Aficamten for Symptomatic Nonobstructive Hypertrophic Cardiomyopathy (ACACIA-HCM)
Authors, Journal, Affiliations, Type, DOI
- Authors: Ahmad Masri (first author), Martin S. Maron, Ankit Bhatia, Maria Luisa Peña-Peña, Mark V. Sherrid, Estevao Lanna Figueiredo, Michael Arad, Roberto Barriales-Villa, Lubna Choudhury, Brian L. Claggett, Caroline J. Coats, Edileide de Barros Correia, Juan Pablo Costabel, Perry M. Elliott, Jorge E. Silva Enciso, Pablo Garcia-Pavia, Junbo Ge, Sheila M. Hegde, Gregory D. Lewis, Renato D. Lopes, Matthew W. Martinez, Mathew S. Maurer, Zi Michael Miao, Michelle Michels, Sumeet S. Mitter, Jesus E. Pino Moreno, Iacopo Olivotto, Anjali T. Owens, Steen H. Poulsen, Florian Rader, Ethan J. Rowin, P. Christian Schulze, Robert Sepp, Nikhil Sikand, Scott D. Solomon, John A. Spertus, Charles Teale Sr, Xiaoyan Zhao, and Carolyn Y. Ho (senior author), plus nine Cytokinetics employees, for the ACACIA-HCM Investigators
- Affiliations: Oregon Health and Science University, Portland (lead); Brigham and Women's Hospital and Harvard Medical School, Boston; Lahey Hospital and Medical Center, Burlington MA; plus sites in Spain, Brazil, Israel, the United Kingdom, Germany, the Netherlands, Denmark, Italy, Hungary, China, Japan and elsewhere; Cytokinetics, South San Francisco (sponsor)
- Journal: New England Journal of Medicine, published online 28 August 2026 (early release; continuous page numbering)
- Type: Phase 3, multinational, double-blind, randomized, placebo-controlled trial
- DOI: 10.1056/NEJMoa2603021
- Registration: NCT06081894 (ACACIA-HCM)
- Funding: Cytokinetics (sponsor); editorial support funded by Cytokinetics
Overview
ACACIA-HCM is the first phase 3 randomized trial to meet its primary efficacy endpoints in symptomatic nonobstructive hypertrophic cardiomyopathy, a phenotype for which no pharmacotherapy had previously shown benefit. In 517 patients across 160 sites in 19 countries, aficamten titrated to a maximum tolerated dose of 20 mg improved the Kansas City Cardiomyopathy Questionnaire clinical summary score by 3.0 points and peak oxygen uptake by 0.67 ml/kg/min over placebo at 36 weeks, with a 57% reduction in NT-proBNP and a 14.1-percentage-point absolute increase in the proportion improving at least one NYHA class. The effect sizes are materially smaller than those aficamten achieves in obstructive disease, and reversible reduction of left ventricular ejection fraction below 50% occurred in 10.5% — threefold the rate in the obstructive trials. The result stands in direct contrast to ODYSSEY-HCM, in which mavacamten missed both co-primary endpoints in the same phenotype.
