Studies have demonstrated a lower incidence of complications after video-assisted thoracoscopic surgery (VATS) lobectomy compared with thoracotomy, but the data on in-hospital and 90-day mortality are inconclusive. This study analyzed whether surgical approach, VATS or thoracotomy, was related to early mortality of lobectomy in lung cancer and determined the differences between in-hospital and 90-day mortality. Data of all patients with non-small cell lung cancer who underwent lobectomy between January 1, 2007, and July 30, 2018, were retrieved from Polish National Lung Cancer Registry. Included were 31 433 patients who met all study criteria. After propensity score matching, 4946 patients in the VATS group were compared with 4946 patients in the thoracotomy group. Compared with thoracotomy, VATS lobectomy was related to lower in-hospital (1.5% vs 0.9%, P = .004) and 90-day mortality (3.4% vs 1.8%, P < .001). Mortality at 90 days was twice as high as in-hospital mortality in both the VATS (1.8% vs 0.9%, P < .001) and thoracotomy groups (3.4% vs 1.5%, P < .001). Postoperative complications were less common after VATS compared with thoracotomy (23.6% vs 31.8%, P < .001). VATS lobectomy is associated with lower in-hospital and 90-day mortality compared with thoracotomy and should be recommended for lung cancer treatment, if feasible. Patients should also be closely monitored after discharge from the hospital, because 90-day mortality is significant higher than in-hospital mortality. Display omitted