Demystifying Idiopathic Interstitial Pneumonia

Careful histopathological evaluation has shown the traditionally clinical diagnosis of idiopathic interstitial pneumonia to be more heterogeneous than once thought. Its subclassification, based on clinicopathological criteria, has important therapeutic and prognostic implications. The most important distinction is the presence of usual interstitial pneumonia, the histopathological pattern seen in idiopathic pulmonary fibrosis. Idiopathic pulmonary fibrosis has a worse response to therapy and prognosis. New insight into the pathophysiology of idiopathic pulmonary fibrosis suggests a distinctly fibroproliferative process, and antifibrotic therapies show promise. Although the clinical and radiographic diagnosis of idiopathic pulmonary fibrosis can be made confidently in some cases, many patients require surgical lung biopsy to determine their underlying histopathological pattern. A structured, clinicopathological approach to the diagnosis of idiopathic interstitial pneumonia, with particular attention to the identification of idiopathic pulmonary fibrosis, ensures proper therapy, enhances prognosis, and allows for further investigation of therapies aimed at the distinct pathophysiology.

The diagnosis and management of idiopathic interstitial pneumonia have challenged physicians since its initial description more than a century ago. Much has changed in our understanding of idiopathic interstitial pneumonia during the past decade, most important being our appreciation of the clinical relevance of histopathology. When faced with a patient suspected of having idiopathic interstitial pneumonia, physicians must have a rigorous and organized approach to its diagnosis, as therapy and prognosis depend on the patient's underlying histopathological pattern. This fact remains unappreciated by many physicians, and as a result, patients often undergo incomplete evaluation and empirical treatment. This review aims to describe idiopathic interstitial pneumonia and its different histopathological subgroups, explain the therapeutic and prognostic importance of careful histopathological distinction, and provide an understanding of the utility of clinical, radiographic, and surgical approaches to the diagnosis.

We reviewed the epidemiology, clinical presentation, pathogenesis, diagnosis, and treatment of idiopathic interstitial pneumonia. After a search of MEDLINE, we identified articles published from 1966 to 2001 using the exploded Medical Subject Heading terms lung diseases and interstitial and pulmonary fibrosis and the keyword terms idiopathic pulmonary fibrosis, idiopathic interstitial pneumonia, usual interstitial pneumonia, desquamative interstitial pneumonia, respiratory bronchiolitis/interstitial lung disease, nonspecific interstitial pneumonia, acute interstitial pneumonia, fibrosis, and inflammation. Identified articles and additional sources such as textbook chapters and meeting abstracts were reviewed for relevant publications. All included reports underwent peer review.

Interstitial pneumonia has been recognized as a chronic progressive lung disease for more than a century. In 1892, Osler¹ described chronic interstitial pneumonia, also called cirrhosis of the lung, as "a fibrinoid change, which may have its starting point in the tissue about the bronchi and blood-vessels, the interlobular septa, the alveolar walls or in the pleura. So diverse are the different forms and so varied the conditions under which this change occurs that a proper classification is extremely difficult." An association with inorganic dust inhalation was by this time well established, but little attention was paid to clinical or pathological subcategorization.

In 1944, Hamman and Rich² described 4 cases of acute diffuse interstitial fibrosis in which "the lungs were the seats of a widespread connective tissue hyperplasia throughout the interstitial structures. The alveolar walls were tremendously thickened; in the early stages of the process crowded with fibroblasts." Although all 4 of these cases were acute in onset and rapidly progressive, the term Hamman-Rich syndrome was used for some time to describe any diffuse idiopathic fibrotic lung disease.

By the mid-1960s, significant progress in the subcategorization of interstitial lung disease had been made, with inherited conditions, collagen-vascular diseases, drugs, and occupational exposures being recognized as potential causes. Those conditions that remained idiopathic were collectively called by various names, including chronic idiopathic interstitial fibrosis, diffuse interstitial fibrosis, diffuse fibrosing alveolitis, Hamman-Rich syndrome, diffuse pulmonary alveolar fibrosis, idiopathic pulmonary fibrosis, and idiopathic interstitial pneumonia. The preferred term at present is idiopathic interstitial pneumonia, and we will use this term for the remainder of the discussion.