Keywords
Hypertrophic cardiomyopathy, nonobstructive HCM, aficamten, cardiac myosin inhibitor, peak oxygen uptake, cardiopulmonary exercise testing, KCCQ-CSS, NT-proBNP, left ventricular ejection fraction, LVOTO, exercise capacity, randomized controlled trial
Key Takeaways
Background and Rationale
- Nonobstructive HCM lacks left ventricular outflow tract obstruction and affects an important subset of all HCM; a proportion of these patients have limiting symptoms and "no effective therapies are available"
- A fundamental feature of HCM is excessive actin–myosin cross-bridge formation causing hypercontractility and diastolic dysfunction; without obstruction, the clinical consequences arise from increased intracardiac pressure, microvascular ischaemia, impaired cardiac reserve and impaired myocardial energetics
- Aficamten is a cardiac myosin inhibitor with a shallow dose–LVEF response relationship and a half-life permitting rapid escalation and reversibility; a potential benefit was seen in the phase 2 REDWOOD-HCM cohort 4
- The trial was designed to test whether that phase 2 signal held in a phase 3 population
Trial Design
- Phase 3, multinational, double-blind, placebo-controlled; 160 sites in 19 countries; randomization 1:1 by interactive web response system
- Stratified by presence or absence of persistent or permanent atrial fibrillation and by presence or absence of intracavitary obstruction
- Treatment period up to 72 weeks plus a 4-week safety follow-up after withdrawal; database locked 24 April 2026
- Dosing: aficamten 5 mg starting dose, with opportunities at weeks 2, 4 and 6 to increase by 5 mg increments to a maximum of 20 mg, guided by locally measured LVEF and evaluated by a cardiologist blinded to assignment
- Dual primary end points: change from baseline to week 36 in KCCQ-CSS and in peak oxygen uptake during cardiopulmonary exercise testing
- Secondary end points tested hierarchically: ≥1 NYHA class improvement, composite z score of peak oxygen uptake and VE/VCO₂ slope, proportional change in NT-proBNP, change in left atrial volume index (in patients without AF or flutter), and time to first adjudicated cardiovascular event
- Events adjudicated by an independent clinical events committee at Duke Clinical Research Institute; data independently verified at an academic core laboratory (Brigham and Women's Hospital)
- Sample size assumed differences of 5 points for KCCQ-CSS (SD 15) and 1.0 ml/kg/min for peak oxygen uptake (SD 3.0), giving ≥90% power at a two-sided alpha of 0.025 per primary end point; no interim analysis
Patients
- 948 screened; 517 randomized — 258 to aficamten and 259 to placebo — between 30 August 2023 and 28 March 2025
- One placebo-randomized patient never received either placebo or aficamten, so the safety population comprises 258 aficamten and 258 placebo
- 31 additional patients were enrolled at 22 Japanese sites between May and September 2025, after the global trial closed, and are not reported here
- Treatment period completed by 228 (88.4%) in the aficamten group and 233 (90.0%) in the placebo group
- At week 36, peak oxygen uptake was missing in 44 patients (8.5%) and KCCQ-CSS in 25 (4.8%)
Eligibility: age 18–85; nonobstructive HCM with maximum LV wall thickness ≥15 mm (≥13 mm with a known disease-causing variant or family history); resting LVOT gradient <30 mmHg and <50 mmHg with Valsalva; LVEF ≥60%.
Baseline Characteristics
| Characteristic |
Aficamten (N=258) |
Placebo (N=259) |
| Age — years |
55.7±16.2 |
54.5±16.0 |
| Female sex — no. (%) |
140 (54.3) |
137 (52.9) |
| Pathogenic or likely pathogenic sarcomere variant — no./total (%) |
124/197 (62.9) |
107/193 (55.4) |
| Family history of HCM — no. (%) |
105 (40.7) |
106 (40.9) |
| Beta-blocker — no. (%) |
203 (78.7) |
197 (76.1) |
| Non-dihydropyridine calcium-channel blocker — no. (%) |
31 (12.0) |
39 (15.1) |
| Implantable cardioverter–defibrillator — no. (%) |
107 (41.5) |
105 (40.5) |
| Paroxysmal atrial fibrillation or flutter — no. (%) |
78 (30.2) |
68 (26.3) |
| Persistent or permanent atrial fibrillation — no. (%) |
20 (7.8) |
19 (7.3) |
| NYHA class II / III — no. (%) |
171 (66.3) / 84 (32.6) |
183 (70.7) / 74 (28.6) |
