Abstract
In 2000, the American Thoracic Society and European Respiratory Society published the first consensus statement providing guidelines on the diagnosis and treatment of idiopathic pulmonary fibrosis (IPF). This statement presented, for the first time, diagnostic criteria for IPF and recommendations for treatment. Results from several studies have reshaped the thinking on IPF, and as a result, the guidelines have been recently revised using an evidence-based approach. Meanwhile, several epidemiologic studies have yielded data that identify potential risk factors and that better define the societal burden of IPF. This chapter summarizes the approach to diagnosing IPF and reviews epidemiologic data on IPF.
Keywords
Diagnosis, Epidemiology, Idiopathic pulmonary fibrosis, Prevalence and incidence, Treatment, Usual interstitial pneumonia
Key Points
- •
Over the last several decades, results from several studies have advanced understanding of idiopathic pulmonary fibrosis (IPF): how it is diagnosed, its basic epidemiologic profile, and occupational or environmental exposures that may increase the risk for developing the disease.
- •
These results have reshaped how IPF is diagnosed, especially by highlighting the accuracy with which a characteristic HRCT identifies a UIP pattern of lung injury: patients with such an HRCT need not have a surgical lung biopsy for IPF to be diagnosed confidently.
- •
Making a diagnosis of IPF is complex, and whether a surgical lung biopsy is indicated or not, diagnostic accuracy is improved with multidisciplinary discussions in centers that specialize in the care of patients with this disease.
Idiopathic pulmonary fibrosis (IPF) is defined as a chronic fibrosing interstitial pneumonia of unknown cause with a histologic pattern of usual interstitial pneumonia (UIP) on surgical lung biopsy. IPF is a lung-limited process that tends to occur in older adults. It is the most common of the idiopathic interstitial pneumonias (IIPs), among which it has the worst prognosis, with median survival estimates ranging from 3 to 5 years after diagnosis.
In 2000, the American Thoracic Society (ATS) and European Respiratory Society (ERS) published the first consensus statement providing guidelines on the diagnosis and treatment of IPF. This statement presented, for the first time, diagnostic criteria for IPF and recommendations for treatment. Over the past decade, results from several studies have reshaped the thinking on IPF, and as a result, the guidelines have been recently revised using an evidence-based approach. Meanwhile, several epidemiologic studies have yielded data that identify potential risk factors and that better define the societal burden of IPF. This chapter summarizes the approach to diagnosing IPF and reviews epidemiologic data on IPF.
Diagnosis of Idiopathic Pulmonary Fibrosis
In recent years, emerging data have helped to refine the diagnostic criteria for IPF. The first collaborative effort, in 2000, among the ATS, ERS, and American College of Chest Physicians resulted in an international consensus statement for the diagnosis of IPF. That statement, formulated on expert opinion and interpretation of available research at the time, held that a definitive diagnosis of IPF required a surgical lung biopsy showing a UIP pattern of lung injury and the following three criteria: (1) exclusion of other known causes of interstitial lung disease (ILD), including drug toxicities, environmental exposures, and collagen vascular diseases; (2) abnormal pulmonary function tests or impaired gas exchange; and (3) imaging consistent with this diagnosis. In the absence of a surgical lung biopsy, a diagnosis of probable IPF required all of the following four major criteria: (1) exclusion of other causes of ILD; (2) abnormal pulmonary function tests or impaired gas exchange; (3) bibasilar reticular abnormalities with minimal ground-glass opacities on high-resolution computed tomography (HRCT); and (4) transbronchial lung biopsy or bronchoalveolar lavage specimens without features to support an alternative diagnosis along with at least three of four minor criteria (age >50 years, the insidious onset of dyspnea, a duration of symptoms greater than 3 months, and bibasilar, inspiratory crackles).
Since that time, additional evidence has shown the value of HRCT in diagnosing IPF: when an experienced radiologist can say with high confidence that the pattern on HRCT is consistent with a histologic UIP pattern, UIP is the histologic pattern identified in more than 90% of cases. Based on these and other data supporting the accuracy of HRCT, a surgical lung biopsy is no longer required for a definitive diagnosis of IPF in cases with a radiologic UIP pattern and a compatible clinical presentation, and the characteristic HRCT pattern of a UIP pattern of lung injury has been defined ( Box 6.1 ). Given a characteristic HRCT pattern, the diagnosis of IPF still requires exclusion of other known causes of ILD, including domestic, occupational, or environmental exposures, connective tissue diseases, and drug toxicities.
