In patients with chronic obstructive pulmonary disease (COPD), differentiating a pulmonary embolism (PE) from an exacerbation of COPD can be difficult, since clinical signs and symptoms of the two conditions overlap. Development of reliable noninvasive or minimally invasive techniques for the diagnosis of PE is, especially in these patients, necessary. In this study we assessed the effect of COPD on the accuracy of the clinical probability estimate (CPE), spiral computed tomographic angiography (SCTA), D-dimer analysis, ventilation perfusion (V/Q) scintigraphy, and pulmonary angiography for the diagnosis of PE. From May 1997 through March 1998, 627 consecutive patients with suspected PE were investigated in six teaching hospitals. In these patients, D-dimer testing, CPE, V/Q scintigraphy, and SCTA and/or pulmonary angiography were performed according to a strict diagnostic protocol. The patients were also independently categorized as having COPD or not. A diagnosis of COPD was established in 91 patients (15%). The prevalence of PE was similar in patients with and without COPD (29% and 31%, respectively), notwithstanding the larger proportion of nondiagnostic V/Q scan results in patients with COPD (46% versus 21%, p < 0.001). The distribution of CPEs, diagnostic value of the D-dimer assay and SCTA, and reproducibility of pulmonary angiography were comparable among patients with and without COPD. The presence of COPD does not affect the diagnostic performance of CPE, D-dimer testing, SCTA, or pulmonary angiography. Furthermore, although more nondiagnostic V/Q scan results can be expected in the presence of COPD, V/Q scintigraphy remains a valuable screening test in patients with COPD.