Abstract
RATIONALE:
Idiopathic pulmonary fibrosis (IPF) begins years before symptoms appear, but the natural history is incompletely understood.
OBJECTIVES:
To describe and quantify IPF disease progression using four pulmonary function test (PFT) parameters.
METHODS:
Two cohorts included 245 adults with subclinical through advanced familial pulmonary fibrosis (FPF) or 347 placebo-treated IPF patients enrolled in two randomized controlled trials (RCTs). A Bayesian joint repeated measures model was fit to describe the observed PFT values as a function of the estimated years since onset (EYO). A latent variable estimated the subject-level chronological age at onset. Onset was assumed to occur when the diffusion capacity for carbon monoxide (DLCO) was 70%-predicted. The relationships between EYO and clinical outcomes not included in the model (e.g., transplant-free survival) were evaluated, adjusting for age and sex.
RESULTS:
In FPF, the DLCO declined steadily starting around EYO −10, reaching 86.8%-predicted by EYO −5 and 45.3%-predicted by +5. The forced vital capacity (FVC) declined later, reaching 98.6%-predicted by EYO −5 and 76.2%-predicted by +5. The annualized decline in FVC was 12-fold greater in the year after EYO +5 (6.14%-predicted) than −5 (0.49%-predicted). There was a 31% higher risk of death or transplant (HR 1.31, 95% CI 1.25–1.37) per 1-year increase in EYO. Similar findings were observed in the RCTs.
CONCLUSIONS:
We identified a conserved sequential decline in lung function in IPF, which has important implications for the design of clinical trials. The DPM provides a powerful framework to investigate the clinical and/or biological processes that initiate and propagate IPF.
Keywords: Idiopathic Pulmonary Fibrosis, Natural History, Lung Diseases, Interstitial
INTRODUCTION
Idiopathic pulmonary fibrosis (IPF) is the most common form of progressive interstitial lung disease (ILD), with an estimated prevalence of nearly 1 per 200 people over age 65 years in the United States.(1) At the time of diagnosis, most individuals already have a high symptom burden (dyspnea, cough), have lost substantial lung function, and have a median survival of less than 5 years.(1–4) Three disease-modifying treatments modestly slow the rate of lung function decline but do not reverse symptoms,(5–7) presenting an urgent need to facilitate earlier diagnosis and treatment.
While all individuals with IPF experience lung function decline over time, a widely accepted view is that the rate of decline varies greatly across individuals.(8–10) Currently, trials of disease-modifying treatments use change in the forced vital capacity (FVC) as the primary endpoint. With high inter-individual variability in the rate of change in FVC, trials require large numbers of patients observed over extended periods and/or population enrichment strategies.(5, 6) Improved understanding of the full disease natural history, including the rate and timing of progression on symptoms, lung function, imaging and other parameters, is needed to improve the design of clinical trials, develop more effective therapies, and offer the potential for prevention trials.
IPF is hypothesized to begin years before symptoms appear.(8, 11–13) The best biomarker of subclinical disease is interstitial lung abnormalities (ILAs) on chest high-resolution computed tomography (HRCT) imaging.(14, 15) ILAs have good construct validity as a precursor of IPF, sharing genetic and environmental risk factors,(11, 13, 16) declining lung function,(17) and increased respiratory morbidity and mortality.(18, 19) Our data suggested that progression from early ILAs to more advanced disease is relatively consistent across individuals.(20)
A Bayesian disease progression model (DPM) is a useful method to study a slowly progressive disease. Instead of following each individual in a cohort over the entire disease development period, which is often not feasible, a DPM provides alignment of individuals observed at various points to synthesize a complete progression trajectory.(21, 22) Our objective was to determine the rate and timing of decline in lung function in IPF using a Bayesian DPM to estimate the degree of deviation from population-normative values on four pulmonary function test (PFT) parameters among individuals with subclinical through advanced familial pulmonary fibrosis (FPF). We further evaluated whether FPF and IPF have similar progression features by applying the DPM to the placebo arms of two large randomized controlled trials (RCTs).(23) Some of these results were presented as an abstract.(24)
METHODS
Cohorts and Data Collection
The Vanderbilt FPF Registry enrolls families in which two or more blood-related members have clinically-diagnosed FPF (“probands”), including at least one with IPF.(25) Asymptomatic first-degree relatives of probands (“relatives”) participated in a prospective study designed to identify individuals with subclinical FPF using serial screening visits with HRCT (Figure E1A).(13, 16, 20) Relatives with subclinical FPF experienced new development or progression of ILAs across HRCTs, or met criteria for clinically-diagnosed FPF at or after the enrollment HRCT. Relatives with clinically-diagnosed FPF had “extensive” ILAs on HRCT or were diagnosed with pulmonary fibrosis by a pulmonologist.(20) Relatives with clinical FPF identified during screening are called “subclinical FPF” for clarity on how they entered the cohort. The DPM analysis included all relatives with subclinical FPF and probands who received medical care at the Vanderbilt ILD clinic between 2003–2024, provided that at least one diffusion capacity for carbon monoxide (DLCO) measurement was available (Figure E1B–C). Relatives without evidence of subclinical FPF on 2 or more HRCTs (i.e., absent ILAs) were comparators in some analyses.
PFTs occurred in Vanderbilt’s PFT laboratory several times annually among probands or at screening visits among relatives. Raw values obtained from the electronic medical record (EMR) were converted to %-predicted using Global Lung Initiative (GLI) formulas.(26–29) A research questionnaire administered at Registry enrollment captured cough and dyspnea (Figure E2).(13, 20) The age at clinical FPF diagnosis is self-reported at enrollment among probands or was the date when a relative met criteria for clinically-diagnosed FPF (Figure E1A).(20) Use of supplemental oxygen was obtained from the EMR at the earliest visit with pulse oximetry (SpO2), recorded as none, with exertion, or at rest. Exposure to pirfenidone or nintedanib was obtained from the EMR and recorded as “never” or “ever” used during observation on PFTs.(5, 6) The timing of lung transplantation or death was obtained from the EMR or self-reporting. Follow-up time for those without a terminal event was censored on the date of the most recent contact through 31-March-2024. PFTs and other measurements collected after lung transplantation were censored. These studies were approved by the Vanderbilt IRB (#020343, #080780).
Placebo-treated participants in the CAPACITY-004/006 phase 3 RCTs evaluating the effect of pirfenidone compared to placebo among patients with IPF were used as an independent cohort.(23) FVC, 1-second forced expiratory volume (FEV1), and DLCO were measured every 12 weeks, and total lung capacity (TLC) was measured at weeks 0, 36 and 72. After week 72, blinded treatment was maintained and PFTs were measured every 12 weeks until the last patient enrolled reached week 72. Baseline six-minute walk SpO2, symptoms, and the occurrence of death or transplant were also available. Data were accessed using the Vivli platform and processed like the FPF cohort (Methods E1).
Statistical Methods
Our goal was to align individuals using a single, continuous latent variable, estimated chronological age at onset, to characterize progression as it relates to four PFT parameters (FVC, FEV1, TLC, DLCO). For model identifiability, the point of “onset” was assumed to occur when the DLCO was at 70%-predicted. DLCO was selected because it had a wide range of values in the population, appeared to become abnormal earlier than other parameters,(20) and a proof-of-concept analysis demonstrated good alignment of the other PFTs when the DLCO value was used to estimate the chronological age at onset using a simple logistic formula (Figure E3).
