ANDREW G. NICHOLSON, THOMAS V. COLBY, ROLAND M. DUBOIS, DAVID M. HANSELL, and ATHOL U. WELLS

Transcription

1 The Prognostic Significance of the Histologic Pattern of Interstitial Pneumonia in Patients Presenting with the Clinical Entity of Cryptogenic Fibrosing Alveolitis ANDREW G. NICHOLSON, THOMAS V. COLBY, ROLAND M. DUBOIS, DAVID M. HANSELL, and ATHOL U. WELLS Departments of Histopathology, Interstitial Lung Disease Unit and Radiology, Royal Brompton Hospital, London, United Kingdom; and Department of Pathology, Mayo Clinic, Scottsdale, Arizona Lone cryptogenic fibrosing alveolitis (CFA) is a progressive interstitial lung disease, with a median survival of 3 to 6 yr from the onset of dyspnea. CFA can be subdivided into prognostically significant histopathologic patterns, including nonspecific interstitial pneumonia (NSIP). We reviewed 78 patients with a clinicopathologic diagnosis of CFA, biopsied between 1978 and 1989, to evaluate the prevalence and prognostic significance of these histopathologic patterns, in particular NSIP. Biopsy appearances were reclassified by two pulmonary histopathologists as usual interstitial pneumonia (UIP) (47%), NSIP (36%), or desquamative interstitial pneumonia (DIP)/respiratory bronchiolitis associated interstitial lung disease (RBILD) (17%). The kappa coefficient of agreement between pathologists was In 67 cases, follow-up was complete to death or 10 yr after biopsy, with 50 deaths during a median follow-up of 42 mo (UIP, 89%; NSIP, 61%, DIP/RBILD, 0%). Survival was highest in DIP/RBILD and higher in NSIP than UIP, p When analysis was confined to patients with UIP or NSIP, the mortality of UIP remained higher, p 0.01, but the 5-yr survival in patients with fibrotic NSIP was only 45%, indicating that this histologic appearance is often associated with a poor outcome. A response to treatment was more frequent in DIP/RBILD than in NSIP (p 0.01) or UIP (p ). This study confirms the prognostic value of subclassifying patients with CFA according to histopathologic pattern. However, in patients with clinically typical CFA, a histologic diagnosis of fibrotic NSIP needs to be interpreted with caution and does not necessarily denote a good outcome. The term lone cryptogenic fibrosing alveolitis (CFA), synonymous with idiopathic pulmonary fibrosis (IPF) (1 3), relates to an idiopathic progressive interstitial lung disease, with a mean survival from the onset of dyspnea of 3 to 6 yr (4 10). The entity of CFA is a major subgroup of the interstitial pneumonias, originally subdivided by Liebow and Carrington into usual interstitial pneumonia (UIP), desquamative interstitial pneumonia (DIP), bronchiolitis obliterans with interstitial pneumonia (some of which are now termed as bronchiolitis obliterans organizing pneumonia [BOOP]), giant cell interstitial pneumonia (now thought of as a hard-metal pneumoconiosis), and lymphoid interstitial pneumonia (viewed by some as a lymphoproliferative disorder) (11). More recently, nonspecific interstitial pneumonia (NSIP) (12), respiratory bronchiolitis associated interstitial lung disease (RBILD) (13), and diffuse alveolar damage, clinically termed acute interstitial pneumonia (AIP) when idiopathic (14), have been recognized as additional patterns (15, 16). How these histologic patterns relate to patients with a clinical diagnosis of CFA remains contentious. Although DIP and UIP were believed by some to represent different stages of (Received in original form March 9, 2000 and in revised form August 11, 2000) Correspondence and requests for reprints should be addressed to Dr. A. G. Nicholson, Department of Histopathology, Royal Brompton Hospital, Sydney St, London SW3 6NP, UK. a.nicholson@rbh.nthames.nhs.uk Am J Respir Crit Care Med Vol 162. pp , 2000 Internet address: CFA, there is now a general consensus that DIP represents a separate entity to UIP and that CFA should be related to the latter histopathologic pattern. However, the relationship of NSIP to CFA remains