Risk registers that get reviewed, not just written

A risk register written at kick-off and opened again at the post-project review is a document, not a control. The common failure is size: forty risks, each with an owner who was assigned rather than consulted, and a score that has not moved since week one. Smaller registers get read. So do ones where each review starts by closing what is no longer credible rather than adding more. What keeps yours alive through delivery, and does anyone outside the project manager read it?