FormalPara Key Summary Points

Why carry out this study?

Autoantibody signatures of patients with idiopathic pulmonary fibrosis (IPF) without diagnosed connective tissue disease may be linked with specific disease characteristics and prognosis.

This post-hoc analysis of data from the ASCEND trial aimed to determine the association between the presence of common autoantibodies and baseline disease characteristics, disease progression and treatment outcomes in patients with IPF who were randomised to receive pirfenidone or placebo.

What was learned from the study?

In the current analysis, baseline characteristics and disease course shared similarities by autoantibody status.

Among patients receiving placebo, no significant differences were observed in the evaluated efficacy endpoints for patients who were ANA+, RF+ or anti-CCP+ when compared with autoantibody-negative patients.

By comparing treatment arms, clinical outcomes and management of patients with IPF remain unchanged in those with autoantibody positivity in the absence of other clinical features of systemic autoimmune rheumatic disease.

Introduction

Interstitial lung disease (ILD) is a large and heterogeneous group of pulmonary disorders; some are associated with an underlying autoimmune aetiology and some are linked to environmental exposures, whereas others have unknown causes [1, 2]. It can be challenging to differentiate between different types of ILD due to overlapping clinical, radiological and pathological presentations [1]. Diagnostic guidelines recommend that patients with suspected idiopathic pulmonary fibrosis (IPF), the most common and severe form of ILD, undergo autoantibody testing as part of the initial evaluation to assess for autoimmune-mediated diseases [3,4,5]. In previous studies, autoantibodies such as antinuclear antibodies (ANA), rheumatoid factor (RF) and anti-cyclic citrullinated peptide (anti-CCP) have been reported in 22–33% of patients with IPF in the absence of connective tissue disease (CTD) on initial evaluation, with one study reporting that up to 10% of patients with IPF progressed to CTD [6,7,8]. However, the link between markers of autoimmunity and disease characteristics is not fully understood and may be an important factor to consider in the initial diagnosis and management of patients with IPF [1, 3] who do not otherwise manifest clinically with CTD.

Previous studies have suggested that differences in autoantibody status may be associated with unique patient profiles and differences in prognosis in patients with IPF, and that the presence of autoantibodies in IPF may represent a novel subgroup of patients, but this evidence is based on limited patient numbers, short follow-up duration and a lack of robust clinical data [6, 7, 9, 10]. Here, in this post-hoc analysis, we aimed to determine the association between autoantibody status and baseline disease characteristics, disease progression and treatment outcomes in patients with IPF who were randomised to receive pirfenidone or placebo in ASCEND, a large, well-characterised Phase III clinical trial.

Methods

The trial design of ASCEND (NCT01366209) has been previously reported [11]. In brief, ASCEND was a randomised, double-blind, placebo-controlled Phase III trial in which 555 patients with IPF were randomised to receive either oral pirfenidone (2403 mg/day; n = 278) or placebo (n = 277) for 52 weeks.

Eligible participants were aged 40–80 years and had a centrally confirmed diagnosis of IPF with findings on high-resolution computed tomography (HRCT) of the chest indicating either definite or possible usual interstitial pneumonia (UIP), with a surgical lung biopsy to confirm the presence of definite or probable UIP in the latter group. Patients with diagnosis of any CTD, including scleroderma, polymyositis/dermatomyositis, systemic lupus erythematosus (SLE) or rheumatoid arthritis, or any known explanation for ILD, including sarcoidosis and hypersensitivity pneumonitis, were excluded [11]. In addition, participants were recruited if they had 50–90% of percent predicted forced vital capacity (%FVC), 30–90% of percent predicted carbon monoxide diffusing capacity (%DLco), forced expiratory volume in 1 s/forced vital capacity (FVC) ratio of ≥ 0.80 and a 6-min walk distance (6MWD) of ≥ 150 m. All participants randomised to pirfenidone or placebo in the ASCEND trial with any autoantibody result at screening were included in this post-hoc analysis.

The ASCEND trial was conducted in compliance with Good Clinical Practice as described in FDA regulations and the 1996 International Council for Harmonisation document, in consistence with the principles stated in the Declaration of Helsinki. The protocol for the ASCEND trial was approved by the institutional review board or ethics committee at each participating centre and all patients provided written informed consent for participation in the trial. No prospective data were collected during this post-hoc analysis; therefore, ethical approval was not required.

Post-Hoc Analysis

Peripheral blood samples collected at screening were used to determine the presence of the following autoantibodies: ANA, RF and anti-CCP. Autoantibody status was defined by the following titres and staining patterns:

  1. 1.

    ANA-positive (ANA+) participants had ANA titre ≥ 1:40 or nucleolar-staining pattern or centromere-staining pattern as defined by previous studies, regardless of RF or anti-CCP positivity [2, 12].