harder to integrate. Although previous series have demonstrated a better outcome with a histopathologic pattern of NSIP than with UIP (17 19), the studied populations were not selected on the basis of a typical clinical presentation of CFA and were not entirely representative of the full spectrum of CFA encountered in clinical practice. A pilot study of 30 patients who met clinical criteria for a diagnosis of CFA showed significantly better prognosis and response to treatment when diagnosed with a NSIP pattern rather than UIP (20). However, these patients were not representative of the larger population of CFA patients described in keynote historical series (5 8). In the computed tomography (CT) era, thoracoscopic biopsy in suspected CFA has been increasingly reserved for the minority of patients with atypical CT features; thus, patients with typical disease are increasingly excluded from histologic studies. The aim of the present study was to evaluate the clinical significance of histopathologic reclassification in the whole spectrum of CFA encountered in clinical practice, before the advent of CT had resulted in a major selection bias. METHODS Between January 1, 1978 and November 31, 1989, 205 patients met all four clinical criteria for a diagnosis of lone CFA: (1) bilateral basal crackles or widespread crackles, most prominent at the bases; (2) abnormalities compatible with bilateral lung fibrosis on chest radiography, in a distribution typical of fibrosing alveolitis; (3) a restrictive functional defect (reduction in TLC below 80% of predicted normal associated with a restrictive FEV 1 /FVC ratio) or isolated reduction in diffusing capacity of the lungs for carbon monoxide (DL CO ); and (4) absence of occupational or environmental cause for pulmonary fibrosis (6). Open lung biopsies were performed in 108 patients: 81 were reviewed (the remaining 27 cases were unavailable or the biopsy had been performed more than 6 mo before referral). Three cases were excluded on histopathologic reevaluation: one was reclassified as sarcoidosis, known to mimic CFA both clinically and radiologically (21), and two were reclassified as chronic obstructive pulmonary disease (COPD) with associated minor interstitial fibrosis. This left 78 cases with histopathologic appearances consistent with UIP, NSIP, DIP, or RBILD. There was no overlap between these 78 patients and patients included in previous histologic studies of CFA from our institution (7, 20); the study population makes up a subgroup of an earlier nonhistologic study in which survival in CFA was compared with survival in the fibrosing alveolitis of systemic sclerosis (22). Slides were reviewed independently by two pulmonary histopathologists (A.G.N., T.V.C.), without knowledge of the clinical data. All sections had been routinely stained with hematoxylin eosin, with additional staining using elastin van Gieson stains. In 22 of 81 cases, biopsies were taken from two separate areas, in 18 cases from separate lobes of which one was a lower lobe. Both biopsies were evaluated together and, when classification differed between the pathologists, a consensus opinion on the overall histopathologic pattern was reached. The pattern of UIP was distinguished by a temporally heter-

2 2214 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL ogeneous pattern of ongoing lung injury (i.e., a variation in the age of fibrosis with foci of fibroblastic proliferation, where there was an abundance of plump spindle cells and little intervening collagen, immediately adjacent to areas of established fibrosis, these characterized by poorly cellular hyalinized collagen). Secondary features such as honeycombing and a subpleural distribution favored diagnosis of UIP. All other histopathologic patterns of interstitial pneumonia showed temporally uniform fibrosis when present, with a diffuse filling of alveoli by macrophages in DIP and a more peribronchiolar distribution in RBILD. In NSIP, the fibrosis and inflammation was either patchy or diffuse, but the pattern of lung injury remained temporally uniform. Cases of NSIP were subdivided into NSIP grade 1 (cellular interstitial pneumonia with little or no fibrosis), NSIP grade 2 (significant fibrosis consisting of mature collagen, sometimes with accompanying fibroblasts, these elements being present in constant proportions throughout the affected interstitium), and NSIP grade 3 (dense interstitial collagen deposition with or without derangement of architecture) (12). Clinical data, including history of previous exposures, smoking, chronic disease, drugs, and clinical examination, were collected, either from the patient s notes or from general practitioner s records. Patients were categorized as nonsmokers, current smokers, or ex-smokers (a minimum of one cigarette a day for a minimum of 1 yr, stopping at least 3 mo before presentation). All but one patient (who was too compromised to undergo full lung function testing) were included in survival analysis. Changes in pulmonary function tests at 4 to 6 mo were evaluated (as close to 6 mo as possible). Lung function measurements had been performed at presentation and during the follow-up period, including FVC and DL CO. Results were expressed as percentages of values predicted from the subject s age, sex, and height. Lung volumes were measured using an Ohio water-seal spirometer (Ohio Instruments, Atlanta, GA). Measures of gas transfer (DL CO ) were made by the single-breath technique using a P.K. Morgan respirometer (P.K. Morgan, Chatham, Kent, UK). Serial trends in pulmonary function tests at follow-up were defined as improvement (a rise of over 15% in FVC or DL CO ), decline (a fall of over 15% in FVC or (DL CO ), or stability (a change of less than 15%) (23, 24). For treatment effects, we analyzed all patients who were on no treatment (n 56) or low-dose prednisolone ( 10 mg daily, n 3) at the time of referral, who subsequently went on to receive high-dose prednisolone (60 mg daily for 4 wk, then reducing by 5 mg/d each week to a maintenance dose of 20 mg on alternate days) (n 30, consisting of: UIP, n 10; NSIP, n 12; DIP, n 8) or combination prednisolone/cyclophosphamide therapy (cyclophosphamide 100 to 120 mg/d; prednisolone 20 mg/d on alternate days) (n 29, consisting of: UIP, n 18; NSIP, n 9; DIP, n 2), as described previously (25). Nineteen patients were excluded (on high-dose therapy at referral, n 7; died too soon to evaluate, n 2 [one 10 d after biopsy, one with a cerebrovascular accident 2 mo after biopsy, before follow-up lung function tests were performed]; incomplete follow-up, n 4; delay in starting treatment after biopsy or no treatment, n 6). Group comparisons were made using unpaired t tests (for normally distributed, continuous variables), Wilcoxon s rank-sum tests (for non-normally distributed variables), and chi-square statistics or Fisher exact tests as appropriate (for comparisons of proportions). Group survival was compared using proportional hazards regression with adjustment for the severity of disease, as judged by DL CO levels, and patient age at the date of biopsy (STATA data analysis software; Computing Resource Center, Santa Monica, CA) (26). A p value of less than 0.05 was regarded as statistically significant. Variation between pathologists in categorizing cases as NSIP, UIP, or DIP was quantified using the kappa coefficient of agreement (unweighted). RESULTS Of the 78 patients who had a prior clinicopathologic diagnosis of CFA, 37 (47%) were reclassified as UIP, 28 (36%) as NSIP, and 13 (17%) as DIP/RBILD (Table 1). The kappa coefficient of agreement between the two pathologists overall was 0.49, and 0.26 in distinguishing between UIP and NSIP3 (the type of NSIP most frequently mistaken for UIP). Patients with DIP/ RBILD were younger than the other two groups (p in both comparisons) and were more likely to be female (p 0.01 versus UIP, p 0.12 versus NSIP), but did not differ in the severity of functional impairment. There