In 1969, Liebow and Carrington³ described the following 5 histopathological subgroups of chronic idiopathic interstitial pneumonia: usual interstitial pneumonia (UIP); diffuse lesions similar to UIP with superimposed bronchiolitis obliterans termed bronchiolitis interstitial pneumonia; desquamative interstitial pneumonia; lymphoid interstitial pneumonia; and giant cell interstitial pneumonia. Hamman-Rich disease was considered an acute form of UIP. Liebow and Carrington suggested that lung biopsy and histopathological classification might help distinguish clinically distinct conditions, as "there are correlations [of histopathological findings] with radiographic appearance, clinical course and prognosis. Clues to etiology and pathogenesis, which currently are scanty, will in time be forthcoming."³

The subgroups identified by Liebow and Carrington³ have evolved into the classification schema used at present for idiopathic interstitial pneumonia as described in 1998 by Katzenstein and Myers.⁴ Their classification schema includes 4 histopathologically distinct subgroups: UIP, desquamative interstitial pneumonia and a closely related pattern termed respiratory bronchiolitis–associated interstitial lung disease (RBILD), acute interstitial pneumonia (formerly Hamman-Rich syndrome), and nonspecific interstitial pneumonia (NSIP). The lymphoid interstitial pneumonia and giant cell interstitial pneumonia subgroups identified by Liebow and Carrington³ are excluded because their causes, a lymphoproliferative condition and a hard-metal pneumoconiosis, respectively, are no longer idiopathic. Bronchiolitis interstitial pneumonia, now known as bronchiolitis obliterans with organizing pneumonia, is also excluded because it is a predominantly intraluminal process. The Katzenstein and Myers⁴ method of histopathological categorization incorporates the temporal heterogeneity of inflammation and fibrosis, the extent of inflammation, the extent of accumulation of intra-alveolar macrophages, and the presence of honeycombing or hyaline membranes.

The American Thoracic Society and the European Respiratory Society have recently revised the classification schema of Katzenstein and Myers⁴ to emphasize the importance of an integrated clinical and pathological approach to the diagnosis of idiopathic interstitial pneumonia.⁵ The classification of the American Thoracic Society and the European Respiratory Society combines the histopathological pattern seen on lung biopsy findings (as described by Katzenstein and Myers) with clinical information to arrive at a final clinicopathological diagnosis. This approach allows for the preservation of existing histopathological and clinical terms while precisely defining the relationship between them. When the terms are the same for the histopathological pattern and the clinical diagnosis (eg, desquamative interstitial pneumonia), it was recommended that the pathologist use the addendum pattern when referring to the appearance on lung biopsy findings (eg, desquamative interstitial pneumonia pattern) and reserve the initial term for the final clinicopathological diagnosis.

The various histopathological subgroups are often clinically indistinguishable (with the exception of acute interstitial pneumonia). Although patients with UIP are generally older (two thirds are older than 60 years at diagnosis) than those with other forms of idiopathic interstitial pneumonia, age is not a reliable predictor of histopathology. Symptoms are identical for all subgroups, with the typical patient reporting progressive dyspnea for months to years and a chronic, nonproductive cough. Physical examination will often reveal digital clubbing and the presence of bibasilar fine inspiratory crackles ("Velcro" crackles). Results of pulmonary function testing reveal restrictive lung disease (decreased forced vital capacity and total lung capacity [TLC]) and abnormal gas exchange (decreased diffusion capacity of carbon dioxide and resting or exercise PaO₂). The most useful clinical tool for distinguishing between subclasses is high-resolution computed tomography (HRCT) of the chest. The diagnostic utility of HRCT is discussed later in this review.

