Abstract
Rationale: From the late 1970s to the early 1990s, studies found that mortality rates for pulmonary fibrosis were increasing. Recent data for mortality from pulmonary fibrosis are unavailable.
Objectives: We sought to determine mortality rates for pulmonary fibrosis in the United States from 1992 through 2003.
Methods: Using data from the National Center for Health Statistics, we calculated age-adjusted mortality rates from the deaths of persons with pulmonary fibrosis and stratified the data to determine differences in mortality rates by age, sex, race/ethnicity, and geography of the decedent. We developed a multivariable model to predict future mortality rates, and we determined the underlying cause of death in patients with pulmonary fibrosis.
Measurements and Main Results: From 1992 to 2003, there were 28,176,224 deaths in the United States and 175,088 decedents with pulmonary fibrosis. The average age- and sex-adjusted mortality rate was 50.8 per 1,000,000 people. The age-adjusted mortality rate increased 28.4% in men (from 40.2 deaths per 1,000,000 in 1992 to 61.9 deaths per 1,000,000 in 2003) and 41.3% in women (from 39.0 deaths per 1,000,000 in 1992 to 55.1 deaths per 1,000,000 in 2003). While increases were significant in both men and women (p < 0.0001), the rate of increase was higher in women (p < 0.0001). The most common cause of death in patients with pulmonary fibrosis was the disease itself.
Conclusions: From 1992 to 2003, mortality rates for pulmonary fibrosis significantly increased. Further investigation is needed to determine the etiology of these trends, which are predicted to continue to increase.
Recent data for mortality from pulmonary fibrosis in the United States are unavailable.
Mortality rates from pulmonary fibrosis increased from 1992 to 2003 and are predicted to increase further in the foreseeable future.
The relative rarity of IPF has challenged investigators who wished to conduct epidemiologic studies by making it difficult to enroll adequately sized study samples (2). Evaluation of death certificate data provides an opportunity to study IPF on a large scale and from an epidemiologic perspective. Using large public record databases, other investigators have shown that mortality rates for pulmonary fibrosis (PF) steadily increased in the United States, England and Wales, Scotland, Australia, and Canada from the late 1970s to the early 1990s (3–5). In addition, it appeared, that over the same time period, an increasing percentage of patients with PF were dying of this disease and not of comorbid conditions (5, 6). More recent mortality data for PF have not been analyzed.
To build on and extend the work of previous investigators and to determine if past trends persist, we used the U.S. Multiple Cause-of-Death (MCOD) mortality (Centers for Disease Control and Prevention, National Center for Health Statistics, http://www.cdc.gov/nchs) database to calculate mortality rates and examine the underlying cause-of-death (UCD) in United States' decedents with PF from 1992 to 2003. Using our calculated mortality rates, we built a predictive statistical model to estimate mortality rates for PF 5 years into the future. Some of the results of these studies have been previously reported in the form of abstracts (7–9).
The National Center for Health Statistics (NCHS) annually compiles data from all death certificates in the United States and releases the figures in yearly public-use tape files. Over two million records are contained within each annual file, and each record contains decedent demographics, MCOD codes (which identify up to 20 conditions related to death), and the ultimate UCD (10). For this study, we analyzed files from 1992 to 2003. Data from more recent years have yet to be compiled by the NCHS.
From 1992 to 1998, the NCHS coded conditions related to death with the 9th revision of the International Classification of Diseases (ICD-9) (11), but, in 1999, it stopped coding with the ICD-9 and began using the 10th revision of the ICD (ICD-10) (12). For each record in the database, the NCHS codes MCOD data in two axes: (1) the entity axis and (2) the record axis. The entity axis contains the cause(s) of death as listed by the death certifier and maintains the order as written on the death certificate. The NCHS derives the record axis code by applying a computerized algorithm called Translation of Axis (TRANSAX) to the entity axis code. Applying this algorithm minimizes repetition and inconsistencies within the entity axis data to produce the more standardized record axis (10).
The NCHS uses another computerized system, called the Automated Classification of Medical Entities (ACME), to code the UCD by applying ICD modification rules to the record axis codes generated by TRANSAX (13). The World Health Organization defines the UCD as “the disease or injury which initiated the train of events leading directly to death, or the circumstances of the accident or violence which produced the fatal injury” (12, 14). Data quality of the national MCOD database is maintained by records review at the state and national level.
