The enigma of placebo effects in drug-refractory epilepsies


Address correspondence to Günter Krämer, Swiss Epilepsy Center, Bleulerstrasse 60, CH-8008 Zurich, Switzerland. E-mail:


Placebo effects in the therapy of epilepsy were already known before the introduction of effective antiepileptic drugs (AEDs). They have physiologic correlates, and are even stronger in other neurologic disorders such as pain. Placebo effects in epilepsy have many facets. Our understanding of this phenomenon has increased in the last two decades: placebo effects are stronger in children than in adults, and may be culture- and setting-dependent; and impressive placebo effects occur in animals with epilepsy as well. More research is needed to fully elucidate the mechanism of placebo effects in epilepsy care, particularly as we go forth with studies addressing the issue of pharmacoresistance.

According to the classical definition by Shapiro and Morris (1978) a placebo is defined as “any therapy or component of therapy used for its nonspecific, psychological, or psychophysiological effect, or that is used for its presumed specific effect, but is without specific activity for the condition being treated.” Placebo effects belong to the most fascinating phenomena in clinical research and have received increased attention in recent years (Benedetti, 2009; Brown, 2013). In the treatment of neurologic diseases and disorders, impressive placebo effects are observed not only in “functional” disorders such as headache or other pain syndromes, but also in epilepsy.

Historical Background

It was the Scottish physiologist, pharmacologist, internist, and neurologist William Cullen (1710–1790), who had studied medicine in Glasgow, London, and Edinburgh and was the successor of the internist and neurologist Robert Whytt in Edinburgh (Kerr et al., 2000), who first used the term placebo in its contemporary meaning ( The fact that placebo responses in epilepsy therapy are not necessarily associated with efficacious drugs, may be elucidated by two historical vignettes.

The first is the story of the so called “Perkins Patent Tractor” introduced by the American physician Elisha Perkins (1741–1799). This device resembled a large spike or nail supposedly created from special and secret metals and was used “like a pointer,” which was moved across diseased body parts and allegedly “pulled out” pathologic conditions, including epilepsy, thereby removing the illness from the body. They were very popular in the United States and England for many years; even famous people like the American President George Washington apparently bought and used this device (Krämer, 2012).

Historical Vignette 1: Perkins BD. The Influence of Metallic Tractors on the Human Body (1798)

“(Benjamin Shreve). I have been witness to an operation of your Tractors on my son, for the relief of epileptic fits, to which he has been subject about 18 months. On the 23 days of the fourth month, he was seized with one, with entire loss of reason. His hands were so clenched together with spasms, that the efforts of James Laurason, whose assistance I called in, and my own, could not open them. In this situation we applied your Tractors to each arm, drawing them from the elbow down to his hand, and to our great surprise his hands soon became perfectly lax, and opened with ease: by continuing the application on his head for a few minutes, he came to his reason, and went to sleep; since which he has had no more signs of them. On all former attacks, they have continued 6 or 8 h, and from 12 to 20 in number. Bleeding and other means have been used, but he was never before relieved of them so immediately. I am confident the Tractors effected the cure.”

(Perkins, 1798)

However, in 1799 the British physician John Haygarth (1740–1827) published a critical study in which he compared the treatment results of a wooden replica tractor and the high-priced original and did not find any difference: His book was entitled “On the Imagination as a Cause and as a Cure of Disorders of the Body” (Haygarth, 1800) and provided the first evidence of a placebo effect in clinical practice. More specific for placebo effects in epilepsy care was another observation published in the early 19th century.

Historical Vignette 2

“Mr. XXX says, that at the time he was physician to La Salpétrière, where he had large numbers of women afflicted with nervous and mental disorders under his care, he was in the habit of selecting every spring and autumn, 30 epileptics, of whose cases he knew the history, the causes and symptoms, and whom he prepared in advance, by exciting their imagination with repeated promises of cure. He then tried, in one set of cases, bleeding, in another, purgatives, in a third, baths of every possible temperature, and so on with counter-irritants, moxas, the hot iron, antispasmodics, narcotics, and even secret quack medicines. The invariable result of all these therapeutical experiments was, that a new medication suspended the attacks for a fortnight, sometimes even for a month or 3 months. But after that time, the attacks reappeared successively in all the patients, and presented entirely the same characteristics as before. Some of these patients he treated for several years in succession; but he did not obtain a single cure amongst them.”