were no other significant differences in age, male:female ratio, or the functional severity of disease between UIP and NSIP, although patients with NSIP were slightly younger (p 0.07) and had greater impairment of DL CO (p 0.11). In two cases, the differential diagnosis included eosinophilic pneumonia, owing to an increase in the number of eosinophils within the chronic inflammatory cell infiltrate, but this diagnosis was not supported by accompanying clinical data, specifically the imaging and blood indices. Patients with NSIP had a slightly shorter duration of dyspnea at the date of biopsy than those with UIP, p 0.07, but no trends were disclosed when patients with DIP/RBILD were compared with the other two subgroups. TABLE 1 CLINICAL DATA AT PRESENTATION, SURVIVAL AND RESPONSE TO TREATMENT IN SUBGROUPS OF INTERSTITIAL PNEUMONIA UIP NSIP DIP/RBILD No. of cases 37 (47%) 28 (36%) 13 (17%) Age* Sex, M:F ratio 33:4 20:8 6:7 Smoking Nonsmoker Current Ex-smoker DL CO, % pred* FVC, % pred* PO 2, kpa, median (range) ( ) ( ) ( ) Duration of dyspnea at biopsy, mo, median (range) (1 67) (0 180) (0 89) Total number of deaths 33/37 (89%) 17/28 (61%) 0/13 (0%) Response to therapy (n 59) at 6 mo 3/28 6/21 8/10 Decline despite therapy (n 59) at 6 mo 8/28 2/21 0/10 * Values are mean SD. With the exclusion of two patients with UIP in whom the duration of dyspnea was not quantifiable. Figure 1. Survival of usual interstitial pneumonia (UIP), fibrotic nonspecific interstitial pneumonia (NSIP), and desquamative interstitial pneumonia (DIP)/respiratory bronchiolitis associated interstitial lung disease (RBILD)/cellular NSIP subgroups of interstitial pneumonia in patients originally classified as cryptogenic fibrosing alveolitis (CFA).

3 Nicholson, Colby, dubois, et al.: Interstitial Pneumonia Histology and Cryptogenic Fibrosing Alveolitis 2215 In 67 cases, follow-up was complete to death or 10 yr after biopsy. Follow-up was censored at less than 10 yr after biopsy in 11 patients, including five living overseas. There were 50 deaths during a median follow-up of 42 mo (UIP 33/37, 89%; NSIP 17/28, 61%; DIP/RBILD 0/13, 0%) (Table 1). As shown in Figure 1, survival was highest in DIP/RBILD/cellular NSIP and higher in fibrotic NSIP (NSIP2 and NSIP3) than in UIP, p ; median survivals were 24 mo for UIP, 52 mo for NSIP and the median follow-up was 96 mo in the 13 patients with DIP/RBILD (all survivors). Differences in survival persisted, p , after controlling for age, smoking history, and percentage predicted DL CO immediately prior to biopsy. Mortality increased independently with increased age (coefficient 0.065/yr, 95% confidence interval [CI] to 0.111, p 0.005) and reduction in DL CO (coefficient 0.026/1% DL CO, 95% CI to 0.049, p 0.020), but was not independently linked to the smoking history. When analysis was confined to patients with UIP and fibrotic NSIP, the mortality of UIP remained significantly higher after adjustment for age, smoking history, and impairment of DL CO, p 0.01; a higher mortality was independently associated with older age and greater impairment of DL CO (p 0.06 and p 0.03, respectively), but not with the smoking history. As shown in Figure 2, survival in patients with NSIP was highest in NSIP1 (n 3) and higher in NSIP2 (n 12) and in NSIP3 (n 13) than in UIP (p 0.02 and p 0.05, respectively). After adjustment for age, smoking history, and percentage predicted DL CO, survival in NSIP2 (p 0.01) and NSIP3 (p 0.03) remained significantly higher than in UIP. Responsiveness to treatment was more frequent in DIP/ RBILD than in NSIP (p 0.01) or UIP (p ) (Table 1); however, there were no significant differences between UIP and NSIP, either in responsiveness to treatment (3/28 versus 6/21), or in the prevalence of decline in lung function (8/28 versus 2/21), at 4 to 6 mo. DISCUSSION The aim of this study was to assess the value of the histologic reclassification of cases presenting with a clinical diagnosis of CFA. The results are