Exercise and Sport in Hypertrophic Cardiomyopathy: What the Safety Evidence Actually Shows

A Comprehensive Evidence Review for Physicians

  • Prepared: 2026-11-16
  • Peg: Stein R, Alô ROB, Ritt LEF, et al. Systematic Review on the Safety of Physical Exercise in Hypertrophic Cardiomyopathy: Recommendations from the Brazilian Society of Cardiology. Arq Bras Cardiol. 2026;123(8):e20260423 [1]

Abstract

For roughly thirty years, a diagnosis of hypertrophic cardiomyopathy (HCM) was treated as a disqualifying condition. The 1985 Bethesda Task Force III report and the 26th Bethesda Conference a decade later framed eligibility as a categorical decision, and the documents that followed — the 2005 European consensus and the 2015 American guideline — kept the restriction in place [2] [3] [1]. The 2023 European Society of Cardiology (ESC) guideline and the 2024 American Heart Association/American College of Cardiology multisociety guideline reversed the default. The American document now states that “for most patients with HCM, universal restriction from vigorous physical activity or competitive sports is not indicated” (Class 3, B-NR) and that mild-to-moderate recreational exercise is beneficial (Class 1, B-R) [4] [5].

The Brazilian Society of Cardiology (SBC) systematic review is the newest attempt to summarise the safety literature behind that reversal. It included eight studies — four randomised trials and four observational studies, 9,744 participants — and found no increased risk of mortality or severe arrhythmic events with exercise, including at vigorous intensity; certainty of evidence ranged from moderate to very low [1]. This review examines that conclusion from the inside. Three findings organise the analysis.

First, the pooled sample is an artefact of arithmetic. A single Korean national insurance cohort of 7,666 adults contributes 78.7% of the 9,744 participants; with the 1,660-participant LIVE-HCM registry, two observational studies supply 95.7% of the people in a “systematic review” that pooled nothing. The four randomised trials that carry the efficacy claim enrolled 290 patients in total and followed them for 12 weeks to 5 months [1] [6].

Second, the safety claim is a claim about absent events, and absent events in small samples are compatible with unacceptable risk. Across the four trials, roughly 80 patient-years of follow-up accrued without a sustained ventricular arrhythmia or death [1]. With zero events, the 95% upper confidence bound on the event rate is approximately 3/n; at 80 patient-years that bound is on the order of 3.8% per patient-year — an order of magnitude above the event rates that justify an implantable defibrillator in this disease.

Third, the benefit claim is real but narrow and surrogate-based. Structured training improves peak oxygen uptake by roughly 1.3 to 2.3 mL/kg/min in low-risk, mostly non-obstructive patients; trial sequential analysis of the peak VO₂ effect is reported as conclusive, but the safety outcomes were never powered and the phenotype enrolled was mild [7] [6] [8].

What has changed is not that exercise has been proven safe. What has changed is that the alternative — universal restriction — was never tested, carried measurable harms, and rested on a denominator nobody ever measured. The defensible position for 2026 is conditional: exercise is recommended for nearly everyone, vigorous exercise and competitive sport are reasonable for carefully characterised low-risk patients after shared decision-making, and the quality of that decision is bounded by the quality of the risk estimate behind it — which, on external validation, remains moderate at best and disagrees substantially between guidelines [9] [10].


1. Introduction — a prohibition that lasted thirty years (and the one sentence that changed)

The modern history of HCM and exercise begins with a pathological observation and an inferential leap. Young athletes who died suddenly were found at autopsy to have hypertrophic hearts; HCM became the archetypal cause of athlete sudden death; and competitive sport was declared incompatible with the diagnosis. The 1985 Bethesda Task Force III report addressed “hypertrophic cardiomyopathy, other myopericardial diseases and mitral valve prolapse” in a document whose structure was explicitly about eligibility for competition [2]. The 26th Bethesda Conference in 1994 repeated the exercise with a task force dedicated to the same three conditions [3]. The SBC review characterises this lineage — the Bethesda Conferences, the 2005 ESC consensus, and the 2015 AHA/ACC document — as having “adopted a restrictive stance regarding recreational or competitive sports of moderate to vigorous intensity” [1].

Note what kind of statement that was. It was not a trial result, a pooled estimate, or even a prospective cohort. It was a consensus judgement made in the absence of a denominator: no one knew how many people with HCM were competing, and therefore no one could compute the risk of competing. The prohibition was, in the strict sense, unfalsifiable — and it survived for three decades because unfalsifiable prohibitions generate no data that could overturn them.

The reversal, when it came, came in two steps. The 2023 ESC cardiomyopathy guideline graded regular low-to-moderate intensity exercise as a Class I recommendation for all able individuals with HCM, allowed high-intensity exercise and competitive sport to be “considered” in genotype-positive/phenotype-negative individuals (Class IIa), permitted it as a Class IIb option in asymptomatic low-risk patients with mild morphology, absent resting or inducible obstruction and no exercise-induced complex ventricular arrhythmia, and stated that it is “not recommended in high-risk individuals and in individuals with left ventricular outflow tract obstruction and exercise-induced complex ventricular arrhythmias” (Class III) [11] [5].

The 2024 American guideline went further, and it did so with an unusual construction. Rather than granting permission, it recommended against the prohibition: “for most patients with HCM, universal restriction from vigorous physical activity or competitive sports is not indicated” (Class 3, B-NR). It graded mild-to-moderate recreational exercise as beneficial (Class 1, B-R), stated that competitive sport of any intensity is reasonable in genotype-positive/phenotype-negative individuals (Class 2a, B-NR), and added a statement found in no other guideline — that ICD placement for the sole purpose of enabling competitive sport should not be performed (Class 3, C-EO) [4] [5].

A note on what “Class 3” means here, because it is routinely misread. In the American system a Class 3 recommendation denotes “no benefit” or “harm,” i.e. a recommendation against doing the thing. The thing being recommended against in the 2024 guideline is blanket restriction, not exercise. This is the single most consequential sentence in the field and it is a negative recommendation about a policy, not a positive recommendation about a behaviour. The corresponding positive recommendations — about who may exercise vigorously, how they should be evaluated, and what should be documented — are weaker (2a, 2b) and rest on observational data.

Two things follow, and they structure this review. First, the paradigm shift is a guideline shift more than an evidence shift. The evidence base did grow — a prospective registry of 1,660 people and four randomised trials of training did not exist in 2015 — but it did not grow by an order of magnitude, and none of the new studies was designed to detect the outcome that justified the old prohibition. What changed was the interpretation of the observational signals, the arrival of trials on functional capacity, and an explicit decision by writing committees to weight patient autonomy and the harms of restriction [12] [13].

Second, the literature is not one literature. It is at least four, with different risks, different exposures, and different outcome ascertainment:

Population Typical exposure studied Typical outcome studied What is actually known
General adults with HCM Self-reported activity tertiles; supervised moderate training All-cause mortality; peak VO₂ Large observational dose-response; small RCTs on surrogates
Competitive athletes with HCM Continued high-intensity training and competition SCD, appropriate ICD therapy, sustained VT Small retrospective/prospective cohorts; almost no events
Genotype-positive, phenotype-negative relatives (G+/P−) Accelerometer MVPA; competitive sport Incident cardiomyopathy, ventricular arrhythmia Large biobank cohort; small athlete series
Children and adolescents School sport; supervised individualised training Symptoms, fitness, BMI, QoL Feasibility studies of 8–20 patients; registry risk-factor analyses

Conflating these is the characteristic error of the field, and it happens in both directions: adult cohort data are used to reassure the parents of a 14-year-old, and data from young athletes with severe hypertrophy are used to restrict a 62-year-old with mild disease and no risk markers. This review keeps them separate and says explicitly whenever evidence from one is being used for another.


2. What hypertrophic cardiomyopathy does under exertion — only the physiology that changes advice

A review for clinicians does not need the full molecular biology of the sarcomere. It needs the four or five mechanisms that determine whether a given patient will become symptomatic, and whether the advice you give is likely to be wrong in a dangerous direction.

Filling pressures rise early, and they rise before symptoms do. In the Danish randomised trial of 12 weeks of supervised moderate training in non-obstructive HCM, the primary outcome was invasively measured pulmonary capillary wedge pressure at just 25 W of exercise. Baseline values were abnormal, and training lowered the 25-W wedge pressure by 2.8 mmHg while the control group rose by 1.2 mmHg (between-group difference 4.0 mmHg, 95% CI 0.7–7.3; P=0.018) [14]. Two things follow. First, the haemodynamic abnormality in HCM is provokable at trivial workloads — this is not a disease that only misbehaves at peak exertion. Second, the abnormality is modifiable by training, which is the strongest mechanistic argument in the entire “exercise as therapy” literature.

The outflow gradient is dynamic, and the highest gradient is often after exercise, not during it. In a prospective study of alcohol septal ablation with invasive exercise haemodynamics, the resting gradient fell from 49 mmHg to 17 mmHg, the peak-exercise gradient from 45 to 32 mmHg, and the post-exercise gradient from 96 to 28 mmHg [15]. The post-exercise value being roughly double the peak-exercise value is the reason a resting echocardiogram plus a Valsalva manoeuvre misses a substantial fraction of obstruction, and the reason exercise echocardiography is recommended in symptomatic patients with a resting or provocable gradient below 50 mmHg [5].

Fibrosis is the arrhythmia substrate, and its extent is graded, not binary. In a 774-patient cohort with cardiac magnetic resonance and 7.4 years of follow-up, patients who reached the SCD-related endpoint had late gadolinium enhancement (LGE) of 13.7 ± 9.4% of left ventricular mass versus 4.9 ± 6.6% in those who did not. Among patients without extensive LGE, LGE of 5% or more conferred roughly seven-fold adjusted SCD risk, and LGE between 5% and 15% identified worse prognosis within both ICD class II and class III categories [16]. This matters for exercise advice because the SBC document uses a 15% threshold to define “extensive” LGE and therefore high risk [1]; the validation data suggest the informative threshold for incremental risk is lower, at 5%.

Apical aneurysm is a distinct phenotype with a distinct risk. A meta-analysis of ten observational cohorts found a pooled prevalence of left ventricular apical aneurysm of 3% (95% CI 2–5%), with 57% of aneurysms smaller than 2 cm. Sudden-death endpoints occurred in 4.71% (95% CI 1.5–9%) of small aneurysms versus 22% (95% CI 15–31%) of aneurysms 2 cm or larger (odds ratio 4.65, 95% CI 2.14–10.10), and thrombus was far more common in the larger lesions (30% vs 2%; OR 10.92, 95% CI 3.75–31.84) [17]. An apical aneurysm therefore changes both the arrhythmic calculus and the exercise calculus — the latter because intense exercise in a patient with an apical cavity and stagnant flow is a plausible thrombotic provocation, an inference rather than a demonstrated risk.

Exercise limitation and arrhythmic risk are partly dissociated, and this is the single most useful physiological fact in the review. Abnormal blood-pressure response to exercise is taught as a sudden-death marker. In a 589-patient cohort that underwent exercise stress testing, 33% had an abnormal response; it was associated with lower achieved percent-predicted VO₂ and with subsequent heart failure hospitalisation (P=0.002), but it was not associated with sudden cardiac death or severe ventricular arrhythmia (P=0.38), and 30% of patients who had an abnormal response plus obstruction still had it after successful myectomy [18]. The ESC and American guidelines have accordingly demoted it from routine risk assessment, while the Japanese guideline retains it [5].

The clinical consequence is worth stating plainly. “Can this patient exercise?” and “is this patient going to die?” are related but not identical questions. The first is governed by filling pressures, gradient, chronotropic competence and pulmonary pressures. The second is governed by fibrosis, genotype, wall thickness, syncope and ventricular arrhythmia on monitoring. A patient can be severely limited and at low arrhythmic risk, or asymptomatic with a large fibrotic burden. Advice that treats the two questions as one will be wrong in both directions.


3. The evidence the restriction was built on: athlete sudden-death registries, and how the denominator was never known

The prohibition rested on registries of sudden death in athletes. Those registries are real, carefully assembled, and — for the purposes of this question — structurally unable to answer it.

The best contemporary American data are the 20-year National Collegiate Athletic Association (NCAA) study. Across 1,102 deaths in NCAA athletes between July 2002 and June 2022, an expert panel adjudicated 143 sudden cardiac deaths, giving an overall incidence of 1 per 63,682 athlete-years (95% CI 1:54,065–1:75,010). Incidence was higher in men than women (1:43,348 vs 1:164,504) and in Black than White athletes (1:26,704 vs 1:74,581), and highest in Division I male basketball players at 1:8,188 athlete-years. The incidence of SCD declined over the study period, with a five-year incidence rate ratio of 0.71 (95% CI 0.61–0.82), while non-cardiovascular deaths were flat [19].

Buried inside that paper is the fact that matters most for HCM. Among the 118 cases with enough information to adjudicate cause, autopsy-negative sudden unexplained death was the most common finding (19.5%), followed by idiopathic left ventricular hypertrophy or possible cardiomyopathy (16.9%), and then HCM (12.7%). Half of all SCD events were exertional [19]. HCM is therefore a real but minority contributor — roughly one in eight adjudicable cases — and a larger fraction of deaths are labelled “possible cardiomyopathy” precisely because the diagnostic boundary in a young athlete’s heart is genuinely uncertain.

The 2024 Lancet review is candid about the wider problem: the precise incidence of sudden cardiac arrest and death among athletes is unknown, because there is no mandatory case reporting and no infrastructure to capture all cases in the general population; causes are age-dependent, with genetic heart disease and unexplained cases predominating in the young and coronary artery disease in masters athletes [20].

The long-distance running data make the same point from a different direction. Among 29,311,597 finishers in US marathons and half-marathons from 2010 to 2023, there were 176 cardiac arrests — an incidence of 0.54 per 100,000 participants (95% CI 0.41–0.70), unchanged from 2000–2009 — but cardiac death fell from 0.39 to 0.20 per 100,000 and case fatality from 71% to 34%. Among the 67 of 128 arrests with a determinable cause, coronary artery disease rather than HCM was the most common aetiology [21]. In other words, in the largest mass-participation endurance population in the world, HCM is not the dominant lethal substrate — and the improvement in survival came from resuscitation, not from exclusion.

Now the inferential problem, which is the heart of this section.

A registry of athlete deaths answers: among athletes who died, what proportion had HCM? The policy question is: among athletes with HCM, what is the annual risk of death? Those are different conditional probabilities, and no registry of the first kind can produce the second. To compute the second you need the denominator — athlete-years accumulated by people carrying the diagnosis — and that denominator was never assembled, because until recently athletes with HCM were removed from sport and therefore from observation.

