Biological Age, Chronological Age, and Survival in Pulmonary Fibrosis: A Causal Mediation Analysis
Janelle Vu Pugashetti, John S. Kim, Swaraj Bose, Ayodeji Adegunsoye, A. Linderholm, Ching‐Hsien Chen, Mary E. Strek, Kevin R. Flaherty, Susan Murray, Chad A. Newton, Shehabaldin Alqalyoobi, Shwu‐Fan Ma, Josyf C. Mychaleckyj, Russell P. Bowler, MeiLan K. Han, Jeffrey L. Curtis, Fernando J. Martínez, Jennifer A. Smith, Imre Noth, Justin M. Oldham
Abstract
Abstract Rationale Accelerated biological aging has been implicated in the development of interstitial lung disease (ILD) and other diseases of aging but remains poorly understood. Objectives To identify plasma proteins that mediate the relationship between chronological age and survival association in patients with ILD. Methods Causal mediation analysis was performed to identify plasma proteins that mediated the chronological age–survival relationship in an idiopathic pulmonary fibrosis discovery cohort. Proteins mediating this relationship after adjustment for false discovery were advanced for testing in an independent ILD validation cohort and explored in a chronic obstructive pulmonary disease cohort. A proteomic-based measure of biological age was constructed and survival analysis performed, assessing the impact of biological age and peripheral blood telomere length on the chronological age–survival relationship. Measurements and Main Results Twenty-two proteins mediated the chronological age–survival relationship after adjustment for false discovery in the idiopathic pulmonary fibrosis discovery cohort (n = 874), with 19 remaining significant mediators of this relationship in the ILD validation cohort (n = 983) and one mediating this relationship in the chronic obstructive pulmonary disease cohort. Latent transforming growth factor-β binding protein 2 and ectodysplasin A2 receptor showed the strongest mediation across cohorts. A proteomic measure of biological age completely attenuated the chronological age–survival association and better discriminated survival than chronological age. Results were robust to adjustment for peripheral blood telomere length, which did not mediate the chronological age–survival relationship. Conclusions Molecular measures of aging completely mediate the relationship between chronological age and survival, suggesting that chronological age has no direct effect on ILD survival.