Clinical History

The identification of an underlying cause is the single most important contribution made by clinical evaluation. Table 46-2 provides a checklist of the more important causes of DLD. A careful occupational history is essential and should include details of all previous occupations, including short-term employment. Asbestos exposure is often extensive in railway rolling-stock construction, shipyard workers, power station construction and maintenance workers, naval boilermen, garage workers (involved in brake lining), and other occupations in which asbestos exposure is overt; generally, workers in these occupations are well aware of their asbestos exposure. However, other workers, including joiners, electricians, carpenters, and construction workers, who handle asbestos in the form of roofing and insulation material, are often unaware of significant exposure. Other occupations associated with DLD include coal mining (coal worker’s pneumoconiosis), metal polishing (hard metal disease), and sandblasting (silicosis).

Table 46-2 Frequently Encountered Diffuse Lung Diseases with Identifiable Underlying Cause

Blood Tests

In most cases, specific diagnostic tests for DLD are confined to autoimmune serology and precipitins against organic antigens. Serologic evidence of rheumatoid arthritis and the other major CTDs should be sought in any patient with apparently idiopathic DLD in whom the diagnosis is uncertain. This is particularly important when the diagnosis appears to be cryptogenic organizing pneumonia, because underlying connective disease is common; when present, prognosis is not always good, and prolonged treatment may be required. In the antisynthetase syndrome, characterized by Jo1 antibody positivity, polymyositis or dermatomyositis, and progressive pulmonary fibrosis, lung disease often precedes systemic manifestations. However, although antibodies to extractable nuclear antigens tend to be disease specific, levels of antinuclear antibodies and rheumatoid factor are often moderately increased in idiopathic pulmonary fibrosis and are less useful diagnostically unless titers are greatly elevated.

The presence of precipitins to organic antigens increases the likelihood of hypersensitivity pneumonitis but should never be considered diagnostic in isolation. Positive precipitins denote exposure to an antigen, with immune recognition, but alone are not indicative of clinically significant DLD; pigeon and other bird breeders often have avian precipitin positivity without overt lung disease. Moreover, in many patients with convincing exposure histories and a histologic diagnosis of hypersensitivity pneumonitis, the appropriate precipitins are not present; avian antigens, for example, vary between bird species and even between individual birds.

Pulmonary Function Tests

Most DLDs are characterized by a restrictive ventilatory defect, a reduction in gas transfer, and variable hypoxia at rest or on exercise. A purely obstructive defect is often present in Langerhans cell histiocytosis and lymphangioleiomyomatosis and is sometimes a feature of fibrotic pulmonary sarcoidosis. A mixed ventilatory defect, which usually denotes an airway-centered component, is often present in hypersensitivity pneumonitis, sarcoidosis, and CTD (in which bronchiolitis or bronchiectasis may coexist with pulmonary fibrosis). Paradoxically, normal lung volumes in association with severe reduction in gas transfer levels are the hallmark of the combination of emphysema and pulmonary fibrosis. The often-cited phenomenon of a marked increase in gas transfer (measured using single-breath techniques) caused by diffuse alveolar hemorrhage is rarely clinically useful because this abnormality persists for only 36 hours after hemorrhage.

The most diagnostically useful pulmonary function pattern is preservation of gas transfer in association with irreversible airflow obstruction, a combination that points strongly toward intrinsic airways disease (i.e. bronchiolitis) rather than emphysema. However, the pattern of pulmonary function impairment seldom makes a major diagnostic contribution.

Chest Radiography

With recent attention focused on the diagnostic value of HRCT, it is often forgotten that the plain chest x-ray film provides invaluable information in diffuse lung disease. The chest radiograph points strongly toward a specific diagnosis in some patients. Sarcoidosis, the most prevalent DLD encountered in clinical practice, can be diagnosed with confidence from the clinical and chest radiographic features at presentation in many patients. HRCT seldom adds useful diagnostic information in this context.

