Placebo or active comparator in chronic migraine trials: Balancing evidence and practicality

Introduction

Position 1: We still need placebo comparators

Definition of standard of care. The use of an active comparator is acceptable if there is robust and reproducible evidence from quality-controlled superiority trials demonstrating its efficacy over placebo, a position also endorsed by the IHS.^17,18 Fortunately, such comparators are available in the treatment of CM and include both specific and non-specific treatments. However, a more critical and complex challenge lies in defining what constitutes the standard of care. Which treatments are considered standard, and is there international consensus on this standard?

CGRP pathway inhibitors could be favorable active comparators due to their specific mechanism of action, efficacy, and better tolerability compared to traditional preventives such as amitriptyline, valproate, and topiramate, which are associated with high dropout rates. ¹⁹ This supports the argument for CGRP pathway inhibitors as a first-line treatment option and, by extension, as standard active comparators in RCTs. ²⁰ However, caution is warranted. Classical treatments like beta-blockers and candesartan may offer comparable efficacy with fewer side effects than amitriptyline, valproate and topiramate.^21,22 Additionally, migraine-specific treatments remain largely inaccessible due to high costs and strict national reimbursement requirements, which often mandate prior failure of multiple other preventive treatments, thereby limiting widespread use. ¹⁹ Labeling them as standard of care presents an ethical dilemma, potentially conflicting with the IHS inclusive mission, as access to those treatments is predominantly restricted to wealthier countries. ²³ Designating them as the standard care could exacerbate existing global disparities in the ability to conduct clinical treatment studies across different regions. Furthermore, benchmarking new treatments against those targeting the CGRP pathway may escalate healthcare costs; in addition to demonstrating efficacy and tolerability, emerging drugs would also need to show a favorable incremental cost-effectiveness ratio compared to these already expensive treatments, further driving up costs.^24,25 The increasing costs would hamper the development of novel cheap treatments accessible to underserved populations. Because migraine treatments lack universal standardization, choosing a global comparator is challenging. Using active comparators instead of placebo may yield inconsistent interpretations on treatment choices partly influenced by regional and individual preferences of clinicians.

Placebo response management. Migraine studies lack hard clinical endpoints such as stroke or death; instead, they rely on measures such as monthly migraine days, monthly headache days, and acute medication days. These outcome measures may be more susceptible to placebo effect than objective clinical outcomes, due to influences like expectation, attention, anxiety, and prior experiences. ²⁶ It is crucial to distinguish between placebo effect, which is the sum of all psychological and contextual factors and the placebo response, which is the observed effect in the placebo group of a trial. The placebo effect also includes non-specific effects, such as natural course of disease, regression to the mean, measurement artifacts, and the Hawthorne effect (behavior changes from being observed). These influences occur in both trial arms, leaving the pharmacological effect as the only element unique to the treatment arm. In migraine studies there is a high correlation (r = 0.85) between placebo and treatment effect, suggesting that only one third of variance is attributed to pharmacological action. ²⁷ This demonstrates that the placebo and treatment response in trials are highly interdependent. In comparator trials, both arms have pharmacological effects and placebo responses, making it even harder to isolate treatment efficacy. This overlap further complicates the interpretation of results, as it becomes challenging to discern whether observed improvements are due to the active treatment or the psychological and contextual factors influencing both groups, leading to potentially inflated estimates of treatment efficacy. Besides, assuming that active comparators are superior to placebo might not be true for any RCT, as the active comparator might underperform because of design flaws, dosing errors, participant adherence issues, unusually high placebo responses, biological differences in the patient population compared to previous RCTs, or insensitivity of the outcome measures.

Scientific validity. Given their precise account on the absolute effect of a treatment and of the placebo response, which is high in individuals with CM, placebo-controlled RCTs have a high internal validity.

Trial feasibility. Comparing a treatment with placebo is overall easier than comparing two active treatments. A relatively small sample size is required to show the effect of an active treatment over placebo. On the contrary, demonstrating the superiority or even non-inferiority of the tested treatment over an active comparator would require large sample sizes especially if an active comparator is highly effective. Besides, designing a RCT with an active comparator would imply issues in the mode of delivery and in the assessment of outcomes between different treatments. Comparing an oral with a parenteral drug would imply the assessment of the different contextual effects and of different parameters of medication adherence.