| KCCQ-CSS |
66.3±16.2 |
65.0±15.4 |
| Median NT-proBNP (IQR) — pg/ml |
906 (508–1529) |
893 (492–1609) |
| Median high-sensitivity troponin I (IQR) — ng/litre |
30 (15–77) |
28 (14–113) |
| Peak oxygen uptake — ml/kg/min |
18.0±5.3 |
18.0±5.1 |
| Percent of predicted peak oxygen uptake |
55.3±13.1 |
54.9±12.7 |
| Resting LVOT gradient — mmHg |
9.1±6.1 |
8.2±4.6 |
| LVOT gradient after Valsalva — mmHg |
11.8±8.8 |
11.4±8.5 |
| LV ejection fraction — % |
67.9±4.0 |
68.4±3.4 |
| Maximum LV wall thickness — mm |
22.3±5.2 |
23.1±5.7 |
| Left atrial volume index — ml/m² |
36.8±10.9 |
37.5±11.9 |
- 441 patients (85.3%) were on background medical therapy for HCM, unchanged at week 36 because the protocol permitted medication changes only for a clinical indication
- Intracavitary obstruction was present in 48 patients (9.3%); 44 (8.5%) had a body-mass index between 35 and 40
Primary End Points
| End point (change to week 36) |
Aficamten |
Placebo |
Difference (95% CI) |
P |
| KCCQ-CSS — points |
+11.4 (9.6–13.2) |
+8.4 (6.6–10.2) |
3.0 (0.5–5.5) |
0.02 |
| Peak oxygen uptake — ml/kg/min |
+0.64 (0.32–0.95) |
−0.03 (−0.35 to 0.28) |
0.67 (0.22–1.11) |
0.003 |
- Both values are model-based least-squares means, after imputation of missing data
- No apparent treatment–covariate heterogeneity across prespecified subgroups for either primary end point
- The KCCQ-CSS benefit was sustained at approximately 5 points greater than placebo through 72 weeks, and symptoms returned toward placebo levels shortly after cessation, corroborating a pharmacological effect
- Subgroups of note: KCCQ-CSS difference +5.2 (1.7–8.7) in men vs +1.1 (−2.5 to 4.8) in women; +11.9 (3.3–20.6) in the 20 vs 19 patients with persistent or permanent AF; −2.7 (−8.2 to 2.9) in the 29 vs 20 patients with KCCQ-CSS >85 — while peak oxygen uptake in that last subgroup rose by 2.6 (0.9–4.3) ml/kg/min
Secondary End Points
| End point |
Aficamten |
Placebo |
Effect (95% CI) |
P |
| ≥1 NYHA class improvement — no. (%) |
108 (41.9) |
72 (27.8) |
common rate difference 14.1 pp (6.0–22.2) |
<0.001 |
| Composite exercise z score |
+0.05 (−0.01 to 0.12) |
−0.09 (−0.14 to −0.05) |
0.14 (0.07–0.22) |
<0.001 |
| NT-proBNP proportional change — geometric LS mean |
0.43 (0.39–0.46) |
1.00 (0.93–1.08) |
ratio 0.43 (0.38–0.47) |
<0.001 |
| Left atrial volume index — ml/m² |
−0.17 (−1.12 to 0.79) |
+1.13 (0.19–2.08) |
−1.30 (−2.64 to 0.04) |
0.06 |
| Time to first cardiovascular event — weeks (restricted mean) |
68.2 (66.2–70.2) |
68.8 (66.9–70.7) |
−0.6 (−3.3 to 2.2) |
— |
- The first three secondary end points in the hierarchical chain were positive; left atrial volume index and time to first cardiovascular event were not
- The NT-proBNP effect — a 57% relative reduction — is numerically similar to the 59% seen with mavacamten in ODYSSEY-HCM
- The trial was not powered to detect differences in clinical outcomes; the cardiovascular-event end point showed no separation
Dose Escalation and Attained Doses
- After the dose-adjustment phase at week 8: 1.6% on 0 mg, 6.2% on 5 mg, 5.8% on 10 mg, 22.9% on 15 mg, 57.8% on 20 mg
- By end of treatment: 3.5% on 0 mg, 6.6% on 5 mg, 10.5% on 10 mg, 23.6% on 15 mg, 55.8% on 20 mg
- The protocol targeted the maximum tolerated dose guided by LVEF, not the minimum effective dose — the strategy the authors identify as the explanation for the higher rate of systolic dysfunction than in obstructive trials
Safety
| End point |
Aficamten (N=258) |
Placebo (N=258) |
| Any adverse event — no. (%) |
207 (80.2) |
217 (84.1) |
| Any serious adverse event — no. (%) |
52 (20.2) |
38 (14.7) |
| Leading to discontinuation of trial drug — no. (%) |
18 (7.0) |
5 (1.9) |
| Leading to interruption — no. (%) |
19 (7.4) |
9 (3.5) |
| LVEF <50% (reversible) — no. (%) |
27 (10.5) |
2 (0.8) |
| LVEF <50% with clinical heart failure — no. (%) |
2 (0.8) |
0 |
| LVEF <40% — no. (%) |
14 (5.4) |
1 (0.4) |
- Serious adverse events during treatment included heart failure in 11 patients (4.3%) vs 1 (0.4%), new atrial fibrillation in 4 (1.6%) vs 3 (1.2%), and ventricular arrhythmias in 7 (2.7%) vs 12 (4.7%)
- Three deaths occurred, all in the placebo group
