OBJECTIVE Acute pulmonary embolism (PE) increases alveolar dead space, which can dilute carbon dioxide content in exhaled breath. This study was undertaken to determine whether the capnogram waveform area from patients with PE is decreased compared with that from patients without PE and to examine the potential role of the capnogram waveform area as a screening test for PE. METHODS Studies were conducted on patients from a large, urban emergency department. Steady-state capnograms were collected from subjects breathing room air, and data were electronically stored for later analysis; arterial blood was collected for blood gas analysis. PE was diagnosed or excluded on the basis of appropriate combinations of isotopic ventilation-perfusion lung scanning, lower-extremity venous Doppler ultrasound studies, and pulmonary angiography. Capnogram areas were measured by tracing waveforms on printed copies with an electronic digitizing tablet. RESULTS The mean capnogram area from patients with PE (n=19) was 28+/-10 mm Hg x sec, significantly less than for patients without PE (n=120), 53+/-17 mm Hg x sec (95% confidence interval [CI] for a difference of 25 mm Hg x sec, 17 to 33 mm Hg x sec). At a cutoff of 50 mm Hg x sec, the test sensitivity was 100% (95% CI, 82% to 100%), with a negative likelihood ratio (LR-) of .17, and the specificity was 53% (95% CI, 44% to 62%), with a positive likelihood ratio (LR+) of 2.1. At a cutoff of 25 mm Hg x sec, the test sensitivity was 42% (95% CI, 20% to 67%; LR-, .6) and the specificity was 97% (95% CI, 92% to 99%; LR+, 12.3). The area under the smoothed receiver operating characteristic curve for the capnogram area was .896 (95% Cl, .80 to .99). CONCLUSIONS The capnogram waveform area may be useful in screening for PE in the ED.