What is the Placebo Effect?
The placebo effect is a very vague and controversial topic that a lot of doctors are especially hesitant to deal with. Research into this mysterious area has been limited so not much is known about what causes the placebo effect. Also, different possibilities of areas where the effect occurs, the nature of the possible mechanisms, and variations in studies and patients. have made it exceptionally difficult to come up with a precise biochemical change that explains the effect. We do, however, have at least some insight into its use.
A placebo takes many forms. It can be an inert, orally consumed substance (sometimes termed a dummy), a procedure or a situation. Sometimes an association and/or situation has placebo characteristics, and in some cases the substance may be active. In all cases a placebo effect occurs only if the substance has an unintended effect.
Back to top
Whence Comes the Placebo Effect?
Let's look briefly at the history. Originally the term 'placebo' was derived from the Latin phrase 'I shall please'(Vernon, 1991). In addition, it had early biblical references and was related to pleasing or deception (Strauss, Cavanaugh 1996). Later on, during the middle ages, it evolved to refer to one who was a flatterer or sycophant (one who artificially lies to satisfy another person) (Vernon 1991). So-called 'placebos' were hired to mourn at funerals and comfort the victimís family (Weihrauch, Gauler, 1999). After the middle ages its more modern definition emerged. In 1785, a medical dictionary defined a placebo as a commonplace medicine or methodology (Strauss, Cavanaugh 1996). In 1811, the definition was modified to refer to 'an epithet to any medium adopted to please rather than benefit the patient'(Strauss, Cavanaugh 1996). During this time various 'placebos' were being used in medical treatment and consisted mainly of 'alternative remedies'. In Chinese medicine, for example, inert substances such as cockroach intestines, animal bones, feathers, animal fat, and so-called herbal remedies were prescribed for multiple ailments and may have exerted limited placebo actions (Vernon 1991). Other cultures had similar practices, like those of witch doctors. At present a placebo is as we know it. Shapiro, in 1964, defined the placebo as 'any therapy or component of therapy that is deliberately used for itís non-specific psychological or psycho-physiological effect but is without specific activity for the treated condition'(Strauss, Canvanaugh, 1996). In other words, an inert substance that supposedly has no intended effect causes some sort of change in the patient that is similar to the response to an active drug. Its use in double blind studies was documented as early as 1932 by Paul Martini, but even in 1912 it was used as a control in caffeine studies by Hollingsworth (Vernon, 1991).
Back to top
What Are They Used For?
There are a wide variety of diseases and conditions where placebos exhibit some effect. Placebos have exhibited effects in treating angina pectoris, hypertension, gastrointestinal disorders, reflux oesophagitis, asthma, allergies, flu symptoms, sleep disorders, sea sickness, and nocturnal enuresis (Weihrauch, Gauler, 1999). This list is obviously not all inclusive. Areas where placebos seem to have the greatest effect are in treating pain and affective disorders such as depression, bipolar disorder, mania, and anxiety disorders. Conditions where placebos generally fail are in the treatment of long term and/or terminal illnesses such as cancer (Vernon, 1991). Also it has no effect in rescuing vital function following a heart attack.(Vernon, 1991). It must also be noted that the effects of a placebo are not always good. Placebos in many patients have been linked to side effects, some of which are severe (fainting, cardiac dysfunction) (Wiehrauch, Gauler,1999). Pain and depressive disorders are our main focus here.
The pharmacological and biological effects of the placebo in affective disorders and pain remain a mystery. In the early 80's, there were some studies done on the ability of placebos to treat pain. One such study was done by Levine and colleagues; they investigated the effects of placebos on the pain that followed oral surgery. Briefly, the central nervous system contains a special class of receptors called opioid receptors. Exogenous opioids that bind to this receptor include alkaloids that are derived from the poppy flower (Vernon, 1991). Morphine and heroin are ligands that bind to this receptor. Activation of opioid receptors, as occurs upon introduction of morphine, creates analgesia. It was later discovered that the brain produces a special class of endogenous opioid ligands called beta endorphins. These messengers appear to be released during injury as a means to alleviate pain naturally. There is still some mystery surrounding the actions of beta endorphins. Levine administered placebos to patients who had just undergone oral surgery. The placebo did in fact block some of the pain that was present (Vernon, 1991). This analgesia was reversed by naloxone, which is an opioid antagonist. This indicated that a placebo in the pain situation might be causing the natural release of beta endorphins by the brain (Vernon, 1991). It is impossible, however, to draw that specific conclusion since naloxone also caused hyperalgesia in patients who hadnít undergone the surgery and who hadnít received a prior placebo. In addition, naloxone failed to block analgesia induced by another type of placebo-specifically hypnosis (Gowdey, 1983). In addition this only applies to pain, what about other areas and conditions? We truly donít know.
Back to top
When Do They Work?
In looking at depression, anxiety, manic and bipolar disorders, no specific biochemical changes have been found as a consequence of placebo action. Studies have indicated several areas that could account for placebo effects, but nothing is for sure. Some theories involve changes in autonomic functioning in the central nervous system, immune system changes, endocrine changes, cognitive effects, classical conditioning, and especially expectations due to environment (Strauss, Cavanaugh, 1996). There is crossover in these categories, such as expectations changing cognitive functions.
