|Home | About | Journals | Submit | Contact Us | Français|
In the American Thoracic Society/European Respiratory Society consensus classification, idiopathic interstitial pneumonias are classified into seven clinicopathologic entities. The classification is largely based on histopathology, but depends on the close interaction of clinician, radiologist, and pathologist. An accurate diagnosis can be very difficult, especially when deciding between idiopathic pulmonary fibrosis and fibrotic nonspecific interstitial pneumonia; better diagnostic markers are needed. The prognosis of idiopathic pulmonary fibrosis is very poor, with median survival of 2–4 yr after the diagnosis, yet the course of individual patients is highly variable. Predicting prognosis in the individual patient is challenging but various clinical and radiologic variables have been identified. According to several recent clinical trials, the natural history of this disease may involve periods of relative stability punctuated by acute exacerbations of disease that result in substantial morbidity or death. Nonspecific interstitial pneumonia is characterized by a distinct histopathologic appearance and a better prognosis than idiopathic pulmonary fibrosis. However, there is still confusion and controversy over the relationship between idiopathic pulmonary fibrosis and fibrotic nonspecific interstitial pneumonia.
The idiopathic interstitial pneumonias (IIPs) are a group of diffuse parenchymal lung diseases of unknown etiology with varying degrees of inflammation and fibrosis. In 1969, Liebow and Carrington published a landmark histopathologic classification schema for the IIPs consisting of five patterns: usual interstitial pneumonia (UIP), bronchiolitis obliterans interstitial pneumonia and diffuse alveolar damage, desquamative interstitial pneumonia, lymphocytic interstitial pneumonia (LIP), and giant cell interstitial pneumonia (1). Katzenstein and Myers emphasized the need for consistent histopathologic criteria for the classification of the IIPs and incorporated several key distinguishing features: the temporal heterogeneity of inflammation and fibrosis, the extent of inflammation, the extent of fibroblastic proliferation, the extent of accumulation of intraalveolar macrophages, and the presence of honeycombing or hyaline membranes (2).
The histopathologic classification of the IIPs has evolved over time, most recently codified in the American Thoracic Society/European Respiratory Society (ATS/ERS) 2002 consensus classification statement (3) (Table 1). This classification separates the IIPs into seven clinicopathologic entities (in order of relative frequency): idiopathic pulmonary fibrosis (IPF; frequency, ~ 47 to 64%), nonspecific interstitial pneumonia (NSIP; frequency, ~ 14 to 36%), respiratory bronchiolitis–associated interstitial lung disease, respiratory bronchiolitis–associated interstitial lung disease/desquamative interstitial pneumonia (frequency, ~ 10 to 17%), cryptogenic organizing pneumonia (frequency, ~ 4 to 12%), acute interstitial pneumonia (frequency < 2%), and LIP (frequency < 2%). Although the histopathologic patterns provide the basis for the ATS/ERS classification, it is no longer the “gold standard” for the classification of the IIPs (3). The ATS/ERS consensus statement emphasizes the importance of dynamic interactions among clinicians, radiologists, and pathologists to arrive at a final clinico-radiologic-pathologic diagnosis. This new multidisciplinary gold standard for the diagnosis of the IIPs is a combined clinico-radiographic-pathologic analysis arrived at by a dynamic, integrated process among clinicians, radiologists, and pathologists (when a surgical lung biopsy is available) that improves diagnostic confidence and interobserver agreement and should be routinely employed in the diagnosis of the IIPs (3, 4).
The discussion of all seven forms of IIP is beyond the scope of this short review; several reviews on these entities have been recently published (5–8). Table 2 summarizes some of the key features of the IIPs. The pathologic, physiologic, and radiologic aspects of these diseases and the approach to treatment are discussed in more detail by other authors in this series. This review will focus on IPF and NSIP.
IPF is the most common interstitial lung disease of unknown etiology. According to the current ATS/ERS definition, IPF is a distinctive type of chronic fibrosing interstitial pneumonia of unknown cause limited to the lungs and associated with a surgical lung biopsy showing a histopathologic pattern of UIP (see Visscher and Myers, pages 322–329).
IPF occurs worldwide and the majority of the patients reported are white (10–14). IPF mainly affects people over 50 yr of age; approximately two-thirds are over the age of 60 yr at the time of presentation. The incidence is estimated at 10.7 cases per 100,000 per year for males and 7.4 cases per 100,000 per year for females (15, 16). The prevalence of IPF is estimated at 20/100,000 for males and 13/100,000 for females (15, 16). The majority of patients with IPF are current or former smokers, and cigarette smoking has been identified in some studies as a risk factor for developing IPF (12).
