INTRODUCTION — Nonspecific interstitial pneumonia (NSIP) is one type of idiopathic interstitial pneumonia (IIP). The other IIPs include usual interstitial pneumonia (UIP)/idiopathic pulmonary fibrosis, (IPF)/cryptogenic fibrosing alveolitis (CFA), desquamative interstitial pneumonia (DIP), respiratory bronchiolitis associated interstitial lung disease (RB-ILD), acute interstitial pneumonia (AIP), lymphocytic interstitial pneumonia (LIP), and cryptogenic organizing pneumonia (COP)/bronchiolitis obliterans organizing pneumonia (BOOP).

NSIP is "nonspecific" in that it presents similarly to other IIPs, but lacks the histopathologic features that characterize the individual disorders. In this review, the etiology, pathogenesis, clinical manifestations, diagnosis, treatment, and prognosis of NSIP are discussed. The pathology of NSIP is described separately. (See "Idiopathic interstitial pneumonias: Clinical manifestations and pathology".)

ETIOLOGY — The term NSIP has been associated with many medical conditions, although a causal link has not been identified. In addition, it may be idiopathic.

HIV infection — NSIP is frequently identified in patients who have AIDS [1-6]. In one observational study of 110 patients with AIDS and interstitial lung disease, NSIP was detected in 38 percent [1]. Nearly half of the patients had normal chest radiographs. A study performed since the introduction of HAART found that 7 percent of HIV patients admitted with ILD had NSIP [2].

Connective tissue disease — Several connective tissue disorders including polymyositis-dermatomyositis, rheumatoid arthritis, Sjogren's syndrome, and systemic sclerosis have been associated with NSIP [7-15]. Although not all patients with ILD in the context of a known collagen vascular disease are biopsied, several studies have shown NSIP to be a common pattern.

• The most common pulmonary presentation in an observational study of 70 patients with polymyositis-dermatomyositis (PD) and interstitial lung disease (ILD) was presumed acute or subacute antibiotic resistant community acquired pneumonia [10]. NSIP was identified in 18 of 22 patients who underwent lung biopsy. (See "Pulmonary disease in dermatomyositis and polymyositis".)

• Among 16 patients with rheumatoid arthritis and ILD, 7 had features most consistent with NSIP [15]. (See "Overview of lung disease associated with rheumatoid arthritis".)

• Of nine patients with Sjogren's syndrome and ILD, three had NSIP [14]. (See "Interstitial lung disease associated with Sjögren's syndrome".)

• In 80 patients with ILD associated with systemic sclerosis, NSIP was identified in 62 (77.5 percent) [12]. In another study of patients with systemic sclerosis, 13 of 19 had NSIP [13]. (See "Evaluation for and diagnosis of lung disease in systemic sclerosis (scleroderma)".)

The presenting manifestation of systemic sclerosis sine scleroderma (a rare illness characterized by the typical vascular features and visceral fibrosis of systemic disease but without skin sclerosis) may be dyspnea with radiographic interstitial lung disease in a pattern of either NSIP or UIP [16]. Patients usually have subtle physical exam findings such as scattered telangiectasia, Raynaud's phenomenon, esophageal reflux, pericardial effusion, and abnormal nailfold capillaroscopy. Antinuclear antibodies are typically of the nucleolar pattern. (See "Classification of scleroderma disorders", section on systemic sclerosis sine scleroderma.)

• At least one of six patients with ILD in the context of microscopic polyangiitis had a lung biopsy showing NSIP [17]. (See "Clinical manifestations and diagnosis of Wegener's granulomatosis and microscopic polyangiitis".)

Many patients who have histologic NSIP have some features of connective tissue disease, but the findings are insufficient for definitive classification. In a case-control study of 28 patients with NSIP who had no clear underlying disease, 88 percent had laboratory or clinical features of undifferentiated connective tissue disease [18]. These features included arthralgias, GERD, Raynaud's, sicca symptoms, and dysphagia, in addition to positive serologic studies (eg, rheumatoid factor, ANA, or antibodies to SSA, SSB, RNP, Jo-1, SCL-70). A larger study with prospectively collected data is needed to better define the relationship between "idiopathic" and "connective tissue related" NSIP. (See "Undifferentiated systemic rheumatic (connective tissue) diseases and overlap syndromes".)

