Skip to main content
. 2020 Apr 28;75(8):648–654. doi: 10.1136/thoraxjnl-2019-213865

Table 4.

Multivariable Cox regression analyses models demonstrating mortality prediction determined by change in various visual CT variables (measured with 5-point ordinal scores) in the combined idiopathic pulmonary fibrosis cohorts (n=211)

Categorical change in visual CT variables HR, 95% CI, P value, Concordance Index
Total ILD extent 1.68, 1.33 to 2.12, 5.9×10–5, 0.70
Ground-glass opacity extent 1.64, 1.28 to 2.09, 0.0004, 0.68
Reticular pattern extent 1.61, 1.26 to 2.05, 0.0007, 0.69
Honeycombing extent 1.90, 1.44 to 2.51, 2.7×10–5, 0.69
Traction bronchiectasis severity 2.14, 1.59 to 2.88, 2.5×10–6, 0.70

P values shown are adjusted for multiple comparisons.

Each visual CT variable was analysed in a separate model adjusted for patient age, gender, baseline disease severity using the diffusion capacity for carbon monoxide and antifibrotic use (never/ever).

ILD, interstitial lung disease.