Definite UIP Pattern
Subpleural, basal predominance
Reticular abnormality
Honeycombing without traction bronchiectasis
Absence of features (listed below) as inconsistent with UIP pattern
Possible UIP Pattern
Same as the criteria for a definite UIP pattern, although honeycombing is not present with or without traction bronchiectasis
Inconsistent With UIP Pattern
Upper-lung or midlung predominance
Peribronchovascular predominance
Extensive ground-glass abnormality (defined as the extent of the ground-glass abnormality is greater than the extent of the reticular abnormality)
Profuse micronodules
Discrete cysts
Diffuse mosaic attenuation or air-trapping (bilateral, in three or more lobes)
Consolidation in bronchopulmonary segments or lobes
For diagnosing IPF, the sensitivity of HRCT is significantly lower than its positive predictive value ; thus, when the characteristic HRCT pattern is absent, a surgical lung biopsy showing a UIP pattern is still required to make a definitive diagnosis of IPF. Histologic criteria have been devised to allow pathologists to categorize findings in surgical lung biopsy specimens as definite, possible, probable, or not UIP ( Box 6.2 ). In addition, the recently published evidence-based guidelines provide a framework for interpreting permutations of HRCT and histologic data ( Table 6.1 ).
Definite UIP Pattern
Evidence of marked fibrosis/architectural distortion with or without honeycombing in a predominantly subpleural/paraseptal distribution
Fibrosis in a patchy distribution
Fibroblastic foci
Absence of features against a diagnosis of UIP (see below)
Probable UIP Pattern
Evidence of marked fibrosis/architectural distortion with or without honeycombing in a predominantly subpleural/paraseptal distribution
Absence of either patchy fibrosis or fibroblastic foci, but not both
Absence of features against a diagnosis of UIP (see below)
OR
Honeycomb changes only
Possible UIP Pattern
Patchy or diffuse involvement of lung parenchyma by fibrosis, with or without interstitial inflammation
Absence of other criteria for a definite UIP patternAbsence of features against a diagnosis of UIP (see below)
Inconsistent UIP Pattern
Hyaline membranes (unless associated with an acute exacerbation of IPF)
Organizing pneumonia or granulomas (unless mild or occasional, respectively, but may otherwise suggest hypersensitivity pneumonitis)
Marked interstitial inflammation away from honeycombing
Predominant airway-centered disease
Other features suggesting an alternative diagnosis
| HRCT Pattern | Surgical Lung Biopsy Pattern | Diagnosis of IPF |
|---|---|---|
| UIP | UIP | IPF |
| Probable UIP | IPF | |
| Possible UIP | IPF | |
| Nonclassifiable fibrosis | IPF | |
| Not UIP | Not IPF | |
| Possible UIP | UIP | IPF |
| Probable UIP | IPF | |
| Possible UIP | Probable IPF | |
| Nonclassifiable fibrosis | Probable IPF | |
| Not UIP | Not IPF | |
| Inconsistent with UIP | UIP | Possible IPF |
| Probable UIP | Not IPF | |
| Possible UIP | Not IPF | |
| Nonclassifiable fibrosis | Not IPF | |
| Not UIP | Not IPF |
Challenges in Diagnosing Idiopathic Pulmonary Fibrosis
A threat to making a confident diagnosis of IPF arises when, as is the case in 12.5–26% of patients who have a multilobe surgical lung biopsy, a UIP pattern is found in samples from one lobe, but a different pattern is found in samples from another lobe (a scenario termed discordant UIP). However, survival in patients with discordant UIP is similar to patients with concordant UIP (i.e., surgical lung biopsy samples from all lobes having UIP patterns). Thus, if a surgical lung biopsy is performed, multiple lobes should be sampled, and a diagnosis of definite IPF can be made when a UIP pattern is identified in any.
