Background: The routine use of multidetector computed tomography has led to increased detection of unsuspected pulmonary embolism (UPE), with questionable benefit for diagnosis and treatment. Objective: The purpose of this work was to compare the clinical characteristics and prognosis of patients with UPE to patients with suspected PE (SPE). Methods: We retrospectively reviewed the charts of patients diagnosed with PE in a community-based university hospital between the years 2002 and 2007. UPE was defined as PE detected on CT scans performed for indications other than suspicion of PE. We compared patients with UPE to patients with SPE for differences in clinical features, electrocardiogram, imaging and echocardiographic findings. We also assessed the long-term outcomes using electronic patient records. Results: Of 500 patients with PE, 408 had SPE and 92 had UPE. Patients with UPE were similar to patients with SPE regarding age and sex distribution. Malignancy was more prevalent in UPE patients (39 vs. 23%, p < 0.0068). UPE patients had significantly less tachypnea (37 vs. 57%, p = 0.0005), dyspnea (47 vs. 87%, p < 0.0001), chest pain (19 vs. 42%, p < 0.0001) and hypoxemia (36 vs. 55%, p = 0.0011). Mortality was higher in UPE patients (70.3 vs. 53%, p = 0.0029). The hazard ratio after adjustment for confounders including age, sex and malignancy was 1.546 (95% CI: 1.139–2.099, p = 0.0052). Conclusions: We suggest that UPE is more prevalent in patients with a malignancy and is associated with higher mortality despite a less severe clinical presentation. UPE may be a marker of poor prognosis.