notable for two features. First, the proportion of patients found to have NSIP was higher than in other series (Table 2) (17 19), but this finding was wholly ascribable to the presence of a large subgroup with fibrotic NSIP. More importantly, the present series documents a much worse outcome in NSIP at 5 yr than previously reported, although survival in NSIP was significantly better than in UIP. It is likely that the poor outcome in NSIP in the present study, compared with previous series, largely results from the selective inclusion of patients meeting clinical criteria for a diagnosis of CFA and undergoing open lung biopsy. In previous series of patients with subgroups of NSIP and UIP, clinical criteria for a diagnosis of CFA were not inclusion criteria (17 19) or a highly selected subgroup were studied (20). Nagai and coworkers reported a cohort in which many patients had clinical and, particularly, radiologic findings suggestive of cryptogenic organizing pneumonia (18). Similarly, in the series of Bjoraker and coworkers, 20% of patients did not have crackles and would not have been included in the large historical series, in which the survival characteristics of the clinical entity of CFA/ IPF were defined (17). Travis and coworkers viewed their cases of fibrotic NSIP as a group of miscellaneous non-uip fibrosing disorders of diverse aetiologies (19); they did not suggest that the clinical entity of CFA was selectively evaluated in their study. More recently, CT scanning has introduced a major selection bias in the decision to proceed to biopsy: Daniil and coworkers emphasized that the selection of patients for biopsy was heavily influenced by some unusual features in their clinical and particularly radiographic work-up (20). Thus, in previous histologic series of patients with UIP and NSIP (17 20) the spectrum of clinically typical CFA was not selectively studied. The major strength of the present study is the description of a cohort in which the diagnosis of CFA was made by clinicians who had defined the entity of CFA in earlier keynote historical series (6, 7), at an institution in which an open lung biopsy was performed whenever possible, by protocol. Thus, our findings can be applied to the classic clinical entity of CFA, widely encountered in routine respiratory practice. The relatively poor prognosis in fibrotic NSIP in the present study highlights the dangers of extrapolating previous reports of a good outcome associated with NSIP to patients with clinically typical CFA. NSIP had most likely been described before Katzenstein and Fiorelli s study (12), with Kitaichi proposing an unclassified group of interstitial pneumonia in 1990 (27), and histopathologic patterns whose descriptions equate to NSIP being described in other series relating response to treatment in patients with CFA (28). Indeed, the term nonspecific interstitial pneumonitis had been used prior to 1990 in relation to patients with human immunodeficiency virus (HIV) infections (29) but, in relation to idiopathic or connective tissue disorder related interstitial pneumonias, NSIP only became established as a diagnostic term after Katzenstein and Fiorelli s study. They described a group of cases of interstitial pneumonia, which could not be classified according to the recognized TABLE 2 PERCENTAGES OF PATIENTS WITH UIP, NSIP, DIP, AND BOOP IN PREVIOUS SERIES Series UIP NSIP Cellular NSIP Fibrotic NSIP DIP/RBILD BOOP Bjoraker, et al. (17) 62% 14% NS NS 10% 2% Nagai, et al. (18) 58% 28% 14% 13% 14% Travis, et al. (19) 55% 29% 7% 22% 16% Nicholson, et al. 47% 36% 4% 32% 17% Figure 2. Survival in the NSIP subgroup, according to the absence (NSIP1) or extent of interstitial fibrosis (NSIP2 and NSIP3). Definition of abbreviations: BOOP bronchiolitis obliterans-organizing pneumonia; DIP/ RBILD desquamative interstitial pneumonia/respiratory bronchiolitis associated interstitial lung disease; NSIP nonspecific interstitial pneumonia; UIP usual interstitial pneumonia.

