Arteriovenous malformations (AVMs) are the leading cause of intracerebral haemorrhage in young adults, but are also diagnosed on brain imaging without having bled when they cause seizures, focal neurological deficits, or present as incidental findings. In such unruptured cases, the long-term risk of intracerebral haemorrhage might be reduced by interventional treatment to embolise or excise the arteriovenous malformation and to treat any associated aneurysms, but evidence of the balance of risks and benefits in randomised trials has been lacking. A Randomized Trial of Unruptured Brain Arteriovenous malformations (ARUBA) was a non-blinded, randomised, parallel-group trial in adults with an unruptured arteriovenous malformation, who had never undergone interventional therapy, and were considered suitable for intervention to eradicate the lesion. The trial compared medical management alone with medical management and interventional therapy (neurosurgery, embolisation, or stereotactic radiotherapy, alone or in combination). 1 Of a planned sample size of 400 (reduced from 800 owing to slow recruitment), 226 patients were randomly assigned at 39 centres in nine countries before the trial was halted by the data and safety monitoring board (DSMB), after a mean follow-up of 33· 3 months, when a prespecified interim analysis showed that medical management alone was superior to interventional therapy in preventing stroke or death. 1

Follow-up of ARUBA to 5 years was specified in the protocol and it had been the intention of the study investigators to try to extend the follow-up to 10 years. However, the National Institute of Neurological Disorders and Stroke (NINDS) declined to fund applications for further follow-up on two occasions, despite the NINDS-appointed trial DSMB concluding in their April 15, 2013, report, which stopped recruitment, that:“There is a compelling need for additional long-term data; therefore extended follow-up will be necessary to determine whether the disparity in event rates will persist over time.” The study investigators did manage to get some funding from another source, but only for a further 2 years of follow-up. In The Lancet Neurology, Jay Mohr and colleagues2 now report the findings from this extended short-term follow-up (mean 50· 4 months