Methods

To explore these associations, data were obtained from the Age, Gene/Environment Susceptibility-Reykjavik (AGES-Reykjavik) study. Details of the study design have been previously published.6 Informed consent was obtained from all participants. Two separate CT scans of the thorax were obtained with a four-row CT scanner (Sensation, Siemens Medical System, Erlangen, Germany) that covered about 95% of the lungs, omitting the superior portion of the apices. The axial images were evaluated by up to three readers (pulmonologists and chest radiologists) blinded to participant-specific information, using a sequential reading method as previously described.1 Scans labelled as ILA, indeterminate and 20% of normal scans were read by two readers and the third reader provided majority opinion on discordantly labelled scans. ILA was defined as ‘non-dependent changes affecting more than 5% of any lung zone, including reticular or ground-glass abnormalities, diffuse centrilobular nodularity, non-emphysematous cysts, honeycombing or traction bronchiectasis’ while ‘focal or unilateral ground-glass attenuation, focal or unilateral reticulation or patchy ground-glass abnormalities (<5% of any lung zone)’ were defined as indeterminate changes. The data on smoking status, health status, activities of daily living (ADLs) and physical activity were obtained from participant questionnaires.6 The question regarding health status was ‘In general, how would you say your health is?’, with the answers ranging from 1 (‘excellent’) to 5 (‘poor’). For statistical analyses, participants were categorised into two groups based on whether they perceived their health as ‘good’ or better. Thus, one group was composed of participants that perceived their health as ‘good’, ‘very good’ or ‘excellent’, while participants that perceived their health as ‘poor’ or ‘fair’ made up the other group. ADLs were assessed by asking participants whether they had difficulties performing the following five activities: bathing, dressing, walking from room to room, transferring out of bed/chair or eating.7 For statistical analyses, participants were grouped into two groups: people independent in all ADLs and people dependent in one or more ADLs. Physical activity was estimated by asking participants how many hours per week, during the last 12 months, they participated in moderate-vigorous intensity physical activity. The possible answers were never, rarely, every week but less than 1 hour, 1–3 hours per week, 4–7 hours per week and more than 7 hours per week. For statistical analyses, people were categorised into two groups: people who never or rarely participated in physical activity and those who did so at least weekly. Comparisons of baseline characteristics were done using Χ² tests, t-tests and Wilcoxon rank-sum tests as appropriate. The analysis of the association of ILA with measures of health and functional status was done using logistic regression modelling. Participants with indeterminate ILA status were excluded from these analyses. The measures of health were assigned to multivariable models adjusted for covariates including age, sex, body mass index (BMI), pack-years of smoking and current smoking status. All statistical analyses were done using R V.3.3.2.