There is a further selection problem. The cases that built the policy were young, male, and often Black basketball players, the subgroup with the highest incidence in the NCAA data (1:8,188) [19]. The risk estimate derived from that subgroup was then applied to a clinic population with a mean age near 60 and a majority of women [22]. Whatever the merits of restriction in a 19-year-old Division I point guard with 32 mm of septum, it was never a proportionate response for a 58-year-old with 16 mm of septum and no risk markers.

Finally, the counterfactual is unobservable by construction. Because restriction was universal, there is no cohort of high-risk athletes who played, and therefore no way to estimate what restriction prevented. The only way to learn was to stop restricting and measure — which is what the last decade did. That is the honest description of how this field changed: not by refuting the old evidence, but by finally generating the missing denominator.


4. The seed review: what the Brazilian Society of Cardiology actually did, and what it actually found

The SBC document is a “rapid systematic review” registered on PROSPERO (CRD420251005992), restricted to English-language publications, with grey literature sought only through the reference lists of included studies, artificial intelligence used to assist screening and data extraction, and searches of PubMed/MEDLINE, Embase and the Cochrane Library to 18 December 2024 [1]. Two independent searches yielded 219 and 236 publications; after de-duplication, eight original studies were included: four randomised trials and four observational studies [1].

Critically, the review did not pool. The methods state that because of clinical and methodological heterogeneity in design, intervention, follow-up and outcome definition, “results were synthesized narratively and presented in descriptive tables,” with heterogeneity explored qualitatively [1]. This is a defensible choice. It also means the document cannot generate a pooled effect estimate, and any reader who takes “8 studies, 9,744 participants” as a meta-analytic sample has misread it.

4.1 The eight included studies

Study (first author, year) Country Design n Mean/median age (y) Exercise exposure Principal outcome reported
Saberi, 2017 [7] USA RCT 136 50 16 weeks moderate-intensity, home + centre Peak VO₂ +1.27 mL/kg/min vs usual activity; no sustained VT, arrest, shock or death
MacNamara, 2023 [23] USA RCT, no non-exercising control 15 completed 48 5 months moderate vs progressive vigorous Peak VO₂ +1.3 mL/kg/min in both; vigorous increased LV end-diastolic volume; no serious arrhythmia
Gudmundsdottir, 2025 [14] Denmark RCT 59 58 12 weeks supervised moderate, 3 h/week Exercise PCWP −2.8 mmHg vs +1.2 mmHg; peak VO₂ +1.8 vs −0.3 mL/kg/min
Basu, 2025 [24] UK RCT 80 46 12 weeks supervised vigorous (70→85% HR reserve) Peak VO₂ +4.1 mL/kg/min; no difference in arrhythmia composite (P=0.99)
Kwon, 2020 [22] South Korea Retrospective national cohort 7,666 59.5 Self-reported activity tertiles (MET-based) All-cause mortality 9.1% → 8.9% → 6.4% across tertiles
Lampert, 2023 (LIVE-HCM) [25] USA (multinational) Prospective cohort 1,660 39 ≥60 h/year at ≥6 METs vs moderate vs sedentary Composite 4.7% vs 4.6%; aHR 1.01, upper 1-sided 95% CL 1.48
Pelliccia, 2020 [26] Italy Retrospective cohort of athletes 88 31 (median) 27 continued training/competition (6–14 h/week); 61 detrained 2 SCA/deaths over 7 ± 5 years (0.3%/year), both outside sport; no difference by training status
Martinez, 2023 [1] USA Retrospective cohort of athletes 40 NR Return to elite/professional sport Low event rate over ~7 years; one non-exertional arrest, one exertional non-arrhythmic syncope

(Table assembled from the seed’s Tables 1 and 2, the primary reports where retrieved, and the Pelliccia cohort publication. The Martinez cohort is reported only in the seed; its separate primary publication was not retrieved for this review, so no detail beyond the seed’s description is asserted.)

Two structural features of this table deserve emphasis. The first is the arithmetic: 7,666 + 1,660 = 9,326 of 9,744 participants, or 95.7%, come from two cohorts. The four randomised trials contribute 290 participants — 3.0% of the total — and the two athlete cohorts contribute 128. The second is that the exposure definitions are not commensurable: two studies use MET-based activity tertiles from questionnaires, one uses a threshold of ≥60 hours per year at ≥6 METs, one uses hours per week of training, and the trials use prescribed heart-rate zones from 60% of reserve to 90–95% of peak. The review’s own methods acknowledge this and decline to pool. That is honest, and it is also fatal to any claim that the eight studies jointly estimate a dose-response.

4.2 What the review found, and how certain it was

The findings are consistent across designs: no study reported an increased risk of mortality or severe arrhythmic events with exercise, syncope was rare, non-sustained ventricular tachycardia occurred in both arms of the trials without significant difference, and a single sustained ventricular tachycardia occurred in the control arm of one trial [1]. In the two athlete cohorts, remaining in competition for up to seven years was not associated with sudden death or clinical deterioration, and the events observed were judged unrelated to exertion [1].

The GRADE architecture is where the document becomes interesting, because the three recommendations carry very different certainty:

Recommendation Content Strength Certainty
1 Regular exercise in HCM, individualised, shared decision-making; risk re-assessed annually (or six-monthly with high-risk markers) Strong Moderate
2 Regular exercise even with high-risk markers, preferably supervised or where an AED is available Weak Very low
3 Does not contraindicate high-intensity exercise or competitive sport; may be considered in carefully selected cases after multidisciplinary evaluation Weak Low

Risk-of-bias judgements were similarly mixed: among the four RCTs, two were high risk, one moderate and one low on RoB 2; among the four observational studies, two were high, one moderate and one low on ROBINS-I [1].

The review’s own discussion concedes more than its abstract does. It states that the four RCTs followed “approximately 80 patient-years” without relevant adverse events; that young adults aged 20–40 — the subgroup at highest SCD risk — were evaluated only in small observational studies; that there are no Brazilian data; and that applicability is limited by access to specialists and cardiac rehabilitation [1].

4.3 Where the certainty grading strains

Moderate certainty for Recommendation 1 is the claim most likely to be over-read, and it is worth reconstructing how it was reached. The review states that the Korean cohort was assessed as low risk of bias and that this finding “resulted in upgrading the certainty of evidence from ‘Low’ to ‘Moderate’ regarding the benefit of exercise on mortality” [1]. Two observations follow, offered here as critical appraisal rather than as findings of the review.

First, a single observational cohort cannot upgrade certainty for an absence-of-harm claim covering a population that includes high-risk patients. GRADE permits upgrading observational evidence for large magnitude, dose-response and plausible residual confounding — and the Korean study does show a dose-response gradient. But the population in that cohort is a national insurance sample with a mean age of 59.5 years, of whom 29.9% were women, drawn from people who attended health check-ups and who reported some physical activity at baseline: the study explicitly excluded subjects reporting no physical activity [22]. The comparison is between tertiles of the active, not between exercise and no exercise. That is a narrower inference than “exercise reduces mortality in HCM,” and it cannot upgrade certainty about competitive athletes or high-risk patients, who are not in the cohort at all.

Second, the review grades Recommendation 3 as weak with low certainty, which is appropriate, but the underlying evidence is thinner than “low” implies: two retrospective cohorts of 88 and 40 athletes, one of which the review itself notes did not include exclusively HCM patients [1]. Forty athletes followed retrospectively is not a base from which to counsel a 22-year-old about a decade of competition; it is a base from which to design the study that would be.

None of this makes the SBC document wrong. Its central conclusion — that exercise should be encouraged after individualised assessment, and that in patients without high-risk criteria the benefits appear to outweigh the risks — is the same conclusion the European and American guidelines reached, and it is probably correct. The point is that the document’s certainty language is doing work that its data cannot support, and a physician who reads “moderate certainty” as “we know” will counsel differently from one who reads it as “one large cohort and four small trials, one of which was at low risk of bias.”


5. What 9,744 participants really means: the shape of the evidence base, and why pooling breadth is not pooling depth

The headline number in the SBC review is 9,744 participants across eight studies. It is a large number for a rare disease, and it is the number that will be quoted. It is also, as a description of the evidence, close to meaningless, because it is a sum of four different kinds of study measuring four different things in four different populations.

Here is the decomposition.

Component Studies Participants Share of 9,744 What it can actually support
National insurance cohort, Korea [22] 1 7,666 78.7% Association between self-reported activity tertile and mortality, ages ~60
LIVE-HCM prospective registry [25] 1 1,660 17.0% Noninferiority of vigorous exercise for a composite arrhythmic endpoint
Athlete cohorts [26] [1] 2 128 1.3% Descriptive event counts in selected low-risk athletes
Randomised trials [7] [23] [14] [24] 4 290 3.0% Effect on peak VO₂ over 12 weeks to 5 months

Two studies contribute 95.7% of the participants. Four contribute 3.0%. And the four that contribute 3.0% are the only ones that can speak to causation, because they are the only ones in which the exposure was assigned.

This is not a criticism of the SBC authors, who were transparent about the design of what they included. It is a criticism of the sentence that gets repeated afterwards. “A systematic review of 9,744 patients found no increased risk” implies a meta-analytic sample in which each patient contributed comparable information about a comparable exposure. What actually exists is one very large administrative cohort, one moderately large prospective registry, two small athlete series, and four small trials — a body of evidence with breadth in one direction and depth in another, and almost nothing in the middle.

5.1 The arithmetic of zero events

The trials enrolled 290 patients and followed them for 12 weeks to 5 months. The SBC review puts the total follow-up at “approximately 80 patient-years” [1]. Across those 80 patient-years there were no deaths, no resuscitated cardiac arrests and no appropriate ICD shocks in either arm of any trial; a single sustained ventricular tachycardia, without complication, occurred in the control arm of one trial [1] [24].

The correct inference from zero deaths and zero resuscitated arrests in 80 patient-years is bounded, and the bound is computed with the rule of three: when zero events are observed in n independent opportunities, the 95% upper confidence limit on the event probability is approximately 3/n.

Applied here, treating patient-years as the unit of observation:

  • 3 / 80 patient-years = 0.0375 per patient-year, i.e. a 95% upper bound of about 3.8% per year.
  • Expressed over five years, that upper bound corresponds to roughly 19% cumulative incidence.
  • To drive the upper bound below 1% per year — approximately the annualised rate implied by the conventional 6% five-year threshold used in risk-score-based ICD decision-making [5] — you would need about 300 patient-years with zero events (3 / 0.01 = 300).

This is arithmetic on reported numbers, not a finding, and it cuts in a specific direction. It does not say exercise caused anything. It says the trial evidence is compatible with an event rate several-fold higher than the rate at which we offer patients an implantable defibrillator. A clinician who reads “no adverse events in the trials” as reassurance about arrhythmic risk is making an inference the data cannot support, and the SBC document — to its credit — does not make that inference explicitly; it makes it implicitly by assigning moderate certainty to a recommendation whose safety component rests on these trials.

There is a second, subtler point about zero-event trials. The trials were not designed to detect arrhythmias; they were designed to detect changes in peak VO₂. Adverse events were collected as safety reporting, usually by site investigators, without systematic adjudication by a blinded committee in most of these studies, and without protocolised rhythm monitoring beyond what was clinically indicated. That is a normal and acceptable design for a phase 2 functional study. It is not a design that can establish the safety of an exposure that is hypothesised to trigger fatal arrhythmias at peak effort, because the exposure of interest — the last 10% of maximal exertion, the sprint finish, the competitive surge — is precisely the part of the dose that a supervised 12-week programme at 60–85% of reserve does not deliver.

5.2 What “moderate” and “vigorous” mean in each study

The single largest threat to the coherence of this literature is that the exposure is defined differently in every study, and the words are not interchangeable.

Study Definition of the higher-intensity exposure
Saberi [7] 60% of heart-rate reserve, progressing to 70%
Gudmundsdottir [14] 60% of maximal work capacity, Borg 12–14
MacNamara (vigorous arm) [23] 4 × 4 min at 90–95% of peak heart rate
Basu [24] 70% of heart-rate reserve, progressing to 85%
Kwon [22] Self-reported; MET-weighted score, tertiles
LIVE-HCM [25] ≥60 hours/year at ≥6 METs
Pelliccia [26] 6–14 hours/week of training; competitive level

A patient at 70% of heart-rate reserve, a patient running four-minute intervals at 92% of peak heart rate, and a patient accumulating 60 hours per year above 6 METs are not doing the same thing, and there is no reason to assume their risk is the same. The competitive athlete’s exposure includes the specific physiology of competition — catecholamine surge, dehydration, heat, the final sprint — that no laboratory-based training protocol reproduces. When the SBC review says exercise at “moderate or vigorous intensity” was not associated with harm, the phrase aggregates a set of exposures that spans roughly 60% to 95% of maximal capacity plus competitive racing [1].

The LIVE-HCM investigators handled this more carefully than anyone, by defining vigorous exercise a priori as ≥6 METs for ≥60 hours per year and separately identifying the 259 participants who competed [25]. That is the standard the rest of the field should meet.


6. The exercise-training trials: what they measured, what they did not, and what the trial sequential analysis says

Four randomised trials underpin the claim that training helps. They are small, short, and consistent in direction. They are also consistent in what they did not measure.

6.1 The four trials

Trial n Duration Intervention Primary outcome Result Safety
Saberi 2017 [7] 136 randomised, 113 completed 16 weeks Moderate, home + centre, 3–7 sessions/week Change in peak VO₂ +1.35 vs +0.08 mL/kg/min; between-group +1.27 (95% CI 0.17–2.37), P=0.02 No sustained VT, arrest, appropriate shock or death in either arm
MacNamara 2023 [23] [27] 15 completed (8 moderate, 7 vigorous) 5 months Moderate vs progressive vigorous (4×4 min at 90–95% peak HR) Peak VO₂, cardiac output, AV O₂ difference +1.3 mL/kg/min overall; vigorous +1.5 vs moderate +1.1, no difference (P=0.628); vigorous increased LV end-diastolic volume by 17 mL No serious arrhythmia or adverse cardiac event; explicitly underpowered for safety
Gudmundsdottir 2025 [14] 59 randomised, 51 completed 12 weeks Supervised moderate, 3 h/week Invasive PCWP at 25 W −2.8 vs +1.2 mmHg; difference 4.0 mmHg (95% CI 0.7–7.3), P=0.018 No signal reported; primary outcome haemodynamic
Basu 2025 [24] 80 randomised, 67 completed 12 weeks Supervised vigorous, 70→85% HR reserve, 3×/week + home Feasibility; exploratory arrhythmia composite Peak VO₂ +4.1 mL/kg/min (95% CI 1.1–7.1); anaerobic threshold +2.3; systolic BP −7.3 mmHg; BMI −0.8 kg/m² No difference in arrhythmia composite (P=0.99); gains largely dissipated at 6 months

Read together, these trials establish three things and fail to establish a fourth.

