Abstract
Background Idiopathic pulmonary fibrosis (IPF) is a progressive lung disease in which circulatory biomarkers have the potential for guiding management in clinical practice. We assessed the prognostic role of serum biomarkers in three independent IPF cohorts: Australian Idiopathic Pulmonary Fibrosis Registry (AIPFR), Trent Lung Fibrosis (TLF) and Prospective Observation of Fibrosis in the Lung Clinical Endpoints (PROFILE).
Methods In the AIPFR cohort, candidate proteins were assessed by ELISA as well as in an unbiased proteomic approach. LASSO (least absolute shrinkage and selection operator) regression was used to restrict the selection of markers that best accounted for the progressor phenotype at 1 year in the AIPFR cohort, and subsequently prospectively selected for replication in the validation TLF cohort and assessed retrospectively in the PROFILE cohort. Four significantly replicating biomarkers were aggregated into a progression index model based on tertiles of circulating concentrations.
Results 189 participants were included in the AIPFR cohort, 205 participants from the TLF cohort and 122 participants from the PROFILE cohort. Differential biomarker expression was observed by ELISA and replicated for osteopontin, matrix metallopeptidase-7, intercellular adhesion molecule-1 and periostin for those with a progressor phenotype at 1 year. Proteomic data did not replicate. The progression index in the AIPFR, TLF and PROFILE cohorts predicted risk of progression, mortality and progression-free survival. A statistical model incorporating the progression index demonstrated the capacity to distinguish disease progression at 12 months, which was increased beyond the clinical GAP (gender, age and physiology) score model alone in all cohorts, and significantly so within the incidence-based TLF and PROFILE cohorts.
Conclusion A panel of circulatory biomarkers can provide potentially valuable clinical assistance in the prognosis of IPF patients.
Abstract
Pathobiologically relevant circulatory biomarkers were found to be associated with IPF progression and mortality, and a statistical model incorporating these markers into a progression index score showed improved prognostication across all outcomes https://bit.ly/37L0oMl
Introduction
Idiopathic pulmonary fibrosis (IPF) is a progressive fibrotic disease of unknown aetiology, with a median survival of ∼2–5 years. While the overall prognosis is poor, the disease course and rate of progression are highly variable [1, 2]. The ability to stratify patients based on their predicted disease course has potential to inform management decisions, including the most appropriate time to commence antifibrotic treatment, potential transplant referral and planning end-of-life care. Identifying patients who are more likely to progress within a short time frame could also assist in stratifying clinical trial enrolment, allowing enrichment of trial populations with patients at greatest risk of decline.
There have been multiple studies in IPF aiming to identify reliable prognostic markers based on demographic data (age, gender and smoking status) [3, 4], clinical and physiological parameters (dyspnoea scores, and baseline and serial lung function) [5–7], as well as specific radiological features on high-resolution computed tomography (HRCT) scanning [8, 9]. While many of these parameters have been able to predict progressive disease either alone or in combination, detectable differences only occur after significant lung damage has occurred. Therefore, there is a need to identify molecules reflecting early underlying cellular and tissue damage; this led us to investigate blood biomarkers as predictors of disease progression and mortality. Many of these molecules have been associated with the pathogenesis of IPF as demonstrated by findings in bronchoalveolar lavage fluid and lung tissue studies [10–12]. Peripheral blood biomarkers are clinically more appealing than those obtained via other invasive procedures. Currently, there is a paucity of peripheral blood protein biomarkers that have been sufficiently replicated across different populations to be informative in clinical practice, even in conjunction with clinical/radiological parameters.
Using three international and well-characterised IPF populations, i.e. the Australian Idiopathic Pulmonary Fibrosis Registry (AIPFR) [13] as the primary cohort, and the Trent Lung Fibrosis (TLF) study [14] and the Prospective Observation of Fibrosis in the Lung Clinical Endpoints (PROFILE) cohort [11] for validation, we sought to assess biomarker profiles of patients with progressive versus stable disease, as markers indicating increased risk of progression at 1 year. We hypothesised that a panel of biomarkers could improve prediction of mortality and disease progression above current clinical scores. We used two methods of biomarker discovery: 1) a hypothesis-driven approach based on a selection of 12 molecules with known pathogenic roles in IPF, as well as 2) an unbiased proteomic approach to screen a large number of candidate biomarkers. After assessment of all these findings in the TLF validation cohort, we were able to formulate a progression index based on tertile levels of four biomarkers that sustained a signal which identified a population at risk of progression and then re-applied this index to test whether the addition of this progression index led to improved prognostic utility above that of the currently used clinical GAP (gender, age and physiology) score model.