Participants’ recruitment. Placebo-controlled RCTs can deprive a group of participants of a potentially effective treatment. However, there is a key distinction between ‘no treatment’ and placebo, especially given the substantial placebo response in migraine trials.^26,27 Moreover, according to IHS guidelines, participants with CM may continue on their stable preventive medication during the study. Thus, even in placebo-controlled RCTs, patients may not be left untreated. Recruitment may also benefit from other designs. For example, including open-label extensions after the primary endpoint ensures all participants eventually receive the active treatment.^28,29 Another strategy may involve incorporating non-pharmaceutical interventions in both the placebo and active comparator groups. Two studies included behavioral advice for patients with CM and MOH, aiming to assess the impact of behavioral interventions alongside the pharmacological effects of onabotulinumtoxinA or eptinezumab compared to placebo.^30,31 The design can be adapted accordingly, for example, comparing treatment A + placebo (control group) versus treatment A + treatment B (intervention group). Furthermore, by carefully selecting the study population, such as patients who failed previous treatments, placebo-controlled trials can yield insights into switching strategies that closely reflect real-world clinical practice. ³² As a final point, it should be noted that individuals with CM often have a history of many preventive treatment failures; therefore, including them in an RCT with an active comparator might pressure them to persist with or initiate pharmacological treatment known to be ineffective for them, thereby introducing possible selection bias.

Clinical relevance. Placebo-controlled RCTs provide valuable evidence not only for evaluating single treatments, but also for assessing combination therapy.

Ethical considerations. The Declaration of Helsinki requires that all study participants receive the best proven therapy, which is often interpreted as excluding placebo; however, this standard is equally compromised in active-comparator trials where patients may receive an investigational drug instead of established treatments. ³³ Since preventive treatment for CM does not influence long-term health outcomes, and placebo-controlled studies are typically short-term with fully informed and consenting participants, the use of placebo in such trials is not ethically unjustified. ³⁴

Regulatory acceptance. The clear demonstration of efficacy and safety of preventive treatments guaranteed by placebo-controlled trials is useful for the first approval of new drugs. As no international standard of care exist for migraine prevention, placebo-controlled RCTs are needed to determine whether newly tested medication can enter the market with adequate efficacy and safety data.

Position 2: We need active comparators

Definition of standard of care. Preventive migraine treatments that were repurposed from different indications, including antidepressants, anti-hypertensive agents, and anti-seizure medications, might not be considered standard of treatment as they were not specifically designed for migraine prevention. OnabotulinumtoxinA, despite being tested for CM ³⁵ and having a pathophysiological rationale for its use, ³⁶ is not migraine-specific. On the other hand, CGRP pathway inhibitors act on a mechanism related to migraine; besides, they have a good efficacy and excellent tolerability. As such, they are about to set a standard of treatment due to their robust and reproducible superiority over placebo in RCTs, and according to the European Medicine Agency definition ¹⁸ endorsed by the International Headache Society. ¹⁷ Current recommendations for RCTs of migraine prevention affirm the importance of performing placebo-controlled trials as there is the possibility that active comparators are less effective than placebo. ¹⁷ This consideration seems to be driven by the lack of specificity of most migraine preventive drugs, which might not act on migraine-specific mechanisms, but also from the very high placebo effect of preventive migraine treatments.^12,26 However, this view might be challenged by CGRP pathway inhibitors as they are not only migraine-specific, but also highly effective and well tolerated. Notably, a drug may appear less effective than placebo if it has a higher discontinuation rate due to adverse events, a problem that seems minimal with CGRP pathway inhibitors in the short and medium term.

Placebo response management. Similar to placebo-controlled RCTs, in trials with active comparators the placebo response is present in both arms. At variance in the active comparator design the actual response to the drugs is mixed with the placebo response. Therefore, RCTs with active comparators assess the global response, including placebo, to the tested treatment compared with the global response to the comparator. This might dilute the effect size of a treatment with respect to its comparator as both arms would include a placebo response.

Scientific validity. As RCTs with active comparators provide data on the relative efficacy and safety of one treatment over another, they are expected to detect mostly small differences between groups, which should be considered when evaluating their results. However, non-inferiority to an active comparator may still justify the clinical use of the treatment.

Trial feasibility. Designing RCTs with an active comparator would imply changes compared with placebo-controlled RCTs. In a two-arm RCT with an active comparator, blinding requires that each participant receive both a placebo and one of the two active treatments, either the tested treatment or the active comparator. A simpler strategy than a blinded trial would be to perform open label RCTs with a pragmatic design. ³⁷ However, pragmatic trials usually compare two treatment strategies that are both clinically valid and are not suitable to test novel drugs. Referring to outcomes, trials with active comparators would not only consider superiority, but also non-inferiority, depending on the profile of the tested treatment with respect to the active comparator. Caution is needed with non-inferiority trials, as weaker incentives for strict trial conduct mean that poor adherence, imprecise measurements, or protocol deviations can artificially favor positive results, increasing the risk of falsely concluding for the efficacy of a treatment. ³⁸ Besides, if newly tested drugs act on the CGRP pathway, the use of active comparators might increase the difficulty in the measurement of effects due to common mechanisms of action.