Hypersensitivity pneumonitis — In some patients with hypersensitivity pneumonitis (also called extrinsic allergic alveolitis), surgical lung biopsy lacks the characteristic poorly formed granulomata and multinucleated giant cells and has an appearance more typical of NSIP [19]. In patients with chronic bird fancier's lung, 13 of 26 had NSIP; most had some multinucleated giant cells, but only one-fourth had granulomata [20]. (See "Classification and clinical manifestations of hypersensitivity pneumonitis (extrinsic allergic alveolitis)".)

Drug-associated NSIP — A number of drugs, including flecainide, amiodarone, methotrexate, carmustine, nitrofurantoin, statins, and chlorambucil have been associated with an NSIP pattern [21-26].

PATHOGENESIS — The understanding of the pathogenesis of idiopathic NSIP and other IIPs is evolving. A potential relationship between NSIP and usual interstitial pneumonia (UIP) is suggested by the commonality of associated diseases, exposures, and genetic mutations, in addition to the observation that individual patients can harbor histopathologic lesions of both UIP and NSIP [27,28]. The exact nature of this association is not known.

Epithelial injury and disregulated repair likely play a role in the pathogenesis of NSIP [29]. Numerous cytokines that potentially play a role have been described including: epimorphin, a cell surface-associated protein involved in epithelial morphogenesis in embryonic organs [30], matrix metalloproteinases [30,31], heat shock protein 47 [32], surfactant protein C [33-37], the coagulation system [38,39], intercellular adhesion molecule-1 [40], IL-4, IL-13, IL-18, interferon-gamma [29,41,42], the pro-fibrotic chemokine, CCL7 and CCL5 [43].

Involvement of the immune system in the pathogenesis of NSIP is suggested by the presence of lymphocytes in the alveolar septae and in bronchoalveolar lavage (BAL) [8,9,44,45]. Dendritic cells (DC), which play a role in the immune response through antigen presentation, have also been noted in greater numbers in biopsies of patients with NSIP compared to UIP and were seen in close proximity to CD4 and CD8 lymphocytes [46].

Fibroblasts are believed to be a key effector cell in fibrotic lung diseases [47]. When comparing fibroblasts from patients with IPF/UIP to those from patients with NSIP, differences in contractility and secretion of transforming growth factor-Beta (TGF-b) and fibronectin are observed [11]. NSIP fibroblasts behave more like normal fibroblasts in these studies. These data suggest that fibroblast phenotype and other factors in the fibroblast microenvironment contribute to the fibroblast function/dysfunction in IIPs.

PATHOLOGY — The histopathology of NSIP is characterized by varied degrees of alveolar wall inflammation and fibrosis in a pattern that suggests temporal homogeneity (table 1) [48,49]. The histopathologic features do not fit the patterns of other IIPs such as usual interstitial pneumonia (UIP), desquamative interstitial pneumonia, respiratory bronchiolitis interstitial lung disease, cryptogenic organizing pneumonia, acute interstitial pneumonia, or lymphocytic interstitial pneumonia. (See "Idiopathic interstitial pneumonias: Clinical manifestations and pathology".)

Although the histopathologic features of NSIP are well defined, the differentiation of NSIP from other IIPs, particularly IPF/UIP, is difficult and even expert pathologists may disagree on a particular biopsy [50,51]. Three separate groups within NSIP have been described: Group I has primarily interstitial inflammation (figure 1), Group II has both inflammation and fibrosis, and Group III has primarily fibrosis (figure 2). Group III is distinguished from UIP by the absence of fibroblast foci and the presence of temporal homogeneity [7].

The classification of NSIP is further complicated by the fact that areas of both NSIP and UIP can be identified in the same patient when biopsies are taken from multiple locations. Studies evaluating patients with multiple lobe biopsies found a pattern of UIP in one lobe and NSIP in another lobe in 13 to 26 percent of patients [27,28]. These mixed NSIP/UIP cases had a similar prognosis to patients with IPF/UIP in all lobes.

