Clinical background
The diagnosis of Brugada syndrome has traditionally rested on a single criterion: a spontaneous type 1 Brugada pattern in the right precordial leads. This leaves a large number of patients in a grey zone, particularly those with a type 2/3 pattern requiring provocation testing, or those with a history of arrhythmic events but a non-diagnostic ECG. The decision on how far to pursue investigation, whether genetic testing should be performed and whether an ICD is indicated then becomes a clinical guess without a structured basis.
The Shanghai score was developed to replace the binary diagnostic criterion with a multiparametric assessment that combines ECG findings, clinical history, family history and genetics. The score is therefore a diagnostic instrument, not a tool for risk stratification of arrhythmic events, even though validation studies have examined its prognostic properties secondarily.
Calculating the Shanghai score for Brugada syndrome
The score is the sum of the single highest-scoring item in each of three categories (ECG, clinical history, family history) plus an addition for genetics:
where genetics contributes 0 or 0.5 points depending on whether a probably pathogenic mutation in a Brugada-associated gene is identified.
The four categories and their point values are:
| Category | Highest score within the category |
|---|---|
| Spontaneous type 1 Brugada pattern | 3.5 |
| Fever- or drug-induced type 1 Brugada pattern | 3.0 |
| Type 2/3 pattern converting to type 1 on provocation testing | 2.0 |
| Unexplained cardiac arrest or documented VF/polymorphic VT | 3.0 |
| Suspected arrhythmic syncope or nocturnal agonal respiration | 2.0 |
| Syncope of unclear mechanism | 1.0 |
| Atrial fibrillation/flutter in a patient <30 years without an alternative cause | 0.5 |
| First- or second-degree relative with confirmed Brugada syndrome | 2.0 |
| Suspected sudden cardiac death (fever-, night- or drug-related) in a first- or second-degree relative | 1.0 |
| Unexplained sudden cardiac death <45 years in a first- or second-degree relative (negative autopsy) | 0.5 |
| Probably pathogenic mutation in a Brugada-associated gene | 0.5 |
The score was not derived from a statistical model but was constructed by expert consensus in the 2016 HRS/EHRA/APHRS/SOLAECE consensus report on J-wave syndromes, in which the scoring was based on the available literature and weighted coefficients from limited datasets [1]. The consensus document states explicitly that the scoring system requires ongoing validation.
Interpretation in practice
| Score | Interpretation | Clinical action |
|---|---|---|
| <2 | Non-diagnostic | Brugada syndrome is unlikely. Consider alternative diagnoses. Follow up if suspicion persists, particularly during febrile episodes. |
| 2–3 | Possible Brugada syndrome | Further investigation is warranted: genetic testing, family screening, consider repeat ECGs or ambulatory monitoring. An ICD is not indicated on the basis of the score alone. |
| ≥3.5 | Probable/definite Brugada syndrome | The diagnosis can be made. Risk stratification for an ICD according to current guidelines (syncope, documented VT/VF, inducibility at EPS). |
A patient with only a drug- or fever-induced type 1 pattern (3.0 points) and no additional history or family history falls into the "possible" band and therefore does not meet the criteria for definite Brugada syndrome. This is a deliberate design feature: the consensus document downgraded the isolated provocation-induced type 1 pattern from being diagnostic to requiring support from additional clinical features [1,4]. The 2022 ESC guideline on ventricular arrhythmias and sudden cardiac death follows this view and states that a drug-induced type 1 pattern strengthens the diagnosis only in the presence of relevant symptoms or a family history [4].
Validation and performance
The first external validation was performed by Kawada et al. in a cohort of 393 patients evaluated for Brugada syndrome at Okayama University in Japan between 1996 and 2016 [2]. Of these, 271 were asymptomatic, 99 had syncope and 23 had documented ventricular fibrillation (VF). In total, 348 patients (88%) reached a score ≥3.5 and were classified as probable/definite Brugada syndrome. Over a median follow-up of 97.3 months, 43 patients suffered VF. The difference in arrhythmic events between the score groups was significant (p = 0.01). Crucially, no patient with a score <3.5 (possible or non-diagnostic) experienced a malignant arrhythmic event during follow-up [2].
When the Shanghai score has been tested as a prognostic instrument for predicting arrhythmic events, discrimination has been moderate. A 2025 systematic review by Gomes et al., which included 16 studies and 11 unique risk models for Brugada syndrome, calculated a pooled c-statistic for the Shanghai score of AUC 0.69 (95% CI 0.61–0.76) in patients in primary prevention, the lowest among the models compared [3]. The same review judged all included models to be at high risk of bias according to PROBAST, with common shortcomings in the form of inappropriate inclusion criteria, low event numbers and inadequate reporting of calibration [3].
A 2025 narrative review concludes that, used prognostically, the Shanghai score shows high sensitivity but low specificity, with c-statistics between 0.69 and 0.80 in various validation attempts [5]. The score has not been adopted in international guidelines as a prognostic tool, and neither the ESC nor the AHA/ACC/HRS recommends it for decisions on an ICD [5].
Limitations
The Shanghai score is a diagnostic instrument and should not be used as the basis for a decision on ICD implantation. Risk stratification in established Brugada syndrome rests on other variables, above all arrhythmic syncope and inducibility at electrophysiological study (EPS), which are the only predictors that guide management under current guidelines [5].
The ECG findings must be interpreted in the absence of confounders. Fever, myocardial ischaemia, electrolyte disturbances and certain drugs can mimic a type 1 Brugada pattern and lead to incorrect scoring. A fever-induced type 1 pattern in a patient with no other history or family history gives 3.0 points and hence the "possible" band, but the fever must be taken into account in the interpretation for the finding to be valid.
The scoring system is based on expert consensus, not on a formal statistical derivation from a defined cohort [1]. The weighting of the individual components is therefore not empirically validated in the same sense as for regression-based risk models. The Kawada validation came from a single centre in Japan with a patient population dominated by asymptomatic individuals, which limits generalisability to other ethnic groups and to patients with a broader range of symptoms [2].
The genetic component assigns 0.5 points for a probably pathogenic mutation, but the penetrance of Brugada-associated mutations, particularly in SCN5A, is incomplete, and many patients with clinically manifest Brugada syndrome have no identifiable mutation. The contribution of genetics to the score is therefore small and may both overestimate and underestimate the diagnostic probability depending on context.
References
- Antzelevitch C et al. J Wave Syndrome Expert Consensus Report: Diagnosis, Risk Stratification, and Management. Heart Rhythm 2016. PMID: 27810171
- Kawada S et al. Shanghai Score System for Diagnosis of Brugada Syndrome: Validation of the Score System and Reclassification of the Patients. JACC Clin Electrophysiol 2018. PMID: 29929664
- Gomes DA et al. Multiparametric models for predicting major arrhythmic events in Brugada syndrome: a systematic review and critical appraisal. Europace 2025. PMID: 40314213
- Behr ER et al. The diagnostic role of pharmacological provocation testing in cardiac electrophysiology: a clinical consensus statement. Europace 2025. PMID: 40165484
- Rattanawong P, Shen WK. Brugada syndrome risk scores: what we've learned and what's next. Front Cardiovasc Med 2025. PMID: 41696545
