are counted as pure cost, and patients read them as evidence that something is happening. If that reading changes expectations, it should change outcomes. So 77 healthy participants were told they would receive fentanyl nasal sprays before thermal pain. The sprays held no fentanyl: some contained to produce a mild burning sensation, and some inert saline. Did the burn change the pain?
The sprays that burned hurt less. Nasal sprays with a side effect led to lower pain than inert sprays without one. How much depended on two things measured separately: an individual’s beliefs about how side effects relate to treatment outcome, and their expectation about what they had received. Imaging during pain showed the engaged, including anterior cingulate cortex and periaqueductal gray.
Treatment side effects adversely impact patients, and they may also function as indicators that an effective treatment has been received. Both can be true. The second possibility has had almost no attention, because a side effect and an active drug normally arrive together and cannot be separated in a clinic.
This preregistered trial separated them. Seventy-seven healthy participants were made to believe they would receive fentanyl nasal sprays before thermal pain in a controlled setting. No spray contained fentanyl. Some contained capsaicin, which induces a mild burning sensation, and some contained inert saline.
After the first session, participants were randomized into two groups for functional MRI. One group continued to believe the sprays could contain fentanyl, while the other was explicitly told that no fentanyl was included. That allows the side effect and the expectation of pain relief to be manipulated independently of each other.
Sprays with a side effect produced lower pain than inert sprays without one. The size of the influence depended on each person’s beliefs about how side effects relate to treatment outcome, and on their about what they had received. Functional MRI indicated involvement of the descending pain modulatory system during pain after a spray with side effects, including the anterior cingulate cortex and the periaqueductal gray. Read the population before drawing a clinical lesson: healthy volunteers, experimental heat pain, and no disease anywhere in the design.
wat. A nasal spray that stung relieved experimental pain better than one that did nothing, because the sting read as proof the drug had arrived.
wat wat
It separates the side effect from the expectation of relief, which are normally welded together. A mild burn worked as a signal that the treatment was real, and the signal did part of the analgesia.
How do we know?
Schenk, Fadai and Büchel ran a preregistered trial in which 77 participants believed they were receiving fentanyl nasal sprays. Half were then randomized to keep believing it, while the others were told no fentanyl was involved, with functional MRI during thermal pain.
What is still unknown
Healthy volunteers and experimental thermal pain, not patients with disease. Nothing here licenses adding a side effect to a treatment, and the imaging locates the effect rather than proving the pathway.
Learn something?
Where to next
Does a placebo still work if you are told up front that it is one?placebo
If a placebo relieves pain, what is actually doing the relieving?placebo
Can being told you slept badly actually make you think worse?placebo