Brugada Syndrome (BrS)

Details

Brugada syndrome is an inherited channelopathy defined by coved ST-segment elevation ≥2 mm with T-wave inversion in the right precordial leads — the type 1 pattern — in the absence of structural heart disease, predisposing to syncope and sudden cardiac death from ventricular fibrillation. It accounts for up to 28% of SCD in structurally normal hearts and ~5–10% of resuscitated cardiac arrest. Arrhythmias cluster at rest, during sleep and after large meals — periods of high vagal tone — rather than during exertion, a distribution that governs both risk assessment and sports advice.

Key Facts

Epidemiology

• Prevalence ~1:2,000 for the type 1 ECG pattern and ~1:500 for types 2 and 3; most common in Asia, then Europe and the United States. In children it falls to ~1:20,000, and phenotypic expression is age-dependent. (sources/brs-jaccep-2022, rating: high)
• The ethnic gradient is pronounced: type 1 ECG prevalence 0.0–0.94% in Asians, 0.15% in Japanese-Americans, 0.0–0.02% in Europeans and 0.005–0.1% in North Americans. (sources/inherited-arrhythmias-nrdp-2020, rating: very high)
• Males account for 80–90% of diagnosed cases, with the disparity appearing only after adolescence — a sex-hormone effect rather than a sex-linked one (mechanism under Pathophysiology).
• Two-thirds are asymptomatic at diagnosis; one-third present with syncope or palpitations, and SCD can be the first manifestation. (sources/channelopathies-jaha-2025, rating: high)
• Conduction disease coexists in a minority — first-degree AV block, RBBB, intraventricular conduction delay and sick sinus syndrome (sources/channelopathies-jaha-2025).

Pathophysiology

Three non-mutually-exclusive hypotheses — a confluence is the likeliest truth, which is why no single treatment targets them all:

(sources/brs-jaccep-2022, rating: high; sources/channelopathies-jaha-2025, rating: high)

• I_to is carried by Kv4.3, encoded by KCND3, whose amplitude is set by the β-subunit KChIP2 and modulated by KCNE3/MiRP2 and KCNE5/MiRP4. Gain-of-function KCNE3 Arg99His and X-linked KCNE5 variants raise I_to and are rare causes of BrS; I_to prominence in the atria also explains the atrial arrhythmias of BrS. (sources/brs-jaccep-2022; sources/membrane-potential-physrev-2021, rating: very high)
• KChIP2 explains the male predominance. Expression is higher in males under androgen-receptor signalling and suppressed by oestrogen, so a male with a given SCN5A loss-of-function variant starts with more I_to and reaches dome loss at a smaller reduction in I_Na — and premenopausal women are partly protected. (sources/membrane-potential-physrev-2021)
• BrS may be a partial cardiomyopathy. CMR increasingly shows structural RVOT abnormality — enlarged ventricular volumes and structural change — suggesting the substrate is regional myocardial disease rather than pure electrophysiology. (sources/scn5a-jaccep-2018, rating: high)
• The mechanism is argued, not settled. The Primer presents both sides: Antzelevitch's repolarisation hypothesis (phase-1 outward shift → heterogeneous epicardial AP-dome loss → phase-2 reentry) rests on the J wave diminishing at fast rates as Ito availability falls, quinidine's Ito-blocking benefit, absence of RVOT conduction delay in a whole-heart model, and the benefit of RVOT ablation destroying the cells with the most prominent notch. Wilde's depolarisation hypothesis (fibrosis, reduced connexin-43 and sodium channels → discontinuous conduction) counters that quinidine's benefit is expected under either model because changing AP morphology improves conduction safety in compromised regions — so the quinidine observation cannot adjudicate the debate. (sources/inherited-arrhythmias-nrdp-2020, rating: very high)
• Early repolarisation is the left-ventricular counterpart. Ito is most prominent in the LV inferior wall, so early repolarisation expresses there with J-point elevation in inferior leads; epicardial mapping shows fractionated signals in 75% of early-repolarisation patients, which the depolarisation camp reads as conduction delay and the repolarisation camp as delayed repolarisation. (sources/inherited-arrhythmias-nrdp-2020)
• SCN5A-BrS carries conduction disease at every level, more so than non-SCN5A BrS, and its prevalence and severity increase with age (sources/scn5a-jaccep-2018).

