Different Placebos, Different Mechanisms, Different Outcomes: Lessons for Clinical Trials

Abstract

Clinical trials use placebos with the assumption that they are inert, thus all placebos are considered to be equal. Here we show that this assumption is wrong and that different placebo procedures are associated to different therapeutic rituals which, in turn, trigger different mechanisms and produce different therapeutic outcomes. We studied high altitude, or hypobaric hypoxia, headache, in which two different placebos were administered. The first was placebo oxygen inhaled through a mask, whereas the second was placebo aspirin swallowed with a pill. Both placebos were given after a conditioning procedure, whereby either real oxygen or real aspirin was administered for three consecutive sessions to reduce headache pain. We found that after real oxygen conditioning, placebo oxygen induced pain relief along with a reduction in ventilation, blood alkalosis and salivary prostaglandin (PG)E2, yet without any increase in blood oxygen saturation (SO2). By contrast, after real aspirin conditioning, placebo aspirin induced pain relief through the inhibition of all the products of cyclooxygenase, that is, PGD2, PGE2, PGF2, PGI2, thromboxane (TX)A2, without affecting ventilation and blood alkalosis. Therefore, two different placebos, associated to two different therapeutic rituals, used two different pathways to reduce headache pain. The analgesic effect following placebo oxygen was superior to placebo aspirin. These findings show that different placebos may use different mechanisms to reduce high altitude headache, depending on the therapeutic ritual and the route of administration. In clinical trials, placebos and outcome measures should be selected very carefully in order not to incur in wrong interpretations.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.

Figures

Fig 1. Rationale of the study.

High altitude headache can be due to at least two pathways. The first is the reduction of blood oxygen saturation (SO2), which triggers compensatory hyperventilation and excessive elimination of carbon dioxide (CO2) which, in turn, increases blood pH (alkalosis). The importance of alkalosis in high altitude headache is shown by the therapeutic effects of acetazolamide, that reduces blood pH. The second pathway is the activation of cyclooxygenase (COX) by hypoxia with an increase in prostaglandin (PG) synthesis which, in turn, induces cerebral vasodilation. High altitude headache can be treated with oxygen or COX inhibitors, like aspirin, and these treatments require two different therapeutic rituals. In the former, oxygen is breathed through a mask, in the latter, aspirin is a pill that is taken orally. This allows the study of the mechanisms of two different placebos associated to two different rituals.

Fig 2. Experimental design.

The sequence of test sessions is shown for all groups, starting from 8:00 AM of Day 1 at sea level to 8:00 PM of Day 3 at 3500 m. Vmin = minute ventilation; PG = prostaglandins; TX = thromboxane.

Fig 3. No-treatment group.

Natural course of SO2, Vmin, pH, headache pain, PGD2, PGE2, PGF2, PGI2, TXA2 (means+SD) for the first 48 h after arrival at 3500 m, showing the stable conditions of all these physiological parameters at high altitude.

Fig 4. Means (+SD) for all the measurements in the oxygen group.

After oxygen administration for three consecutive sessions, real oxygen was replaced with placebo oxygen. Note that placebo mimicked the effects of real oxygen for headache pain, Vmin, pH and PGE2, but it had no effect on SO2, PGD2, PGF2, PGI2, TXA2. *P

Fig 5. Means (+SD) for all the measurements in the aspirin group.

After aspirin administration for three consecutive sessions, real aspirin was replaced with placebo aspirin. Note that placebo mimicked the effects of real aspirin for headache pain, PGE2, PGD2, PGF2, PGI2, TXA2, but it had no effect on SO2, Vmin, pH. *P

Fig 6. Means (+SD) for all the measurements in the mixed group.

After oxygen administration for three consecutive sessions, real oxygen was replaced with placebo aspirin. Note that none of the parameters was affected by placebo aspirin. *P

Fig 7. Differences of the means and 95% CI for all the parameters analyzed in the present study.

Horizontal bars crossing the 0 line and white diamonds are statistically nonsignificant. Horizontal bars not crossing the 0 line and black diamonds are significant. From top to bottom: i) placebo oxygen effect is represented by the difference between placebo oxygen an no-treatment, ii) placebo aspirin effect is the difference between placebo aspirin and no-treatment, iii) the difference between the two placebos is the difference between placebo oxygen and placebo aspirin, iv) the mixed effect is represented by the difference between placebo aspirin after oxygen conditioning and no-treatment.

Fig 8. Model that explains the findings of the present study.

The repeated association between a conditioned stimulus (CS) and an unconditioned stimulus (US) leads to a conditioned response (CR), in which the CS alone (the placebo) is capable of inducing the same unconditioned response (UR) of the US. In the case of oxygen treatment, the CS (the ritual of the oxygen mask) induces a CR characterized by a decrease in Vmin and pH. In the case of aspirin treatment, the CS (the ritual of the pill) induces a CR characterized by a decrease in prostaglandins (PG) and thromboxane (TX). The present study suggests that PGE2 might be the final common pathway of the analgesic effect. In fact, both placebos decrease PGE2.