in Group A, with indications for therapy. Interestingly, we observed the same trends as Hunninghake and coworkers (11) when subdividing the initial pulmonary function tests according to the further course of disease. In Group A, patients who improved with therapy tended to have a lower initial VC, FEV₁, and DL_CO than those remaining stable, although this difference was not statistically significant (data not shown). In Group B, those patients who deteriorated as a group tended to have poorer initial pulmonary function. This difference was close to statistical significance when these patients were compared with patients who improved (VC: p = 0.07; TLC: p = 0.08; FEV₁: p = 0.09; data not shown). The observed changes in lung function at follow-up, subdivided according to patients' clinical outcome are shown in Table 2. Those patients who deteriorated in Group A and B had significantly lower pulmonary function parameters than those who improved, and in part also when compared with those who remained stable.

The vast majority of patients (89.6%) had radiographic thoracic manifestations of sarcoidosis (Table 3). There were significantly more patients with Stage 0 and Stage I disease in Group B, with no indication for therapy (65% versus 37.8% in Group A; p < 0.05), and considerably more patients with Stage II disease in Group A (40.5% versus 20%). The sarcoidosis patients in our study population who deteriorated had disease of more advanced radiologic stages, resembling the data presented by Hunninghake and coworkers (11). The majority of patients suffered from symptoms attributable to their disease. Cough (43.2% versus 25%, p = n.s.) and dyspnea on exertion (46% versus 32.5%, p = n.s.) were more frequent in Group A, whereas the expression of other symptoms did not differ between the two groups.

Evaluation of BAL and Serologic Parameters

In Group A, a slight increase in BAL lymphocytes could be noted, whereas the CD4⁺/CD8⁺ lymphocyte ratio was higher in Group B (Table 1). The numbers of HLA-DR-positive cells did not differ between the two groups, but IL-2R-positive cells were slightly more numerous in Group A. None of these differences was of statistical significance.

The spontaneous release of TNF- by cultured BAL immune cells is shown in Figure 1. It was insignificantly greater in Group A (1,872 ± 428 pg/ml) than in Group B (1,561 ± 449 pg/ml), but in both groups differed significantly from that in the control group (220 ± 37 pg/ml, p < 0.005).

The serum levels of neopterin were significantly higher in Group A than in Group B (21.0 ± 4.5 nmol/L versus 9.3 ± 1.0 nmol/L, p = 0.010; Table 1). The ACE value was slightly increased in Group A (25.4 ± 2.7 U/L versus 22.2 ± 2.2 U/L, p = n.s.) as was the serum level of sIL-2R (1,870 ± 263 U/ml versus 1,274 ± 216 U/ml, p = n.s.).

Correlations Between BAL and Serologic Parameters

Since both TNF- and neopterin are regarded as markers of macrophage/monocyte activation, we were interested in whether there was a correlation between the TNF- spontaneously released by AM and the serum neopterin level in our study population. Neither in Group A (p > 0.5, r_s = 0.01) nor in Group B (p > 0.3, r_s = 0.16) could a correlation be established. There was no correlation between the serum sIL-2R level and the percentage of IL-2R-positive lymphocytes (LY) in either group (Group A: p = 0.17, r_s = 0.28; Group B: p > 0.5, r_s = 0.09). The other correlations between the investigated BAL and serologic parameters yielded no significant results (data not shown). The serum levels of sIL-2R and of ACE showed no correlation with each other (Group A: p = 0.27, r_s = 0.21; Group B: p = 0.22, r_s = 0.22).

Prognostic Parameters

In Group A, seven of 37 patients (18.9%) had a relapse. Four of them (10.8%) worsened during tapering of corticosteroids and required an increased dosage and prolonged treatment. The others deteriorated within 3 mo after cessation of corticosteroid treatment and required reinitiation of therapy. In Group B, nine of 40 patients (22.5%) had disease progression within 6 mo after BAL and required initiation of corticosteroid therapy. To seek markers that could be of prognostic value for predicting disease progression, we evaluated the BAL (Figure 2) and serologic parameters (Figure 3) named earlier, in correlation with the occurrence of disease progression. For this purpose the sarcoidosis patients were divided into subgroups with elevated ("high"; Figures 2 and 3, dark bars) or normal ("normal"; Figures 2 and 3, light bars) levels of the investigated parameters.

View larger version (38K): [in this window] [in a new window]   Figure 2.   Cytologic and functional BAL parameters. Sarcoidosis patients with indications for therapy (Group A) and without indications for therapy (Group B) were divided into subgroups either with high (dark bars) or normal (light bars) levels of the investigated parameters. Relapses during tapering of corticosteroids or after cessation of therapy in Group A, or progressive disease requiring therapy in Group B, served as endpoints as determined by a follow-up, and are given as a percentage on the ordinate. The numbers of relapses or of cases of progressive disease within the total number of patients in each subgroup is given at the tops of the bars. Group B patients with high levels of spontaneous TNF- release by AM had a significantly greater risk of disease progression (*p = 0.025) than those with normal TNF- values. In Group A, a high level of TNF- release was also associated with a higher risk of relapse, although this was not statistically significant (p = 0.4).