Several radiographic features have useful positive predictive values (PPVs). The presence of hilar lymphadenopathy on chest radiography is particularly strongly predictive of sarcoidosis in the correct clinical context, although the radiographic differential diagnosis includes tuberculosis and malignancy, especially when hilar lymphadenopathy is asymmetric. Pleural effusions are an occasional feature of CTD, lymphangioleiomyomatosis, and asbestos-related disease, as well as disease processes that sometimes mimic DLD, including heart failure, infection, and malignancy. The distribution of disease on chest radiography is also diagnostically useful; granulomatous diseases (tuberculosis, sarcoidosis, hypersensitivity pneumonitis) tend to be more prominent in the middle to upper zones, whereas fibrotic diseases (IPF, fibrotic nonspecific interstitial pneumonia, asbestosis) have a predominantly lower-zone distribution.

However, apart from a large subgroup of patients with sarcoidosis, diagnoses based on chest radiography are seldom confident. Chest radiography is sometimes insensitive; in one often-quoted series (Epler et al.), 10% of patients with biopsy-proven DLD had normal chest radiographic appearances. The superior diagnostic accuracy of HRCT over chest radiography has been documented in numerous series, and the increased confidence associated with an HRCT diagnosis is a considerable aid to management. The classification of chest radiographic abnormalities as “nodular, reticulonodular, or reticular” provides relatively little diagnostic information. Predominantly basal honeycombing on chest radiography (invariably associated with honeycombing on HRCT) may be diagnostically useful in increasing the likelihood of IPF but is radiographically overt in surprisingly few IPF patients. It is now generally accepted that, except in patients with obvious sarcoidosis, routine HRCT is almost always warranted in DLD, although occasional exceptions exist, including elderly patients with obvious lower-zone honeycombing on chest radiography, indicative of IPF, and patients with long-standing pulmonary fibrosis on serial chest radiography and an obvious underlying cause, such as coal mining.

Two chest radiographic appearances pose particular diagnostic difficulties. Persistent unexplained multifocal consolidation has usually been treated unsuccessfully as for community-acquired pneumonia and has a wide differential diagnosis. Serial chest radiography tends to be more useful than HRCT in refining investigation, because the crucial diagnostic distinction lies between fixed infiltrates (nonbacterial infection, including tuberculosis, alveolar cell carcinoma, and other malignant processes) and changing infiltrates, in which these diagnoses are effectively excluded. However, immunologically mediated diseases, including eosinophilic pneumonia, cryptogenic organizing pneumonia, and vasculitic disorders, may give rise to either fixed or evanescent radiographic abnormalities, and a histologic diagnosis is often warranted. Diffuse alveolar filling processes giving rise to widespread air-space consolidation are generally indicative of life-threatening disease. Although this picture may represent DLD (e.g., acute interstitial pneumonitis, acute eosinophilic syndromes, drug-induced pulmonary infiltration), it is essential to broaden the differential diagnosis beyond DLD to include diffuse pulmonary infection, toxic inhalation, severe aspiration, opportunistic infection (especially Pneumocystis pneumonia), diffuse alveolar hemorrhage syndromes, mitral stenosis, and most importantly, heart failure. In both these radiographic presentations, successful management often depends on consideration of a wide differential diagnosis from the outset.

High-Resolution Computed Tomography

High-resolution CT has been the most important diagnostic advance in DLD in the last two decades. Numerous studies have confirmed the overall diagnostic accuracy of HRCT against findings at surgical biopsy, with a striking increase in sensitivity and specificity for individual diseases, compared with chest radiography. However, academic series understate the impact of HRCT, because the most important benefit has been to increase clinician confidence in noninvasive diagnosis, with a corresponding reduction in the numbers of patients needing to undergo surgical biopsy. Before HRCT, diagnoses based on clinical data and chest radiographic findings were seldom confident, and management was necessarily tentative in many cases. The combination of clinical and HRCT information now provides a confident first-choice diagnosis in the majority of patients, and in many other cases, the realistic differential diagnosis is shortened to two or three disorders. This allows surgical biopsy to be reserved for patients in whom the distinction between a small group of possible disorders has important management implications and for occasional patients in whom HRCT appearances are not suggestive of any single disorder. Thus, routine surgical biopsy, as a diagnostic “gold standard,” can no longer be justified in the HRCT era.