We included in this study files from any decedent with “pulmonary fibrosis” in the record axis; thus, from 1992 to 1998, we captured all decedents with either (or both) ICD-9 codes 516.3 (IPF) or 515 (postinflammatory PF [PIPF]). Because of the change in coding from ICD-9 to ICD-10, from 1999 to 2003, we captured all decedents with ICD-10 code J84.1 (a code that comprises both IPF and PIPF).
Because our aim was to calculate mortality rates in IPF, we identified and excluded decedents with axis codes for any condition that might be associated with PF (i.e., we excluded files of decedents with known-cause PF). To gain the clearest understanding of the data, we performed three separate mortality rate analyses using increasingly stringent case definitions for IPF. In the first analysis, we included all decedent records with a code for PF, regardless of what other codes were present (e.g., conditions that might be associated with IPF). In the second analysis, we excluded records that contained codes for both PF and either connective tissue disease, asbestosis, and/or radiation fibrosis. Selecting these codes for exclusion allowed us to compare our results with those of Mannino and colleagues who used identical exclusion criteria in their study (5). From this analysis, we also evaluated the UCD in decedents with PF; details of the UCD determination are outlined in Appendix A. In the third analysis, we applied the strictest diagnostic criteria for IPF and excluded records that contained codes for both PF and any other condition known to cause PF, including connective tissue diseases, radiation fibrosis, asbestosis, pneumoconiosis (including coal workers' pneumoconiosis, silicosis, talcosis, and berylliosis), sarcoidosis, and/or extrinsic allergic alveolitis (hypersensitivity pneumonitis). Specifics of the exclusion methodology used in the second and third analyses appear in Appendix B. We used July 1st intercensal population estimates (from 1992 to 1999) and July 1st population projections (from 2000 to 2003), obtained from the U.S. Census Bureau (15) to determine denominators for corresponding yearly mortality rates. We used the 2000 U.S. Census population to standardize mortality rates (15).
We further analyzed mortality rates by stratifying the data according to year of death, age, sex, race/ethnicity, and geography (resident state of the decedent). We used Poisson multivariable regression analysis to evaluate differences in mortality rates based on age and sex groupings and year of death. The model included second-order interactions between the age and sex groupings and the year of death. With our final model, we predicted mortality rates in decedents with PF for the year 2008.
All data were analyzed using GENMOD with SAS version 9.1 (SAS Institute, Cary, NC). The GENMOD procedure was used to perform the Poisson regression analysis. Mortality rates were calculated using Microsoft Office Excel 2003 SP2 (Microsoft, Redmond, WA). We were not required to obtain institutional review board approval for this study because all data contained in the database files have been de-identified and are of public record.
From 1992 to 2003, there were 28,176,224 deaths in the United States. Of these deaths, a total of 175,088 records contained a diagnostic code for PF. From 1992 to 1998, 89,469 records contained a code for PIPF (ICD-9 code 515), 1,198 records contained a code for IPF (ICD-9 code 516.3), and 6 records contained codes for both. After 1998, a total of 84,415 records contained a code for PF (ICD-10 code J84.1).
Using the entire subset of 175,088 decedents with a diagnostic code for PF, the age-adjusted mortality rate increased 29.4% (from 49.7 deaths per 1,000,000 in 1992 to 64.3 deaths per 1,000,000 in 2003) in men and 38.1% (from 42.3 deaths per 1,000,000 in 1992 to 58.4 deaths per 1,000,000 in 2003) in women.
When connective tissue disease, asbestosis, and radiation fibrosis were excluded from the original analysis, 9,236 records were removed (connective tissue disease, n = 8,500; asbestosis, n = 718; radiation fibrosis, n = 7; and both connective tissue disease and asbestosis, n = 11). After excluding these conditions, the age-adjusted mortality rate over the study period increased 28.4% in men and 41.3% in women (Figure 1).
Figure 1. Actual number of deaths per year and age-adjusted mortality rates in decedents with pulmonary fibrosis per 1,000,000 population, 1992 through 2003. Mortality rates are standardized for age to the 2000 U.S. Census population.