View larger version (39K): [in this window] [in a new window]   Figure 3.   Serologic parameters. Patients were grouped and evaluated as described in the legend for Figure 2. Sarcoidosis patients with no indication for therapy (Group B) and high levels of sIL-2R had a significantly greater risk of disease progression (*p = 0.017) than those with a normal sIL-2R level. In both groups, patients with a high serum level of neopterin had a greater risk of relapse or disease progression, although this difference was not statistically significant (Group A: p = 0.5; Group B: p = 0.3).

In Group A, 25.0% of the patients with increased production of TNF-, in comparison with only 7.7% of those with normal levels of TNF-, experienced a relapse (p = n.s.). In Group B, 43.8% of the patients with high levels of TNF- had disease progression within 6 mo after BAL, whereas only 8.3% of those patients with normal TNF- levels did so. This difference is statistically significant (p = 0.025). Consequently, those patients in Group B who deteriorated showed a significantly greater TNF- release by AM (3,725 ± 1,506 pg/ml) than those who remained stable or improved (933 ± 323 pg/ml, p < 0.05). In Group A, the TNF- release by AM was increased to a similar extent in both of these different clinical-outcome subgroups (1,682 ± 712 pg/ml and 1,916 ± 505 pg/ml, respectively). The cytologic BAL parameters investigated in the study disclosed no significant differences, but there was a slightly reduced risk of relapse in those patients with a high percentage of BAL lymphocytes (12.5% [high] versus 23.8% [normal]) or a high CD4⁺/CD8⁺ lymphocyte ratio (0% [high] versus 16% [normal]) in Group A. Thus, neither the CD4⁺/ CD8⁺ lymphocyte ratio nor the percentage of lymphocytes seem to be suited to the prediction of relapse or disease progression in sarcoidosis.

The most striking result with regard to the serological parameters examined in the study (Figure 3) is the fact that 42.1% of the patients in Group B who had a high sIL-2R concentration experienced disease progression, whereas none with a normal serum level of sIL-2R did. This difference is statistically significant (p = 0.017). Consequently, the initial serum sIL-2R was significantly elevated in those patients in Group B who deteriorated (2,289 ± 712 U/ml), as compared with showing normal values in those patients who remained stable or improved (936 ± 113 U/ml, p < 0.01). In Group A there were no significant differences between the initial serum sIL-2R and the clinical outcome because both subgroups had similar values (1,456 ± 584 U/ml and 1,939 ± 293 U/ml, respectively). Nevertheless, 19.0% of patients with high but only 11.1% of those with normal levels of sIL-2R had a relapse. The analysis of the other serologic parameters revealed no significant differences, but there was a trend toward a greater risk of a relapse or disease progression in those sarcoidosis patients with a high neopterin value (23.8% versus 8.3% in Group A and 36.4% versus 15.4% in Group B). The analysis of ACE levels yielded no correlation with the course of the disease.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The aim of the present study was to evaluate the predictive value of BAL and serum markers in patients with newly diagnosed sarcoidosis. We focused our interest on those patients who had no indication for corticosteroid treatment at the time of BAL (Group B), because for this group such parameters could be extremely helpful in determining the correct time point for initiation of therapy. We compared these patients with those who had received therapy according to generally accepted clinical criteria after BAL [11] (Group A). Both groups were subdivided into patients with high and those with normal levels of the chosen parameters in order to enable the evaluation of their predictive value.

In our study population we were able to demonstrate that patients with no indication for therapy (Group B) but with high levels of AM TNF- release had a significantly greater risk of disease progression than those with normal TNF- production. The sarcoidosis patients with indications for therapy and high levels of TNF- release also had a higher risk of relapse, although this trend was not statistically significant. This observation is in accordance with earlier findings (3, 18, 19) that AM of patients with active sarcoidosis spontaneously release greater amounts of TNF- than those of patients with inactive disease. In our study population, 24 of 37 patients (64.9%) with indications for therapy (Group A) had increased TNF- production, whereas in Group B only 16 of 40 patients (40%) showed an exaggerated release of this cytokine. However, in those patients with no clinical indication for therapy at the time of diagnosis, this parameter can serve as a prognostic guide, helping to identify a subgroup of patients with a significantly greater risk of deterioration in the following 6 mo (Figures 1 and 2). Our results corroborate earlier findings (12) demonstrating that functional immune parameters and clinical status may dissociate at the time of observation. In addition, we were able to demonstrate with this prospective study that disease activity in sarcoidosis, as reflected by exaggerated AM TNF- release, may predict the likelihood of disease progression in those patients without mandatory indications for therapy at the time of their presentation.

