Patient demographics and medical history were abstracted from medical records. Descriptive data are reported as frequencies, means with SDs, or medians with ranges. Groups were compared using a χ2 test of association, independent t tests, Mann-Whitney U test, or Fisher exact test depending on level of measurement and distribution of data. Variables included in bivariate analysis to assess association with GERD included age, gender, FEV1, FVC, TLC, residual volume, diffusing capacity of the lung for carbon monoxide (Dlco), mean LES pressure, presence of decreased LES pressure, presence of decreased peristalsis, body mass index (BMI), prednisone use, theophylline use, proton-pump inhibitor use, histamine type-2 blocker use, symptoms of heartburn and/or acid regurgitation during the study, and diabetes status. Variables with a p value ≤ 0.1 on bivariate analysis were included in multivariate logistic regression analysis. Two models were generated, one with all candidate variables and one utilizing stepwise regression. The appropriateness of the models was assessed using Hosemer-Lemeshow goodness-of-fit test. There was no indication that the data did not fit the models. Analysis was performed using statistical software (SPSS v13.0; SPSS; Chicago, IL); p < 0.05 was considered significant.