Paul W Noble Department of Medicine, Division of Pulmonary, Allergy and Critical Care, Duke University Medical Center, Durham, North Carolina 27710, USA , Eric B Meltzer Department of Medicine, Division of Pulmonary, Allergy and Critical Care, Duke University Medical Center, Durham, North Carolina 27710, USA.

Acute exacerbations of IPF
Japanese physicians were the first to describe acute, unexpected deterioration in patients with IPF. This phenomenon has been called the “acute exacerbation” or, more euphemistically, the “terminal complication” of IPF ^{[24, 25]}. Acute exacerbations in IPF (AE-IPF) are characterized by sudden worsening of respiratory symptoms accompanied by hypoxemia and the appearance of new radiographic infiltrates. It is important to clinically distinguish AE-IPF from pulmonary embolism, congestive heart failure, pneumothorax and infection. Acute exacerbations can occur in patients with known IPF, but AE-IPF also presents as the initial manifestation of IPF ^[26]. The yearly incidence of AE-IPF is between 10 and 15% of all patients who are at risk. This estimate is based on data derived from the placebo-control groups of two large randomized clinical trials ^{[21, 27]}. AE-IPF is recognized by clinical criteria established in the first published case series from Japan ^[24]. The diagnosis, in the setting of known IPF, requires all of the following: a) acute worsening of dyspnea within the last month; b) deterioration from baseline of either vital capacity or gas exchange (usually documented by a wide A-a gradient); c) new radiographic infiltrates; and d) the absence of alternative causes for clinical deterioration ^[28]. The development of AE-IPF in patients without known IPF can be recognized when consensus criteria for IPF (see below) are satisfied and the patient presents with acute respiratory failure. The radiographic features of AE-IPF are ground glass opacification superimposed on typical interstitial markings. The histopathology of AE-IPF can demonstrate usual interstitial pneumonia (UIP; see below) with superimposed diffuse alveolar damage (DAD), characterized by alveolar epithelial injury and hyaline membranes. An alternate histopathology of AE-IPF consists of UIP with superimposed organizing pneumonia ^[26]. The prognosis of AE-IPF is poor. Several small series of AE-IPF report that the mortality of this condition ranges between 78 to 96% ^[29-31]. However, this data is skewed by methods that used autopsy results for case finding. Still, an association is reported between the need for mechanical ventilation and mortality in AE-IPF. The usual approach to treating AE-IPF is to use corticosteroids but the benefit of such has not been established.

Pulmonary hypertension and IPF
Pulmonary hypertension has been reported to occur in 32 to 84% of patients with IPF. The exact prevalence remains unclear because triggers for the evaluation of pulmonary pressures and the best method to detect pulmonary hypertension in IPF remain unsettled ^[32]. Diffusion capacity is strongly correlated with pulmonary hypertension, being inversely related ^{[33, 34]}. However, the severity of restrictive physiology has little bearing on the prevalence or degree of pulmonary hypertension. Several studies have demonstrated a lack of correlation between pulmonary artery pressure and forced vital capacity (FVC) ^{[33, 35]}. Right-heart catheterization is the best diagnostic test for pulmonary hypertension but the implementation of such invasive testing is difficult to justify in the absence of data demonstrating a benefit to treatment of pulmonary hypertension in IPF. Still it is clear that the presence of pulmonary hypertension in IPF adversely affects survival ^{[33, 36]}.

IPF and emphysema
Several groups have described a syndrome in which IPF coexists with pulmonary emphysema ^[37-39]. This comes as no surprise since both diseases are associated with a history of exposure to cigarette smoke. Combined IPF and emphysema is characterized by upper lobe emphysema and lower lobe fibrosis. Physiologic testing of these patients reveals preserved lung volume indices contrasted by markedly impaired diffusion capacity. The incidence of combined IPF and emphysema remains unknown but smaller case series suggest that this subgroup may comprise up to 35% of patients with IPF ^[38].

Combined IPF and emphysema is a strong determinant of secondary pulmonary hypertension ^[39]. In addition, combined IPF and emphysema has major effects on measures of physiologic function, exercise capacity and prognosis. The composite physiologic index (CPI) was derived to capture the effect of emphysema on IPF. CPI is simple to calculate, CPI = 91 – (0.65 percent predicted DLCO) – (0.53 percent predicted FVC) + (0.34 percent predicted FEV₁), and has been shown to reflect the extent of disease more accurately than single physiologic indices ^[38]. CPI is also a powerful predictor of mortality.

