Lobectomy is not advocated for papillary thyroid carcinoma (PTC) with high-risk features, although there is no high-level evidence showing that this is an inferior strategy. This study aimed to examine the association between the extent of surgery and survival of patients with PTC and high-risk features. Consecutive patients with PTC and at least one high-risk feature treated in 2000-2012 were included in the study. High-risk features were defined as: primary tumour larger than 4 cm, gross extrathyroidal extension, macroscopic multifocality, and confirmed nodal metastasis including pathological lateral neck metastasis (pN1b) or more than five central lymph node metastases. Cox proportional hazards models were employed to measure the association between the extent of surgery and disease-specific survival (DSS) in the whole cohort and in a matched-pair analysis. Among a total of 2059 patients with high-risk features, 1224 underwent lobectomy and 835 had total thyroidectomy. Patients who underwent total thyroidectomy had significantly higher rates of bilateral cancer than those who had a lobectomy (79.4 versus 2.7 per cent respectively), macroscopic multifocality (80.8 versus 32.8 per cent) and bilateral neck metastasis (30.9 versus 3.3 per cent) (all P < 0.001). With a median follow-up of 93 months, multivariable analysis showed that the extent of surgery was not associated with DSS in the whole cohort (hazard ratio 1.36, 95 per cent c.i. 0.75 to 2.48; P = 0.310). After 1 : 1 case-control matching of 528 patients, no significant difference between lobectomy and total thyroidectomy groups was observed with respect to the 10-year DSS rate (94.3 versus 95.2 per cent respectively; P = 0.323) or 10-year recurrence-free survival rate (75.8 versus 79.2 per cent; P = 0.784). Lobectomy was not associated with significantly worse outcomes for patients with PTC and high-risk features.