BACKGROUND: Uveitis associated with juvenile idiopathic arthritis is a cause of major ocular morbidity. A substantial proportion of children are refractory to systemic methotrexate and TNF inhibitors. Our aim was to study the safety and efficacy of tocilizumab in children with juvenile idiopathic arthritis-associated uveitis refractory to both methotrexate and TNF inhibitors. METHODS: This multicentre, single-arm, phase 2 trial was done following a Simon's two-stage design at seven tertiary hospital sites in the UK. Patients aged 2-18 years with active juvenile idiopathic arthritis-associated uveitis were eligible. All patients had been on a stable dose of methotrexate for at least 12 weeks and had not responded to treatment with a TNF inhibitor. Patients weighing 30 kg or more were treated with 162 mg subcutaneous tocilizumab every 2 weeks for 24 weeks, and participants weighing less than 30 kg were treated with 162 mg every 3 weeks for 24 weeks. The primary outcome was treatment response defined as a two-step decrease, or decrease to zero, from baseline in the level of inflammation (anterior chamber cells) at week 12, per the standardisation of uveitis nomenclature criteria. A phase 3 trial would be justified if more than seven patients responded to treatment. An interim analysis was planned to assess whether the trial would be stopped for futility, with futility defined as two or fewer treatment responses among ten participants. Adverse events were collected up to 30 calendar days after treatment cessation. The primary analysis was done in the intention-to-treat population and the safety analysis was done in all patients who started the treatment. This trial is registered with the International Standard Randomised Controlled Trial Number registry (ISRCTN95363507) and EU Clinical Trials Register (EudraCT 2015-001323-23). FINDINGS: 22 participants were enrolled to the trial between Dec 3, 2015, and March 9, 2018, and 21 participants received treatment. One participant was found to be ineligible immediately after enrolment and was therefore withdrawn. Seven of 21 (median unbiased estimate of proportion 34% [95% CI 25-57]) responded to treatment (p=0·11). Safety results were consistent with the known safety profile of tocilizumab. INTERPRETATION: The primary endpoint was not met, and thus the results do not support a phase 3 trial of tocilizumab in patients with juvenile idiopathic arthritis-associated uveitis. Importantly, data on the use of tocilizumab in clinical practice is now captured in national registries. Despite this trial not meeting the threshold required to justify a larger phase 3 trial, several patients responded to treatment; as such, tocilzumab might still be a therapeutic option in some children with uveitis refractory to anti-TNF drugs, given the absence of other treatment options. FUNDING: Versus Arthritis and the National Institute for Health Research Clinical Research Network: Children.