Pulmonary embolism (PE) is a common and lethal yet treatable condition. Several authors have reported on the diagnostic value of combinations of arterial blood gas (ABG) and other clinical data (i. e., prediction rules), and have claimed that these combinations can be safely used to exclude PE. The purpose of this investigation was to evaluate the diagnostic value of ABG measurement and to attempt to validate the ABG prediction rules published by these various authors for the assessment of patients with suspected PE. Two hundred ninety-three consecutive patients referred for imaging to investigate suspected PE were approached to participate in the investigation. ABG and other clinical data were obtained from consenting and eligible patients before an outcome classification (PE versus non-PE) was performed. None of the ABG data or prediction rules had sufficient negative predictive value, specificity, or likelihood ratios to be useful in the management of patients with suspected PE. We conclude that ABG data alone or in combination with other clinical data are not useful in the assessment of suspected PE.