The DPM (Figure E4) is a joint repeated measures mixed-effects model that simultaneously models observed PFT parameters values as a function of the subject-level estimated years since onset (EYO). The data input were each observed PFT parameter (k) value (in %-predicted) for subject, i at visit, j, and visit timing (ti,j, chronological age at visit, in years). The hierarchical DPM estimates αi, the subject-level chronological age at onset (i.e., DLCO 70%-predicted). The EYO at a given visit is ti,j-αi. The PFT parameter-specific timing (θk) and rate (βk) of decline, relative to onset, are estimated. A random effect, Mi,k, accounts for variability in the subject- and PFT parameter-specific inherent lung function. The DPM makes several key assumptions. First, lung function decline follows a logistic decay shape. Within this general form, the Bayesian prior assumptions about the timing and rate of decline on each PFT cover a wide range of possibilities. For example, each PFT (except DLCO) was assumed to be between 2%- and 98%-predicted at onset, and to decline between 0–30%-predicted within one year after being at 80%-predicted. Second, conditional inherent lung function, subjects have similar rates of progression at any given EYO. The subject- and PFT parameter-specific inherent lung function (Mi,k) is assumed to be correlated within and across healthy subjects; this is clinically reasonable and supported by data from relatives without subclinical FPF (Figure E5). All Bayesian prior assumptions are in Table E1.
Computation utilized the rjags package in R (example program in Methods E2). The joint posterior distributions of all model parameters were fit using Markov chain Monte Carlo method,(30) with the posterior means and 95% equal-tailed credible intervals for the estimates reported. Model performance is indicated by the Gelman-Rubin convergence and the uncertainty of the estimated chronological age at onset (αi), calculated as the per-subject width of the 95% credible interval across dependent samples. Sensitivity analyses in the FPF cohort compared DPM results when using different prior assumptions. Secondary analyses assessed DPM results with the cohort stratified by sex, smoking status, and MUC5B minor (risk) allele (rs35705950-T) carrier status (FPF cohort).(16).
To determine if EYO is associated with progression on outcomes not included in the DPM, logistic regression estimated the odds of having a cough, and proportional odds regression estimated the odds of having worse dyspnea or oxygen use by EYO at outcome assessment. Cox proportional hazards models estimated the risk of lung transplant or death by EYO at the earliest PFT visit. All models included chronological age and sex as covariates. Transplant-free survival probability was visualized using Kaplan-Meier plots, with subjects categorized by increasing EYO at the earliest PFT. Analyses were performed using R V.4.3.4 and SAS V.8.3. P-values <0.05 were considered significant.
RESULTS
Participant Characteristics
The FPF Cohort included 245 subjects (184 probands, 61 subclinical FPF). Among probands, the mean chronological age at first PFT was 64.2 ±10.0 years, 119 (63%) were male, 79 (53%) were ever-cigarette smokers, and 127 (71%) were MUC5B carriers (Table 1). The mean first-recorded FVC was 72.0 ±19.5 and DLCO 48.7 ±16.7 %-predicted. The mean dyspnea score was 2.5 ±1.8, and 56 (56% of those reporting) had a regular cough. Among relatives with subclinical FPF, the mean age at the first PFT was 60.5 ±9.2 years, 25 (41%) were male, 18 (30%) were ever-smokers, and 26 (43%) were MUC5B carriers. Pulmonary function was preserved, with FVC 99.9 ±17.2 and DLCO 83.8 ±16.3 %-predicted. Among subclinical FPF, 22 (36%) developed ILAs after enrollment, 36 (59%) had early ILAs at enrollment that progressed, and 3 (5%) had extensive ILAs at enrollment. The most common FPF subtype was IPF, present in 151 (81%) of probands, and 13 (65%) of 20 relatives who developed clinically-diagnosed FPF during our longitudinal study. Approximately half of those with clinically-diagnosed FPF were exposed to antifibrotic medications. Most probands were observed until death or lung transplant while most subclinical FPF were censored at the most recent contact. Among 347 placebo-treated IPF patients in the RCTs, the mean age at enrollment was 66.6 ±7.7 years, 252 (73%) were male, the FVC was 80.8 ±15.8 and DLCO 59.2 ±12.0 %-predicted (Table 1, Table E2).
Table 1.
Participant Characteristics
| Familial Pulmonary Fibrosis (n=245) | IPF Patients in CAPACITY Pooled Placebo Arms (n=347) | ||
|---|---|---|---|
| Characteristic | Relatives (Subclinical FPF) (n=61) | Probands (Clinical FPF) (n=184) | |
| Chronological age at first PFT, years | 60.5 (9.2) | 64.2 (10.0) | 66.6 (7.7) |
| Male Sex | 25 (41%) | 116 (63%) | 252 (72.6%) |
| Race/Ethnicity | n=142 | NA | |
| Ever Smoker | 18 (30%) | 76 (52%); n=146 | 232 (67%) |
| MUC5B Genotype | n=176 | NA | |
| Baseline PFT, %-predicted | |||
| Enrollment Dyspnea Score, range 0–5 | 0.8 (1.3) | 2.4 (1.8); n=98 | 1.7 (0.98); n=337 |
| Regular Cough | 12 (20%) | 54 (55%); n=98 | 304 (87.6%) |
| Oxygen Use | n=52 | n=171 | |
| ILA Status at Enrollment HRCT | NA | NA | |
| Chronological age at Diagnosis, years | 63 (57–66); n=20$ | 65 (58–70) | NA |
| Pulmonary Fibrosis Subtype | n=20$ | ||
| Antifibrotic Medication Use$ | n=20 | ||
| Terminal Events | |||
| Median Observation Time, years | |||
Data are expressed as N (%), mean (standard deviation), or median (q1-q3).
Reported as “Other” due to small numbers in racial/ethnic strata of Hispanic, Asian, and Black increasing the risk of participant identification.
Only those 20 first degree relatives who were diagnosed with clinical pulmonary fibrosis by March 2024 were assigned a pulmonary fibrosis subtype or eligible for antifibrotic medication prescription. In addition, probands enrolled in the Registry as early as 2003, and relatives as early as 2008. Disease-modifying antifibrotic drugs became available in 2014; those whose observation time ended before 2014 did not have the opportunity to receive treatment.
The end of follow-up is the date of lung transplant, death (in those not transplanted), or the last contact in the study or medical records as of March 2024.
FPF: familial pulmonary fibrosis; PFT: pulmonary function test; DLCO: diffusion capacity of the lung for carbon monoxide; FVC: forced vital capacity; FEV1: forced expiratory volume in 1 second; TLC: total lung capacity; HRCT: high-resolution chest CT scan; IPF: idiopathic pulmonary fibrosis; IIP: idiopathic interstitial pneumonia
NA indicates data are not available or characteristic not relevant for this group.
Physiologic Progression
The majority of the FPF cohort had >1 PFT visit, with a median of 2 (1–3) visits per subject in subclinical FPF and 4.5 (2–10) in probands (Table E3). Onset (EYO 0) was defined as DLCO 70%-predicted and was estimated to occur at a median chronological age of 64.6 (56.3–72.3) years. PFT visits occurred between EYO −32.8 (i.e., an estimated 32.8 years before this subject’s DLCO is expected to reach 70%-predicted) and +11.3. In the RCTs, the estimated age at onset was 63.0 (57.4–68.8) years, and PFT visits occurred between EYO −15.0 and +12.7 (Table E4; Figure E6).
After alignment by EYO, there was a notable difference in the timing and rate of decline on each PFT parameter in FPF (Figure 1A, Table 2). The FVC does not begin to decline until around onset, with the annualized change escalating sharply several years after onset. For example, at EYO −5, the estimated value of FVC was 98.6 (95% credible interval 98.0–99.2) %-predicted, with annualized change of −0.49 (0.33–0.65) %-predicted. By EYO +5, the estimated value of FVC was 76.2 (72.4–80.0) %-predicted, with annualized change −6.14 (5.39–6.95) %-predicted. The FEV1 followed a similar trajectory. On the other hand, TLC and DLCO begin to decline earlier. At EYO −5, TLC has an estimated value of 86.4 (83.7–89.0) with annualized change −1.70 (1.57–1.82) %-predicted, while at EYO +5 the estimated value was 61.6 (59.1–64.1) with annualized change −3.32 (2.91–3.76) %-predicted. At EYO −5, DLCO had an estimated value of 86.8 (85.8–87.8) and annualized change −2.56 (2.48–2.62) %-predicted, while at EYO +5 the estimated value was 45.3 (43.1–47.3) with annualized change −5.07 (4.72–5.41) %-predicted. The RCTs cohort demonstrated similar findings (Figure 1B).