- LVEF: least-squares mean between-group difference −4.4% (−5.1 to −3.6) at week 36; returned to baseline after washout (−0.5%, −1.3 to 0.2)
- A reduction to <40% triggered a protocol-mandated interruption in 7 patients (2.7%), 2 of whom discontinued permanently; 2 patients with LVEF <50% had a concurrent serious heart-failure event
- No apparent between-group difference in new-onset atrial or ventricular arrhythmia
Interpretation
- The benefit is real but smaller than in obstructive disease. The authors attribute the gap to pathophysiology: in obstructive HCM aficamten produces "rapid resolution of excess afterload," whereas nonobstructive HCM has "more-complex myocardial pathologic characteristics, including multifactorial drivers of symptoms and adverse events with exercise"
- They argue the KCCQ-CSS improvement nonetheless exceeds that seen in other pivotal heart-failure trials on a population-mean basis, and note concordance between patient-reported and physician-assessed change
- They compare with ODYSSEY-HCM directly: more than 20% of mavacamten-treated patients had treatment interrupted for systolic dysfunction, against 3% in ACACIA-HCM — offered as a within-nonobstructive comparison of the two agents
- Positioning: nonobstructive HCM patients "have no proven medical or invasive treatment options other than heart transplantation"
Limitations of the document
- Efficacy was assessed at 36 weeks only, so durability is unknown; the authors defer this to the ongoing FOREST-HCM open-label extension (NCT04848506)
- The trial was not powered for clinical outcomes — heart-failure hospitalisation, death and the composite cardiovascular-event end point cannot be interpreted as treatment effects
- Exercise echocardiography was not required, so patients with exercise-induced LVOT obstruction could have been enrolled. The authors argue this is unlikely given the mean post-Valsalva gradient below 12 mmHg, high NT-proBNP levels and mildly dilated left atria
- KCCQ-CSS is patient-reported and subject to residual measurement variability despite double-blinding; the placebo group improved by 8.4 points
- 31 Japanese patients enrolled after global closure are not reported, so the published population is not the full randomized population
- Cross-trial comparison with ODYSSEY-HCM is uncontrolled — different drug, different titration target, different baseline symptom burden — so the apparent superiority of aficamten over mavacamten in this phenotype is an inference, not a tested hypothesis
- The achieved doses were high by design. 55.8% reached 20 mg at end of treatment, and the LVEF <50% rate of 10.5% is roughly threefold that in the obstructive aficamten trials, which targeted the minimum effective dose
- Editorial support for an earlier draft was provided by a medical communications company funded by the sponsor
Key Concepts Mentioned
Key Entities Mentioned
- entities/Aficamten — the drug under study; extends its evidence base from obstructive to nonobstructive HCM
- entities/HCM — the parent disease and the nonobstructive phenotype
- entities/Mavacamten — the comparator by implication, via ODYSSEY-HCM
- Cytokinetics — sponsor. No wiki page exists and none is warranted on the strength of a single funding disclosure
Wiki Pages Updated
wiki/sources/aficamten-acaciahcm-nejm-2026.md — created (this page)
wiki/entities/Aficamten.md — ACACIA-HCM results, nonobstructive indication and safety profile added; two contradictions added
wiki/entities/Mavacamten.md — ACACIA-HCM contrast with ODYSSEY-HCM added
wiki/entities/HCM.md — nonobstructive management bullet and the "myosin inhibition works only where contractility is the problem" organising principle revised; contradiction added
wiki/concepts/HCM-Management.md — nonobstructive mechanism row, myosin inhibitor section and contradictions revised
wiki/concepts/LVOTO.md — "efficacy depends on there being obstruction to relieve" revised
wiki/concepts/HCM-Mechanism.md — the "nonobstructive symptoms are not contractility-driven" and "failure of myosin inhibition" passages revised
wiki/sourceindex.md — index row
wiki/wikiindex.md — Aficamten and Mavacamten rows revised
log.md — ingest entry