They establish that training improves functional capacity in mildly symptomatic, mostly non-obstructive HCM. The effect sizes range from +1.3 to +4.1 mL/kg/min. The largest, +4.1 mL/kg/min in the British vigorous-intensity trial, is roughly twice the 1.7 mL/kg/min between-group gain produced by aficamten in obstructive disease over 24 weeks, which is a striking result for 12 weeks of supervised exercise [24] [28]. The smallest, and the most methodologically conservative, is the Danish trial’s +1.8 mL/kg/min in the exercise arm against −0.3 in controls [14].

They establish a plausible central mechanism. The Danish trial’s primary endpoint was not fitness but invasively measured filling pressure during light exercise, and training lowered it by 4 mmHg relative to control while also improving peak cardiac index (+0.38 vs −0.85 L/min/m²) and ventilatory efficiency (VE/VCO₂ difference 2.0, 95% CI 0.6–3.5) [14]. The Dallas trial added the complementary finding that vigorous training increased left ventricular end-diastolic volume by 17 mL, suggesting genuine cardiac rather than purely peripheral adaptation [23]. This is the strongest mechanistic case in the field, and it is worth being precise about what it means: training appears to improve diastolic reserve and stroke-volume recruitment, which is exactly the abnormality that limits these patients.

They establish that the benefit is not durable without continued training. In the British trial, most of the gain had dissipated by six months after the programme ended [24]. Saberi’s group reported the same pattern [1]. Any prescription that consists of a 12-week course and a discharge letter is therefore a prescription for a transient benefit.

They do not establish safety for hard outcomes. All four were powered on function. The largest had 136 patients; the smallest had 15. None had a prespecified, adjudicated arrhythmia endpoint with adequate power; the British trial’s arrhythmia composite was explicitly exploratory [24]. At 290 patients across four trials, the pooled safety database is smaller than a single phase 2 oncology study, applied to an exposure whose principal theoretical hazard is a rare fatal event.

6.2 The meta-analyses and what the trial sequential analysis actually says

Four meta-analyses have now pooled this literature, and their convergence on efficacy is more impressive than their convergence on anything else.

The meta-analysis with trial sequential analysis (TSA) included four RCTs and 290 patients, predominantly low-risk and non-obstructive, with a mean LVEF of 66% and mild functional limitation. Exercise improved peak VO₂ by a mean difference of 1.81 mL/kg/min (95% CI 1.01–2.61, P<0.01), reduced maximal wall thickness by 0.56 mm (95% CI −0.97 to −0.16) and BMI by 0.65 kg/m² (95% CI −1.05 to −0.25), with no significant change in LVEF or outflow gradient and no increase in arrhythmias, syncope or SCD. The authors state that the peak VO₂ effect was “confirmed by TSA,” and conclude that larger multicentre trials including higher-risk phenotypes are needed [6].

A second meta-analysis and meta-regression of eight studies (three RCTs, 205 low-to-moderate-risk patients) found a larger peak VO₂ gain of 2.3 mL/kg/min (95% CI 0.76–3.83), predicted peak VO₂ +8.1% (2.7–13.4), exercise time +1.0 min, peak heart rate +5.2 bpm, heart-rate reserve +5.5 bpm, BMI −1.12 kg/m² and resting heart rate −3.1 bpm; a lower baseline heart-rate reserve predicted a larger training response, and there were no adverse events reported [8].

A third, including seven studies (three RCTs, four observational) and 9,702 patients, reported improved peak VO₂, reduced BMI (−0.67 kg/m²), reduced wall thickness (−0.56 mm) and improved quality-of-life physical score (SMD 0.66), with the mortality benefit judged to be driven mainly by observational data [29]. A fourth, of eight studies and 2,217 patients, reported a major adverse cardiovascular event risk ratio of 1.01 (P=0.97) and a peak VO₂ gain of 1.76 mL/kg/min [30].

The TSA result deserves a careful reading, because it is being cited as if it settled the question. Trial sequential analysis asks whether the accumulated information exceeds the “required information size” needed to detect or exclude an effect of a given magnitude, adjusting for repeated testing. Applied to peak VO₂ across 290 patients, with a consistent effect and modest variance, it is entirely plausible that the required information size is reached — a continuous surrogate with a mean difference of 1.8 and a standard deviation around 3 requires on the order of a few hundred patients. That is what “confirmed by TSA” means here: the functional-capacity question is answered.

It does not mean the safety question is answered, and no TSA was performed on hard outcomes in these reports. Even if one had been, the required information size for a rare event is not a few hundred patients; it is tens of thousands of patient-years. The TSA finding and the safety finding are about different quantities, and the sentence “exercise training in HCM is safe, confirmed by trial sequential analysis” is a category error that appears in secondary summaries of this literature.

6.3 What the trials actually enrolled, and who they did not

The pooled population is the limitation. Mean age 46–58, mean LVEF 66%, predominantly non-obstructive, NYHA class I–II, with mean maximal wall thickness in the mid-teens to low twenties. The TSA meta-analysis states this explicitly: “all studies primarily involved mild, non-obstructive phenotypes, limiting generalisability” [6].

The patients whom clinicians worry about most are precisely those absent from the trials: the 19-year-old with 30 mm of septum; the patient with a resting gradient of 80 mmHg; the patient with an ICD who has had appropriate therapy; the patient with an apical aneurysm; the child. For all of these, the trials provide indirect evidence at best, and the guidelines’ recommendations are explicitly consensus-based [1].

One detail from the Danish trial is worth flagging for practice. The overall Kansas City Cardiomyopathy Questionnaire score did not differ between groups, but the physical limitation subscale improved by 8.4 points [14]. Twelve weeks of training made patients fitter and improved their physical limitation without changing their global quality of life. That is a realistic expectation to set, and it is a useful counter to the enthusiasm that follows a +4 mL/kg/min result.


7. The prospective vigorous-exercise cohort (LIVE-HCM): design, event counts, confidence intervals, and the criticism

LIVE-HCM is the most important single study in this field, and it is important precisely because it was designed to answer the right question and because its results are more equivocal than the headlines suggested.

7.1 Design

The Lifestyle and Exercise (LIVE) consortium enrolled 1,660 individuals aged 8 to 60 years with either a clinical diagnosis of HCM or genotype-positive status without left ventricular hypertrophy, excluding those with conditions precluding exercise, across experienced centres. Exposure was amount and intensity of physical activity: 252 participants (15%) were sedentary, 709 (43%) exercised moderately, and 699 (42%) exercised vigorously, of whom 259 (37% of the vigorous group) competed. The primary prespecified composite endpoint was death, resuscitated sudden cardiac arrest, arrhythmic syncope, or appropriate ICD shock, adjudicated by an events committee blinded to exercise category [25].

That design — prospective, prespecified, blinded adjudication, with a noninferiority boundary declared in advance — is a genuinely higher standard than anything else in this literature. It is the study that a sceptical reviewer would have asked for.

7.2 Results

Measure Non-vigorous (n=961) Vigorous (n=699)
Composite endpoint 44 (4.6%) 33 (4.7%)
Rate per 1,000 person-years 15.3 15.9
Adjusted hazard ratio 1.00 (reference) 1.01
Upper 95% one-sided confidence limit — 1.48 (prespecified noninferiority boundary 1.50)

Two numbers in that table are routinely misreported, and both matter.

The first is the event rate. 15.9 events per 1,000 person-years is 1.59% per year. At that rate, the observed 4.7% implies roughly three years of average follow-up — a derived figure, since the abstract does not report follow-up duration. That is not a small number for a population with a mean age of 39: it is the annual rate of death, resuscitated arrest, arrhythmic syncope or appropriate shock. Describing this as a “low event rate” population, as is often done, is inaccurate. What the study shows is that the rate is the same in vigorous and non-vigorous exercisers, not that the rate is low.

The second is the size of the noninferiority margin. The trial was designed to declare vigorous exercise noninferior if the upper one-sided 95% confidence limit of the hazard ratio fell below 1.50. The observed upper limit was 1.48 — just inside the boundary. The study therefore excluded, with 95% confidence, a 50% relative increase in events; it did not exclude a 40% increase, a 30% increase, or a 20% increase. For a shared decision about competitive sport in a 20-year-old, a 30% relative increase in a serious arrhythmic endpoint is not a negligible quantity. The result is best stated as: LIVE-HCM provides moderate assurance that vigorous exercise does not roughly double risk, and no assurance at all that it is risk-neutral.

7.3 The criticism

Four criticisms of LIVE-HCM are legitimate and should be taught alongside the result.

Volunteer and healthy-adherer bias. Participants were people who volunteered for a study about exercise, recruited through centres and advocacy channels, and who were exercising at the time of enrolment. People who had previously had an event during exercise, or who felt unwell when exercising, were less likely to enrol. The comparison group of “non-vigorous” exercisers is itself active. This design cannot capture the counterfactual of the patient who would have had an event had they exercised.

Exclusion of the highest-risk patients. Individuals with conditions precluding exercise were excluded, which in practice means the cohort is depleted of exactly the patients about whom the guidelines are cautious. The result generalises to the population that already looked like it could exercise.

Limited power for the subgroups that matter. With 77 events total, the study cannot meaningfully stratify by age, by LGE burden, by wall thickness above 30 mm, or by genotype, and the 259 competitive athletes are far too few to analyse separately with any precision. Every subgroup statement from this cohort is hypothesis-generating.

Ascertainment depends on the ICD. Part of the composite is appropriate ICD shock, which is only observable in patients who have a device. Patients at higher perceived risk are more likely to have devices, which biases ascertainment upward in the high-risk group and can mask a true exposure effect — or, depending on how risk and exercise correlate, exaggerate one.

The fairest summary is that LIVE-HCM substantially reduced the prior probability of a large exercise-related hazard, moved the field from “prohibited” to “permitted after evaluation,” and left the question of a modest hazard entirely open. Anyone who tells a patient that the safety question is settled is overstating a noninferiority trial with a 50% margin.


8. The observational dose-response literature: more activity, lower mortality — and the four confounders that could produce it

The Korean national cohort is the single largest piece of evidence in this field, and it is the study that most directly supports the “exercise is good for you” half of the argument.

8.1 What the cohort found

From the National Health Insurance Service database, 7,666 adults with HCM who underwent health check-ups between 2009 and 2016 and who reported some physical activity at baseline were followed for a mean of 5.3 ± 2.0 years. Mean age was 59.5 years and 29.9% were women. Across tertiles of a MET-weighted activity score, all-cause mortality fell from 9.1% to 8.9% to 6.4% (P for trend 0.0144) and cardiovascular mortality from 4.7% to 3.8% to 2.7% (P for trend 0.0023). Comparing the highest with the middle tertile, the hazard ratio was 0.78 (95% CI 0.63–0.95) for all-cause mortality and 0.75 (95% CI 0.54–1.03) for cardiovascular mortality [22].

Read the confidence interval on the cardiovascular mortality estimate carefully. It crosses 1. The statistically robust finding is a reduction in all-cause mortality of about 22%; the cardiovascular-specific reduction is directionally consistent but not significant. Since the hypothesised mechanism is cardiovascular, that is a meaningful weakness rather than a technicality.

8.2 The four confounders

Confounder 1: the sedentary were excluded. The study excluded subjects who reported no physical activity at baseline [22]. The comparison is therefore between the least active tertile of people who exercise and the most active tertile — not between exercisers and non-exercisers. This is a defensible design choice (it reduces reverse causation from people too ill to exercise), but it has a consequence that is rarely stated: the study cannot tell you whether exercising is better than not exercising, only whether exercising more is better than exercising less, among people who exercise. The SBC review uses this study to upgrade certainty about “the benefit of exercise on mortality” [1]; the study’s own design supports a narrower claim.

Confounder 2: exercise capacity is a marker of disease severity, not only a modifier of it. This is the deepest problem and it is not fixable by adjustment. In a cohort of 253 low-risk patients with stable exercise habits, those doing more than 30 MET-hours/week had lower BMI, lower maximal wall thickness (15.9 vs 18.9 mm), better diastolic function, less LGE, and fewer symptoms than sedentary patients, with no difference in arrhythmic burden or adverse events [31]. The most economical explanation is not that vigorous exercise thinned the septum; it is that people with milder disease are the people who can and do exercise vigorously. The SBC review itself notes external data showing that patients achieving >100% of predicted METs have very low event rates while those below 85% have significantly higher risk [1]. Functional capacity is a prognostic marker. Any observational study that treats it as an exposure will find that the able do better, and will misattribute the cause.

Confounder 3: self-report and the health-check-up population. Activity was assessed by questionnaire, once, at baseline, with a MET-weighted score. Self-reported activity correlates modestly with measured activity and is biased by recall and by social desirability. More importantly, the cohort consists of people who attended a national health screening programme — a selected population with better health engagement than the average patient with HCM. And the HCM diagnosis itself came from administrative coding, not adjudication.

Confounder 4: the population is not the population at risk of the outcome we fear. Mean age 59.5 years. The deaths in this cohort are predominantly cardiovascular deaths in older adults — heart failure, coronary disease, stroke — not arrhythmic deaths in the young. Even if the association is causal, it is causal for a different outcome than the one that justified the prohibition. Extrapolating a mortality benefit in 60-year-olds to a competitive 22-year-old is an inversion of the direction in which this literature is usually misapplied.

8.3 What the exposure actually is, in 2026

There is an irony in studying the health effects of exercise in a population that has been told not to exercise. Accelerometry in 203 patients with HCM and 37 genotype-positive phenotype-negative individuals showed that patients took 5,254 steps per day versus 6,573 in the genotype-positive group, spent 3.4% versus 4.5% of the day in moderate-to-vigorous activity, and were sedentary for more than 80% of the day in 61% versus 35% of cases. Symptomatic and obstructive patients were the least active; notably, asymptomatic obstructive patients were as inactive as symptomatic ones. Sixty-two per cent recalled receiving exercise guidance, and 59% reported reducing their activity as a result [32].

That last figure is the one to remember. The exposure in these cohorts is partly iatrogenic. When a study finds that the least active patients do worse, some of the causal pathway may run through the advice those patients were given.

