Cluster phenotypes in a non-idiopathic pulmonary fibrosis fibrotic interstitial lung diseases cohort in Singapore

BACKGROUND: Non-idiopathic pulmonary fibrosis fibrosing interstitial lung diseases (F-ILDs) may demonstrate a progressive disease trajectory similar to idiopathic pulmonary fibrosis (IPF). We aimed to identify novel F-ILD phenotypes in a multi-ethnic South-East Asian population. METHODS: F-ILD subjects (n=201) were analysed using unsupervised hierarchical cluster analysis and their outcomes compared against IPF (n=86). RESULTS: Four clusters were identified. Cluster 1 (n=53, 26.4%) comprised older Chinese males with high body mass index (BMI) and comorbidity burden, higher baseline forced vital capacity (FVC) percentage predicted and lower diffusing capacity of the lung for carbon monoxide (DLCO) percentage predicted. They had similar mortality to IPF. Cluster 2 (n=67, 33.3%) had younger female non-smokers with low comorbidity burden, groundglass changes on high-resolution chest computed tomography (HRCT) and a positive anti-nuclear antibody (ANA) titre ≥1:160. They had lower baseline FVC and higher DLCO, low mortality and slower lung function decline. Cluster 3 (n=42, 20.9%) consisted male smokers with low comorbidity burden, emphysema on HRCT and high baseline lung function. They had low mortality and slow lung function decline. Cluster 4 (n=39, 19.4%) was the highest risk and comprised of mainly Indians with high BMI. They had the highest proportion of ischemic heart disease (IHD) and previous pulmonary tuberculosis. Subjects had the lowest baseline lung function, highest mortality, and fastest lung function decline. Survival differences across clusters remained significant following adjustment for treatment. CONCLUSIONS: We identified four distinct F-ILD clinical phenotypes with varying disease trajectories. This demonstrates heterogeneity in F-ILD and the need for complementary approaches for classification and prognostication beyond ATS/ERS guideline diagnosis.