Scary IPAF Save

Interstitial pneumonia with autoimmune features (IPAF) describes patients who have interstitial lung disease (ILD) plus clinical, serologic, or morphologic features of autoimmunity but who do not meet criteria for a defined connective tissue disease. The 2015 ERS/ATS criteria proposed IPAF as a research classification, not a distinct disease, intended to create more uniform cohorts for study (Fischer et al., European Respiratory Journal, 2015).

Consider a 40-year-old woman referred by pulmonology. She has an NSIP pattern on HRCT, new-onset Raynaud's phenomenon, and a nucleolar ANA at 1:1,280. On paper she meets IPAF criteria. If no defined diagnosis emerges, familiar questions remain: whom to treat, with what, and who should lead management? Outcome data to guide those decisions have been limited, and most of them come from referral centers. Enter Abstract 1619.

The authors built what they describe as the first population-based incident IPAF cohort: 113 patients from a geographically defined area, diagnosed between 2010 and 2019 and followed through June 30, 2025. They estimated overall survival with Kaplan-Meier methods. For ED visits, hospitalizations, and first oxygen prescription, they estimated cumulative incidence while treating death as a competing risk. Cox models were adjusted for age, sex, and year of IPAF incidence.

Within 5 years of IPAF incidence:

• 42% died
• 69% were hospitalized
• 89% had an ED visit
• 14% were newly prescribed supplemental oxygen

Predictors of outcomes (HR, 95% CI)

Death:

• Prior oxygen prescription: 2.70 (1.19–6.16)
• Prior hospitalization: 1.87 (1.09–3.21)
• Age: 1.77 (1.40–2.23)
• Liver disease: 1.77 (1.01–3.12)
• Pulmonary vasculopathy: 1.67 (1.02–2.71)

Oxygen prescription:

• Rheumatoid factor ≥30: 11.02 (2.96–41.02)
• Family history of liver disease: 3.17 (1.12–8.99)

Hospitalization:

• PR3: 19.93 (2.06–192.38)
• ANA centromere pattern: 6.55 (1.45–29.53)
• Hematologic cancer: 3.58 (1.05–12.2)
• Prior hospitalization: 2.93 (1.74–4.91)
• Prior oxygen prescription: 2.58 (1.05–6.32)
• Liver disease: 2.11 (1.24–3.57)
• Age: 1.33 (1.09–1.62)

ED visits:

• PR3: 10.85 (1.29–91.18)
• Mechanic's hands: 9.18 (1.18–71.59)
• Anti-Jo-1: 5.06 (1.04–24.64)
• Non-White race: 2.68 (1.11–6.50)
• Diabetes mellitus: 2.11 (1.22–3.63)

Notably, about 90% of the cohort had received glucocorticoids at some point.

Some caution is warranted. With 113 patients and relatively few events, the serologic hazard ratios are imprecise.  These associations should be treated as hypothesis-generating, not as validated risk markers. Prior hospitalization and prior oxygen use may reflect underlying disease severity rather than independent causes of poor outcomes. The utilization outcomes also appear to be all-cause, not specific to ILD, which partly explains the very high ED-visit rate. Finally, the cohort comes from a single region, so generalizability is uncertain.

This population-based study supports the view that IPAF is not a benign entity. The authors argue for early risk stratification using patient-specific factors and for closer monitoring of patients with high-risk features. My own view goes a step further. The study does not establish the best treatment, but rheumatology should be an active co-manager of these patients, working alongside pulmonology much as we do in ILD associated with myositis or systemic sclerosis. 

Practically, that means:

• Re-evaluating periodically for an evolving defined CTD
• Monitoring pulmonary function and imaging
• Paying particular attention to patients with features this cohort flagged, such as older age, oxygen need, prior hospitalization, pulmonary vasculopathy, liver disease, or antisynthetase or ANCA-related serologies