[More] [Minimize]In the third analysis, we applied the most rigorous definition of IPF and removed all records with codes for connective tissue disease (n = 8,477), asbestosis (n = 718), radiation fibrosis (n = 7), pneumoconioses including sarcoidosis (n = 1,284), extrinsic allergic alveolitis (n = 93), or any combination of these diagnostic codes (n = 34); this resulted in the exclusion of 10,613 records. With these records excluded, we found the mortality rate increased 27.5% in men (from 48.1 deaths per 1,000,000 in 1992 to 61.2 deaths per 1,000,000 in 2003) and 40.8% (from 38.7 deaths per 1,000,000 in 1,992 to 54.5 deaths per 1,000,000 in 2003) in women. Furthermore, according to this analysis, for each year from 1992 to 2003, the age-adjusted mortality rates for both men and women with PF were within 1 death per 1,000,000 people of the rates calculated in the second analysis. Because the second and third analyses yielded similar results, we used the exclusion criteria applied in the second mortality rate analysis (i.e., ICD-9 515 or 516.3 or ICD-10 J84.1 less any records that contained codes for connective tissue diseases, radiation fibrosis, and/or asbestosis) to perform all additional analyses. This allowed for a direct comparison of our results with those previously reported for 1979 to 1991 (5).
When we stratified the dataset by age and sex, mortality rates in people with PF increased with increasing age (p < 0.0001) and increased from 1992 to 2003 (p < 0.0001). Among men with PF, from 1992 to 2003, mortality rates increased 13.9% in the youngest stratum (45–54 yr) and 28.6% in the oldest stratum (> 85 yr) (Table 1). Among women with PF, mortality rates increased 28.8% in the youngest stratum and 45.8% in the oldest stratum (Table 1). Over the length of the study period, mortality rates were significantly higher among men than women (p < 0.0001); however, over the study period, mortality rates accelerated more steeply in women than in men (p < 0.0001).
Men | Women | |||||
|---|---|---|---|---|---|---|
| Age Strata (yr) | Mortality Rate, deaths per 1,000,000 persons (2003) | Percent Increase (from 1992) | Mortality Rate, deaths per 1,000,000 persons (2003) | Percent Increase (from 1992) | ||
| 45–54 | 17.2 | 13.9 | 13.4 | 28.8 | ||
| 55–64 | 66.7 | 10.8 | 45.6 | 28.4 | ||
| 65–74 | 268.5 | 24.7 | 152.6 | 38.5 | ||
| 75–84 | 721.6 | 42.4 | 397.6 | 47.6 | ||
| > 85 | 1,256.7 | 28.6 | 793.1 | 45.8 | ||
In our multivariable Poisson regression analysis, we predicted that mortality rates for men older than 65 years with PF and for women of all ages with PF will be significantly higher in 2008 than in 2003 (Table 2).
Men | Women | |||||
|---|---|---|---|---|---|---|
| Age Strata (yr) | Predicted Mortality Rate per 1,000,000 | 95% Confidence Interval | Predicted Mortality Rate per 1,000,000 | 95% Confidence Interval | ||
| 45–54 | 18 | 16–19 | 17* | 15–18 | ||
| 55–64 | 71 | 67–75 | 52* | 48–56 | ||
| 65–74 | 306* | 295–317 | 185* | 177–193 | ||
| 75–84 | 827* | 802–852 | 494* | 478–510 | ||
| > 85 | 1,380* | 1,320–1,443 | 942* | 908–977 | ||
Over the study period, age-adjusted mortality rates increased 35.4% (to 78.0 deaths per 1,000,000 in 2003) in non-Hispanic white men and 46.7% (to 67.2 deaths per 1,000,000 in 2003) in non-Hispanic white women (Figures 2 and 3). Overall, age-adjusted mortality rates in Hispanics were lower than for white, non-Hispanics. Age-adjusted mortality rates and percentage increases in mortality rates from 1992 to 2003 were similar between Hispanic men and women with PF: the age-adjusted mortality rate increased 27.2% (to 23.8 deaths per 1,000,000 in 2003) in Hispanic men and 32.4% (to 24.9 deaths per 1,000,000 in 2003) in Hispanic women. Mortality rates declined 7.0% (to 25.1 deaths per 1,000,000 in 2003) in non-Hispanic, black men but increased 17.8% (to 28.5 deaths per 1,000,000 in 2003) in non-Hispanic, black women with PF. Mortality rates declined 12.2% (to 21.5 deaths per 1,000,000 in 2003) in nonwhite, nonblack, non-Hispanic men and 3.2% (to 18.4 deaths per 1,000,000 in 2003) in nonwhite, nonblack, non-Hispanic women with PF (Table 3).
Figure 2. Age-adjusted mortality rates in men with pulmonary fibrosis per 1,000,000 men by ethnicity and race from 1992 through 2003.
[More] [Minimize]
Figure 3. Age-adjusted mortality rates in women with pulmonary fibrosis per 1,000,000 women by ethnicity and race from 1992 through 2003.
