Background: The ATLAS trial compared axitinib versus placebo in patients with locoregional renal cell carcinoma (RCC) at risk of recurrence after nephrectomy. Patients and methods: In a phase III, randomized, double-blind trial, patients had >50% clear-cell RCC, had undergone nephrectomy, and had no evidence of macroscopic residual or metastatic disease [independent review committee (IRC) confirmed]. The intent-to-treat population included all randomized patients [pT2 and/or Nþ, any Fuhrman grade (FG), Eastern Cooperative Oncology Group status 0/1]. Patients (stratified by risk group/country) received (1: 1) oral twice-daily axitinib 5 mg or placebo for 3 years, with a 1-year minimum unless recurrence, occurrence of second primary malignancy, significant toxicity, or consent withdrawal. The primary end point was disease-free survival (DFS) per IRC. A prespecified DFS analysis in the highest-risk subpopulation (pT3, FG 3 or pT4 and/or Nþ, any T, any FG) was conducted. Results: A total of 724 patients (363 versus 361, axitinib versus placebo) were randomized from 8 May 2012, to 1 July 2016. The trial was stopped due to futility at a preplanned interim analysis at 203 DFS events. There was no significant difference in DFS per IRC [hazard ratio (HR) ¼ 0.870; 95% confidence interval (CI): 0.660-1.147; P ¼ 0.3211). In the highest-risk subpopulation, a 36% and 27% reduction in risk of a DFS event (HR; 95% CI) was observed per investigator (0.641; 0.468-0.879; P ¼ 0.0051), and by IRC (0.735; 0.525-1.028; P ¼ 0.0704), respectively. Overall survival data were not mature. Similar adverse events (AEs; 99% versus 92%) and serious AEs (19% versus 14%), but more grade 3/4 AEs (61% versus 30%) were reported for axitinib versus placebo. Conclusions: ATLAS did not meet its primary end point; however, improvement in DFS per investigator was seen in the highest-risk subpopulation. No new safety signals were reported. Trial registration number: NCT01599754.