Genetics

• SCN5A encodes Na_V1.5, the predominant cardiac sodium channel. Loss of function reduces peak I_Na, slows the phase 0 upstroke and shortens action potential duration, predominantly in RVOT epicardium; the temperature dependence of Na_V1.5 is why fever precipitates events.
• SCN5A is found in only ~20% of probands and is now curated as definitively disease-causing through the broad "SCN5A-related cardiac rhythm disorder" designation (ClinGen, 10/08/2025). (sources/arrhythmia-genetics-mgenetik-2025, rating: high; sources/clingen-summary-2026-05-09)
• The 2020 reappraisal supports only SCN5A. At least 20 other genes had been reported, almost all from candidate-gene studies in single individuals or small families; testing for rare-variant burden against controls and ClinGen curation left SCN5A as the only gene with supported involvement, in ~20% of probands. Sporadic presentation, low penetrance and phenotype-positive genotype-negative relatives argue for complex rather than Mendelian inheritance. (sources/inherited-arrhythmias-nrdp-2020, rating: very high)
• GWAS localised common susceptibility variants near SCN5A and HEY2, and >4 of 6 risk alleles makes a spontaneous type 1 ECG more likely — direct evidence for the polygenic model and a nascent route to event prediction. (sources/inherited-arrhythmias-nrdp-2020)
• ClinGen has disputed or refuted nearly every other gene ever attributed to BrS:

• SLMAP, SEMA3A and SCNN1A were additionally reported by Campuzano et al. (2019) but have no ClinGen validation. See concepts/ClinGen-Gene-Disease-Validity.
• H558R (rs1805124) is a common modifier, not a cause. The polymorphism repairs abnormal gating kinetics and improves Na_V1.5 trafficking in pathogenic-variant carriers, reducing phenotypic severity; epigenetically it lowers SCN5A promoter methylation and raises cardiac expression, which may prevent VF. Paradoxically the G allele is also associated with QTc prolongation in population studies — a pleiotropic, context-dependent effect. (sources/genetics-va-fcvm-2022, rating: medium)
• The architecture is substantially polygenic. GWAS identifies 21 independent loci across 12 chromosomal regions, including HEY2 (a transcription factor central to RVOT development) and MAPRE2 — well beyond ion-channel genes. This explains familial clustering without an identifiable variant and the ~50% penetrance in SCN5A-positive families. (sources/brs-jaccep-2022; sources/gwas-arrhythmias-cmp-genes-2025, rating: high)
• The BrS polygenic risk score predicts the ECG, not the events. OR 2.12 per SD for a spontaneous type 1 pattern and for a positive sodium-channel-blocker provocation test, with unusual cross-ancestry portability into Japanese patients — but no prediction of VF or SCD. Susceptibility to the phenotype and susceptibility to arrhythmic death are dissociated. (sources/gwas-arrhythmias-cmp-genes-2025)

ECG Patterns and Diagnosis

(sources/channelopathies-jaha-2025, rating: high)