(Althaus, 1866)

These observations were made well before the introduction of potassium bromide as the first effective antiepileptic drug in 1857. Mr. XXX was the eminent French psychiatrist and early epileptologist Jean-Étienne Dominique Esquirol (1772–1840) who had worked at the Salpêtrière Hospital in Paris with Philippe Pinel and had been appointed director of the famous Charenton Hospice near Paris in 1825. In 1815 he had introduced the terms “grand mal” and “petit mal”—originally used by patients—as well as the concept of symptomatic epilepsy in the medical nomenclature (Krämer, 2012).

Placebo Effects in Modern Epileptology

Placebo-controlled, double-blind randomized clinical trials (RCTs) are the gold standard for trials in the modern age of evidence-based medicine. However, for not fully understood reasons, some of the newer antiepileptic drugs (AEDs) were unable to demonstrate a significant advantage in responder rates over placebo in add-on-trials (e.g., carisbamate [Halford et al., 2011], perampanel [French et al., 2012]) and even in monotherapy-trials in newly diagnosed patients (e.g., oxcarbazepine [Novartis AG (2013), data on file]). On the other hand, in a recent meta-analysis of all add-on placebo-controlled AED trials, patients in the placebo arm had a sevenfold increased risk of sudden unexpected death in epilepsy (SUDEP) compared with those randomized to the active arm (Ryvlin et al., 2011), which raises ethical issues.

After the approval of several new AEDs and in the absence of head-to-head comparative trials, the use of a placebo-corrected response rate in the clinical evaluation of AEDs was proposed for the first time in 1999 (Cramer et al., 1999). This placebo-corrected response rate has now been used repeatedly (e.g., Beyenburg et al., 2010, 2012). One of the confounding problems with this approach has been identified as a statistically significant increase in the responder rates over the years. A recent meta-analysis showed increasing responder rates over time in both the placebo and the active medication groups (responder rates were significantly higher in more recent than in older studies). This remained significant only for placebo when using linear regression and was associated with a nonsignificant trend toward a reduced size of the treatment effect in the more recent trials (Rheims et al., 2011). Another interesting finding is that placebo responses in AED trials tend to be more pronounced in children than in adults. A meta-analysis on this topic revealed an almost doubled rate in children than in adults (19% vs. 9.9%, p < 0.001), whereas no significant difference was observed in the response to active treatment (Rheims et al., 2008).

It is also significant that placebo responses to AEDs are not restricted to human studies but can also be observed in animal-based experimental epileptology. In refractory canine epilepsy, a placebo-controlled add-on trial failed to demonstrate a superiority of levetiracetam (Muñana et al., 2012); a meta-analysis of three other studies evaluating new AEDs in development noted a placebo effect with decreased seizure frequency compared with baseline in 79% of the dogs in the studies (Muñana et al., 2010).


Not only in nonrandomized studies, but also in double-blind, placebo-controlled RCTs, the preferences of patients and physicians can influence the evaluation of treatment effectiveness. Internal and external validity of RCTs are impacted by various biases related to patient and physician preferences, and influence of patient and physician expectations on trial outcomes might be much less trivial than expected, both in open-label and double-blind, placebo-controlled RCTs (Rheims & Ryvlin, 2012). Included among other possible explanations for placebo responses is that the population studied was different in an important manner (e.g., more susceptible to psychological suggestion) and that environmental settings may influence the magnitude of a placebo effect (Spilker, 1986). Many other factors contribute to rising responder rates in the placebo arms of AED trials. In addition to a true placebo effect, regression to the mean and expectations of investigators and patients, which might also differ by region, are to be mentioned. Patients in resource-poor countries might have greater incentives to remain in a trial or to report improvements to obtain better access to an otherwise not available drug treatment (Friedman & French, 2012), and other cultural factors may also be important. The extremely important consideration of placebo effects must be incorporated into all trials of new therapeutics as the problem of pharmacoresistant epilepsy is addressed in coming years.


The author has no conflicts of interest to disclose. I confirm that I have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.