Materials and methods
Study participants
The AIPFR is a national multicentre, prospective registry of IPF patients across Australia, collating comprehensive longitudinal data paired with a biobank of plasma and serum [13]. IPF diagnoses from physicians around Australia were centrally reviewed according to the 2011 American Thoracic Society/European Respiratory Society/Japanese Respiratory Society/Latin American Thoracic Association IPF guidelines for patients recruited to the AIPFR [15]. The TLF cohort recruited IPF patients from hospitals across England and Wales diagnosed by thoracic radiologists following HRCT review [14]. Blood samples from both the AIPFR and TLF cohorts were analysed at the same centre for both ELISA and proteomics (Harry Perkins Institute for Medical Research, Perth, Australia).
PROFILE is a multicentre, prospective cohort study of incident cases of IPF from secondary and tertiary centres in the UK, diagnosed by multidisciplinary discussion according to current diagnostic criteria (ClinicalTrials.gov: NCT01134822) [11]. The retrospective dataset generated from the PROFILE cohort was used as an additional replication cohort for the data generated by AIPFR and TLF. In the PROFILE cohort, the participants were missing biomarker values for POSTN (see supplementary table S1 for details of biomarkers), which impacted on subsequent score aggregation.
This study was approved by the Royal Perth Hospital Ethics Committee (HREC/2011-138) and the Sydney Local Health Network (HREC/15/RPAH/28). Ethical approval for the TLF study was granted by the Nottingham Research Ethics Committee (REC 09/H0403/59), while ethical approval for the PROFILE study was granted by the Royal Free Hospital Research Ethics Committee (REC 10/H0720/12) and PROFILE (Central England) Northampton Research Ethics Committee (REC 10/H0402/2). No patients were on pirfenidone or nintedanib at the time of blood collection.
Definitions
Our primary outcome of interest was progression status at 1 year, with date of consent being “time zero” across all three cohorts. In the AIPFR and TLF cohorts, disease stability was assessed in the 12 months following blood collection. “Progressive disease” was defined as a fall in forced vital capacity (FVC) ≥10% and/or diffusing capacity of the lung for carbon monoxide (DLCO) ≥15% and/or death within 12 months from the time of blood collection, while “stable disease” was defined as the absence of progression over the same time frame. Classification and analysis of progression at 1 year was restricted to those with biomarker data and at least two lung function tests available.
Secondary outcomes included mortality and progression-free survival (PFS). Time to mortality was defined as the time from baseline (consent) to date of death. PFS was defined as the time to death or progression from consent.
Biomarker identification
ELISA analysis
12 biomarkers were selected on the basis of previous studies and measured by ELISA in serum using commercial kits according to the manufacturers' instructions. A list of all ELISA markers can be found in supplementary table S1.
Proteomic analysis
Plasma proteomics was performed in two stages as previously described [16, 17]. In brief, isobaric tags for relative and absolute quantitation (iTRAQ) and multiple reaction monitoring (MRM) were carried out to identify differential expression of circulating proteins in patients with stable and progressive IPF. First, iTRAQ was performed on pooled plasma samples from stable and progressive patients. Peptides that were differentially expressed were then quantified and validated using targeted mass spectrometry with MRM for all individual patient samples.
Protein identification and quantification were performed using ProteinPilot 4.5 beta software (AB Sciex, Macclesfield, UK) and spectra searched against the human SwissProt database as previously described [16]. A list of all proteomic metabolites can be found in supplementary table S1.
Statistical analyses
Identification of candidate prognostic biomarkers
The AIPFR cohort was used for initial exploratory analyses. The ELISA panel and proteomics were assessed in this cohort for candidate biomarkers predictive of clinical outcomes. Specifically, LASSO (least absolute shrinkage and selection operator) was used to shrink the selection to those that best account for this effect in AIPFR and key biomarkers associated with progression at 1 year were selected.
