INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Nonspecific interstitial pneumonia/fibrosis (NSIP) has recently been defined by Katzenstein and Fiorelli (1) as an idiopathic interstitial pneumonia with a pathologic pattern distinct from usual, desquamative, and acute interstitial pneumonia (Hamman-Rich syndrome) (2). The pathologic features of NSIP include interstitial inflammation and/or fibrosis, the proportion of which could vary over a spectrum ranging from pure chronic inflammatory cell infiltrate to fibrosis without any interstitial cellular infiltration. However, the most characteristic finding in NSIP is that the lesions are temporally uniform, as opposed to the temporal heterogeneity seen in usual interstitial pneumonia (UIP). Although NSIP was defined by exclusion of the cases that cannot be pathologically categorized into one of the three abovementioned well-defined types of idiopathic interstitial pneumonia, it appeared to have some distinguishing clinical features from UIP, and more importantly, a better prognosis (1, 6). NSIP was also characterized by the possible association with etiologic factors, including underlying connective tissue disease, inhalation of organic dust, or slowly resolving acute lung injury.

The aim of this study was to describe the clinical, radiologic, and evolutionary profile of NSIP in a retrospective series of 12 patients in order to determine if NSIP really deserves to be individualized as a clinicopathologic entity.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Selection of Cases

We reviewed the pathologic slides of the cases of idiopathic interstitial pneumonia from the files of the Department of Pathology of Louis Pradel Hospital that had been classified as UIP, desquamative interstitial pneumonia, acute interstitial pneumonia, or any other descriptive unspecific diagnosis (thus excluding for example cases labeled as bronchiolitis obliterans organizing pneumonia). We selected the cases with temporally uniform lesions consisting of cellular interstitial inflammatory infiltration and/or fibrosis, consistent with a pathologic diagnosis of NSIP (1).

Clinical Analysis

For all patients we retrospectively reviewed the clinical records, pulmonary function tests, and chest X-ray films simultaneous with lung tissue sampling. High-resolution computed tomography (HRCT) of the chest was reviewed by three of us (D.R., V.C., J.F.C.). Exposures to inhaled organic antigens were investigated, as well as their chronologic correlation with pulmonary symptoms during the follow-up period. We also reviewed routine laboratory tests, including erythrocyte sedimentation rate, C-reactive protein level, and differential blood cell counts, as well as bronchoalveolar lavage (BAL) cell counts and electrophoretic study of the proteins of the sera, which were performed in all patients. Serologic titers for organic antigens were also measured in all patients. In some of the patients the serum complement level, circulating immune complexes, Latex and Waaler Rose reactions, and auto-antibodies, including anti-nuclear antibodies, antibodies against soluble nuclear antigens, and antineutrophil cytoplasmic antibodies, were measured.

Pathologic Analysis

Bouin-fixed tissue sections, 4-µm thick, cut and routinely stained with hematoxylin and eosin, were reviewed by two pathologists (R.L., A.V.D.). The cases were classified into one of three distinct patterns according to the relative amounts of inflammation and fibrosis, as proposed by Katzenstein and Fiorelli (1). Briefly, group 1 was characterized by a cellular interstitial pneumonia with relatively little fibrosis; group 2 exhibited cellular interstitial infiltrates with a significant amount of fibrosis; and group 3 was characterized by dense interstitial collagen deposition or scarring, resulting in more severe derangement of lung architecture.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

We identified 12 consecutive patients with NSIP, which was diagnosed based on the pulmonary specimen obtained by open (n = 6) or thoracoscopic (n = 6) lung biopsy in our institution between 1979 and 1996. All the patients underwent biopsy as a diagnostic procedure for idiopathic interstitial lung disease. All the patients had been asked for any possible organic exposure, including occupational or domestic exposure to birds, fungus, etc., and none of them clearly fulfilled the clinical criteria for a diagnosis of hypersensitivity pneumonitis (HP) (7) before the biopsy. None of them had a well-characterized connective tissue disease at the time of pulmonary biopsy.

