Pneumologie 2019; 73(S 01)
DOI: 10.1055/s-0039-1678098
Posterbegehung (P09) – Sektion Klinische Pneumologie
Neue Entwicklungen bei ILD
Georg Thieme Verlag KG Stuttgart · New York

Predicting mortality in idiopathic pulmonary fibrosis (IPF) patients treated with Nintedanib[*]

J Behr
1   Medizinische Klinik und Poliklinik V, University of Munich (Lmu) and Asklepios Klinik München-Gauting, Member of the German Center for Lung Research, Germany
,
W Tang
2   Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, Connecticut, USA
,
S Menjoge
2   Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, Connecticut, USA
,
S Stowasser
3   Boehringer Ingelheim International GmbH, Ingelheim am Rhein, Germany
,
J Korell
2   Boehringer Ingelheim Pharmaceuticals, Inc., Ridgefield, Connecticut, USA
› Author Affiliations
Further Information

Publication History

Publication Date:
19 February 2019 (online)

Also available at
 
 

    Rationale Exploration of predictors of mortality in IPF is important both clinically and to future drug development. By conducting a pooled analysis of a large cohort of patients treated with nintedanib 150 mg b. i.d from the TOMORROW and INPULSIS trials the association of FVC decline, exacerbations, and baseline predictors with mortality was investigated.

    Methods A stratified Cox proportional hazards model with backward stepwise variable selection was used to evaluate the association between baseline predictors, FVC % predicted (%pFVC) decline from baseline, acute exacerbations, and all-cause mortality, where %pFVC decline and acute exacerbations were treated as time-varying covariates. To illustrate the potential application of the model in the clinical setting, predictive analysis of one-year survival was performed using simulated %pFVC data from an IPF disease progression model developed from the same trials.

    Results A total of 723 patients treated with nintedanib 150 mg b. i.d were included in the analysis. Among the baseline variables investigated, older age, status as a current smoker, lower %pFVC, and lower DLco % predicted were associated with an increased risk of death. Additionally, an increased risk of death was observed for patients having experienced one or more exacerbations, with a hazard ratio (HR) of 7.7 (95% confidence interval [CI], 3.5 – 16.9). Lastly, compared with a %pFVC absolute decline from baseline at any time during the study of < 5%, a decline ≥ 5% but < 10% was associated with an increased risk of death, with an HR of 3.3 (95% CI, 1.6 – 7.1), as was a decline ≥ 10%, with an HR of 3.9 (95% CI, 1.8 – 8.7). The observed associations were further illustrated in the predictive analysis of one-year survival.

    Conclusion In this pooled analysis of IPF patients treated with nintedanib from the TOMORROW and INPULSIS trials, we developed a model for estimating risk of mortality based on a set of important baseline and time-varying predictors. Our analyses confirm the importance of baseline age, lung function, and smoking status as predictors of mortality, and provide additional evidence regarding the strong association of FVC decline and acute exacerbations with death.

    Compared to reference level:

    1: Age<65

    2: Non-smoker

    3: %p-FVC>91.05

    4: %p-DLco>53.97

    5: %p-FVC Decline<5


    #

    * *  presented at ATS 2018