Univariable logistic regression and Cox regression analyses were performed for individual biomarkers to determine their relationship with progression status at 1 year, mortality and PFS outcomes. All univariable results are included in the supplementary material only. Multivariable adjustments for demographic and physiological parameters (age, gender, smoking status and baseline FVC) were also performed for each biomarker relative to each outcome. The false discovery rate (FDR) for all sets of analyses was controlled to 5% using Hochberg's method.
The ability of each biomarker to predict progression status was summarised using the area under the receiver operating characteristic curve (AUC) and 95% confidence intervals. Multivariable regression models were also estimated with all biomarkers (separately for ELISA and proteomics) using the LASSO estimation procedure to control for overfitting.
Development and validation of a progression index
The TLF cohort was selected as an independent cohort for the replication of all biomarker analyses obtained in the AIPFR cohort. A nonparametric Wilcoxon rank sum test was used to determine absolute differences in significant biomarkers between progressive and stable groups within this second cohort.
Associations were tested against progressor status as continuous data or as tertiles for ELISA panel concentration and mean peptide sequence protein ratio. Specifically, median values were compared between stable and progressors, and proteomic values were averaged across the contributing peptide sequences to estimate the concentration of protein. The biomarkers that proved significant were segregated into discrete tertiles based on circulating concentration levels and were aggregated into a progression index (0–8) from which they were assigned an index category (0–2, 3–5 or 6–8). The ability of each biomarker to predict progression status using the tertiles was summarised using the AUC and 95% confidence intervals, and compared relative to the GAP score. Cox proportional hazards models were used to test associations with time to mortality and PFS, and proportional hazards assumptions were confirmed. A p-value of 0.05 defined the significance threshold. All analyses were performed using Stata version 16.0 SE (StataCorp, College Station, TX, USA).
All progression index models were adjusted for age, gender, baseline FVC and body mass index (BMI). Smoking data were not collected in the detail needed as part of the TLF study and therefore we did not include smoking in the multivariable models where comparisons were made across studies, in order to keep adjustments consistent. Unadjusted estimates of the progression index with disease progression, adjusted estimates as reported (FVC % pred, age, gender and BMI) and adjusted sensitivity estimates for DLCO (DLCO, age, gender and BMI) are included in the supplementary material only, to demonstrate comparable findings across TLF, AIPFR and pooled in multilevel analysis (supplementary figure S1).
Results
Patient demographics
Of the participants enrolled in the AIPFR at the time of this analysis, 189 had blood available for the investigation: 136 (72%) with stable disease and 53 (28%) with progressive disease. Demographic characteristics of this Australian cohort demonstrated predominantly males (n=136 (72%)), mean±sd age 69±8 years, FVC 82±19% predicted and DLCO 49±15% predicted (table 1 and supplementary figure S2).
The TLF cohort comprised 211 IPF cases. Six individuals did not have serum available and so 205 individuals were included: 138 with stable disease (67%) and 67 with progressive disease (33%). Demographic characteristics demonstrated predominantly males (n=152 (74%)), mean±sd age 73±9 years, FVC 85±19% predicted and DLCO 44±16% predicted (table 1 and supplementary figure S2).
The PROFILE cohort comprised 122 IPF cases with relevant biomarker data available: 68 with stable disease (56%) and 54 with progressive disease (44%). Demographic characteristics demonstrated predominantly males (n=96 (79%)), mean±sd age 71±8 years, FVC 81±19% predicted and DLCO 45±15% predicted (table 1 and supplementary figure S2).
In the PROFILE cohort, five out of 122 had a record of antifibrotics (pirfenidone) at baseline (4.1%) and 14 out of 122 were prescribed during the course of 1 year follow-up (11.5%) with a median (interquartile range (IQR)) time to starting of 179 (132–201) days, with a further five prescribed after 1 year. Overall, for the 19 starting antifibrotics, the median (IQR) time to starting was 192 (161–367) days. There was no evidence of antifibrotics in concomitant medications in the AIPFR and TLF cohorts.