Pathologic Analysis

By definition, the lung biopsy specimens of the 12 cases were characterized by a temporally uniform interstitial inflammatory infiltration and/or fibrosis sparing the alveolar spaces. Three cases were identified as group 1, eight cases as group 2, and one as group 3. Although particularly conspicuous in group 1, interstitial cell infiltrates were present in all cases, consisting mainly in lymphocytes and plasma cells. The distribution of the infiltrate was patchy in five cases (42%) and diffuse in seven cases (58%). In six cases, a peribronchiolar accentuation of the lymphoid infiltrates was present (50%). Fibrosis of all alveolar septa was found in cases from groups 2 and 3. The alveolar lumen was respected in all cases.

Focal bronchiolitis obliterans-organizing pneumonia was present in five cases (42%). Fibroblast foci were not observed. Marked hyperplasia of type 2 pneumocytes lining the alveolar lumen was noted in nine cases (75%). Nodular lymphoid aggregates with or without germinal centers were found in eight cases (66%). Small, loosely formed granulomas consisting of a cluster of epithelioid cells, giant cells, and cholesterol clefts were seen in six cases (50%). Focal mural bronchiolitis was observed in 7 cases (58%). Adventitial vascular fibrosis was marked in eight cases (66%).

Clinical Analysis

Characteristics of patients. The patients were six males and six females, aged 52.5 ± 11.8 yr (Table 1). Six (50%) were current or ex-smokers. The mean interval between the onset of disease and the pathologic diagnosis was 31 ± 58 mo (range, 1- 164); two patients presented moderate symptoms for several years (isolated cough, followed by slowly progressive dyspnea), leading to a long interval between the onset of disease and the biopsy. Two of the patients (Patients 10 and 5) had already received corticosteroid treatment at the time of the lung biopsy, for 2 mo and 12 yr, respectively. The main symptom was dyspnea, which was present in all patients. Other common symptoms were cough (67%), fatigue (58%), and weight loss (42%). The mean weight loss was 6.6 ± 4.2 kg (range, 3-12). Inspiratory crackles were present in 92% and squeaks in 25%.

Etiologic inquiries and diagnosis. An underlying connective tissue disease was eventually diagnosed in three patients, including rheumatoid arthritis (Patient 7), primary Sjögren's syndrome (Patient 8), and idiopathic inflammatory myopathy (Patient 9). In Patients 8 and 9, the pulmonary symptoms preceded the connective tissue disease by 8 yr and a few months, respectively. In one additional case (Patient 4), overlapping connective tissue disease was also suspected (but not confirmed) because of the association of arthralgia, discrete sclerodactylia, complaints of dryness of the mouth and eyes (but negative Schirmer test), and biologic abnormalities (Table 2). HP was eventually diagnosed in two patients with occupational exposure (Patients 10 and 11). Patient 10 was a farmer who ground barley in the open air and was occasionally exposed to moldy hay. Patient 11 was exposed to cheese fungus during his work washing cheese, and precipitins against Penicillium were slightly positive. Although both cases were not typical of HP to inhaled organic antigens, this was the clinical diagnosis eventually admitted. Some occasional exposures were found in seven other patients, but without chronologic correlation between the exposure and the onset of the pulmonary disease. Two successive episodes of acute respiratory distress syndrome requiring endotracheal intubation and mechanical ventilation had occurred in Patient 12, 2 yr and 12 wk, respectively, before the diagnosis of NSIP. Extensive investigations were performed, showing no evidence of any etiologic agent, except late seroconversion for Legionella pneumophila (with titer increasing from 1/32 to 1/256), but this pathogen was not isolated from BAL. Five patients were receiving drugs, some of which are known to cause lung toxicity, but in none of them could iatrogenic pneumonitis ultimately be diagnosed, because the onset of the disease and/or relapse(s) did not chronologically correlate with drug therapy or with a modification of the treatment.

In summary, NSIP was idiopathic and solitary in half of the patients. The other cases were associated with connective tissue disease (Patients 7-9) and possible etiologic conditions, including significant organic dust exposure (Patients 10 and 11) and prior acute lung injury (Patient 12).