Acute interstitial pneumonia presents much differently, with the acute onset of dyspnea and cough and the rapid development of respiratory failure. The clinical picture is usually indistinguishable from that of acute respiratory distress syndrome, with hypoxemia, diffuse crackles, and bilateral alveolar opacities present on a chest x-ray film. The age of onset is typically younger than for the other subgroups of idiopathic interstitial pneumonia. A few cases of chronic relapsing acute interstitial pneumonia have been described, usually leading to progressive deterioration in lung function.⁶

The relative distribution of the 4 histopathological subgroups can be estimated from several case series of patients with idiopathic interstitial pneumonia (Table 1).^7-10 These studies suggest that UIP is the most common (50%-60%), with NSIP (14%-36%) and desquamative interstitial pneumonia/RBILD (10%-17%) less common, and acute interstitial pneumonia quite rare (0-2%). The histopathological patterns described herein can be seen in conditions other than idiopathic interstitial pneumonia.⁴ This is especially true of NSIP, which has been described in a variety of conditions, including collagen-vascular disease, hypersensitivity pneumonitis, recent acute lung injury, and drug reactions.¹² Therefore, before the diagnosis of idiopathic interstitial pneumonia is made, a careful clinical evaluation to rule out such associated conditions is paramount.

Usual interstitial pneumonia is characterized by a heterogeneous, predominantly subpleural distribution of involvement, a pattern quite distinguishable even on low-power magnification (Figure 1A). Temporal heterogeneity is seen, with areas of end-stage fibrosis and honeycombing (thickened collagenous septa surrounding airspaces lined by bronchial epithelium) abutting areas of active proliferation of fibroblasts and myofibroblasts. These discrete areas of acute fibroblastic proliferation have been termed fibroblastic foci and are essential to the histopathological diagnosis of UIP (Figure 1B). Generally, minimal interstitial inflammation is seen, and if present in significant amounts, the histopathological diagnosis should be reconsidered.

Nonspecific interstitial pneumonia evolved as a categorization reserved for idiopathic interstitial pneumonia that did not meet the criteria for UIP, desquamative interstitial pneumonia/RBILD, or acute interstitial pneumonia. It is characterized by varying degrees of inflammation and fibrosis, with some forms primarily inflammatory (cellular NSIP) and others primarily fibrotic (fibrotic NSIP).¹² Although NSIP may have significant fibrosis, it is usually of temporal uniformity, and fibroblastic foci and honeycombing, if present, are rare (Figure 1C). The temporal uniformity is distinctly different from that seen in UIP. Nonetheless, fibrotic NSIP can be difficult to distinguish reliably from UIP, and significant interobserver variability exists even among expert histopathologists.⁹

Desquamative interstitial pneumonia and RBILD, once considered distinct entities, are arguably different manifestations of the same condition.^13,14 Desquamative interstitial pneumonia/RBILD is quite distinct in histopathological appearance from UIP, with a much more homogeneous geographical distribution. Significantly less septal thickening is seen, and fibroblastic foci and honeycombing are absent. The most striking feature of these two conditions is the filling of alveoli by pigment-laden macrophages (Figure 1D). These macrophages were originally and incorrectly thought to be desquamated alveolar epithelial cells; thus the name desquamative interstitial pneumonia. If the macrophage infiltrate is predominantly peribronchiolar, the term RBILD is used. If widespread alveolar involvement is found, the process is called desquamative interstitial pneumonia.

Acute interstitial pneumonia is a diffuse, fibroproliferative process that is indistinguishable from the organizing stage of diffuse alveolar damage, the histopathological corollary of acute respiratory distress syndrome. Widespread proliferation of fibroblasts and myofibroblasts is seen, with minimal inflammatory infiltrate (Figure 1E). The fibrotic changes are temporally homogeneous and may be accompanied by the presence of hyaline membranes. With time, end-stage fibrosis develops, and large cystic airspaces resembling honeycomb form. These cystic spaces, however, are lined with alveolar epithelium, not bronchial epithelium as seen in UIP.

