SESSION TITLE: Interstitial Lung Disease
SESSION TYPE: Original Investigation Slide
PRESENTED ON: Wednesday, October 30, 2013 at 02:45 PM - 04:15 PM
PURPOSE: Current guidelines from the American Thoracic Society (ATS) recommend the use of FEV1 to grade the severity of restrictive lung diseases such as idiopathic pulmonary fibrosis (IPF), while the majority of validated composite prediction scores utilize FVC. Our goal was to compare the performance characteristics of FEV1 to FVC in a large cohort of patients with IPF.
METHODS: We performed a review of all patients diagnosed with IPF who were seen at the Inova Fairfax Hospital Interstitial Lung Disease Clinic between January 2000 and November 2012. Area Under the Receiver Operator Characteristic (AUROC) curve was calculated for both FVC and FEV1. Survival was evaluated using Kaplan-Meier curves. Variables significant in univariate analysis were entered into models using Cox proportional hazard regression after assessment for co-linearity. Model performance was assessed using the c-statistic.
RESULTS: We prospectively followed 556 patients referred to an advanced lung disease clinic for management of IPF. Mean age and BMI were 66.4±10.2 and 29.3±6.2, respectively. AUROC for FVC and FEV1 were 0.602 (p=0.001) and 0.604 (p<0.001), respectively, and grading of severity for both variables was significantly associated with mortality in Kaplan-Meier analysis (log-rank 0.002 for FVC and <0.001 for FEV1). In univariate analysis, age, BMI, DLCO, FVC, FEV1, and FEV1/FVC were significantly associated with mortality. After adjustment using Cox-regression, all variables except FEV1/FVC remained significant. FEV1 and FVC were entered into the model individually, as they were highly correlated at baseline. Overall model performance was excellent and was not significantly changed by the inclusion of FEV1 or FVC (AUROC 0.720 [p<0.001] with FEV1 and 0.724 [p<0.001] with FVC).
CONCLUSIONS: In a very large cohort of IPF patients, FVC and FEV1 demonstrated similar performance characteristics. Our multivariate models performed comparably and demonstrated similar predictive capacity as existing more complicated composite prediction scores, such as ROSE.
CLINICAL IMPLICATIONS: Given the similar predictive capacities of FVC and FEV1, our findings would support the use of FEV1 or FVC for predicting survival in patients with IPF.
DISCLOSURE: The following authors have nothing to disclose: Robert Walter, Aaron Holley, Steven Nathan
No Product/Research Disclosure Information