Interstitial lung abnormalities and aging-related factors

Abstract

Introduction: Interstitial lung diseases (ILD), such as idiopathic pulmonary fibrosis (IPF), are diseases of elderly people that are characterised by pulmonary deposition of fibrous tissue and often have a poor prognosis. Interstitial lung abnormalities (ILA) are radiologic changes that are similar in appearance to ILD but are characterised in cohort study participants without known ILD. Prior research findings, including risk factors parallels such as advanced age, suggest that ILA are likely a precursor to ILD in some cases and are themselves associated with poor health outcomes. Due to the strong links that ILA and ILD share with advanced aging, the aims of the thesis were to assess the relationship of ILA with aging-related biological markers and outcomes in four papers. These were functional status, pulmonary and extra-pulmonary malignancies, a variety of blood proteins and leukocyte telomere length. Methods: Data from three cohort studies, the Age/Gene-Environment Susceptibility-Reykjavik (AGES-Reykjavik) study, the Genetic epidemiology of COPD (COPDGene) study and the Framingham Heart study were used. CT images of participants had been manually read with regards to ILA status. Outcomes were ascertained from the cohorts’ phenotyping data and from electronic medical records. The associations of three self-reported measures of health and functional status with ILA were assessed with logistic regression. The relationship of ILA with diagnoses of and mortality from pulmonary and non-pulmonary malignancies was assessed with Gray’s tests and proportional hazards models. The associations of ILA, ILA progression and single nucleotide polymorphisms associated with pulmonary fibrosis with protein markers were assessed with adjusted regression models of single proteins. Adaptive LASSO modelling of bootstrap data samples was used to find sets of proteins predictive of ILA and ILA progression. The relations of ILA with leukocyte telomere length and telomere length with mortality among those with ILA were modelled with regression and Cox proportional hazards models. Results: In the AGES-Reykjavik cohort, participants with ILA were less likely to be independent in activities of daily living (odds ratio (OR) 0.70, 95% confidence interval (CI) 0.55-0.90), to perceive their health as good or better (OR 0.66, CI 0.52-0.82) and to be regularly physically active (OR 0.72, CI 0.56-0.91). The cumulative incidences of lung cancer diagnoses and lung cancer mortality were higher among participants with ILA than others (p < 0.001), but such differences were not found for other cancers. In adjusted Cox proportional hazards models, ILA were associated with diagnoses of lung cancer (hazard ratio (HR) 2.77, CI 1.76-4.36) and mortality from lung cancer (HR 2.89, CI 1.80-4.66) but not with diagnoses of, or mortality from, cancers excluding lung cancer. In adjusted analyses of single serum proteins and ILA in AGES-Reykjavik, 287 proteins were significantly associated with ILA. The most significant associations were for Surfactant protein B (SFTPB), Secretoglobin family 3A member 1 (SCGB3A1) and WAP four-disulfide core domain protein 2 (WFDC2). An eight-protein model of ILA status was created based on inclusion in 200 adaptive LASSO models in bootstrap data samples which had an area under the receiving operator characteristic curve (AUROC) value of 0.880 after bootstrap validation. Single-protein results for SFTPB, SCGB3A1 and WFDC2 and the eight-protein model were validated in COPDGene (validated AUROC 0.826). Similarly, 121 proteins were associated with ILA progression in single protein models in AGES-Reykjavik and multi-protein modelling yielded a four-protein model which had a validated AUROC of 0.824. In analyses of fibrosis-associated genetic polymorphisms, the rs35705950 polymorphism near the MUC5B gene was found to be associated with levels of SFTPB (β 0.26, p = 8×10-18). As for analyses of ILA and leukocyte telomere length, shorter telomere length was associated with increased odds of ILA in COPDGene (OR 2.2, CI 1.5-3.4) and AGES-Reykjavik (OR 2.6, CI 1.4-4.9) and ILA was associated with shorter telomere length in FHS (β -767 base pairs, CI -76 - -1584). Adjusted Cox proportional hazards models did not demonstrate a significant association of telomere length with mortality among those with ILA in COPDGene nor in AGES-Reykjavik. Conclusions: These cohort study-based data demonstrate the associations of ILA with several aging-related markers and outcomes. These are subjective markers of functional status, diagnoses of and mortality from pulmonary malignancies and telomere length in leukocytes. Additionally, many novel protein biomarkers of ILA are proposed, some of them previously related to aging. These findings improve knowledge of biological markers and epidemiological outcomes associated with ILA as a possible marker of early pulmonary fibrosis.