Identification of candidate prognostic biomarkers in the AIPFR cohort
To determine the association of biomarkers with clinical outcomes, empirical estimates of the AUC were made for each biomarker and those with AUC >0.6 were identified (supplementary table S2). Due to the large number of potential predictors, LASSO with 10-fold cross-validation was utilised to select the subset of variables that best predicted progression at 1 year. A total of 11 biomarkers were shrunk to zero <5 times, including six ELISA biomarkers (ENRAGE, ICAM1, OPN, POSTN, SPD and VCAM1) and five proteomic biomarkers (A2GL, APOE, GELS, PEDF and SAA4) (supplementary table S2).
For prediction of PFS, the biomarkers ICAM1 (p=0.007), OPN (p=0.0002) and SPD (p=0.0003) were associated with worse outcomes on multivariable Cox analysis, following Hochberg FDR adjustment (table 2), while an increase in mortality by multivariable Cox regression was observed with significant differences in OPN (p<0.0001) and POSTN (p=0.030).
TLF cohort
The TLF cohort was selected to validate biomarker analyses from the AIPFR cohort. Notably, in the TLF cohort, several of the ELISA biomarkers identified in the AIPFR cohort demonstrated a significant difference between the progressive and stable groups, including OPN (p=0.0080), MMP7 (p=0.0015), ICAM1 (p=0.0001) and POSTN (p=0.0001), but none of the proteomics markers (table 3).
We tested clinical outcomes using composite scores from the four replicating biomarkers confirmed as significant in the TLF cohort. The distributions of values for these four validated biomarkers were used to define tertiles (table 4 and figure 1), with assignment of an index value for low (0), medium (1) and high (2) tertiles. Estimates of the AUC were made for each biomarker tertile showing further discrimination of progressor status (supplementary table S4). Univariable analysis of tertile 3 relative to tertile 1 identified individual biomarkers (OPN, MMP7, ICAM1 and POSTN) as significantly associated with disease progression at 12 months, with OR 3.13 (95% CI 1.42–6.90) for OPN to OR 4.57 (95% CI 2.12–9.87) for POSTN (supplementary table S5).
PROFILE cohort
Analysis in the retrospective PROFILE dataset of the significant ELISA biomarkers (OPN, MMP7, ICAM1 and POSTN) showed significant differences in two of the ELISA biomarkers (MMP7 and ICAM1) between progressive and stable groups (p=0.033 and p=0.035, respectively) (supplementary table S6).
Development and validation of a progression index
These four candidate ELISA biomarkers (i.e. OPN (osteopontin), MMP7 (matrix metallopeptidase-7), ICAM1 (intercellular adhesion molecule-1) and POSTN (periostin)) were used to develop a clinically relevant progression index which was applied to all three cohorts. Thus, tertiles from the four biomarkers were aggregated into possible progression indexes from 0 to 8 based on the sum of tertiles, which were grouped into three clinically interpretable categories of sufficient sample size to test associations with clinical outcomes: 0–2 (26% of participants in TLF, 42% in AIPFR and 73% in PROFILE), 3–5 (49% in TLF, 53% in AIPFR and 27% in PROFILE) and 6–8 (25% in TLF, 5% in AIPFR and no participants in PROFILE) (supplementary table S7 and supplementary figure S3). In PROFILE, POSTN was not measured and was therefore missing for all participants, which unsurprisingly affected progression index frequencies when aggregated.
In the TLF cohort, progression index categories 3–5 and 6–8 were predictive of disease progression in adjusted analyses compared with people in the lowest score category (0–2) (OR 2.75, 95% CI 1.02–7.44; p=0.013 and OR 11.27, 95% CI 3.99–31.77; p<0.001, respectively). Relative to a score of 0–2, scores of 3–5 or 6–8 were associated with a 2.55-fold and 5.86-fold increase in the adjusted risk of mortality, respectively (hazard ratio (HR) 2.55, 95% CI 1.41–4.61; p=0.002 and HR 5.86, 95% CI 3.20–10.73; p<0.001, respectively). Significant association was also observed for scores of 6–8 regarding the outcome of PFS (HR 1.74, 95% CI 1.10–2.73; p=0.017) (table 5 and figure 2).
In the AIPFR cohort, a progression index score of 3–5 compared with 0–2 was predictive of disease progression in adjusted analyses (OR 2.30, 95% CI 1.04–5.09; p=0.04) and PFS (HR 1.61, 95% CI 1.07–2.42; p=0.023), and a progression index score of 6–8 was strongly predictive of PFS (HR 5.09, 95% CI 2.65–9.80; p<0.001) (table 5 and figure 2). Increasing progression index scores were not significantly associated with overall mortality in AIPFR, but trends were in the same direction but with wide confidence limits (score 3–5: HR 1.70, 95% CI 0.83–3.45; score 6–8: HR 2.92, 95% CI 0.37–22.91) (table 5 and figure 2).
In adjusted pooled analysis of TLF and AIPFR, progression index category 6–8 was significantly associated with a 54% increase in the odds of disease progression (OR 1.54, 95% CI 1.34–1.77; p<0.001), a 5.81-fold increase in the risk of mortality (HR 5.81, 95% CI 3.47–9.72; p<0.001) and a prediction of PFS (HR 2.29, 95% CI 1.60–3.27; p<0.001). Significant associations were also observed for progression index category 3–5 in disease progression (OR 1.17, 95% CI 1.05–1.29; p=0.003) and mortality (HR 2.23, 95% CI 1.42–3.51; p<0.001) (table 5 and figure 2).
In the PROFILE cohort, the range of scores was limited by the missing biomarker (POSTN) data, but progression index category 3–5 was still predictive of mortality relative to category 0–2 (HR 2.23, 95% CI 1.02–4.85; p=0.043) (supplementary table S8). However, a significant association with disease progression was not observed in adjusted analyses (OR 2.06, 95% CI 0.59–7.24).
Addition of the progression index to the GAP model
In TLF, the GAP score and the progression index were used to calculate the AUC for the prediction models to determine the independent and combined outcome prediction power (table 6). The GAP score is used clinically to stage IPF and predict mortality. In the TLF cohort, the GAP score predicted disease progression at 12 months with AUC 0.57 (95% CI 0.49–0.66), while the progression index gave AUC 0.73 (95% CI 0.66–0.80). When combined with the GAP score, the progression index improved the AUC compared with the GAP score alone (AUC 0.71, 95% CI 0.64–0.78; p<0.001).
For the AIPFR cohort, the AUC for disease progression at 12 months using the GAP score alone was weak (AUC 0.49, 95% CI 0.38–0.59), while it improved using the progression index alone, although not significantly (AUC 0.59, 95% CI 0.48–0.69). The combined GAP score and progression index did not improve this (AUC 0.55, 95% CI 0.45–0.66) (table 6). In the PROFILE cohort, the GAP score offered an AUC for disease progression at 12 months of 0.55 (95% CI 0.45–0.65), while the progression index alone gave AUC 0.62 (95% CI 0.57–0.70). Combination of the progression index with the GAP score significantly improved prediction capacity above GAP alone (AUC 0.60, 95% CI 0.50–0.70; p=0.049) (table 6).
Discussion
In this multicentre study, we have shown differential biomarker expression in patients with progressive IPF compared with stable disease. In the AIPFR cohort, we derived a predictive model based on differences in expression of OPN, SPD, ICAM1 and MMP7. In the replication TLF cohort, the ELISA biomarkers OPN, MMP7, ICAM1 and POSTN showed significant differences between progressive and stable IPF. Using these four meaningful ELISA biomarkers, a progression index was generated according to tertile thresholds in TLF, which were associated with progression at 1 year, mortality and PFS. Furthermore, when combined in a statistical model with the GAP score this progression index improved the clinical predictive model for the identification of IPF progression in two of the cohorts tested. Notably, we were able to replicate the majority of our findings across three large, international, longitudinal and well-characterised IPF disease cohorts.
Our panel of predictive biomarkers has been previously implicated in the pathogenesis of IPF. Alveolar epithelial injury and impaired restitution are thought central to the development of this condition, and all these significant markers are known to be expressed by alveolar epithelial cells; their increased levels emphasise ongoing alveolar epithelial cell damage in IPF [18, 19]. Indeed, OPN, MMP7, ICAM1 and POSTN have been previously associated with IPF progression, in line with our findings [20–27]. Furthermore, OPN and MMP7 overexpression may also reflect amplified fibroblast activity and increased extracellular matrix deposition in IPF lungs, while elevated ICAM1 expression may highlight the inflammatory and immune dysregulation of this condition [22, 28, 29].