Pulmonary function tests. All patients received a spirometry test and measurement of total lung capacity (TLC) (Figure 1). Transfer factor (TLCO) and transfer coefficient (KCO) were measured in 11 of 12 patients (Table 3). A restrictive ventilatory defect as shown by impairment of TLC was found in 11 patients (92%). In addition, five patients (42%) had impairment of FEV₁/FVC and FEF_25-75/FVC values, suggesting a slight obstructive defect. Impairment of TLCO was found in all 11 tested patients (100%), while KCO was lowered in only 5 of 11 patients (45%). Seven patients (58%) were hypoxemic at rest (Pa_O2 less than 10 kPa). At exercise the Pa_O2 significantly decreased by 0.5 kPa or more in eight patients (67%). Increased alveolar-arterial oxygen tension (Aa_O2) while breathing 100% oxygen (Aa_O2 > 18 kPa) was found in 2 of 11 patients tested (both had peribronchial patchy areas of alveolar consolidation as a predominant feature on HRCT (see below).

View larger version (33K): [in this window] [in a new window]   Figure 1.   Pulmonary function tests at diagnosis and at last measurement.

Imaging findings. Chest X-ray showed diffuse infiltrative opacities in all cases (Figure 2). Eleven patients underwent HRCT (Table 4), showing parenchymal opacities involving both lungs in all cases, symmetrically in most cases (n = 9), and left or right lung preferentially in two cases. The lower lobes were more frequently involved in parenchymal lesions in eight cases, and upper, middle, and lower lobes were equally involved in the other cases. The most common finding on HRCT was bilateral ground-glass opacities present in nine patients (82%); in two patients there was juxtaposition of segments involved and uninvolved by these ground-glass opacities. Alveolar opacities were seen in seven patients (64%), but a typical pattern of alveolar consolidation was observed in only one patient, whereas irregular alveolar opacities with a patchy and/or peribronchovascular distribution were seen in seven patients (Figure 3). Abnormal thickening of the interlobular and intralobular septa was seen in five (45%) and two (18%) patients, respectively. Linear nonseptal opacities were noted in five patients (45%). Irregularities of septal and pleural lines were found in five patients (45%). Localized subpleural honeycombing was present in only one patient (Patient 3). Slight nodular opacities were noted in three patients (27%) but were not prominent. Bronchial dilation and thickening of the bronchial walls were present in one patient. Enlarged mediastinal lymph nodes were observed in one patient. Pleural effusion was not present in any patient.

View larger version (146K): [in this window] [in a new window]   Figure 2.   Bilateral infiltrative opacities predominating in the lower lobes on chest X-ray (Case 8).

View larger version (160K): [in this window] [in a new window]   Figure 3.   High-resolution computed tomography demonstrating ground-glass opacities in the middle and right lower lobes, with patchy areas of alveolar opacities of the right lower lobe (Case 4).

Biology. The BAL differential cell pattern was markedly lymphocytic (lymphocytes more than 25%) in four patients, markedly neutrophilic (neutrophils more than 10%) in four patients and mixed in four patients (Table 2).

Examination and cultures of lung tissue for bacteria, mycobacteria, fungi, parasites, and viruses were negative in all patients. The same studies on BAL fluid revealed Pseudomonas in Patient 1, Streptococcus pneumoniae in Patient 2, Coronavirus in Patient 10, and Haemophilus in Patient 11, but these findings were not confirmed by further studies on subsequent BAL. Serology for human immunodeficiency virus-1 (HIV-1) and hepatitis B virus were negative in the eight patients studied, except Patient 5 who had isolated anti-HBc antibodies.

None of the patients had peripheral eosinophilia. The C-reactive protein level was higher than 5 mg/L in five of nine patients tested. The fibrinogen concentration was elevated in three patients. Alpha-2 globulin concentrations were higher than the normal range in four patients. High levels of seric complement C1q were found in two of five patients studied, of C3 in one of nine patients, of C4 in none of seven patients, and of CH50 in six of nine patients. Precipitins for bird antigens were found in two patients, for hay in three patients, and for Penicillium in one patient. Antibodies against native desoxyribonucleic acid were absent in all eight patients studied, as were anti-Sm, anti-RNP, anti-SSb (anti-La), anti-endoplasmic reticulum, anti-sulfate oxidase, and anti-mitochondria antibodies. Antineutrophil cytoplasmic antibodies were negative in all nine patients studied. Angiotensin-converting enzyme levels were within the normal range in the sera of the four patients studied. Creatine phosphokinase and aldolase levels were increased in two of five patients studied. Serum creatinine was normal in all patients.