CLINICAL ASSESSMENT — The assessment of patients with suspected NSIP requires a synthesis of clinical, radiographic, and histopathologic data [52,53]. In many patients with a history suspicious for NSIP, a high resolution CT scan will show characteristic findings, but surgical biopsy is usually necessary for confirmation. Once it is clear that the patient has ILD, it is important to characterize the pattern and look for clues regarding a possible etiology [54]. (See "Approach to the adult with interstitial lung disease" and "Idiopathic interstitial pneumonias: Clinical manifestations and pathology".)

Clinical characteristics — The most common symptoms of NSIP are dyspnea and cough that have developed subacutely over weeks to months [8,9,45]. About one-third of patients have fever or flu-like symptoms. Depending on whether an underlying collagen vascular disease is present, patients may also report dry mouth or eyes, joint or muscle pain, muscle weakness, Raynaud's phenomenon, rash, or other skin changes.

Patients should be asked about exposure to airborne organic antigens (eg, barns, birds, hot tubs, humidifiers, mold, plastic manufacture) [48]. A complete medication list should be reviewed for medications associated with NSIP and risk factors for HIV infection should be assessed. (See 'Drug-associated NSIP' above.)

Physical examination — The majority of patients have bibasilar crackles, but only 10 to 35 percent have clubbing [55]. Careful attention should be paid to identifying stigmata of collagen vascular disease. Nail bed capillaroscopy may also be helpful if systemic sclerosis is suspected.

Pulmonary physiology — Pulmonary function testing typically demonstrates a restrictive ventilatory defect and a decrease in gas transfer manifested by a decrease in DLCO and/or desaturation during ambulatory oximetry [48]. While pulmonary function testing is not essential for making the diagnosis of NSIP, monitoring FVC and DLCO is helpful for assessing disease progression, response to therapy, and prognosis [56].

Chest imaging studies — Chest radiographs typically show increased interstitial markings with a basilar predominance. Patients with suspected NSIP should also undergo high resolution computed tomography (HRCT). In addition to aiding the diagnosis, serial HRCT scans may be useful to assess disease improvement or progression [57-61]. (See "High resolution computed tomography of the lungs".)

The most frequent HRCT findings in NSIP include increased reticular markings, traction bronchiectasis, lobar volume loss, and ground glass opacification (GGO) [53,62]. Although less common than the other radiographic findings, bibasilar GGO that spares the subpleural region is the most characteristic HRCT appearance; it is often associated with manifestations of fibrosis such as volume loss, reticular pattern, and/or traction bronchiectasis. These changes predominantly occur in the lower lung zones. Honeycomb change is rare with NSIP; a predominance of honeycombing compared with ground glass attenuation is more suggestive of UIP [63-65]. The HRCT features of idiopathic NSIP compared to connective tissue disease related NSIP are similar [7-9,14,19,44,45,57-61,66-75].

Unlike IPF/UIP where HRCT can at times supplant the need for a surgical lung biopsy, the ability of HRCT to make an accurate diagnosis of NSIP is more limited [76]. This is in part related to variable inter reader interpretations of HRCT [77]. Several series have evaluated the ability of HRCT to make a diagnosis of NSIP (as confirmed by surgical lung biopsy). The accuracy ranged from 66 to 68 percent [65,70]. Given the prognostic and treatment implications of NSIP versus IPF/UIP, a surgical lung biopsy for confirmation should be strongly considered even though the HRCT suggests NSIP.

Bronchoalveolar lavage — In some reports, bronchoalveolar lavage (BAL) shows a higher percentage of lymphocytes in patients with NSIP compared with UIP [8,9,44,45], however, this is not always the case [66,78]. In addition, several other ILDs have BAL lymphocytosis so this is not specific for NSIP. Therefore, BAL cell counts cannot be utilized to differentiate NSIP from other IIPs. BAL may be helpful in excluding other causes of interstitial opacities, such as hemorrhage, infection and malignancy. As a result, BAL is not typically used for the diagnosis of ILD unless malignancy or opportunistic infection is likely, or transbronchial biopsy is being performed to obtain diagnostic tissue specimens. (See "Role of bronchoalveolar lavage in diagnosis of interstitial lung disease".)

