Airway Obstruction in Bronchial Sarcoidosis* Outcome With Treatment Franck Lavergne, MD; Christine Clerici, MD, PhD; Danièle Sadoun, MD; Michel Brauner, MD; Jean-Paul Battesti, MD; and Dominique Valeyre, MD Study objective: Airway obstruction (AO) in sarcoidosis is reported to be associated with respiratory symptoms, increased morbidity, and an increased mortality risk. Because AO in sarcoidosis may result from several causes, the therapeutic benefit of corticosteroids is difficult to determine. The aim of this study was to evaluate the therapeutic response of AO attributable to sarcoid granulomas in the bronchial wall. Patients: We selected 11 patients who had sarcoidosis with AO (defined as FEV 1 /vital capacity [VC] < 70%) associated with sarcoid granulomas on an endobronchial biopsy. Exclusion criteria were history of asthma, smoker or exsmoker, stage 4 disease, evidence of extrinsic compression by enlarged lymph nodes, and localized endobronchial stenosis seen during fiberoptic bronchoscopy. Interventions: We compared the results of pulmonary function tests and clinical, radiologic, and biological findings at baseline with those obtained at the time of the last pulmonary function tests available, between the sixth and 12th months of treatment. Eight patients took oral corticosteroids (20 to 60 mg/d initially), one received IV methylprednisolone pulses, another took oral hydroxychloroquine, and the last one received IM methotrexate. Measurements and results: With treatment, FEV 1 and FEV 1 /VC significantly improved in eight patients (72%), normalized in four patients, and was unchanged in the remaining three patients. The mean FEV 1 increased from 60.8 10.8% to 76 13.7% of the predicted value (p < 0.02). VC did not change significantly. FEV 1 /VC increased from 76.1 6.4% to 87.6 10.7% of the predicted value (p < 0.01). Dyspnea on exertion and other clinical findings were attenuated in 10 patients; the chest radiograph improved in 9 patients, and normalized in 5 patients. The mean serum angiotensin-converting enzyme level decreased from 112 48 to 58 40 IU/mL (p < 0.05), and normalized in four patients. Conclusion: The present study indicates that AO caused by sarcoid granulomas in the bronchial wall can be either partially or completely reversed by treatment with a concomitant attenuation of pulmonary symptoms. (CHEST 1999; 116:1194 1199) Key words: bronchi; obstructive lung diseases; respiratory function tests; sarcoidosis; treatment outcome Abbreviations: AO airway obstruction; SACE serum angiotensin-converting enzyme; VC vital capacity *From Service de Pneumologie (Drs. Lavergne, Sadoun, Battesti, and Valeyre), Service de Physiologie (Dr. Clerici), and Service de Radiologie (Dr. Brauner), Hôpital Avicenne (Assistance Publique-Hôpitaux de Paris), UFR Bobigny, Bobigny, France. Manuscript received December 1, 1998; revision accepted June 9, 1999. Correspondence to: Dominique Valeyre, MD, Service de Pneumologie, Hôpital Avicenne, 125, route de Stalingrad, 93009 Bobigny Cédex, France; e-mail: dominique.valeyre@avc.ap-hopparis.fr In pulmonary sarcoidosis, airway obstruction (AO) is reported in 5 to 63% of cases, 1 9 depending on the criteria of obstruction used by different authors. In a longitudinal study, AO was considered to be the most common presenting pulmonary physiologic abnormality in sarcoidosis 7 and occurred at all stages, with its frequency increasing from stage 1 to stage 4. 1 Although AO was mainly the consequence of small airway disease, large airways were also involved, as shown by an FEV 1 /vital capacity (VC) ratio 70% in 25% of the cases. AO in sarcoidosis is reported to be associated with increased morbidity, respiratory symptoms, 10 and increased mortality risk: Viskum and Vestbo 11 demonstrated that patients with an FEV 1 /VC ratio 70% of the predicted value had an increased mortality risk of 1.9 compared with patients with an FEV 1 /VC 70% of the predicted value. AO in sarcoidosis may be observed in different situations 12 : it may result from narrowing of the 1194 Clinical Investigations
