Mavacamten is the first-in-class CMI. It revolutionized the treatment of patients with oHCM. Safety and efficacy of mavacamten were derived from rigorously conducted clinical trials, which usually included core-lab confirmation of echocardiographic diagnosis of oHCM. However, since commercial use is not subjected to the same rigorous patient selection and monitoring, there is a need to understand the safety of mavacamten in a large, unselected patient cohort. In this current study, patients with oHCM treated with mavacamten had frequent acute care encounters, with the majority related to heart failure and/or atrial arrhythmias. Patients had significant utilization of additional pharmacotherapies including oral anticoagulation, anti-arrhythmic drugs, and diuretics. In addition, a significant number of patients discontinued mavacamten during follow-up even when a stringent 90-day interruption period was used.

Mavacamten was approved by the FDA in 2022 and by the EMA in 2023. Due to its potential to cause systolic dysfunction and heart failure, it was required to have a REMS program in the United States or metabolizer status genotyping and monitoring in Europe^3,4. The REMS monitoring program is onerous and uses echocardiography as a surrogate to mavacamten drug-level, since the drug level was used in the pivotal phase III EXPLORER-HCM trial, but was not required by the FDA for commercial use. In addition, drug-drug interactions due to mavacamten metabolism pathways require dispensing pharmacies to review patient’s medications and supplements at each dispense. In Europe, a distinct monitoring program is recommended which also includes CYP genotyping.

The instituted REMS monitoring program is distinct from how the clinical trials were run, and as such, commercial use of mavacamten requires further scrutiny. Despite almost 3 years since the approval of mavacamten in the United States, there is a paucity of large-scale evaluations of patient outcomes. The single-center studies are informative, but they suffer from small sample size, and selection and reporting biases^12–15. In 2023 a real-world short-term retrospective cohort study of 659 patients followed for 118 days showed high adherence to mavacamten¹⁶. In addition, the severe adverse events from mavacamten are expected to be uncommon, therefore, a large representative sample is required.

Recent data published from the FDA REMS program showed a 6.1% discontinuation rate of mavacamten after the first fill, 4.6% of patients had LVEF <50%, 1.3% had heart failure hospitalization, and 0.3% had both LVEF <50% and heart failure hospitalization. However, these data do not account for patients who discontinued mavacamten permanently due an adverse event. As such, there inherent survival and reporting biases to such data since there is no obligation to report on patients who simply discontinued mavacamten without plans to resume and therefore the need to fulfill the REMS program requirement⁸. A better suited database for this question is one that allows for understanding of the events that occur to patients who permanently discontinued or temporarily held mavacamten.

Mavacamten has already been adopted by the American College of Cardiology/American Heart Association guidelines as a second-line therapy for symptomatic oHCM, adjacent to disopyramide and septal reduction therapy (SRT), while the European Society of Cardiology guidelines recommend it as a second-line therapy prior to proceeding with SRT^3,4. The EXPLORER-HCM and VALOR-HCM trials have shown the ability of mavacamten to improve exercise capacity, symptoms, and avert the need for SRT^(8,18). However, only VALOR-HCM trial based dose titration solely on echocardiographic monitoring, which resembles real-world practice. In VALOR-HCM, over 112-128 weeks of mavacamten treatment, 15 (13.8%) patients developed LVEF <50%⁹.

In the current study, our goal was not to re-define the incidence of LVEF <50% on mavacamten. We sought to study healthcare utilization in the acute care settings as well as new diagnoses of interest and discontinuation rates. Another goal is to use pharmacy claims to study disease progression and adverse events. In HCM, symptoms suggestive of heart failure directly overlap with symptoms related to HCM; therefore, other tools are required to study disease evolution in patients. An important surrogate of patients’ well-being in HCM is the new use of loop diuretics, which signals the transition to a more complex disease state. While loop diuretics use should not equate heart failure or progressive end-stage HCM, it is a reasonable surrogate of how the disease is evolving. In our study, loop diuretics were newly introduced in 154 patients (6.0%) after mavacamten had been initiated.

Aside from heart failure, AF and AFL, and their respective treatments, are emerging as a significant contributor to morbidity while on mavacamten treatment, and beyond. Developing new-onset AF/AFL mandates lifelong anticoagulation in patients with HCM irrespective of other risk factors, and as such, the morbidity associated with this diagnosis carries over a lifetime. AF/AFL are commonly encountered in HCM, with a prevalence as high as 20-30%, and a typical incidence of 2% per year^17–20. It remains unclear if there is an association between mavacamten, or cardiac myosin inhibition, and AF/AFL. However, there are some clues that the incidence of new-onset AF/AFL might be higher than expected, especially with how effective mavacamten is in lowering LVOT gradients and improving left atrial volume²¹. In a study of human atrial tissue, mavacamten inhibited the α-myosin heavy chain in the atria at a similar rate to the β-myosin heavy chain found in the ventricle, with resultant reduction in atrial twitch amplitude and possibly depressing atrial contractility²². Whether these findings translate into humans with HCM remains to be seen.

Data form clinical trials provide further insights. AF and AFL occurred in 7.7% and 7.1% of patients on mavacamten compared to 6.5% and 0% on placebo in the EXPLORER-HCM and VALOR-HCM trials, respectively. The long term extension study MAVA-LTE showed 33 (14.3%) events of AF/AFL with 18 (7.8%) being new-onset over 166 weeks of follow up, and the long-term extension study of VALOR-HCM showed new-onset AF/AFL in 11 patients (10.2%) over 128 weeks^7,9. In the current study, new AF/AFL was common and occurred in 125 patients (5.1%), leading to a 7.6% annualized rate of AF/AFL in the first year. The majority of new-onset AF/AFL episodes required a rhythm control strategy with an initiation of oral anticoagulation. Combined with prior findings from mavacamten trials, these findings represent an urgent need for an in-depth study of this phenomenon. Further data are needed to understand if AF/AFL episodes are related to duration of mavacamten exposure, metabolizer status, hemodynamic change, effect on atrial myosin, or if these are simply representations of a common arrhythmia that occurs in patients with oHCM.

Finally, discontinuations of mavacamten were substantial. We chose to focus on 90 days or more in order to have a more stringent and specific criteria. Using this cut-off, 412 (16.9%) patients had discontinued mavacamten after their first episode without an additional fill, while an additional 85 (3.5%) individuals had more than 90-day interruption after an initial episode but restarted mavacamten for at least one additional fill. These longer-term discontinuations likely signal permanent discontinuation of mavacamten. While temporary hold of mavacamten typically lasts 4-6 weeks for LVEF<50%, we felt that it is rather difficult to discern these short-term temporary holds.

Our study has multiple unavoidable limitations. First, using claims data provided access to a large representative sample across the country at the expense of granular phenotypic, laboratory and echocardiographic data. The goal of this analysis is to quantify patterns of mavacamten use, complications, and acute care utilization. Despite almost 3 years of commercial availability of mavacamten, no large-scale real-world studies have been conducted. Secondly, while the sample size is large enough for a disease like HCM on a specialty drug such as mavacamten, this database does not capture all patients in the United States who received mavacamten. Thirdly, we are unable to assess mortality in this database. The Symphony database was specifically used to access details of pharmacy claims that provided us with significant insights into complications and morbidity while on mavacamten using fill claims of new medications. In addition, we manually reviewed all the acute care claims codes and timelines, classifying them into the categories in Table 3. Finally, we did not include a control group due to the inherent heterogeneity of patients who are selected to receive mavacamten vs other traditional therapies oHCM.