In a retrospective study, Pueringer and associates (20) compared TNF-, IL-2, and prostaglandin E₂ (PGE₂) production by AM with the clinical status and course of sarcoidosis during the 3 mo preceding BAL, and could not demonstrate a correlation between those functional immune parameters and any clinical finding. However, we were recently able to show that increased release of transforming growth factor- (TGF-) by AM is predictive for spontaneous remission in untreated patients (21). Moreover, increased IL-2 release by BAL cells, in combination with an increased percentage of BAL lymphocytes, denotes to some extent a future deterioration of pulmonary function (22). These findings indicate that functional immune parameters such as cytokine release may be able to predict the future course of sarcoidosis, instead of describing the patient's present or past clinical status. The observation of heterogeneous AM TNF- release in study groups with active and/or inactive disease (3, 20), with a good prognosis (21), and with or without therapy (20) led us to conduct the present study correlating this functional immune parameter with the course of sarcoidosis after diagnosis. As expected, Group A and Group B did not differ in TNF- release (Figure 1). Within the groups, however, increased TNF- release marked clinical courses with a higher risk of relapse in Group A and significantly higher risk of deterioration requiring therapy in Group B (Figure 2). Thus, the analysis of AM cytokine release can help to identify subgroups of sarcoidosis patients with different risk patterns for deterioration or failure of therapy. This information cannot be extracted from lung function tests (Table 1), as observed previously by other groups (20, 23). We suggest that this subgroup of sarcoidosis patients should be more closely monitored in order to prevent severe organ damage.

The role of the CD4⁺/CD8⁺ lymphocyte ratio as a prognostic marker in sarcoidosis is still a controversial topic. Some authors (24) claim that a high-intensity T-cell alveolitis indicates poor prognosis and functional deterioration. In contrast, others (25, 26) found a high CD4⁺/CD8⁺ lymphocyte ratio to be an indicator of good prognosis. Several investigators have focused their interest on the prognostic value of the BAL lymphocyte percentage and have obtained divergent results. Some suggest that an intensive lymphocitic alveolitis indicated by a high percentage of BAL lymphocytes predicts functional deterioration (27), whereas others suggest that high lymphocyte counts may herald an alleviation of the disease (25, 26). Hollinger and colleagues demonstrated that a very high percentage of BAL lymphocytes (> 35%) predicts a good steroid response (28). In our study as well as in others (24), neither the lymphocyte count nor the CD4⁺/CD8⁺ lymphocyte ratio yielded any prognostic information. We propose that the disparity in the findings of the different studies may at least in part be due to racial and sampling differences among the patients (29), which influence the mode of presentation of sarcoidosis, leading to differences in the immunologic characteristics of study groups.

Considering the serum levels of neopterin, ACE, and sIL-2R, only the latter was of significant prognostic value. In our study group, 42.1% of the sarcoidosis patients with no indication for therapy and increased serum levels of sIL-2R experienced disease progression, whereas none of those with normal serum levels did. Consequently, those untreated patients who deteriorated had a significantly increased initial serum level of sIL-2R as a group, whereas those who remained stable or improved had normal values. In Group A, both of the two clinical-outcome subgroups had an increased initial sIL-2R. Although it was not significant, a similar trend toward a higher risk of relapse in patients with a high sIL-2R could be observed in those patients with indications for therapy (Figure 3). These findings are in accordance with previous findings (30) that patients with active sarcoidosis have significantly higher values of sIL-2R than those with dormant or inactive disease. In addition, our study revealed that patients without mandatory indications for therapy at presentation, but with increased serum sIL-2R concentrations, have a significantly greater risk of disease progression. In contrast to previously published results (30, 31), we could not demonstrate a positive correlation between serum sIL-2R and ACE, corroborating earlier findings by our group (32). On the basis of our data, it can be concluded that the measurement of sIL-2R cannot be replaced by that of ACE, and that sIL-2R is a prognostic serum marker for the spontaneous course of sarcoidosis and may even guide therapeutic interventions.

The predictive values of serum neopterin and ACE are still under consideration; although some investigators suggest that a rising ACE level can predict radiographic relapses of sarcoidosis (33), others have demonstrated that ACE does not appear to be of great prognostic value (34, 35). In our study ACE did not yield prognostic information. Interestingly, patients in both Group A and Group B who had a high serum level of neopterin had a greater risk of deterioration, although this difference did not reach statistical significance. Both serum neopterin concentration and AM TNF- reflect the activation of cells of the macrophage/monocyte lineage. Therefore a correlation between these parameters could have been expected, but was not found. This indicates that an increased neopterin level is not due to the activation of AM, but rather to that of cells in other compartments of the body. Thus, TNF- measurement in tissue-culture supernatants cannot be replaced by measurement of serum neopterin.

In conclusion, we have been able to demonstrate that sarcoidosis patients with no clinical indication for prednisolone therapy at the time of diagnosis, and with high levels of spontaneously released AM TNF-, have a significantly increased risk of disease progression. We conclude that this parameter, obtained by BAL, not only reflects the activity of sarcoidosis but is also useful for predicting the future course of the disease. Additionally, we report that patients with no indication for therapy and high levels of sIL-2R also have a significantly greater risk of deterioration. It is important to mention that a sole increase in one of these parameters is no indication for corticosteroid therapy. However, we propose that this group of patients should be closely monitored in order to detect ongoing disease progression as early as possible, in order to prevent functional deterioration. Whereas the measurement of TNF- release in a cell-culture system is at present only suitable for a few centers, the serum marker sIL-2R can be easily used to detect the subgroup of patients with a significantly greater risk of progression of sarcoidosis.