bronchial wall due to either granulomatous lesions or their resulting fibrotic scarring, 13 18 compressive enlarged lymph nodes, 14,16 or airway distortion attributable to pulmonary fibrosis. 3,4 Until now, in spite of the poor prognosis of AO, few studies have evaluated its evolution during treatment. In fact, most authors have examined nonhomogeneous groups of patients with AO resulting from different mechanisms, 19,20 which makes drawing definite conclusions difficult. 21 Because the therapeutic response probably depends on the cause of AO, the aim of this study was to consider the effect of treatment exclusively on AO attributable to bronchial sarcoid granulomas. Subjects Materials and Methods Eleven subjects were retrospectively selected (see inclusion and exclusion criteria below) from among 300 patients with sarcoidosis who were followed in the Department of Respiratory Medicine, Avicenne Hospital, Bobigny, France. Demographic data are given in Table 1. The diagnosis was established on the following: (1) clinical and radiographic signs compatible with sarcoidosis; (2) evidence of noncaseating granulomas in one or several tissue samples; and (3) exclusion of other possible causes of granulomatous disease, particularly tuberculosis. Study inclusion criteria were as follows: (1) the presence of AO, ie, obstructive ventilatory defect, which was defined as FEV 1 /VC 70%, without bronchodilator response 22 ; and (2) the presence of sarcoid granulomas in an endobronchial biopsy specimen. Exclusion criteria were history of asthma, smoker or ex-smoker, stage 4 disease (ie, fibrosis), evidence of extrinsic compression by enlarged lymph nodes, and/or localized endobronchial stenosis noted at fiberoptic bronchoscopy. Study Design To evaluate the effect of treatment, we compared the results of pulmonary function tests at baseline (ie, just prior to the initiation of treatment) with those obtained at the time of the last pulmonary function tests, between the sixth and 12th months of treatment. Clinical, radiologic, and biological data were also considered. Pulmonary Function Testing FEV 1 and relaxed VC were measured with a spirometer (Gould Instrument Systems, Inc; Valley View, OH). The functional residual capacity was determined using a standard helium dilution method. All values were expressed at body temperature, barometric pressure, and saturated with water vapor under these conditions; and as percentage of predicted normal values determined by the European Respiratory Society, according to sex, age, height, and ethnic group. 22 Radiologic Examination Table 1 Baseline Clinical and Radiologic Characteristics and SACE Level for the 11 Patients Studied* Radiographic stages of the disease were defined as follows: stage 1 intrathoracic lymphadenopathy alone; stage 2 intrathoracic lymphadenopathy and pulmonary infiltration; stage 3 pulmonary infiltration alone with no signs of fibrosis; and stage 4 pulmonary fibrosis. Seven patients underwent a highresolution CT scan of the lung at baseline and after 6 to 12 months of treatment. Two of them also underwent a localized helical CT scan with 2-mm thick slices every 2 mm, to assess luminal abnormalities of the bronchi. 23,24 Treatment Characteristics Data Age, yr Mean SD 42 9 Range 28 57 Sex Male 3 (27) Female 8 (73) Ethnic group White 6 (55) Black 5 (45) Duration of sarcoidosis 2 yr 8 (73) 2 yr 3 (27) Range 0 10 Dyspnea 10 (91) Cough 5 (45) Wheeze 5 (45) Systemic symptoms 7 (64) Weight loss 5 (45) Fatigue 6 (55) Fever 2 (18) Extrathoracic sarcoidosis 6 (55) Radiographic stage 1 2 (18) 2 7 (64) 3 2 (18) 4 0 (0) Elevated SACE level 7 (78) Fiberoptic bronchoscopy findings Normal 0 (0) Thickened mucosa 9 (82) Typical granulations 4 (36) CT scan of the lung Bronchial wall thickening 7 (100) Diffuse luminal narrowing 4 (57) *Data are presented as No. (%), unless otherwise indicated. Cutaneous (three patients), peripheral lymph nodes (three patients), liver (one patient), bone (one patient), muscle (one patient), and hypothalamus (one patient). See the text for definition of stages. Normal value, 52 IU/mL; measured in 9 of 11 patients. Mean SD, 112 48 IU/mL. Performed in 7 of 11 patients. Eight patients took oral corticosteroids at mean daily dosages of 0.6 mg/kg (range, 20 to 60 mg/d) at the beginning, with dose tapering after 1 to 12 months (mean, 4 months). One other patient received IV methylprednisolone pulses, 500 to 1,000 mg twice monthly, because of psychologic intolerance of previous CHEST / 116 / 5/ NOVEMBER, 1999 1195
