by internal mammary artery ligation produced 65% to 91% subjective improvement in four uncontrolled trials, but later was found in two placebo-controlled trials to have no effect (see review 2 ). Similarly, the implantation of the internal mammary artery into the heart for revascularization in angina patients yielded subjective improvement in 85% of patients, but this was later shown to have no correlation with graft patency or cardiac revascularization. 2 Interestingly, some patients showed objective evidence of improved exercise tolerance or cardiac function with these procedures, suggesting that the placebo response can also affect physiological function. With respect to the study of shunting for iNPH by Kahlon and colleagues, 3 I apologize for misquoting the duration of the effect of the shunting (i.e., 96% of the operated patients did indeed report subjective improvement at 6 months, but around half that number at 5 years). However, it is important to recognize that the Kahlon study contained a very interesting "nocebo" group (neither the doctor nor the patient thought that the patient had been treated), that is, a set of patients who were referred for shunting, but did not meet criteria for either the lumbar infusion test or cerebrospinal fluid removal test, and therefore did not receive operations. Interestingly, at the 6-month follow-up, 10% of the nocebo patients had improved on a walking step test, 24% on a timed walking test, 48% on a reaction time test, and 28% on a memory test. At 5 years, 1 of the 4 nocebo patients who was followed up still showed improved walking and reaction times. So, the improvement in subjective or even objective performance after shunting procedures for iNPH must be compared to a noisy baseline, in which substantial numbers of unoperated patients may show improvement, and a very strong surgical placebo effect. Together, these can more than account for the results of the uncontrolled iNPH studies.As I stated in my editorial, no drug would ever be used without evidence for efficacy in a randomized, placebocontrolled, clinical trial, particularly if it had a known rate of serious adverse events of greater than 10% in the first 6 months. The same standard should be applied to shunting procedures for iNPH: There should be a moratorium on such procedures until such time as randomized, placebo-controlled studies, such as I outline in my editorial, have been done.