Placebo response not all in the mind

NEW YORK — A new brain imaging technique has revealed that patients with depression have clear physical responses to both drugs and placebos, but the responses are dramatically different, perhaps explaining how the 'placebo effect' works.The new technique — quantitative electroencephalography (QEEG) — developed at University of California, Los Angeles (UCLA), was used to examine brain electrical activity in patients being treated for depression with placebo. These scans were compared with scans from other patients treated with antidepressant medication. Andrew Leuchter and colleagues used the imaging technique to examine QEEG power and cordance — a new measure that reflects cerebral perfusion and which is sensitive to the effect of antidepressant medication (Am J Psychiatry 2002, 159:122-129).Patients who responded to placebo showed increased activity in the prefrontal cortex of the brain, while those who responded to standard antidepressant medication showed suppressed activity in that area. Researchers in the field have previously linked the prefrontal cortex to processes affecting diverse areas of cognition, including working memory, information processing, behavioral organization and attention. "People have known for years that if you give placebos to patients with depression or other illnesses, many of them will get better," said Andrew Leuchter, lead author of the study and director of adult psychiatry at the UCLA Neuropsychiatric Institute and Hospital. "What this study shows, for the first time, is that people who get better on placebo have a change in brain function, just as surely as people who get better on medication. We now know that placebo is, very definitely, an active treatment condition."The current study enrolled 51 individuals with major depression who were assigned to one of two independent, nine-week, double-blind placebo-controlled studies, using either Prozac (Fluoxetine) or Effexor (Venlafaxine) as the antidepressant medication. A series of five QEEG recordings were performed on each study subject during the course of treatment. In addition, each individual's condition was monitored weekly by a research nurse to address safety concerns about dispensing placebo alone to patients with significant depression. After nine weeks, the blind was broken and subjects were classified as medication responders, placebo responders, medication non-responders or placebo non-responders. Overall, 52% (13 of 25) of the subjects receiving the antidepressant medications responded to treatment, compared with 38% (10 of 26) of those receiving placebos.The imaging studies showed distinct changes in the brain's prefrontal cortex between the two groups. "The immediate outcome was they were virtually indistinguishable," Leuchter said. "At eight weeks you couldn't tell them apart in terms of mood ratings. What happened at eight weeks plus a day was a bit different. Some of the placebo responders, when told they were on a placebo, had a deterioration of their mood. In fact, most of them did. Within a month, most of the placebo responders had enough depressive symptoms that they actually ended up on medications."Once people realized they were not taking real drugs, Leuchter said, the placebo effect stopped. The QEEG scans, interpreted through a computer, showed something even more interesting. "Brain function changed at different rates," Leuchter said. "The medications' effect was virtually immediate. Within 48 hours, we saw there was suppression of prefrontal activity in medication responders. It was an immediate and dramatic effect of medication." But it took two weeks for any changes to be seen in the brains of people who responded to placebo, and when the change did come, it was an increase of activity in that part of the brain, not the decrease seen with the drugs.Quantitative electroencephalography has been around for some time. With the technique, an EEG signal is digitized and analyzed by computer to determine the energy in the signal. The advance that the group from UCLA made was the development of the cordance mapping method, where the QEEG information is processed for additional variables, such as blood flow. The group has used the same technique to examine changes in brain function in response to medication and Leuchter said he believed it might have some utility in predicting which antidepressant is likely to work for an individual patient. "These findings show us that there are different pathways to improvement for people suffering from depression," Leuchter said. "Medications are effective, but there may be other ways to help people get better. If we can identify what some of the mechanisms are that help people get better with placebo, we may be able to make treatments more effective." He added: "What our study suggests is that medications are effective in relieving symptoms and placebo can be effective for some patients in acute relief of symptoms but that they are working through different and perhaps complementary mechanisms. The placebo effect is something that is very physically distinct."The team considered that their most surprising finding was that the changes in brain function seen during medication-related improvement were the opposite of the changes seen during placebo-related improvement. They believe that this is the first time that treatments for depression that appear to yield the same clinical result have been shown to be associated with distinct and different changes in brain function. It has been recognized since 1955 that people suffering from depression often experience a substantial reduction in symptoms when given a placebo. In that year, Henry K. Beecher published a study 'The Powerful Placebo' in Journal of the American Medical Association. Reviewing 15 clinical trials, Beecher claimed that, on average, about one in three patients found relief from placebo alone. This observation led some researchers to propose that up to 75% of the apparent efficacy of antidepressant medicine might be attributable to the placebo effect. The new study from UCLA sheds light on the mechanism, and might indicate ways for doctors to use the placebo response — when patients get better from a sham treatment or drug — to improve standard treatments.Although some specialists have challenged the placebo effect for years, in the minds of most physicians and in the public consciousness, it remained a fact until May 2001. That's when Peter Gøtzsche and Asbjørn Hróbjartsson of the University of Copenhagen concluded in the New England Journal of Medicine that "there is no justification for the use of placebos" in medical practice. They had pooled data from 114 previously published clinical trials in a range of conditions that compared patients who received placebos with those who got no treatment whatsoever. Sifting the numbers through statistical sieves, the doctors found no significant overall difference in how the two groups fared.There was good reason to doubt that new charge that placebos are worthless, however. "Their own data show that placebo is significantly better for pain than no treatment is," observed Walter Brown, a psychiatrist at Brown University, Providence, Rhode Island, US. That result seems credible, because 27 of the 114 trials measured the effect on pain. But the remaining trials lumped into the analysis looked at 39 other maladies, ranging from infertility and compulsive nail biting to marital discord, orgasmic difficulties and fecal soiling. For each of these problems, the number of patients was too small to allow any firm conclusion except that placebos do much more for some illnesses than for others.Other meta-analyses have shown measurable placebo effects for depression, asthma and phobic anxiety. Parkinson's disease recently joined that list. In August 2001, Jon Stoessl and colleagues at the University of British Columbia, Vancouver reported in Science that they could see the effect of placebo treatment in brain scans of people with Parkinson's disease. They used positron emission tomography (PET) to estimate dopamine activity in the diseased areas of six patients' brains after they were injected with either inactive saline solution or apomorphine, a drug that mimics dopamine. When the subjects were given placebo, their brains released as much dopamine (which is suppressed in Parkinson's disease) as when they were given active drugs, Stoessl said. This study, however, looked at the release of neurochemicals deep within the brain using PET, a different application than the group from UCLA.Until the latest study from the UCLA team, this was one of very few studies ever to look beyond whether or not a placebo works to see how it works. One fact not reported in this paper is what happened at the end of the study, when subjects were told what they really had received. Most of the placebo treated subjects had a deterioration of their mood, and ended up receiving medication. This is consistent with the traditional view of the placebo response, that it does not yield enduring benefit. One thing appears certain. Until more studies of this type are done, the mechanism of the placebo effect will remain a mystery.

Placebo response not all in the mind

Imaging the brain during drug or placebo treatment for depression shows different but similarly effective responses.

Interested in reading more?

Become a Member of