oral corticosteroids. Corticosteroids were contraindicated in the remaining two patients (diabetes in one, and osteonecrosis of the hip under previous steroid therapy in the other); one of them took hydroxychloroquine, 400 mg/d po, and the other one received IM methotrexate, 10 mg/wk. Treatment was started within 4 weeks of baseline, and the patients were followed for at least 6 months under therapy. All of them were still taking their medication at the time of the final pulmonary function evaluation (between the sixth and 12th months of treatment). Recently occurring or worsened dyspnea concomitant with AO was the reason for treatment in 10 cases; in addition, for 6 of the 10 patients, the treatment was of prime necessity because of extrathoracic involvement (weight loss or hypothalamic localization). In the 11th patient, treatment was initiated because of a dramatic radiographic worsening concomitant with the occurrence of AO. Four patients had already required treatment for sarcoidosis prior to the onset of AO, but treatment had been withdrawn for 3 to 12 months when AO appeared. Statistical Analysis FEV 1, VC, FEV 1 /VC values, and serum angiotensin-converting enzyme (SACE) level were compared at baseline and after 6 to 12 months of treatment using a paired t test. A p value of 0.05 was considered significant. Baseline Results Demographic, clinical, biological, and radiologic data at baseline are reported in Table 1. In most of these patients, sarcoidosis was of recent onset, and all but one were dyspneic. None had stage 0 (normal chest radiography) or stage 4 disease. In all cases, CT scan (when available) and fiberoptic bronchoscopy showed abnormalities. Baseline pulmonary function test results are shown in Table 2. AO was severe (FEV 1 50% predicted) in four patients. Four patients had a restrictive defect, defined as total lung capacity 80% predicted. Follow-up Pulmonary Function Tests: The FEV 1 improved by 10% predicted with a concomitant increase in the FEV 1 /VC ratio in eight patients (72%), and did not change significantly in the other three (Fig 1, top, A and center, B), although in one of them the FEV 1 /VC ratio normalized. FEV 1 and the FEV 1 /VC ratio normalized in four cases. The mean FEV 1 increased from 1,835 59 ml (60.8 10.8% predicted) to 2,304 81 ml (76 13.7% predicted; p 0.02). The mean VC did not change significantly. The FEV 1 /VC ratio increased from 76.1 6.4% predicted to 87.6 10.7% predicted on average (Fig 1, center, B; p 0.01); it decreased in two patients, in whom VC improved but FEV 1 did not change significantly. Clinical and Radiologic Findings: Dyspnea on exertion was attenuated in all but one patient, as were other clinical findings (fever, cough, weight loss, cutaneous lesions, etc). Radiographic improvement was observed in nine patients (two stage 1, five stage 2, two stage 3), and involved both pulmonary infiltrates and lymph nodes. In five of them (two stage 1 and three stage 2), the chest radiographic appearances completely normalized. The chest radiographs did not change in the remaining two patients (both stage 2), in whom AO was not attenuated. SACE Level: The mean SACE level decreased Table 2 Complementary Examinations at Baseline* Patient No. Radiographic Stage SACE Level, IU/mL FEV 1,mL (% pred) VC, % pred FEV 1 /VC, % (% pred) RV/TLC, % pred 1 3 33 1,880 (67) 82 68 (84) 115 2 2 95 1,830 (50) 64 63 (78) 125 3 2 110 2,120 (66) 92 57 (68) 88 4 1 150 1,280 (47) 67 53 (64) 160 5 2 95 3,360 (74) 84 69 (84) 101 6 2 146 1,990 (63) 89 60 (73) 93 7 1 143 2,160 (73) 88 62 (80) 78 8 2 193 1,520 (72) 101 58 (74) 68 9 2 ND 1,230 (49) 61 67 (83) 127 10 2 50 1,700 (65) 79 62 (78) 110 11 3 ND 1,120 (43) 62 57 (71) 151 Mean 112 1,835 (60.8) 79 61.5 (76.1) 110.7 SD 48 59 (10.8) 12.9 4.9 (6.4) 27.4 *RV/TLC residual volume/total lung capacity ratio; pred predicted; ND not done. See text for definition of radiographic stages. Normal value, 52 IU/mL. 1196 Clinical Investigations