4 2216 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL subsets at that time. They were, however, particular in stating that this was not a specific disease entity, because some cases were associated with collagen vascular diseases, some with exposure to environmental allergens, and some had a history of acute lung injury, the histopathologic pattern perhaps reflecting incomplete resolution of diffuse alveolar damage. Despite this, the pattern that they described showed greater response to treatment and a more favorable prognosis than patients with a histologic pattern of UIP (12). With incorporation of the histopathologic pattern of NSIP into the classification of idiopathic interstitial pneumonias (16), we wished to assess Katzenstein and Fiorelli s grading system in a group of patients with idiopathic NSIP. When cases were subgrouped according to the degree of fibrosis, there were significant differences between the fibrotic (NSIP2 and NSIP3) and nonfibrotic (NSIP1) groups, although the number of cases classified as NSIP1 is small. These data especially support the recent findings of Travis and coworkers who found that cellular, i.e. nonfibrotic NSIP, had a significantly better prognosis than patients with fibrotic NSIP (19), and that fibrotic NSIP had a worse prognosis than DIP/RBILD. However, our 5-yr survival for fibrotic NSIP (grades 2 and 3 combined) was much worse than that described by Travis and coworkers (45% versus 90%), and this difference in outcome is likely to reflect the nature of the population in the present study, that of clinically typical CFA. In this series, cases were also reviewed independently by two histopathologists, in order to evaluate the reproducibility of the classification system. A kappa value of 0.49 is regarded as clinically acceptable and is comparable to that seen in the scoring of CT patterns in CFA (i.e., which morphologic pattern predominates) (30), but it highlights the fact that, despite its gold standard status, assessment of open lung biopsies is subject to interobserver variation when CFA is subgrouped into patterns of interstitial pneumonia. Furthermore, the kappa value for distinguishing between UIP and NSIP3 was only 0.26, underlying the difficulty in making this distinction in some cases. With the histopathologic differences at the fibrotic end of the NSIP spectrum less well defined than at the cellular end, our analyses support suggestions that it may be more appropriate to refer to such cases as end-stage lung or interstitial lung with honeycomb change rather than attempting further subclassification (16). A response to treatment was more frequent in DIP/RBILD than in NSIP and UIP, supporting the contention that the small percentage of patients with CFA who responded to corticosteroid therapy in previous studies may have had patterns other than UIP (7). It has been observed that an early response to treatment (7) and improvement or stability in lung function indices after 1 yr of treatment (31) are powerful predictors of long-term survival in CFA. Our data show that, whereas responsiveness to treatment appeared to identify patients with a high chance of having DIP/RBILD, the distinction between NSIP and UIP was not strongly linked to changes in lung function indices after 4 to 6 mo of therapy. Thus, the clinician cannot rely upon short-term changes in clinical status to draw conclusions on likely histopathologic group, or longer-term survival to discriminate consistently between UIP and NSIP, when the majority of patients with NSIP have fibrotic disease. However, it is not yet clear whether prolonged aggressive treatment is more beneficial in fibrotic NSIP than in UIP. Even before recognition of NSIP, variability in the clinical outcomes of patients with lone CFA had long been recognized, a finding which has often been accounted for by the idea that variations in the histologic patterns related to early and late phases of the same disease: in particular DIP being an early phase of UIP (4, 6, 32 34). However, most groups now regard UIP and DIP as distinct conditions (17, 35), a view supported by recent studies on the progression of