[More] [Minimize]Men | Women | |||||
|---|---|---|---|---|---|---|
| Ethnicity/Race | Mortality Rate, Deaths per 1,000,000 persons (2003) | Percent Change (from 1992) | Mortality Rate, Deaths per 1,000,000 persons (2003) | Percent Change (from 1992) | ||
| White, non-Hispanic | 78.0 | 35.4 | 67.2 | 46.7 | ||
| Hispanic | 23.8 | 27.2 | 24.9 | 32.4 | ||
| Black, non-Hispanic | 25.1 | −7.0 | 28.5 | 17.8 | ||
| Other, non-Hispanic | 21.5 | −12.2 | 18.4 | −3.2 | ||
Age- and sex-adjusted mortality rates (averaged over the 12-yr study period) varied greatly by state (Figure 4). States with the lowest mortality rates included New Jersey (34.3 per 1,000,000), New York (34.3 per 1,000,000), Nevada (35.5 per 1,000,000), and the District of Columbia (40.4 per 1,000,000). States with the four highest mortality rates included New Mexico (69.5 per 1,000,000), Vermont (69.3 per 1,000,000), North Carolina (68.3 per 1,000,000), and South Carolina (68.1 per 1,000,000).
Figure 4. Age-adjusted mortality rates per 1,000,000 by state for 1992–2003 and 1979–1991. (A) A comparison of states with persistently low mortality rates. Upper set of map, 1992–2003. Mortality: white areas, 0–50 per million; gray areas, 51–60 per million; black areas, 61–70 per million. Lower map, 1979–1991. Age-adjusted mortality rates: white areas, 25 to <34 per million; gray areas, 34 to <40 per million; black areas, 40 to 70 per million. (B) A comparison of states with persistently high mortality rates. Upper map, 1992–2003. Mortality: white areas, 0–50 per million; gray areas, 51–60 per million; black areas, 61–70 per million. Lower set of map, 1979–1991. Age-adjusted mortality rates: white areas, 25 to <34 per million; gray areas, 34 to <40 per million; black areas, 40 to 70 per million. (Graphic from 1979–1991 reproduced by permission from Reference 5).
[More] [Minimize]In decedents with PF, the UCD averaged annually (± SD) was pulmonary fibrosis in 60.0 ± 2.2%, ischemic heart disease in 8.5 ± 0.6%, lung cancer in 2.9 ± 0.4%, pneumonia in 2.4 ± 0.5%, cerebrovascular disease in 1.3 ± 0.2%, congestive heart failure in 1.1 ± 0.3%, pulmonary embolism in 0.37 ± 0.05%, and other causes in 23.4 ± 1.3% (Figure 5).
Figure 5. Annual average underlying cause of death in decedents with pulmonary fibrosis from 1992 through 2003.
[More] [Minimize]Our results provide comprehensive data on mortality in patients with PF in the United States who died between 1992 and 2003. We analyzed over 28,000,000 death files, identified over 175,000 decedents with PF, and found that mortality rates rose steadily from 1992 to 2003. Over this time period, we found that the average age- and sex-adjusted mortality rate for PF was 50.8 per 1,000,000 people in the general population.
Recent mortality data from other countries are not available for comparison. However, when the survival duration for a disease is short and the case fatality rate is high (as in IPF), the incidence of a disease generally reflects its mortality rate (16). As anticipated for IPF—a disease with short survival and no effective therapy (17, 18)—the mortality rates we found mirror the recently described increasing incidence (19–21).
Two recent epidemiologic investigations have reported an increase in the incidence of IPF compared with the rate found by Coultas and colleagues in the late 1980s (19). Using a national health care claims database, Raghu and colleagues (20) determined the incidence of IPF (using broad case-finding criteria), in persons over the age of 18, was 160 per 1,000,000 population compared with 90 per 1,000,000 population reported by Coultas and colleagues. Using a national longitudinal general practice database from the United Kingdom, Gribbin and colleagues (21) found the crude incidence of IPF more than doubled (from 27.3 per 1,000,000 person-years to 67.8 per 1,000,000 person-years) between 1990 and 2003.
By using MCOD mortality data from 1979 through 1991, Mannino and colleagues found that the average age-adjusted mortality rate for PF was 33.0 per 1,000,000 individuals in the general population (an increase of 14% over the study period) (5). By applying the same exclusion criteria described by these investigators, we identified a 34% increase in mortality rates for PF from 1992 to 2003. Like Mannino and colleagues, we found mortality rates that were significantly higher in men than in women, rates that increased with increasing age, and rates that steadily increased over time (5). In addition, we found mortality rates in women climbing at a significantly faster pace than in men.