from 112 48 to 58 40 IU/mL (p 0.05; Fig 1, bottom, C). The SACE level completely normalized with treatment in four patients, and decreased to just above the normal value in three patients. In one patient, although it decreased, it remained elevated. Comparison of SACE levels was not possible for three patients: SACE was measured only before treatment in one patient (33 IU/mL), only after treatment in another patient (118 IU/mL), and was not determined in the third patient. Among the three patients whose AO was not significantly attenuated, SACE levels were normal before treatment in two patients, and remained elevated after treatment in the third patient. Discussion In the present study, AO attributable to bronchial wall involvement by sarcoid granulomas was substantially attenuated by treatment in 72% of cases, indicating that AO can be either completely or partially reversed by treatment. This improvement was always associated with the disappearance or clear attenuation of pulmonary symptoms. Numerous studies have assessed airway function in sarcoidosis 1 9 and demonstrated the presence of AO in many patients. However, little emphasis has been placed on the evolution of AO during treatment, and contradictory results have been reported. Some studies showed that corticosteroids were ineffective. Olsson et al 25 and Lewis and Horak 12 observed no clinical improvement in three and one treated patients, respectively. Miller et al 26 described two patients with severe AO; one patient experienced no improvement with corticosteroids, whereas the other patient improved only after prolonged treatment with high-dose corticosteroids, 80 mg/d. However, in these cases, AO was associated with single or, more often, multiple bronchostenoses, 25,26 which are usually insensitive to treatment. 27 Similarly, Renzi et al 20 found no beneficial effect of 4 months of steroid therapy on the small airway disease detected in seven patients, but there was no evidence of functional involvement of the large airways in these patients (the FEV 1 /VC ratio was normal). In contrast, Smellie et al 28 found that treatment with corticosteroids, 20 to 30 mg/d, led to improvement of Figure 1. Individual values (Œ) for FEV 1 (top, A), FEV 1 /VC ratio (center, B), and SACE level (bottom, C) at baseline and 6 to 12 months after starting treatment. The mean ( SD; f) for each variable was significantly different after 6 to 12 months of treatment: p 0.02 for FEV 1 ;p 0.01 for FEV 1 /VC ratio; and p 0.05 for SACE level. In bottom, C, the dashed line represents the upper limit of the normal range: 52 IU/mL. CHEST / 116 / 5/ NOVEMBER, 1999 1197
FEV 1 in three of six patients who had low FEV 1 values at baseline, but the FEV 1 /VC ratio at that time was not given. Colp et al 19 reported that AO (defined as an FEV 1 70% of forced VC) was attenuated in 10 (all nonsmokers) out of 22 steroid-treated patients (4 of these patients deteriorated), and even in 5 out of 21 untreated patients (7 deteriorated in this group), during a follow-up period of at least 6 months. The authors concluded that, while the course of AO in sarcoidosis is variable, corticosteroid therapy may exert a beneficial effect in nonsmokers. However, they did not discuss the details of the AO encountered in this subgroup of patients or statistically analyze their results. Other isolated cases have been published with contrasting results. 15,29 All of these studies recruited nonhomogeneous groups of patients, with different sites and causes of AO, which could explain their discrepant results. Our findings are more favorable, probably because of the selection of the patients and, especially, the mechanism involved (granulomatous process). In this study, all patients had a ventilatory obstructive defect located in the large airways, with an FEV 1 /VC ratio 70%. We selected patients with AO due to bronchial wall disease, proved by bronchoscopic and histologic findings (ie, the presence of granuloma in bronchial mucosa) and confirmed by a CT scan of the lung (presence of peribronchial thickening) in all seven patients so examined. AO was not likely to be caused by other mechanisms, such as extrinsic compression by lymph nodes or airway distortion related to fibrosis. Indeed, chest radiography showed no stage 4 disease and no atelectasis, and no sign of extrinsic compression was found during fiberoptic bronchoscopy. Furthermore, CT scans of the lungs, available for seven patients, showed no sign of major distortion and no compressive lymph nodes. Lastly, AO could not be attributed to airway hyperreactivity associated with sarcoidosis as described by some authors 30 because we excluded patients who had a bronchodilator response or a history of asthma. Most of our patients had sarcoidosis of recent onset: the median duration of disease was 2 years. Their chest radiographs showed that a majority had stage 1 or 2 (nine patients) disease, few had stage 3 (two patients), and none had stage 4. In addition, for most of them, the disease presentation at baseline was severe, with systemic symptoms (64%), extrapulmonary involvement (55%), and substantially elevated SACE level (78%). Corticosteroids are more likely to suppress the growth of existing granulomas and to prevent the development of new ones, and are less effective against fibrotic lesions. 9 Thus, the favorable results obtained in these severe presentations probably cannot be extrapolated to all cases of AO in sarcoidosis. It is highly relevant that two patients in this study had stage 1 symptomatic disease associated with AO (severe in one patient) and that they improved dramatically under treatment, suggesting that airways can be substantially involved in so-called radiologically benign sarcoidosis. A high percentage (90%) of our patients with AO complained of dyspnea. This finding is in accordance with the study by De Remee and Andersen, 10 which showed a strong link between dyspnea and slowed expiratory airflow. Although it is difficult to ascertain whether dyspnea was related to AO in our patients, several findings strongly support this hypothesis: (1) dyspnea was not likely to be due to a restrictive pattern, which was found in only four patients and was moderate; and (2) although dyspnea decreased in all patients whose AO was attenuated, their VC did not change significantly. Because of the small number of patients, it was not possible to identify factors predictive of therapeutic reversibility of AO. No statistical correlation was found between attenuated AO and baseline data as a function of age, duration of disease, radiographic stage, FEV 1 /VC ratio, or the FEV 1 value. AO severity was not predictive of a therapeutic response. In addition, changes in the SACE level were not correlated with those of FEV 1 or FEV 1 /VC ratio. Among the three patients in whom FEV 1 did not change significantly, two had normal SACE levels at baseline (SACE level had not been measured in the third), whereas all patients whose AO was attenuated had elevated SACE levels ( 2 times the normal value) at baseline (available in all but one). Although this relationship did not reach statistical significance, a low baseline SACE level tended to be predictive of a poor response to treatment. We can add that the treatment may have been suboptimal in the three nonresponder patients: two patients received corticosteroids at low doses, 20 and 30 mg/d, and the third patient received irregular IV pulses of methylprednisolone. Long-term outcome has not been specifically studied in this series. Nine patients were followed after the end point of the study (for a median duration of 29 months); seven of them demonstrated a persistent benefit in terms of their airflow, even a few months after concluding the treatment in two cases. In conclusion, the presence of AO in a patient suffering from sarcoidosis should be investigated, at least by means of fiberoptic bronchoscopy and CT scan of the lungs, to assess the mechanism involved. Treatment should be discussed if AO is thought to be related to granulomatous airway involvement. Future studies should focus on the prevalence and evolution of AO attributable to other mechanisms, particularly fibrosis-related airway distortion. 1198 Clinical Investigations
