Background
Interstitial pneumonia with autoimmune features (IPAF) is a conceptual entity proposed to identify patients with interstitial pneumonia and features suggestive of connective tissue disease (CTD), but not meeting established classification criteria for CTD [
1]. The traditional serologic and clinical features of connective tissue disease (CTD) were included in the proposed criteria of IPAF. The findings from high-resolution computed tomography (HRCT), histopathology and other diagnostic modalities such as pulmonary function tests and echocardiography which were consistent with CTD, were included in the morphologic domain of IPAF [
2].
IPAF criteria were applied to patients previously diagnosed as undifferentiated-CTD interstitial lung disease (UCTD-ILD) [
3]. According to previous publications, majority of IPAF patients were shown to be female and had no or little smoking history [
4,
5]. Furthermore, the most frequent ILD pattern on HRCT was non-specific interstitial pneumonia (NSIP) [
6,
7]. The clinical outcomes of IPAF have been compared to other ILD types. From the study by Oldham et al., the IPAF cohort showed worse survival than the CTD-ILD group, while showing slightly better survival than the idiopathic pulmonary fibrosis (IPF) group [
3]. The study by Ahmad et al. found no significant difference between IPAF and IPF [
7].
However, in order to recommend the classification of IPAF for wider use among clinicians, more clinical data needs to be accumulated. A series of studies on IPAF were published after the expert consensus in 2015, however, little is known about longitudinal clinical outcomes of IPAF, and the question whether IPAF has better survival than IPF remains to be answered. Furthermore, the acute exacerbation (AE) of ILD has been reported to be a significant negative prognostic factor associated with worse survival in CTD-ILD and IPF [
8‐
10]. Nevertheless, it has not been evaluated for IPAF in previous studies.
In the present study, we compared clinical characteristics, survival and ILD exacerbation of IPAF patients to the CTD-ILD and IPF groups, from the ILD cohort of a single institution.
Discussion
We applied the definition of IPAF in patients with ILD from the cohort at our institution, and further compared clinical characteristics and clinical outcomes of this cohort to patients with CTD-ILD and IPF from the same institution. Our study has its strength in comparing longitudinal clinical outcomes, including ILD exacerbation between the different ILD groups. In the present study, IPAF showed distinct clinical characteristics when compared to other ILD types. In addition, the IPAF group showed better survival and less exacerbation events when compared to the IPF group.
Of the three domains of the IPAF criteria, features of the serologic domain was most commonly observed (90.7%), thereafter, features of the morphologic domain (81.5%). This is different from the results of a previous study by Yoshimura et al. in which features of the morphologic domain was most commonly observed (96.9%), followed by the serologic domain [
17], and consistent with the results of the study on IPAF by Ahmad et al. [
7].
When baseline clinical characteristics were compared, the IPAF group in our study showed some similarities to previous studies. In our study, NSIP was the most frequent HRCT pattern, which was consistent with previous studies [
7,
17]. Idiopathic NSIP is often associated with autoimmune features [
6], and a high proportion of NSIP in the IPAF group suggests a possible correlation with autoimmune features.
Furthermore, the percentage of female patients was more than 50% for both the IPAF group and the CTD-ILD group. A female predominance in the IPAF group of our study is consistent with results of previous publications [
5,
6]. However, Ahmad et al. demonstrated a slight male predominance in the IPAF group as well as the IPF group (female percentage 49%) [
7]. Regarding smoking history, our study showed a lower proportion of ever smokers than previous studies. (27.8% vs 34.8 and 38.8%) [
7,
18].
In addition to comparison of baseline clinical characteristics, our study focused on longitudinal clinical outcomes of IPAF: AE-ILD and mortality. First, the proportions of patients who experienced exacerbations in the IPAF group at different time periods were significantly lower when compared to the IPF group. Moreover, despite statistical insignificance, the time to first exacerbation in the IPAF group was longer than that of the IPF group. In IPF, AE-ILD is a known significant negative prognostic factor [
8], and was found to be a significant factor associated with mortality in our multivariate analysis.
Secondly, our study demonstrated that the IPAF group showed significantly better survival than the IPF group. Ahmad et al. showed that there was no significant difference in OS between patients with IPAF and patients with IPF [
7]. In addition, Oldham et al. showed that there was no statistically significant difference in OS between the IPAF and IPF cohort. On the other hand, both studies by Ahmad et al. and Oldham et al. showed that the survival of patients with IPAF was not better than that of patients with CTD-ILD [
3,
7].
