Crossover trial design

Crossover trial designs (Figure 31.2) are less often used in clinical practice and are implemented typically only in Phase 1 or 2 of development. Examples of these trials, such as bioequivalence and drug-drug interaction, commonly appear in the literature. In fact, numerous journal articles and textbooks are devoted to the formulation and analysis of these designs. In these trials, patients are allocated randomly to study arms (or sequences). Each patient within the sequence receives a treatment followed by one or more treatments these treatments can be replicated. The treatments can consist of single or multiple doses. 

Controlled trials can be conducted under matched pairs, parallel, crossover, group comparisons, or mixed design conditions. Parallel and crossover trial designs each have advantages and disadvantages. Both can be used to compare two or more treatments, one of which may be placebo. In a crossover trial, all treatments to be compared are administered to every enrolled subject in a carefully designed and blinded sequence with an interim drug washout period. Each subject receives all treatments and thus serves as his or her own control. 

Crossover design A clinical trial design in which patients receive, in sequence, the treatment (or the control), and then after a specified time, receive the opposite arm of the trial. This allows patients to serve as their own controls. Randomization should be... 

Finally, trials can follow parallel or crossover study designs. In a parallel trial, patients are assigned to a therapy that they remain on, and they are compared with patients in alternate therapy groups. In a crossover trial, patients switch or change therapy assignments during the course of the trial. 

In some instances it may be better to design the trial so that each patient provides his own control - by having various treatments in turn. This is known as a "crossover " trial By such means we may sometimes make... 

Dextromethorphan for the treatment of neuropathic pain a double-blind randomised controlled crossover trial with integral n-of-1 design, Pain 1994, 59, 127-133. 

Crossover designs, however, have their own deficiencies. Assessment of safety endpoints (when they are not known to be directly related to the magnitude of drug concentrations) is difficult conventions such as intent-to-treat analyses are not available. Moreover, if each treatment period is lengthy, then the time required to complete a crossover study could be prohibitive. In addition, lengthy crossover trials could lead to increased patient dropout prior to the subsequent treatment period, thereby reducing the efficiency of the trial. 

CONTAK-CD (10) 2003 490 NYHA II to IV LVEF <35% QRS >120 NSR ICD indication Single blind, originated as a crossover trial but changed midway to a parallel design No significant change in a composite of mortality, CHF hospitalization, and ventricular tachyarrhythmias requiring device therapy at 3 to 6 months... 

Jones, B. and Kenward, M.G. 1989. Design and Analysis of Crossover Trials. New York, Chapman and Hall. 

Erlotinib (Tarceva) took a differenf approach to a pivotal trial. In the study of Tarceva for patients with advanced, previously treated non-small cell lung cancer, where survival was the primary endpoint, patients were randomized to receive either Tarceva or placebo. Tarceva clearly prolonged overall survival (6.7 vs. 4-7 months, p = 0.001). In addition, Tarceva improved progression-free survival, and improved fime-fo-deferiora-tion of patients reported symptoms (cough, dyspnea, and pain). Obviously, this was a definitive trial (drug vs. placebo) however, it could be a controversial trial as it was a survival trial with no provisions to crossover to the active Tarceva. It is doubtful fhaf many trials with this noncrossover design (with a survival endpoint) will be done in the future. 

---