Many studies have addressed the therapeutic responsiveness of idiopathic interstitial pneumonia. Corticosteroids have been the mainstay of therapy since their initial release for clinical use in 1948.¹⁵ Unfortunately, no prospective randomized placebo-controlled trial has examined the role of corticosteroids for idiopathic interstitial pneumonia. Retrospective reviews have suggested a clinical response to corticosteroids in 16% to 57% of patients (Table 2).^16-18 These studies did not distinguish between histopathological subgroups and relied on the clinical diagnosis (not confirmed by biopsy findings) of idiopathic interstitial pneumonia in most patients. They also used a variety of subjective (eg, dyspnea scale) and objective (eg, forced vital capacity and radiographic improvement) outcome measures to document response, many of which are of questionable use as meaningful end points. Two recent studies of 38 and 41 patients with idiopathic interstitial pneumonia confirmed by surgical lung biopsy findings reported a 26% and 27% improvement, respectively, in clinical, radiographic, and physiological variables with corticosteroids.^19,20 These data have been used to justify the use of corticosteroids in patients with idiopathic interstitial pneumonia, and most patients receive them.^31,32

Attempts to predict who will respond to treatment have been largely disappointing. The results of several large studies have been reviewed by Mapel et al.¹⁵ Clinical factors possibly associated with responsiveness to corticosteroid therapy include younger age, lesser degree of dyspnea, and female sex. A stronger association has been suggested with the degree of inflammation on open lung biopsy findings. Of more than 100 biopsy samples reviewed, 64% of patients with primarily inflammatory biopsy specimens responded to corticosteroids, compared with 23% of patients with primarily fibrotic patterns. Unfortunately, most patients in these early studies did not undergo biopsy, and when biopsy specimens were obtained, they were not classified by histopathological subgroup. Nonetheless, this association suggested an important role for the histopathological subclassification of idiopathic interstitial pneumonia in predicting responsiveness to therapy.

Relatively few studies address the treatment of specific histopathological subgroups of idiopathic interstitial pneumonia (Table 2). The best data on corticosteroid responsiveness in UIP and desquamative interstitial pneumonia were published in 1978 by Carrington et al.²¹ They retrospectively reviewed patients with UIP and desquamative interstitial pneumonia and found that when treated with corticosteroids, 11% and 62%, respectively, improved as measured by a combination of clinical, physiological, and radiographic indices. The same patients, when untreated during a different period, had 0% and 22% improvement, respectively.²¹ The criteria used for histopathological classification at that time did not distinguish UIP from NSIP, likely augmenting the response rate for UIP. More recent studies have found similarly poor response rates to corticosteroids for patients with UIP (8%-17%).^9,22,23 No prospective treatment studies exist for NSIP or RBILD. However, retrospective reviews of observational studies in which most patients received corticosteroids have suggested excellent responsiveness for these diseases.^9,12,14,23 Respiratory bronchiolitis–associated interstitial lung disease appears to resolve in most patients with smoking cessation alone.^13,14 No good data exist on corticosteroid therapy for acute interstitial pneumonia, with one study of 29 patients showing no significant difference in survival with or without treatment.²⁴ Taken together, these studies demonstrate a distinct difference in corticosteroid responsiveness on the basis of histopathological subgroup.

Because of the often poor responsiveness of idiopathic interstitial pneumonia to corticosteroids, immunomodulatory agents have been tried with mixed results (Table 2). A prospective, double-blind, randomized placebo-controlled trial of 27 patients with idiopathic interstitial pneumonia compared azathioprine sodium plus prednisone with prednisone alone and showed no statistically significant difference in lung function at 1 year or in survival, although a trend favored the azathioprine group.²⁵ A randomized controlled trial of 43 patients with idiopathic interstitial pneumonia compared cyclophosphamide (anhydrous) plus low-dose prednisolone with standard-dose prednisolone alone and also showed no statistically significant difference in improvement or survival between the 2 groups, although a combined end point of time to change of treatment and survival was better for the cyclophosphamide group.²⁶ A trial of single-agent therapy with cyclophosphamide in 19 patients with UIP in whom corticosteroid therapy had failed showed no significant response to therapy and significant toxic effects.²⁷