    1. a.

      Participants with high ANA+ levels (H-ANA+) had ANA titre ≥ 1:160 or a positive nucleolar-staining pattern or centromere-staining pattern (regardless of titre level), both of which were independent of RF or anti-CCP positivity; a previous study suggests the ≥ 1:160 cutoff would be likely to exclude 95% of individuals without systemic sclerosis, SLE or Sjogren’s syndrome [12].

    2. b.

      Participants with low ANA+ levels had ANA titre ≥ 1:40 to < 1:160 and absence of both nucleolar-staining pattern and centromere-staining pattern, regardless of RF or anti-CCP positivity; a previous study suggests that an ANA cutoff level of 1:40 could have diagnostic value, and a survey of laboratories participating in the College of American Pathologists’ Proficiency Testing Programme suggests that a majority of US laboratories use this traditional cutoff for reporting ANA positivity [12, 13].

  2. 2.

    RF-positive (RF+) participants had RF titre ≥ 20 IU/mL.

  3. 3.

    Anti-CCP–positive (anti-CCP+) participants had anti-CCP titre ≥ 20 IU/mL.

  4. 4.

    Autoantibody-negative (AAb−, triple negative) participants had ANA titre < 1:40, “negative” titre with an absence of nucleolar-staining pattern or centromere-staining pattern, and negative RF (< 20 IU/mL) and negative anti-CCP (< 20 IU/mL) titres.

The endpoints of the current analysis focused on participants who were classified as ANA+, H-ANA+, RF+ and/or anti-CCP+ and AAb−. The endpoints included:

  1. 1.

    Summaries of the demographic and baseline characteristics organised by autoantibody subgroup and/or treatment arm.

  2. 2.

    Changes in %FVC from baseline to Week 52, determined using a fixed effect rank analysis of covariance, where the outcome variable was standardised ranked change from baseline, and fixed effect was either the participant ANA status, treatment arm or RF/anti-CCP status. The ranked baseline %FVC was included as a covariate (deaths were ranked worst according to time until death).

  3. 3.

    Changes in FVC volume (litres) from baseline to Week 52, determined using the same approach as for the %FVC and by ranking the relative change in volume defined as (Week 52 FVC volume − baseline FVC volume)/baseline FVC volume.

  4. 4.

    Estimation of progression-free survival (PFS), defined as first occurrence of death, confirmed ≥ 10% decline from baseline in %FVC or confirmed ≥ 50 m decline from baseline in 6MWD. The decline in either %FVC or 6MWD was confirmed at two consecutive assessments at least 6 weeks apart. PFS was analysed using the product limit method log-rank test and a proportional hazards model with treatment as a covariate. ANA status or RF/anti-CCP status was used to estimate the hazard ratio (HR) and Kaplan–Meier estimates were used to summarise PFS time.

Data from participants receiving placebo were analysed to determine the effect of autoantibody status (ANA+, H-ANA+, RF+ and/or anti-CCP+, or AAb−) on the course of disease, whereas data from participants receiving pirfenidone versus those receiving placebo were analysed to assess the impact of autoantibody status (ANA+ or AAb−) on response to treatment. P values for autoantibody-positive groups versus the AAb− group were calculated using Pearson’s chi-squared test. Comparisons of some subgroups of placebo- or pirfenidone-treated participants (e.g., those who were H-ANA+ and RF+ and/or anti-CCP+) were not included in the analysis due to their small sample size.

Results

Autoantibody Analysis

All 555 enrolled participants from the ASCEND trial had autoantibody data available for analysis (Table 1). In total, 514 participants were tested for the presence of ANA, of whom 47.5% (244/514) were classed as ANA+ and 16.1% (83/514) of participants were further categorised as H-ANA+. Additionally, 10.8% (60/555) participants were classed as RF+ and/or anti-CCP+. Overall, 52.5% (270/514) of participants were classed as AAb−, i.e., had a confirmed negative status for all the tested autoantibodies.

Table 1 Autoantibody results at screening in the ASCEND trial
Full size table

Baseline demographics and characteristics by autoantibody status (ANA+, H-ANA+, RF+ and/or anti-CCP+ and AAb−) for all participants are presented in Table 2 and for participants receiving pirfenidone or placebo are presented in Table 3. Key baseline demographics and characteristics were broadly similar between participants who were ANA+ and those who were AAb− (Table 2). However, there was a higher proportion of American Indian or Alaska Native participants in the ANA+ (8.6%) and RF+ and/or anti-CCP+ (8.3%) groups versus the AAb− (3.3%) group. There was also a greater proportion of women in the H-ANA+ (27.7%) and RF+ and/or anti-CCP+ (31.7%) groups versus the AAb− (20.0%) group (Table 2). Additionally, when compared with the AAb− group, the H-ANA+ group had a lower proportion of current smokers (51.8% vs. 64.1%), a lower proportion of patients requiring supplemental oxygen (19.4% vs. 30.4%) and a lower 6MWD (400.0 m vs. 424.0 m) (Table 2).