CT findings in these two patterns: cases of DIP, which have a patchy distribution on CT, do not invariably progress to the characteristic subpleural distribution of UIP (36, 37). Our data support the latter view, not only because of the differences in prognosis between the subsets, but also because there were no significant differences between DIP and UIP in the duration of dyspnea before biopsy or severity of lung function impairment, indicating that DIP is unlikely to be a precursor of UIP. DIP is more appropriately grouped with RBILD, in view of their histopathologic and prognostic similarities (13, 38). In conclusion, this study confirmed the prognostic value of subclassifying cases of lone CFA as UIP, NSIP, DIP, and RBILD. However, the outcome in patients with fibrotic NSIP was substantially worse than previously reported. This finding may be partially ascribable to difficulties in making the distinction between UIP and NSIP in advanced disease but is also likely to reflect the selection of patients considered to have the clinical entity of CFA, rather than the larger group of patients who are found to have NSIP at biopsy. Our findings indicate that in patients with clinically typical CFA, a histologic diagnosis of fibrotic NSIP needs to be interpreted with caution and does not necessarily denote a good outcome. Our study also emphasizes the importance of viewing NSIP in the context of the clinical presentation. Acknowledgment : The authors would like to thank our colleagues in the Department of Thoracic Surgery, led by P. Goldstraw, for undertaking the biopsies and providing the tissue which made this work possible. References 1. Crystal RG, Fulmer JD, Roberts WC, Moss ML, Line BR, Reynolds HY. Idiopathic pulmonary fibrosis: clinical, histologic, radiographic, physiologic, scintigraphic, cytologic, and biochemical aspects. Ann Intern Med 1976;85: Raghu G. Idiopathic pulmonary fibrosis: a rational clinical approach. Chest 1987;92: Crystal RG, Bitterman PB, Rennard SI, Hance AJ, Keogh BA. Interstitial lung diseases of unknown cause: disorders characterized by chronic inflammation of the lower respiratory tract. N Engl J Med 1984;310: Scadding JG, Hinson KFW. Diffuse fibrosing alveolitis (diffuse interstitial fibrosis of the lungs): correlation of histology at biopsy with prognosis. Thorax 1967;22: Stack BH, Choo-Kang YF, Heard BE. The prognosis of cryptogenic fibrosing alveolitis. Thorax 1972;27: Turner-Warwick M, Burrows B, Johnson A. Cryptogenic fibrosing alveolitis: clinical features and their influence on survival. Thorax 1980;35: Turner-Warwick M, Burrows B, Johnson A. Cryptogenic fibrosing alveolitis: response to corticosteroid treatment and its effect on survival. Thorax 1980;35: Tukiainen P, Taskinen E, Holsti P, Korhola O, Valle M. Prognosis of cryptogenic fibrosing alveolitis. Thorax 1983;38: Hubbard R, Johnston I, Britton J. Survival in patients with cryptogenic fibrosing alveolitis: a population-based cohort study. Chest 1998;113: Johnston IDA, Prescott RJ, Chalmers JC. British Thoracic Society study of cryptogenic fibrosing alveolitis: current presentation and initial management. Thorax 1997;52: Liebow AA, Carrington CB. The interstitial pneumonias. In: Simon M, Potchen EJ, Lemay E, editors. Frontiers in pulmonary radiology. New York: Grune and Stratton; p Katzenstein AA, Fiorelli RF. Nonspecific interstitial pneumonia/fibrosis: histologic features and clinical significance. Am J Surg Pathol 1994; 18: Myers JL, Veal CF Jr, Shin MS, Katzenstein ALA. Respiratory bronchiolitis causing interstitial lung disease: a clinicopathologic study of six cases. Am Rev Respir Dis 1987;135: Katzenstein A-L, Myers JL, Mazur MT. Acute interstitial pneumonia: a