Mannino and colleagues suggested that the percentage increase in mortality rates from PF in men and larger percentage increase in rates in women from 1979 to 1991 may be related to smoking patterns in the United States (5). By 1955, the percentage of men who smoked in the population was decreasing, whereas the percentage of women who smoked was still on the rise (22). From 1965 to 1991, the percentage of men who were ever-smokers fell, whereas the percentage of women who were ever-smokers was stable (23). Epidemiologic studies suggest that smoking is a risk factor for IPF (24). Given this and the average age at IPF diagnosis (65 yr), the historical data on smoking tendencies, and what is known about the age at which most people begin smoking, we can speculate that the sex-specific mortality rates we observed may also be related to historical smoking patterns. If this is the case, we can expect mortality rates for women with PF to continue to rise out of proportion to rates for men for several more years.
It is also possible that the persistently rising mortality rates found in our study might reflect an increase in the clinical recognition of PF instead of an increase in the true incidence of—or mortality from—disease. During the 1990s, high-resolution computed tomography (HRCT) scans became commonplace in the evaluation of patients with interstitial lung disease. As the technology advanced, HRCT proved to be a simple, noninvasive, relatively inexpensive (compared with surgical lung biopsy) method for making a confident diagnosis of IPF in many patients (25). In the late 1990s, the first large, multinational, therapeutic trial for IPF was conceptualized and began enrolling patients—a factor that, no doubt, raised awareness for the disease (18). Although both of these developments may have increased the clinical recognition of disease, neither can explain the rise in mortality rates for PF noted by investigators before the 1990s (3–5). Whatever the cause, the overall burden of disease from PF is significantly higher than that reported over a decade ago (5).
Using Poisson multivariable regression, we created a model to predict future mortality rates for PF. By 2008, we predict mortality rates for PF to increase significantly in men older than 65 and in women older than 45 years. Because we do not know which factors have led to a rise in mortality rates from PF (such as the incidence of disease, changes in clinical recognition of disease, or changes in diagnostic coding), this prediction model only holds true if these unknown factors continue to have the same influence on the mortality rates in the future. If these factors persist, as the population ages we can expect the absolute numbers of deaths from PF to continue to rise.
Our study is the first large-scale study to examine age-, sex-, race-, and ethnicity-stratified mortality rates in decedents with PF. Consistent with previous studies (1, 5), we found that the age-adjusted mortality rate was higher among whites than blacks and is increasing at a faster pace in whites than in other racial or ethic groups. Although previous studies have suggested that whites are more likely to be diagnosed with IPF than blacks (5, 26), little is known about the differences in the incidence, mortality rates, and changes in these rates between ethnic groups. We found that age-adjusted mortality rates in Hispanics were lower than in white non-Hispanics, but the rates in Hispanics increased from 1992 to 2003. We also found that rates increased steadily in black, non-Hispanic women. In a retrospective analysis of IPF in New Zealand, the incidence of disease was lower in those of Maori or Polynesian descent than in those of European descent (27). Differences in race and ethnicity may play a role in susceptibility to IPF (or in gene–environment interactions), and these findings warrant further investigation.
Like a number of other investigators (3–5, 21), we found geographic variation across the United States in the distribution of deaths among people with PF. Although the overall mortality rate in patients with PF has increased, there remain a number of areas in the United States with below average age-adjusted mortality rates. Geographic variation may develop for any of three reasons: (1) there are true differences in rates, (2) the identification of disease is different among local clinicians (perhaps because the diagnostic criteria are not explicit or are interpreted differently), and (3) there are differences in the rate of diagnostic test use (such as HRCT or lung biopsy) (1, 28). Although we were unable to determine the underlying cause of the geographic variation by using this dataset, these differences warrant further exploration. If this variation represents true differences in mortality rates, it would argue that environmental factors may play a key role in the pathogenesis of IPF.