Table 2 Baseline demographics and characteristics for all patients receiving pirfenidone or placebo from the ASCEND trial, grouped by autoantibody status
Full size table
Table 3 Baseline demographics and characteristics for participants receiving pirfenidone or placebo from the ASCEND trial, split by autoantibody status and treatment arm
Full size table

Disease Course in Placebo-Treated Participants

Overall, the disease course in placebo-treated participants was similar regardless of autoantibody status. Numerically, a greater proportion of participants in the ANA+ and H-ANA+ groups had a decline from baseline to Week 52 of ≥ 10% in %FVC or death (48.7% and 55.9%, respectively) or in FVC volume or death (48.7% and 47.1%, respectively) compared with the AAb− group (%FVC or death: 42.0%; FVC volume or death: 42.0%) (Fig. 1). However, there were no statistically significant differences in decline in %FVC or death or in FVC volume or death between the AAb− group and the ANA+ and H-ANA+ groups (Fig. 1). The proportions of patients with decline in %FVC or death or in FVC volume or death, in the RF+ and/or anti-CCP+ group were similar to those in the AAb− groups, again with no statistically significant difference (Fig. 1). There was no difference in PFS between the ANA+ group versus the AAb− group [HR (95% confidence interval [CI]): 1.14 (0.78 to 1.66); P = 0.5] (Fig. 2A) or between the H-ANA+ group versus the AAb− group [HR (95% CI): 1.22 (0.69 to 2.17); P = 0.5] (Fig. 2B). Due to small sample sizes, analysis of PFS for the H-ANA+ and RF+ and/or anti-CCP+ groups was not performed.

Fig. 1
figure 1

Placebo-treated participants with decline from baseline to Week 52 of ≥ 10% in %FVC or in FVC volume, or death, stratified by ANA+, H-ANA+, RF+ and/or anti-CCP+ and AAb− status. P values were calculated using Pearson’s chi-squared test. %FVC percent predicted forced vital capacity, AAb autoantibody, ANA antinuclear antibody, anti-CCP anti-cyclic citrullinated peptide, FVC forced vital capacity, H-ANA+ high antinuclear antibody titre, RF rheumatoid factor

Full size image
Fig. 2
figure 2

PFS in placebo-treated participants, stratified by A ANA+ versus AAb− and B H-ANA+ versus AAb− status. AAb autoantibody, ANA antinuclear antibody, CI confidence interval, H-ANA+ high antinuclear antibody titre, HR hazard ratio, PFS progression-free survival

Full size image

Response to Pirfenidone Treatment

Clinically relevant trends towards a treatment effect for pirfenidone over placebo were observed for patients who were ANA+. Numerically lower proportions of participants with ANA+ who received pirfenidone than who received placebo had decline in %FVC or death, or decline in FVC volume or death, although this difference did not reach statistical significance for either endpoint (%FVC or death: pirfenidone 37.2% vs. placebo 48.7%, P = 0.093; FVC volume or death: pirfenidone 35.7% vs. placebo 48.7%, P = 0.053) (Fig. 3). In the AAb− group, there was a statistically significant benefit of pirfenidone over placebo for both decline in %FVC or death and decline in FVC volume or death (both endpoints: pirfenidone 29.1% vs. placebo 42.0%, P = 0.039) (Fig. 3). PFS was statistically significantly higher for participants receiving pirfenidone compared with those receiving placebo in both the ANA+ group [HR (95% CI): 0.56 (0.37 to 0.86); P = 0.007; Fig. 4A] and the AAb− group [HR (95% CI): 0.50 (0.32 to 0.78); P = 0.002; Fig. 4B]. Due to small sample sizes, analysis of PFS for the H-ANA+ group and the RF+ and/or anti-CCP+ group was not performed.

Fig. 3
figure 3

Proportion of participants treated with pirfenidone or placebo with decline from baseline to Week 52 of ≥ 10% in A %FVC or death or B FVC volume or death stratified by ANA+ and AAb−. P values were calculated using Pearson’s chi-squared test. %FVC, percent predicted forced vital capacity, AAb autoantibody, ANA antinuclear antibody, d day, FVC forced vital capacity

Full size image
Fig. 4
figure 4

PFS for A pirfenidone versus placebo in participants with ANA+ and B pirfenidone versus placebo in participants with AAb−. AAb autoantibody, ANA antinuclear antibody, CI confidence interval, d day, HR hazard ratio, PFS progression-free survival

Full size image

Discussion

In this post-hoc analysis of data from the ASCEND trial, we report the impact of autoantibody status on disease progression and treatment responses in patients with IPF. We did not identify any prominent pulmonary physiologic differences in baseline characteristics of participants with IPF who were classed as positive for commonly tested autoantibodies versus AAb−. Some differences were observed between the subgroups of participants with high ANA levels versus AAb−, such as a higher proportion of women and American Indians or Alaska Natives, and a lower proportion of ever smokers and supplemental oxygen users; however, other baseline clinical characteristics (%FVC, %DLco) were comparable between groups. Our findings support those of previous studies of patients with IPF that found broadly similar results in baseline demographics, pulmonary function tests and definite UIP pattern on HRCT based on the presence of autoantibodies [6, 14,15,16].

