A critical analysis of the scientific evidence behind international guidelines related to cardiac arrhythmias

BACKGROUND: Guidelines have become very important in assisting with decision making in clinical practice. However, few studies have analyzed the level of evidence (LOE) underlying guidelines critically. This study aims to assess the accuracy of the referenced literature that has led to recommendations with a level of evidence A (LOE-A) rating. METHODS AND RESULTS: The latest updates of the practice guidelines related to arrhythmia posted on the European Society of Cardiology (ESC) web site were analyzed. The referenced literature for LOE-A recommendation was analyzed to reassess the proposed grading scheme for LOE-A. Furthermore, the clearly defined positive (Class I) and negative (Class III) recommendations with correctly referenced LOE-A were assessed. A median of 5.4% of all recommendations per guideline (interquartile range 4.9% to 9.7%) were categorized as LOE-A, but only 3.7% (IQR 3.4% to 4.9%) were accurately referenced as LOE-A. In total, 27 of 698 recommendations (median 1.2% per guideline [IQR 0.95% to 3.7%]) were correctly referenced as Class I or III LOE-A recommendations implying definite evidence-based positive or negative conclusion. CONCLUSIONS: Our findings raise the question of the accuracy of LOE-A in medical guidelines in general and highlight the importance of a critical use of all recommendations. Moreover, they underline the need for improving the guideline-writing process. Further randomized double-blinded and/or crossover-designed studies should focus on areas with a gap in the evidence, such as existing but not yet convincing (LOE-B) or conflicting (Class II) evidence