Once upon a time, treatment of hypertrophic cardiomyopathy (HCM) was based on observational data, personal experiences, and strong opinions. Then one day, 2 pharmaceutical companies came along and developed 2 marvelous drugs that enthralled everyone. In hospitals all over the world, cardiologists engaged in studies with these drugs, leaving the entire HCM community, including authors of this editorial, with enthusiasm and belief in a new era for management of this disease.
Growing up in Denmark, the country of the worldfamous fairytale writer Hans Christian Andersen, who in 1837 wrote the fairy tale “The Emperor’s New Clothes,” we have learned to ask the question: “Is the emperor wearing any clothes?” Are myosin inhibitors in fact new wonder drugs?
In this debate, we congratulate Maron et al[1] for the important work presented in this issue of JACC. Maron et al[1] present prespecified post hoc analyses of the randomized controlled SEQUOIA-HCM (Phase 3 Trial to Evaluate the Efficacy and Safety of Aficamten Compared to Placebo in Adults With Symptomatic oHCM) trial investigating the effect of treatment with the myosin inhibitor aficamten vs placebo in patients with obstructive hypertrophic cardiomyopathy (hHCM).[2] The investigators report data about the proportions of patients who obtained one or more of clinically relevant objective measures and patient reported outcomes. At 24 weeks, aficamten treatment showed remarkable and significant improvements in all outcomes compared to placebo. Comprehensive analyses of different outcome measures from the SEQUOIA-HCM trial support clinical decision-making. We can now discuss the expected response to treatment in more detail with our patients. Two in 7 patients (29%) will experience symptom relief on aficamten treatment, whereas 61% and 76%, respectively, will get hemodynamic and biomarker responses that we believe are beneficial effects that may improve long-term left ventricular remodeling and prognosis.[3] However, we are still awaiting documentation of these long-term effects and their clinical relevance for patients with oHCM.
Maron et al[1] report improvements of any one of the outcome measures on aficamten treatment in 97% of patients. Is this the effect of aficamten alone? Looking at the placebo group, we find that 59% of patients had similar improvements, downscaling the true effect of aficamten to 38% of patients. This shows that it is highly important to analyze and interpret results of the placebo group in randomized trials. Failing to do so reduces the trial to observational analyses. We congratulate the authors, investigators, and sponsors for the great efforts that everyone has put into screening, randomization, and blinding in the myosin inhibitor trials. The strengths of these trials are the randomization and blinding; hence, the treatment effect is the difference between the 2 treatment arms. As such, Maron et al[1] report the treatment effects relative to placebo in their Table 3: complete hemodynamic response 61%, symptom reduction 29%, peak oxygen uptake 22%, and N-terminal pro–B-type natriuretic peptide 76%. These numbers reflect that the closer we measure to the mechanism of action of aficamten (eg, the left ventricular outflow tract gradient and N-terminal pro–B-type natriuretic peptide), the more likely we are to observe an effect. Oxygen uptake and symptoms are more complex measures that are also affected by age, comorbidity, lifestyle, body weight, concomitant medication, and degrees of cardiomyopathy, which tend to reduce the signal-to-noise ratio in these analyses.
There seems to be an apparent treatment effect in the placebo group of the SEQUIA-HCM trial. A similar placebo effect was observed in the EXPLORER-HCM (Clinical Study to Evaluate Mavacamten [MYK-461] in Adults With Symptomatic Obstructive Hypertrophic Cardiomyopathy) trial comparing mavacamten vs placebo. The complex primary combined endpoint ($1.5 mL/kg/min increase in venous oxygen partial pressure and $1 NYHA functional class improvement, or $3.0 mL/kg/min increase in venous oxygen partial pressure and no worsening of NYHA functional class improvement) was observed in 37% and 17% of patients in the treatment and placebo groups, respectively. [4] We acknowledge the authors for providing exactly the same analysis in the SEQUIA-HCM trial (see their Table 2), in which they found essentially equal numbers of 42% and 14%, respectively. The SEQUOIA-HCM and EXPLORER-HCM cohorts were very similar; however, there were small differences in screening procedures and baseline characteristics that may have affected the results. Both trials compared changes from baseline to treatment between active treatment and placebo groups. Small differences in baseline characteristics and patient adaptation to the complex cardiopulmonary exercise test setting may have affected the primary endpoints.
One way to handle this issue is to design crossover trials with randomly assigned treatment sequence similar to the randomized trial of metoprolol vs placebo in oHCM (the TEMPO [The Effect of Metoprolol in Patients With Hypertrophic Obstructive Cardiomyopathy] trial).[5] This allows paired analyses to compensate for interpatient variations and direct comparison of outcomes between treatment groups with reduced statistical noise as opposed to comparing changes calculated from baseline.
There is a large degree of day-to-day fluctuations in symptoms and left ventricular outflow tract (LVOT) gradients in oHCM patients. They often report “good and bad days.” Eligibility criteria for both the EXPLORER-HCM and SEQUOIA-HCM comprised a certain degree of impairment of parameters that form the primary endpoints (eg, Kansas City Cardiomyopathy Questionnaire score, LVOT gradients, NYHA functional class scores, and peak oxygen uptake). Thus, there was a risk of including patients mostly on their “bad days,” excluding them on their “good days,” and allowing both treatment arms to result in spontaneous improvement due to day-to-day fluctuations. The question is whether this effect in the placebo group blurred the true effect of myosin inhibitors; ideally, future trials should seek to minimize this phenomenon of regression towards the mean.
Traditionally, trial procedures are performed in a single visit, but measurements of highly fluctuating parameters may benefit from repeated evaluation and prolonged observation. The importance of this was evident in the VALOR-HCM (A Study to Evaluate Mavacamten in Adults With Symptomatic Obstructive HCM Who Are Eligible for Septal Reduction Therapy) trial (mavacamten vs placebo in surgical candidates), in which 23% of patients were no longer eligible for septal reduction therapy after 16 weeks of placebo treatment.[6] This suggests that approximately 1 in 4 patients eligible for septal reduction therapy should never undergo invasive treatment. However, this may not be true in real-life practice. Such observations from trials underscore the value of repeated evaluations during titration of medical treatment and assessment for septal reduction therapy to get an idea of the general level of symptoms and LVOT gradients before deciding on an invasive strategy. Importantly, the results are a warning against fast-lane evaluation for invasive treatments.
Is the emperor wearing any clothes, or is it a placebo? The myosin inhibitors aficamten and mavacamten do have significant and clinical meaningful effects in all available clinical trials. However, the substantial treatment effects of placebo in these randomized trials leaves us with a question about the magnitude of these treatment effects. Ideally, modification of trial designs, screening procedures, and selection of endpoints could minimize the treatment effects in the placebo groups in future trials. We should refrain from using the same highly fluctuating parameters in both eligibility criteria and endpoints. When used, such parameters should be evaluated repeatedly and possibly on different days to reduce the impact of extreme results. Statistical noise can potentially be minimized using crossover trials to address the interpatient variability, and by comparing actual measurements rather than changes in measurements between treatments groups. Other improvements to trial designs may be relevant as well and, finally, we must develop clinically feasible endpoints that are of relevance to patients’ everyday life because that is what we are trying to improve.
J Am Coll Cardiol. 2024;84(19):1832–1834 © 2024 American College of Cardiology Foundation
Comments