Side effects can enhance treatment response through expectancy effects: an experimental analgesic randomized controlled trial

Abstract

In randomized controlled trials, medication side effects may lead to beliefs that one is receiving the active intervention and enhance active treatment responses, thereby increasing drug-placebo differences. We tested these hypotheses with an experimental double-blind randomized controlled trial of a nonsteroidal anti-inflammatory drug with and without the addition of atropine to induce side effects. One hundred healthy volunteers were told they would be randomized to either combined analgesics that might produce dry mouth or inert placebos. In reality, they were randomized double blind, double-dummy to 1 of the 4 conditions: (1) 100 mg diclofenac + 1.2 mg atropine, (2) placebo + 1.2 mg atropine, (3) 100 mg diclofenac + placebo, or (4) placebo + placebo, and tested with heat-induced pain. Groups did not differ significantly in demographics, temperature producing moderate pain, state anxiety, or depression. Analgesia was observed in all groups; there was a significant interaction between diclofenac and atropine, without main effects. Diclofenac alone was not better than double-placebo. The addition of atropine increased pain relief more than 3-fold among participants given diclofenac (d = 0.77), but did not enhance the response to placebo (d = 0.09). A chain of mediation analysis demonstrated that the addition of atropine increased dry mouth symptoms, which increased beliefs that one had received the active medication, which, in turn, increased analgesia. In addition to this indirect effect of atropine on analgesia (via dry mouth and beliefs), analyses suggest that among those who received diclofenac, atropine directly increased analgesia. This possible synergistic effect between diclofenac and atropine might warrant future research.

Figures

Figure 1. Study design and proceedings

Participants were only aware of the two boldened randomization groups, ie. diclofenac+ atropine and placebo + placebo. VS= vital signs, Q= Questionnaires, P.V. = psychological variables, Expectations= measure of participant's expectations of relief, S.E.= side effects, Beliefs= beliefs about treatment assignment

Figure 2. Effects of diclofenac and atropine on pain relief

The levels of analgesia, defined as the difference between pre- minus post- treatment pain ratings, are illustrated for the four treatment groups. Diclofenac is a Non Steroidal Anti-inflammatory Drug (NSAID) reducing heat pain, atropine is an antimuscarinic agent without known analgesic effects, that was given to induce a side effect in the form of a dry mouth. Error bars = SEM.

Figure 3. Structural equation model predicting analgesia

Standardized estimates are shown on the paths connecting the variables, and proportions of variance accounted for in each variable are shown above each variable box. The variables shown in circles represent unexplained variance in each measured variable. * p < 0.05. ** p < 0.01.