Identification and Prognosis of Patients With Interstitial Pneumonia With Autoimmune Features

Nikhil Jiwrajka, MD; Giorgos Loizidis, MD; Karen C. Patterson, MD; Maryl E. Kreider, MD, MSCE; Cheilonda R. Johnson, MD, MHS; Wallace T. Miller, Jr, MD; Eduardo Jose Mortani Barbosa, Jr, MD; Namrata Patel, MD; Michael F. Beers, MD; Leslie A. Litzky, MD; Michael D. George, MD, MSCE; Mary K. Porteous, MD, MSCE


J Clin Rheumatol. 2022;28(5):257-264. 

In This Article


Identification of Patients Classifiable as IPAF

The ILD registry included 456 patients with MDD diagnoses of IPF (n = 125), CTD-ILD (n = 133), cryptogenic organizing pneumonia (n = 10), idiopathic NSIP (n = 9), lymphoid interstitial pneumonia (n = 1), and unclassifiable ILD (n = 178) (Figure 1). The baseline characteristics of these groups are included in Supplementary Table 1 ( After medical record review and application of CTD and IPAF classification criteria, we classified 60 patients as IPAF, 113 as CTD-ILD, and 126 as IPF. Patients classified as IPAF primarily were comprised patients who previously had MDD diagnoses of unclassifiable ILD (n = 27) or CTD-ILD (n = 26; 10 of whom had an MDD diagnosis of "undifferentiated connective tissue disease") (Supplementary Figure 2,

Figure 1.

Flow diagram of the study cohort. COP = cryptogenic organizing pneumonia; iNSIP, idiopathic NSIP; LIP, lymphoid interstitial pneumonia. "Other" includes patients with hypersensitivity pneumonitis, chronic aspiration, drug-induced ILD, respiratory bronchiolitis (RB)-ILD, granulomatous-lymphocytic (GL)-ILD, lymphangioleiomyomatosis (LAM), Birt-Hogg-Dubé syndrome, IgG4-related disease, eosinophilic pneumonia, Langerhans cell histiocytosis, and acute interstitial pneumonia.

The most common CTD-ILDs after reclassification were systemic sclerosis (27.4%, n = 31), rheumatoid arthritis (26.5%, n = 30), ASyS/anti-MDA5 dermatomyositis (15.9%, n = 18), other dermatomyositis/polymyositis (8.8%, n = 10), and Sjögren syndrome (8.8%, n = 10) (Supplementary Figure 3, Notably, although 19 patients with CTD-ILD (16.8%) were diagnosed with ILD prior to their CTD-ILD diagnosis, all met the criteria for CTD-ILD within 6 months of their baseline visit—none were classifiable as IPAF. Among those with CTD-ILD, patients with ASyS/anti-MDA5 dermatomyositis were more likely to present with ILD either prior to or concurrently with their CTD diagnosis (Supplementary Figure 4,

Characteristics of Patients Classified as IPAF, CTD-ILD, and IPF

Of the 60 patients classified as IPAF, 57 (95.0%) fulfilled serologic domain criteria, 49 (81.7%) fulfilled morphologic domain criteria, and 24 (40.0%) fulfilled clinical domain criteria (Figure 2, Supplementary Table 2, The most common clinical criteria were Raynaud phenomenon (25.0% of all IPAF cases, n = 15) and inflammatory arthritis (16.7%, n = 10). These were also the most common clinical criteria in patients with CTD-ILD, present in 44.2% and 46.0% of patients, respectively (Supplementary Figure 5, The most common serologic criteria in patients with IPAF were abnormal antinuclear antibody (ANA; 51.7%, n = 31) and positive anti-CCP (26.7%, n = 16). Most patients who fulfilled morphologic criteria had NSIP on HRCT (48.3%, n = 29) or NSIP on histopathology (16.7%, n = 10). The majority of patients with IPAF (60.0%, n = 36) fulfilled criteria via the combination of the serologic and morphologic domains. Although patients classified as IPAF underwent comprehensive serologic testing, a substantial proportion (56.7%, n = 34) had missing non-Jo1 antisynthetase antibodies. Notably, cytoplasmic staining was reported on ANA testing in several of these patients (n = 4) (Supplementary Table 3,

Figure 2.

Clinical, serologic, and morphologic domain composition of patients classified as IPAF. Overlapping regions of the Venn diagram indicate the proportion of all IPAF patients with a particular IPAF domain combination. The smaller offset circles depict the percentage and count of all IPAF patients who fulfilled criteria for a given IPAF domain. Bar graphs indicate the percentage of all IPAF patients with a particular domain criterion. Counts of each domain criterion are provided in Supplementary Table 2, Antisynthetase antibodies consisted of anti-Jo1 (n = 5), anti–PL-12 (n = 2), and anti-EJ (n = 1). Raynaud, Raynaud phenomenon; other, Scl-70, PM-Scl, or MDA-5; OP, organizing pneumonia; LymPlaCyt infiltrate, diffuse lymphoplasmacytic infiltration.