An example of the expectation phenomenon occurred in observations by Beecher of WWII injured soldiers and civilians at home who were injured by other means. War injured soldiers appeared to be more analgesic compared to normal injured individuals (Gowdey,1983). Beecher attributed this to the expectancy or excitement the soldiers felt, knowing they were headed home because of their injury (Gowdey,1983). The other injured group had to leave home to go to the hospital and were anxious and hyperalgesic as a result.
Classical conditioning also ties into expectancy. In guinea pigs, pairing cylcophosphamide (an immune system activator) administration, with saccharin (not active in the immune system) caused an immune response to saccharin when it was later administered alone(Tausk, 1998). In humans, cancer patients receiving chemotherapy have been known to get nauseated (to the point of vomiting in many cases) upon entering the room or hospital where they receive the chemotherapy(Tausk, 1998). Autonomic functions seem to get activated by these factors. Mothers have been known to produce milk in response to hearing their infant crying (Strauss, Cavanaugh 1996). These factors also may be playing a role in pain and depression.
Doctor- patient relationships also plays a role in the placebo effect. The attitude a doctor has when prescribing medication or a placebo when treating depression (and other disorders) may affect how the patient responds (Gowdey, 1983). This is a complicated phenomenon. The patient may not want to disappoint the doctor and will report relief of depressive symptoms even when given a placebo that had no effect. There are poorer response rates to placebos when the doctor is cold or indifferent to the patient (Gowdey, 1983). Going back to the expectancy issue, patients who are warned of post-surgical pain prior to any type of surgery are more algesic following surgery (Gowdey, 1983).
There are also issues concerning the nature of a placebo. For instance patients respond more to red and blue pills than other colors (Gowdey, 1983). Very large pills or unusually small pills have a higher response rate -likely due to their perceived potency (Gowdey, 193). A similar effect is seen for injections.
People who are most susceptible to the placebo effect are those that are more anxious, younger, and those who have less experience taking medication and adhere to the guidelines. Adverse side effects were a bit higher in women, especially during their menstrual cycle. People less likely to respond to placebos are those that may suspect they are on one, older people, and those who have more experience taking medication and are less likely to follow specified guidelines (Gowdey, 1983).
So far, most of what has been discussed is behavioral. We have no explanation for the why these factors are present. For example, the biochemical changes that expectancy causes, other than autonomic arousal, are unknown. It must also be noted that the above statements are only generalized. Many variations are present depending on the condition and nature of treatment. There just arenít any specific rules that determine if a placebo effect will occur or what it will be.
Back to top
Placebos and Disorders
It gets more confusing when looking at the affective disorders. Placebos are most effective in anxiety disorders such as panic attacks, obsessive compulsive disorder, phobias and generalized anxiety (Piercy et al, 1996). Related are depression and mania, including bipolar disorders. In all cases, studies have revealed some effectiveness from placebos. Many studies use placebos as a control(Tausk, 1998). Response rates to placebos were as low as 7% yet as high as 64% in many of these disorders (Piercy et al, 1996). Disorders requiring lithium treatment (manic disorders, bipolar) were most susceptible to the placebo effect (Keck et al, 2000). The placebo may somehow activate an independent chemical pathway in the brain that dampens the anxiety or manic attacks. We donít know for certain. Also, every study in this area is different (Keck et al, 2000). For example, drug washout times vary in different studies and they affect the results, making it difficult to draw generalized conclusions (Keck et al,2000).
Overall, it has been established that the placebo effect is a real thing that studies must take into consideration, especially those related to affective disorders. However we really donít understand the physical and chemical basis for the placebo effect. Until we do, the placebo effect is nothing more than a control in experiments to test drugs.
Back to top
A Few Words on Placebo Studies
Researchers are beginning to understand that there is a difference between curing symptoms and improving the quality of life of the patient. Where before placebos had been considered successful if overt behavior such as panic attacks or sleep disturbances were reduced, now more consideration is being directed towards the patientís description of their life after placebo. Recent studies have demonstrated that improvement in quality of life, as measured by an assessment test, can differentiate placebo response from true medication response (Clary, Mardekian, Pollack, Rapaport & Wolkow, 2000). Unfortunately for placebo theorists, the studies found that medication served to improve quality of life over placebos, but you never know what future studies may reveal.
In studies that involve placebos, evaluation of effectiveness is even more difficult than in the administration of Ďvalidí drugs alone (Dobrilla & Scarpignato, 1994). The uninformed researcher can easily misinterpret the data, and as mentioned, even their attitudes toward the patient can drastically affect their response and reports. There are a number of factors that can create false impressions of placebo effects that both researcher and reader need to take into consideration during a study. Knowing them, or at least having a general idea of what they stand for, will aid in a more competent and valid analysis of the results.