ACKNOWLEDGMENT: The authors would like to thank D. Malka, MD, and J. Jacobson for technical assistance. References 1 Lamberto C, Saumon G, Loiseau P, et al. Respiratory function in recent pulmonary sarcoidosis with special reference to small airways. Bull Eur Physiopathol Respir 1985; 21:309 315 2 Dines DE, Stubbs SG, McDougall JC. Obstructive disease of the airways associated with stage I sarcoidosis. Mayo Clin Proc 1978; 53:788 791 3 Levinson RS, Metzger LF, Stanley NN, et al. Airway function in sarcoidosis. Am J Med 1977; 62:51 59 4 Miller A, Teirstein AS, Jackler I, et al. Airway function in chronic pulmonary sarcoidosis with fibrosis. Am Rev Respir Dis 1974; 109:179 189 5 Kaneko K, Sharma OP. Airway obstruction in pulmonary sarcoidosis. Bull Eur Physiopathol Respir 1977; 13:231 240 6 Miller A, Chuang M, Teirstein AS, et al. Pulmonary function in stage I-II pulmonary sarcoidosis. Ann NY Acad Sci 1976; 278:293 300 7 Harrison BDW, Shaylor JM, Stokes TC, et al. Airflow limitation in sarcoidosis: a study of pulmonary function in 107 patients with newly diagnosed disease. Respir Med 1991; 85:59 64 8 Sharma OP, Johnson R. Airway obstruction in sarcoidosis: a study of 123 nonsmoking black American patients with sarcoidosis. Chest 1988; 94:343 346 9 Newman LS, Rose CS, Maier LA. Sarcoidosis. N Engl J Med 1997; 336:1224 1234 10 De Remee RA, Andersen HA. Sarcoidosis: a correlation of dyspnea with roentgenographic stage and pulmonary function changes. Mayo Clin Proc 1974; 49:742 745 11 Viskum K, Vestbo J. Vital prognosis in intrathoracic sarcoidosis with special reference to pulmonary function and radiological stage. Eur Respir J 1993; 6:349 353 12 Lewis MI, Horak DA. Airway function in sarcoidosis. Chest 1987; 92:582 584 13 Longcope WT, Freiman DG. A study of sarcoidosis based on a combined investigation of 160 cases, including 30 autopsies, from the Johns Hopkins Hospital and the Massachussetts General Hospital. Medicine (Baltimore) 1952; 31:1 32 14 Turiaf J, Rose Y, Basset F. La sarcoïdose bronchique. Les Bronches 1963; 13:587 609 15 Benatar SR, Clark TJH. Pulmonary function in a case of endobronchial sarcoidosis. Am Rev Respir Dis 1974; 110: 490 496 16 Kalbian VV. Bronchial involvement in pulmonary sarcoidosis. Thorax 1957; 12:18 23 17 Udwadia ZF, Pilling JR, Jenkins PF, et al. Bronchoscopic and bronchographic findings in 12 patients with sarcoidosis and severe or progressive airways obstruction. Thorax 1990; 45: 272 275 18 McCann BG, Harrison BDW. Bronchiolar narrowing and occlusion in sarcoidosis: correlation of pathology with physiology. Respir Med 1991; 85:65 67 19 Colp C, Park SS, Williams MH. Pulmonary function follow-up of 120 patients with sarcoidosis. Ann NY Acad Sci 1976; 278:301 307 20 Renzi DG, Renzi PM, Lopez-Majano V, et al. Airway function in sarcoidosis: effect of short-term steroid therapy. Respiration 1981; 42:98 104 21 Du Bois RM. Corticosteroids in sarcoidosis: friends or foe? Eur Respir J 1994; 7:1203 1209 22 European Respiratory Society. Standardized lung function testing. Eur Respir J 1993; 6(suppl 16):1 100 23 Naidich DP, Funt S, Ettenger NA, et al. Hemoptysis: CTbronchoscopic correlations in 58 cases. Radiology 1990; 177: 357 362 24 Lenique F, Brauner M, Grenier P, et al. CT assessment of bronchi in sarcoidosis: endoscopic and pathologic correlations. Radiology 1995; 194:419 423 25 Olsson T, Björnstad-Pettersen H, Stjernberg NL. Bronchostenosis due to sarcoidosis: a cause of atelectasis and airway obstruction simulating pulmonary neoplasm and chronic obstructive pulmonary disease. Chest 1979; 75:663 666 26 Miller A, Bracon LK, Teirstein AS. Stenosis of main bronchi mimicking fixed upper airway obstruction in sarcoidosis. Chest 1985; 88:244 248 27 Sadoun D, Tandjaoui H, Hirshi S, et al. Sténoses bronchiques proximales au cours de sarcoïdoses médiastino-pulmonaires sans signe radiographique de fibrose: étude de 10 cas [abstract]. Rev Mal Respir 1995; 12(suppl 2):A85 28 Smellie H, Apthorp GH, Marshall R. The effect of corticosteroid treatment on pulmonary function in sarcoidosis. Thorax 1961; 16:87 91 29 Hadfield JW, Page RL, Flower CDR, et al. Localized airway narrowing in sarcoidosis. Thorax 1982; 37:443 447 30 Bechtel JJ, Starr T III, Dantzker DR, et al. Airway hyperreactivity in patients with sarcoidosis. Am Rev Respir Dis 1981; 124:759 761 CHEST / 116 / 5/ NOVEMBER, 1999 1199