Further complicating a confident diagnosis of IPF includes a recent study that utilized imaging and pathology from a prior multicenter study of IPF with a focus on discordance between imaging and a histologic diagnosis of UIP. Of the 241 HRCT from IPF cases that were reviewed, 75 cases (31.1%) had HRCT findings that were inconsistent with IPF based on criteria noted earlier ( Box 6.1 ). Of these inconsistent imaging cases, 94.7% were found to have either definite or probable UIP pattern on pathologic review—and were termed discordant cases. Discordant cases were then compared to concordant cases (imaging with a UIP pattern and pathology with either a definite or a probable UIP pattern). Discordant cases were younger and had less of a smoking history, smaller lung volumes, and slightly higher FEV 1 /FVC ratio than those in the concordant group. In addition, no significant difference in survival was found. Among the discordant cases, the inconsistent CT findings were most often the result of a diffuse mosaic pattern/air-trapping (71.8%), followed by a diffuse cradiocaudal distribution (39.4%), diffuse axial distribution (26.8%), predominance of signs in the upper zone or midzone of the lungs (23.9%), extensive ground-glass abnormalities (22.5%), and peribronchovascular predominance (12.7%). Thus, these findings further complicate the diagnosis, indicate that the term “inconsistent with IPF” is misleading, and stress the importance of surgical lung biopsy in those patients without a definite UIP—and even inconsistent with UIP—pattern on imaging.
As noted, making a diagnosis of IPF can be difficult, but the accuracy and confidence of an IPF diagnosis increase with multidisciplinary discussions among clinicians, radiologists, and pathologists. In a study of 58 consecutive cases of suspected IIP, Flaherty and colleagues sequentially gave expert clinicians, radiologists, and pathologists more and more information about a case of ILD and then allowed them to discuss their impressions as a group. As more information was divulged and cross-disciplinary discussions took place, the level of agreement about the diagnosis (and the degree of certainty in that diagnosis) improved. Not surprisingly, centers with expertise in IPF are more accurate at diagnosing IPF than community-based, referral practices. For unclear reasons, early referral to such a center seems to improve survival in patients with IPF.
Epidemiology of Idiopathic Pulmonary Fibrosis
Background
Epidemiology is defined as “the study of the distribution and determinants of health-related states or events (including disease),” and the goal of epidemiologists is to apply findings to control diseases or health issues. Specific objectives include determining the extent and effects of disease: by defining its prevalence, incidence, and mortality; by identifying its risk factors or causes; and by examining its natural history and prognosis. This information then allows for the evaluation of preventive and therapeutic interventions and builds a foundation for policies and regulatory decisions to be made that alleviate the burden of disease.
The relative rarity of IPF has challenged investigators with an interest in its epidemiology and, before 1990, discouraged large-scale epidemiologic studies from being performed. Although Leibow and Carrington first defined a UIP pattern in 1969, IPF was not given a diagnostic code in the International Classification of Diseases (ICD) until the ninth revision (ICD-9) at the end of the 1970s. This coding system gave investigators an opportunity to use ICD-coded mortality data to study the burden of IPF at the population level. Johnston and colleagues did so and published their results of mortality (discussed later) in 1990.
Mortality and Mortality Trends Over Time
Disease-specific mortality is calculated by determining the number of deaths per year resulting from a specific cause, divided by the number of persons alive in the midyear population. In a disease that is lethal, and when survival is short, as occurs in IPF, mortality serves as a surrogate for the incidence of disease.
Death certificate and census data are vital statistics recorded in several countries, and these data have provided investigators with the means to study mortality and mortality trends over time. Little is known about the validity of death certificate ICD coding in IPF, so results from studies using death certificate data should be interpreted with some caution. Investigators in the United Kingdom found that in 23 decedents with a diagnosis of IPF (ICD-9 code 516.3) recorded on a death certificate, 19 (83%) had premortem clinical information confirming either definite or possible IPF. Conversely, among 45 patients with a premortem diagnosis of IPF (ICD-9 code 516.3), IPF was recorded on the death certificate only about 50% of the time. Before the ICD-10 coding system (which combines IPF and postinflammatory pulmonary fibrosis [PIPF]), diagnostic transfer (or coding IPF as PIPF on the death certificate) was also reported to occur commonly. Of 20 decedents coded with PIPF (ICD-9 code 515) on the death certificate, nearly 50% had IPF (ICD-9 code 516.3) diagnosed before death. These data suggest IPF is likely underrecorded as the cause of death, and a significant proportion of decedents coded as dying from PIPF died of IPF. Although these findings are based on a small number of decedents from the United Kingdom, Coultas and Hughes identified similar issues in mortality data from New Mexico.