Figure 1. Expected pulmonary function decline as a function of estimated years since onset in idiopathic pulmonary fibrosis.
The expected function on each pulmonary function test (PFT) parameter (in %-predicted, y-axis) is plotted as a function of estimated years since onset (EYO, x-axis) in A) patients with familial pulmonary fibrosis (FPF) or B) patients with idiopathic pulmonary fibrosis (IPF) randomized to the pooled placebo arm of the CAPACITY 004 or 006 randomized controlled trials (RCTs). The solid lines reflect the posterior mean, and the shaded region corresponds to the 95% equal-tailed credible interval. The vertical solid line indicates EYO 0 (onset), when the DLCO is expected to be at 70%-predicted and other PFT parameters are at their respective intercepts, which are shown on the figure as colored numbers (see also Table E4). The grid below the plot shows the number (N) of unique subjects whose first PFT visit fell in a given interval of EYO, and the median and interquartile range for the uncertainty (i.e., the per-subject width of the 95% credible interval across dependent samples) of the estimated chronological age at onset for those subjects. In addition, the total number of visits with DLCO, spirometry (FVC, FEV1), or TLC are shown for each interval of EYO (each subject can contribute multiple visits in a single interval and in more than one interval). PFT: pulmonary function test; EYO: estimated years since onset; DLCO: diffusion capacity of the lung for carbon monoxide; FVC: forced vital capacity; FEV1: forced expiratory volume in 1 second; TLC: total lung capacity
Table 2.
Estimated pulmonary function test values and 1-year decrease across a range of estimated years since onset in Familial Pulmonary Fibrosis
| EYO | DLCO | FVC | FEV1 | TLC | ||||
|---|---|---|---|---|---|---|---|---|
| Value | 1-year dec. | Value | 1-year dec. | Value | 1-year dec. | Value | 1-year dec. | |
| −10 | 94.9 (94.0–95.7) | 1.10 (1.03–1.17) | 99.7 (99.5–99.9) | 0.11 (0.06–0.16) | 99.9 (99.7–99.9) | 0.06 (0.03–0.10) | 92.7 (90.5–94.6) | 0.99 (0.84–1.12) |
| −9 | 93.8 (92.9–94.6) | 1.32 (1.26–1.37) | 99.6 (99.3–99.8) | 0.14 (0.08–0.22) | 99.8 (99.6–99.9) | 0.08 (0.04–0.13) | 91.7 (89.4–93.8) | 1.11 (0.97–1.24) |
| −8 | 92.4 (91.5–93.4) | 1.58 (1.53–1.61) | 99.5 (99.1–99.7) | 0.19 (0.12–0.28) | 99.7 (99.5–99.9) | 0.11 (0.06–0.18) | 90.6 (88.1–92.8) | 1.24 (1.10–1.36) |
| −7 | 90.9 (89.9–91.8) | 1.87 (1.85–1.87) | 99.3 (98.8–99.6) | 0.26 (0.17–0.38) | 99.6 (99.3–99.8) | 0.16 (0.09–0.24) | 89.3 (86.8–91.7) | 1.38 (1.24–1.50) |
| −6 | 89.0 (88.0–90.0) | 2.20 (2.17–2.21) | 99.0 (98.5–99.4) | 0.36 (0.24–0.50) | 99.5 (99.1–99.7) | 0.22 (0.13–0.33) | 88.0 (85.3–90.4) | 1.53 (1.40–1.66) |
| −5 | 86.8 (85.8–87.8) | 2.56 (2.48–2.62) | 98.6 (98.0–99.2) | 0.49 (0.33–0.65) | 99.2 (98.8–99.6) | 0.31 (0.19–0.45) | 86.4 (83.7–89.0) | 1.70 (1.57–1.82) |
| −4 | 84.2 (83.4–85.2) | 2.95 (2.82–3.07) | 98.2 (97.3–98.8) | 0.66 (0.47–0.86) | 98.9 (98.3–99.4) | 0.43 (0.28–0.61) | 84.7 (82.0–87.4) | 1.87 (1.73–2.00) |
| −3 | 81.3 (80.5–82.1) | 3.36 (3.17–3.55) | 97.5 (96.5–98.4) | 0.88 (0.65–1.13) | 98.5 (97.7–99.1) | 0.60 (0.41–0.82) | 82.9 (80.1–85.6) | 2.05 (1.90–2.20) |
| −2 | 77.9 (77.4–78.5) | 3.77 (3.52–4.03) | 96.6 (95.3–97.7) | 1.18 (0.90–1.47) | 97.9 (96.9–98.7) | 0.84 (0.59–1.11) | 80.8 (78.0–83.6) | 2.23 (2.06–2.41) |
| −1 | 74.2 (73.9–74.5) | 4.17 (3.85–4.50) | 95.4 (93.9–96.8) | 1.57 (1.24–1.91) | 97.1 (95.8–98.1) | 1.16 (0.85–1.49) | 78.6 (75.8–81.4) | 2.41 (2.21–2.62) |
| 0 | 70.0 (NA) | 4.53 (4.16–4.92) | 93.9 (92.0–95.6) | 2.06 (1.68–2.46) | 95.9 (94.3–97.3) | 1.59 (1.21–1.99) | 76.2 (73.4–79.0) | 2.59 (2.36–2.84) |
| +1 | 65.5 (65.1–65.8) | 4.83 (4.42–5.27) | 91.8 (89.6–93.9) | 2.68 (2.23–3.13) | 94.3 (92.4–96.0) | 2.16 (1.71–2.63) | 73.6 (70.9–76.3) | 2.77 (2.49–3.06) |
| +2 | 60.6 (59.8–61.4) | 5.04 (4.61–5.50) | 89.1 (86.5–91.6) | 3.42 (2.91–3.95) | 92.1 (89.8–94.3) | 2.90 (2.36–3.45) | 70.8 (68.2–73.5) | 2.94 (2.62–3.27) |
| +3 | 55.6 (54.3–56.8) | 5.16 (4.73–5.61) | 85.7 (82.7–88.6) | 4.28 (3.70–4.89) | 89.2 (86.4–91.9) | 3.80 (3.17–4.47) | 67.9 (65.3–70.5) | 3.09 (2.73–3.46) |
| +4 | 50.4 (48.7–52.1) | 5.17 (4.77–5.58) | 81.4 (78.0–84.8) | 5.21 (4.54–5.92) | 85.4 (82.1–88.6) | 4.86 (4.13–5.66) | 64.8 (62.3–67.4) | 3.21 (2.83–3.63) |
| +5 | 45.3 (43.1–47.3) | 5.07 (4.72–5.41) | 76.2 (72.4–80.0) | 6.14 (5.39–6.95) | 80.6 (76.7–84.4) | 6.01 (5.16–6.94) | 61.6 (59.1–64.1) | 3.32 (2.91–3.76) |
| +6 | 40.2 (37.7–42.6) | 4.86 (4.59–5.12) | 70.1 (65.8–74.3) | 6.96 (6.15–7.86) | 74.6 (70.1–78.9) | 7.13 (6.17–8.18) | 58.3 (55.7–60.8) | 3.40 (2.97–3.86) |
| +7 | 35.4 (32.6–38.0) | 4.58 (4.39–4.73) | 63.1 (58.4–67.8) | 7.55 (6.71–8.50) | 67.4 (62.4–72.4) | 8.04 (7.01–9.16) | 54.9 (52.3–57.5) | 3.44 (3.01–3.91) |
| +8 | 30.8 (27.9–33.6) | 4.22 (4.12–4.29) | 55.6 (50.5–60.6) | 7.81 (6.99–8.74) | 59.4 (53.8–64.9) | 8.57 (7.54–9.69) | 51.4 (48.7–54.1) | 3.46 (3.03–3.92) |
| +9 | 26.6 (23.6–29.5) | 3.83 (3.80–3.84) | 47.8 (42.4–53.1) | 7.70 (6.95–8.54) | 50.8 (44.8–56.7) | 8.61 (7.66–9.63) | 48.0 (45.0–50.8) | 3.44 (3.02–3.88) |
| +10 | 22.7 (19.8–25.7) | 3.42 (3.33–3.47) | 40.1 (34.6–45.6) | 7.23 (6.59–7.95) | 42.2 (36.1–48.4) | 8.16 (7.35–9.04) | 44.5 (41.4–47.6) | 3.38 (2.99–3.80) |
| +11 | 19.3 (16.4–22.2) | 3.01 (2.86–3.12) | 32.8 (27.5–38.3) | 6.50 (5.91–7.14) | 34.1 (28.0–40.2) | 7.30 (6.58–8.10) | 41.1 (37.8–44.4) | 3.30 (2.94–3.68) |
Data presented are the posterior mean and 95% credible interval for the estimated value and 1-year decrease in each pulmonary function test (PFT) parameter for each given year since onset (EYO), in %-predicted. The 1-year decrease is the year-to-year difference in the estimated value for each PFT parameter. Please note that the 95% credible interval indicates a 95% probability that the true value of a parameter falls within this range, given the Bayesian prior assumptions and observed data from the whole cohort. Individual-level variability in the timing of initial progression (i.e., αi, the subject-level chronological age at onset) and inherent lung function (i.e., the random effect Mi,k) are accounted for in the DPM and are effectively “removed” in these population-level estimates. DLCO: diffusion capacity of the lung for carbon monoxide; FVC: forced vital capacity; FEV1: forced expiratory volume in 1 second; TLC: total lung capacity.
