Interstitial Pneumonia with Autoimmune Features (IPAF): time to redefine the classification criteria

KEYWORDS:

• Interstitial pneumonia with autoimmune features
• undifferentiated connective tissue disease
• diagnosis
• autoantibodies
• idiopathic inflammatory myopathies

The research classification named ‘Interstitial Pneumonia with Autoimmune Features’ (IPAF) was proposed to include patients with Interstitial Lung Disease (ILD) associated with clinical and/or serological characteristics that are not sufficient to meet classification criteria for any specific Autoimmune Rheumatic Disease (ARD) [1]. IPAF criteria were proposed by an expert consensus, serving as a potential common platform for further improvements.

IPAF resembles the concept of Undifferentiated Connective Tissue Disease (UCTD), which is widely debated in rheumatology settings [2], with the interesting difference that IPAF does not necessarily require seropositivity. The UCTD criteria were proposed in 1999, showing a progression toward ARDs in 20–60% of patients over 3–10 years, mainly toward Systemic Lupus Erythematosus and Rheumatoid Arthritis (RA) [2]. Over the last 20 years, the classification criteria for ARDs have been revised several times considering research findings. These revisions improved performance (mainly sensitivity) for ARDs classification, reducing the number of patients considered UCTD, and also the UCTD rate of progression, as some patients were already classifiable for a definite ARD.

This data could be explained, at least in part, by a selection bias: a nuanced, if not totally absent, autoimmune clinical picture, could be underestimated by physicians and the patient, who tend to seek out the specialist deemed most useful for the management of the main clinical problem (ILD). A tight collaboration between rheumatologists and pulmonologists could be useful for diagnostic purposes, allowing for the correct classification of patients, and could also be useful in research, highlighting ARD patients in which ILD is the first, the main, or even the sole clinical manifestation. It is quite surprising that the sole ARD for which ILD is included in the validated classification criteria is systemic sclerosis (SSc).

As currently proposed, IPAF criteria allow the inclusion of patients in which the ‘autoimmune feature’ could be stochastically associated: clearly if ANA positivity with a titer of 1:320 can be recognized in about 3% of healthy subjects [3], a similar proportion of patients with Nonspecific Interstitial Pneumonia (NSIP) could be seropositive with the same ANA titer, without this being related to the ILD. Conversely, the autoimmune feature could be the expression of an early/incomplete form of ARD. An overhaul of IPAF criteria could limit the first possibility and improve recognition of the second. The studies have shown that patients enrolled with the current IPAF criteria progress toward ARDs in 20–30% [4–6]. In particular, a recent manuscript reports the largest prospective cohort of IPAF patients (191 subjects), followed for a mean time of 31 months, proving a progression toward ARDs in 24.1% [4]. The rate of progression can be a possible ‘gold standard’ to evaluate the actual effectiveness of IPAF criteria in including patients with early (or at least incomplete) ARDs, and a revision of IPAF criteria could increase this proportion.

Below, we suggest modifications to the IPAF criteria.

The clinical domain is composed of inflammatory arthritis or polyarticular morning stiffness >60 minutes, Raynaud’s Phenomenon (RP), digital edema, distal digital tip ulcerations, palmar telangiectasias, Mechanic’s Hands (MH), Gottron’s Sign (GS). Firstly, these clinical items might be difficult to recognize for not trained physicians, and a close collaboration between rheumatologists and pulmonologists could be needed. Moreover, IPAF criteria showed a progression mainly toward Idiopathic Inflammatory Myopathies (IIMs) and primary Sjögren’s Syndrome (pSS) [4-,6]. A symptomatic ILD preceding myositis or sicca syndrome is relatively common for IIMs and pSS [7,8]. While IIMs could be recognized by some clinical signs such as MH and GS, the main pSS symptom (sicca syndrome) is not mentioned. Its introduction would probably improve the recognition of pSS patients presenting with ILD as the first clinical symptom [8]. Another important point is to balance the level of sensitivity and specificity when enrolling patients. Myalgia was defined as nonspecific for the recognition of ARDs [1], however ILD patients with myalgia showed a high proportion of positivity for Myositis Specific and Myositis Associated Antibodies (MSA/MAA), therefore this symptom could merit inclusion [9]. Morning stiffness is also quite nonspecific, although included. Since 2016, we have considered this item present if associated with an increase in Erythrosedimentation Rate or C Reactive Protein level [10]. On the other hand, distal digital tip ulcerations, as well as digital edema and palmar telangiectasias are generally very specific for SSc [11]. Their recognition in ILD patients is more likely to be associated with a definite diagnosis. To confirm this, prospective studies on IPAF cohorts involving rheumatologists did not report this item [5,6,10,12,13], and almost never report the presence of GS. Considering the mean age of IPAF patients (and ILD patients in general) the presence of GS associated with ILD should be considered for classification at least as a probable IIM, according to the 2017 criteria of Lundberg IE et al. [14].