The pathogenesis of IPF remains unknown. During the past decade, there has been a shift away from the pathogenesis theory of generalized inflammation progressing to widespread parenchymal fibrosis toward a paradigm of disordered fibroproliferation and alveolar epithelial cell function (17). A complete discussion of the current understanding of the pathogenesis of IPF is found elsewhere in this monograph.
Most patients present with gradual onset (> 6 mo) of dyspnea and/or a nonproductive cough. Constitutional symptoms are rare. Bibasilar fine inspiratory crackles (so-called Velcro crackles) are the most frequent physical examination finding and digital clubbing is seen in 25 to 50% of patients. Features of right heart failure and peripheral edema develop late in the clinical course.
Pulmonary function tests often reveal restrictive impairment (decreased static lung volumes), reduced diffusing capacity for carbon monoxide, and arterial hypoxemia exaggerated or elicited by exercise (see Martinez and Flaherty, pages 315–321).
Bronchoalveolar lavage (BAL) fluid cellular analysis shows increased neutrophils and eosinophils; lymphocytosis is not a usual feature.
The chest roentgenogram typically reveals diffuse reticular opacities (see Misumi and Lynch, pages 307–314). The chest radiograph lacks diagnostic specificity. In addition, the chest radiograph correlates poorly with the histopathologic pattern, the anatomic distribution of disease, and the severity of disease, with the exception of honeycombing, which is quite specific for IPF. High-resolution computed tomography (HRCT) is significantly more sensitive and specific for the diagnosis of IPF and has replaced conventional chest radiography as the preferred imaging method (Table 2) (19). The presence of extensive ground-glass opacities, nodules, upper lobe or mid-zone predominance of findings, and significant hilar or mediastinal lymphadenopathy should question the radiographic diagnosis of IPF. HRCT can make a confident, highly specific diagnosis of IPF in half to two-thirds of patients with IIP (20, 21).
The definitive diagnosis of IPF requires a histopathologic pattern of UIP found on surgical lung biopsy (3, 22). The specific features of UIP are described by Visscher and Myers (9). In addition to UIP pattern, the diagnosis of IPF requires the following: (1) exclusion of other known causes of interstitial lung disease, (2) characteristic abnormalities on HRCT scan, and (3) a restrictive ventilatory defect and/or impaired gas exchange (22). In the absence of a surgical lung biopsy, a presumptive diagnosis of IPF can be made based on clinical, radiologic, and physiologic criteria (Table 3). In the immunocompetent adult, the presence of all major diagnostic criteria and at least three of the four minor criteria increases the likelihood of a correct clinical diagnosis of IPF (20).
Increasingly, HRCT has been shown to provide a high degree of specificity in the diagnosis of IPF. Several studies have demonstrated that, when present, these clinical criteria are highly specific (> 90%) for the presence of UIP on surgical lung biopsy (20, 21). Their sensitivity, however, is only around 50%. Therefore, in a subset of patients with IPF, the diagnosis can be confirmed with a high degree of confidence even in the absence of surgical lung biopsy (3, 22).
In rare cases of IPF, the HRCT pattern may be more informative than the surgical lung biopsy pattern. Biopsies that show a fibrosing NSIP pattern in patients with clinical and radiographic features consistent with IPF (especially honeycomb lung) tend to behave like, and should therefore be regarded as, IPF. This further emphasizes the importance of a combined clinico-radiographic-histopathologic diagnostic approach to IPF.
The natural history of IPF is not well defined. While the course of IPF is typically described as one of relentless decline in respiratory function, the course of individual patients is highly variable and some patients remain stable for extended periods of time without treatment (Figure 1).
Estimates of survival in IPF are dependent on the time point from which they are calculated: time of first radiographic abnormality, time of onset of symptoms, or time of diagnosis. Basilar reticular opacities are often visible on chest imaging studies several years before the development of symptoms. Among asymptomatic patients with IPF (diagnosed by radiographic abnormalities found on routine chest X-ray screening and lung biopsy showing UIP), symptoms developed about 1,000 d after the recognition of the radiologic abnormality (23). Many studies have reported that the median duration of symptoms before the diagnosis of IPF is 1 to 2 yr (23–31). Thus, the diagnosis of IPF may be delayed 5 or more yr from the time of onset. Compared with survival from the time of diagnosis, median survival from the development of symptoms is approximately 48 mo shorter, illustrating the impact the time point chosen can have on survival estimates (Figure 2). The median survival in most studies performed after the development of new IIP classification is between 2 and 4 yr (average, 3 yr) after the diagnosis is made, and the 5-yr survival is between 20 and 40% (24–31). Survival of the prevalent cases is longer than that of incident cases, likely because prevalence studies exclude those who died quickly of aggressive disease (i.e., survival bias) (32).