Model performance was good in both cohorts, with median uncertainty in the estimated chronological age at onset of 4.5 (3.3–13.4) years in the FPF cohort and 6.1 (4.1–10.3) years in the RCTs cohort (Table E4). Uncertainty was high among those with PFTs at EYO less than −10 (24.7 years in FPF; 30.8 in RCTs); uncertainty improved with advancing EYO and was 3.8 years among those with PFT at EYO 0 to +5 (Figure 1, grid). Gelman-Rubin convergence was near 1, indicating good performance (Table E5).
Sensitivity analyses in the FPF cohort demonstrated that the cohort’s median estimated chronological age at onset, and mean rate and timing of PFT parameter decline were little-changed when using different prior assumptions (Table E5). In the secondary analyses, there was little difference across subgroups (sex, smoking, MUC5B risk allele carrier status) in the rate of decline on any PFT parameter. For example, in the year after FVC is 80%-predicted, it is estimated to decline by 4.8%-predicted (4.1–5.6) in males and 4.6%-predicted (3.7–5.7) in females (Table E6). The timing of decline on volume measurements (FVC, FEV1, TLC) relative to DLCO may differ across subgroups. In the FPF cohort, females, MUC5B non-carriers, and ever-smokers were approximately 5%-predicted higher on lung volumes at onset than their counterparts. Similar findings were observed in the RCTs cohort by smoking status (Table E7).
Clinical Progression
In the FPF cohort, most PFT visits among relatives with subclinical FPF who were not yet clinically-diagnosed with FPF occurred at EYO<0 (green dots in Figure E6A), while the clinically-diagnosed relatives (blue dots) had visits at EYO −10 to +10, and most probands (pink dots) had visits at EYO>0. In Figure 2, the observed PFT values and corresponding HRCT images obtained at various EYO from several individuals with FPF are superimposed on the cohort’s estimated trajectories, demonstrating the evolution of ILAs (Subjects A, B) toward the characteristic architectural distortion (Subject C) of advanced IPF. In the FPF cohort, with every 1-year increase in EYO, the odds of having higher (worse) dyspnea increased by 13% (covariate-adjusted OR 1.13, 1.07–1.18), the odds of having a cough increased by 11% (aOR 1.11 95% CI 1.06–1.17), and the odds of using supplemental oxygen increased by 37% (aOR 1.37, 1.25–1.50) (Figure 3). Increasing EYO at first PFT corresponded with shorter transplant-free survival time, which was ≥16.8 years among those with EYO≤−10, 7.9 (95% CI 7.1–9.1) among −10<EYO≤0 years, 4.8 (4.1–5.9) among 0<EYO≤+5 years, and 1.5 (1.2–1.9) among those with EYO>+5 years (Figure 4). With every 1-year increase in EYO, there was a 31% higher risk of death or transplant (covariate-adjusted HR 1.31, 95% CI 1.25–1.37). Similar relationships between EYO and transplant-free survival (Figure E7), dyspnea, or oxygen (Figure 3) were observed in the RCTs. Combined, this suggests good alignment of DPM estimates with clinical observations about the relative timing of clinical diagnosis, progressive of symptoms and imaging, initiation of supplemental oxygen, and the occurrence of lung transplantation or death.
Figure 2. Observed pulmonary function and ILAs in 3 FPF subjects at varying phases in disease development.
The observed PFT values (colored shapes) from 3 subjects with FPF are superimposed on the FPF cohort’s posterior mean (dotted lines) and 95% credible intervals (shaded areas) on each pulmonary function test (PFT) parameter. The 95% credible interval indicates a 95% probability that the true value of the unknown PFT parameter falls within this range, given the Bayesian prior assumptions and observed data from the whole cohort. Individual-level variability in the inherent lung function, apparent in the observed PFT values, is accounted for in the DPM by the random effect, Mi,k, and is not reflected by the population-level posterior mean and corresponding 95% credible intervals. The black shapes indicate the timing of a high-resolution chest computed tomography (HRCT) scan. Patient A (circles) was a relative with subclinical FPF with an HRCT scan at EYO −8.2 that showed mild reticulation limited to the extreme lung bases. Patient B was a relative with subclinical FPF with an HRCT at EYO −0.6 that showed multifocal reticulation in the lower lobes, without architectural distortion. Patient C was a proband with an HRCT at EYO +4.0 that showed peripheral and basilar predominant reticulation with architectural distortion that progressed to end-stage disease by the time of another HRCT at EYO +7.8. The HRCT images show the characteristic progression of ILAs from very minor/focal to advanced disease. PFT: pulmonary function test; EYO: estimated years since onset; DLCO: diffusion capacity of the lung for carbon monoxide; FVC: forced vital capacity; FEV1: forced expiratory volume in 1 second; TLC: total lung capacity
Figure 3. The association between estimated years since onset and dyspnea, cough or oxygen use.
Each panel shows dot plots and superimposed box-whisker plots among A) FPF or B) RCTs participants. Each participant is a dot located at the value of their clinical outcome response (dyspnea, cough, oxygen; Y-axis) and estimated years since onset (EYO, X-axis) at the time of the outcome assessment. The vertical line in the box is the median, the edges of the box are the lower and upper quartiles, and the whiskers are 1.5*IQR of EYO at assessment among participants in each outcome category. The chronological age- and sex-adjusted odds of having a worse outcome response per 1-year increase in EYO, 95% confidence intervals, and p-values are at the top of each panel. The dyspnea outcome has a range of 0 to 5, where 5 indicates worse (more severe) dyspnea.
Figure 4. Lung transplant-free survival after first PFT visit – FPF Cohort.
The Kaplan-Meier survival probability is shown by time since the first PFT visit among FPF Cohort participants stratified by estimated years since onset (EYO) at the time of the first PFT visit. The mean and standard deviation of the chronological age at first PFT visit are shown for each strata of EYO. The number at risk for each strata are in the table below the plot. The crosses represent censored non-events while the steps represent lung transplants or deaths. PFT: pulmonary function test; EYO: estimated years since onset.
DISCUSSION
IPF is characterized by progressive decline in lung function culminating in death from respiratory failure. In this study, we used a unique cohort of individuals with subclinical through advanced disease to define the rate and timing of progression on four PFT parameters across the disease development period. We identified a conserved sequence whereby DLCO and TLC deviate from population-normative values earlier and at a relatively steady rate, while FVC and FEV1 undergo accelerated decline during more advanced disease. This pattern was also observed among IPF patients randomized to the placebo arms of two large RCTs. The latent parameter used to align subjects to study PFT progression, EYO, was strongly associated with clinical indicators of progression (dyspnea, cough, hypoxemia, and transplant-free survival) that were not included in the DPM.
