Drug*placebo interaction effect may bias clinical trials interpretation: hybrid balanced placebo and randomized placebo-controlled design

Abstract

Background: Conventional randomized placebo-controlled study design assumes the absence of drug*placebo interaction. We hypothesized the presence of such an interaction and that conventionally estimated drug effect might be biased. The objectives of the study were to determine the drug*placebo interaction effect (main) and compare conventionally estimated and interaction model-estimated drug effects (secondary).

Methods: We used a hybrid of balanced placebo and randomized placebo-controlled designs. Four hundred eighty healthy volunteers were randomized to three groups. The first received hydroxyzine (25 mg) described as hydroxyzine or placebo, the second received placebo described as hydroxyzine or placebo, and the third received hydroxyzine and placebo described as unknown; each in a randomized crossover design. Seven participants failed to crossover. Group assignment was concealed from participants and study coordinators. Coordinators were blinded to group and intervention assignment. Participants and coordinators were deceived as to study objectives. Main outcomes were mean area-under-the-curve of drowsiness (therapeutic outcome) and mouth-dryness (adverse outcome), self-reported on 100 mm visual analog scale over 7 h. Drug, placebo, placebo + interaction, and total effects were estimated using analysis of covariance by comparing received hydroxyzine/told placebo to received placebo/told placebo, received placebo/told hydroxyzine to received placebo/told placebo, received hydroxyzine/told hydroxyzine to received hydroxyzine/told placebo, and received hydroxyzine/told hydroxyzine to received placebo/told placebo, respectively. Drug effect was also conventionally estimated in the third group.

Results: Mean (SD) age was 31.4 (6.6) years, 65% were males. There was significant difference between placebo + interaction effect and placebo effect for both drowsiness and mouth-dryness with a mean difference (95% confidence interval) of 35.1 (5.6 to 64.6) and 23.8 (2.4 to 45.2) mm*hr, respectively. Total effect was larger than the sum of drug and placebo effects for drowsiness (139.7 (109.8 to 169.6) vs. 99.1 (68.2 to 130.0) mm*hr) and mouth-dryness (63.6 (41.1 to 86.1) vs. 34.7 (11.1 to 58.4) mm*hr). Conventionally estimated drug effect was larger than interaction model-estimated drug effect for drowsiness (69.2 (45.5 to 92.8) vs. (58.3 (31.6 to 85.0) mm*hr) and mouth-dryness (19.9 (5.3 to 34.5) vs. 9.5 (-9.2 to 28.1) mm*hr).

Conclusions: There is significant and important drug*placebo interaction effect that may bias conventionally estimated drug effect.

Trial registration: ClinicalTrial.gov identifier: NCT01501591 (registered December 25, 2011).

Keywords: Balanced placebo design; Biased drug effect size; Drug placebo interaction; Effect modification; Randomized placebo-controlled trials.

Figures

Fig. 1

Models of Positive and Negative Drug*Placebo Interaction Effect in Balanced Placebo and Conventional Randomized Placebo-Controlled Designs. a and b depict balanced placebo design. c and d depict conventional randomized placebo-controlled design. A positive drug*placebo interaction model is presented in a and c. A negative drug*placebo interaction model is presented in b and d. Hydroxyzine is used as the model drug. a & b Under the balanced placebo design, ooutcome measures in received placebo/told placebo represent none-specific passive changes (baseline); in received placebo/told hydroxyzine, placebo effect and baseline; in received hydroxyzine/told placebo, drug effect and baseline; and in received hydroxyzine/told hydroxyzine, drug, placebo, drug*placebo interaction effects, and baseline. The total effect of ingesting a hydroxyzine capsule is the difference between received hydroxyzine/told hydroxyzine and received placebo/told placebo and consists of drug, placebo, and drug*placebo interaction effects. The interaction model-estimated drug effect is the difference between received hydroxyzine/told placebo and received placebo/told placebo and consists of the drug effect only. The model predicts two different estimates of the placebo effect: placebo effect-1, the difference between received hydroxyzine/told hydroxyzine and received hydroxyzine/told placebo, which consists of the placebo and the drug*placebo interaction effects; and placebo effect-2, the difference between received placebo/told hydroxyzine and received placebo/told placebo, which consists of the placebo effect only. Placebo effect-1 is larger than placebo effect-2 if there is positive interaction effect (a) and smaller if there is negative interaction effect (b). c & d Under the conventional randomized placebo-controlled design, outcome measures in received hydroxyzine/told unknown represent the drug effect, part of the placebo effect (because of participants’ knowledge that there is 50% chance of receiving hydroxyzine) and consequently part of the drug*placebo interaction effect, and baseline; and in received placebo/told unknown, part of the placebo effect (because of participants’ knowledge that there is 50% chance of hydroxyzine) and baseline. Conventionally estimated drug effect is the difference between received hydroxyzine/told unknown and received placebo/told unknown and consists of the drug effect and part of the interaction effect. It is larger than the interaction model-estimated drug effect if there is positive interaction effect (compare a and c) and smaller if there is negative interaction effect (compare b and d)

Fig. 2

Flow of Participants through the Study

Fig. 3

Time Course of Self-Reported Drowsiness and Mouth-dryness According to Intervention or Type of Effect. a to d mean unadjusted scores on continuous scale (a and b) and binary scale (c and d) per intervention. Squares indicate receiving 25 mg hydroxyzine, described as hydroxyzine (closed square with continuous line), as placebo (open square with interrupted line), or as unknown (closed square with dotted line). Circles indicate receiving placebo, described as placebo (closed circle with continuous line), as hydroxyzine (open circle with interrupted line), or as unknown (closed circle with dotted line). e and f mean adjusted scores per type of effect. Closed diamonds indicate total effect; closed squares interaction model-estimated drug effect, open squares conventionally estimated drug effect, and closed circle placebo effect. T-bars indicate standard errors

Fig. 4

Mean Area-Under-the-Curve According to Intervention or Type of Effect. a to c adjusted mean area-under-the-curve after receiving 25 mg hydroxyzine (black bars), described as hydroxyzine (H/H), as placebo (H/P), or as unknown (H/U); or placebo (grey bars), described as hydroxyzine (P/H), as placebo (P/P), or as unknown (P/U). d to f adjusted interaction model-estimated drug effect (black bars), drug*placebo interaction effect (dotted bars), placebo effect (white bars), conventionally estimated drug effect (vertical strips bars), and total effect (horizontal strips bars). T-bars indicate standard errors. * p <0.05, ** p <0.001