New insight into the natural history of IPF comes from several recent well-designed clinical trials (33–35). Although these trials confirmed the high mortality associated with this disease, the mean change in lung function was surprisingly small. In a randomized, placebo-controlled trial of IFN-γ1b, subjects who survived to Week 72 had a decrease in mean FVC % predicted from 64.5 to 61.0%, and a decrease in mean DlCO% predicted from 37.8 to 37.0% (33). Similarly, a randomized, placebo-controlled trial of acetylcysteine demonstrated a mean reduction of VC during 12 mo of 190 ml (7.5% of baseline) and mean decline in DlCO of 0.70 mmol/min/kPa. (13.3% of baseline) in the placebo group (34). Finally, in a randomized, placebo-controlled trial of pirfenidone, a decline in VC of only 130 ml was observed in the placebo group at 9 mo (35).
The discordance between the rate of decline in pulmonary function and mortality suggests that the clinical course of IPF may be less a gradual decline and more a steplike process, with periods of relative stability punctuated by periods of acute decompensation that may be associated with high mortality (Figure 1). These acute decompensations have received increasing attention over the last several years and have been termed acute exacerbations of IPF.
Acute deterioration in IPF—that is, an abrupt and unexpected worsening of the underlying lung disease—may occur secondary to infections, pulmonary embolism, pneumothorax, or heart failure (36). Often, however, there is no identifiable cause for the acute decline, and these episodes are called “acute exacerbations” or “accelerated phase” of IPF (37–40).
Acute exacerbations of IPF are increasingly recognized as common and highly morbid clinical events (35, 37–44) (Table 4). Martinez and colleagues retrospectively analyzed the course of 168 patients with IPF (45). Over a median period of 76 wk, 21% of these patients died, and 47% of these deaths followed an acute deterioration in the patient's clinical status. Azuma and colleagues prospectively reported 35 patients with untreated IPF and found a 14% incidence (five cases) of acute exacerbation (35). The impact of acute exacerbation on mortality was unclear as the number of cases was small and only one of the patients died. Rice and colleagues in a review of autopsy findings conclude that acute exacerbations associated with a histopathologic pattern of diffuse alveolar damage may be a common terminal event (46). Recent studies have suggested mortality rates from 20 to 86% (35, 41–43).
Acute exacerbation of IPF has not been well defined clinically. Most diagnostic criteria are modeled after those of Kondoh and include acute subjective worsening, new radiographic abnormalities, and evidence of impaired gas exchange (38). Because of the importance of acute exacerbations of IPF to our understanding of the natural history the disease, a consensus definition should be developed and the etiology, risk factors, pathogenesis, treatment, prognosis, and predictors need to be studied. We propose that the definition of acute exacerbation of IPF require subjective and objective evidence of deterioration and exclusion of common alternative etiologies, such as infection, pulmonary embolism, heart failure, and pneumothorax (Table 5).
Predicting survival based on baseline clinical variables has proven challenging. A number of parameters at the time of diagnosis have been proposed as predictors of worse survival in IPF: increasing age, male sex, degree of dyspnea, smoking history, severity of lung function and radiographic abnormality, neutrophilia or eosinophilia on BAL, honeycombing on HRCT, and the extent of fibroblastic foci on surgical lung biopsy specimens (24, 47, 48). Among the demographic and physiologic variables, patient age at the time of diagnosis and reduced lung function appear to be relatively consistent prognostic markers, but even these are inconsistent.
A study by Flaherty and colleagues suggests that the HRCT pattern adds important prognostic information. Among patients with IPF, a HRCT consistent with definite IPF was associated with worse survival than a HRCT that was indeterminate (median survival, 2.08 [95% confidence interval (CI), 1.30–3.98] vs. 5.76 [95% CI, 4.03 to not available], respectively) (49).
Recently, oxygen desaturation during 6 min of walking has been demonstrated to predict survival in IPF (50). End of exercise SpO2, change in SpO2 with exercise, walk distance, and walk velocity were also correlated with survival in a 5-yr follow-up study in patients with IPF (51). Considering its simplicity and practicality, the prognostic value of this test should be investigated further.
Composite indices containing multiple parameters may be more accurate than single parameters. King and colleagues generated a clinico-radiologic-physiologic score from a large cohort of patients with biopsy-proven IPF using clinical, radiographic, and physiologic (including exercise test) features to predict survival (13). Wells and colleagues derived a composite physiologic index (CPI) from simple spirometry and DlCO and demonstrated that CPI was linked to mortality more closely than the individual pulmonary function test values (52). The CPI score is easier to generate than the clinico-radiologic-physiologic score because no radiographic scoring or exercise data are required. However, the validity of these composite indices requires more study.
Finally, several studies have shown that serial changes in dyspnea and lung function are predictive of survival (27–29, 53). Change in FVC during the initial 6 mo seems to be better than change in DlCO as a predictor of mortality in patients with moderately severe disease (27–29).
The treatment of IPF is discussed by elsewhere in this issue by Walter and coworkers (pages 330–338).
For many years, it was recognized that lung biopsy samples from some patients with IIP do not fit into any well-defined histologic pattern. In 1994, Katzenstein and Fiorelli assigned the term “nonspecific interstitial pneumonia” to this histopathologic group (55). Importantly, the NSIP pattern is found on surgical lung biopsies from nonidiopathic forms of interstitial lung disease, such as connective tissue disease, hypersensitivity pneumonitis, poorly resolved diffuse alveolar damage, LIP, and a variety of exposures. It has been suggested that NSIP may represent an injury pattern common to many settings, rather than a specific disease, and that idiopathic NSIP (i.e., the clinical condition) may represent subclinical forms of these alternative diagnoses.
The concept of NSIP as a separate disease entity is attractive, however, because it can explain some of the historical heterogeneity in the clinical behavior of IPF. Many patients previously labeled as having IPF had cellular biopsies (prominent lymphoplasmacytic inflammation), BAL lymphocytosis, a dramatic response to steroids, and better long-term prognosis. On re-review, many of these steroid-responsive, cellular, lymphocyte-predominant cases represented NSIP (i.e., their surgical lung biopsy showed NSIP pattern, not UIP pattern). Based on this observation, the ATS/ERS classification schema for the IIPs included idiopathic NSIP as a provisional clinical diagnosis and recommended further study and characterization of this condition.
Nicholson has proposed that fibrotic NSIP pattern may be a relatively “inactive” UIP pattern (56). However, in a study comparing lung biopsy and subsequent explant tissue in patients with IIP, Katzenstein and colleagues found no explants showing UIP pattern to be preceded by biopsy findings of the NSIP pattern, which was interpreted as evidence against the evolution from NSIP to IPF (57). Furthermore, surgical lung biopsy specimens of early stages from patients with IPF who were detected by CT scan showed the same temporal heterogeneity (i.e., UIP pattern) as seen in advanced cases of IPF; NSIP pattern was not observed (26). Based on these data, current evidence suggest that, in general, idiopathic fibrosing NSIP pattern represents a distinct clinical condition, and not an early (or inactive) stage of IPF (58). The finding of both UIP and NSIP pattern in multiple biopsies from the same patients (59, 60) raises the possibility that NSIP pattern may, on occasion, be a nonspecific manifestation of IPF. However, more data are required to clarify this issue.
NSIP occurs worldwide and the majority of the patients reported are white. The incidence and prevalence are unknown. The mean age of patients with NSIP is a decade younger than the patients with IPF (median age of onset is 40 and 50 yr respectively) (3, 23, 25, 29, 53). NSIP appears to be more common in women, although this needs further clarification. There is not an association with cigarette smoking.
The pathogenesis of NSIP is unknown. Many investigators suspect it to be an autoimmune disease because of the common appearance of this pattern in patients with connective tissue disease, especially systemic sclerosis and polymyositis-dermatomyositis. As mentioned previously, it is also possible that idiopathic NSIP may represent subclinical forms of other nonidiopathic interstitial lung diseases (e.g., hypersensitivity pneumonitis). Clearly, this issue needs additional study.