Participants’ recruitment. From the point of view of individuals with CM, the potential inclusion in the placebo arm of a RCT might be a reason to refrain from participation. Having active comparators in RCTs may facilitate inclusion of individuals.

Clinical relevance. At variance with placebo-controlled RCTs, which are focused on the absolute effect of treatments, those with active comparators would align with clinical practice by providing information on the added value of new treatments over current care.

Ethical considerations. Having a standard of treatment raises an ethical issue as to whether the presence of a placebo arm ensures participants’ protection in RCTs. As CM is associated with the highest degree of disability in the migraine spectrum, ³⁹ both clinicians and affected individuals require timely and effective interventions. There might be an ethical issue in denying or delaying effective treatments to individuals with CM who are those most in need of those treatments. Additionally, active comparators such as CGRP pathway inhibitors would help to better address the individuals’ need for effective treatments. Trials of new compounds compared with treatments targeting the CGRP pathway would encourage finding the best possible solutions for migraine prevention and optimize its control. ²⁰

Regulatory acceptance. While current recommendations advise against the use of active comparators for RCTs of migraine preventive treatments, this might not be true for regulatory agencies. According to the European Medicine Agency ⁴⁰ and the American Food and Drug Administration, ⁴¹ placebo control is considered necessary only for first-in-class treatments, while active comparators can be adopted when there is a standard of care for diseases.

Final considerations: reconciling perspectives

Figure 1 summarizes the arguments in favor of placebo control groups as well as active comparators in RCTs for CM prevention. Individuals with CM typically exhibit a high placebo response, which should be quantified for any preventive treatment; this consideration supports the use of placebo controls in trials. However, ethical concerns may arise when comparing preventive migraine treatments with placebo in individuals with CM. Regulatory agencies may approve new treatments when tested against either placebo or the standard of care. Should CGRP pathway inhibitors be recognized as a standard of care for CM prevention, they could serve as active comparators in RCTs; however, it remains challenging to regard these drugs as a definitive standard of care for migraine prevention. Using an active comparator in RCTs could mitigate the ethical issue of withholding treatment from individuals with CM who have a significant need for effective management of their condition.

OPEN IN VIEWER

Figure 1.

Characteristics of randomized controlled trials with placebo control or an active comparator for chronic migraine prevention.

Shifting from traditional placebo-controlled RCTs to RCTs with active comparators would imply a change in the design of those studies. Should CGRP pathway inhibitors be considered as active comparators, new treatments would face challenges in their testing, considering the high efficacy and tolerability of CGRP pathway inhibitors that makes it difficult to find treatments with higher efficacy and safety at least in the short term. Nevertheless, CM is a severe condition for which it is important to set ambitious prevention goals ²⁰ and comparisons against placebo might lead to the approval of treatments of modest efficacy – potentially lower than CGRP pathway inhibitors. Another solution to this problem is to set a clinical minimal important difference for treatment effect in individuals with CM. ⁴² A potential compromise is a three-arm RCT comparing the investigational treatment, placebo, and an active comparator, testing superiority over placebo and non-inferiority to the active comparator. However, this design requires larger sample sizes, involves participants receiving placebo, and retains the methodological limitations inherent to non-inferiority trials, as discussed previously. Open-label extension after a double-blind comparison with placebo is an alternative way to maximize participation to placebo-controlled RCTs. Open-label extensions might be useful to test long-term efficacy and safety of CM preventive treatments in a controlled setting, even if they do not make comparisons between treatments and placebo or among different treatments. For trials on CM, the duration of double-blind placebo-controlled phases might not exceed the minimal duration of 12 weeks ¹⁷ to show the efficacy of a treatment against placebo, while open-label phases without placebo control might provide safety information. An interesting RCT design that explicitly accounts for the interaction between the pharmacological effects and placebo responses is the balanced placebo design ⁴³ which allows expectancy effects to be disentangled from pharmacological effects in both the placebo and treatment groups. This design option may be particularly relevant in chronic migraine research, where outcomes are largely subjective. ⁴⁴

As an alternative, if allowed by their mechanism of action and by interactions, new treatments may be tested as add-on to current treatments, so that participants in the placebo group are not without treatment; this is particularly relevant for individuals with CM, as they are treated by several clinicians with combined agents to achieve optimal prevention. A key point in designing RCTs with active instead of placebo comparators should encourage the development of strategies to provide meaningful results instead of discouraging the search for novel migraine treatments.

In conclusion, the role of the placebo group in future of trials of CM prevention might be reconsidered. Those trials should prioritize the optimization of treatment in individuals with CM and participants’ protection while maintaining feasibility and rigorous methods for the ascertainment of outcomes.

Key findings

Chronic migraine is highly disabling, and new targeted therapies challenge the ethics of placebo-controlled trials.