In the absence of a randomized controlled trial, it is important to obtain as much evidence as possible by other methods on whether inferior vena cava (IVC) filters reduce mortality in patients who undergo pulmonary embolectomy. Therefore, this retrospective cohort study based data from the National Inpatient Sample 2009 to 2014 was undertaken. We assessed in-hospital all-cause mortality in stable and unstable (in shock or on ventilator support) patients with acute pulmonary embolism who underwent pulmonary embolectomy. International Classification of Diseases-9-Clinical Modification (ICD-9-CM) codes were used to identify patients. Co-morbidities were assessed by the updated Charlson co-morbidity index. A time-dependent analysis was performed to control for immortal time bias. In stable patients who underwent pulmonary embolectomy, mortality with an IVC filter was 50 of 1,212 (4.1%) compared with 202 of 755 (27%) with no IVC filter (p <0.0001). In unstable patients, mortality with an IVC filter was 108 of 598 (18%) compared with 179 of 358 (50%) with no IVC filter (p <0.0001). Mortality was reduced with IVC filters only if the filters were inserted in the first 4 or 5 days. Co-morbid conditions and immortal time bias could not explain these results. We conclude that both stable and unstable patients who underwent pulmonary embolectomy had a lower mortality with IVC filters if inserted in the first 4 or 5 days.