8.4 The synthesis the observational data can support

Combining the Korean cohort, LIVE-HCM, the athlete cohorts and the accelerometry data, the defensible observational claim is this: in adults with HCM who are capable of exercise, higher self-reported activity is associated with lower all-cause mortality, and vigorous activity is not associated with a large increase in arrhythmic events. The first half of that sentence is a prognostic association confounded by disease severity; the second half is a genuine safety finding from a prospective registry with a generous noninferiority margin. Neither half justifies the statement that exercise prevents sudden death, and no study in this literature has shown that it does.


9. The contrary signal: studies linking high-intensity activity to arrhythmia, and why they are not dismissed

A review that only presented the reassuring studies would be advocacy. There is a real signal in the opposite direction, it comes from a well-conducted cohort, and it deserves to be stated with its numbers.

9.1 Twenty per cent of sudden deaths occurred during high-intensity activity

A Korean tertiary-centre study identified 75 patients with documented sudden cardiac death events from a cohort of 2,619 patients with HCM evaluated between 2005 and 2023. Activity at the time of the event was classified as high-intensity (≥6 METs) or low-to-moderate. Fifteen of the 75 deaths (20%) occurred during high-intensity activity. Those patients were much younger — median age 25 years (IQR 16–43) versus 56 years (IQR 48–64), P<0.001 — and had higher ESC sudden-death risk scores (median 4.9% vs 2.4%, P=0.023) but fewer ventricular arrhythmias during exercise testing. There were no differences between groups in the degree of hypertrophy, outflow obstruction, systolic or diastolic function, or genetic findings. In multivariable analysis, younger age was the only independent risk factor. Among patients who survived an initial high-intensity-related event, recurrent events were triggered by subsequent high-intensity activity [33].

Four things follow, and they are not comfortable.

First, the arithmetic of event timing. If 20% of sudden deaths occur during an activity that occupies a small fraction of a patient’s waking hours, the instantaneous hazard during that activity is correspondingly elevated. The study does not report exposure time, so the hazard ratio cannot be computed — this is a limitation, not a finding, and I state it as an inference. But the direction is unambiguous: exercise is not a neutral backdrop against which deaths happen to occur; it is, for a minority of patients, the moment at which they occur.

Second, the phenotype of the exercise-related deaths is the phenotype our risk tools handle worst. These patients were young, had fewer exercise-induced arrhythmias, and did not differ in hypertrophy, obstruction or genotype. Their calculated ESC risk was higher than that of patients who died at rest or at moderate intensity (median 4.9% vs 2.4%), so the model moved in the right direction — but a median five-year risk of 4.9% sits below the threshold at which most clinicians would implant a defibrillator. The calculator ranked them higher without ranking them high enough to act on. That is the strongest single argument against using a risk score as a clearance certificate for a young athlete.

Third, the recurrence finding matters for counselling. That recurrent events in survivors were again triggered by high-intensity activity is a small number of patients but a coherent pattern, and it argues that a patient who has had an exertional event and returns to the same intensity is re-entering the same exposure.

Fourth, the finding cuts against a simple “dose” model. If risk simply scaled with cumulative exercise, the high-intensity deaths would look like the high-risk phenotype. They do not. They look like young people with a substrate that ordinary exercise testing does not expose.

9.2 The genotype signal in athletes is weaker than expected

In a retrospective series of 59 sports-active patients with HCM (median 6 hours/week, 67.8% competing), genotype-positive and genotype-negative patients had similar ventricular arrhythmia rates (20.7% vs 23.3%, P=0.807) and similar event-free survival (P=0.685); 84.6% of arrhythmias were isolated non-sustained ventricular tachycardia, and hard events were few. The ESC 5-year risk score was modestly higher in genotype-positive patients (1.87% vs 1.54%, P=0.048) but most participants were low risk, and 88.1% remained active at last follow-up [34].

This is a small study and should not be over-read. Its value is that it undercuts two intuitions at once: that genotype-positive athletes are necessarily the high-risk group, and that continuing sport inexorably produces events. Neither is supported at this sample size.

9.3 Device data provide the most direct measure of exertional arrhythmia

Implantable defibrillators give us something registries of death cannot: a continuous, device-adjudicated record of ventricular arrhythmia in people who are exercising.

In a subcutaneous ICD sports registry of 280 patients (median age 46, 72% male), 205 (73%) engaged in sport. Over 38 ± 13 months, 27 patients (13%) received appropriate shocks; 20 of those 27 (74.1%) occurred at rest and 7 (25.9%) during peak performance. All patients who received appropriate shocks had been implanted for secondary prevention. Inappropriate shocks occurred in 27 patients (13.1%), of which 12 (44.4%) occurred during peak physical activity, mostly from T-wave oversensing. There was no difference in the type or intensity of sport between patients who did and did not receive shocks [35].

The interpretation is double-edged. On one side, three-quarters of appropriate shocks happened at rest — exercise was not the dominant trigger even in a sporting population. On the other, a quarter happened at peak effort, in a population where peak effort occupies a small fraction of time. And the inappropriate-shock burden during sport is real: 12 patients received a shock during exertion because the device misread a T wave. For a competitive athlete, an inappropriate shock during competition is not a minor adverse event; it is a fall, a loss, and a reason to stop.

9.4 Why the contrary signal is not dismissed

The temptation in 2026 is to treat the Korean sudden-death study as an outlier and the registry data as reassuring. That would be a mistake, for three reasons.

It is the only study in this literature that examines the circumstances of death rather than the activity level of survivors. Cohort studies of activity and mortality are, structurally, studies of people who did not die. A study that starts from deaths and works backwards asks a different and complementary question, and it should not be discarded because it is inconvenient.

Its findings are consistent with the physiology. Catecholamine surge, dynamic obstruction, diastolic filling failure, ischaemia in a hypertrophied bed, and fibrosis as the arrhythmic substrate are all maximally engaged at peak effort, not at 60% of reserve. A model in which exercise is uniformly protective would have to explain why the moment of maximal physiological stress is not also the moment of maximal arrhythmic risk.

It is consistent with the device data, in which 26% of appropriate shocks and 44% of inappropriate shocks occurred during peak exertion [35].

The honest synthesis is that exercise probably has two effects that run in opposite directions: a chronic benefit mediated by fitness, filling pressures, body composition and mental health, and an acute hazard concentrated in the minutes of maximal exertion, in a small subgroup we cannot reliably identify. The guidelines’ shift is a judgement that the first outweighs the second for most patients. That is a reasonable judgement. It is not the same as evidence that the second does not exist, and the SBC’s own weak, low-certainty recommendation for competitive sport reflects exactly that tension [1].


10. Risk stratification is the gate: HCM Risk-SCD, the AHA/ACC approach, LGE, apical aneurysm, exercise blood pressure

Every exercise recommendation in both guidelines is conditional on risk. “Low-risk patients may” is the operative sentence, and it is worth exactly as much as the risk estimate behind it. This section examines that estimate.

10.1 The two frameworks are not the same measurement

Feature ESC 2023 approach [11] [5] AHA/ACC 2024 approach [4] [5]
Core method Validated risk-prediction model (HCM Risk-SCD) producing a 5-year risk estimate, combined with clinical judgement Qualitative, multifactorial assessment of established risk factors and “risk modifiers,” with shared decision-making
Role of the calculator Central; a high calculated risk is itself an indication to consider an ICD Supportive; not the primary determinant
Major risk factors named Unexplained syncope; maximal wall thickness ≥30 mm; high calculated risk; family history of SCD; non-sustained VT Similar factors, weighted clinically rather than summed
Effect on ICD volume More restrictive More permissive: 58% of an external cohort classified as class IIa vs 29% under ESC [9]
Concordance — Fleiss’ kappa 0.34 (95% CI 0.29–0.40) across the three frameworks [9]

The concordance figure is the headline. In 200 Brazilian patients with HCM assessed under the 2014 ESC, 2023 ESC and 2024 AHA/ACC systems, agreement on primary-prevention ICD recommendation was fair at best (kappa 0.34). The American algorithm placed 58% of patients in class IIa while only 29% reached that recommendation under the 2023 ESC model; endpoints occurred in 8% over 9.4 ± 6.5 years. The 2014 ESC guidelines had the highest accuracy (77%, 95% CI 71–83) and negative predictive value (96%, 95% CI 90–98), but all guidelines had low positive predictive value [9].

That last clause is the crux and it deserves to be said in plain language. A high negative predictive value means the frameworks are good at telling you who will not have an event. A low positive predictive value means most patients flagged as high risk will not have an event either. A gate that opens for many and admits few events is a gate that produces a lot of ICDs per death prevented — which may still be the right trade, but it is not “risk stratification” in the sense that a patient hearing “you are high risk” imagines.

10.2 Calibration in a large Japanese registry

The REVEAL-HCM registry followed 3,611 consecutive patients and reported five-year cumulative incidence of SCD-equivalent events by guideline class: 23.8%, 7.2%, 5.7% and 2.3% for 2024 AHA/ACC classes 1, 2a, 2b and 3; and 23.8%, 2.9%, 9.3% and 2.6% for 2023 ESC classes 1, 2a, 2b and 3 [10].

Two things stand out. The American framework did not separate classes 2a and 2b meaningfully (P=0.101). The European framework actually reversed them — class 2b patients had a higher five-year event rate (9.3%) than class 2a patients (2.9%), P=0.545 — which is a failure of ordinality, not merely of precision. And the highest-risk phenotype by far was end-stage HCM with ejection fraction below 50%, at 18.5% five-year incidence, followed by mid-ventricular obstruction at 6.9% and non-obstructive HCM at 4.7% [10].

The lesson is not that one guideline is better. It is that both frameworks are ordering patients with modest accuracy, and that a phenotype the calculator handles poorly — end-stage disease with systolic dysfunction — dominates the risk in any real clinic.

10.3 What actually improves the estimate

Late gadolinium enhancement. In 774 patients with CMR and 7.4 years of follow-up, 46 (5.9%) reached an SCD-related endpoint; those patients had LGE of 13.7 ± 9.4% versus 4.9 ± 6.6% (P<0.001). C-statistics for SCD prediction were 0.68 for the 2011 ACC/AHA model, 0.64 for 2014 ESC, 0.76 for 2020 AHA/ACC and 0.78 for the 2022 ESC model — so the calculators have improved but remain in the 0.7–0.8 range. Adding LGE helped: among patients without extensive LGE, LGE ≥5% carried roughly seven-fold adjusted SCD risk, and LGE of 5–15% identified worse prognosis within both ICD class II and class III [16].

The practical implication is that the SBC’s 15% threshold for “extensive” LGE [1] is probably set too high for the purpose of discriminating risk, and that a patient with 8% LGE and an otherwise borderline profile is not the same patient as one with 2%.

Apical aneurysm. Pooled prevalence 3% (95% CI 2–5%), with 57% of aneurysms under 2 cm. Sudden-death endpoints occurred in 4.71% of small aneurysms versus 22% of those ≥2 cm (OR 4.65, 95% CI 2.14–10.10), and left ventricular thrombus in 30% versus 2% (OR 10.92, 95% CI 3.75–31.84) [17]. An apical aneurysm larger than 2 cm is therefore a phenotype in which both the arrhythmic and the thrombotic calculus change, and exercise advice should reflect that rather than being driven by wall thickness.

Left ventricular systolic dysfunction. End-stage HCM (EF <50%) carried an 18.5% five-year SCD-event incidence in REVEAL-HCM [10] — by some distance the highest-risk group in that registry. This is a group in whom exercise is usually limited by heart failure rather than by arrhythmia, and in whom cardiac rehabilitation logic (supervised, low-intensity, monitored) is more appropriate than athlete logic.

Exercise blood pressure response. As discussed in Section 2, abnormal blood-pressure response was associated with lower functional capacity and with heart-failure hospitalisation but not with sudden death or ventricular arrhythmia in a 589-patient cohort (P=0.38), and persisted after myectomy in 30% of those who had it with obstruction [18]. The ESC and American guidelines have demoted it from routine risk assessment; the Japanese guideline retains it [5]. For exercise advice specifically, an abnormal blood-pressure response remains useful — not as a sudden-death marker, but as evidence that the patient’s haemodynamic reserve is poor and that the exercise prescription should be built around supervised testing rather than around a target heart rate.

Cardiopulmonary exercise testing as a risk tool. The RoMa classification, derived from peak heart rate and oxygen pulse, was externally validated in 292 patients with a mean follow-up of about six years: 68 composite events occurred, and RoMa classes II–IV carried hazard ratios of 3.89 to 5.37 versus class I, independent of genotype, ejection fraction, sex and obstruction [36]. This is a promising adjunct because it measures the physiological reserve that the anatomical scores ignore, and because it comes from the same test used to write the exercise prescription.

Children are a separate problem. In 446 children and teenagers with HCM from 35 centres, conventional adult risk factors — family history of SCD, abnormal blood pressure response, non-sustained VT on ambulatory monitoring — were not associated with lethal arrhythmic events; the factors that were, in multivariable analysis, included age at ICD placement, posterior wall thickness z-score, and an outflow gradient below 30 mmHg [37]. Applying the adult calculator to a child is therefore not merely imprecise; it may be using the wrong variables.

10.4 The consequence for exercise advice

If the risk estimate is fair-to-moderate, then the exercise recommendation built on it is fair-to-moderate. That is not a reason to abandon the recommendation; it is a reason to be honest about what “low risk” means and to reassess it. The SBC recommends annual reassessment, or six-monthly in patients with high-risk markers, and immediate reassessment with new symptoms [1]. Given that the calculators have C-statistics around 0.75 and that risk factors change — fibrosis accumulates, gradients rise with age, ventricular function declines — an exercise clearance issued once at diagnosis and never revisited is indefensible.


