The safety and efficacy of neratinib monotherapy were compared with that of lapatinib plus capecitabine in patients with human epidermal growth factor receptor-2-positive (HER2+), locally advanced/metastatic breast cancer and prior trastuzumab treatment.Patients received neratinib 240mg/d continuously (n=117) or lapatinib 1250mg/d continuously plus capecitabine 2000mg/m 2 per day on days 1–14 of each 21-d cycle (n=116). The primary aim was to demonstrate non-inferiority of neratinib for progression-free survival (PFS).The non-inferiority of neratinib was not demonstrated when compared with lapatinib plus capecitabine (hazard ratio, 1.19; 95% confidence interval, 0.89–1.60; non-inferiority margin, 1.15). Median PFS for neratinib was 4.5months versus 6.8months for lapatinib plus capecitabine and median overall survival was 19.7months versus 23.6months. Objective response rate (neratinib, 29% versus lapatinib plus capecitabine, 41%; P=0.067) and clinical benefit rate (44% versus 64%; P=0.003) were lower for the neratinib arm but consistent with previously reported results. In both treatment arms, diarrhoea was the most frequently reported treatment-related adverse event of any grade (neratinib, 85% versus lapatinib plus capecitabine, 68%; P=0.002) and of grade 3/4 (28% versus 10%; P<0.001), but was typically managed with concomitant anti-diarrhoeal medication and/or study treatment modification. Importantly, neratinib had no significant skin toxicity.The results are considered as inconclusive since neither inferiority nor non-inferiority of treatment with neratinib versus lapatinib plus capecitabine could be demonstrated. The study confirmed relevant single-agent clinical activity and acceptable overall tolerability of neratinib in patients with recurrent HER2+ advanced breast cancer.