Clinical and serological predictors of autoimmune etiology in patients with interstitial lung disease: a 10-year retrospective study.

objectiveInterstitial lung disease (ILD) may represent the first manifestation of an underlying autoimmune disorder. This study aimed to determine the prevalence of connective tissue disease-associated ILD (CTD-ILD) and interstitial pneumonia with autoimmune features (IPAF) among patients referred for rheumatologic evaluation, and to identify clinical predictors of autoimmune etiology.

methodsWe retrospectively evaluated 226 consecutive patients diagnosed with ILD between 2010 and 2020 and referred to the Rheumatology clinic for etiological assessment. Demographic characteristics, clinical features, laboratory and imaging findings were analyzed. Univariate and multivariable logistic regression models were used to determine independent factors associated with autoimmune interstitial lung disease (AI-ILD), defined as CTD-ILD or IPAF. To address potential circularity and incorporation bias, five pre-specified sensitivity analyses were performed.

resultsAmong the study population, 50.9% were classified as AI-ILD (30.1% CTD-ILD and 20.8% IPAF), while 49.1% had non-rheumatologic ILD. Sjögren's disease was the most common underlying rheumatic disease (41.1%). A usual interstitial pneumonia (UIP) pattern was present in 49% of patients. In multivariable analysis, ocular involvement (dry eye confirmed on ophthalmologic examination; OR 8.884, 95% CI 3.475-22.712, p<0.001) and ANA positivity (any titer; OR 4.163, 95% CI 2.114-8.197, p<0.001) were independently associated with autoimmune etiology. Absence of comorbidities was also independently associated (OR 2.052, 95% CI 1.093-3.852, p=0.025). High-titer ANA (titer≥1:320) was observed exclusively in AI-ILD patients (0/111 non-rheumatologic vs 69/115 AI-ILD). These findings were consistent across all five sensitivity analyses (AUC range 0.710-0.857).

conclusionsAmong patients referred for rheumatologic evaluation, autoimmune etiologies accounted for approximately half of ILD cases. Confirmed ocular involvement and ANA positivity were independently associated with autoimmune ILD across all model specifications and five pre-specified sensitivity analyses. These data support their use in clinical triage within similar referral settings. Key Points • High Prevalence of Autoimmunity: Nearly half of the patients referred for rheumatologic evaluation following an ILD diagnosis were found to have an underlying autoimmune etiology. • The Dominance of Sjögren's: Contrary to some previous reports, Sjögren's disease emerged as the most frequent connective tissue disease in this cohort, likely reflecting the value of systematic sicca screening. • Consistent Predictors: Confirmed ocular involvement and ANA positivity were independently associated with AI-ILD across five sensitivity analyses in this referral population. • The 1:320 Threshold: High-titer ANA (≥1:320) appeared exclusively in AI-ILD patients in this cohort. This finding requires external validation before it can inform broader clinical practice.