Background:The quantification of hypoxia based on resting partial pressure of arterial oxygen (PaO2) may underestimate hypoxia related to activities of daily living or sleep and thus not accurately reflect pulmonary hypertension (PH). The aim of the present study was to investigate the association of resting PaO2 with percent time of SpO2 below 90% (T90) and 88% (T88) in 24 hours. We also evaluated the capacity of hypoxia measures to predict PH in patients with idiopathic pulmonary fibrosis (IPF).Method:This cross-sectional study included 27 patients with IPF presenting PaO2 ≥ 55 mmHg and not receiving home oxygen therapy. All were submitted to blood gas measurement, 24-h oximetry, and transthoracic Doppler echocardiography to estimate systolic pulmonary artery pressure (SPAP). Patients were divided into three groups according to resting PaO2: 55-55.9 mmHg (A); 60-60.9 mmHg (B); ≥ 70 mmHg (C). PH was defined as “likely” if SPAP > 50 mmHg, and as possible for SPAP between 37 and 50 mmHg.Results:T90 and T88 in Groups A, B, and C were as follows: 59.9±29% and 44.1±34%; 49.3±34% and 29.9±31%; 17.1±25% and 8.8±18% respectively, with significant differences between the groups for both T90 (p ≤ 0.01) and T88 (p = 0.02). PaO2 was inversely correlated with T90 (r = -0.398; p = 0.04) and T88 (r = -0.351; p = 0.07). Hypoxia variables did not correlate with SPAP, and were not able to predict PH.Conclusion:Percent time of SpO2 below 90% and 88% in 24 hours revealed periods of severe hypoxia even in patients with borderline-normal resting PaO2. However, none of the present hypoxia variables was capable of predicting PH.