Appraisal of the 2024 revision of the level of evidence grading system in European Society of Cardiology Guidelines

The European Society of Cardiology (ESC) introduced evidence-based clinical practice guidelines in the field of cardiovascular medicine more than 30 years ago.^1,2 European Society of Cardiology Guidelines have proved invaluable to inform healthcare providers and patients with the goal of improving cardiovascular care and advancing implementation of established therapies.

Since publication of its first guideline in 1994¹ on the prevention of coronary heart disease, the ESC has continuously refined its processes related to guidelines, including the creation of a dedicated Clinical Practice Guidelines (CPG) Office in 2001, publication of a strict Declaration of Interest policy, and, more recently, the inclusion of patient representatives.

Of note, ESC Guidelines are typically developed by dedicated task forces with all recommendations required to obtain at least 75% of task force member votes. Documents undergo an intensive review process (on average 3 rounds generating an average of more than 3000 reviewer comments) involving experts and representatives from ESC National Cardiac Societies. As the systematic evaluation of existing evidence is the very basis for any new guideline recommendation, the ESC has recognized the importance of involving clinical methodology experts who are included as representatives in the Clinical Practice Guidelines Committee and in every ESC Guideline Task Force, highlighting the ESC’s commitment to methodological rigour.

European Society of Cardiology class of recommendation and level of evidence grading system

European Society of Cardiology Guidelines have relied on a Class of Recommendation (COR) and Level of Evidence (LOE) grading system to classify the quality of evidence supporting specific recommendations since more than 20 years.³ Although this system has proved useful and is widely accepted by the cardiovascular community, it has intrinsic limitations as any grading system. Therefore, the ESC CPG Committee appointed a methodology task force in 2022 to revise criteria in the evaluation of trials for therapy and prevention (Task Force 1)⁴ and diagnostic methods and predictions models (Task Force 2),⁵ whose output is detailed in two manuscripts in this edition of the European Heart Journal.

Highlights of the proposed revisions

The 2024 revision of the LOE grading system for ESC Guideline recommendations represents a significant step forward for future guidelines due to numerous reasons:

1. It recognizes differences in study design and methodology required to establish evidence for therapy and prevention when compared with evidence for diagnosis tests or prediction models, with the latter not necessarily requiring randomized controlled trials (RCTs).

2. It provides specific criteria to evaluate the quality of RCTs, meta-analyses, and non-randomized studies.

3. It provides specific methodological guidance on the assessment of heterogeneity, power, and sources of relevant bias in RCTs including small study effects.

4. It specifies measures to assess the strength of evidence against the play of chance.

5. It clearly distinguishes between evidence derived from randomized studies when compared with observational studies.

6. It recommends that only RCTs should be included in meta-analyses, which inform ESC Guideline recommendations.

7. It proposes a comprehensive assessment of evidence that includes the absolute magnitude of observed benefits, including the number needed to treat (NNT), the type of outcome, and the balance between benefit and risk of an intervention.

8. While LOE reflects the level of confidence to declare an intervention as truly superior (or non-inferior), COR classifies recommendations according to their clinical relevance.

Strengths of the revised grading system

In this context, the revised ESC LOE grading system introduces several noteworthy changes designed to enhance the methodological underpinning of ESC Guideline recommendations for therapy and prevention as outlined by the first task force manuscript in this issue of the journal:

• Enhanced criteria for LOE A applying to conclusive evidence—LOE A now requires conclusive evidence from ≥2 adequately powered RCTs free from major bias or meta-analyses with stringent statistical thresholds (e.g. P-value <.005 for superiority for frequentist statistics, 99.75% for Bayesian posterior probability, and an upper limit of the two-sided 95% confidence interval that is <1.15 for non-inferiority trials). This ensures a robust foundation for high-grade recommendations.

• Sub-division of LOE B—The original three-tiered system (A, B, and C) has been expanded to include LOE B1 (suggestive but not conclusive evidence) and LOE B2 (limited evidence). This distinction highlights differences in the quality and robustness of evidence, with LOE B1 recommendations requiring at least one adequately powered RCT free of major bias or meta-analysis with sufficient evidence against the play of chance (e.g. P-value <.05 for superiority for frequentist statistics), while LOE B2 encompasses large high-quality non-randomized studies meeting strict design criteria or meta-analyses of small RCTs.

