Purpose: The utility of arterial blood gas levels in excluding the diagnosis of acute pulmonary embolism (PE) was evaluated.
Methods: Data are from the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED). PE was diagnosed or excluded by pulmonary angiography. Among 330 patients with no prior cardiopulmonary disease, 130 patients had PE and 200 did not. Among 438 patients with prior cardiopulmonary disease, 147 had PE and 291 did not. Definitions were low PaO2 (< 80 mm Hg), low PaCO2 (< 35 mm Hg), and high alveolar-arterial oxygen gradient (P(A-a)O2 [> 20 mm Hg]).
Results: Among patients with no prior cardiopulmonary disease who had values of the PaO2 and PaCO2 that were not low and a P(A-a)O2 that was normal, 16 of 42 or 38% (95% confidence interval [CI] = 24 to 54%) had PE. Among patients with prior cardiopulmonary disease who had PaO2 and PaCO2 values that were not low and a P(A-a)O2 that was normal, 4 of 28 or 14% (95% CI = 4 to 33%) had PE. Other combinations of blood gas levels and the P(A-a)O2 gradient, failed to exclude PE in larger percentages of patients.
Conclusion: With various combinations of the PaO2 of 80 mm Hg or more, the PaCO2 of 35 mm Hg or higher, and the P(A-a)O2 gradient of 20 mm Hg or less, PE could not be excluded in more than 30% of patients with no prior cardiopulmonary disease and PE could not be excluded in more than 14% of patients with prior cardiopulmonary disease. Blood gas levels, therefore, are of insufficient discriminant value to permit exclusion of the diagnosis of PE.