Placebo controlled trials in neuromyelitis optica are needed and ethical.

Currently, there are no approved treatments for NMO. All therapeutic studies in NMO have been either small, retrospective case series or uncontrolled prospective studies. Such studies are susceptible to inherent biases. As a consequence, conclusions regarding efficacy and safety from these studies may be erroneous. The optimal method for assessing therapeutic efficacy is the prospective, controlled trial with random treatment assignment that has the potential to control for multiple sources of bias. There is a significant unmet need for well-designed clinical trials in NMO. Successfully conducted, well-controlled NMO trials that show proof of benefit will lead to regulatory approval and subsequent acceptance by payers resulting in broad therapeutic availability. The most direct method to prove efficacy is to compare an active treatment vs. no treatment or placebo control. However, because of the devastating nature of the disease some clinicians are reluctant to expose potential study patients to the risk of no treatment. The primary ethical concern in the case of placebo-control in NMO clinical trials rests on the relative merits of answering the scientific question regarding efficacy compared to the relative risk of exposure to harm in the placebo-control group. This article outlines the case for clinical equipoise in NMO by addressing the uncertainty regarding the relative scientific and clinical merits of current empirically used treatments and showing that a placebo arm is consistent with competent medical care. Because no currently available treatment has proven benefit, and because all therapies are known to potentially cause harm, placebo-control is not only ethical but is in some ways preferable to active comparator or add-on study designs. Without well-designed, placebo-controlled trials, NMO patients may not have access to new treatments and will never know whether the therapies that they may be currently taking have risk to benefit profiles that clearly favor their use.