• Refinement of LOE C—LOE C now applies to preliminary evidence from non-randomized studies without adequate control of major sources of bias, single under-powered RCTs, or expert consensus.

• Acknowledgment of exceptions—The revised system accommodates rare scenarios where a single, large RCT or clinically obvious interventions (e.g. defibrillation for ventricular fibrillation) may suffice for LOE A classification. These exceptions emphasize clinical practicality while maintaining methodological rigour.

The updated LOE grading system for ESC recommendations in the domain of therapy and prevention offers several advantages that enhance the robustness of ESC Guideline recommendations. Clarity and precision of future guideline recommendations are enhanced due to the stringent methodological criteria required for LOE A and the sub-division of LOE B providing clearer distinctions between varying levels of evidence quality, which will improve interpretation by clinicians and patients. In addition, the emphasis on methodological rigour will further enhance the interpretation of existing evidence relying on adequately powered RCTs and robust meta-analyses for high-grade recommendations. At the same time, the revision offers room for flexibility allowing for exceptions to enable practical clinical decision-making in scenarios where traditional evidence hierarchies may not apply.

The second task force manuscript in this issue of the journal introduces a new grading system in order to improve ESC Guideline recommendations in the domain of diagnostic tests and prediction models. The document applies the ESC LOE A, B, and C grading systems to classify evidence as conclusive, suggestive, or preliminary, respectively. Moreover, the assessment of data quality and adequate diagnostic or predictive ability as assessed by published methodological guidelines (TRIPOD, PROBAST, and QUADAS) is emphasized.

Implications of the revised grading system

In order to be successfully implemented, the updated grading system will require careful review by future task forces, and ESC CPG Committee will provide comprehensive task force training on the revised criteria ensuring consistent application across task forces. In addition, healthcare providers, researchers, patients, and readers of ESC Guidelines will need to adapt to the revised grading system.

Moreover, future task forces will continue to be faced with difficult issues related to the extrapolation of data from single drug or device interventions to class effects, recommendations related to rare diseases with low levels of evidence, and innovative interventions early in development as opposed to mature therapies and preventive measures.

Of note, some recommendations may experience downgrading in guideline updates not only in view of new evidence but also related to the more stringent grading criteria, which will require careful attention and communication by task forces. Future work by ESC CPG Committee may also consider the need for periodic reassessment of the new LOE grading systems and the importance of patient-reported outcomes and quality-of-life measures in the evidence evaluation process.

The 2024 revision of the ESC LOE grading system represents a commendable effort of the ESC CPG Committee and all authors to refine evidence evaluation in ESC Guidelines. By emphasizing methodological rigour and clinical relevance, the revised framework undoubtedly sets a new standard for future ESC Guideline recommendations with important implications for clinical practice and therapy implementation. All of us should look forward to see the revised LOE grading system being operational in the 2026 edition of ESC Guidelines.

Declarations

Disclosure of Interest

S.W. reports research, travel and/or educational grants to the institution from Abbott, Abiomed, Alnylam, Amicus Therapeutics, Amgen, Astra Zeneca, Bayer, B. Braun, Bioanalytica, Biotronik, Boehringer Ingelheim, Boston Scientific, Bristol Myers Squibb, Cordis Medical, CorFlow Therapeutics, CSL Behring, Daiichi Sankyo, Edwards Lifesciences, Fumedica, GE Healthcare, Guerbet, IACULIS, Inari Medical, Janssen AI, Johnson & Johnson, Medalliance, Medtronic, MSD Merck Sharp & Dohme, Neovii Pharmaceutica, Neutromedics AG, Novartis, Novo Nordisk, OM Pharma, Optimapharm, Orchestra BioMed, Pfizer, Philips AG, Sanofi-Aventis, Servier, Shockwave Medical, Siemens Healthcare, Sinomed, SMT Sahajan and Medical Technologies, Vascular Medical, V-Wave. S.W. serves as advisory board member and/or member of the steering/executive group of trials funded by Abbott, Amgen, Abiomed, Edwards Lifesciences, EnCarda Inc., Medtronic, Novartis, Sinomed with payments to the institution but no personal payments. He is also member of the steering/executive committee group of several investigator-initiated trials that receive funding by industry without impact on his personal remuneration.