Lung cancer and IPF
An association between IPF and lung cancer was theorized based upon the simultaneous finding of IPF and lung cancer in autopsy studies dating back several decades ^[40]. A small number of epidemiologic reports helped advance the notion that IPF is an independent risk factor for lung cancer ^{[41, 42]}. Yet, studies examining multiple causes of death, utilizing information obtained from death certificates, failed to confirm an association between pulmonary fibrosis and lung cancer ^[43]. A retrospective case-control study took advantage of the British general-practice database and identified 890 cases of IPF. Compared to 5,884 controls, a seven-fold increase in lung cancer was observed in IPF patients ^[44]. Unsettled issues regarding the association between fibrosis and lung cancer include questions regarding the underlying mechanism of this association; as well as questions regarding the difference in cancer subtype and location in patients with pulmonary fibrosis as compared to the general population ^[45].

Etiology and pathogenesis

Etiology
The term “idiopathic” suggests there are no known causes for IPF. Diagnostic criteria for IPF require exclusion of known causes of interstitial lung disease. However, an environmental etiology for IPF is supported by evidence from several sources ^[46]. A relationship between environmental exposures and IPF is plausible, has been consistently demonstrated by casecontrol studies and is analogous to known disease, such as asbestosis, in which environmental material is associated with pulmonary fibrosis. Technical obstacles to epidemiologic research have prevented the definitive determination of a causal link between environmental exposure and IPF. Research in this area is hampered by the low prevalence of IPF. Case-control studies, though convenient, are flawed due to selection bias and recall bias.

Meanwhile, cigarette smoking is consistently associated with IPF. A recent study of familial pulmonary fibrosis looked at 309 affected individuals ^[16]. After adjusting for age and sex, this cohort demonstrated a strong association between smoking and IPF (odds ratio [OR], 3.6; 95% confidence interval [CI], 1.3-9.8).

A multi-center case-control study conducted in the United States included 248 patients with IPF and 491 matched control subjects ^[47]. This study demonstrated significant associations between IPF and a) cigarette smoking (OR, 1.6; 95% CI, 1.1-2.4); b) silica exposure (OR, 3.9; 95% CI, 1.2-12.7); and c) exposure to livestock (OR, 2.7; 95% CI, 1.3-5.5). Other associations failed to reach statistical significance.

Several intriguing reports suggest the involvement of herpesvirus and/or hepatitis C virus in the etiology of IPF ^[48-50]. However, another study demonstrated viral infection limited to IPF patients receiving corticosteroids, suggesting that infection is simply a marker of immunosuppression rather than an etiologic agent of fibrosis ^[51].

Pathogenesis
While pathogenetic mechanisms are incompletely understood, the currently accepted paradigm proposes that injury to the alveolar epithelium is followed by a burst of pro-inflammatory and fibroproliferative mediators that invoke responses associated with normal tissue repair. For unclear reasons, these repair processes never resolve and progressive fibrosis ensues. This theory is aligned with the most recent scientific data and has been summarized elsewhere ^[52-54]. A thorough appraisal of the scientific evidence is beyond the scope of this review. However, a few recent advances are worth mentioning.

The origin of pathological fibroblast foci within the IPF lesion remains puzzling. Possibilities include differentiation of resident fibroblasts, recruitment of circulating fibroblast precursors and transdifferentiation of epithelial cells into pathological fibroblast phenotypes. An animal model has demonstrated bone marrow-derived cells assuming a fibroblastic phenotype and migrating to the lung in substantial numbers following alveolar epithelial cell injury ^[55]. A separate group of researchers observed migration of fibrocytes to the lungs of animals in a model of epithelial injury [56]. Fibrocytes are circulating cells of hematopoietic origin which are thought to play a role in normal and pathological wound repair ^[57]. Fibrocytes were previously identified in a variety of fibrotic lesions, including hypertrophic dermal scars and abnormal airways in asthmatics ^[57].

Transdifferentiation of epithelial cells to a mesenchymal phenotype is a well documented phenomenon that takes place during embryogenesis. Epithelial-to-mesenchymal transition (EMT) is a similar process which has recently been demonstrated as an important pathway mediating fibroproliferation in certain renal diseases ^[58]. Pulmonary researchers have now demonstrated coexpression of epithelial and mesenchymal markers in histologic specimens obtained from patients with IPF, suggesting a role for EMT in pulmonary fibrosis ^[59]. In addition, animal models of pulmonary fibrosis demonstrate the possibility of EMT in the lung. One study employed a beta-galactosidase reporter cell to trace epithelial cell lineage during the development of experimentally-induced fibrosis. Mesenchymal markers were noted almost exclusively in cells of epithelial lineage ^[60].

