In the phase 3 CELESTIAL trial, cabozantinib improved overall survival (OS) and progression-free survival (PFS) compared with placebo in patients with previously treated advanced hepatocellular carcinoma (HCC). This subgroup analysis evaluated cabozantinib in patients who had received sorafenib as the only prior systemic therapy. CELESTIAL randomised (2:1) patients with advanced HCC and Child-Pugh class A liver function to treatment with cabozantinib (60 mg daily) or placebo. Eligibility required prior treatment with sorafenib, and patients could have received ≤2 prior systemic regimens. The primary endpoint was OS. Outcomes in patients who had received sorafenib as the only prior therapy were analysed by duration of prior sorafenib (<3 months, 3 to <6 months and ≥6 months). Of patients who had received only prior sorafenib, 331 were randomised to cabozantinib and 164 to placebo; 136 patients had received sorafenib for <3 months, 141 for 3 to <6 months and 217 for ≥6 months. Cabozantinib improved OS relative to placebo in the overall second-line population who had received only prior sorafenib (median 11.3 vs 7.2 months; HR=0.70, 95% CI 0.55 to 0.88). This improvement was maintained in analyses by prior sorafenib duration with longer duration generally corresponding to longer median OS-median OS 8.9 vs 6.9 months (HR=0.72, 95% CI 0.47 to 1.10) for prior sorafenib <3 months, 11.5 vs 6.5 months (HR=0.65, 95% CI 0.43 to 1.00) for 3 to <6 months and 12.3 vs 9.2 months (HR=0.82, 95% CI 0.58 to 1.16) for ≥6 months. Cabozantinib also improved PFS in all duration subgroups. Safety data were consistent with the overall study population. Cabozantinib improved efficacy outcomes versus placebo in the second-line population who had received only prior sorafenib irrespective of duration of prior sorafenib treatment, further supporting the utility of cabozantinib in the evolving treatment landscape of HCC. NCT01908426.