Other. Schirmer test was positive in two of four patients tested (Patients 8 and 9). Nailfold capillaroscopy was performed in four patients, showing abnormalities of the capillary vessels consistent with microvasculitis in Patient 4. Ten patients underwent cardiac echographic evaluation, showing slightly elevated systolic pulmonary arterial pressure in two patients (35 and 45 mm Hg) and moderate aortic and mitral leak in Patient 3.

Treatment and clinical outcome. All 12 patients were initially treated with corticosteroids alone, with a daily dose of 60 mg prednisone or prednisolone orally for 2 (two patients), 4 (two patients), or 6 wk (eight patients), followed by a progressive decrease of the daily dose. Seven patients (58%) subsequently received immunosuppressive agents, because of incomplete clinical recovery with corticosteroids alone or in order to decrease the daily dose of corticosteroids required to control disease activity. Immunosuppressive agents included azathioprine (n = 6), cyclophosphamide (n = 2), methotrexate (n = 1), chlorambucil (n = 1), colchicine (n = 1), and hydroxychloroquine (n = 1). One patient also received intravenous polyclonal gamma globulins as a long-term treatment of her idiopathic inflammatory myopathy (this was started long after pulmonary disease had improved).

Mean duration of follow-up after the pathologic diagnosis was 50.2 ± 40.2 months (range, 10-159) (Table 1). The whole duration of the survey from the onset of the symptoms to the last follow-up was 81.5 ± 63.4 mo (range, 16-198). All patients were alive at last follow-up. Ten patients (83%) improved according to clinical and lung function test criteria, whereas disease progressed in the remaining two patients (Patients 1 and 5), leading to chronic respiratory insufficiency (with 42 and 34 mo of follow-up, respectively) (Table 4). The poor outcome in these two patients could not be correlated with specific clinical or radiologic features at the time of the diagnosis of NSIP (Table 4). In Patient 1, the radiologic features eventually evolved to a pattern similar to that seen in patients with UIP, with subpleural accentuation of the interstitial lesions, but without honeycombing. In Patient 5, the CT scan of the chest showed the association of ground-glass opacities, thickening of intralobular septa, and patchy peribronchiolar opacities. The biopsies in these two patients were of good size and quality. Patients 1 and 5 were identified as pathologic groups 1 and 3, respectively.

Clinical improvement was partial in three patients and complete in seven, with only moderate functional and/or radiologic sequelae in six of them. The evolution of pulmonary function parameters (Table 3) is illustrated in Figure 2. Interestingly, relapses of NSIP occurred in four patients (33%), either while decreasing (n = 3) or after stopping (n = 1) the treatment. Multiple relapses occurred in three patients. In Patient 8, three consecutive relapses occurred while slowly decreasing the corticosteroid treatment below a daily dose of 10 mg prednisone, but adding hydroxychloroquine allowed us to decrease prednisone to below 5 mg/d. In Patient 9, a complete clinical recovery from the pulmonary disease was obtained without any relapse, despite multiple relapses of the muscular symptoms of chronic idiopathic inflammatory myopathy.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

It is clear that NSIP is not a new pulmonary disease. It has previously received other names (as nonclassifiable interstitial pneumonia), and it may be recognized in older descriptions of interstitial lung disease, either idiopathic or not. However, NSIP has been individualized only recently by Katzenstein on precise characteristic pathologic criteria (1, 2). Since the definition of NSIP was first based on the exclusion from the other groups by pathologic criteria, one can expect its pathologic and clinical heterogeneity. However, it might also represent a clinically relevant distinct entity. This study was an attempt to identify common features, allowing NSIP to be considered as a distinct entity.

By definition, the pathologic lesions in our patients were temporally uniform. Temporal uniformity is the main feature differentiating NSIP from UIP (8); other differences between these entities include the prominence of the interstitial inflammatory lesions in NSIP, with the absence of honeycomb areas.