The clinical presentation of NSIP generally consists of the subacute onset of dyspnea or nonproductive cough. Constitutional symptoms are rare, although a low-grade fever is sometimes reported. The median duration of symptoms before diagnosis varies from 6 to 18 mo. Clubbing is found in about 10% of patients. In contrast to IPF, serologic abnormalities (antinuclear antibodies and rheumatoid factor) may be positive in low titer. More than half of the patients have an increased percentage of lymphocytes in the BAL fluid. Pulmonary function tests often reveal restrictive impairment, reduced diffusing capacity for carbon monoxide, and arterial hypoxemia exaggerated or elicited by exercise (18). The radiologic and histopathologic findings will be discussed by other authors in this issue (9, 19).
As with all IIPs, the diagnosis of NSIP depends on a combination of clinical, radiologic, and histopathologic findings. NSIP is characterized pathologically by varying degrees of inflammation and fibrosis and by temporal homogeneity (i.e., uniformity of fibrosis), which is distinct from UIP pattern. Nonetheless, fibrotic NSIP pattern can be difficult to distinguish from UIP pattern, and significant interobserver variability exists even among expert histopathologists (30, 61). Therefore, even though the current ATS/ERS classification is based on histopathologic pattern, additional diagnostic methods, such as microarray analysis, are needed to better distinguish the two diseases. (62).
Patients with NSIP have a better outcome than patients with IPF, with many showing improvement after treatment with corticosteroids. Importantly, there are no controlled data on which to judge the effectiveness of therapy. Retrospective reviews of patients with NSIP show around a third will improve pulmonary function with therapy and the majority of the rest will stabilize (30, 63). Patients with NSIP may require the addition of immunosuppressive agents as the initial response to corticosteroids alone may not be adequate. Patients with moderate to severe impairment are at particular risk for progressive or poorly responsive disease (53). It has been suggested that patients with fibrosing NSIP be treated with the combination of corticosteroid and immunosuppressant therapy to prevent the development of an irreversible fibrosis (44). Further studies are needed to compare the effect of corticosteroid therapy with corticosteroid and immunosuppressant combination therapy for idiopathic fibrosing NSIP (44).
The clinical course of idiopathic NSIP has not been well studied. The prognosis of cellular NSIP is excellent with few reported deaths (29, 53). The course of fibrotic NSIP is worse, with reported median survival ranging from 6 to 13.5 yr (25, 29, 49, 53, 60). Travis and colleagues reported the 10-yr survival of fibrotic NSIP to be 35% (31). Nicholson and colleagues reported a 5-yr survival of only 43% (30). Nagai and colleagues reported deterioration of lung function despite treatment in 5 of 17 patients, and 2 of them died (48). Kondoh and colleagues showed that 25% worsened or died after a mean follow-up of 92 mo (44). Further studies with long-term follow-up in larger number of patients are required to clarify the course and prognostic marker in this disease.
Following the development of the ATS/ERS consensus classification, we have been able to classify the IIPs more clearly. HRCT has become central to the diagnosis of the IIPs, and a multidisciplinary approach to the diagnosis of these conditions is paramount. IPF has the worst prognosis among the chronic IIPs. However, the clinical course of IPF may not be as indolent and steadily progressive as classically described. Periods of relative stability may instead be punctuated with acute exacerbations of disease that cause new lung injury, acute decline, and, frequently, death. Predicting prognosis in the individual patient with IPF is challenging but various clinical and radiologic variables have been identified. NSIP has been proposed as a separate entity among the IIPs. There is little evidence to support the idea that NSIP progresses to IPF; rather, NSIP may be a unique form of IIP, although further clarification of the relationship of NSIP to collagen vascular disease, hypersensitivity pneumonitis, and other nonidiopathic forms of interstitial lung disease is needed. NSIP appears more responsive to corticosteroid and immunomodulatory therapy, and has better survival than IPF.
Supported, in part, by National Institutes of Health grant U10 HL080685.
Conflict of Interest Statement: D.S.K. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. H.R.C. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. T.E.K. has served on advisory boards for Actelion (compensation: 2003 = $11,725; 2004 = $9,940; 2005 = $15,000), for InterMune (2003 = $21,000; 2004 = $15,000; 2005 = $20,000), and for GlaxoSmithKline (2004 = $12,625; 2005 = $10,000), and has served as a consultant for Nektar, Alexza, AstraZeneca, Biogen, Centocor, Fibrogen, Genzyme, Human Genome Sciences, Merck, and CoTherix, and none of these exceeded $10,000 per company per year in any of the preceding 3 yr.