Background This study investigated the utility of the alveolar - arterial oxygen pressure difference (AaDO ₂) in predicting the short-term prognosis of acute pulmonary embolism (PE). Methods and Results This study retrospectively enrolled 114 consecutive patients with acute PE, diagnosed by either spiral computed tomography or high probability ventilation - perfusion lung scans. During the first 24 h of admission, all patients had initial artery blood gas collected under room air. Patient exclusion criteria were chronic lung disease, septic emboli, and moderate and low probability lung scans. Patients were assigned to 2 groups based on either 30-day death or a 30-day composite event. Receiver operating characteristic analyses was used to determine the AaDO₂ cut-off value for predicting primary and composite endpoints. Statistical analysis demonstrated significant differences in AaDO₂ between the 30-day composite endpoint group and the 30-day composite event-free survival group (p=0.012). The AaDO₂ had a strong trend between the 30-day death group and the survival group (p=0.062). The best cut-off value for AaDO₂ was 53 mmHg and using this, the positive predictive value for 30-day death was 25% and the negative predictive value was 92%. For the 30-day composite endpoint, the positive predictive value for AaDO₂ was 35%, and the negative predictive value was 84%. In this study, thrombocytopenia was also an indicator of poor prognosis for patients with acute PE. Conclusion The AaDO₂ measurement is a highly useful and simple measurement for predicting short-term prognosis in patients with acute PE. It has high negative predictive value and moderate positive predictive value for 30-day death and 30-day composite event. Aggressive thrombolytic treatment strategies should be considered for patients with an initial poor prognostic parameter (ie, AaDO₂ ≥53 mmHg). (Circ J 2006; 70: 1611 - 1616)