To account for the potential for a variable timing of the onset and rate of progression across PFT parameters, the DPM jointly models the PFT parameter-specific onset and rate of progression along with the subject-specific EYO, which is also estimated from the data. Once aligned, a systematic progression on the observed PFTs across a decades-long disease development period is evident. We identified an acceleration in the annualized change in FVC from ~1%-predicted during the ~15 years before onset to over 5%-predicted annually by ~4 years after onset. Several studies also support within-subject FVC acceleration. For example, larger interval FVC decline was associated with shorter transplant-free survival.(31–34) Using a cluster analysis of FVC trajectory, Fainberg and colleagues found that the two clusters with the largest FVC decline had the lowest baseline DLCO values.(9) Previous studies have also conceptually supported the DPM results, indicating an early decline in DLCO and TLC.(11) There were subtle between-cohort differences, with less acceleration in the rate of DLCO decline with advancing EYO in the RCTs cohort compared to the FPF cohort. This may be related to differences in the types of subjects included in the FPF vs RCTs cohorts. The RCTs cohort included a narrower range of EYO, with very early or advanced disease underrepresented.
The DPM estimates past and future disease progression for individual subjects, conditional upon the per-subject onset and inherent lung function. The accuracy of long-term prediction depends on the degree to which modeling assumptions are met. First, we assume that PFT decline follows a logistic decay. This could mask a stepwise decline, which may occur in some patients who experience acute disease exacerbations.(35) Further study is needed to understand the utility of more flexible decline functions. Second, we assume that subjects do not have different rates of progression after accounting for the timing of initial progression and the inherent lung function. Subject-level comorbidities or other biological factors may result in between-subject differences in inherent lung function or progression.(10) We explored this possibility in our secondary analysis. In FPF, lung volumes, particularly TLC, were lower for a given level of DLCO in MUC5B carriers than non-carriers, which could suggest genetic differences in the onset of progression.(25, 36, 37) In both cohorts, the TLC was higher at a given level of DLCO in ever-smokers, which may be due to comorbid emphysema.(38, 39) Despite subtle differences in the timing, the rate of progression on each PFT was similar across subgroups. We did not study pulmonary hypertension, which may impact the DLCO. Further research is needed to understand how well the DPM predicts the long-term progression in individuals from various subgroups.
Our observation that FVC progression accelerates sharply during advanced disease has important implications for clinical trial design. FVC is currently the primary endpoint to support new drug labeling. With little progression in early disease and accelerated decline later, FVC will not be a useful endpoint for secondary prevention trials but could still play a key role in trials enrolling patients with more advanced disease.(15, 40–42) Using the baseline value of FVC and/or DLCO to identify those likely to experience high-magnitude FVC progression may offer advantages over proposed alternative prognostic enrichment strategies using genomic, proteomic, and radiomic signatures.(43–45)
There are several important limitations. First, the median uncertainty in the per-subject chronological age at onset was 4.5 years in the FPF and 6.2 years in the RCTs cohort. PFT parameters are near-normal in early disease and may, therefore, have higher test-to-test variability.(46) Both model ignorance and data noise likely contribute to the high uncertainty in the timing of onset during very early disease. Incorporating additional parameters, such as quantitative imaging or fibrosis-associated biomarkers may improve the understanding of both the subclinical timeline and disease mechanisms.(12, 16, 47–50) The narrow credible intervals suggest the DPM has high precision and confidence in the unknown PFT progression parameters in both cohorts. Second, the Bayesian DPM is a hybrid of cross-sectional and longitudinal design in that most individuals were observed longitudinally, but none were observed over the entire disease development period, which should be considered during interpretation. Third, we did not adjust for use of disease-modifying treatment in the FPF cohort. The RCTs cohort was intentionally limited to individuals who were not exposed to these medications before or during the PFT visits used for modeling.(5, 6) Fourth, a similar pattern of progression was observed in FPF and in IPF patients enrolled in RCTs, but differences in the underlying disease biology between these entities remains possible. Finally, individuals with truly subclinical disease (as defined in the FPF cohort) were unrepresented in the RCTs cohort, which nonetheless included physiologically-mild IPF (DLCO>70%).(23)
In conclusion, we identified a conserved sequential decline in lung function in patients with familial or idiopathic pulmonary fibrosis. The Bayesian DPM provides a novel and powerful framework to investigate the relative timing of clinical and/or biological processes that initiate and propagate IPF. Improved understanding of the outcome-specific progression at different points in the disease development period will enhance the design of clinical trials by aligning entry criteria that are informative about an individual’s location in the disease development period with the therapy’s mechanistic target and endpoint(s) on which progression is measured. Collectively, this is an important step toward developing new and better treatments for pulmonary fibrosis, including those that prevent or delay onset of symptomatic disease, a key priority in the field.
Supplementary Material
This article has an online data supplement, which is accessible at the Supplements Tab. Artificial Intelligence Disclaimer: No artificial intelligence tools were used in writing this manuscript.
At a Glance Commentary:
Scientific Knowledge on the Subject: While individuals with idiopathic pulmonary fibrosis (IPF) experience progressive symptoms and lung function decline over time, the rate of decline across individuals is widely accepted to vary greatly.
What This Study Adds to the Field: We developed a Bayesian latent variable repeated measures model to describe the physiological disease progression of IPF over the entire disease development period. This model estimated a latent subject-specific alignment parameter which allowed characterization of a conserved sequential decline in four pulmonary function test parameters that occurs from very early (subclinical) through advanced disease. The better understanding of IPF progression will enable us to make evidence-based decisions regarding the design of future clinical trials, including selection of the most relevant endpoints.
ACKNOWLEDGEMENTS
This publication is based on research using data from data contributor Roche that has been made available through Vivli, Inc. Vivli has not contributed to or approved, and is not in any way responsible for, the contents of this publication. Dr. Quintana is an employee of Berry Consultants, LLC, a statistical consulting firm that specializes in the design, conduct, oversight, and analysis of adaptive and platform clinical trials. A subject-level dataset from the FPF cohort corresponding to the example statistical code provided in the Online Supplement will be available on dbGaP.
Funding Support:
NIH/NHLBI R56HL166941 (MLS), NIH/NHLBI R01HL179065 (MLS), NIH/NHLBI P01HL172729 (TSB), NIH/NHLBI P01HL092870 (TSB), NIH/NHLBI R01HL151016 (TSB), NIH/NHLBI R01HL175555 (TSB), NIH/NHLBI R01HL145372 (JAK), and Boehringer Ingelheim Pharmaceuticals, Inc.