Surgical lung biopsy — As the clinical features, radiographic appearance, and BAL are not specific for NSIP, a definitive diagnosis requires a surgical lung biopsy [48]. This can be accomplished by thoracotomy or video assisted thoracic surgery. Ideally, samples are obtained from more than one lobe. (See "Role of lung biopsy in the diagnosis of interstitial lung disease".)

TREATMENT — The first step in the treatment of NSIP is to remove any inciting drugs or inhalational exposures. This alone may be adequate treatment. For patients with NSIP complicating connective tissue diseases, specific therapies for those diseases may guide treatment of the NSIP. (See "Pulmonary disease in dermatomyositis and polymyositis" and "Overview of lung disease associated with rheumatoid arthritis" and "Interstitial lung disease associated with Sjögren's syndrome" and "Evaluation for and diagnosis of lung disease in systemic sclerosis (scleroderma)" and "Clinical manifestations and diagnosis of Wegener's granulomatosis and microscopic polyangiitis".)

The following discussion will emphasize the treatment in idiopathic NSIP. A few patients, less than 20 percent, improve or stabilize without therapy; consideration of this approach should be limited to patients with mild disease [44,66]. Patients with more severe initial disease or untreated patients who progress are treated with glucocorticoids, with or without the addition of other immunosuppressive drugs such as azathioprine, cyclophosphamide, and cyclosporine [7-9,11,44,49,63,66,68,79-81].

Assessing the response — The response to therapy should be assessed three to six months after its initiation, or sooner, if the patient has more severe initial disease or reports worsening symptoms. A favorable response includes a reduction in dyspnea and cough, radiographic clearing, and physiologic improvement, as assessed by forced vital capacity, total lung capacity, diffusion capacity for carbon monoxide, and both resting and exercise gas exchange. These issues are discussed in detail separately in the discussion on idiopathic pulmonary fibrosis; the principles also apply to NSIP. (See "Treatment of idiopathic pulmonary fibrosis".)

Glucocorticoids — The optimal dose and duration of glucocorticoid therapy is not known. Most studies of patients with idiopathic NSIP consist of a small number of patients and a variety of regimens [7-9,44,49,63,66,68,79-81].

We typically begin with prednisone at 1 mg/kg ideal body weight per day up to a maximum of 60 mg/day for one month followed by 40 mg/day for an additional two months. For patients with severe disease requiring hospitalization, some authors have used pulse intravenous methylprednisolone [80]. The usual regimen is 1000 mg/day for three days followed by oral prednisone as dosed above. Rarely, patients need additional pulse doses [80,81].

In patients who respond or stabilize with treatment, the prednisone dose is gradually tapered, aiming to reach 5 to 10 mg daily or on alternate days, by the end of 12 months, with attempted cessation after at least one year of therapy. Some patients relapse when prednisone is tapered or discontinued [8]. Such patients can be maintained for a longer period on low-dose prednisone or, if relapse occurs at higher prednisone doses, they can be treated with another immunosuppressive agent.

Glucocorticoid treatment is associated with a variety of side effects [82]. Thus, there needs to be a careful and ongoing assessment of risks and benefits. (See "Major side effects of systemic glucocorticoids".)

Pneumocystis prophylaxis is not routinely recommended for all patients taking glucocorticoids. However, in patients with underlying lung disease such as NSIP, it is reasonable to administer prophylaxis in patients on moderate to high doses of glucocorticoids (eg, prednisone 20 mg daily and above) since acquiring Pneumocystis could significantly worsen lung function. In addition, it is difficult to have to rule out Pneumocystis each time the patient experiences a decline in pulmonary function. (See "Treatment and prevention of Pneumocystis carinii (P. jirovecii) pneumonia in non-HIV-infected patients", section on 'Prophylaxis'.)

Other immunosuppressive drugs — Patients with more severe early disease, an inadequate response or intolerance to glucocorticoids, or relapsing disease have been treated with a variety of additional immunosuppressive drugs, such as azathioprine, cyclophosphamide, and cyclosporine [8,10,80,83]. Methotrexate is generally not chosen in idiopathic NSIP because of its potential for pulmonary toxicity. (See "Methotrexate-induced lung injury".)

