Table 1

Emphysema-like lung on CT and incidence of airflow limitation at 5-year follow-up

Model*	Cases/at risk	Per cent emphysema>upper limit of normal (dichotomous)†		Per cent emphysema log transformed (continuous)
Model*	Cases/at risk	OR (95% CI)‡	p Value	OR per SD (95% CI)‡	p Value
Incident prebronchodilator airflow limitation§	224/2045
Unadjusted		1.86 (1.12 to 3.09)	0.016	1.64 (1.39to 1.94)	<0.001
Minimally adjusted		2.07 (1.22 to 3.51)	0.007	1.48 (1.23 to 1.77)	<0.001
+age, sex, race/ethnicity
Fully adjusted		2.62 (1.47 to 4.67)	0.001	1.92 (1.55 to 2.38)	<0.001
+height, weight, CT scanner, smoking history, initial FEV₁
Extended		2.41 (1.32 to 4.40)	0.004	1.61 (1.28 to 2.03)	<0.001
+initial FEV₁/FVC
Incident postbronchodilator airflow limitation§	53/1915
Unadjusted		2.47 (1.03 to 5.94)	0.043	1.91 (1.35 to 2.69)	<0.001
Minimally adjusted		2.56 (1.05 to 6.25)	0.040	1.79 (1.23 to 2.60)	0.003
+age, sex, race/ethnicity
Fully adjusted		4.38 (1.63 to 11.74)	0.003	2.59 (1.65 to 4.08)	<0.001
+height, weight, CT scanner, smoking history, initial FEV₁
Extended		4.63 (1.71 to 12.53)	0.003	2.36 (1.47 to 3.79)	<0.001
+initial FEV₁/FVC

*Logistic regression models were sequentially adjusted. The minimally adjusted model includes study baseline age, sex and race/ethnicity. The fully adjusted model additionally includes height, weight, CT type, smoking status, pack-years, urinary cotinine and FEV₁ per cent predicted at the initial spirometry exam. The extended model also includes initial FEV₁/FVC ratio.
†The ULN for per cent emphysema on cardiac CT was defined according to reference equations developed in the same cohort accounting for age, gender, race/ethnicity, height, weight, current smoking and scanner parameters.8
‡ORs reported for presence versus absence of per cent emphysema>ULN or per SD (3.1%) of log-transformed per cent emphysema, with 95% CI.
§Airflow limitation was defined as FEV₁/FVC<0.700.10
ULN, upper limit of normal.