Its important to keep in mind, when reading such cautions, that things such as suggestion, biofeedback and forms of self-healing are indeed valid and are under investigation, but these are too often confused with placebo effect in medical experiments, which is a relevant change in behaviour accredited to the administration of an imitation treatment, like a sugar pill said to be a drug, or ultrasound with the machine turned off (Dobrilla & Scarpignato, 1994).. This specific form of placebo effect is what the rest of this page will be looking at, due to it's comparative convenience of administration and observation. It is important to understand this, for some of the conditions mentioned below that can disprove the effectiveness of placebos are excellent proof for the effectiveness of suggestion and what are called placebo effects outside the experimental realm. Kiene and Kienle (1997) provided us with the list of categories, and we have altered their explanations for easier digestion.
These factors are often incredibly difficult to pick out in a study, and often appear completely by accident. Placebos are tricky creatures and it takes a whole team of objective researchers to be able to get a good angle on their effect. A lone researcher can be unwittingly cursed by his own biases, and placebo administerers can unintentionally throw off a whole study with their attitude. Such are the hazards of the times and the beauty of scientific advancement. As we are learning these things we can develop tighter and tighter controls for ever more accurate results.
Back to top
Placebo Studies and Anxiety
First of all, here are the basics of the raging controversy over the right or wrong of placebos.
On the down side, some placebo-controlled trials may indeed be withholding effective treatment from the patients who need it for the sake of the testing; therefore, many people advocate that placebo-controlled trials are only ethical when there is no known effective treatment. The rationale behind this is that the interest of medical science should really be to compare new treatments with known treatments, rather than new treatments with nothing (Preston, Materson, Reda, & Williams, 2000). This belief, of course, is held by those people who believe that placebo is nothing.
On the up side, many new drugs have proven ineffective or even toxic, so whatís the harm in a little placebo? These people spit back at the non-believers that ANY new trial should be of ethical concern, as it could harm a patient as easily as a placebo. Controlled trials using placebos and those that focus on the specific placebo effects have been credited as being the most scientifically rigorous means of evaluating new treatments. Without placebo control groups, two different drugs could not be adequately compared (Preston et al.2000). Placebo effects must be taken into account to determine the actual effectiveness of the specific drug characteristics so they are not credited for something a 'mere' placebo could do.
There are a few circumstances where everyone agrees placebo use is A-OK. One is those situations where placebo responses are high and relatively close to the response rates for proven effective therapies (though no one thought to mention those placebo studies had to start with no data sometime). Another is when those established treatments carry a known risk (Preston et al. 2000).
Studies on hypertension fit these criteria and are good examples of placebo-controlled experiments.
Back to top
A Study In Hypertension Patients (Preston et al. 2000)
187 patients out of 1105 were randomly picked to receive a placebo in an attempt to relieve stage 1 and 2 hypertension. 30% of the patients achieved complete treatment success by reaching the goal blood pressure during the study and maintained it for one year. There was no difference in treatment response with respect to age, gender, resting pulse rate, or baseline laboratory values (things like electrolyte levels, cholesterol levels) Patients did, however, respond to different 'doses'; 41% to the lowest, 19% to the middle and 40% to the highest. This was seen as either a function of simply passing time, regression toward the mean, or additional effects related to the patientís expectations that the 'dose' was increasing.
The adverse effects normally shown to be associated with treatment were also seen in the placebo patients. Often a lower percentage displayed them but sometimes the placebo group beat out the one receiving 'real' treatment. An example of this phenomenon is seen in the following table, taken from an aforementioned study in panic disorder patients.
Back to top
(Clary, Mardekian, Pollack, Rapaport & Wolkow, 2000).
A Study in Panic Disorder Patients (Beck, Mellergard, Ottosson, Rosenberg & Rosenberg, 1991)
123 patients were diagnosed as having panic disorder or phobias with panic attacks. There were 3 random groups of 41, treated with either alprazolam, imipramine or placebo. Pre-treatment, each patient received a standard diagnosis, was rated on several scales for panic attacks, anxiety, phobia and depression and demographic data was collected. To control for noncompliance, plasma concentrations of active drugs were taken at the beginning, end, and near middle of the study. The patients were interviewed and assessed weekly and were also asked to keep records of self-ratings and symptoms.
Sadly, there was a significantly higher dropout rate among the placebo people than the drug people. However, all placebo completers were reduced to zero panic attacks by the end of the study. The placebo response was also evaluated and determined to be just as good as the response to active medicine. The researchers found no age, sex, marital or work differences between placebo responders and nonresponders. However, among the placebo dropouts (nonresponders) the distress levels had been recorded as higher, with more severe syndromes and symptoms recorded at baseline and throughout the study. There was a similar pattern found between active medication responders and nonresponders. The less severe patients had a higher response rate. (Beck, Mellergard, Ottosson, Rosenberg & Rosenberg, 1991)
These differences can have several interpretations. The responders may have a better cognitive grasp on their symptoms and so process them better, or the responder's lower number of disabling symptoms may make them better suited to cope with the attacks.
Another factor that may play into placebo response that has been debated in many studies, this one included, is the personality of the responder vs. the nonresponder. The evidence suggests that the strength of personality, the ego, is of great importance for response. This evidence might lead to future personality evaluations prior to placebo administration. The path has promise.
Back to top