11. Competitive athletes: what the 2024 AHA/ACC and 2023 ESC guidelines now permit, where they disagree, and what shared decision-making has to document

11.1 The three guideline positions, side by side

Era / document Recreational exercise Competitive sport and athletes
1985 Task Force III; 26th Bethesda Conference (1994) Not framed as therapy; the question was eligibility Categorical eligibility decision for HCM [2] [3]
2005 ESC consensus; 2015 AHA/ACC Restrictive Restrictive, as characterised by the SBC review [1]
2020 AHA/ACC [38] Recommended, with individualised assessment Individualised assessment; vigorous competitive sport framed as requiring expert evaluation and shared decision-making. Superseded in 2024 [4]
2023 ESC [11] [5] Class I (C): recommended in all able individuals IIa (C) genotype-positive/phenotype-negative; IIb (B) asymptomatic low-risk with mild morphology, no obstruction, no complex exercise-induced VA; III (C) high-risk, or LVOTO with exercise-induced complex VA
2024 AHA/ACC multisociety [4] [5] Class 1 (B-R): beneficial, in keeping with general population guidelines 2a (B-NR) competitive sport of any intensity in genotype-positive/phenotype-negative; Class 3 (B-NR) “universal restriction from vigorous physical activity or competitive sports is not indicated”; Class 3 (C-EO) ICD should not be implanted solely to permit competitive sport

Classes and levels of evidence for the ESC and 2024 AHA/ACC rows are taken verbatim from a published systematic comparison of the three guidelines [5]. The 2020 row summarises the position the 2024 document explicitly updated [4]; the SBC review characterises the pre-2024 documents (Bethesda, the 2005 ESC consensus and the 2015 AHA/ACC guideline) as restrictive [1], and the 2020 guideline is best read as the transitional document between the categorical prohibition and the 2024 reversal. The 1985 and 1994 rows are cited for the existence and framing of those eligibility documents, not for specific wording.

11.2 Where they genuinely disagree

The two contemporary guidelines agree on more than the commentary suggests. Both grade mild-to-moderate exercise as beneficial, both explicitly decline to exclude genotype-positive phenotype-negative individuals from high-intensity sport, and both use the same broad definitions of intensity based on percentage of maximal heart rate [5].

The disagreement is narrow and consequential, and it concerns the high-risk patient and the patient with obstruction:

  • The ESC prohibits; the AHA individualises. The European document states affirmatively that high-intensity exercise, including competitive sport, “is not recommended in high-risk individuals and in individuals with left ventricular outflow tract obstruction and exercise-induced complex ventricular arrhythmias” (Class III). The American document declines to prohibit for most patients and instead requires comprehensive evaluation and shared decision-making with an expert, without a categorical exclusion for obstruction [5]. For a patient with a provocable gradient of 60 mmHg and non-sustained ventricular tachycardia on exercise testing who wants to play competitive basketball, the two documents point in different directions.
  • The AHA addresses the ICD-for-sport question; the ESC does not. The American Class 3 statement that ICD placement solely to permit competitive sport should not be performed is the only explicit guidance on a question that arises weekly in sports cardiology clinics [5].
  • The risk framework underneath differs, as Section 10 describes, and that difference propagates into who is “low risk” in the first place. The same patient can be class IIa under one framework and class 3 under another [9].

11.3 What the athlete conversation has to cover

Shared decision-making is named in every one of these documents and specified in none of them. The following is a synthesis of what the evidence supports and is labelled as such — it is not a guideline recommendation.

What should be documented. The diagnosis and phenotype (maximal wall thickness, resting and provocable gradient, ejection fraction, LGE burden, apical aneurysm if present); the calculated risk where a calculator applies, and its known limitations; holter findings; exercise test findings including blood-pressure response and achieved workload; the specific sport, its intensity and its competitive level; the presence of an AED and an emergency action plan at training and competition venues; and the content of the discussion, including the residual uncertainty.

What has to be said out loud. That the evidence for safety in competitive athletes comes from cohorts of a few dozen to a few hundred people, that the largest prospective study excluded a 50% increase in risk rather than a 20% increase [25], that a minority of sudden deaths occur during high-intensity activity and those patients are often young with unremarkable risk profiles [33], and that the patient’s own tolerance for uncertainty is a legitimate input.

What the patient is entitled to hear about the alternative. That restriction is not risk-free. Adults with HCM are less active than the general population and have higher BMI [39]; objectively measured activity is markedly lower than in genotype-positive relatives, with 59% of patients reporting that they reduced activity after receiving guidance [32]; and restrictions negatively affected emotional well-being in most surveyed patients [39]. For young athletes, permanent disqualification is one of the identified risk factors for greater psychological morbidity after diagnosis [40].

11.4 The institutional dimension

An eligibility decision is not only a clinical decision. It determines whether a school, a club, a national federation or an insurer accepts the patient, and clinicians are frequently asked to sign a document that transfers risk. Two observations from the retrieved literature are relevant. First, athletes who are genotype-positive and phenotype-negative frequently seek clearance after being disqualified elsewhere — 19% of one series of 274 such athletes did so — and in that series no cardiac event or death occurred over more than 1,300 patient-years, leading the authors to describe restriction based solely on a positive genetic test as “genetic discrimination” [41]. Second, critical reviews on the conservative side of this debate argue that the observational cohorts are too small and too selected to support liberalisation in young athletes [42]. Both positions are defensible on the current data, which is precisely why the decision belongs to a properly informed patient rather than to a form.


12. Return to play, ICDs in athletes, and the contact-sport problem

12.1 Return to play after diagnosis

Return to play is a different question from initial participation, because it follows an event — a diagnosis made at screening, a syncope, an ablation, a myectomy, or a cardiac arrest. The literature here is thin and mostly retrospective.

In the Italian athlete cohort of 88 patients with HCM followed for 7 ± 5 years, 27 continued training and competition while 61 substantially reduced or stopped. Two patients had sudden cardiac arrest or death (0.3% per year), both outside sport, and 22% reported symptoms; freedom from the combined endpoint did not differ between those who continued and those who detrained (log-rank P=0.264) [26]. The authors are appropriately careful: similar results may not be seen in younger, more racially diverse, or more severely affected populations. This is the best available direct evidence that continuation is not associated with a large excess of events, and it is a retrospective cohort of 88 people, 92% male and 98% white.

For the patient who returns after septal reduction, the haemodynamic data are encouraging. Seven months after alcohol septal ablation in 24 patients, wedge pressure at 75 W fell from 33 ± 9 to 24 ± 10 mmHg, resting gradient from 49 to 17 mmHg, post-exercise gradient from 96 to 28 mmHg, and KCCQ clinical summary score rose from 65.4 to 76.7, with NYHA class III falling from 54% to 13% [15]. A patient whose obstruction has been mechanically relieved has a different exercise physiology than before, and re-evaluation rather than permanent restriction is the logical posture — an inference from haemodynamic data, not a tested return-to-play strategy.

After surgical myectomy, the picture is more mixed. In 295 patients with cardiopulmonary testing before and after myectomy, 128 (43%) were peak-VO₂ non-responders, and non-response was associated with higher pre-operative percent-predicted VO₂, older age, female sex, dyslipidaemia, lower BMI — and lack of enrolment in cardiac rehabilitation. Non-responders had higher long-term mortality (adjusted HR 1.77, 95% CI 1.06–3.34) [43]. The rehabilitation association is observational and confounded by indication, but it is the closest thing to a trial signal in the post-myectomy population, and it supports referring these patients to structured rehabilitation rather than simply clearing them.

12.2 The athlete with an ICD

Two facts define this problem. The device works: in the subcutaneous ICD sports registry, 13% of 280 patients received appropriate therapy over about three years, and a quarter of those therapies occurred during peak exertion [35]. And the device misfires: 13.1% received inappropriate shocks, 12 of them during peak activity, mostly from T-wave oversensing [35].

For counselling, that means three things. An ICD is not a guarantee — it treats ventricular arrhythmia, it does not prevent it, and an arrhythmic syncope before the shock arrives is still a fall at speed. Inappropriate shocks are common enough in sport that programming and lead choice deserve as much attention as the eligibility decision. And because all appropriate shocks in that registry occurred in secondary-prevention patients, the primary-prevention athlete who has never had an event is a different risk category from the survivor.

The American guideline’s Class 3 statement — that an ICD should not be implanted solely to permit competitive sport — closes a loophole that was being used in exactly the wrong direction: escalating therapy to enable an exposure rather than reducing the exposure to match the therapy [5].

12.3 Contact sport

There are no retrieved data quantifying device or lead failure as a function of contact-sport participation in HCM. The S-ICD registry found no difference in shock occurrence by type or intensity of sport [35], which addresses arrhythmic events but not mechanical trauma. The reasonable position — explicitly an inference — is that the decision in contact sport turns on device integrity and the consequences of a shock-induced collapse in a collision environment, that subcutaneous systems avoid transvenous lead fracture but not blunt trauma, and that the patient should be told the mechanical risk is not quantified rather than being told it is absent.


13. Children and adolescents: school sport, individualised training, and the measurable harm of blanket restriction

Paediatric HCM is where the adult evidence is least transferable and where the cost of being wrong is highest — in both directions.

13.1 The risk model does not transfer

In a retrospective multicentre cohort of 446 children and teenagers aged 20 or younger with idiopathic HCM who had an implantable defibrillator, 152 (34%) had experienced a lethal arrhythmic event (defined as a secondary-prevention device or a primary-prevention device that had delivered appropriate therapy). The conventional adult risk factors were not associated with those events: family history of sudden death, abnormal blood-pressure response to exercise, and ventricular tachycardia on ambulatory monitoring were all non-significant. The factors that did emerge were age at ICD placement, posterior wall thickness z-score, and an outflow gradient below 30 mmHg — a variable that in adults is reassuring and in this cohort was not [37].

That single result should change how clinicians talk to parents. When a family asks “what is my child’s risk?”, the honest answer is that the adult calculator has not been validated in children, that the variables it uses behave differently before puberty, and that the decision is being made on expert consensus plus phenotype. The 2026 review of paediatric exercise and sport participation makes the same point in its conclusion: the relationship between activity and events is “more nuanced than once believed,” and the negative consequences of uniform restriction have become increasingly clear [44].

13.2 The feasibility evidence is small but real

The only retrieved trial of structured exercise in paediatric HCM enrolled eight subjects aged 8–18, all NYHA class I or II, with a median septal thickness of 2.3 cm. The programme was 16 weeks of three 30-minute aerobic sessions plus two strength sessions per week, with a target heart-rate zone derived from baseline cardiopulmonary exercise testing and a consumer wearable used to titrate intensity. Self-reported adherence ranged from 1.25% to 96.25% — a range that tells you most of what you need to know about feasibility in adolescents. All four subjects who completed the intervention improved exercise performance and health-related quality of life, and ambulatory rhythm monitoring revealed no significant arrhythmia [45].

Four completers is not evidence of efficacy. It is evidence of possibility: that a supervised, individually prescribed programme can be delivered to children with HCM, that it does not generate arrhythmia on monitoring, and that the main obstacle is adherence rather than safety. That is enough to justify a proper trial and enough to justify not telling a family that exercise is forbidden.

13.3 The harm of restriction is real but not uniform

The strongest evidence for harm comes from congenital heart disease rather than HCM. In 110 children followed for a mean of 8.4 years, activity restriction was associated with overweight (risk ratio 2.60, 95% CI 1.34–3.54) and obesity (risk ratio 4.08, 95% CI 1.42–7.38) after adjustment for baseline weight; among the 92 children at healthy weight at baseline, the obesity risk ratio was 6.14 (95% CI 2.54–8.82). Exercise intolerance without restriction was not significantly associated with either outcome, so the signal was specific to the restriction itself rather than to cardiac limitation [46].

The HCM-specific paediatric data are more equivocal, and this is worth stating because it is usually omitted. In a chart review of 109 children with inherited arrhythmia syndromes or cardiomyopathies, some level of restriction was documented in 80% of phenotype-positive children versus 37% of phenotype-negative children, yet BMI did not differ over time by restriction status, and the proportion overweight or obese at follow-up was consistent with the general Canadian paediatric population. Thirty-eight per cent were nonetheless playing moderate- or high-dynamic sport after diagnosis [47].

The reconciliation is probably that restriction is one of several inputs, and that in a contemporary North American cohort the background obesogenic environment is strong enough to swamp it. That does not make restriction harmless. It means the harm is not automatic, and that a blanket claim that restriction causes obesity in children with HCM overstates the HCM-specific evidence while understating the general one.

What is not equivocal is the fitness and activity deficit. In a cross-sectional study of 100 children and adolescents with inherited cardiac disease and 107 controls, peak VO₂ was 32.2 vs 40.2 mL/kg/min (z-score −1.49 vs −0.16, P<0.001) and moderate-to-vigorous activity 42 vs 48 min/day (P=0.009), after adjusting for beta-blockade and using modern paediatric reference models [48]. In a separate cohort with accelerometry, patients accumulated 35 vs 55 min/day of moderate-to-vigorous activity and had lower quality-of-life scores; 51% had modified their activity because of the diagnosis, and modification was associated with worse health-related and cardiac-specific quality of life [49].

13.4 A defensible paediatric position

Synthesising the above — and labelling this as consensus and inference rather than trial evidence:

  • Do not use the adult risk calculator as if validated; state its limits to the family [37].
  • Do not impose blanket exclusion from school physical education. Most school PE is moderate-intensity and time-limited, and the alternative is a sedentary adolescent with a lower fitness trajectory [48].
  • Use cardiopulmonary exercise testing to build an individual prescription where the child is motivated; the feasibility data support supervised moderate aerobic plus strength training with heart-rate guidance [45].
  • Address weight, diet and activity together when restriction is unavoidable, because the CHD literature shows the obesity signal is concentrated in restricted children [46].
  • Treat competitive sport as a separate, higher-stakes decision requiring the multidisciplinary process described in Section 11, not as an extension of PE class.
  • Screen for the psychological consequences. In young athletes diagnosed with cardiac disease, permanent disqualification was one of the identified risk factors for greater psychological morbidity, and the process unfolded in four stages from identity threat through grief to adaptation [40]. A 16-year-old who is told to stop playing is not making a scheduling change.

14. Masters athletes, genotype-positive phenotype-negative relatives, pregnancy, and the post-arrest patient

14.1 Masters athletes

The 2026 joint European Association of Preventive Cardiology/American College of Cardiology consensus on masters athletes with abnormal cardiovascular findings is an expert-based document, and it is explicit about its own evidentiary basis: guidelines for these conditions are generally derived from symptomatic sedentary patients, which limits their applicability to highly trained individuals. It addresses atrial fibrillation, bradyarrhythmias, ventricular arrhythmias, coronary atherosclerosis, aortic dilatation, myocardial fibrosis, and exercise-induced arrhythmogenic cardiomyopathy, and calls for dedicated randomised trials and international registries [50].

HCM is not among the seven domains the consensus organises itself around — an inference from its published scope rather than a statement in the document — which means the 55-year-old with HCM who wants to keep racing falls between two stools: the HCM guidelines, built on clinic populations with a mean age near 60 but little competitive-athlete data, and the masters consensus, built for athletes but not organised around HCM.

Two things are nevertheless clear for this patient. First, the aetiology of sudden death shifts with age: coronary artery disease rather than HCM dominates among masters athletes [20], so a masters athlete with HCM still needs their coronaries assessed, and attributing every symptom to the cardiomyopathy is a diagnostic error waiting to happen. Second, the physiology of masters sport is different — lower peak heart rate, greater reliance on stroke volume, more atrial fibrillation — and exercise prescriptions written for a 40-year-old do not transfer.