Several lines of investigation have demonstrated that circulatory proteins are associated with patient outcome, increased mortality and disease severity [20–27]. Furthermore, studies have quantified clinically relevant circulating biomarkers in IPF to create a predictive performance index [22, 30, 31]. Richards et al. [22] defined a threshold of 203 and 4 ng·mL−1 for ICAM1 and MMP7, respectively, with plasma concentrations higher than these defining thresholds being associated with significantly lower median survival times in IPF.
Similar to our study, Ashley et al. [30] used unbiased proteomics to identify relevant biomarkers associated with disease progression in IPF. The group identified biomarker thresholds for six analytes involved in proteolysis, angiogenesis and immune function. They derived an index score that correlated with disease progression; however, limitations of the Ashley et al. [30] study are the small sample size and lack of a validation cohort.
A more recent study by Adegunsoye et al. [31] also defined circulating threshold concentrations of 47 and 3 ng·mL−1 for OPN and MMP7, respectively, which they found positively correlated with transplant-free survival. In addition, the group generated a clinical-molecular signature-risk score based on several biomarkers, age and FVC % pred, classifying IPF patients in “low-risk” and “high-risk” groups (the latter had worse transplant-free survival and increased mortality risk). Although there is merit to generating a single threshold value, we believe tertiles enable interpretable stratification, reflect the biomarker distribution more closely and minimise the impact of outlying values.
A limitation of our study was the application of the progression index to retrospective PROFILE study data, as there were insufficient data from some biomarkers to undertake a complete analysis of the progression index. Another limitation of the study was the inconsistency of the timing of pulmonary function test data in the AIPFR cohort, as these tests were collected for clinical follow-up rather than mandated at specific time-points as in a clinical trial. However, replication of the AIPFR data demonstrated generalisability and this real-world scenario translates into a possible strength. The reproducibility of OPN, MMP7, ICAM1 and POSTN, and the progression index, across international, multicentre cohorts that were prospectively (AIPFR and TLF) and retrospectively (PROFILE) analysed supports the robustness of our findings. A further limitation of the study was the lack of data for CA125 and CA19-9 in the initial discovery analysis of the AIPFR cohort, which therefore prevented the inclusion of these biomarkers in the progression index. It is likely that future biomarker studies will measure CA125 and CA19-9 in IPF, and it should be possible to build these into the current model. Indeed, as current research matures, advanced radiological image scoring, telomere length and related gene polymorphisms, as well as epithelial basal cell proteomics, can be incorporated into such prediction indices.
In conclusion, this study is in line with current literature and adds to rapidly evolving work that has demonstrated elevated circulating levels of OPN, MMP7, ICAM1 and POSTN in IPF. The progression index has provided us a new way of assessing IPF disease progression by using several well-known biomarkers in an index, with the scope to add and reassess the addition of new clinically relevant markers. Our reproducible findings across different sites and cohorts support additional validation in larger datasets, strengthening the potential prognostic value of these circulatory molecules and associated scores in clinical practice.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary material ERJ-01181-2021.SUPPLEMENT
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-01181-2021.Shareable
Acknowledgements
We would like to thank the co-ordinators of the Australian Idiopathic Pulmonary Fibrosis Registry (AIPFR) in each state for the collection and entry of data and coordination of sample collection, and Svetlana Baltic and Marisa Ryan (University of Western Australia and Institute for Respiratory Health, Perth, Australia) who collected and processed the blood samples. We would like to extend our gratitude to the statisticians C. Oldmeadow and L. Leigh from the Hunter Medical Research Institute (Newcastle, Australia), who designed the joint modelling to define the IPF groups, and carried out some of the statistical analyses. The proteomics analyses were performed in facilities provided by the Lotterywest State Biomedical Facility-Proteomics node and Bioplatforms Australia at the Harry Perkins Institute for Medical Research (Perth, Australia).
Footnotes
This article has supplementary material available from erj.ersjournals.com
Conflict of interest: B. Clynick has nothing to disclose.
Conflict of interest: T.J. Corte reports grants, personal fees and nonfinancial support from Boehringer Ingelheim, grants and personal fees from Roche, grants from Galapagos, Actelion and Bayer, outside the submitted work.