The clinical presentation (1, 9, 10) and the findings of the lung function tests were characteristic of interstitial lung disease. The HRCT features were heterogeneous, including ground-glass opacities, patchy alveolar opacities, linear nonseptal opacities, and thickening of the septal and pleural lines. Although none of these are specific, their combination formed a pattern distinguishing the imaging features in most of these patients from those of the other interstitial lung disorders, and especially HP and UIP. Diffuse ground-glass opacities evocative of HP were not as prominent as in HP and were associated with alveolar opacities. Subpleural honeycombing was present in only one patient and was localized. Although imaging features of the idiopathic cases were heterogeneous, the most striking pattern was the association of patchy and peribronchovascular areas of alveolar consolidation with the interstitial images. The pattern of bilateral patchy areas of opacity was also reported by Park and colleagues (10), and is probably related to the patchy peribronchiolar distribution of the pathologic lesions (2). Similar findings have also been reported in patients with polymyositis and pulmonary involvement (11).

Some factors were considered to play a definite etiologic role in half of the patients only after a pathologic diagnosis was made. Three patients eventually developed an associated connective tissue disease. Cellular interstitial pneumonia has been reported as one of the pathologic patterns of pulmonary involvement in rheumatoid arthritis (12) and polymyositis (13). NSIP was also noted in patients with various connective tissue diseases (1). Interestingly, NSIP antedated the systemic manifestations in two patients, showing that NSIP may be the first manifestation of a connective tissue disease and precede the characteristic features allowing its recognition and diagnosis.

All the patients in this series had been asked about any possible pathogenic exposures before the lung biopsy, but none of them convincingly fulfilled the clinical criteria for HP (7). However, investigations for possible exposures were performed or further pursued after the pathologic diagnosis. Significant occupational exposures to organic dust were found and thought to be etiologic in two cases. Moreover, slight-to-moderate occasional organic exposures were found in six additional patients, but whether these exposures played a role in the pathogenesis of the disease is unknown. Thus, a NSIP pattern is sometimes associated with organic exposure. None of the cases described here exhibited the pathologic association of granulomas, cellular bronchiolitis, and interstitial infiltrates, a triad evocative of HP (14), though not always present in patients with a clinical diagnosis of HP (15). However, diagnosing HP remains difficult in some cases (16), and NSIP might be an occasional atypical pattern of unrecognized HP (1).

Finally, one patient had a recent history of acute respiratory distress syndrome, confirming that NSIP might be related to acute lung injury and slowly resolving diffuse alveolar damage in some cases (1). However, despite careful investigation, no associated disorder or definite exposure was found in half of the cases, and these remained idiopathic.

Most patients in this series responded to treatment, but the response was not as dramatic as it is in bronchiolitis obliterans-organizing pneumonia (17). It usually required a daily dose of 1 mg/kg of corticosteroids for at least 4 to 6 wk, as well as immunosuppressors in some cases. Relapses occurred while decreasing treatment, especially in patients with idiopathic NSIP. No mortality was observed over a follow-up period of more than 6 yr. The favorable outcome in this and other series (1, 6, 9, 10) is in contrast to that in patients with UIP, in whom the prognosis is poor and only a small proportion of the patients respond to therapy (18, 19). The good prognosis of NSIP may be explained at least in part by the predominance of cellular inflammation over interstitial fibrosis, a feature that has been previously associated in IPF with a better response to corticosteroid therapy and a more favorable long-term prognosis (20). Since in the latter studies NSIP was not distinguished from UIP, some cases with a good response to corticosteroids might indeed correspond to NSIP. Similarly, BAL lymphocytosis was present in 8 of 12 of our patients with NSIP, and has been associated with a better prognosis in IPF at a time when NSIP was not distinguished from UIP (26). Finally, the better prognosis of NSIP also may be related in some cases to its association with connective tissue disease, where the prognosis of interstitial lung involvement has been described as better than that of IPF (29), despite some exceptions (13, 34, 35). In two patients, the disease progressed, leading to chronic respiratory insufficiency, and radiologic features compatible with end-stage IPF in one of them. Although the biopsies were of good size and quality, we cannot rule out the possibility that these cases might represent early UIP with inadequate and nonrepresentative biopsies, as has been suggested (5, 8).

This study shows that in patients with apparently idiopathic interstitial lung disease and NSIP on lung biopsy, etiologic factors may be found in some patients after careful evaluation or during follow-up. However, NSIP remains idiopathic in a number of patients. Most patients respond to treatment, and prognosis is good, in contrast with UIP. We thus consider that idiopathic NSIP deserves to be individualized as a distinct clinicopathologic syndrome.