The proportion of decedents dying of, rather than with, PF has also increased over time. Panos and colleagues (6) examined six case series (n = 326) of decedents with IPF published between 1964 and 1983 and reported that 38.7% of decedents with IPF died of respiratory failure. Mannino and colleagues (5) found that, from 1979 to 1991, 50% of patients with PF died of their disease—in our study that number was 60%. This may result from advances in the treatment of other conditions that afflict patients with PF (e.g., cardiovascular disease), whereas effective treatment for PF remains elusive. As importantly, changes in the classification schema for idiopathic interstitial pneumonias may also explain at least a portion of these findings. In 1997, Katzenstein and Myers (29) proposed a classification system whereby the term “idiopathic pulmonary fibrosis” was reserved for patients with a histopathologic diagnosis of usual interstitial pneumonia, therefore excluding other histopathologic lesions with a more favorable prognosis. As a result, cohorts diagnosed with IPF by this classification system are more homogenous and have poorer prognoses than cohorts established with previous systems in which subjects with less ominous histopathologic lesions were (as we know now) misclassified as having IPF. From this dataset, we are unable to determine if coding differences, more sensitive diagnostic testing, changes in classification schema, or a true change in the proportion of patients dying of PF account for these changes.
We faced hurdles common to studies using large ICD-coded or death certificate databases. For example, we were forced to rely on death certifiers to correctly identify cases and then accurately code these conditions on the death certificate. Coding also creates potential problems: To identify decedents with PF, from 1992 to 1998, we had to combine relevant diagnostic codes for IPF including PIPF (ICD-9 code 515) and IPF (ICD-9 code 516.3); however, after 1998, the ICD collapsed these two entities into one diagnostic code (ICD-10 J84.1), which may or may not reflect a group of patients identical to those who would have been coded with ICD-9 515 or 516.3 before 1998.
Coultas and Hughes (30) shed some light on a few of the disparities of coding PF at the state level. They found that PF was never recorded in their state's mortality database unless a patient carried a diagnostic code for PF before death. To complicate things further, they noted that decedents with a clinical diagnosis of IPF (ICD-9 code 516.3) before death were coded only as PIPF (ICD-9 code 515) in their database. Among patients with a clinical diagnosis of PIPF (ICD-9 code 515) before death, only 5% carried that code into the state mortality database. If the results of this study (which used data from only one county in one state) hold true for national data, we believe it is appropriate to combine (as we did in our analyses) ICD-9 codes 515 and 516.3 when attempting to determine mortality rates from IPF. In addition, these coding aberrancies (which have been reported by other investigators) (4) suggest that PF is underreported on death certificates. Thus, the mortality rates we report may in fact underestimate the true burden of disease; however, with the available data, we cannot determine whether differences and changes in the mortality rates we found are due to true shifts in the disease or simply differences in coding.
In an attempt to isolate cases of IPF from known-cause PF, we performed three analyses using increasingly rigorous case definitions for IPF. These analyses produced similar mortality rates and identified few records that included codes for both PF (ICD-9 codes 515 or 516.3, or ICD-10 code J84.1) and any additional codes that might have identified an etiology for PF. Thus, about decedents with ICD-9 codes 515 or 516.3 or ICD-10 code J84.1, we may make either of two assumptions: (1) PF was idiopathic in the majority of decedents or (2) coding patterns are inadequately sensitive to determine the underlying etiology of PF from a death certificate. Because mortality rates were similar between our three analyses (i.e., no matter how we defined IPF), we believed justified in using data from the second analysis—which excluded decedents with connective tissue diseases, radiations fibrosis, and/or asbestosis—that allowed us to compare our data with those from previous studies (5).
Death rates for PF continue to rise steadily and are predicted to continue to rise for the foreseeable future. Rates rise with increasing age, are highest among older people, and are consistently higher in men than women. However, mortality rates in women with PF are climbing more rapidly than in men. Among patients with PF, death from PF has outpaced other causes, such as ischemic heart disease, lung cancer, and pneumonia.
Although once considered an orphan disease, our results suggest that PF should no longer be considered a rarity. The mortality rate from PF is now higher than recent mortality rates from a number of malignancies, including acute myeloid leukemia (31), multiple myeloma (32), and bladder cancer (33). These findings indicate an important and growing problem and provide an argument for more resources focused on the pathobiology of and therapy for this disease.
From 1992 to 1998, UCD diagnostic codes examined included the following: ischemic heart disease (ICD-9 codes 410 through 414.9), heart failure (ICD-9 codes 428 through 428.9), lung cancer (ICD-9 codes 162 through 162.9), pulmonary embolism (ICD-9 code 415.1), pneumonia (ICD-9 codes 480 through 487.8) and cerebrovascular disease (ICD-9 codes 430 through 438). After 1998, diagnostic codes included the following: pulmonary fibrosis (ICD-10 code J84.1), ischemic heart disease (ICD-10 codes I20 through I25), heart failure (ICD-10 codes I50 through I50.9), lung cancer (ICD-10 codes C34 through C34.9), pulmonary embolism (ICD-10 codes I26 through I26.9), pneumonia (ICD-10 code J09 through J18.9), and cerebrovascular disease (ICD-10 codes I60 through I69.8).
