The Association of demographic and physiological variables and risk of subsequent disease progression in idiopathic pulmonary fibrosis
O'Riordan, Thomas Gerard
MetadataShow full item record
Background. Efficient clinical trial design in idiopathic pulmonary fibrosis (IPF) is hindered by incomplete understanding of the natural history of IPF and absence of robust predictive models. A retrospective analysis of the RAINIER clinical trial for simtuzumab, an inhibitor of lysl-oxidase-like-2 (LOXL-2), was conducted to identify clinically predictive variables associated with IPF disease progression. The RAINIER trial was discontinued for lack of efficacy. Because the trial did not demonstrate efficacy, it was considered appropriwte to include both control and simtuzumaab treated subjects. Methods. RAINIER was a randomized, double-blind, phase 2 clinical trial (NCT01769196) conducted at 181 respiratory clinics in 14 countries. Patients with IPF diagnosed within 3 years prior to screening or with evidence of clinical worsening were enrolled between March 2013 and June 2015 and randomized 1:1 to simtuzumab or placebo. Standard demographic and physiological baseline variables, 2 baseline exploratory endpoints (serum LOXL2 and fibrosis score from quantitative high-resolution computed tomography and the change in forced vital capacity (L) over 14 weeks (prodromal ΔFVC) from placebo-treated patients, were used to generate predictive models for weeks 14–66 for change in forced vital capacity (FVC), time to disease progression (as a composite of mortality and categorical decrease in FVC), death, and time to first all-cause and adjudicated respiratory hospitalization. A backward process selected variables (threshold p≤0.05) using a linear mixed model for ΔFVC. A stepwise process selected variables (entry \threshold p=0.10, stay threshold p=0.05) in survival analysis for time-to-event outcomes. Performance of the predicted models was evaluated by a 10-fold cross-validation method. Results. Of 544 subjects enrolled in the study,, 501 remained in the study at 14 weeks. An increased prodromal ΔFVC was associated with increased risk for decreased time to death, (HR=0.936, p=0.017, first adjudicated respiratory hospitalization (HR=0.980, p=0.001) and categorically defined disease progression (HR =0.936, p< 0.0001) over 52 weeks but was paradoxically associated with a reduced rate of decline in FVC over the same period (Pearson correlation coefficient [R] =---2.060, p<0.001). Prodromal ΔFVC was not significant in the model of time to death or all-cause hospitalization. Serum LOXL2 at baseline significant predictor of mortality but not of other endpoints Conclusion. In a well-characterized clinical trial population, an increased rate of prodromal decline in FVC was associated with an increased risk of respiratory death, hospitalization and categorically defined disease progression, but was not associated with an increased rate of subsequent FVC decline over the subsequent 52 weeks. Research Funding Source: Gilead Science, Inc.
- Health services