An overlooked feature of pulmonary fibrosis is the presence of increased angiogenic activity, reminiscent of tumorigenesis. This has been well-established in both animal and human studies ^{[61, 62]}. An imbalance between angiogenic chemokines (IL-8 and ENA-78) and angiostatic chemokines (IP-10) has been proposed to explain angiogenesis in the development of progressive pulmonary fibrosis ^[63].

Diagnostic methods and criteria

Radiographic findings
The chest roentograph is abnormal in most patients with IPF (Figure 1). Nevertheless, approximately ten percent of patients with histologically proven IPF have a normal roentograph. In these cases, high-resolution computed tomography (HRCT) will reveal evidence of the disease that has been missed by a plain roentograph ^[1].

Roentographic images of IPF can demonstrate reticular markings (net-like curvilinear opacities). These markings are distributed in a bilateral, asymmetric pattern with predilection toward the lower lobes. A particular pattern of course reticular lines juxtaposed between areas of focal round translucency is known as honeycombing. Roentographic honeycombing emerges late in the course of disease and portends poor prognosis ^[64].

Use of standard roentographs lacks diagnostic accuracy. A correct diagnosis of IPF (true positive) is made by a roentograph in less than 50% of cases. Furthermore, the radiographic interpretation of interstitial patterns has the characteristic of poor interobserver agreement. Studies examining this particular test characteristic have reported meager 70% concordance amongst radiologists ^{[65, 66]}.

The development of HRCT revolutionized the diagnostic evaluation of interstitial lung disease. HRCT utilizes x-ray technology, along with computerized algorithms, to construct images of virtual thin axial slices through the chest. These high-fidelity images allow a detailed examination of the pulmonary parenchyma. Subsequently, interobserver agreement and overall diagnostic accuracy has been improved with this technology. HRCT permits identification of alternate patterns of diseases. The primary role of HRCT during the diagnostic evaluation of interstitial lung diseases has become the discrimination of “typical” radiographic IPF from that of other ILD.

The “typical” appearance of IPF on HRCT consists of patchy, predominantly peripheral, predominantly subpleural and necessarily bibasilar reticular opacification (Figure 2). Ground glass infiltrates can occupy no more than scant, limited areas of the images. Regions of dense reticulation may demonstrate secondary involvement of medium-sized airways that is known as “traction bronchiectasis”. The presence of subpleural honeycombing (defined on HRCT as palisades of small, round translucencies), traction bronchiectasis and thickened interlobular septae increases specificity for a diagnosis of IPF. Together, these findings constitute a radiographic pattern that is termed “confident” or “certain” IPF ^[67].

Certain characteristics need be absent from HRCT in order to label the images as consistent with the “confident” IPF pattern. The HRCT cannot feature predominant ground glass opacities, nodular infiltrates or significant lymphadenopathy. A predominance of upper lobe infiltrates is also inconsistent with “confident” IPF. These findings may suggest alternative diagnoses. For example, ground glass implies heart failure, non-specific interstitial pneumonia, cryptogenic organizing pneumonia, desquamative interstitial pneumonia, respiratory bronchiolitis-associated interstitial lung disease or hypersensitivity pneumonitis. A pattern of fine nodules might suggest hypersensitivity pneumonitis, granulomatous infection or lymphangitic spread of malignancy. Upper lobe disease is found in pulmonary Langerhans’ cell histiocytosis, hypersensitivity pneumonitis, several pneumoconioses, sarcoidosis, ankylosing spondylitis, rheumatoid nodules and eosinophilic pneumonia syndromes. Significant hilar lymphadenopathy is associated with sarcoidosis, infection and malignancy.

Several studies examined the accuracy of HRCT utilizing histopathology as the “gold standard” ^{[68, 69]}. Studies have demonstrated specificity exceeds 90% for the “confident” HRCT pattern. Thus in the right clinical setting, it is possible to make the diagnosis of IPF by HRCT alone. In such cases HRCT obviates the need for lung biopsy.

However, testing for the “confident” HRCT pattern is not a sensitive tool for case finding of IPF. The full spectrum of the “confident” HRCT pattern can only be found in 4-out-of-5 cases of biopsy-proven IPF. In other biopsy-proven IPF cases, a less specific reticular pattern is seen on HRCT which has been called “possible” IPF (Figure 3). The radiographic pattern of “possible” IPF requires surgical lung biopsy to confirm the diagnosis. Sometimes biopsy identifies an alternative diagnosis.