Histospecific diagnosis aside, HRCT sometimes plays an important role in detecting DLD. The superior sensitivity of HRCT over chest radiography has been an invariable finding in studies of a wide range of disorders. This feature of HRCT is particularly useful when symptoms or lung function impairment are associated with normal chest radiographic appearances. In CTD, pulmonary involvement is now the leading cause of death, and early treatment of progressive lung disease is desirable. By identifying limited pulmonary fibrosis, HRCT allows clinicians to select patients requiring more intensive monitoring, even when immediate treatment is not warranted. In workers previously exposed to asbestos, HRCT often discloses pulmonary fibrosis, which is obscured on chest radiography by concurrent pleural disease. However, it should be stressed that the sensitivity of HRCT occasionally creates its own difficulties. When interstitial abnormalities are limited, their clinical significance is sometimes difficult to rationalize; disease evident only on HRCT should not be extrapolated, in terms of natural history and management, to the more extensive disease described in historical clinical series. It is essential that HRCT findings are integrated with other clinical and investigative features and not interpreted in isolation.

The distinction between predominantly inflammatory and predominantly fibrotic disease can generally be made with reasonable confidence from HRCT. Anatomic distortion and reticular abnormalities are strongly indicative of irreversible fibrotic disease, and this is invariably true of honeycomb change (Figure 46-1). Consolidation is usually reversible, although it may occasionally represent dense fibrosis, especially in sarcoidosis. Ground-glass attenuation is often more difficult to interpret. In early work, this HRCT sign was shown to identify a substantial increase in the likelihood of significant inflammation, especially in the absence of concurrent reticular abnormalities. However, it is now clear that ground-glass attenuation denotes fine fibrosis in many cases, especially in sarcoidosis and nonspecific interstitial pneumonia, in which ground-glass attenuation is the cardinal HRCT feature (Figure 46-2). Traction bronchiectasis is a key HRCT discriminator because it invariably indicates underlying fibrosis. Thus, reversible inflammatory disease is likely only when ground-glass attenuation is not associated with traction bronchiectasis or admixed with reticular abnormalities.

Figure 46-1 High-resolution computed tomography (HRCT) scan in patient with idiopathic pulmonary fibrosis (IPF). There is prominent subpleural honeycombing with no ground-glass attenuation. In the correct clinical context, this image is virtually pathognomonic of IPF.

Figure 46-2 HRCT scan in patient with nonspecific interstitial pneumonia. Predominant abnormalities are ground-glass attenuation, with variable admixed fine reticular abnormalities, and traction bronchiectasis. Notably, there is no honeycomb change.

A wide range of HRCT profiles encompassing the distribution and pattern of DLD strongly suggests individual diseases. Box 46-1 summarizes the cardinal findings in common DLDs. As with its other applications, it is essential that HRCT findings be integrated with the pretest diagnostic probability, distilled from the history, clinical signs, previous natural history, or treatment course and the investigative findings, especially chest radiography, pulmonary function tests, and serology for autoimmune disease and environmental antigens. The role of HRCT in diagnosis is critically dependent on the presence or absence of a likely cause. In patients with appropriate environmental antigen or drug exposures (hypersensitivity pneumonitis, pneumoconioses, drug-induced lung disease), malignant disease (lymphangitis carcinomatosis), clinical or serologic evidence of CTD, or a heavy smoking history (Langerhans cell histiocytosis, RBILD), the diagnostic weighting required from HRCT can be reduced. In these contexts, HRCT appearances that are merely compatible (and not classic) often allow a sufficiently confident diagnosis to obviate diagnostic surgical biopsy.