At baseline, the mean age of patients with IPAF (61.9 years) was similar to that of patients with CTD-ILD (58.7 years; p = 0.11) but significantly lower than that of patients with IPF (70.3 years; p < 0.01 vs. IPAF) (Table 1). Patients with IPAF and CTD-ILD were more likely to be female (63.3% and 70.8%) than patients with IPF (31.0%; p < 0.01 vs. IPAF) and more likely to be Black (20.0% and 23.9%) than patients with IPF (2.4%; p < 0.01 vs. IPAF). Duration of ILD at the first clinic visit, as well as measures of baseline pulmonary function, were similar between groups; 71.7% of patients with IPAF had seen a rheumatologist during follow-up, compared to 95.6% of patients with CTD-ILD (p < 0.01) and 16.7% of patients with IPF (p < 0.01); 23.7% of patients with IPAF demonstrated UIP. Notably, patients with IPAF were more likely to undergo lung biopsy (33.3%; p < 0.01 vs. CTD-ILD or IPF). Although patients with IPAF were less likely to be on immunosuppression at their baseline visit compared to patients with CTD-ILD, (48.3% vs. 70.8%, p < 0.01), these patients were equally likely to be treated with immunosuppression during follow-up.

Transplant-free Survival Analysis

During follow-up, 72 patients died (16 IPAF, 22 CTD-ILD, and 34 IPF patients), and 20 underwent lung transplantation (3 IPAF, 4 CTD-ILD, and 13 IPF patients), with similar median follow-up times across all 3 groups (Figure 3). Patients with CTD-ILD had longer unadjusted transplant-free survival than patients with IPF (log-rank p = 0.02). We did not detect statistically significant differences in rates of transplant-free survival among patients with different types of CTD-ILD, although transplant-free survival of CTD-ILD patients with myositis/ASyS-ILD appeared longer than that of other types of CTD-ILD (Supplementary Figure 6,

Figure 3.

Transplant-free survival of patients classified as IPAF versus patients with CTD-ILD versus patients with IPF. Survival analysis excludes those patients whose first ILD clinic visit occurred prior to 2011 as described under Materials and Methods. Patients with IPF demonstrated worse transplant-free survival compared to patients with CTD-ILD (log-rank p = 0.02). There was no statistically significant difference in transplant-free survival between patients classified as IPAF and patients with CTD-ILD (log-rank p = 0.07 vs. IPAF) or patients with IPF (log-rank p = 0.80 vs. IPAF). Median follow-up is reported as the median time (in years) [interquartile range]. Cumulative transplant-free survival is shown at year 4 (95% confidence interval).

There was no statistically significant difference in unadjusted transplant-free survival between patients with IPAF and patients with either CTD-ILD (p = 0.07 vs. IPAF) or IPF (p = 0.80 vs. IPAF). Results were similar in age- and sex-adjusted analyses (Supplementary Table 4, Pairwise comparison of transplant-free survival rates between IPAF and different CTD-ILDs revealed significantly longer transplant-free survival for patients with myositis/ASyS-ILD versus IPAF (log-rank p < 0.01), but no differences with respect to other CTD-ILDs (Supplementary Figure 6, Transplant-free survival of IPAF was modestly improved in sensitivity analyses in which CTD-ILD patients with ASyS without idiopathic inflammatory myopathy were instead classified as IPAF (Supplementary Figure 7,

In univariate analysis among patients with IPAF, male sex was associated with significantly worse transplant-free survival (hazard ratio [HR], 4.58 [1.77–11.87]; p < 0.01) and remained a significant predictor of death or transplant even after adjusting for age and baseline FVC (HR, 4.17 [1.56–11.13]; p < 0.01) (Table 2). The presence of UIP was not associated with transplant-free survival among patients with IPAF.

Patients Classified as IPAF who Developed CTD Diagnoses During Follow-up

We identified 6 patients (10.0%) classified as IPAF who subsequently fulfilled criteria for a CTD 1.2 to 10.7 years after their initial date of ILD diagnosis (Supplementary Table 5, All of these patients had been evaluated by a rheumatologist at the time of IPAF classification. Antisynthetase syndrome was the most frequent (50.0%, n = 3) eventual CTD diagnosis in these patients. In all of these cases, extrapulmonary manifestations developed later during follow-up, although in some instances, missed follow-up appointments led to delays in diagnosis.