Short- and long-term response to corticosteroid therapy in chronic beryllium disease

Chronic beryllium disease (CBD) is a granulomatous disorder that affects the lung after exposure to beryllium. The present study reports short- and long-term evolution of granulomatous and fibrotic components in eight patients with severe CBD receiving corticosteroid therapy. Eight patients with confirmed CBD were studied at baseline, after initial corticosteroid treatment (4–12 months), at relapse and at the final visit. Beryllium exposure, Glu69 (HLA-DPB1 genes coding for glutamate at position β69) polymorphism, symptoms, pulmonary function tests (PFT), serum angiotensin-converting enzyme (SACE) and high-resolution computed tomography (HRCT) quantification of pulmonary lesions were analysed. The CBD patients were observed for a median (range) of 69 (20–180) months. After stopping beryllium exposure, corticosteroids improved symptoms and PFT (vital capacity +26%, diffusing capacity of the lung for carbon monoxide +15%), and decreased SACE level and active lesion HRCT score. In total, 18 clinical relapses occurred after the treatment was tapered and these were associated with SACE and active lesion HRCT score impairment. At the final visit, corticosteroids had completely stabilised all parameters including both HRCT scores of active lesions and fibrotic lesions in six out of eight patients. Corticosteroids were beneficial in chronic beryllium disease. They were effective in suppressing granulomatosis lesions in all cases and in stopping the evolution to pulmonary fibrosis in six out of eight patients.

C hronic beryllium disease (CBD) is a granulomatous disorder that mainly affects the lung after exposure to beryllium (Be). Susceptibility to CBD has been associated with variants of the HLA-DPB1 gene coding for glutamate at position b69 (Glu 69 ) [1].
The primary process of CBD is characterised, at the histopathological level, by an interstitial mononuclear cell inflammation and non-necrotising epithelioid cell granulomas in the lung. Over time, areas of fibrosis surround the granulomas. The development of extensive pulmonary fibrosis is a major complication of CBD and the main cause of mortality [2]. Medical literature suggests that, unlike sarcoidosis, once CBD is clinically overt it continues to progress even after a long interruption of Be exposure [3]. This observation is probably linked to the low-level clearance of Be and its particularly prolonged persistence in the lung [4].
In CBD, the inflammatory responses to Be can establish a state of acquired Be antigen-specific, cell-mediated immunity, as indicated by an abnormal lymphocyte proliferative response to Be (BeLPT) on peripheral blood or bronchoalveolar lavage (BAL) cells [5]. Medical treatment of CBD is aimed at controlling or reducing the immunological reaction to Be. In addition to stopping exposure, corticosteroid therapy is the cornerstone of treatment [6]. Previous clinical series have shown the efficacy of corticosteroids in improving symptoms and pulmonary function [7][8][9][10][11][12][13][14]. However, nowadays, BeLPT is necessary to diagnose CBD, and previous reports may have included some cases with diseases different to CBD, such as sarcoidosis.
The aim of the present retrospective study was to investigate the short-and long-term response to corticosteroid therapy on the granulomatosis and fibrotic components of CBD.

Patients
Eight patients with confirmed CBD were recruited and followed up by an investigator (D. Valeyre) in the Dept of Respiratory Medicine, Avicenne Hospital (Bobigny, France) between 1988 and 2004. Five of the patients have been reported previously [15]. Demographic data are given in table 1. All patients stopped smoking during the study. The diagnosis was based on the following criteria [16]: 1) occupational history showing exposure to Be; 2) evidence of noncaseating granulomatosis in tissue samples; and 3) positive BeLPT.

Investigations
Symptoms, physical examination, pulmonary function tests (PFT; see online supplementary data), chest radiography and serum angiotensin-converting enzyme (SACE) activity were recorded on the chart of each patient. The Be exposure level (see online supplementary data) and Glu 69 polymorphism encoded by HLA-DPB1 genes were also recorded.
High-resolution computed tomography scores All patients underwent high-resolution computed tomography (HRCT) of the lung at baseline, after 4-12 months of corticosteroid therapy and at the end of the study. Opacities such as ground-glass opacities, micronodules, nodules and alveolar consolidations were considered as active inflammatory granulomatosis lesions and quantified in active lesion score; while opacities such as linear opacities, traction bronchiectasis, lobular distortions, bulla formations, cysts and honeycombings were considered as fibrotic lesion expressions and quantified in fibrosis lesion score (see online supplementary data).