Among placebo-treated participants, there was no difference in PFS between the AAb− group and the ANA+ or H-ANA+ groups, despite participants who were ANA+ or H-ANA+ being more likely to exhibit a non-statistically significant decline of ≥ 10% in %FVC or FVC volume compared with those who were AAb−. Several studies have previously examined the associations between autoantibody status and outcomes in patients with IPF [6, 9, 14, 16, 17]; however, ours is the largest analysis reporting post-hoc prospective data from a randomised controlled trial. Although data on other autoantibodies were not available for this analysis, an analysis of patients with IPF who were part of the Pulmonary Fibrosis Foundation Patient Registry suggested that baseline characteristics and clinical outcomes were generally similar among patients regardless of baseline seropositivity status across a wide range of autoantibodies, including ANA, RF, anti-CCP, anti-Smith and anti-myositis antibodies [18].

Our analysis, based on a large, well-characterised clinical trial population, indicates that participants responded to pirfenidone treatment regardless of their ANA status, with a treatment effect of pirfenidone over placebo for PFS in both the ANA+ and the AAb− groups. A significant treatment effect was also observed for ≥ 10% percent predicted FVC decline or death and ≥ 10% FVC volume decline or death from baseline to Week 52 in the AAb− group, although this clinically relevant trend in the ANA+ group did not reach statistical significance. These findings mirror those shown in a smaller, 6-month follow-up, retrospective observational study that also reported a pirfenidone treatment effect irrespective of autoantibody status [15].

Diagnostic guidelines recommend that patients with suspected IPF undergo autoantibody testing as part of the initial evaluation, but no consensus was reached about which autoantibodies should be included in screening panels [3]. Initial screening for a broad range of autoantibodies is not deemed mandatory for all patients with suspected IPF, although it could be useful in cases where other potential causes of ILD are clinically suspected [3]. The prognostic value of autoantibodies in patients diagnosed with IPF is not yet fully understood [1], and, as such, it will be important to consider further which autoantibodies should be included in the initial diagnostic screen [3].

Apart from its possible link with autoimmune diseases, ANA positivity has also been described as a factor of ageing, as its prevalence generally increases with age and it also correlates with shorter telomere length, a marker of biological age [19, 20]. Cellular senescence associated with ageing has been described in the pathogenesis of IPF, a disease of the aged population; whether the high prevalence of ANA positivity is merely an association or has pathological implication is unclear [21]. Nevertheless, therapeutic response to pirfenidone remains unaffected by autoantibody positivity in this study.

There are several limitations of this analysis. Firstly, it was a post-hoc exploratory analysis with only 52 weeks of follow-up available. Secondly, data on autoantibodies other than ANA, RF and anti-CCP, change in positivity or titre, and whether participants went on to develop systemic autoimmune rheumatic diseases, were not collected. Thirdly, the small number of participants in certain groups (e.g., H-ANA+, and RF+ and/or anti-CCP+) may limit the interpretation of the data and impact the observed outcomes; therefore, these results should be interpreted with caution. Fourthly, we present here available data for the three commonly evaluated autoantibodies that were included during screening for ASCEND. Diagnostic guidelines suggest screening for the presence of these antibodies in patients with suspected IPF, with evaluation using a full antibody panel reserved for cases where other autoimmune diseases are clinically suspected [3]. We recommend that screening for other disease-related autoantibodies be included in future studies in order to broaden potential analyses. Finally, although we used the placebo arm of the ASCEND trial as a proxy for the natural disease progression in IPF in this analysis, we cannot rule out the presence of a placebo effect in these patients, which could have influenced disease outcomes.

Conclusion

This post-hoc analysis of data from the ASCEND trial in patients with IPF indicates that disease course did not differ by ANA, RF or anti-CCP autoantibody status, although patients in the H-ANA subgroup have differences in certain baseline demographics compared with antibody-negative patients. Importantly, we observed a treatment benefit for pirfenidone regardless of ANA status, particularly in relation to PFS. This analysis underscores that, while some patients with IPF may have autoantibody positivity, in the absence of other clinical features of systemic autoimmune rheumatic disease, clinical outcomes and management remain unchanged.