. Author manuscript; available in PMC: 2023 Jun 1.

Published in final edited form as: Thorax. 2022 Aug 5;78(6):566–573. doi: 10.1136/thorax-2021-218139

Table 3.

Associations of baseline telomere length with high attenuation areas

Model	No. Participants	Mean percent change in Exam 1 HAAs (95% CI)	P-value	% Longitudinal change in HAAs per 10 years (95% CI)	P-value

Absolute telomere length^*	4,488	−6.51 (−10.24 to −2.62)	0.001	0.32 (−4.68 to 5.35)	0.90
Age-adjusted telomere length percentile cutoffs
Above 5^th percentile	4,267	0.0 (REF)		0.0 (REF)
Below 5^th percentile	221	0.44 (−2.55 to 3.52)	0.78	0.87 (−2.82 to 4.56)	0.65
Above 10^th percentile	4,042	0.0 (REF)		0.0 (REF)
Below 10^th percentile	446	0.16 (−2.00 to 2.36)	0.88	1.41 (−1.28 to 4.10)	0.30
Above 25^th percentile	3,369	0.0 (REF)		0.0 (REF)
Below 25^th percentile	1,119	1.56 (0.04 to 3.11)	0.04	0.11 (−1.74 to 1.96)	0.91

Abbreviations: CI=confidence intervals; HAAs=high attenuation areas

Reported per standard deviation increment of log-transformed telomere length

Exam 1 HAAs model: Adjusted for scanner parameters, principal components of genetic ancestry, and baseline age, sex, self-reported race/ethnicity, smoking status, cigarette pack-years height, weight, and percent emphysema.

Longitudinal HAAs model: Adjusted for scanner parameters and principal components of genetic ancestry. Baseline age, sex, self-reported race/ethnicity, smoking status, cigarette pack-years were also adjusted for including their interaction terms with “time since initial HAAs assessment.” Time-varying covariates height, weight, percent emphysema, and cigarettes smoked per day were also adjusted for in the model.