The serological domain includes almost all the known autoantibodies. However, a cytoplasmic ANA pattern is not included. This pattern is quite common in patients with IIM [15], and in our opinion, it could have the same weight as a nucleolar or anti-centromere pattern, therefore the inclusion of patients at any titer, if present, should be allowed. The presence of anti-La is not useful, as it has been proved that the presence of this autoantibody without anti-Ro is not associated with autoimmune conditions [16]. The serological domain also includes anti-tRNA synthetase antibodies (AtRSA), but it is well-recognized that antisynthetase syndrome can be present with ILD as the sole clinical manifestation [7]. Therefore, these autoantibodies are not useful to the aims of the IPAF criteria, and they should be substituted with other non-AtRSA MSA/MAAs, such as Tif1γ, anti Ku, NXP2, SAE1 and Mi2. The major concern is that these autoantibodies are evaluated, at least in clinical practice, using immunoblotting, with the risk of false positivity. Finally, Anti-Neutrophil Cytoplasm Antibodies (ANCA) were excluded due to their specificity in recognizing vasculitides. However, the IPAF criteria are useful as they suggest an autoimmune pathway and can recognize autoimmune conditions with potential systemic involvement, therefore the exclusion of vasculitides can be questioned. As a final point on the serological domain, the authors of the IPAF criteria stated that reevaluating autoantibodies during the follow-up of these patients may not be useful [1]. It is generally true that autoantibody positivity precedes the diagnosis of ARDs, however the possibility of some fluctuations in seropositivity should be taken into account. These are generally possible for autoantibodies associated with disease activity, and cannot be excluded for other antibodies of which our current knowledge is limited (e.g. MSA/MAA). To confirm this, prospective studies on IPAF patients proved the progression of patients toward ARDs in patients that were seronegative at the first assessment and with specific seropositivity during the follow-up [5].

The Morphological Domain of IPAF, while not excluding the possibility, limits the enrollment of patients with a Usual Interstitial Pneumonia (UIP) pattern, deemed by the authors not to be sufficiently associated with ARDs. However, UIP is the most common radiological pattern in RA-ILD, in longstanding SSc-ILD and in vasculitis, ILD preceding sicca syndrome in pSS commonly shows a UIP pattern, and it can also be recognized in up to 20% of IIMs [17]. We studied UIP patients with and without only one IPAF domain (clinical or serological), calling this group ‘UIPAF.’ Our UIPAF patients progressed toward ARDs in a proportion similar to classic IPAF, and significantly higher than Idiopathic Pulmonary Fibrosis [5,18]. For these reasons, we suggest completely removing the morphological domain, and substituting it with an ‘instrumental domain.’

The instrumental domain could include some very simple, rapid and inexpensive tools for the diagnosis of ARDs. Functional gland impairment (mainly with unstimulated whole saliva test and Schirmer’s test) can be positive in pSS patients without significant symptoms, so may be suggestive of the disease [19]. Nailfold videocapillaroscopy is also a useful tool, assessing RP as possibly associated with scleroderma spectrum disorders (SSD). However, it could show alterations (bushy capillaries or even a scleroderma pattern) also in IIM, regardless of the presence of RP [20,21]. In our opinion, NVC should be performed on all ILD patients, in order to exclude the presence of a defined SSD in RP patients, but also to find elements suggesting IIM in patients without RP.

In conclusion, IPAF is a research classification, and it cannot be considered a diagnosis. This classification has the great merit of stimulating close collaboration between pulmonologists and rheumatologists, sharing knowledge that contributes to clear improvements not only in research, but also in daily practice. IPAF criteria need an overhaul, in light of the current literature. The clinical and serological domains can be reviewed, the morphological domain could be substituted with an ‘instrumental’ domain, including NVC and tests for gland impairment, that ideally should be performed on all ILD patients A potential revision of these criteria could be started in late 2023, after the upcoming publication of the classification criteria for Antisynthetase Syndrome by the CLASS project [22]. New IPAF criteria could be validated by examining the rate of progression toward ARDs and compared with the current criteria (24.1%) [4].

Conflict of Interest: GS reports personal fees from Boehringer Ingelheim outside the submitted work; CV is part of the F. Hoffmann-La Roche Ltd. and Boehringer Ingelheim Scientific board. He has received consulting fees and/or speaker fees from AstraZeneca, Boehringer Ingelheim, Chiesi, F. Hoffmann-La Roche Ltd and Menarini. DS declares no conflict of interest.

Declaration of interest

G Sambataro reports personal fees from Boehringer Ingelheim outside the submitted work; C Vancheri is part of the F. Hoffmann-La Roche Ltd. and Boehringer Ingelheim Scientific board. He has received consulting fees and/or speaker fees from AstraZeneca, Boehringer Ingelheim, Chiesi, F. Hoffmann-La Roche Ltd and Menarini. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded.