Weight loss and outcomes in subjects with progressive pulmonary fibrosis: data from the INBUILD trial

Respiratory Research

Table 3 Association between change in weight (slope) and risk of outcomes over the whole INBUILD trial

	Acute exacerbation or death	ILD progression	ILD progression or death
N included
Placebo	330	327	327
Nintedanib	331	326	326
Longitudinal sub-model^a
Estimated change in weight (kg) (95% CI) with placebo	− 1.60 (− 2.08, − 1.12)	− 0.76 (− 1.34, − 0.19)	− 0.93 (− 1.50, − 0.37)
Estimated change in weight (95% CI) difference for nintedanib vs placebo	− 1.51 (− 2.19, − 0.84)	− 2.07 (− 2.84, − 1.31)	− 2.00 (− 2.76, − 1.24)
p-value	< 0.001	< 0.001	< 0.001
Time-to-event sub-model^b
n (%) with event
Placebo	64 (19.4)	156 (47.7)	177 (54.1)
Nintedanib	46 (13.9)	109 (33.4)	129 (39.6)
Hazard ratio (95% CI) for nintedanib vs placebo	0.60 (0.41, 0.89)	0.68 (0.52, 0.89)	0.64 (0.50, 0.81)
p-value	0.011	0.004	< 0.001
Association between change in weight (slope) and risk of event, hazard ratio (95% CI)
Per 1 kg decrease	1.08 (1.03, 1.14)	0.96 (0.91, 1.01)	1.00 (0.96, 1.04)
Per 4 kg decrease	1.38 (1.13, 1.68)	0.83 (0.68, 1.02)	1.01 (0.87, 1.18)
Per 10 kg decrease	2.23 (1.36, 3.65)	0.64 (0.39, 1.05)	1.03 (0.70, 1.53)
p-value	0.002	0.08	0.88

UIP usual interstitial pneumonia, ILD interstitial lung disease
^aRandom effects normal linear model of weight (kg) with HRCT pattern (UIP-like pattern or other fibrotic patterns) and weight at baseline as predictor variables, a separate slope assumed for the nintedanib and placebo groups, and trajectories modelled by a linear trend with an unstructured variance–covariance matrix assumed
^bProportional hazard model with a piecewise exponential baseline hazard, stratified by HRCT pattern, with treatment as a predictor variable and the endogenous time-dependent covariate of weight (kg) as estimated slope of the longitudinal response

ISSN: 1465-993X