Is it worthwhile to screen patients with type 2 diabetes mellitus for subclinical Cushing's syndrome?