By contrast, as discussed, the diagnostic role of HRCT is much greater in idiopathic disease, but unless the pretest diagnostic probability is high, confident diagnosis requires the presence of typical HRCT appearances. The optimal diagnostic use of HRCT in apparently idiopathic DLD can be distilled in a simple pragmatic algorithm, as follows:

Figure 46-3 HRCT scan in patient with sarcoidosis. Abnormalities suggesting the diagnosis include multiple, well-defined nodules, surrounding bronchovascular beading (white arrows), micronodules (especially in left anterior lung, black arrows), septal thickening (by granulomatous infiltration), and areas of dense consolidation, which may be reversible (coalescence of granulomas) or irreversible (fibrosis). This combination of abnormalities is virtually pathognomonic of sarcoidosis.

The weighting given to HRCT in the diagnosis of DLD varies from case to case but can usefully be considered in three categories. In some patients, HRCT appearances are virtually pathognomonic; this includes many cases of IPF, Langerhans cell histiocytosis, sarcoidosis, and lymphangitis carcinomatosis. Often, HRCT findings are diagnostic when combined with clinical information. A good example is the combination of widespread ground-glass attenuation (often with poorly defined centrilobular nodules) in combination with mosaic attenuation, which may be strongly indicative of hypersensitivity pneumonitis in nonsmokers with a compatible exposure history but may also represent RBILD in smokers (Figure 46-4). Third, even when not conclusive alone, HRCT may be invaluable when considered with diagnostic surgical biopsy. The histologic entity of nonspecific interstitial pneumonia (NSIP) is found in a variety of clinicoradiologic contexts, including entities overlapping clinically with IPF, fibrosing organizing pneumonia, and hypersensitivity pneumonitis; HRCT evaluation is key to distinguishing among these variants.

Figure 46-4 HRCT scans in patient with hypersensitivity pneumonitis. A, Abnormalities on inspiratory scan show widespread ground-glass attenuation, often with poorly formed centrilobular nodules, and are often extremely subtle; appearances may be virtually normal. B, On expiratory scan, striking regional variation in lung attenuation is often apparent, with areas of darker lung representing gas trapping (from bronchiolitic component of disease).

Bronchoscopic Procedures

Although endobronchial and transbronchial biopsies are straightforward and relatively noninvasive procedures, the volume of tissue taken is small, with only bronchial and peribronchial tissue sampled. Thus, both procedures have a high yield in diseases with peribronchial distribution, especially sarcoidosis and lymphangitis carcinomatosis. Occasionally, transbronchial biopsy findings can help confirm a diagnosis of hypersensitivity pneumonitis, although bronchoalveolar lavage fluid tends to be more rewarding in this regard. Bronchoscopic biopsy procedures have little or no diagnostic value in the idiopathic interstitial pneumonias.

Hemorrhage and pneumothorax (with transbronchial biopsy) are the important risks associated with bronchoscopic procedures. Major hemorrhage is rare, but pneumothoraces complicate transbronchial biopsies in 1% to 2% of procedures, although intercostal tube drainage is not always required.

In the 1980s, many regarded bronchoalveolar lavage (BAL) as an important part of the diagnostic algorithm in DLD. The distinction between a BAL neutrophilia (suggestive of IPF) and a BAL lymphocytosis (as in sarcoidosis or hypersensitivity pneumonitis) was held to be particularly useful. With experience, however, it became apparent that diagnostic distinctions based on BAL fluid in large groups of patients were insufficiently reliable in individual patients. The advent of HRCT also limited the role of BAL, which had been more influential in the pre-HRCT era, when most noninvasive diagnoses were tentative. There are currently no published evaluations of the diagnostic value added by BAL, once HRCT findings are taken into account.