Conflict of interest: H.E. Jo reports other (travel support and lecture fees) from Boehringer Ingelheim and Roche, outside the submitted work.
Conflict of interest: I. Stewart has nothing to disclose.
Conflict of interest: I.N. Glaspole reports personal fees for advisory board work from Boehringer Ingelheim and Roche, outside the submitted work.
Conflict of interest: C. Grainge reports personal fees for advisory board work from Boehringer Ingelheim and Roche, outside the submitted work.
Conflict of interest: T.M. Maher reports grants, personal fees and nonfinancial support from UCB, grants and personal fees from GlaxoSmithKline and AstraZeneca, personal fees from Boehringer Ingelheim, Roche, Bayer, Prometic, Samumed, Galapagos, Celgene, Indalo, Pliant, Blade Therapeutics, Respivant, Novartis and Bristol-Myers Squibb, other (stock options) from Apellis, outside the submitted work.
Conflict of interest: V. Navaratnam has nothing to disclose.
Conflict of interest: R. Hubbard has nothing to disclose.
Conflict of interest: P.M.A. Hopkins has nothing to disclose.
Conflict of interest: P.N. Reynolds has nothing to disclose.
Conflict of interest: S. Chapman reports personal fees for advisory board work from Boehringer Ingelheim and Roche, outside the submitted work.
Conflict of interest: C. Zappala has nothing to disclose.
Conflict of interest: G.J. Keir has nothing to disclose.
Conflict of interest: W.A. Cooper has nothing to disclose.
Conflict of interest: A.M. Mahar has nothing to disclose.
Conflict of interest: S. Ellis has nothing to disclose.
Conflict of interest: N.S. Goh reports grants from the National Health Medical Research Council (NHMRC) (APP1066128, APP114776) and the Centre of Research Excellence in Pulmonary Fibrosis (CRE-PF), Australia (NHMRC GNT116371; 2017-2021), during the conduct of the study; the PROFILE study was funded by the Medical Research Council (G0901226) and GlaxoSmithKline R&D (CRT114316).
Conflict of interest: E. De Jong has nothing to disclose.
Conflict of interest: L. Cha has nothing to disclose.
Conflict of interest: D.B.A. Tan has nothing to disclose.
Conflict of interest: L. Leigh has nothing to disclose.
Conflict of interest: C. Oldmeadow has nothing to disclose.
Conflict of interest: E.H. Walters has nothing to disclose.
Conflict of interest: R.G. Jenkins reports other (sponsored research agreements) from AstraZeneca, Biogen and Galecto, outside the submitted work; is on advisory boards for Boehringer Ingelheim, NuMedii, Promedior and Redex; reports lecture fees and consultancy for AstraZeneca, Boehringer Ingelheim, GlaxoSmithKline, Heptares, Pliant and Roche; and is a trustee for Action for Pulmonary Fibrosis.
Conflict of interest: Y. Moodley reports personal fees for advisory board work from Boehringer Ingelheim and Roche, outside the submitted work.
Support statement: The study is supported by the National Health Medical Research Council (NHMRC) grants (APP1147776 and APP1066128) and the Centre of Research Excellence in Pulmonary Fibrosis (GNT116371; supported by the NHMRC, Lung Foundation Australia and industry sponsors including: Boehringer Ingelheim, Roche Products Pty Ltd, Galapagos and Bristol-Myers Squibb Australia). Lung Foundation Australia has established the Australian Idiopathic Pulmonary Fibrosis Registry with the generous support of unrestricted educational grant from Foundation Partners Roche Products Pty Ltd and Boehringer Ingelheim. The PROFILE study was funded by the Medical Research Council (G0901226) and GlaxoSmithKline R&D (CRT114316), and was sponsored by Nottingham University and the Royal Brompton and Harefield NHS Foundation Trust. Additionally, R.G. Jenkins is supported by an NIHR Research Professorship (RP-2017-08-ST2-014); T.M. Maher is supported by a National Institute for Health Research Clinician Scientist Fellowship (CS-2013-13-017). Funding information for this article has been deposited with the Crossref Funder Registry.
- Received April 25, 2021.
- Accepted August 3, 2021.
- Copyright ©The authors 2022. For reproduction rights and permissions contact permissions{at}ersnet.org