The most defensible position, labelled as inference: treat the masters athlete with HCM like any other patient with HCM for risk stratification, add age-appropriate coronary and aortic evaluation, accept that the competitive-sport evidence in this age group is essentially absent, and use cardiopulmonary exercise testing to set intensity rather than a formula.

14.2 Genotype-positive, phenotype-negative relatives

This is the population in which the evidence has improved most in the last two years, and in which the old advice was most clearly disproportionate.

In 84,699 UK Biobank participants with whole-genome sequencing and accelerometer data — including 3,979 genotype-positive phenotype-negative carriers of cardiomyopathy-associated variants — followed for a median of eight years, greater moderate-to-vigorous activity was associated with a lower risk of adverse cardiovascular outcomes irrespective of genotype. In carriers, the hazard ratios at the optimal activity level versus zero were 0.68 (95% CI 0.58–0.79) for atrial fibrillation, 0.58 (0.47–0.71) for heart failure, 0.49 (0.24–1.00) for myocardial infarction and 0.35 (0.12–0.99) for stroke. Activity of 100–400 minutes per week was generally associated with the lowest risk. Cardiac remodelling on CMR was similar in carriers and non-carriers. Higher activity was associated with a lower risk of incident cardiomyopathy (HR 0.03, 95% CI 0.00–0.98) and with no increase in ventricular arrhythmia (HR at 400 minutes/week versus zero, 0.98; 95% CI 0.83–1.14) [51].

In athletes specifically, a Mayo Clinic series of 274 genotype-positive phenotype-negative athletes across all levels of sport — 84% long QT, 6% arrhythmogenic cardiomyopathy, 3% HCM — accumulated more than 1,300 patient-years of follow-up with no cardiac event or death. Nineteen per cent had sought return-to-play approval after being disqualified elsewhere, and 27% were managed with an intentional non-therapy strategy. The authors conclude that restriction based solely on a positive genetic test “should be viewed as genetic discrimination” [41].

Objective activity data reinforce how far this group sits from the phenotype-positive population: genotype-positive phenotype-negative individuals took 6,573 steps per day versus 5,254 in HCM patients and were sedentary for more than 80% of the day in 35% versus 61% of cases [32].

The clinical implication is straightforward and now well supported: a positive genetic test alone is not a reason to restrict exercise. The caveat is that these cohorts are still selected — biobank participants are healthier than average, and the Mayo series is a referral population with intensive surveillance — so the guidance applies to a carrier with a normal echocardiogram, no fibrosis, no family history of juvenile sudden death, and a variant whose pathogenicity has been properly classified.

14.3 Pregnancy

Pregnancy is the special population with the least exercise-specific evidence and the most consequential counselling.

The largest retrieved dataset is an administrative propensity-matched analysis from the TriNetX network: 10,936 individuals with HCM of reproductive age, matched to 3,399 patients in each group. Pregnancy was associated with comparable all-cause mortality over ten years (adjusted OR 0.89, 95% CI 0.70–1.14, P=0.37), no difference in major adverse cardiovascular events (adjusted OR 1.02, 95% CI 0.85–1.22) and no difference in arrhythmic events (adjusted OR 0.93, 95% CI 0.73–1.18), including in the obstructive subgroup [52].

That is reassuring about pregnancy itself and says nothing about exercise during pregnancy — no retrieved study addresses structured exercise training in pregnant patients with HCM, and the guidelines’ reproductive recommendations are consensus-based [5]. The practical position, labelled as inference: continue moderate activity in an uncomplicated pregnancy, avoid the supine position and valsalva-heavy resistance work as pregnancy advances, stop with any exertional syncope, chest pain or arrhythmia, and treat the postpartum period as a period of haemodynamic flux rather than a return to baseline. Competitive sport during pregnancy is a decision without an evidence base in either direction.

14.4 The post-arrest patient

There is no retrieved evidence on return to sport after cardiac arrest in HCM, and the guidelines do not provide a specific pathway. Three principles follow from the data reviewed above and are labelled as inference: the survivor has demonstrated that they have a substrate capable of producing ventricular fibrillation, so risk stratification has changed qualitatively; the recurrence data from the Korean sudden-death study suggest that survivors who returned to high-intensity activity had recurrences triggered by that activity [33]; and the device registry shows that secondary-prevention patients account for all the appropriate shocks during sport, which is exactly this population [35]. Return to competitive sport after an HCM-related arrest should therefore be treated as an exceptional decision, made by a multidisciplinary team with an explicit emergency action plan, and not as the default outcome of a successful resuscitation.


15. Pharmacotherapy and exercise capacity: beta-blockers, myosin inhibitors, and the prescription question nobody has answered

The interaction between drug therapy and exercise capacity is where the “exercise is safe” conversation most often goes wrong in practice — not because a drug makes exercise unsafe, but because a drug can make exercise feel impossible, and the patient then concludes that their disease has progressed.

15.1 Beta-blockers reduce obstruction and symptoms but do not improve exercise capacity

The best evidence is a double-blind, placebo-controlled crossover trial in 29 patients with obstructive HCM and NYHA class II or worse, comparing metoprolol with placebo over two consecutive two-week periods (TEMPO, [53]). Metoprolol lowered the resting gradient (25 vs 72 mmHg, P=0.007), the peak-exercise gradient (28 vs 62 mmHg, P<0.001) and the post-exercise gradient (45 vs 115 mmHg, P<0.0001); it reduced the proportion in NYHA class III or worse from 38% to 14% and improved the KCCQ overall summary score modestly (76.2 vs 73.8, P=0.039). Measures of exercise capacity, peak oxygen consumption and NT-proBNP did not differ between arms [54].

That is a precise and clinically useful result. Beta-blockade in obstructive HCM is a symptom and gradient intervention; it is not a fitness intervention. The corollary is that a patient whose breathlessness persists on a beta-blocker may still have a gradient worth treating, and that a patient whose exercise capacity has fallen on a beta-blocker may be experiencing a drug effect rather than disease progression.

The MAPLE-HCM exercise analysis sharpened this considerably. In a prespecified analysis of 175 randomised patients comparing aficamten with metoprolol (MAPLE-HCM, [55]), metoprolol was associated with a large (≥3 mL/kg/min) reduction in peak VO₂ in 20.7% of patients, versus 2.4% with aficamten (odds ratio 10.6), while a large improvement occurred in 20.5% with aficamten versus 3.7% with metoprolol (odds ratio 6.8). Aficamten also improved submaximal exercise efficiency (VE/VCO₂ slope −2.8), anaerobic threshold (+76 mL/min), peak workload (+8 W) and post-exercise oxygen-uptake recovery (−11 seconds) [56].

15.2 Myosin inhibitors improve exercise capacity in obstructive disease, and not in non-obstructive disease

Trial Population n Duration Peak VO₂ result
EXPLORER-HCM, mavacamten [57] Obstructive, symptomatic 251 30 weeks +1.4 mL/kg/min (95% CI 0.6–2.1) vs placebo; primary composite met in 37% vs 17%
SEQUOIA-HCM, aficamten [28] [58] Obstructive, symptomatic 282 24 weeks +1.8 vs 0.0 mL/kg/min; between-group +1.7 (95% CI 1.0–2.4)
ODYSSEY-HCM, mavacamten [59] [60] Non-obstructive, symptomatic 580 48 weeks +0.52 vs +0.05; between-group +0.47 (95% CI −0.03 to 0.98), P=0.07

The contrast between the first two rows and the third is the most important pharmacodynamic fact in this review. Removing outflow obstruction — whether by drug, myectomy or ablation — improves exercise capacity, because obstruction is a mechanical load that limits stroke volume. In non-obstructive disease there is no such load to remove, and mavacamten did not produce a statistically significant improvement in peak VO₂ or symptoms, with more reductions in ejection fraction and more treatment interruptions than placebo [59].

The mechanistic detail adds a second layer. In EXPLORER, mavacamten improved a range of CPET parameters beyond peak VO₂ — peak VE/VCO₂ ratio (−2.2), peak METs (+0.4), circulatory power (+372.9 mL/kg/min × mmHg), peak end-tidal CO₂ (+2.0 mmHg), exercise time (+0.7 min) and the VE/VCO₂ slope (−2.6) [61]. In SEQUOIA, aficamten shortened the time for VO₂ to fall by 12.5% after exercise by 8 seconds (95% CI −12 to −5), with an odds ratio of 3.7 for a ≥15-second improvement and a number needed to treat of 4.8; this recovery metric predicted heart-failure hospitalisation and death in a referral cohort (HR 1.54 per 15-second increase) [62]. These are not fitness effects alone; they are measures of how quickly the heart recovers from the work it has just done.

15.3 What this means for the exercise prescription

Three conclusions, each of which is an inference from the data above rather than a tested strategy.

No drug makes exercise safe. None of these trials was designed or powered to assess arrhythmic risk during exercise, and none reported a reduction in sudden death. The SEQUOIA and EXPLORER populations were highly selected, followed for 24–30 weeks, and monitored in a way that real-world exercise is not. A patient on a myosin inhibitor still has HCM and still has a substrate.

Beta-blocker dose is a modifiable determinant of exercise capacity, and it is often overlooked. If a patient reports exercise intolerance and is on a beta-blocker, the drug is a candidate explanation before the disease is blamed — particularly in non-obstructive disease, where the drug offers less gradient benefit. Deprescribing or dose reduction is a decision that requires the same shared process as any other, but it should at least be on the list.

The drug and the exercise prescription are competing interventions for the same outcome. In obstructive HCM, a myosin inhibitor produces roughly 1.4–1.8 mL/kg/min of peak VO₂ gain over 24–30 weeks [57] [28]; supervised exercise produces 1.3–4.1 mL/kg/min over 12 weeks to 5 months [7] [24]. No trial has compared them, combined them, or tested whether exercise training after myosin inhibition produces additional benefit. That is a conspicuous gap given that both are now standard options.


16. What a clinician should actually prescribe, and what will not tell you it is going wrong

This section is a synthesis of consensus and inference, not trial evidence. No randomised trial has tested a specific exercise prescription against another in HCM. What follows is built from the protocols used in the four trials, the published prescription frameworks, and the physiology in Section 2.

16.1 Before prescribing: the minimum dataset

Domain Test What you are looking for
Symptoms Structured history Exertional syncope or presyncope; chest pain; palpitations with dizziness
Rhythm 24–48 h ambulatory monitoring Non-sustained VT; burden; atrial fibrillation
Structure Echocardiography at rest, with Valsalva Maximal wall thickness; resting gradient; systolic function
Provocation Exercise echocardiography if resting/provocable gradient <50 mmHg Provocable obstruction; exercise-induced mitral regurgitation
Fibrosis Cardiac MRI with LGE quantification LGE extent, including the 5–15% range [16]; apical aneurysm [17]
Physiology Cardiopulmonary exercise test Peak VO₂, anaerobic threshold, VE/VCO₂ slope, heart-rate reserve, blood-pressure response
Risk Guideline framework, documented Explicit statement of which framework and what it estimated [9]

Two of these deserve emphasis because they are the ones most often skipped. Exercise echocardiography changes management in symptomatic patients with a resting gradient below 50 mmHg, and the post-exercise gradient can be double the peak-exercise value [15]. Cardiopulmonary exercise testing does three things at once: it establishes the baseline for a prescription, it identifies chronotropic incompetence and a low heart-rate reserve (which predicts a larger training response [8]), and it provides the variables for the RoMa classification, which independently predicted outcomes in external validation [36].

16.2 The prescription

Component Recommendation Basis
Frequency 3–5 days/week aerobic; 2 days/week resistance Conventional prescription; the trials used 3 sessions/week up to near-daily [7] [14] [24]
Intensity Moderate: 40–60% heart-rate reserve, or 60–70% of peak heart rate, RPE 11–14. Progress to 70–85% reserve if asymptomatic and low risk Trial protocols spanned 60% of reserve to 90–95% of peak; the moderate band is the best-evidenced starting point [7] [14]
Time 20–30 min initially, progressing to 45–60 min; interval format (4 × 4 min) is an alternative for the fitter patient [23] [7]
Type Aerobic (walking, cycling, swimming, rowing) plus dynamic resistance at moderate load; avoid heavy valsalva straining if a significant gradient is present [63]
Progression Increase duration before intensity; reassess at 8–12 weeks [63]
Supervision Supervised or AED-equipped setting for patients with any high-risk marker; home-based with remote monitoring acceptable for low-risk patients SBC Recommendation 2 [1]; trial designs
Reassessment Annually, or six-monthly with high-risk markers, or immediately with new symptoms [1]

The “exercise should be prescribed and titrated like a drug” framing is not decorative: it comes from a prescription framework developed specifically for HCM, which notes that these patients are on average less active than the general population and that personalised prescription is “an exception rather than the rule” [63]. A personalised moderate-intensity prescription delivered after full evaluation improved cardiopulmonary fitness without adverse events in a prospective cohort, and physically active patients had substantially better peak VO₂ than sedentary ones (32.9 vs 25.2 mL/kg/min) with a similar prevalence of ventricular arrhythmia [64].

16.3 What will not tell you it is going wrong

This is the part of the consultation that is hardest to teach and most worth teaching.

Symptoms are a lagging indicator. Patients adapt by reducing effort. A patient who reports no symptoms may simply have stopped doing anything that provokes them — 59% of patients in one accelerometry cohort reduced their activity after receiving guidance [32]. Asking “are you symptomatic?” without asking “what are you actually doing each week?” will miss this entirely.

A stable echocardiogram does not mean stable risk. Fibrosis accumulates; gradients rise with age and with volume status; end-stage disease with ejection fraction below 50% carries an 18.5% five-year event incidence and can develop in a patient whose last scan was reassuring [10]. The variables that matter for exercise advice are not all in the standard report.

A single exercise test at diagnosis does not describe the patient two years later. The trials show that training gains dissipate within six months of stopping [24], and the guidelines’ own reassessment intervals assume that risk and capacity change [1].

The device is not a safety net for the exercise prescription. Appropriate shocks occur during peak exertion in a quarter of device recipients who receive them, and 44% of inappropriate shocks in one registry occurred during peak activity [35]. An ICD changes the consequence of an arrhythmia; it does not make the arrhythmia benign, and a shock at speed is still a collapse.