In recent years, however, use of BAL has increased in some centers. In the 2002 ATS/ERS recommendations, compatible BAL findings (i.e., no lymphocytosis) are a requirement for the noninvasive diagnosis of IPF. Although not formally retained in the 2011 recommendations, this criterion reflects the diagnostic value of BAL findings when HRCT appearances are suggestive of IPF but not definitive. However, a BAL lymphocytosis has an even greater diagnostic impact in some patients with sarcoidosis or hypersensitivity pneumonitis. When significant fibrosis supervenes in these disorders, HRCT appearances often become atypical, and IPF is frequently the preferred diagnosis, before performance of BAL. A BAL lymphocytosis in the setting of fibrotic DLD may be an important justification for diagnostic surgical biopsy. Thus, BAL continues to play a useful diagnostic role in a significant subset of patients, when clinical and HRCT features are inconclusive, although adding little to diagnosis in the majority of DLD patients.

Surgical Biopsy

Surgical biopsy, formerly performed as an open procedure but now widely obtained using video-assisted thoracoscopic surgery (VATS), was once regarded as the diagnostic reference standard and, until recently, was advocated as a routine diagnostic procedure by some authorities. However, routine surgical biopsy is impractical outside referral centers, and in the 1980s, even before HRCT had a significant diagnostic impact, it was performed in less than 15% of patients with IPF in the United Kingdom. Even in referral centers, number of diagnostic surgical biopsies has been radically reduced with the application of HRCT. In one UK referral center, the Royal Brompton Hospital, biopsy was performed in more than 50% of IPF patients in the 1980s but in less than 25% in the mid-1990s. Moreover, it is increasingly clear that open or VATS biopsy is not a true diagnostic gold standard. Variation between 10 thoracic pathologists in assigning a histologic diagnosis in DLD was recently found to be considerable, with agreement only moderate (kappa coefficient of agreement of 0.38). In more than 20% of biopsies, the first-choice diagnosis was assigned with low confidence.

Thus, although biopsy procedures add invaluable information in select patients, and a pattern of usual interstitial pneumonia is usually diagnostically definitive, a histologic diagnosis alone should no longer be viewed as a diagnostic gold standard.

Sometimes, the histologic diagnosis is at odds with clinical and HRCT information and must be integrated with other information. In some patients with fibrotic hypersensitivity pneumonitis, a histologic pattern of usual interstitial pneumonia (which is normally indicative of IPF) is disclosed at biopsy, despite clinical, HRCT, and BAL features of hypersensitivity pneumonitis and an indolent course during follow-up. Similarly, in a cohort of more than 100 patients with IPF or fibrotic NSIP (Flaherty et al., 2003), a combination of histologic and HRCT findings provided more accurate prognostic information than either modality alone. Thus, surgical biopsy is now best viewed, as with HRCT, as a diagnostic “silver standard” that can often be avoided when HRCT and clinical features are typical of an individual DLD.

The morbidity and mortality associated with surgical biopsy in DLD are low in patients with an adequate pulmonary reserve but increase significantly in those with severe disease. In one series (Utz et al.), patients with advanced IPF (mean gas transfer <35% predicted) had a mortality ascribable to biopsy of 15%. Although this figure is generally regarded as an overstatement of the risk of the procedure, based on other series and widespread anecdotal experience, a surgical biopsy should not be performed unless central to management if the gas transfer is less than 30% of predicted. Moreover, in advanced idiopathic fibrotic disease, the prognostic value of a histospecific diagnosis diminishes. In another series (Latsi et al.), mortality was identical in IPF and fibrotic NSIP when the gas transfer was less than 35%, despite major differences in survival in less severe disease.

Thus, in younger patients (<60) presenting with a typical clinical picture of severe IPF and HRCT features suggestive of fibrotic NSIP, immediate referral for consideration of lung transplantation is warranted, without a histologic diagnosis. The exception is the patient presenting with overwhelmingly severe acute DLD, in which the diagnosis is unclear, and realistic differential diagnoses include acute interstitial pneumonia, severe infection (e.g., opportunistic infection), and malignancy. BAL fluid may be required to exclude infection, and occasionally a histologic diagnosis is required to plan management. Both procedures can be performed in ventilated patients, and if disease is slightly less severe, elective mechanical ventilation may be warranted to investigate appropriately.