No randomized trials have assessed these therapies. All series have included concurrent glucocorticoid therapy. The choice of additional agent varies with the clinical setting.

Azathioprine — We suggest the addition of azathioprine (AZA) in patients who have responded to glucocorticoids but need a glucocortioid-sparing agent, have an incomplete response, or have not responded, but are not rapidly progressing. AZA is usually started at 50 mg/day and increased by 25 mg increments every 7 to 14 days up to 1.5 to 2.0 mg/kg per day, but not exceeding a maximum total dose of 200 mg/day.

The most common side effects of AZA at these doses include gastrointestinal intolerance, bone marrow suppression, and infection. Divided doses may be used to decrease gastrointestinal toxicity. These and other side effects should be discussed with the patient. (See "Pharmacology and side effects of azathioprine when used in rheumatic diseases".)

The toxicity of AZA is predominantly related to the activity of thiopurine methyltransferase (TPMT) [84]. Both genotyping and functional assays for this enzyme are now available commercially. Many clinicians perform one test or another before initiating azathioprine therapy. These issues are discussed in detail separately. (See "Pharmacology and side effects of azathioprine when used in rheumatic diseases", section on 'Metabolism'.)

Azathioprine has been associated with an increased risk of bacterial and opportunistic infections, but some of the increased risk may be attributable to concomitant glucocorticoid therapy [85]. The role of prophylaxis for Pneumocystis in this setting has not been evaluated, but we suggest prophylaxis during combined therapy with glucocorticoid and azathioprine. (See "Treatment and prevention of Pneumocystis carinii (P. jirovecii) pneumonia in non-HIV-infected patients", section on 'Prophylaxis'.)

Cyclophosphamide — For patients who have severe initial disease or have progressed despite glucocorticoids and azathioprine, limited data from case series suggest that treatment with cyclophosphamide (CYC) may have a beneficial effect. Some authors have used a combination of initial pulse dose glucocorticoids and cyclophosphamide [80,81].

Although the optimal dose of CYC in NSIP is unknown, the usual oral dose is 1.5 to 2 mg/kg per day up to a maximum of 200 mg/day, given as a single dose. The dose is adjusted based on the total white blood cell count, aiming to keep it above 4000/microL. (See "Treatment of idiopathic pulmonary fibrosis" and "General toxicity of cyclophosphamide and chlorambucil in inflammatory diseases".)

CYC can also be given intravenously once a month to decrease the risk of bladder toxicity. In patients with normal renal function, the initial dose of intermittent intravenous CYC is usually 750 mg/m2 BSA (calculator 1). In patients who are obese or elderly (eg, age greater than 70 years), the initial dose should be reduced to 500 mg/m2 of BSA. The dose should also be reduced to 500 mg/m2 BSA in the setting of renal dysfunction (especially if the estimated creatinine clearance is less than 40 mL/minute). Subsequent dosing is based on response to therapy and monitoring of hematologic toxicity.

Adequate hydration should be given prior to and for 24 hours after infusion of CYC to reduce the risk of hemorrhagic cystitis. In addition, mercaptoethane sulfonate (mesna), which binds and detoxifies acrolein, is used to minimize the risk of bladder toxicity. (See "General principles of the use of cyclophosphamide in rheumatic and renal disease".)

A trial of at least three to six months of cyclophosphamide is needed to ensure an adequate opportunity for clinical response. Because of toxicity with long-term use, cyclophosphamide is usually discontinued after 6 to 12 months or transitioned to a less toxic agent such as azathioprine for maintenance.

Cyclophosphamide has important hematologic, bladder, and gonadal toxicity. These should be fully discussed with the patient. CYC has also been reported to be a risk factor for Pneumocystis pneumonia, so we suggest administering prophylaxis [86]. (See "General toxicity of cyclophosphamide and chlorambucil in inflammatory diseases" and "Treatment and prevention of Pneumocystis carinii (P. jirovecii) pneumonia in non-HIV-infected patients", section on 'Prophylaxis'.)