Recently, antifibrotic agents have been tried alone or in combination with corticosteroids in patients with UIP (Table 2). Colchicine was compared with prednisone in 26 patients with biopsy-proved UIP who were followed up for 1 year.^22,28 Although improvement was not directly reported, a trend was found toward a lower failure rate (eg, clinical decline, drug intolerance) and a lower incidence of serious adverse effects with colchicine. A recent study of another agent, interferon gamma-1b, showed significantly better lung function (total lung capacity and PaO₂ at rest and with exercise) at 12 months in patients treated with interferon gamma-1b and prednisolone compared with prednisolone alone.²⁹ Interest is increasing in these and other antifibrotic agents for reasons we will discuss (see "Abnormal Wound Healing" subsection of the "UIP: Pathophysiological Paradigm Shift" section).

Lung transplantation should be considered in all patients with progressive disease who are unresponsive to medical treatment. A consensus guideline outlining general indications and contraindications for lung transplantation recommends referral for transplantation when patients have progressive symptoms, a vital capacity below 60% to 70% of that predicted, and a diffusing capacity corrected for alveolar volume below 50% to 60%.³³ Unfortunately, patients with idiopathic interstitial pneumonia are often referred for transplantation too late, with 30% or more of such patients dying while on the waiting list.³⁴ A recent review suggests that survival may be optimized when patients are referred on the basis of a combination of diffusion capacity and HRCT fibrosis score.³⁵ Single-lung transplantation is equal to bilateral lung transplantation³⁶ and leads to improvement in exercise capacity, pulmonary function, and gas exchange.³⁷ Most important, lung transplantation has been shown to improve survival in patients with pulmonary fibrosis.³⁸

Several excellent retrospective studies have compared survival between histopathological subgroups of idiopathic interstitial pneumonia (Table 3). Excluding acute interstitial pneumonia, which is usually clinically distinct and has a poor prognosis (less than half of patients survive),²⁴ the most important prognostic information is the presence or absence of UIP. In 1978, Carrington et al²¹ compared survival in 93 patients with UIP and desquamative interstitial pneumonia and found a dramatically better 5-year survival for patients with desquamative interstitial pneumonia (95%) than for those with UIP (55%.) More recently, Bjoraker et al⁷ reviewed 97 cases of idiopathic interstitial pneumonia and found an even worse 5-year survival for patients with UIP (20%) compared with those with NSIP (70%) and desquamative interstitial pneumonia/RBILD (80%). The patients with UIP in the study by Carrington et al²¹ likely included many who would have been classified as having NSIP by Bjoraker et al,⁷ which may explain the higher UIP survival rate reported by Carrington et al. The dramatic difference in survival between UIP and the histopathological subgroups desquamative interstitial pneumonia, RBILD, and NSIP has been substantiated by several other studies.^{8,9,12-14,23,39} The further classification of NSIP into cellular and fibrotic subgroups appears to be important prognostically, with cellular NSIP having a significantly better prognosis than fibrotic NSIP.^9,10 A recent study of patients with UIP showed that, unlike in NSIP, the extent and severity of interstitial cellularity or fibrosis are not predictive of survival.⁴⁰ Instead, the specific histopathological features that predicted survival are the degree of granulation tissue deposition in the alveolar space and the extent of young connective tissue present within the fibroblastic foci.⁴⁰

It is now clear that distinguishing UIP from the other histopathological subgroups of idiopathic interstitial pneumonia has important therapeutic and prognostic implications. For years, the prevailing pathophysiological theory in interstitial lung disease has held that generalized inflammation eventually leads to widespread parenchymal fibrosis and progressive decline in pulmonary function. This model seems to hold true for many of the corticosteroid-responsive interstitial lung diseases (eg, desquamative interstitial pneumonia/RBILD, cellular NSIP, hypersensitivity pneumonitis, eosinophilic pneumonia, sarcoidosis, and organizing pneumonia), but only for a few, if any, patients with UIP. The historical belief that UIP represents the final common pathway for all forms of idiopathic interstitial pneumonia is likely incorrect. Instead, UIP appears to be a distinct pathophysiological entity characterized by minimal inflammation and chronic fibroproliferation due to abnormal parenchymal wound healing.⁴¹

