In clinical trials of pharmaceutical agents, comparison of the agent being tested with a placebo is a time-honored, accepted method to evaluate the efficacy and safety of the agent. But this methodology has not become standard in the evaluation of a new surgical procedure being proposed as improved therapy. Yet, expectations by patients undergoing a surgical procedure are at least as great as in those willing to try a new medication. A common rationalization is that the greater the perceived risk, the greater the expectation that the procedure must be more rewarding. Thus, it is logical to anticipate that surgical procedures could result in a greater placebo effect than that encountered with drug testing. In spite of such reasoning, placebo-controlled surgery (ie, utilizing sham surgery controls) is rarely employed. Typically, new surgical procedures are open-label trials that never proceed to a rigorous controlled surgical trial in which sham surgery is the comparator. Admittedly, a major justification to avoid sham surgery is that such surgery may incorporate the inherent risks of any surgical procedures, such as anesthesia, infection, blood loss, incision errors, pain, discomfort, time for healing, and even the emotional stress of hospitalization. This has led to the alternative of comparing a new surgical procedure with the best medical treatment available. But open-label controls such as these cannot account for the mental mechanism of rationalization (ie, a satisfactory outcome would justify the ordeal of a surgical procedure). Thus, open-label best-medical comparison trials are less than ideal. Even the ingenious temporal-lobe controlled surgical trial,1 in which blinded raters reviewed subjects' written statements about epileptic seizures, still cannot account for the fact that subjects knew that they either had the surgical procedure or were still in the 1-year medical treatment waiting phase of the study. A placebo effect could have influenced their statements.
Another limiting fact is that some surgical procedures require observing immediate results during the operative procedure before the operation can be completed. This is the typical situation for functional stereotaxic neurosurgery employed to arrest abnormal involuntary movements or bradykinesia on awake patients. Such operations are not suitable candidates for a comparative sham surgical procedure. Thus, only operations that are expected to produce delayed benefit would be the candidates for sham surgery controls. Still, these have rarely been performed.
Transplant surgery for Parkinson disease is an ideal procedure for having sham surgical controls. Probably hundreds of operations have been performed for adrenal medulla autografts, with varying degrees of reported success. It was only after several years went by and autopsies showed no tissue survival that those performing and evaluating the procedure realized that these operations were failures. Motor improvement in patients with Parkinson disease is seen regularly in the clinic and in controlled clinical trials evaluating medications, and thus a placebo effect should have been anticipated. Transplants of fetal dopaminergic tissue have also been performed in dozens of trials, typically showing benefit. The benefit could be explained by the increase in [18F]fluorodopa uptake in positron emission tomographic scans. The first sham-controlled surgical trial for such procedures2 confirmed success in reducing Parkinson disease signs and symptoms in the practically defined "off" state, but only in subjects aged 60 years or younger. Older subjects had the same degree of positron emission tomographic scan improvement but without clinical benefit, indicating that [18F]fluorodopa uptake alone is insufficient to account for clinical benefit. The study also showed that approximately one third of the sham-operated group, both younger and older subjects, rated themselves as moderately to markedly improved. This points out the powerful placebo effect from simply having the operation, and this must be taken into account when trying to evaluate any open-label surgical procedure.
Fahn S. The Place of Placebos/Sham Surgery in Clinical Trials. Arch Neurol. 2002;59(2):325–326. doi: