The origins of idiopathic pulmonary fibrosis (IPF) have eluded us for decades. Basic science and translational approaches have provided much-needed insights into the biological pathways that contribute to lung fibrosis, yet despite these advances, IPF remains an idiopathic disease. Perhaps our meager success is the result of a too-narrow focus on symptomatic, clinically advanced disease.

Consider the cardiologists’ approach to studying cardiovascular disease. In 1778, Edward Jenner linked the clinical syndrome of angina pectoris to coronary atherosclerosis (1). For the next 150 years, studies of cardiovascular disease focused exclusively on symptomatic patients. It was not until the mid-20th century that forward-thinking clinical investigators began to study unaffected individuals, permitting the identification of the now-classic risk factors of hypertension, diabetes, cigarette smoking, and hyperlipidemia in longitudinal epidemiologic studies. In recent years, important insights into cardiovascular disease have arisen from the study of asymptomatic (subclinical) cardiovascular disease, such as coronary artery calcification on computed tomography (CT) imaging (2).

Establishing a diagnosis of IPF in our traditional symptom-linked fashion is akin to diagnosing coronary artery disease only after a patient presents with a myocardial infarction. Since IPF is often diagnosed many years after symptoms arise (3), it is no surprise that the phenotype of interest has long been the symptomatic patient. A different approach is needed, however, if we are to examine the biological events that underlie the early development of this disease. Methods to identify the presence of subclinical interstitial lung disease (ILD) must first be investigated, and the face, construct, and predictive validity of proposed subclinical ILD phenotypes should then be established.

A fledgling (but growing) body of evidence suggests that subclinical ILD does indeed exist and can be identified by high-resolution CT (HRCT). Subclinical ILD is present in 2% of adult autopsy cases and has been found in a surprising number of older adults undergoing abdominal CT imaging (4, 5). Asymptomatic adults in kindreds of familial pulmonary fibrosis commonly have HRCT imaging abnormalities demonstrating usual interstitial pneumonia or another ILD pattern (6). Subclinical ILD has also been identified on HRCT in high-risk adults, such as cigarette smokers and those with rheumatologic disease (7–9). Subclinical ILD has even been identified in community-dwelling adults sampled without regard to respiratory symptoms or risk factors (10).

In this issue of the Journal (pp. 756–762), Doyle and colleagues build upon this evidence base by examining the construct validity of one definition of subclinical ILD (11). The authors examined 2,404 adult smokers who underwent HRCT imaging of the chest and 6-minute-walk testing in the COPDGene Study, an NHLBI-funded multicenter observational study with the primary aim of identifying novel genetic variants associated with the risk of chronic obstructive pulmonary disease. After accounting for potential confounding factors, the authors found that subclinical ILD on HRCT (termed “interstitial lung abnormalities”) was associated with a 19-m shorter distance on 6-minute-walk testing.

There were a number of strengths of the study, including a multicenter design, a well-considered multivariable modeling approach, and consistent findings in sensitivity and subgroup analyses. Questions that remain unanswered include residual confounding by body habitus, the clinical significance of a difference of 19 m in community-based smoking adults (a difference smaller than the previously reported minimal clinically important difference in ILD [12]), and whether the exercise physiology of subclinical ILD might resemble that of clinically evident ILD, such as increased dead space ventilation and ventilatory inefficiency. Nonetheless, the authors’ work provides evidence that subclinical ILD on HRCT may have a functional (even if unrecognized) impact.

A number of questions still remain about subclinical ILD. For instance, it is not clear what proportion of subclinical ILD cases eventually progress to clinical ILD. Long-term longitudinal studies will be needed to answer this question. Moreover there is no consensus on (and little evidence supporting) a unified definition of subclinical ILD. The authors defined an interstitial lung abnormality as an area on HRCT having ground glass and/or reticular abnormalities as determined by visual inspection using a “sequential reading method” previously developed in this cohort (13). While this method has good face validity, agreement between expert readers has not been established, and the authors’ data suggest that interreader agreement is only fair at best. In addition, the authors were unable to determine if subclinical ILD was present or absent in one-third of their cohort, limiting the generalizability of their findings. Since the radiologic appearance of subclinical ILD is not fully characterized, the value of ascertaining subclinical ILD by visual inspection is questionable. Whether alternate methods, such as semiautomated quantitative measures of lung attenuation (as commonly used to measure pulmonary emphysema), might improve ascertainment of subclinical ILD remains unknown. Finally, the impracticality of obtaining surgical lung biopsies in otherwise healthy individuals with mild disease presents a challenge to studying the biology of this new entity.

These limitations, however, do not justify ignoring subclinical ILD. On the contrary, they should spur targeted funded efforts to advance our understanding of subclinical ILD. Not too long ago, it was widely believed that hypertension was a compensatory response and that hyperlipidemia was an epiphenomenon. Once methods to reliably ascertain subclinical ILD are developed, large-scale observational studies will be able to identify new, modifiable risk factors for IPF, perhaps allowing us to finally achieve secondary prevention of this terrible disease.

References

Section:

ChooseTop of pageReferences <<CITING ARTICLES