The finding that the IPAF group showed better survival compared to the IPF group should be considered in conjunction with the difference in ILD exacerbation incidence and proportion of UIP pattern observed on the initial HRCT. We first hypothesized that ILD exacerbation would be the major factor contributing to the difference of overall survival between the ILD groups, because AE-ILD is a strong negative prognostic factor in IPF [
8]. Furthermore, compared to the IPF group, the proportion of patients who experienced AEs was significantly less in the IPAF group and this difference in proportion was repeatedly shown in year 1, 3 and 5. In addition, IPAF was independently associated with better survival compared to IPF, with other factors such as AE-ILD also adjusted in the analysis. However, the definition of AE in IPF was also applied to non-IPF groups in the present study [
14], and this may have affected the analysis. Further studies, including more detailed evaluation of AEs in non-IPF ILD are needed to clarify other intrinsic factors associated with survival in IPAF.
We initially assumed that the proportion of UIP pattern observed on HRCT would influence the clinical outcome, because the UIP pattern has been reported to be associated with worse survival [
19,
20]. In our study, survival comparison between the 15 IPAF patients with UIP pattern and the IPF group showed no significant difference (
p = 0.08). The IPAF group tends to have a longer survival duration than the IPF group, however, the difference was not statistically significant. We believe that a future prospective study focusing on comparing these two specific subgroups may give a clearer comparative result.
We also attempted to compare the IPAF group to groups of other ILD types with different levels of autoimmune features. When the diagnostic criteria of IPAF were first recommended by ATS/ERS, they focused on ILD with autoimmunity, not meeting established criteria of CTD [
1]. From our study, the IPAF group showed some clinical findings similar to CTD-ILD, such as sex, smoking history and time to first exacerbation. We believe that the similarities came from autoimmune tendencies shared by the two disease entities. Furthermore, IPAF patients should be placed under longitudinal surveillance for future occurrence of CTD [
18,
21]. We think that patients who were categorized as IPAF at the time of initial diagnostic work up, should be re-evaluated on a regular basis for the possibility of definitive CTD-ILD. We also compared clinical characteristics of the IPAF group to the IPF group with seropositivity, considering the possibility that the latter group may have some autoimmune features. However, the seropositive IPF group showed no clinical similarities with the IPAF and the CTD-ILD groups in terms of sex, smoking history, radiologic findings and clinical outcomes. In addition, the seropositive IPF group did not show distinct clinical characteristics from the seronegative IPF group. Despite the serologic findings, we believe that the IPAF group and IPF group with autoimmune antibody share little similarity in clinical characteristics.
IPAF is a relatively recently defined disease category of ILD, and the applicability of the criteria has been repeatedly discussed [
21‐
23]. From our study, arthralgia and Raynaud’s phenomenon were the two most common findings among the criteria of the clinical domain. This finding was similar to a previous study [
7]. However, distal digital fissuring (“mechanic’s hands”), unexplained fixed rash on the digital extensor surfaces (Gottron’s sign), and distal digital tip ulceration were absent or less frequent signs in our study. We assume that some findings related to the clinical domain are more suggestive of CTD rather than fulfilling the diagnostic criteria for IPAF. As was suggested by Ahmad et al., these signs often lead to the diagnosis of CTD after referral to rheumatologists [
7]. On the other hand, there were 10 patients from our ILD cohort who showed dry mouth or eye symptoms and also showed seropositivity or radiologic/histopathologic findings coherent to the morphologic domain of IPAF, but not fulfilling the diagnostic category of IPAF. Dry eye or dry mouth are clinical symptoms related to Sjogren’s disease [
24]. Such symptoms are subjective, because they are non-visual findings and can vary depending on patients’ general condition. However, the symptoms of dry eye and dry mouth can be clinically relevant if other rheumatologic findings are present and should be checked for relevance to autoimmune features. We think that clinical findings such as dry eye or dry mouth could be carefully considered for inclusion into the clinical domain if concurrent autoimmune features are present.
The present study has some limitations. First, this is a single institution retrospective study, and the possibility of selection bias exists. However, the study patients were enrolled consecutively and many of the previous studies were also single institution-based studies [
5,
7]. Secondly, categories other than CTD-ILD and IPF were not included in the study. CTD-ILD and IPF were two major ILD categories in our ILD cohort, while cases of disease such as hypersensitivity pneumonitis and sarcoidosis were not sufficient in number for comparison. In addition, we focused on the two groups (CTD-ILD and IPF), as they show different levels of autoimmune tendencies [
15,
16]. Lastly, the treatment data in our study included only antifibrotics and short-term systemic steroids, and did not include detailed analysis of other treatment modalities such as lung transplant or immunosuppresives. In addition, baseline comorbidities of patients were not described. We think that a future study including more detailed treatment data and comorbidities is necessary.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.