The histopathological observations of temporal heterogeneity and focal proliferation of fibroblasts and myofibroblasts (fibroblastic foci) have led to the hypothesis that UIP is a result of ongoing diffuse microscopic alveolar epithelial injury and abnormal wound healing.⁴¹ Usual interstitial pneumonia is characterized by fibroblast/myofibroblast migration and proliferation to sites of injury, decreased myofibroblast apoptosis, and increased activity of and response to fibrogenic cytokines such as transforming growth factor β₁, tumor necrosis factor α, platelet-derived growth factor, and insulin-like growth factor. In addition, inappropriate reepithelialization of the alveolar epithelium and impaired remodeling of the extracellular matrix appear to be characteristic. The lack of any significant inflammatory stimulus in this fibroproliferative model may explain the poor response of UIP to anti-inflammatory therapies such as corticosteroids.

To emphasize the distinct pathophysiology and clinical course of UIP, the American Thoracic Society and the European Respiratory Society have recommended reserving the term idiopathic pulmonary fibrosis (IPF) for the clinical condition characterized by progressive dyspnea, cough, restrictive lung disease, and the histopathological pattern of UIP.⁴² This is an important semantic change that emphasizes the clinical relevance of histopathology. Idiopathic pulmonary fibrosis has been used historically to refer to a wide variety of interstitial pulmonary conditions, and it is still commonly used to refer to idiopathic interstitial pneumonia irrespective of the histopathological subgroup. It is important that the profession attempts to clarify and define the nomenclature used for idiopathic interstitial pneumonia so that clinicians and researchers will communicate clearly and effectively. The American Thoracic Society/European Respiratory Society definition of IPF will be used throughout the remainder of this article.

The patient with suspected idiopathic interstitial pneumonia presents a formidable diagnostic challenge to physicians, and an organized approach is critical. Our approach to the overall diagnosis of IPF is presented in Figure 2. The clinically important distinction to be made when diagnosing idiopathic interstitial pneumonia is whether the patient has IPF (as defined by the presence of UIP). A full description of the diagnosis of IPF requires a general discussion of the approach to interstitial lung disease, which is beyond the scope of this review. Several excellent reviews of the general approach to the diagnosis of interstitial lung disease have been published in the past several years.^43-47 Our discussion will focus specifically on the ability of different clinical modalities to distinguish IPF from the other histopathological subgroups of idiopathic interstitial pneumonia.

The history and results of physical and laboratory evaluations have little diagnostic accuracy for IPF.⁴⁸ In all chronic forms of idiopathic interstitial pneumonia, patients usually present with progressive dyspnea on exertion and dry cough. Fine inspiratory crackles are common on the physical examination, with digital clubbing less frequent but often present. Pulmonary function testing commonly demonstrates restrictive lung disease (decreased forced vital capacity and total lung capacity) and abnormal gas exchange (decreased diffusion capacity of carbon dioxide and resting or exercise PaO₂).

Unfortunately, no bronchoalveolar lavage findings are specific to IPF. Using the cellular constituency of the bronchoalveolar lavage fluid from patients with idiopathic interstitial pneumonia to predict the underlying histopathological features has been of little clinical utility.^49-51 Patients with IPF demonstrate a nonspecific increase in bronchoalveolar lavage levels of neutrophils, eosinophils, and, less commonly, lymphocytes.⁵² Several studies have suggested that a predominance of bronchoalveolar lavage lymphocytes predicts corticosteroid responsiveness and improved survival.^51,53 This finding has led to the recommendation that isolated bronchoalveolar lavage lymphocytosis should suggest alternative, corticosteroid-responsive diagnoses such as NSIP, sarcoidosis, hypersensitivity pneumonitis, and bronchiolitis obliterans with organizing pneumonia.⁴² Transbronchial biopsy is limited by the small size of the specimen obtained and the lack of histological preservation due to mechanical crushing of the tissue. Because of these factors, transbronchial biopsy specimens have poor diagnostic accuracy for IPF.^54-56