Because the ICD-10 coding system combined PIPF and IPF into one diagnostic code (J84.1), researchers calculating mortality with data after 1998 are likely including some decedents with progressive, fibrosing ILD that is not IPF. Investigators who have used this diagnostic code (J84.1) and who have systematically excluded cases with known-cause pulmonary fibrosis (PF) have termed this entity general PF. Other investigators have not excluded concurrent conditions that may result in PF and have termed this entity IPF clinical syndrome (IPF-CS). Regardless of the precise diagnosis (i.e., IPF vs. other progressive, fibrotic ILD) for decedents in such studies, trends in mortality reveal that PF is a daunting public health problem.
Johnston and colleagues were the first to calculate IPF (previously termed cryptogenic fibrosing alveolitis in the United Kingdom) mortality in a large-scale epidemiologic study. They found that in England and Wales, from 1979 to 1988, deaths from IPF (ICD-9 code, 516.3) more than doubled. IPF-associated mortality was more common in men (odds ratio [OR] = 2.24; 95% confidence interval [CI] = 2.11–2.38) and increased progressively with age: the risk of IPF in those aged 75 years or more was eight times the risk for those aged 45–54 years. Greater mortality was found in the central, industrialized areas of England and Wales, suggesting that environmental/occupational exposures could be a risk factor for IPF.
Hubbard and colleagues extended on the work of Johnston and colleagues by investigating available mortality data for both IPF (ICD-9 code 516.3) and PIPF (ICD-9 code 515) from England, Wales, Scotland, Germany, Australia, New Zealand, Canada, and the United States from 1979 to 1992. They found that mortality from IPF was the highest in England and Wales and rates were increasing not only in these countries but also in Scotland, Australia, and Canada. Mortality from IPF was stable in New Zealand and Germany and had decreased over time in the United States. Mortality from PIPF was the highest in the United States and increased over the study period in the United States, the United Kingdom, Canada, and Australia, with stable mortality again noted in New Zealand and Germany. The increase in mortality from PF could not be explained by diagnostic transfer (e.g., a change in coding practices from PIPF to IPF, or PIPF to IPF over time), although systematic diagnostic transfer (always coding IPF as PIPF, because of different terminology and coding rules) may have explained the higher mortality of PIPF in the United States.
Mannino and colleagues examined PF mortality data in the United States from 1979 to 1991. To capture all decedents with IPF, given the coding issues noted earlier, these investigators defined PF by combing ICD-9 diagnostic codes 515 (PIPF) and 516.3 (IPF) and eliminated cases with concurrent diagnostic codes for conditions with known associations with PF, including radiation fibrosis, collagen vascular diseases, or asbestosis. Over that 13-year period, age-adjusted mortality from PF increased 4.7% in men (from 48.6 per million to 50.9 per million) and increased 27.1% in women (from 21.4 per million to 27.2 per million). PF-associated mortality increased with increasing age. States with the highest PF-associated mortality were in the west and southeast, and those with the lowest rates were in the midwest and northeast.
Data from the 1990s to early 2000 show a further increase, and acceleration, in PF-associated mortality. Using US mortality data from 1992 to 2003 and applying methods similar to those of Mannino and colleagues, Olson and colleagues analyzed more than 28 million death records and found that the age-adjusted and standardized (to the year 2000) mortality increased 29.4% in men (from 49.7 deaths per million to 64.3 deaths per million) and 38.1% in women (from 42.3 deaths per million to 58.4 deaths per million) ( Fig. 6.1 ). Rates increased with increasing age. Compared with the previous decade, mortality over this period increased at a significantly faster pace in both men and women. From 1992 to 2003, rates rose more steeply in women than in men. Mortality was greater among white non-Hispanics than black non-Hispanics or Hispanics, suggesting that race and ethnicity may also play a role in the susceptibility to IPF.