Cyclosporine — Data are limited on the use of cyclosporine in NSIP. In two small observational series, the addition of cyclosporine led to stabilization or clinical improvement in seven of nine patients who had not responded to prednisone alone and one patient who was refractory to both prednisone and cyclophosphamide [66,83]. Until more data are available, consideration of cyclosporine is warranted only in refractory cases. (See "Pharmacology and side effects of cyclosporine and tacrolimus".)

Lung transplant — Patients with severe NSIP that is progressive and disabling despite immunosuppressive therapy may be considered for lung transplant [87]. Three patients with fibrotic NSIP did well after living-donor lobar transplantation without recurrence of disease in the transplanted lung [88]. The indications are similar to those for other lung diseases. (See "Lung transplantation: General guidelines for recipient selection".)

PROGNOSIS — The overall prognosis and response to therapy for idiopathic NSIP is favorable compared to idiopathic pulmonary fibrosis/usual interstitial pneumonitis (IPF/UIP) [44,49,53,56,63,68,72]. Approximately two-thirds of treated patients with idiopathic NSIP are stable or improved in long-term follow-up, but a 15 to 26 percent mortality at five years is noted [8,44,48,63,66,80,81,89].

The results on serial pulmonary function testing may be a better prognostic indicator than baseline histopathology [56,74,90]. This was suggested in a series of patients with IPF/UIP (n = 61) and NSIP (n = 43) [56]. The patients with NSIP had longer median survival (56 versus 33 months), but this difference did not appear until after two years of follow-up. Histopathology (UIP versus NSIP) predicted prognosis as a baseline factor and at six months when changes in pulmonary physiology were examined. After 12 months of follow-up, performance on pulmonary function tests predicted subsequent mortality while histopathology was no longer predictive. Change in DLCO was the strongest determinant of mortality at 12 months of follow-up.

SUMMARY AND RECOMMENDATIONS

• Nonspecific interstitial pneumonia (NSIP) is a distinct clinical entity within the idiopathic interstitial pneumonias (IIP). (See 'Introduction' above.)

• The histopathologic lesion characteristic of NSIP is thickening of the alveolar septum with varying degrees of inflammatory cell opacity and fibrosis in a pattern suggesting temporal homogeneity. The fibroblast foci typical of usual interstitial pneumonia (UIP) are rarely seen. (See 'Pathology' above.)

• Disease onset is typically in the fifth or sixth decades of life, with men and women affected equally. Most patients present with the insidious onset of dyspnea and cough. (See 'Clinical assessment' above.)

• NSIP may be idiopathic or may be associated with connective tissue disease, certain drugs, human immunodeficiency virus infection, and hypersensitivity pneumonitis. (See 'Clinical assessment' above.)

• Bilateral interstitial opacities are the most typical finding on chest radiograph. High resolution computed tomography (HRCT) usually shows diffuse ground glass opacities. Volume loss, reticular pattern and traction bronchiectasis may be seen, but honeycombing is generally absent. (See 'Chest imaging studies' above.)

• We recommend histopathologic confirmation of NSIP by open or thoracoscopic lung biopsy. (See 'Clinical assessment' above.)

• Once the diagnosis of NSIP has been made, every effort should be made to identify any exposure or concomitant systemic disease that is contributory. (See 'Treatment' above.)

• Patients who have mild disease and minimal symptoms may be observed for a period of time without treatment. Symptoms and pulmonary function should be reassessed every three to six months and therapy is initiated if there is progression. (See 'Treatment' above.)

• We suggest initial treatment with glucocorticoids for the majority of patients with moderate to severe disease and for patients with mild disease whose symptoms have progressed or pulmonary function has declined (Grade 2C). The usual initial treatment is oral prednisone at 1 mg/kg/day. Pulse methylprednisolone may be used for more severe disease. Additional immunosuppressive therapy is sometimes helpful for more severe disease, relapsing disease, or as a steroid sparing agent. (See 'Treatment' above.)

• We suggest Pneumocystis prophylaxis for any patient with NSIP receiving moderate to high dose glucocorticoid (20 mg or more per day), azathioprine, or cyclophosphamide (Grade 2C). (See "Treatment and prevention of Pneumocystis carinii (P. jirovecii) pneumonia in non-HIV-infected patients", section on 'Prophylaxis' and "Major side effects of systemic glucocorticoids".)