Treatment
All patients received oral corticosteroids (prednisone) at mean initial daily dosage schedules of 0.8 mg?kg -1 ?day -1 (40-70 mg?day -1 ), with bisphosphonate as prevention of osteoporasis (when authorised for use in therapeutics). Treatment was started within 4 weeks of baseline. Further adjustment of corticosteroid therapy was based on the response of pulmonary symptoms, radiology and PFT. Occupational exposure to Be was ceased for all patients.

Study design
Response to corticosteroid therapy was evaluated according to clinical, radiological, functional and SACE data. Improvement (defined by a clear decrease of abnormal clinical, radiological and functional manifestations) allowed the reduction of corticosteroid treatment. Conversely, corticosteroids were increased again when two of the following criteria were present: worsening of clinical symptoms, chest radiograph opacities or PFT. The results at baseline and at the first improvement following initial treatment (4-12 months) were compared. Finally, data obtained at the final visit were  analysed. The relapse study design is available in the online supplementary data.

Statistical analysis
A nonparametric method was used, and all data are expressed as median (range). The Wilcoxon signed-rank test was performed to evaluate the significance of changes after treatment. A p-value ,0.05 was considered significant.

Data at baseline
Demographic, clinical, radiological, biological and Be exposure data at baseline are reported in tables 1 and 2. All patients had dyspnoea (level oII of New York Heart Association score) and weight loss (-9 (2-13) kg). Four patients had fever (.38.5uC) with cough, and three patients had chest pain. Physical examination revealed crackles in six patients and digital  clubbing in two. The tuberculin test was negative in all six patients tested.
SACE level was abnormal in all patients (90 (44-262 IU?mL -1 ), normal level ,41 UI?mL -1 ). Noncaseating granulomas were found in lung biopsies from seven patients (video-thoracoscopy: n55; transbronchial biopsy: n52) and in a liver biopsy from the remaining patient. Patients had a positive BeLPT in peripheral blood (five out of eight patients) and/or BAL (four out four patients).

Evolution after 4-12 months of corticosteroid treatment
The maximal clinical and functional improvement was reachedafter 4-12 months of corticosteroid treatment. Dyspnoea decreased in all patients (New York Heart Association classification improved by one level), and all other clinical findings disappeared (fever, cough, and weight loss). Crackles disappeared in four out of six patients.

Relapses after decrease of corticosteroids
Results concerning the CBD relapses in the eight patients are available in the online supplementary data.

End of the study period
The patients did not present any serious complications requiring treatment to stop. Cushing's syndrome was noted in patient seven.
The steroid threshold dose decreased with time. All patients were still treated with prednisone (12 (7-60) mg?day -1 ) for a median (range) period of 69 (20-180) months. The SACE level was normal. DL,CO and HRCT scores ( fig. 3) were stabilised in all but two patients. By contrast, in patients four and five, DL,CO decreased despite corticosteroid treatment ( fig. 3a). On HRCT, these patients presented a direct increase of the fibrosis score compared with other patients (fig. 3b). Interestingly, the appearance of their fibrotic lesions was particular with diffuse multicysts and thin reticulation with honeycomb lesions similar to those of idiopathic pulmonary fibrosis ( fig. 4).