Footnotes
Descriptor Number: 9.23 Interstitial Lung Disease
REFERENCES
- 1.Raghu G, Chen SY, Yeh WS, Maroni B, Li Q, Lee YC, Collard HR. Idiopathic pulmonary fibrosis in US Medicare beneficiaries aged 65 years and older: incidence, prevalence, and survival, 2001–11. Lancet Respir Med 2014; 2: 566–572. [DOI] [PubMed] [Google Scholar]
- 2.Kreuter M, Wuyts WA, Wijsenbeek M, Bajwah S, Maher TM, Stowasser S, Male N, Stansen W, Schoof N, Orsatti L, Swigris J. Health-related quality of life and symptoms in patients with IPF treated with nintedanib: analyses of patient-reported outcomes from the INPULSIS(R) trials. Respir Res 2020; 21: 36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Farrand E, Iribarren C, Vittinghoff E, Levine-Hall T, Ley B, Minowada G, Collard HR. Impact of Idiopathic Pulmonary Fibrosis on Longitudinal Health-care Utilization in a Community-Based Cohort of Patients. Chest 2021; 159: 219–227. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Raghu G, Remy-Jardin M, Richeldi L, Thomson CC, Inoue Y, Johkoh T, Kreuter M, Lynch DA, Maher TM, Martinez FJ, Molina-Molina M, Myers JL, Nicholson AG, Ryerson CJ, Strek ME, Troy LK, Wijsenbeek M, Mammen MJ, Hossain T, Bissell BD, Herman DD, Hon SM, Kheir F, Khor YH, Macrea M, Antoniou KM, Bouros D, Buendia-Roldan I, Caro F, Crestani B, Ho L, Morisset J, Olson AL, Podolanczuk A, Poletti V, Selman M, Ewing T, Jones S, Knight SL, Ghazipura M, Wilson KC. Idiopathic Pulmonary Fibrosis (an Update) and Progressive Pulmonary Fibrosis in Adults: An Official ATS/ERS/JRS/ALAT Clinical Practice Guideline. Am J Respir Crit Care Med 2022; 205: e18–e47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.King TE Jr., Bradford WZ, Castro-Bernardini S, Fagan EA, Glaspole I, Glassberg MK, Gorina E, Hopkins PM, Kardatzke D, Lancaster L, Lederer DJ, Nathan SD, Pereira CA, Sahn SA, Sussman R, Swigris JJ, Noble PW, Group AS. A phase 3 trial of pirfenidone in patients with idiopathic pulmonary fibrosis. N Engl J Med 2014; 370: 2083–2092. [DOI] [PubMed] [Google Scholar]
- 6.Richeldi L, du Bois RM, Raghu G, Azuma A, Brown KK, Costabel U, Cottin V, Flaherty KR, Hansell DM, Inoue Y, Kim DS, Kolb M, Nicholson AG, Noble PW, Selman M, Taniguchi H, Brun M, Le Maulf F, Girard M, Stowasser S, Schlenker-Herceg R, Disse B, Collard HR, Investigators IT. Efficacy and safety of nintedanib in idiopathic pulmonary fibrosis. N Engl J Med 2014; 370: 2071–2082. [DOI] [PubMed] [Google Scholar]
- 7.Richeldi L, Azuma A, Cottin V, Kreuter M, Maher TM, Martinez FJ, Oldham JM, Valenzuela C, Clerisme-Beaty E, Gordat M, Wachtlin D, Liu Y, Schlecker C, Stowasser S, Zoz DF, Wijsenbeek MS, Investigators F-IT. Nerandomilast in Patients with Idiopathic Pulmonary Fibrosis. N Engl J Med 2025; 392: 2193–2202. [DOI] [PubMed] [Google Scholar]
- 8.Ley B, Collard HR, King TE Jr. Clinical course and prediction of survival in idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2011; 183: 431–440. [DOI] [PubMed] [Google Scholar]
- 9.Fainberg HP, Oldham JM, Molyneau PL, Allen RJ, Kraven LM, Fahy WA, Porte J, Braybrooke R, Saini G, Karsdal MA, Leeming DJ, Sand JMB, Triguero I, Oballa E, Wells AU, Renzoni E, Wain LV, Noth I, Maher TM, Stewart ID, Jenkins RG. Forced vital capacity trajectories in patients with idiopathic pulmonary fibrosis: a secondary analysis of a multicentre, prospective, observational cohort. Lancet Digit Health 2022; 4: e862–e872. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Selman M, Carrillo G, Estrada A, Mejia M, Becerril C, Cisneros J, Gaxiola M, Perez-Padilla R, Navarro C, Richards T, Dauber J, King TE Jr., Pardo A, Kaminski N. Accelerated variant of idiopathic pulmonary fibrosis: clinical behavior and gene expression pattern. PLoS One 2007; 2: e482. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hunninghake GM, Hatabu H, Okajima Y, Gao W, Dupuis J, Latourelle JC, Nishino M, Araki T, Zazueta OE, Kurugol S, Ross JC, San Jose Estepar R, Murphy E, Steele MP, Loyd JE, Schwarz MI, Fingerlin TE, Rosas IO, Washko GR, O’Connor GT, Schwartz DA. MUC5B promoter polymorphism and interstitial lung abnormalities. N Engl J Med 2013; 368: 2192–2200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Rosas IO, Ren P, Avila NA, Chow CK, Franks TJ, Travis WD, McCoy JP Jr., May RM, Wu HP, Nguyen DM, Arcos-Burgos M, MacDonald SD, Gochuico BR. Early interstitial lung disease in familial pulmonary fibrosis. Am J Respir Crit Care Med 2007; 176: 698–705. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Salisbury ML, Hewlett JC, Ding G, Markin CR, Douglas K, Mason W, Guttentag A, Phillips JA 3rd, Cogan JD, Reiss S, Mitchell DB, Wu P, Young LR, Lancaster LH, Loyd JE, Humphries SM, Lynch DA, Kropski JA, Blackwell TS. Development and Progression of Radiologic Abnormalities in Individuals at Risk for Familial Interstitial Lung Disease. Am J Respir Crit Care Med 2020; 201: 1230–1239. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hatabu H, Hunninghake GM, Richeldi L, Brown KK, Wells AU, Remy-Jardin M, Verschakelen J, Nicholson AG, Beasley MB, Christiani DC, San Jose Estepar R, Seo JB, Johkoh T, Sverzellati N, Ryerson CJ, Graham Barr R, Goo JM, Austin JHM, Powell CA, Lee KS, Inoue Y, Lynch DA. Interstitial lung abnormalities detected incidentally on CT: a Position Paper from the Fleischner Society. Lancet Respir Med 2020; 8: 726–737. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Podolanczuk AJ, Hunninghake GM, Wilson KC, Khor YH, Kheir F, Pang B, Adegunsoye A, Cararie G, Corte TJ, Flanagan J, Gudmundsson G, Hariri LP, Hatabu H, Humphries SM, Kaul B, Kim JS, Konigshoff M, Kropski JA, Lee JS, Luo F, Lynch DA, Martinez FJ, Montesi SB, Moodley Y, Oldham JM, Piciucchi S, Putman RK, Richeldi L, Rosas IO, Salisbury ML, Salvatore MM, Selman M, Seo JB, Song JW, Thomson CC, Vivero M, Wain LV, Wijsenbeek M, Schwartz DA, Ryerson CJ. Approach to the Evaluation and Management of Interstitial Lung Abnormalities: An Official American Thoracic Society Clinical Statement. Am J Respir Crit Care Med 2025; 211: 1132–1155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kropski JA, Pritchett JM, Zoz DF, Crossno PF, Markin C, Garnett ET, Degryse AL, Mitchell DB, Polosukhin VV, Rickman OB, Choi L, Cheng DS, McConaha ME, Jones BR, Gleaves LA, McMahon FB, Worrell JA, Solus JF, Ware LB, Lee JW, Massion PP, Zaynagetdinov R, White ES, Kurtis JD, Johnson JE, Groshong SD, Lancaster LH, Young LR, Steele MP, Phillips Iii JA, Cogan JD, Loyd JE, Lawson WE, Blackwell TS. Extensive phenotyping of individuals at risk for familial interstitial pneumonia reveals clues to the pathogenesis of interstitial lung disease. Am J Respir Crit Care Med 2015; 191: 417–426. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Araki T, Putman RK, Hatabu H, Gao W, Dupuis J, Latourelle JC, Nishino M, Zazueta OE, Kurugol S, Ross JC, San Jose Estepar R, Schwartz DA, Rosas IO, Washko GR, O’Connor GT, Hunninghake GM. Development and Progression of Interstitial Lung Abnormalities in the Framingham Heart Study. Am J Respir Crit Care Med 2016; 194: 1514–1522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Putman RK, Hatabu H, Araki T, Gudmundsson G, Gao W, Nishino M, Okajima Y, Dupuis J, Latourelle JC, Cho MH, El-Chemaly S, Coxson HO, Celli BR, Fernandez IE, Zazueta OE, Ross JC, Harmouche R, Estepar RS, Diaz AA, Sigurdsson S, Gudmundsson EF, Eiriksdottir G, Aspelund T, Budoff MJ, Kinney GL, Hokanson JE, Williams MC, Murchison JT, MacNee W, Hoffmann U, O’Donnell CJ, Launer LJ, Harrris TB, Gudnason V, Silverman EK, O’Connor GT, Washko GR, Rosas IO, Hunninghake GM, Evaluation of CLtIPSEI, Investigators CO. Association Between Interstitial Lung Abnormalities and All-Cause Mortality. JAMA 2016; 315: 672–681. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Podolanczuk AJ, Oelsner EC, Barr RG, Bernstein EJ, Hoffman EA, Easthausen IJ, Stukovsky KH, RoyChoudhury A, Michos ED, Raghu G, Kawut SM, Lederer DJ. High-Attenuation Areas on Chest Computed Tomography and Clinical Respiratory Outcomes in Community-Dwelling Adults. Am J Respir Crit Care Med 2017; 196: 1434–1442. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Salisbury ML, Markin C, Fadely T, Guttentag AR, Humphries SM, Lynch DA, Kropski JA, Blackwell TS. Progressive Early Interstitial Lung Abnormalities in Persons at Risk for Familial Pulmonary Fibrosis: A Prospective Cohort Study. Am J Respir Crit Care Med 2024; 210: 1441–1452. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Wang G, Berry S, Xiong C, Hassenstab J, Quintana M, McDade EM, Delmar P, Vestrucci M, Sethuraman G, Bateman RJ, Dominantly Inherited Alzheimer Network Trials U. A novel cognitive disease progression model for clinical trials in autosomal-dominant Alzheimer’s disease. Stat Med 2018; 37: 3047–3055. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Bateman RJ, Xiong C, Benzinger TL, Fagan AM, Goate A, Fox NC, Marcus DS, Cairns NJ, Xie X, Blazey TM, Holtzman DM, Santacruz A, Buckles V, Oliver A, Moulder K, Aisen PS, Ghetti B, Klunk WE, McDade E, Martins RN, Masters CL, Mayeux R, Ringman JM, Rossor MN, Schofield PR, Sperling RA, Salloway S, Morris JC, Dominantly Inherited Alzheimer N. Clinical and biomarker changes in dominantly inherited Alzheimer’s disease. N Engl J Med 2012; 367: 795–804. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Noble PW, Albera C, Bradford WZ, Costabel U, Glassberg MK, Kardatzke D, King TE Jr., Lancaster L, Sahn SA, Szwarcberg J, Valeyre D, du Bois RM, Group CS. Pirfenidone in patients with idiopathic pulmonary fibrosis (CAPACITY): two randomised trials. Lancet 2011; 377: 1760–1769. [DOI] [PubMed] [Google Scholar]
- 24.Salisbury ML, Huang X, Markin C, Fadely T, Guttentag AR, Kropski JA, Wu P, Blackwell TS. Disease Progression Modeling Identifies a Conserved Pattern of Lung Function Decline in Familial Pulmonary Fibrosis. Am J Respir Crit Care Med 2025; 211: A3236. [Google Scholar]
- 25.Liu Q, Zhou Y, Cogan JD, Mitchell DB, Sheng Q, Zhao S, Bai Y, Ciombor KK, Sabusap CM, Malabanan MM, Markin CR, Douglas K, Ding G, Banovich NE, Nickerson DA, Blue EE, Bamshad MJ, Brown KK, Schwartz DA, Phillips JA 3rd, Martinez-Barricarte R, Salisbury ML, Shyr Y, Loyd JE, Kropski JA, Blackwell TS. The Genetic Landscape of Familial Pulmonary Fibrosis. Am J Respir Crit Care Med 2023; 207: 1345–1357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Bowerman C, Bhakta NR, Brazzale D, Cooper BR, Cooper J, Gochicoa-Rangel L, Haynes J, Kaminsky DA, Lan LTT, Masekela R, McCormack MC, Steenbruggen I, Stanojevic S. A Race-neutral Approach to the Interpretation of Lung Function Measurements. Am J Respir Crit Care Med 2023; 207: 768–774. [DOI] [PubMed] [Google Scholar]
- 27.Hall GL, Filipow N, Ruppel G, Okitika T, Thompson B, Kirkby J, Steenbruggen I, Cooper BG, Stanojevic S, contributing GLINm. Official ERS technical standard: Global Lung Function Initiative reference values for static lung volumes in individuals of European ancestry. Eur Respir J 2021; 57. [DOI] [PubMed] [Google Scholar]
- 28.Stanojevic S, Graham BL, Cooper BG, Thompson BR, Carter KW, Francis RW, Hall GL, Global Lung Function Initiative Twg, Global Lung Function Initiative T. Official ERS technical standards: Global Lung Function Initiative reference values for the carbon monoxide transfer factor for Caucasians. Eur Respir J 2017; 50. [DOI] [PubMed] [Google Scholar]
- 29.“Official ERS technical standards: Global Lung Function Initiative reference values for the carbon monoxide transfer factor for Caucasians.” Stanojevic Sanja, Graham Brian L., Cooper Brendan G., Thompson Bruce R., Carter Kim W., Francis Richard W. and Graham L. Hall on behalf of the Global Lung Function Initiative T (LCO) working group. Eur Respir J 2017; 50: 1700010. [DOI] [PubMed] [Google Scholar]; Eur Respir J 2020; 56. [Google Scholar]
- 30.Quintana M, Shrader J, Slota C, Joe G, McKew JC, Fitzgerald M, Gahl WA, Berry S, Carrillo N. Bayesian model of disease progression in GNE myopathy. Stat Med 2019; 38: 1459–1474. [DOI] [PubMed] [Google Scholar]
- 31.Collard HR, King TE Jr., Bartelson BB, Vourlekis JS, Schwarz MI, Brown KK. Changes in clinical and physiologic variables predict survival in idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2003; 168: 538–542. [DOI] [PubMed] [Google Scholar]
- 32.Flaherty KR, Andrei AC, Murray S, Fraley C, Colby TV, Travis WD, Lama V, Kazerooni EA, Gross BH, Toews GB, Martinez FJ. Idiopathic pulmonary fibrosis: prognostic value of changes in physiology and six-minute-walk test. Am J Respir Crit Care Med 2006; 174: 803–809. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Salisbury ML, Xia M, Zhou Y, Murray S, Tayob N, Brown KK, Wells AU, Schmidt SL, Martinez FJ, Flaherty KR. Idiopathic Pulmonary Fibrosis: Gender-Age-Physiology Index Stage for Predicting Future Lung Function Decline. Chest 2016; 149: 491–498. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Oldham JM, Neely ML, Wojdyla DM, Gulati M, Li P, Patel DC, Palmer SM, Todd JL, Investigators I-PR. Changes in Lung Function and Mortality Risk in Patients With Idiopathic Pulmonary Fibrosis. Chest 2025. [DOI] [PubMed] [Google Scholar]
- 35.Collard HR, Ryerson CJ, Corte TJ, Jenkins G, Kondoh Y, Lederer DJ, Lee JS, Maher TM, Wells AU, Antoniou KM, Behr J, Brown KK, Cottin V, Flaherty KR, Fukuoka J, Hansell DM, Johkoh T, Kaminski N, Kim DS, Kolb M, Lynch DA, Myers JL, Raghu G, Richeldi L, Taniguchi H, Martinez FJ. Acute Exacerbation of Idiopathic Pulmonary Fibrosis. An International Working Group Report. Am J Respir Crit Care Med 2016; 194: 265–275. [DOI] [PubMed] [Google Scholar]
