30 Syntax score- a simple way to stratify complex coronary disease?. (30th September 2020)

• Record Type: Journal Article
• Authors: Mailey, J; Murphy, C
• Is Part Of: Heart. Volume 106(2020)Supplement 4
• Page Start: A20
• Page End: A21
• Publication Date: 2020-09-30
• Subjects: Heart -- Diseases -- Treatment -- Periodicals; Cardiology -- Periodicals; 616.12
• DOI: 10.1136/heartjnl-2020-ICS.30
• Languages: English
• ISSNs: 1355-6037

Abstract

Abstract : Background: The Syntax Score (SS) is an angiographic risk stratification tool to grade coronary disease complexity and aid revascularisation strategy decision making. Calculated scores are divided into tertiles (0–22, 23–32, >32) and guidelines advocate that decisions should be based on this tertile (1). Syntax Score calculation is dependent on the subjective interpretation of coronary angiograms by Interventional Cardiologists (ICs) and is therefore prone to variability. Our institution consists of 13 ICs performing on average >5000 coronary procedures per year. We sought to compare syntax scoring amongst ICs. Methods: 10 patients with multi-vessel coronary artery disease who had previously been discussed by a Multi-Disciplinary team (MDT) were retrospectively identified. Eight ICs were asked to interpret the coronary angiograms and calculate a SS for each patient. Scoring was performed independently, with no knowledge of the score that had been used at the MDT discussion. Interobserver variability between raters for the raw syntax score was assessed using intraclass correlation coefficient. Interobserver variability between raters with respect to tertile was assessed using Fleiss kappa. Results: All 8 ICs completed scores for each patient. Mean calculated SS was 30.2 (± 6.7 SD). 21.3% of scores were 1st tertile (0–22), 45.0% were 2nd tertile (23–32) and 33.7% were 3rd tertile (>33). Mean range of calculated SS between raters was 18.5 (range 9.5 - 29). 40.0% of scores were divided across all 3 tertiles with the remaining 60.0% divided across 2 tertiles. Intraclass correlation coefficient of raw SS was 0.55 suggesting only fair agreement. The interobserver strength of agreement with regard to tertiles was also only fair (Fleiss Kappa=0.25; 95% CI [0.07, 0.43]). The interobserver strength of agreement improved to moderate when comparing tertile 1 vs tertile 2 or 3 together (Fleiss Kappa=0.47; 95% CI [0.17, 0.77]). Only 63.1% of IC tertile scores agreed with that quoted at the initial MDT. This improved to 82.3% when comparing to tertile 1 vs tertile 2 or 3. Conclusion: Overall we demonstrated only fair to modest repeatability in SS calculation amongst individual ICs. This is similar to findings in other studies, though it is known improvements are possible with intensive training (2). We advise that significant variability in SS calculations should be considered within all institutions unless validated training systems are in place. In the absence of these systems we suggest the weight of importance attributed to quoted SS in MDT decision making should be guarded.