Image: Getty Images
The placebo effect baffles patients, confounds clinicians and frustrates drug developers. Until now, relatively little empirical evidence existed for the biological mechanisms that underlie the effect. But recently, researchers have begun approaching the challenge with methodological rigor. This new area of investigation, straddling basic and clinical realms, has evolved largely because of the novel, detailed window of observation offered by modern imaging technologies. "What we're getting," says Harvard Medical School's Ted Kaptchuk, "is good preliminary evidence that describes the hardwiring of the placebo effect--that is, the impact of symbolic treatment, and how it's mediated through the neurobiology of the brain to produce physical effects in illnesses."
Understanding the biological basis of the placebo effect has potentially wide-ranging implications. Knowing the power of placebo may help scientists and philosophers to better characterize an age-old question. "We had this Cartesian split of mind and body. It's taken us a long time to get back to having ... due respect for mind-body effects," says Linda Engel, acting director, National Center for Complementary and Alternative Medicine's division of extramural research and training. Engel co-organized a workshop on the placebo that NCCAM held in November 2000. Based in part on conclusions from this workshop, the National Institutes of Health plans several new, soon-to-be-announced grants for the coming fiscal year.1
Philosophical implications aside, understanding the biology of placebo could also help improve healthcare and help the struggling pharmaceutical industry to develop better, more effective medications. "[The placebo effect] is crushing whole areas of drug development," says Kaptchuk, an assistant professor. While clinicians like him work to maximize people's self-healing capacity, the drug industry attempts to minimize the placebo effect to demonstrate the usefulness of medications.
"[Clinicians] have looked, particularly in psychiatric disorders, for factors that might predict response to placebo to try to minimize placebo response," says Gil Block, therapeutic area medical leader for the central nervous system, pain, and infectious disease at AstraZeneca. "To date, nothing has really been clearly identified." Factors include experiment site selection, different levels of quality-care follow-up, learning effects, and the increasing incidence of patients enrolling in multiple studies, according to Block. But isolating and eliminating those factors has proven extremely difficult. Block notes, for example, that it is not unexpected to have 50% of depression trials fail, meaning the placebo responders do as well or better than those taking active antidepressant medication.
Image: Courtesy of Predrag Petrovic
PAIN AND PLACEBO Researchers have already reported intriguing findings. For years, opioids were known to play a significant role in the placebo effect in some contexts. In 1978, scientists demonstrated that an opioid blocker called naloxone could abolish the placebo effect in a pain experiment.2 A 2001 study demonstrated that the simple interaction between patient and practitioner has measurable effects in the brain. Researchers in Italy, led by Fabrizio Benedetti, found that "open" injections of analgesics (patient-witnessed) were significantly more effective and less variable than "hidden" injections (patients were ignorant of injections).3 Furthermore, they showed that by blocking the opioids, naloxone greatly reduces this open-injection placebo effect, suggesting that the therapist-patient interaction activates the endogenous opioid systems. "We are beginning to understand what happens in the patient's brain when he or she interacts with his or her therapist," says Benedetti, a physiology professor at the University of Turin Medical School.
Earlier this year, a Swedish research group caught the relationship of pain and placebo in action using positron emission tomography (PET), and showed that placebo and opioid analgesia may have a common neural mechanism.4 Nine subjects were exposed to standardized, brief, nonharmful, painful experiences. They received either no treatment or an injection of either placebo or analgesic. Subsequent PET scans indicated activation of the rostral anterior cingulate cortex for both treatments, though analgesic did provide more pain relief.
Swedish study leader Martin Ingvar says his research is partially motivated by shortcomings in patient care and a troubled drug industry. "In spite of more and more effective medicines, patients are complaining about less and less [therapeutic] effects," says Ingvar, a cognitive neurophysiology professor. "Explaining all drug effects by pure molecular mechanisms will underexpose the effects of drugs."4
PLACEBO, PARKINSON, AND DEPRESSION But the opioid-placebo connection, though still largely a mystery, may turn out to be more straightforward than cases of other types of placebo effects. Another study, which appeared in May, also compared brain scans of persons on placebo versus those on active medication.5 But here, investigators didn't have a specific system like the opiate system to investigate.
Researchers gave 17 severely depressed men either placebo or fluoxetine (Prozac). Scans taken at one and six weeks showed that both groups exhibited increased activity in the cortex and decreased activity in limbic regions, but only patients given fluoxetine experienced changes in brain stem, striatum, and hippocampus. "Drug is placebo plus," explains lead author Helen Mayberg, a psychiatry and neurology professor, University of Toronto. The unique brain areas activated with an active drug aren't particularly surprising, says Mayberg, since they include known binding sites for fluoxetine and serotonin, a neurotransmitter that could have a role in placebo-treated depression.
In August 2001, Parkinson disease researchers reported that the release of another neurotransmitter, dopamine, may be triggered by the placebo effect.6 Again using PET, investigators provided in vivo evidence that endogenous dopamine was released in the striata of six patients with Parkinson disease. Senior author Jon Stoessl notes that the findings indicate the importance of carefully controlling for placebo in Parkinson disease, and that dopamine release could contribute to the placebo effect in other conditions.
Of course, as with any technique, imaging studies have limitations. The depression study, for example, could not rule out the possibility that patients got well as a result of spontaneous remission rather than the placebo effect. Theoretically, suspected placebo responders might have gotten well without any medication, active or not.
The work continues. Ingvar and his colleagues are studying the underlying biology of the effect in mood disorders and so-called modern living maladies like irritable bowel syndrome. A major future goal, says Benedetti, is to understand why not all subjects respond to placebo. "[These studies] begin to say that we cannot only look at a phenomenology of people responding to imitation treatments," explains Kaptchuk. "But we can actually look at the hardwiring that possibly accounts for what we're seeing. And that's an important phenomenon."
Eugene Russo (erusso@the-scientist) is a contributing editor.
1. The science of the placebo: Toward an interdisciplinary research agenda, available online at http://placebo.nih.gov.
2. J.D. Levine et al., "Mechanism of placebo analgesia," Lancet, 2:654-7, 1978.
3. M. Amanzio et al., "Response variability to analgesics: a role for non-specific activation of endogenous opioids," Pain, 90:205-15, 2001.
4. P. Petrovic et al., "Placebo and opioid analgesia - Imaging a shared neuronal network," Science, 295:1737-40, 2002.
5. S. Mayberg et al., "The functional neuroanatomy of the placebo effect," American Journal of Psychiatry, 159:728-37, 2002.
6. R. de la Fuente-Fernandez et al., "Expectation and dopamine release: Mechanism of the placebo effect in Parkinson's disease," Science, 293:1164-6, 2001.