What to tell patients will signal a problem: exertional syncope or near-syncope, chest pain that is new or different, palpitations with dizziness, an unexplained fall in exercise tolerance over weeks, or a new inability to complete a previously manageable route. Each of these should trigger reassessment rather than a change of training plan.


17. Research agenda: the trial that has not been done

The field’s central question has never been tested. No randomised trial has assigned patients to liberalised vigorous exercise or competitive sport versus restriction and measured hard outcomes. Everything in Sections 6 through 12 is either a surrogate trial, an observational cohort, or a consensus judgement.

It is worth being concrete about why, because the arithmetic explains the literature’s shape.

To detect a hazard ratio of 1.5 for a serious arrhythmic endpoint with 80% power at a two-sided alpha of 0.05 requires roughly 190 events. If the control-group event rate is about 1% per year and the exposed-group rate 1.5% per year — an average of 1.25% per year — 190 events require approximately 15,000 patient-years, or roughly 3,000 patients followed for five years. HCM has a reported prevalence of roughly 1 in 500 by echocardiographic screening, and up to 1 in 200 by genetic analysis [1], of whom a minority are competitive athletes; no single country can supply that sample, and no sponsor has an incentive to fund it.

This calculation is arithmetic on assumed rates, not a finding; the assumptions are stated so that a reader can substitute their own.

What could actually be done, offered as a design proposal rather than a recommendation:

  1. An international prospective registry with mandatory event reporting and a prespecified exposure definition. LIVE-HCM proved this is feasible with 1,660 participants; a registry an order of magnitude larger, with device-measured activity rather than questionnaires, would give usable precision within a decade [25].
  2. A registry-based randomised trial embedded in that cohort, randomising equipoise cases to liberalised versus conventional advice — the design that solved the sample-size problem in cardiology elsewhere.
  3. A core outcome set and standardised exposure definitions. MET-hours per week plus device-measured time above threshold, with competitive sport reported separately. The current literature cannot be pooled because “vigorous” means four different things [1].
  4. Adequately powered safety reporting in every future trial, with independent adjudication and protocolised rhythm monitoring, rather than adverse-event counts collected as an afterthought.
  5. Paediatric and pregnancy-specific studies, where the evidence is thinnest and the counselling stakes are highest [45] [52].
  6. Trials of exercise alongside myosin inhibition, since both interventions now target the same outcome and no one has tested the combination [28].
  7. Implementation research on deconditioning, because the largest modifiable harm in this field may be the advice itself [32] [39].

18. Limitations of this review

This review is built on a systematic review of eight heterogeneous studies, and it inherits that review’s limitations while adding its own.

The seed’s constraints. The SBC document is a rapid systematic review, English-language only, with grey literature sought only through reference lists and with artificial intelligence assisting screening and extraction [1]. It synthesised narratively and did not pool. Its evidence base is dominated by one administrative cohort. Its “moderate certainty” for the primary recommendation rests substantially on a single study assessed as low risk of bias, in a population that excludes the sedentary and the young.

Source access. The full text of the seed was retrieved from PubMed Central and read in full, so statements about the seed’s methods, included studies, GRADE ratings and recommendation wording are drawn from the document itself rather than from its abstract. Most primary studies were read as structured abstracts from PubMed rather than as full texts; where a detail was not in the abstract, it is not asserted. The full texts of the 2023 ESC and 2024 AHA/ACC guidelines could not be retrieved (the publisher returned HTTP 403), so guideline recommendations, classes and levels of evidence are taken from a published systematic comparison of those guidelines that reproduces them verbatim in tabular form [5], supplemented by the seed’s description and by the guidelines’ own abstracts [11] [4] [38].

What was not done. No independent risk-of-bias assessment, no meta-analysis, no formal GRADE assessment, no assessment of publication bias, and no systematic screening of the grey literature were performed for this review. The searches (39 distinct PubMed queries, logged with hit counts) were purposive rather than exhaustive, and a different reviewer would have selected a different set of primary sources. Several 2026 publications are cited from abstracts alone. One athlete cohort (Martinez) is described only as reported in the seed, because its primary publication was not retrieved; that is stated in the relevant table.

Conceptual limits. “Vigorous,” “high-intensity,” “competitive” and “athlete” are used inconsistently across the literature and, where possible, this review reports each study’s own definition. Where evidence from one population is applied to another — adult cohort data to children, athlete data to clinic patients, obstructive trial data to non-obstructive patients — the transfer is flagged. The risk-score arithmetic in Section 5 is arithmetic on reported numbers and is labelled as such. The FITT prescription in Section 16 and the return-to-play discussion in Section 12.3 are consensus and inference, not trial evidence.

What this review cannot tell you. It cannot tell you the annual risk of competitive sport for a specific 19-year-old with 28 mm of septum, no fibrosis and a pathogenic variant. Nobody can, on current data. What it can tell you is why nobody can, what the best available estimates are, and how wide the uncertainty around them remains.


19. Bottom line

The prohibition has been reversed, and the reversal is justified by the shape of the evidence rather than by its weight. The old restriction rested on registries of athlete deaths that could not compute a denominator, applied to a clinic population ten times older than the cases that generated them. It survived because it was untestable. When it was finally tested — by LIVE-HCM prospectively, by the athlete cohorts retrospectively, by four randomised trials of training — no large hazard appeared, and the harms of restriction became visible. The 2024 American guideline’s Class 3 statement against universal restriction is the correct reading of that history [4] [5].

What the evidence actually supports. Supervised moderate exercise improves peak VO₂ by roughly 1.3–2.3 mL/kg/min in mildly symptomatic, mostly non-obstructive HCM, lowers exercise filling pressures, and improves physical limitation; the effect is real, modest, and decays without continued training [7] [14] [6]. Vigorous exercise is not associated with a large increase in arrhythmic events in a prospective cohort, with the important qualification that the cohort excluded a 50% increase rather than a 20% one [25]. In genotype-positive phenotype-negative individuals, moderate-to-vigorous activity is associated with better cardiovascular outcomes and no increase in ventricular arrhythmia in a biobank cohort of nearly 4,000 carriers [51].

What it does not support. It does not support the claim that exercise prevents sudden death. It does not support the claim that the safety question is settled. The totality of randomised safety data is about 80 patient-years, which bounds the event rate at roughly 3.8% per year — an order of magnitude above the threshold at which we implant defibrillators. It does not support extrapolating adult clinic data to children, in whom the adult risk factors do not even predict events [37]. And it does not support the claim that any drug makes exercise safe.

The three things a clinician should carry away. First, the gate is the risk estimate, and the risk estimate is fair-to-moderate at best: two guidelines agree only fairly on who should get an ICD (kappa 0.34), both have low positive predictive value, and one of them reverses its own intermediate risk classes on external validation [9] [10]. Second, the patient in front of you is probably doing less than they should, partly because of advice they have already received — 59% of patients in one cohort reduced activity after guidance, and the least active include asymptomatic patients with obstruction [32]. Third, restriction has a cost that is measurable in fitness, weight, quality of life and psychological morbidity, and that cost is borne disproportionately by the young [48] [46] [40].

The residual uncertainty, stated plainly. A minority of sudden deaths in HCM occur during high-intensity activity, in patients who are young and whose standard risk profile is unremarkable [33]. That signal has not been explained away, and it is the reason the correct posture is individualised permission after evaluation and shared decision-making — not a new universal permission any more than the old universal prohibition. The field has moved from “no” to “yes, if.” Getting the “if” right is the work of the next decade, and it will require the trial described in Section 17 that no one has yet funded.