• Brugada is a pseudo-RBBB, not a true RBBB. The discriminator is a wide S wave in leads I and V6 — present in true RBBB, absent in Brugada — and Brugada ST-T changes are spontaneously dynamic. When a genuine RBBB coexists the Brugada pattern is concealed within it; the Chiale manoeuvre (right apical ventricular pacing with appropriately timed AV intervals) unmasks it. (sources/rbbb-ccr-2021, rating: medium)
• High precordial lead positions (V1/V2 at intercostal spaces 2–4) raise diagnostic sensitivity ~1.5-fold; test both standard and high leads. (sources/brs-jaccep-2022, rating: high)
• The Shanghai score integrates ECG type, clinical history, family history and genotype and accounts for phenocopies (RVOT compression, ischaemia, electrolyte disturbance, drug effect): ≥3.5 probable/definite, 2–3 possible, <2 non-diagnostic. A drug-induced type 1 scores only 2 points, so it requires additional features to reach a diagnosis. Full points table: concepts/Shanghai-Score-System. (sources/channelopathies-jaha-2025)
• Sodium-channel-blocker provocation is indicated for a type 2/3 pattern or a suggestive clinical or family history; agents and protocol are below. False positives run ~8% with high-dose ajmaline in BrS families (Tadros) but ~27% in AVNRT patients and 4.5% of healthy controls (Hasdemir) — protocols are not standardised, and a drug-induced type 1 in isolation remains insufficient for a definitive diagnosis. (sources/brs-jaccep-2022)
• Ajmaline is the most effective provocation agent, but its sensitivity and specificity are unknown; vagal loading — recording after a meal or exercise — is an alternative when drugs are unavailable. Additional clinical criteria are required to call a drug- or fever-induced type 1 a diagnosis. (sources/inherited-arrhythmias-nrdp-2020, rating: very high)
• Echocardiography in all patients to exclude structural disease; CMR in complex cases for RVOT delineation (sources/brs-jaccep-2022).
• Genetic testing is Class I when a spontaneous type 1 pattern is present, because it enables cascade screening. Only SCN5A is tested routinely; other genes only with a genetics expert, when ≥2 family members are affected and SCN5A is negative. (sources/arrhythmia-genetics-mgenetik-2025; sources/brs-jaccep-2022)
• Genetic testing is not indicated for an isolated type 2 or 3 pattern — those patterns are not diagnostic and are not an independent indication. Testing is triggered by a confirmed type 1 or by a clinical suspicion meeting diagnostic criteria. (sources/genetic-test-aha-2020, rating: high)
• Family screening for all first-degree relatives: adults get standard and high-lead ECG ± provocation, and one negative provocation is adequate. Children get an ECG at age 3 and every 3 years to 15, with provocation deferred past 15 because of paediatric adverse-event risk. (sources/brs-jaccep-2022)

Risk Stratification

(sources/brs-jaccep-2022, rating: high)

• Two established markers: spontaneous type 1 ECG (2–6× relative risk for resuscitated arrest or SCD across large cohorts) and cardiogenic syncope (2.5–5×). Non-cardiogenic syncope confers no excess risk — a distinction the history has to make.
• As a rule of thumb the Primer puts an asymptomatic spontaneous type 1 at ~1%/year malignant arrhythmia risk, against 3–7%/year in symptomatic patients depending on presentation — the spread that decides who is offered a device. (sources/inherited-arrhythmias-nrdp-2020, rating: very high)
• A progressive fractionated ECG in the right precordial leads is described as probably the most reliable risk marker in asymptomatic patients; programmed stimulation may add information only with a low-intensity protocol. (sources/inherited-arrhythmias-nrdp-2020)
• "Brugada burden" is a third axis. Spatial burden (type 1 changes beyond the right precordial leads) and temporal burden (a higher proportion of spontaneous type 1 ECGs over follow-up, greater ST burden on 24-hour Holter) each independently predict events.
• Age and sex are not independent after adjustment for ECG pattern and syncope; patients diagnosed at ≥55 have event rates comparable to the general population. Family history of SCD is inconsistently predictive, though early familial SCD (first-degree relative <35) may confer risk (Sieira).
• Additional ECG markers are unvalidated in large adjusted studies — fragmented QRS, QRS duration, S-wave duration, rJ interval, early repolarisation, T_peak-end, QTc, signal-averaged ECG. Atrial fibrillation has been reported as an independent predictor (Calò).
• Programmed ventricular stimulation is, at best, a tie-breaker. FINGER (n=1,029) and PRELUDE (n=308) failed to show predictive utility; a pooled analysis (Sroubek, n=1,312) found OR 2.7 with low positive predictive value; more recent pooling supports a small role for a non-aggressive protocol limited to double extrastimuli from the right ventricular apex — not triple extrastimuli or multiple sites.
• No score performs adequately. The Sieira score does not stratify intermediate-risk patients (Probst) and the Honarbakhsh model awaits external validation.

Conservative Management

• Avoid aggravating drugs (brugadadrugs.org), cocaine, cannabis and excessive alcohol. (sources/VA-SCD-ESC-2022, rating: very high)
• Treat fever promptly — it unmasks the phenotype and precipitates events, especially in children — and correct hypokalaemia, hyperkalaemia and metabolic acidosis. (sources/brs-jaccep-2022, rating: high)
• No competitive sports restriction is supported by data. Participation is reasonable after expert assessment, with hydration prioritised and heat exhaustion, dehydration and exercise during febrile illness avoided. This aligns with the physiology: arrhythmias occur at high vagal tone, and exercise-triggered events are not a characteristic feature. (sources/competitive-sports-aha-2025, rating: very high)

Antiarrhythmic Therapy

• Quinidine inhibits I_to (its principal BrS action), I_CaL and I_Na, prolonging the effective refractory period and abolishing VF inducibility at programmed stimulation in ~90% (Belhassen, n=60). Indications: recurrent ICD shocks, ICD declined or contraindicated, and co-management of atrial fibrillation. Low dose (≤600 mg/day) improves tolerability; ~1/3 stop for side effects; regular blood counts are required. (sources/channelopathies-jaha-2025, rating: high; sources/brs-jaccep-2022)
• QUIDAM was underpowered and negative — zero severe arrhythmic events in the hydroquinidine arm, no demonstrated benefit. Clinical use rests on observational and mechanistic data, not on randomised evidence.
• Intravenous isoproterenol is the acute treatment for electrical storm and short-coupled PVC-triggered VF, potentiating I_CaL. PDE III inhibitors (cilostazol, milrinone) are the alternative, working the same way.

Device Therapy

• Secondary prevention ICD is Class I after resuscitated cardiac arrest, and primary prevention is recommended for BrS with cardiogenic syncope on either a spontaneous or provoked type 1 pattern.
• Asymptomatic spontaneous type 1: close follow-up rather than primary prevention ICD unless additional high-risk features are present — an expert shared decision. Uncertain syncope may warrant an implantable loop recorder.
• The complication burden is the argument against liberal implantation: 3.3%/year inappropriate shocks plus 4.5%/year other complications (lead malfunction, infection, psychological sequelae) in a meta-analysis of 1,539 patients, against an SCD incidence of ~0.19%/year without an ICD and ~0.10%/year with one (Probst, n=1,613).
• Subcutaneous versus transvenous: ~15% of BrS patients fail initial S-ICD sensing screening. S-ICD is preferred in the young with no pacing requirement; dual-chamber transvenous systems are preferred when sinus node dysfunction or atrial arrhythmia discrimination is needed.

Catheter Ablation

• Combined epicardial and endocardial approach: epicardial substrate modification in the anterior RVOT plus endocardial elimination of triggering PVCs, with sodium-channel-blocker provocation during the procedure to unmask additional substrate. Proposed endpoint is resolution of J-point elevation despite provocation.
• Pappone series (n=135): acute ECG normalisation in 100%, maintained in 133/135 (98.5%) at a median of 10 months.
• The original description (9 severely symptomatic patients) ablated an area of abnormal low-amplitude, late-to-very-late (>200 ms) fractionated activity in the RVOT; the ECG normalised in 8 and events were prevented. A larger cohort extended this to asymptomatic patients, with the ECG pattern abolished and no longer inducible by ajmaline or flecainide, and no recurrence in 39 symptomatic patients over ~10 months. (sources/inherited-arrhythmias-nrdp-2020, rating: very high)
• Indications: recurrent ICD shocks uncontrolled by drugs, and patients who decline or cannot have an ICD. Data are insufficient for asymptomatic patients — where it is Class III; the Primer explicitly discourages prophylactic ablation in asymptomatic patients, given the complication risk (tamponade, coronary injury) against a low event rate. (sources/brs-jaccep-2022; sources/inherited-arrhythmias-nrdp-2020)

Drug-Induced Brugada Syndrome

• Definition: a normal pretreatment ECG that develops a type 1 pattern after drug exposure; the genetic substrate may be present but previously silent. (sources/drug-arrhythmia-aha-2020, rating: very high)
• Culprits act by I_Na blockade unless stated: antiarrhythmics (ajmaline, pilsicainide, flecainide, procainamide, propafenone); tricyclics (2.3–15.3% in overdose); bupivacaine, propofol, cocaine, lithium, oxcarbazepine; and alcohol by I_CaL blockade. Induction reaches 39–48.2% with ajmaline in a high-pretest population; the maintained list is at brugadadrugs.org.
• The risk is an order of magnitude lower than spontaneous BrS — SCD ~0.08%/year against 0.8–1.2%/year for asymptomatic spontaneous type 1 — and it is often asymptomatic, though VT/VF can appear weeks to years after starting the drug and fever precipitates it.
• Ambulatory monitoring after drug withdrawal is mandatory, because many such patients turn out to have spontaneous type 1 patterns (sources/drug-arrhythmia-aha-2020).
• Management: stop the agent; monitor; cardiovert or defibrillate VT/VF, with amiodarone for recurrence — procainamide is contraindicated, since it adds I_Na blockade. ICD for prior cardiac arrest, sustained VT, or syncope with a spontaneous type 1. Asymptomatic drug-induced BrS needs avoidance and observation only.

SCB Provocation Testing — 2025 EHRA Consensus

Indications and contraindications, each carrying >90% consensus (sources/pharmacological-provocation-europace-2025, rating: high):

• Protocol: ajmaline preferred where available (class 1A, t½ ~5 min, maximum 1 mg/kg or 100 mg over 5–10 min); high precordial lead positions throughout; continuous ECG or at minimum every 30–60 s; isoproterenol immediately available. Higher-risk patients — pre-existing AV conduction disturbance, SCN5A variants — should be tested in a catheter laboratory with temporary pacing standby.
• A positive test in a family member without an SCN5A variant may reflect polygenic susceptibility, not monogenic BrS: common variants predict ajmaline response independently. What a drug-induced type 1 means for the family is therefore genuinely uncertain.

ESC 2022 Recommendations

(sources/VA-SCD-ESC-2022, rating: very high)

Emerging Therapies

• MOG1 overexpression — AAV9 delivery of the Na_V1.5 chaperone enhanced cell-surface trafficking and reversed the arrhythmic phenotype in SCN5A-G1746R knock-in mice, working around the ~6,048 bp SCN5A coding sequence that exceeds AAV's ~4.7 kb capacity. In SCN5A p.D1275N knock-in mice with dilated cardiomyopathy, sinus node dysfunction and distal conduction disease it improved all three, with ≥90% ventricular cardiomyocyte transduction. (sources/gene-therapy-arrhythmia-2025, rating: high; sources/gene-therapy-arrhythmia-cjca-2026)
• SCN10A-short is a second, independent route to more I_Na. An intronic enhancer drives a small protein from the last seven exons; it, rather than Na_V1.5, carries the I_Na increase associated with SCN10A intronic SNPs. AAV9-SCN10A-short restored upstroke velocity, peak I_Na density and conduction velocity to wild type in SCN5A haploinsufficient mice — and because it is small and needs no full SCN5A copy, it is plausible for the ~80% of BrS that is genotype-negative or has a trafficking-independent defect. (sources/gene-therapy-arrhythmia-cjca-2026)
• The caveat: MOG1 addresses trafficking-limited Na_V1.5 and SCN10A-short addresses current amplitude, so the choice depends on the defect a given variant produces — and neither has been tested in a large animal or a human. No BrS gene therapy is in clinical trial, and the review notes the bar is high: most patients already do well on quinidine, epicardial RVOT ablation or an ICD, and the population is largely asymptomatic.
• ARumenamide-787 inhibits I_to and enhances I_Na, with promising effects in canine BrS/ERS models; RNA-based and noncoding RNA therapeutics remain preclinical. (sources/channelopathies-jaha-2025)

Contradictions / Open Questions

• The pathophysiology is unresolved, and treatment has chosen a side. Depolarisation, repolarisation and neural-crest hypotheses remain unreconciled. Ablation targets the depolarisation substrate; quinidine targets repolarisation. Any guideline endorsement of ablation implicitly assumes conduction disease without the debate having been settled. (sources/channelopathies-jaha-2025, rating: high; sources/brs-jaccep-2022, rating: high)
• The same observation is claimed by both camps. Quinidine's benefit is cited by the repolarisation hypothesis as Ito blockade and by the depolarisation hypothesis as improved conduction safety from altered AP morphology — so the one therapeutic fact the field agrees on cannot distinguish the two models. The Primer presents this disagreement in the authors' own words rather than resolving it. (sources/inherited-arrhythmias-nrdp-2020, rating: very high)
• Polygenic architecture versus monogenic counselling. GWAS suggests common variants explain most cases; penetrance in SCN5A-positive families is only ~50%; relatives without the variant may still have clinical BrS. Family screening cannot rest on genetic testing alone. (sources/brs-jaccep-2022)
• The PRS predicts the ECG phenotype but not arrhythmic events, so it can identify polygenic electrophysiological susceptibility without saying who needs an ICD. The GWAS architecture explains why ~80% of BrS is genotype-negative; it does not resolve risk. (sources/gwas-arrhythmias-cmp-genes-2025, rating: high)
• SCN5A testing yields a positive result in the minority. It is causative in only 20–30% of probands, so a Class I recommendation generally returns a negative — and a negative result does not exclude the diagnosis. (sources/channelopathies-jaha-2025; sources/VA-SCD-ESC-2022, rating: very high)
• Programmed stimulation is recommended despite failing its registries. FINGER and PRELUDE showed no predictive utility, yet ESC retains Class IIb. Whether PVS adds anything once all non-invasive ECG markers are modelled jointly has never been tested in a large prospective study.
• Ablation is endorsed for shocks and forbidden for the asymptomatic, with nothing in between. ESC upgraded ablation for refractory recurrent shocks to Class IIa while ablation in asymptomatic BrS is Class III — a binary that does not fit the intermediate-risk patient with a spontaneous type 1, a family history of SCD and no personal event.
• The ICD may do net harm in asymptomatic BrS. Annual SCD without an ICD (~0.19%) is lower than the combined annual complication rate (3.3% inappropriate shocks + 4.5% other), which is the arithmetic behind withholding primary prevention devices from patients without additional risk features.
• The only randomised drug trial was underpowered. QUIDAM did not demonstrate benefit over placebo for hydroquinidine, and the entire pharmacologic recommendation rests on observational and mechanistic evidence. (sources/brs-jaccep-2022)
• ClinGen has refuted or disputed nearly the whole historical gene list. Of the ~20+ genes attributed to BrS, almost all non-SCN5A genes are now Disputing or Refuting — TRPM4 and SCN3B refuted in 12/2025. A positive panel result in these genes should not be called causative, and the 21-locus polygenic model is the mechanistic explanation for why they were ever associated. (sources/clingen-summary-2026-05-09, rating: high)

Connections

• Related to entities/SCN5A — the gene, the channel and its variant interpretation
• Related to entities/Early-Repolarization-Syndrome — the overlapping phenotype, shared I_to mechanism and Shanghai scoring
• Related to concepts/Shanghai-Score-System — the full diagnostic points table
• Related to concepts/Pharmacological-Provocation-Testing — the provocation procedure in general use
• Related to concepts/Channelopathies — the wider inherited arrhythmia group
• Related to concepts/ClinGen-Gene-Disease-Validity — how the BrS gene list collapsed
• Related to concepts/Genetic-Testing-in-Arrhythmias — who to test, what the panel should contain, and what a negative result means
• Related to concepts/Variant-Reclassification — why old BrS gene reports are no longer causal
• Related to concepts/GWAS-Cardiac-Genetics — the polygenic architecture
• Related to concepts/Polygenic-Risk-Score — what a PRS can and cannot tell a patient
• Related to concepts/Cardiac-Action-Potential — phase 1 I_to and the AP dome
• Related to concepts/Electrical-Remodeling — the substrate that develops with age
• Related to concepts/Electrical-Storm — the acute presentation and its drug management
• Related to concepts/Drug-Induced-Arrhythmia — the provoking drug list
• Related to concepts/RBBB — pseudo-RBBB and the Chiale manoeuvre
• Related to concepts/Sudden-Cardiac-Death — the outcome being prevented
• Related to concepts/Sports-Cardiology-SDM — the 2025 sports framework
• Related to concepts/Cardiogenetic-Centers — where provocation and family screening should happen
• Related to concepts/iPSC-Derived-Cardiomyocytes — the model system for mechanism work

Sources

Core reviews

• sources/brs-jaccep-2022
• sources/channelopathies-jaha-2025

Genetics

• sources/arrhythmia-genetics-mgenetik-2025
• sources/clingen-summary-2026-05-09
• sources/gwas-arrhythmias-cmp-genes-2025
• sources/scn5a-jaccep-2018
• sources/genetic-test-aha-2020
• sources/genetics-va-fcvm-2022
• sources/inherited-arrhythmias-nrdp-2020

Diagnosis and provocation

• sources/pharmacological-provocation-europace-2025
• sources/rbbb-ccr-2021

Management

• sources/VA-SCD-ESC-2022
• sources/drug-arrhythmia-aha-2020
• sources/competitive-sports-aha-2025

Physiology and emerging therapy

• sources/membrane-potential-physrev-2021
• sources/gene-therapy-arrhythmia-2025
• sources/gene-therapy-arrhythmia-cjca-2026