For the second analysis, for the years 1992 to 1998, a record was excluded if it contained a code for diffuse connective tissue disease or rheumatoid arthritis (ICD-9 codes 710 through 710.9, or 714 through 714.9), radiation fibrosis (ICD-9 code 508.1), and/or asbestosis (ICD-9 code 501). For records after 1998, a record was excluded if it contained an ICD-10 code for diffuse connective tissue disease (ICD-10 codes M32 through M35.0, M35.1, M35.5, M35.8, M35.9, or M36.0) or rheumatoid arthritis (ICD-10 codes M05 through M05.9, M06 through M06.9, or M08 through M08.9), radiation fibrosis (ICD-10 code J70.1), and/or asbestosis (ICD-10 code J61).
In the third analysis, exclusion criteria included those used for the second analysis, together with several others: From 1992 to 1998, records with ICD-9 codes for coal workers' pneumoconiosis (ICD-9 code 500), silicosis or talcosis (ICD-9 code 502), berylliosis and other inorganic dusts (ICD-9 code 503), unspecified pneumonconiosis (ICD-9 code 505), sarcoidosis (ICD-9 code 135), and/or extrinsic allergic alveolitis (ICD-9 codes 495 through 495.9) were excluded. After 1998, records with the following ICD-10 codes were excluded: coal workers' pneumoconiosis (ICD-10 code J60), silicosis or talcosis (ICD-10 code J62 through J62.8), berylliosis and other inorganic dusts (ICD-10 code J63 through J63.8), unspecified pneumoconiosis (ICD-10 code J64), pneumoconiosis associated with tuberculosis (ICD-10 code J65), sarcoidosis (ICD-10 codes D86 through D86.9), and/or extrinsic allergic alveolitis (ICD-10 codes J67 through J67.9).
| 1. | American Thoracic Society; European Respiratory Society. Idiopathic pulmonary fibrosis: diagnosis and treatment: international consensus statement. Am J Respir Crit Care Med 2000;161:646–664. |
| 2. | Coultas DB, Hubbard R. Epidemiology of idiopathic pulmonary fibrosis. In: Lynch JP, editor. Idiopathic pulmonary fibrosis. New York: Marcel Dekker; 2004. pp. 1–30. |
| 3. | Hubbard R, Johnston I, Coultas DB, Britton J. Mortality rates from cryptogenic fibrosing alveolitis in seven countries. Thorax 1996;51:711–716. |
| 4. | Johnson I, Britton J, Kinnear W, Logan R. Rising mortality from cryptogenic fibrosing alveolitis. BMJ 1990;301:1017–1021. |
| 5. | Mannino DM, Etzel RA, Parrish RG. Pulmonary fibrosis deaths in the United States, 1979–1991: an analysis of multiple-cause mortality data. Am J Respir Crit Care Med 1996;153:1548–1552. |
| 6. | Panos RJ, Mortenson RL, Niccoli SA, King TE. Clinical deterioration in patients with idiopathic pulmonary fibrosis: causes and assessment. Am J Med 1990;88:396–404. |
| 7. | Olson AL, Swigris J, Lezotte D, Norris J, Wilson C, Brown KK. Mortality rates of decedents with pulmonary fibrosis (PF) steadily increased over the period from 1992 to 2003 [abstract]. Am J Respir Crit Care Med 2007;175:A142. |
| 8. | Olson AL, Swigris J, Lezotte D, Norris J, Wilson C, Brown KK. The underlying cause-of-death (UCD) in decedents with pulmonary fibrosis (PF), 1992 to 2003 [abstract]. Am J Respir Crit Care Med 2007;175:A369. |
| 9. | Olson AL, Swigris J, Lezotte D, Norris J, Wilson C, Brown KK. Mortality rates predicted to increase in decedents with pulmonary fibrosis (PF) [abstract]. Am J Respir Crit Care Med 2007;175:A143. |
| 10. | Centers for Disease Control and Prevention; National Center for Health Statistics. Public use data tape documentation: multiple cause of death for ICD-9 1992 data. Hyattsville, MD: U.S. Department of Health and Human Services; 1994. |
| 11. | U.S. Public Health Service. International classification of diseases, 9th revision. Washington, DC: U.S. Government Printing Office; 1980. DHHS Publication No. (PHS) 80-1260. |