5 Nicholson, Colby, dubois, et al.: Interstitial Pneumonia Histology and Cryptogenic Fibrosing Alveolitis 2217 clinicopathologic, ultrastructural, and cell kinetic study. Am J Surg Pathol 1986;10: Katzenstein ALA, Myers JL. Idiopathic pulmonary fibrosis: clinical relevance of pathologic classification. Am J Respir Crit Care Med 1998; 157: Katzenstein ALA, Myers JL. Nonspecific interstitial pneumonia and the other idiopathic interstitial pneumonias: classification and diagnostic criteria. Am J Surg Pathol 2000;24: Bjoraker JA, Ryu JH, Edwin MK, Myers JL, Tazelaar HD, Schroeder DA, Offord KP. Prognostic significance of histopathological subsets in idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 1998;157: Nagai S, Kitaichi M, Itoh H, Nishimura K, Izumi T, Colby TV. Idiopathic nonspecific interstitial pneumonia/fibrosis: comparison with idiopathic pulmonary fibrosis and BOOP. Eur Respir J 1998;12: Travis WD, Matsui K, Moss J, Ferrans VJ. Idiopathic nonspecific interstitial pneumonia: prognostic significance of cellular and fibrosing patterns survival comparison with usual interstitial pneumonia and desquamative interstitial pneumonia. Am J Surg Pathol 2000;24: Daniil ZD, Gilchrist FC, Nicholson AG, Hansell DM, Harris J, Colby TV, du Bois RM. A histologic pattern of nonspecific interstitial pneumonia is associated with a better prognosis than usual interstitial pneumonia in patients with cryptogenic fibrosing alveolitis. Am J Respir Crit Care Med 1999;160: Padley SP, Padhani AR, Nicholson A, Hansell DM. Pulmonary sarcoidosis mimicking cryptogenic fibrosing alveolitis on CT. Clin Radiol 1996;51: Wells AU, Cullinan P, Hansell DM, Rubens MB, Black CM, Newman- Taylor AJ, du Bois RM. Fibrosing alveolitis associated with systemic sclerosis has a better prognosis than lone cryptogenic fibrosing alveolitis. Am J Respir Crit Care Med 1994;149: Chinet T, Dusser D, Labrune S, Collignon MA, Chretien J, Huchon GJ. Lung function declines in patients with pulmonary sarcoidosis and increased respiratory epithelial permeability to 99mTc-DTPA. Am Rev Respir Dis 1990;141: Wells AU, Rubens MB, du Bois RM, Hansell DM. Serial CT in fibrosing alveolitis: prognostic significance of the initial pattern. Am J Roentgenol 1993;161: Johnson MA, Kwan S, Snell NJ, Nunn AJ, Dabyshire JH, Turner-Warwick M. Randomised controlled trial comparing prednisolone alone with cyclophosphamide and low dose prednisolone in combination in cryptogenic fibrosing alveolitis. Thorax 1989;44: Cox DR, Oakes D. Analysis of survival data. London: Chapman and Hall; Kitaichi M. Pathologic features and the classification of interstitial pneumonia of unknown etiology. Bull Chest Dis Res Inst Kyoto Univ 1990; 23: Moolman JA, Bardin PG, Rossouw DJ, Joubert JR. Cyclosporin as a treatment for interstitial lung disease of unknown aetiology. Thorax 1991;46: Ognibene FP, Masur H, Rogers P, Travis WD, Suffredini AF, Feuerstein I, Gill VJ, Baird BF, Carrasquillo JA, Parillo JE, et al. Nonspecific interstitial pneumonitis without evidence of Pneumocystis carinii in asymptomatic patients infected with human immunodeficiency virus (HIV). Ann Intern Med 1988;109: Collins CD, Wells AU, Hansell DM, Morgan RA, MacSweeney JE, du Bois RM, Rubens MB. Observer variation in pattern type and extent of disease in fibrosing alveolitis on thin section computed tomography and chest radiography. Clin Radiol 1994;49: Hanson D, Winterbauer RH, Kirtland SH, Wu R. Changes in pulmonary function test results after 1 yr of therapy as predictors of survival in patients with idiopathic pulmonary fibrosis. Chest 1995;108: Crystal RG, Gadek JE, Ferrans VJ, Fulmer JD, Line BR, Hunninghake GW. Interstitial lung disease: current concepts of pathogenesis, staging and therapy. Am J Med 1981;70: Dunnill MS. Pulmonary fibrosis. Histopathol 1990;16: Fishman AP. UIP, DIP and all that. N Engl J Med 1978;298: Carrington CB, Gaensler EA, Coutu RE, Fitzgerald MX, Gupta RG. Natural history and treated course of usual and desquamative interstitial pneumonia. N Engl J Med 1978;298: Hartman TE, Primack SL, Kang EY, Swensen SJ, Hansell DM, McGuinness G, Muller NL. Disease progression in usual interstitial pneumonia compared with desquamative interstitial pneumonia: assessment with serial CT. Chest 1996;110: Akira M, Yamamoto S, Hara H, Sakatani M, Ueda E. Serial computed tomographic evaluation in desquamative interstitial pneumonia. Thorax 1997;52: Yousem SA, Colby TV, Gaensler EA. Respiratory bronchiolitis-associated interstitial lung disease and its relationship to desquamative interstitial pneumonia. Mayo Clin Proc 1989;64: