The "nondipper" elderly: a clinical entity or a bias?