| 12. | World Health Organization. ICD-10: international statistical classification of diseases and related health problems, 10th revision. Geneva, Switzerland: World Health Organization; 2003. |
| 13. | Israel RA. Automation of mortality data coding and processing in the United States of America. World Health Stat Q 1990;43:259–262. |
| 14. | Redelings MD, Sorvillo F, Simon P. A comparison of underlying cause and multiple causes of death. Epidemiology 2006;17:100–103. |
| 15. | U.S. Census Bureau. Intercensal population estimates. Available from: http://www.census.gov (accessed November 17, 2006). |
| 16. | Gordis L. Measuring the occurrence of disease: II. Mortality. In: Gordis L, editor. Epidemiology, 3rd ed. Philadelphia: Elsevier Saunders; 2004. pp. 48–70. |
| 17. | Demedts M, Behr J, Buhl R, Costabel U, Dekhuijzen R, Jansen HM, MacNee W, Thomeer M, Wallaert B, Laurent F, et al. High-dose acetylcysteine in idiopathic pulmonary fibrosis. N Engl J Med 2005;353:2229–2242. |
| 18. | Rhagu G, Brown KK, Bradford WZ, Starko K, Noble PW, Schwartz DA, King TE Jr. A placebo-controlled trial of interferon gamma-1b in patients with idiopathic pulmonary fibrosis. N Engl J Med 2004;350:125–133. |
| 19. | Coultas DB, Zumwalt RE, Black WC, Sobonya RE. The epidemiology of interstitial lung diseases. Am J Respir Crit Care Med 1994;150:967–972. |
| 20. | Raghu G, Weycker D, Edelsberg J, Bradford WZ, Oster G. Incidence and prevalence of idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2006;174:810–816. |
| 21. | Gribbin J, Hubbard RB, Le Jeune I, Smith CJ, West J, Tata LJ. The incidence and mortality of idiopathic pulmonary fibrosis and sarcoidosis in the UK. Thorax 2006;61:980–985. |
| 22. | Centers for Disease Control and Prevention. Achievements in public health, 1900–1999: tobacco use—United States, 1990–1999. MMWR Morb Mortal Wkly Rep 1999;48:986–993. |
| 23. | Centers for Disease Control and Prevention. Surveillance for selected tobacco-use behaviors: United States, 1900–1994. MMWR CDC Surveill Summ 1994;43:1–43. |
| 24. | Taskar VS, Coultas DB. Is idiopathic pulmonary fibrosis an environmental disease? Proc Am Thorac Soc 2006;3:293–298. |
| 25. | Lynch JP, Saggar R, Weigt SS, Zisman DA, White ES. Usual interstitial pneumonia. Semin Respir Crit Care Med 2006;27:634–651. |
| 26. | Lynch D. Imaging of diffuse parenchymal lung diseases. In: Schwarz MI, King TE, editors. Interstitial lung disease, 4th ed. Hamilton, ON, Canada: BC Decker; 2003. pp. 75–113. |
| 27. | Young LM, Hopkins R, Wilsher ML. Lower occurrence of idiopathic pulmonary fibrosis in Maori and Pacific Islanders. Respirology 2006;11:467–470. |
| 28. | Birkmeyer JD. Primer on geographic variation in health care. Eff Clin Pract 2001;4:232–233. |
| 29. | Katzenstein ALA, Myers JL. Idiopathic pulmonary fibrosis: clinical relevance of pathologic classification. Am J Respir Crit Care Med 1998;157:1301–1315. |
| 30. | Coultas DB, Hughes MP. Accuracy of mortality data for interstitial lung diseases in New Mexico, USA. Thorax 1996;51:717–720. |
| 31. | National Cancer Institute. SEER cancer statistics review, 1975–2003. Acute myeloid leukemia. Available from: http://seer.cancer.gov/statfacts/html/amyl.html?statfacts_page=amyl.html&x=13&y=18 (accessed February 11, 2007). |
| 32. | National Cancer Institute. SEER cancer statistics review, 1975–2003. Myeloma. Available from: http://seer.cancer.gov/csr/1975_2003/results_merged/sect_18_myeloma.pdf (accessed February 11, 2007). |
| 33. | National Cancer Institute. SEER cancer statistics review, 1975–2003. Cancer of the urinary bladder. Available from: http://seer.cancer.gov/statfacts/html/urinb.html?statfacts_page=urinb.html&x=19&y=16 (accessed February 11, 2007). |
