Abstract OBJECTIVES This study compares the uniportal with the 3-portal video-assisted thoracic surgery (VATS) by examining the data collected in the Italian VATS Group Database. The primary end point was early postoperative pain; secondary end points were intraoperative and postoperative complications, surgical time, number of dissected lymph nodes and length of stay. METHODS This was an observational, retrospective, cohort, multicentre study on data collected by 49 Italian thoracic units. Inclusion criteria were clinical stage I–II non-small-cell lung cancer, uniportal or 3-portal VATS lobectomy and R0 resection. Exclusion criteria were cT3 disease, previous thoracic malignancy, induction therapy, significant comorbidities and conversion to other techniques. The pain parameter was dichotomized: the numeric rating scale ≤3 described mild pain, whereas the numeric rating scale score >3 described moderate/severe pain. The propensity score-adjusted generalized estimating equation was used to compare the uniportal with 3-portal lobectomy. RESULTS Among 4338 patients enrolled from January 2014 to July 2017, 1980 met the inclusion criteria; 1808 patients underwent 3-portal lobectomy and 172 uniportal surgery. The adjusted generalized estimating equation regression model using the propensity score showed that over time pain decreased in both groups (P < 0.001). There was a statistical difference on the second and third postoperative days; odds ratio (OR) 2.28 [95% confidence interval (CI) 1.62–3.21; P < 0.001] and OR 2.58 (95% CI 1.74–3.83; P < 0.001), respectively. The uniportal-VATS group had higher operative time (P < 0.001), shorter chest drain permanence (P < 0.001) and shorter length of stay (P < 0.001). CONCLUSIONS Data from the Italian VATS Group Database showed that in clinical practice uniportal lobectomy seems to entail a higher risk of moderate/severe pain on second and third postoperative days.