DISCUSSION
The present study has shown that corticosteroid therapy at high doses makes the control of active granulomatosis inflammation possible and achieves a clear response of symptoms, PFT, HRCT and SACE in severe CBD. Over the long term, steroid treatment prevented the decline of pulmonary function in six patients by preventing fibrosis progression; however, it did not modify the fibrotic lesions pre-dating corticosteroid initiation. In two cases, despite clear short-term responses to corticosteroids, the issue with time was different with the progressive development of serious fibrotic lesions.
Many clinical series and case reports of CBD have indicated that corticosteroids were effective in reducing symptoms and improving lung function [7][8][9][10][11][12][13][14]. However, in these earlier reports, BeLPT tests were not practised, although this immunological assessment leads to a greater specificity of the diagnosis. In addition, the case definition was prone to potential misclassification, and some cases may have followed a different course as they had different aetiology, such as sarcoidosis. In the present study, all patients had confirmed pulmonary CBD, based on a new gold standard [17]. A recent study examined the effect of long-term corticosteroid therapy on the natural history of six patients with CBD confirmed by a positive BeLPT [18]. The current findings are more relevant because of the radiological distinction between granulomatous inflammation and irreversible fibrosis by using both HRCT scores and SACE measurement and a longer survey period. The same classification as sarcoidosis has been used [19], and the lesions separated into two categories: 1) active granulomatosis  inflammatory lesions (ground-glass opacities, micronodules and alveolar consolidations); and 2) pulmonary fibrotic lesions (linear opacities, traction bronchiectasia, bulla formations and honeycombings).
The current study showed the importance of inflammatory granulomas in the seriousness of baseline presentation and relapses. The disease presentation at baseline was invariably severe and needed steroid treatment. With corticosteroid therapy, the improvement of symptoms and PFT were associated with a decrease of the active lesion score on HRCT and SACE, a marker of macrophage activation. DANILOFF et al. [20] have previously shown that HRCT abnormalities (nodule, ground-glass opacities) were correlated with lymphocytosis in BAL and suggested that HRCT was a meaningful index of the inflammatory activity in CBD. The response to corticosteroids was not examined in the present study.
Before corticosteroid initiation, patients presented with initial linear and reticular opacities and bulla formation suggesting early fibrotic lesions. This was probably related to the natural evolution of several granulomas to fibrosis. STOECKLE et al. [13] observed that, in patients with CBD, pulmonary nodular and granular densities on chest radiographs tended to be replaced by linear densities and lung retraction. The delay between the onset of the illness and corticosteroid therapy favoured fibrosis emergence [13]. In a recent autopsy study [2], all patients showed end-stage interstitial fibrosis with varying degrees of cystic and honeycomb change. Little is known about the mechanisms that govern the transition from granulomatous inflammation to increasing fibrosis in CBD. In some CBD patients, histology revealed that an increasing number of fibroblasts infiltrated around granulomas [17,21].
Corticosteroids were more likely to suppress the growth of existing granulomas and to prevent the development of new ones, and were less effective against fibrotic lesions. STOECKLE et al. [13] described several clinical cases where treatment with steroids appeared to have interrupted the progressive fibrosis. The current study patients were followed after steroid onset for a median duration of 69 months. Six of the study patients demonstrated a persistent initial benefit in terms of pulmonary function and radiological opacity. In particular, fibrosis opacities on HRCT did not worsen in these six patients. Corticosteroids seemed to limit the progressive fibrosis.
However, it is known that the rate of progression of CBD is variable [13,18,22]. The present data indicate that after initiating corticosteroid treatment, pulmonary fibrosis did not progress in six cases while it clearly developed in the other two. This worsening of fibrosis lesions may be the consequence of an insufficient control of insidious granulomatous lesions. However, it may also be the consequence of an evolution of fibrosis. FERRIS [23] suggested that the nonreversible fibrosis may develop despite suppressed inflammation by steroids. Several parameters may have participated in the progression to pulmonary fibrosis in the two patients. One of these is the total survey period, as patients with progressive fibrosis were surveyed for a longer time. Although these two patients were Glu 69 homozygous, it was not possible to draw a conclusion about the role of this factor in the progression to fibrosis owing to the small number of patients ( fig. 3).
In conclusion, corticosteroid treatment in patients suffering from serious chronic beryllium disease improved symptoms, pulmonary function tests and radiology by acting on inflammatory granulomas. The control of inflammatory granulomatosis limited the fibrotic evolution as long as doses were monitored under the control of clinical examination, serum angiotensin-converting enzyme and highresolution computed tomography scaning. However, corticosteroids seemed insufficient to stop this poor evolution for some patients. The understanding of the mechanisms that govern the fibrosis in chronic beryllium disease could help to control its emergence in other granulomatosis, such as sarcoidosis.
b) a)