Classification of Disease

Given the pivotal importance to management of identifying the likely or observed pattern of disease behavior, with a view to effective management, clinicians need a formalized disease behavior classification of DLDs, as discussed earlier. This approach (1) provides a better understanding of treatment principles by patients and non-DLD clinicians because this classification sidesteps the opaque terminology of DLD, (2) deals with the problem of diagnostic overlap (e.g., probable IPF, possible NSIP, possible hypersensitivity pneumonitis), and (3) provides a simple framework for rational treatment for the large subset of patients who cannot be classified using the current classification approach. An ATS/ERS committee is constructing a preliminary disease behavior classification along these lines.

To achieve a useful system, clinicians need to confront three uncertainties highly relevant to routine practice. First, this type of approach is not a replacement for the current diagnostic classification. Best diagnosis and management practice requires that the two classifications be viewed as complementary. A confident histospecific diagnosis establishes which patterns of disease behavior are possible. A diagnosis of IPF, for example, is essentially a statement that management should be as for inexorably progressive fibrosis. A diagnosis of hypersensitivity pneumonitis clarifies that all patterns of disease behavior are possible, and therefore best management should be strongly influenced by many other considerations, as discussed earlier.

A second problem is that in DLD, the current ethos is “definite classification at a single point in time.” An approach in which disease behavior is the basis for management is limited because such a system is tentative at presentation and becomes increasingly certain as short-term disease behavior, including change with initial therapy, is taken into account. This is actually an advantage because the use of a disease behavior classification perfectly mirrors the way in which experienced clinicians evaluate DLD, with reappraisal of initial impressions and therapies at follow-up. However, formalization of this approach undoubtedly requires a change in mindset.

Third, it is important to consider that although the inflammation/fibrosis dichotomy is the most therapeutically relevant treatment distinction, new antifibrotic agents may ultimately result in partial reversibility of some fibrotic disorders. It may now be timely to design a disease behavior classification around “reversible” versus “irreversible” disease, rather than inflammation and fibrosis. Such an approach would make a classification of disease behavior relevant to rare disorders, such as pulmonary alveolar proteinosis and Langerhans cell histiocytosis, and to clinicians who manage chronic lung and systemic diseases in general, who will appreciate the straightforward rationale of therapies applied to a framework of reversible/irreversible disease.

Optimal Diagnosis and Regional Expertise

The major shift within DLD to multidisciplinary diagnosis and away from a histologic gold standard does create important practical difficulties outside referral centers. Robust multidisciplinary diagnosis is a process of debate, disagreement, and the reconciliation of divergent views, and this approach cannot work well if one of the three disciplines, whether clinical, radiologic, or histologic, is hegemonic. However, away from referral centers, it is highly unlikely that there will be equivalent expertise in all three disciplines; even within some referral centers, this ideal is not entirely achieved. The act of questioning in a public forum nonetheless undoubtedly leads to improved diagnostic practice, but the importance of maximizing expertise should not be overlooked.

Decision to Perform Biopsy

No authoritative guidance on the indications for performing a diagnostic lung biopsy changes that the decision is often nuanced and difficult to make clinically, even before patient concerns are considered. Clinicians know what constitutes a high-risk biopsy and generally know what constitutes a low-risk biopsy. However, many biopsies fall somewhere in-between. Moreover, patients are involved in decisions, especially those associated with increased risk, although expecting a patient to decide whether to undergo a biopsy is unreasonable. One helpful approach is to outline for patients the following five conversations about biopsy and to inform them where they lie in this spectrum:

The five conversations just summarized do allow a clear statement of the balance of benefit and risk and provide most patients with the information they need to participate in decision making. By contrast, an isolated discussion of the risks of a diagnostic surgical biopsy, although satisfying the demands of clinical governance, fails to make the discussion relevant to the key medical considerations or to engage the patient fully.