The chest radiograph lacks diagnostic specificity in interstitial lung disease, with a correct diagnosis made in less than 50% of cases and only a 70% interobserver agreement.^57,58 Sensitivity is better, but up to 10% of patients with histopathologically confirmed interstitial fibrosis have normal findings on chest radiographs.^57-60 In addition, the chest radiograph correlates poorly with the histological pattern (eg, UIP or NSIP), the anatomic distribution, and the severity of the disease, with the exception of honeycombing, which is quite specific.^61-64 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.⁵⁸

The characteristic radiographic features of IPF on HRCT have been well described (Figure 3).^65-67 These include patchy, predominantly basilar, subpleural reticular opacities; absent or limited ground-glass opacities; traction bronchiectasis and bronchiolectasis; and honeycombing. Atypical findings include extensive ground-glass opacities, nodules, upper-lobe or midzone predominance of findings, and significant hilar or mediastinal lymphadenopathy. The presence of more ground-glass opacities than reticular changes is particularly confounding, as its histopathological correlation is quite variable.⁶⁸ Thus, the presence of significant ground-glass opacities on HRCT should bring in to question the radiographic diagnosis of IPF.

The use of HRCT chest findings to diagnose IPF has been widely investigated (Table 4).^58,69-72 Most published studies have used surgical lung biopsy as the comparative gold standard. The radiographic and histopathological criteria used for the diagnosis of IPF in these studies have varied by institution and over time. The most recent studies of HRCT have reported sensitivities of 43% to 78% and specificities of 90% to 97% for the confident radiographic diagnosis of IPF.^73-76 Inclusion of less confident radiographic diagnoses improves sensitivity, but leads to significant numbers of false-positive findings (ie, lower specificity.) On the basis of these data, HRCT can be used to make a confident, highly specific diagnosis of IPF in half to two thirds of patients with idiopathic interstitial pneumonia.

Two recent studies have addressed the accuracy of the combined clinical and radiographic diagnosis of IPF. In both studies, clinicians were masked to the results of the surgical lung biopsies, which were used as the gold standard for diagnosis. Raghu et al⁷⁵ prospectively studied 59 patients who were referred for evaluation of new-onset interstitial lung disease, all of whom underwent diagnostic transbronchial or surgical lung biopsies. A clinical diagnosis of IPF was made if a patient met all of the following criteria: (1) absence of clinical features suggestive of infectious, neoplastic, collagen-vascular, occupational, environmental, or drug-related conditions or hereditary diseases known to be associated with IPF; (2) age greater than 50 years; (3) insidious onset of exertional dyspnea of greater than 6 months' duration; (4) bibasilar end-expiratory crackles; (5) restrictive lung defect without coexisting airflow obstruction, decreased diffusing capacity, and increased alveolar-arterial oxygen gradient at rest or with exercise; (6) chest radiograph or HRCT with characteristic findings of IPF; and (7) transbronchial biopsy findings or bronchoalveolar lavage cellular profile lacking features to support a specific alternative diagnosis. When compared with the histopathological diagnosis, the clinical diagnosis of IPF had a sensitivity of 62% and a specificity of 97%. Hunninghake et al⁷³ performed a similar study of 91 patients suspected of having IPF. No predetermined criteria for clinical diagnosis were described. The confident clinical diagnosis of IPF had a sensitivity of 48% and a specificity of 89%. These 2 studies argue that, when the clinical and radiographic diagnoses are consistent, IPF can be confidently diagnosed. They also demonstrate that this consistency is present in only about half of patients with biopsy-confirmed UIP.

