The predictors of in-flight hypoxia in interstitial lung disease remain unknown. The hypoxic challenge test (HCT) is widely used to evaluate fitness to fly. We assessed the determinants of a positive HCT in patients with ILD.
183 patients underwent HCT in 2005-2011 (74 males; mean age 57.6; 73 ex-smokers) of which 126 had idiopathic interstitial pneumonia, 26 hypersensitivity pneumonitis, 23 sarcoid and 8 other ILDs. 164 had lung function tests on the same day, 124 had an echocardiogram within 1 year. The HCT was performed using an inhaled gas mixture containing 15% oxygen. From BTS guidelines, the HCT was positive (in-flight oxygen required) if PaO2 < 6.6 kPa on 15% oxygen.
Median PaO2 on air was 10.3 kPa (range 8.2-14.3), with median percent predicted FVC 68.1% and DLCO 38.3%. On univariate logistic regression, variables associated with a positive HCT were age (p=0.04), PaO2 on air (p<0.0001), FVC%, DLCO% and composite physiologic index (CPI) (p for all<0.00001). Pulmonary hypertension (PH) on echocardiogram was also predictive, although less strongly (p=0.015). On multivariate analysis, variables remaining significantly associated with a positive HCT were PaO2 on air (p<0.0001), lung function markers including CPI (p<0.0001), or in separate models DLCO (p<0.0001), and FVC (p=0.006), while PH on echo was no longer predictive (p=0.13). On ROC analysis, area under the curve was 0.80 for PaO2 on air, 0.75 for DLCO%, 0.72 for CPI, 0.83 for combined CPI and PaO2, and 0.85 for combined DLCO and PaO2.
Our findings highlight the potential of PaO2 on air, DLCO and CPI levels as non-invasive predictors of fitness to fly. Identifying the best combination requires further prospective evaluation.