Design issues in chemosensory trials.

Randomized clinical trials in taste and smell disorders have been infrequent, in part because of difficulties in both the design and the conduct of such trials. Hyposmia and hypogeusia, impairments of smell and taste, respectively, have a variety of causes, such as upper respiratory infection, head trauma, or laryngectomy. Once definitive diagnosis is established, eligibility and exclusion criteria may relate to etiology. Since the sense of smell is more acute at younger ages and since women can generally identify odors more accurately than men, either prestratification or poststratification should be considered. A control or comparison group is essential because, occasionally, chemosenses spontaneously return to normal. Patient recruitment may be difficult, depending to a large extent on self-referrals, with the exception of head injury cases. To enroll an adequate number of patients in a reasonable length of time usually requires participation of multiple clinical centers. Objective measurements of taste and smell must be reliably obtained before and after intervention. Design issues specific to chemosensory trials are discussed in the context of an example, a factorial design trial of surgery and steroids in patients with olfactory defects.