The American Thoracic Society and the European Respiratory Society recently published a consensus statement describing major and minor criteria for the clinical diagnosis of IPF, given in Table 5.⁴² The presence of all 4 major criteria and at least 3 of 4 minor criteria increases the likelihood of a correct clinical diagnosis. The panel noted that in the absence of surgical lung biopsy findings, the diagnosis of IPF remains unproved. Ultimately, individual clinicians and patients must decide the degree of diagnostic specificity with which they are comfortable. If patients do not meet these major and minor criteria, we believe that surgical lung biopsy should be pursued in all patients except those too frail or ill to tolerate it.¹¹

In 1980, the largest retrospective review of open-lung biopsy in patients with infiltrative lung disease reported a diagnostic yield of 92% and overall morbidity and mortality of 2.5% and 0.3%, respectively.⁷⁷ Recent studies have confirmed this diagnostic yield.^78,79 Thoracoscopy was introduced in the early 1980s as a potential alternative to thoracotomy.⁸⁰ Many articles have since evaluated the role of video-assisted thoracoscopic surgery in the diagnosis of interstitial lung disease^81-83 and compared the thoracoscopic and open approaches with lung biopsy.^78,79,84-87 Video-assisted thoracoscopic surgery appears to have an equivalent diagnostic yield, equal or lower morbidity, and equal or decreased length of stay compared with the open surgical approach. In addition, thoracoscopy often allows the surgeon to view and sample more distant areas of the lung.

The choice between video-assisted thoracoscopic surgery and open-lung biopsy should generally be made by the thoracic surgeon on the basis of individual patient characteristics, but most patients today undergo video-assisted thoracoscopic surgery. However, video-assisted thoracoscopic surgery is contraindicated in the following 3 important situations: inability to tolerate single-lung ventilation (eg, severe hypoxemia or high airway pressures), coagulopathy, and pleural adhesions or scarring from previous thoracic surgery or pleurodesis. If a histopathological diagnosis is essential in these patients, a limited thoracotomy is usually the surgical approach of choice.

The past few years have seen extensive redefinition of idiopathic interstitial pneumonia, with an increased understanding of the clinical importance of histopathological subclassification. In addition, we have begun to understand the underlying pathophysiology to a much greater extent, and in so doing realized that the old paradigm of inflammation leading to fibrosis might not apply to all conditions. How does this relate to the care of patients with idiopathic interstitial pneumonia in the year 2003?

First, all patients with suspected idiopathic interstitial pneumonia should undergo careful evaluation, with an in-depth history and physical examination, pulmonary function testing, bronchoscopy, and HRCT by physicians experienced in the care of patients with idiopathic interstitial pneumonia. Many patients can be classified as having IPF with a high degree of specificity without the need for surgical lung biopsy.

Second, patients whose diagnosis is unclear after a thorough clinical and radiographic evaluation should undergo surgical lung biopsy to establish the underlying histopathological pattern. It is quite clear that IPF (as defined by the presence of UIP) does not respond to corticosteroids in most patients. Given the significant adverse effects of corticosteroids, the diagnosis of IPF should lead to careful consideration of the decision to use corticosteroids and prompt the addition of a second agent. The American Thoracic Society and the European Respiratory Society have recommended that initial therapy for IPF involve corticosteroids and immunomodulatory agents (cyclophosphamide or azathioprine) for 6 months, with progressive clinical decline or substantial adverse effects of treatment prompting cessation or modification of therapy.⁴²

Third, physicians need to counsel their patients with IPF early with regard to prognosis, as it is distinct from the other forms of idiopathic interstitial pneumonia. Referral for lung transplantation should be pursued early in those patients with progressive disease that is unresponsive to therapy who are candidates.

Last, clear establishment of the diagnosis of IPF allows patients who are refractory to standard therapy to be enrolled in clinical trials of antifibrotic agents and other new therapies. As our understanding of the pathophysiology of IPF grows, we will surely develop promising new agents that may significantly affect the morbidity and mortality of patients with this devastating disease.

Corresponding author and reprints: Talmadge E. King, Jr, MD, San Francisco General Hospital, Room 5H22, 1001 Potrero Ave, San Francisco, CA 94110 (e-mail: tking@medsfgh.ucsf.edu).

Accepted for publication June 12, 2002.