References

  1. Stein R, Alô ROB, Ritt LEF, et al. Systematic Review on the Safety of Physical Exercise in Hypertrophic Cardiomyopathy: Recommendations from the Brazilian Society of Cardiology. Arq Bras Cardiol. 2026;123(8):e20260423. doi:10.36660/abc.20260423. PMID 42788594; PMCID PMC13671148. https://pmc.ncbi.nlm.nih.gov/articles/PMC13671148/
  2. Maron BJ, Gaffney FA, Jeresaty RM, McKenna WJ, Miller WW. Cardiovascular abnormalities in the athlete: recommendations regarding eligibility for competition. Task force III: Hypertrophic cardiomyopathy, other myopericardial diseases and mitral valve prolapse. J Am Coll Cardiol. 1985;6(6):1215-7. doi:10.1016/s0735-1097(85)80203-5. PMID: 4067096.
  3. Maron BJ, Isner JM, McKenna WJ. 26th Bethesda conference: recommendations for determining eligibility for competition in athletes with cardiovascular abnormalities. Task Force 3: hypertrophic cardiomyopathy, myocarditis and other myopericardial diseases and mitral valve prolapse. J Am Coll Cardiol. 1994;24(4):880-5. doi:10.1016/0735-1097(94)90844-3. PMID: 7930220.
  4. Ommen SR, Ho CY, Asif IM, Balaji S, Burke MA, Day SM, et al. 2024 AHA/ACC/AMSSM/HRS/PACES/SCMR Guideline for the Management of Hypertrophic Cardiomyopathy: A Report of the American Heart Association/American College of Cardiology Joint Committee on Clinical Practice Guidelines. Circulation. 2024;149(23):e1239-e1311. doi:10.1161/CIR.0000000000001250. PMID: 38718139.
  5. Sanghvi MM, Dhall E, Anwar A Chahal C, O’Mahony C, Mohiddin SA, Savvatis K, et al. Hypertrophic cardiomyopathy management: a systematic review of the clinical practice guidelines and recommendations. Eur Heart J Qual Care Clin Outcomes. 2025;11(7):919-933. doi:10.1093/ehjqcco/qcae117. PMID: 39749832.
  6. Correia VM, Morales Ribeiro V, Henkes Machado R, Defante MLR, de Souza K, de Oliveira Silva Filho JR, et al. Exercise training in hypertrophic cardiomyopathy: a meta-analysis of randomised trials with trial sequential analysis. Heart. 2025. doi:10.1136/heartjnl-2025-326377. PMID: 41320293.
  7. Saberi S, Wheeler M, Bragg-Gresham J, Hornsby W, Agarwal PP, Attili A, et al. Effect of Moderate-Intensity Exercise Training on Peak Oxygen Consumption in Patients With Hypertrophic Cardiomyopathy: A Randomized Clinical Trial. JAMA. 2017;317(13):1349-1357. doi:10.1001/jama.2017.2503. PMID: 28306757.
  8. Bayonas-Ruiz A, Gimeno-Blanes JR, Muñoz-Franco FM, Sabater-Molina M, Bonacasa B. Exercise Training in Hypertrophic Cardiomyopathy: A Systematic Review, Meta-analysis, and Meta-regression. Sports Med Open. 2026;12(1). doi:10.1186/s40798-026-01059-0. PMID: 42455405.
  9. Scolari FL, Garbin HI, de Carvalho GD, Rodrigues FT, de Menezes RA, Correia EB, et al. Low Agreement Among Guidelines for Primary Prevention Implantable Cardioverter-Defibrillator Recommendations in Hypertrophic Cardiomyopathy. Am J Cardiol. 2025;236:86-91. doi:10.1016/j.amjcard.2024.11.007. PMID: 39557320.
  10. Amano M, Kitaoka H, Yoshikawa Y, Sakata Y, Dohi K, Tokita Y, et al. Validation of Guideline Recommendation on Sudden Cardiac Death Prevention in Hypertrophic Cardiomyopathy. JACC Heart Fail. 2025;13(6):1014-1026. doi:10.1016/j.jchf.2024.12.006. PMID: 40088231.
  11. Arbelo E, Protonotarios A, Gimeno JR, Arbustini E, Barriales-Villa R, Basso C, et al. 2023 ESC Guidelines for the management of cardiomyopathies. Eur Heart J. 2023;44(37):3503-3626. doi:10.1093/eurheartj/ehad194. PMID: 37622657.
  12. Zampieri M, Saberi S, D’Ascenzi F, Papadakis M, Pelliccia A, Olivotto I. Hypertrophic cardiomyopathy: changing the paradigm of exercise prescription and competitive sport participation. Eur Heart J. 2026;47(8):902-915. doi:10.1093/eurheartj/ehaf770. PMID: 41065586.
  13. McKinney J, Isserow M, Wong J, Isserow S, Moulson N. New Insights and Recommendations for Athletes With Hypertrophic Cardiomyopathy. Can J Cardiol. 2024;40(5):921-933. doi:10.1016/j.cjca.2024.02.008. PMID: 38369259.
  14. Gudmundsdottir HL, Axelsson Raja A, Rossing K, Rasmusen H, Snoer M, Andersen LJ, et al. Exercise Training in Patients With Hypertrophic Cardiomyopathy Without Left Ventricular Outflow Tract Obstruction: A Randomized Clinical Trial. Circulation. 2025;151(2):132-144. doi:10.1161/CIRCULATIONAHA.124.070064. PMID: 39513255.
  15. Dybro AM, Rasmussen TB, Nielsen RR, Ladefoged BT, Pedersen ALD, Terkelsen CJ, et al. Effects of Alcohol Septal Ablation on Exercise Hemodynamics, Symptoms, and Quality of Life in Obstructive Hypertrophic Cardiomyopathy. J Am Heart Assoc. 2025;14(16):e041328. doi:10.1161/JAHA.125.041328. PMID: 40827012.
  16. Wang J, Yang S, Ma X, Zhao K, Yang K, Yu S, et al. Assessment of late gadolinium enhancement in hypertrophic cardiomyopathy improves risk stratification based on current guidelines. Eur Heart J. 2023;44(45):4781-4792. doi:10.1093/eurheartj/ehad581. PMID: 37795986.
  17. Linden K, Green N, Muir A, Malcolmson JW, Mohiddin SA, O’Mahony C. Size Matters. Prevalence and Clinical Implications of Left Ventricular Apical Aneurysms in Hypertrophic Cardiomyopathy: A Systematic Review and Meta-Analysis. J Am Heart Assoc. 2025;14(15):e037234. doi:10.1161/JAHA.124.037234. PMID: 40728182.
  18. Smith ED, Tome J, Mcgrath R, Kumar S, Concannon M, Day SM, et al. Exercise hemodynamics in hypertrophic cardiomyopathy identify risk of incident heart failure but not ventricular arrhythmias or sudden cardiac death. Int J Cardiol. 2019;274:226-231. doi:10.1016/j.ijcard.2018.07.110. PMID: 30075967.
  19. Petek BJ, Churchill TW, Moulson N, Kliethermes SA, Baggish AL, Drezner JA, et al. Sudden Cardiac Death in National Collegiate Athletic Association Athletes: A 20-Year Study. Circulation. 2024;149(2):80-90. doi:10.1161/CIRCULATIONAHA.123.065908. PMID: 37955565.
  20. Kim JH, Martinez MW, Guseh JS, Krishnan S, Gray B, Harmon KG, et al. A contemporary review of sudden cardiac arrest and death in competitive and recreational athletes. Lancet. 2024;404(10468):2209-2222. doi:10.1016/S0140-6736(24)02086-5. PMID: 39616000.
  21. Kim JH, Rim AJ, Miller JT, Jackson M, Patel N, Rajesh S, et al. Cardiac Arrest During Long-Distance Running Races. JAMA. 2025;333(19):1699-1707. doi:10.1001/jama.2025.3026. PMID: 40159341.
  22. Kwon S, Lee HJ, Han KD, Kim DH, Lee SP, Hwang IC, et al. Association of physical activity with all-cause and cardiovascular mortality in 7666 adults with hypertrophic cardiomyopathy (HCM): more physical activity is better. Br J Sports Med. 2021;55(18):1034-1040. doi:10.1136/bjsports-2020-101987. PMID: 32967852.
  23. MacNamara JP, Dias KA, Hearon CM Jr, Ivey E, Delgado VA, Saland S, et al. Randomized Controlled Trial of Moderate- and High-Intensity Exercise Training in Patients With Hypertrophic Cardiomyopathy: Effects on Fitness and Cardiovascular Response to Exercise. J Am Heart Assoc. 2023;12(20):e031399. doi:10.1161/JAHA.123.031399. PMID: 37830338.
  24. Basu J, Nikoletou D, Miles C, MacLachlan H, Parry-Williams G, Tilby-Jones F, et al. High intensity exercise programme in patients with hypertrophic cardiomyopathy: a randomized trial. Eur Heart J. 2025;46(19):1803-1815. doi:10.1093/eurheartj/ehae919. PMID: 40037382.
  25. Lampert R, Ackerman MJ, Marino BS, Burg M, Ainsworth B, Salberg L, et al. Vigorous Exercise in Patients With Hypertrophic Cardiomyopathy. JAMA Cardiol. 2023;8(6):595-605. doi:10.1001/jamacardio.2023.1042. PMID: 37195701.
  26. Pelliccia A, Caselli S, Pelliccia M, Musumeci MB, Lemme E, Di Paolo FM, et al. Clinical outcomes in adult athletes with hypertrophic cardiomyopathy: a 7-year follow-up study. Br J Sports Med. 2020;54(16):1008-1012. doi:10.1136/bjsports-2019-100890. PMID: 32532845.
  27. ClinicalTrials.gov. NCT03335332: registered study record. National Library of Medicine. https://clinicaltrials.gov/study/NCT03335332
  28. Maron MS, Masri A, Nassif ME, Barriales-Villa R, Arad M, Cardim N, et al. Aficamten for Symptomatic Obstructive Hypertrophic Cardiomyopathy. N Engl J Med. 2024;390(20):1849-1861. doi:10.1056/NEJMoa2401424. PMID: 38739079.
  29. Amin AM, Shehada W, Elgebaly M, Ayman D, Ibrahim MM, Abdelazeem B, et al. Efficacy and safety of exercise for hypertrophic cardiomyopathy: a systematic review and meta-analysis. Proc (Bayl Univ Med Cent). 2026;39(4):669-681. doi:10.1080/08998280.2026.2667666. PMID: 42269051.
  30. Borrelli F, Tarditi F, Andreis A, Mabritto B, Benso AS, Toso E. Safety and efficacy of physical activity in hypertrophic cardiomyopathy patients: a systematic review and metanalysis. Front Cardiovasc Med. 2025;12:1739956. doi:10.3389/fcvm.2025.1739956. PMID: 41907051.
  31. Pelliccia A, Ferrera A, Daniello CD, Di Gioia G, Maestrini V, Monosilio S, et al. Clinical phenotype and outcomes of low-risk HCM patients who maintained stable exercise habits during follow-up. Int J Cardiol. 2026;461:134676. doi:10.1016/j.ijcard.2026.134676. PMID: 42468694.
  32. Schoonvelde SAC, Zwetsloot PP, Hirsch A, Schinkel AFL, Knackstedt C, Germans T, et al. Accelerometry-defined physical activity and quality of life in hypertrophic cardiomyopathy. Heart. 2026;112(6):333-340. doi:10.1136/heartjnl-2025-326217. PMID: 40887133.
  33. Lee HJ, Gwak SY, Kim K, Cho I, Shim CY, Ha JW, et al. Factors associated with high-intensity physical activity and sudden cardiac death in hypertrophic cardiomyopathy. Heart. 2025;111(6):253-261. doi:10.1136/heartjnl-2024-324928. PMID: 39848653.
  34. Myon F, Delatour M, Behar N, Carré F, Donal E, Schnell F. Genotype status and arrhythmic risk in sport participants with hypertrophic cardiomyopathy. Eur Heart J Open. 2026;6(4):oeag127. doi:10.1093/ehjopen/oeag127. PMID: 42662846.
  35. Nesti M, Notarstefano P, Segreti L, Giaccardi M, Russo V, Solarino G, et al. Subcutaneous ICD sports safety registry (SISS registry). Int J Cardiol. 2025;443:133945. doi:10.1016/j.ijcard.2025.133945. PMID: 41052579.
  36. Willixhofer R, Mapelli M, Baracchini N, Campana N, Capovilla TM, Nava A, et al. RoMa: A Cardiopulmonary Exercise Testing Based Risk Tool in Hypertrophic Cardiomyopathy. J Am Heart Assoc. 2026;15(8):e046438. doi:10.1161/JAHA.125.046438. PMID: 41954058.
  37. Balaji S, DiLorenzo MP, Fish FA, Etheridge SP, Aziz PF, Russell MW, et al. Risk factors for lethal arrhythmic events in children and adolescents with hypertrophic cardiomyopathy and an implantable defibrillator: An international multicenter study. Heart Rhythm. 2019;16(10):1462-1467. doi:10.1016/j.hrthm.2019.04.040. PMID: 31026510.
  38. Ommen SR, Mital S, Burke MA, Day SM, Deswal A, Elliott P, et al. 2020 AHA/ACC Guideline for the Diagnosis and Treatment of Patients With Hypertrophic Cardiomyopathy: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2020;142(25):e558-e631. doi:10.1161/CIR.0000000000000937. PMID: 33215931.
  39. Reineck E, Rolston B, Bragg-Gresham JL, Salberg L, Baty L, Kumar S, et al. Physical activity and other health behaviors in adults with hypertrophic cardiomyopathy. Am J Cardiol. 2013;111(7):1034-9. doi:10.1016/j.amjcard.2012.12.018. PMID: 23340032.
  40. Asif IM, Price D, Fisher LA, Zakrajsek RA, Larsen LK, Raabe JJ, et al. Stages of psychological impact after diagnosis with serious or potentially lethal cardiac disease in young competitive athletes: a new model. J Electrocardiol. 2015;48(3):298-310. doi:10.1016/j.jelectrocard.2014.12.018. PMID: 25680944.
  41. Martinez KA, Bos JM, Tobert KE, Giudicessi JR, Ackerman MJ. Outcomes and Burdens to Return-to-Play for Phenotype Negative Athletes With a Genetic Heart Disease. JACC Clin Electrophysiol. 2025;11(8):1708-1717. doi:10.1016/j.jacep.2025.03.013. PMID: 40372328.
  42. Drezner JA, Malhotra A, Prutkin JM, Papadakis M, Harmon KG, Asif IM, et al. Return to play with hypertrophic cardiomyopathy: are we moving too fast? A critical review. Br J Sports Med. 2021;55(18):1041-1047. doi:10.1136/bjsports-2020-102921. PMID: 33472848.
  43. Smith JR, Layrisse V, Medina-Inojosa JR, Berg JD, Ommen SR, Olson TP. Predictors of exercise capacity following septal myectomy in patients with hypertrophic cardiomyopathy. Eur J Prev Cardiol. 2020;27(10):1066-1073. doi:10.1177/2047487319898106. PMID: 31967491.
  44. Russell WW, Edelson JB. Exercise and Sport Participation in Pediatric Hypertrophic Cardiomyopathy: A Changing Paradigm. Curr Treat Options Cardiovasc Med. 2026;28(1):18. doi:10.1007/s11936-026-01150-5. PMID: 42553283.
  45. Wanner J, Burstein D, Paridon S, Shah M, O’Malley S, Lin KY, et al. Feasibility of an Individualized Exercise Training Program for Children with Hypertrophic Cardiomyopathy. Pediatr Cardiol. 2026;47(2):795-802. doi:10.1007/s00246-025-03843-3. PMID: 40234258.
  46. Stefan MA, Hopman WM, Smythe JF. Effect of activity restriction owing to heart disease on obesity. Arch Pediatr Adolesc Med. 2005;159(5):477-81. doi:10.1001/archpedi.159.5.477. PMID: 15867123.
  47. Christian S, Somerville M, Giuffre M, Atallah J. Physical activity restriction for children and adolescents diagnosed with an inherited arrhythmia or cardiomyopathy and its impact on body mass index. J Cardiovasc Electrophysiol. 2018;29(12):1648-1653. doi:10.1111/jce.13713. PMID: 30106210.
  48. Souilla L, Werner O, Huguet H, Gavotto A, Vincenti M, Pasquie JL, et al. Cardiopulmonary Fitness and Physical Activity Among Children and Adolescents With Inherited Cardiac Disease. JAMA Netw Open. 2025;8(2):e2461795. doi:10.1001/jamanetworkopen.2024.61795. PMID: 39998828.
  49. Christian S, Somerville M, Taylor S, Spence JC, Giuffre M, Atallah J. The impact of physical activity modification on the well-being of a cohort of children with an inherited arrhythmia or cardiomyopathy. Cardiol Young. 2020;30(5):692-697. doi:10.1017/S1047951120000803. PMID: 32286207.
  50. Eijsvogels TMH, Kim JH, Aengevaeren VL, D’Ascenzi F, Churchill TW, Dineen EH, et al. Masters Athletes With Abnormal Cardiovascular Findings: A Clinical Consensus Statement of the European Association of Preventive Cardiology of the ESC and the American College of Cardiology. J Am Coll Cardiol. 2026;88(3):389-411. doi:10.1016/j.jacc.2026.03.025. PMID: 41941267.
  51. Ajufo E, Kany S, Jurgens SJ, Churchill TW, Guseh JS, Aragam KG, et al. Physical Activity and Cardiovascular Outcomes in Phenotype-Negative Cardiomyopathy Variant Carriers. JAMA Cardiol. 2026;11(1):45-56. doi:10.1001/jamacardio.2025.4466. PMID: 41205222.
  52. Abdeldayem J, Abdelfattah OM, Chaabo O, El Haddad D, Sharma G, Ahnert AM, et al. Long-Term Impact of Pregnancy on Clinical Outcomes in Individuals With Hypertrophic Cardiomyopathy. JACC Adv. 2025;4(1):101426. doi:10.1016/j.jacadv.2024.101426. PMID: 39811753.
  53. ClinicalTrials.gov. NCT03532802: registered study record. National Library of Medicine. https://clinicaltrials.gov/study/NCT03532802
  54. Dybro AM, Rasmussen TB, Nielsen RR, Andersen MJ, Jensen MK, Poulsen SH. Randomized Trial of Metoprolol in Patients With Obstructive Hypertrophic Cardiomyopathy. J Am Coll Cardiol. 2021;78(25):2505-2517. doi:10.1016/j.jacc.2021.07.065. PMID: 34915981.
  55. ClinicalTrials.gov. NCT05767346: registered study record. National Library of Medicine. https://clinicaltrials.gov/study/NCT05767346
  56. Lewis GD, Garcia-Pavia P, Masri A, Merkely B, Nassif ME, Peña-Peña ML, et al. Exercise Performance With Aficamten vs Metoprolol in Obstructive Hypertrophic Cardiomyopathy: The MAPLE-HCM Randomized Clinical Trial. JAMA Cardiol. 2026;11(8):736-746. doi:10.1001/jamacardio.2026.1730. PMID: 42307914.
  57. Olivotto I, Oreziak A, Barriales-Villa R, Abraham TP, Masri A, Garcia-Pavia P, et al. Mavacamten for treatment of symptomatic obstructive hypertrophic cardiomyopathy (EXPLORER-HCM): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet. 2020;396(10253):759-769. doi:10.1016/S0140-6736(20)31792-X. PMID: 32871100.
  58. ClinicalTrials.gov. NCT05186818: registered study record. National Library of Medicine. https://clinicaltrials.gov/study/NCT05186818
  59. Desai MY, Owens AT, Abraham T, Olivotto I, Garcia-Pavia P, Lopes RD, et al. Mavacamten in Symptomatic Nonobstructive Hypertrophic Cardiomyopathy. N Engl J Med. 2025;393(10):961-972. doi:10.1056/NEJMoa2505927. PMID: 40888717.
  60. ClinicalTrials.gov. NCT05582395: registered study record. National Library of Medicine. https://clinicaltrials.gov/study/NCT05582395
  61. Wheeler MT, Olivotto I, Elliott PM, Saberi S, Owens AT, Maurer MS, et al. Effects of Mavacamten on Measures of Cardiopulmonary Exercise Testing Beyond Peak Oxygen Consumption: A Secondary Analysis of the EXPLORER-HCM Randomized Trial. JAMA Cardiol. 2023;8(3):240-247. doi:10.1001/jamacardio.2022.5099. PMID: 36652223.
  62. Campain J, Griskowitz C, Newlands C, Claggett BL, Kulac IJ, McGinnis S, et al. Characterization and Application of Novel Exercise Recovery Patterns That Reflect Cardiac Performance: A Substudy of the SEQUOIA-HCM Trial. Circulation. 2025;152(14):990-1002. doi:10.1161/CIRCULATIONAHA.124.073585. PMID: 40910168.
  63. Cavigli L, Olivotto I, Fattirolli F, Mochi N, Favilli S, Mondillo S, et al. Prescribing, dosing and titrating exercise in patients with hypertrophic cardiomyopathy for prevention of comorbidities: Ready for prime time. Eur J Prev Cardiol. 2021;28(10):1093-1099. doi:10.1177/2047487320928654. PMID: 33611579.
  64. Cavigli L, Ragazzoni GL, Vannuccini F, Targetti M, Mandoli GE, Senesi G, et al. Cardiopulmonary Fitness and Personalized Exercise Prescription in Patients With Hypertrophic Cardiomyopathy. J Am Heart Assoc. 2024;13(20):e036593. doi:10.1161/JAHA.124.036593. PMID: 39392151.

Leave a Comment