- 36.Allen RJ, Stockwell A, Oldham JM, Guillen-Guio B, Schwartz DA, Maher TM, Flores C, Noth I, Yaspan BL, Jenkins RG, Wain LV, International IPFGC. Genome-wide association study across five cohorts identifies five novel loci associated with idiopathic pulmonary fibrosis. Thorax 2022; 77: 829–833. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Moll M, Peljto AL, Kim JS, Xu H, Debban CL, Chen X, Menon A, Putman RK, Ghosh AJ, Saferali A, Nishino M, Hatabu H, Hobbs BD, Hecker J, McDermott G, Sparks JA, Wain LV, Allen RJ, Tobin MD, Raby BA, Chun S, Silverman EK, Zamora AC, Ortega VE, Garcia CK, Barr RG, Bleecker ER, Meyers DA, Kaner RJ, Rich SS, Manichaikul A, Rotter JI, Dupuis J, O’Connor GT, Fingerlin TE, Hunninghake GM, Schwartz DA, Cho MH. A Polygenic Risk Score for Idiopathic Pulmonary Fibrosis and Interstitial Lung Abnormalities. Am J Respir Crit Care Med 2023; 208: 791–801. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Flaherty KR, Fell C, Aubry MC, Brown K, Colby T, Costabel U, Franks TJ, Gross BH, Hansell DM, Kazerooni E, Kim DS, King TE Jr., Kitachi M, Lynch D, Myers J, Nagai S, Nicholson AG, Poletti V, Raghu G, Selman M, Toews G, Travis W, Wells AU, Vassallo R, Martinez FJ. Smoking-related idiopathic interstitial pneumonia. Eur Respir J 2014; 44: 594–602. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Zhao A, Gudmundsson E, Mogulkoc N, van Moorsel C, Corte TJ, Vasudev P, Romei C, Chapman R, Wallis TJM, Denneny E, Goos T, Savas R, Ahmed A, Brereton CJ, van Es HW, Jo H, De Liperi A, Duncan M, Pontoppidan K, De Sadeleer LJ, van Beek F, Barnett J, Cross G, Procter A, Veltkamp M, Hopkins P, Moodley Y, Taliani A, Taylor M, Verleden S, Tavanti L, Vermant M, Nair A, Stewart I, Janes SM, Young AL, Barber D, Alexander DC, Porter JC, Wells AU, Jones MG, Wuyts WA, Jacob J. Mortality surrogates in combined pulmonary fibrosis and emphysema. Eur Respir J 2024; 63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Kim JS, Montesi SB, Adegunsoye A, Humphries SM, Salisbury ML, Hariri LP, Kropski JA, Richeldi L, Wells AU, Walsh S, Jenkins RG, Rosas I, Noth I, Hunninghake GM, Martinez FJ, Podolanczuk AJ. Approach to Clinical Trials for the Prevention of Pulmonary Fibrosis. Ann Am Thorac Soc 2023; 20: 1683–1693. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Montesi SB, Gomez CR, Beers M, Brown R, Chattopadhyay I, Flaherty KR, Garcia CK, Gomperts B, Hariri LP, Hogaboam CM, Jenkins RG, Kaminski N, Kim GHJ, Konigshoff M, Kolb M, Kotton DN, Kropski JA, Lasky J, Magin CM, Maher TM, McCormick M, Moore BB, Nickerson-Nutter C, Oldham J, Podolanczuk AJ, Raghu G, Rosas I, Rowe SM, Schmidt WT, Schwartz D, Shore JE, Spino C, Craig JM, Martinez FJ. Pulmonary Fibrosis Stakeholder Summit: A Joint NHLBI, Three Lakes Foundation, and Pulmonary Fibrosis Foundation Workshop Report. Am J Respir Crit Care Med 2024; 209: 362–373. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Raghu G, Ghazipura M, Fleming TR, Aronson KI, Behr J, Brown KK, Flaherty KR, Kazerooni EA, Maher TM, Richeldi L, Lasky JA, Swigris JJ, Busch R, Garrard L, Ahn DH, Li J, Puthawala K, Rodal G, Seymour S, Weir N, Danoff SK, Ettinger N, Goldin J, Glassberg MK, Kawano-Dourado L, Khalil N, Lancaster L, Lynch DA, Mageto Y, Noth I, Shore JE, Wijsenbeek M, Brown R, Grogan D, Ivey D, Golinska P, Karimi-Shah B, Martinez FJ. Meaningful Endpoints for Idiopathic Pulmonary Fibrosis (IPF) Clinical Trials: Emphasis on ‘Feels, Functions, Survives’. Report of a Collaborative Discussion in a Symposium with Direct Engagement from Representatives of Patients, Investigators, the National Institutes of Health, a Patient Advocacy Organization, and a Regulatory Agency. Am J Respir Crit Care Med 2024; 209: 647–669. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Jacob J, Bartholmai BJ, Rajagopalan S, van Moorsel CHM, van Es HW, van Beek FT, Struik MHL, Kokosi M, Egashira R, Brun AL, Nair A, Walsh SLF, Cross G, Barnett J, de Lauretis A, Judge EP, Desai S, Karwoski R, Ourselin S, Renzoni E, Maher TM, Altmann A, Wells AU. Predicting Outcomes in Idiopathic Pulmonary Fibrosis Using Automated Computed Tomographic Analysis. Am J Respir Crit Care Med 2018; 198: 767–776. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Thillai M, Oldham JM, Ruggiero A, Kanavati F, McLellan T, Saini G, Johnson SR, Ble FX, Azim A, Ostridge K, Platt A, Belvisi M, Maher TM, Molyneaux PL. Deep Learning-based Segmentation of Computed Tomography Scans Predicts Disease Progression and Mortality in Idiopathic Pulmonary Fibrosis. Am J Respir Crit Care Med 2024; 210: 465–472. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Bowman WS, Newton CA, Linderholm AL, Neely ML, Pugashetti JV, Kaul B, Vo V, Echt GA, Leon W, Shah RJ, Huang Y, Garcia CK, Wolters PJ, Oldham JM. Proteomic biomarkers of progressive fibrosing interstitial lung disease: a multicentre cohort analysis. Lancet Respir Med 2022; 10: 593–602. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Yadav H, Scanlon PD, Aksamit TR, Zhang Z, Pennington KM, Lim KG, Niven AS. Intersession Variability in Single-Breath Diffusing Capacity of Carbon Monoxide. Chest 2025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Ho JE, Gao W, Levy D, Santhanakrishnan R, Araki T, Rosas IO, Hatabu H, Latourelle JC, Nishino M, Dupuis J, Washko GR, O’Connor GT, Hunninghake GM. Galectin-3 Is Associated with Restrictive Lung Disease and Interstitial Lung Abnormalities. Am J Respir Crit Care Med 2016; 194: 77–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Armstrong HF, Podolanczuk AJ, Barr RG, Oelsner EC, Kawut SM, Hoffman EA, Tracy R, Kaminski N, McClelland RL, Lederer DJ. Serum Matrix Metalloproteinase-7, Respiratory Symptoms, and Mortality in Community-Dwelling Adults. MESA (Multi-Ethnic Study of Atherosclerosis). Am J Respir Crit Care Med 2017; 196: 1311–1317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Misharin AV, Morales-Nebreda L, Reyfman PA, Cuda CM, Walter JM, McQuattie-Pimentel AC, Chen CI, Anekalla KR, Joshi N, Williams KJN, Abdala-Valencia H, Yacoub TJ, Chi M, Chiu S, Gonzalez-Gonzalez FJ, Gates K, Lam AP, Nicholson TT, Homan PJ, Soberanes S, Dominguez S, Morgan VK, Saber R, Shaffer A, Hinchcliff M, Marshall SA, Bharat A, Berdnikovs S, Bhorade SM, Bartom ET, Morimoto RI, Balch WE, Sznajder JI, Chandel NS, Mutlu GM, Jain M, Gottardi CJ, Singer BD, Ridge KM, Bagheri N, Shilatifard A, Budinger GRS, Perlman H. Monocyte-derived alveolar macrophages drive lung fibrosis and persist in the lung over the life span. J Exp Med 2017; 214: 2387–2404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Habermann AC, Gutierrez AJ, Bui LT, Yahn SL, Winters NI, Calvi CL, Peter L, Chung MI, Taylor CJ, Jetter C, Raju L, Roberson J, Ding G, Wood L, Sucre JMS, Richmond BW, Serezani AP, McDonnell WJ, Mallal SB, Bacchetta MJ, Loyd JE, Shaver CM, Ware LB, Bremner R, Walia R, Blackwell TS, Banovich NE, Kropski JA. Single-cell RNA sequencing reveals profibrotic roles of distinct epithelial and mesenchymal lineages in pulmonary fibrosis. Sci Adv 2020; 6: eaba1972. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.