RELATIONSHIP BETWEEN EXERCISE-INDUCED OXYGEN DESATURATION AND DLCO IN PATIENTS WITH ADVANCED LUNG DISEASE

Abstract

SESSION TITLE: Diffuse Lung Disease SESSION TYPE: Original Investigation Posters PRESENTED ON: 10/10/2018 01:00 pm - 02:00 pm PURPOSE: Exercise induced desaturation (EID) is associated with increased morbidity in patients with advanced lung disease. EID predictors have been well studied in patients with COPD, but not in other lung disorders. The aim was to determine whether the percent-of-predicted diffusion capacity of the lung for carbon monoxide (DLCO) could be used to predict desaturation on exercise oximetry in patients with advanced lung disease. We specifically studied patients with interstitial lung disease (ILD) or pulmonary hypertension. METHODS: This was a retrospective chart review on patients referred to our Advanced Lung Disease Center. Patient demographics, WHO functional class, and oxygen saturation during exercise oximetry were reviewed. The DLCO for each patient was corrected for age, gender, and height by expressing DLCO as a percentage of the predicted value using the Miller reference equation (Miller et al. AJRCCM 1983). Data was analyzed in SAS®.All patients underwent treadmill exercise oximetry. Based on exercise oximetry, two groups were defined: desaturators (DS) and non-desaturators (NDS). A patient was considered a DS if there was a drop in standard pulse oximetry (SpO2) ≥ 4%, or a nadir ≤ 88% requiring supplemental oxygen. RESULTS: There were 86 patients (30 males and 56 females). The mean age was 60.42 years (SD 15.7). 22 patients had pulmonary arterial hypertension, 23 patients had Group II/III/IV pulmonary hypertension, 27 had ILD, and 14 had other diagnoses. There were 49 desaturators with a mean percent-of-predicted DLCO of 45.8% (SD 16.1). There were 37 non-desaturators with a mean percent-of-predicted DLCO of 57.5 % (SD 19.3). Logistic regression was used to examine whether increasing percent-of-predicted DLCO was associated with not desaturating during the exercise oximetry. There was a significant association between percent-of-predicted DLCO and desaturation (p<0.0059). For a 1 unit increase in percent-of-predicted DLCO, the Odds Ratio of not desaturating was 1.040 (95% confidence interval: 1.011–1.069); for a 5 unit increase, the Odds Ratio was 1.215 (95% CI: 1.058–1.396). The area under the ROC curve (AUC) was 0.672 (95% CI: 0.558–0.786). An optimal cut-off point of percent-of-predicted DLCO ≥ 50 was determined using the minimum Euclidean distance from the ROC curve to (0,1). CONCLUSIONS: Impaired gas exchange is the most common functional abnormality in patients with advanced lung diseases. Oxygen desaturation occurs only during exercise. As the disease progresses, desaturation will occur at rest. The percent-of-predicted DLCO was associated with underlying EID in patients with ILD or pulmonary hypertension. CLINICAL IMPLICATIONS: A low percent-of-predicted DLCO of <50 is a good predictor of EID in patients with ILD and/or pulmonary hypertension. Future research should be carried out to validate this cut point on an independent sample. DISCLOSURES: No relevant relationships by Zubair Hasan, source=Web Response No relevant relationships by Nina Kohn, source=Web Response No relevant relationships by Shilpa Malik, source=Web Response No relevant relationships by Ronak Shah, source=Web Response No relevant relationships by Arunabh Talwar, source=Web Response No relevant relationships by Sameer Verma, source=Web Response No relevant relationships by Abhinav Vulisha, source=Web Response