Preventive treatment of unruptured intracranial aneurysms is often performed but has never been proved beneficial as compared to conservative management. In a context of uncertainty, the 'best treatment' that can be offered to each individual is a chance to be treated and thus to be protected from rupture of the aneurysm, and an equal chance not to be treated, and hence to be exempted from possible immediate complications, using randomization. Such action is optimal unless or until an independent committee with privileged access to data judges that, given the comparative outcome of the 2 groups, preventive treatment or conservative management, is generally warranted. Potential reasons to interrupt such a study are reviewed, including insufficient recruitment, poor compliance, excessive cross-overs, unacceptable iatrogenia, and treatments being convincingly different or equivalent. We conclude that insufficient recruitment is the sole realistic event that could lead to premature interruption. This review may provide a deeper understanding of the principles justifying the necessity of the study.