033 Reproducibilty of the syntax score in a real world setting: implications for the choice of contemporary revascularisation strategy

Abstract

<h3>Introduction</h3> The 2 year results from the Syntax (The Synergy between Percutaneous Coronary Intervention with TAXUS and Cardiac Surgery) study have suggested that percutaneous coronary intervention may be an acceptable alternative to CABG in subjects with a low (&lt;22) and intermediate (23–32) Syntax Score (SS). Patients with a high SS (≥33), however, appear to have a worse outcome with PCI and it has been suggested that surgery should remain the gold standard in this cohort. The Syntax investigators report high reproducibility scores for this tool; however, it would be interesting to know if this consistency of scoring could be reproduced in a real world scenario. This is important, given that some centres are now routinely incorporating the SS into their cardiac MDT decision-making processes. It would also be valuable to know if operator experience significantly affects the reliability of this scoring system. <h3>Methods</h3> Using a cardiac database, 20 patients with multivessel coronary disease were selected at random for angiographic re-reporting using the SS. Eight middle grade cardiologists of differing levels of experience were asked to independently score these angiograms, after completion of the online teaching module at the http://www.syntaxscore.com Web site. Operators were selected to provide a range of middle grade experience and comprised two groups: four operators with &lt;1000 diagnostic angiograms performed; four operators with &gt;2000 diagnostic angiograms performed. <h3>Results</h3> The mean time to perform the Syntax score was 9.7±5.4 min and the variation between operators is shown (Abstract 33 Figure 1). The less experienced the operator the higher the Syntax score tended to be (&lt;1000: 34±21 vs &gt;2000: 28±10; p&gt;0.05). Within the group of experienced operators, the correlation coefficient for the SS was stronger (intra-observer correlation coefficient: 0.5 for &lt;1000 and 0.6 for &gt;2000), but there was still little agreement in terms of allocation to the tertiles described in the Syntax trial. In only 15% (3/20) of patients did all four operators score the patient into the same tertile. There was a majority agreement on tertile in a further 20% (4/20), but in the remaining 65% (13/20) no more than two of the four operators agreed on the same tertile of risk. <h3>Conclusion</h3> There is significant heterogeneity of scoring amongst relatively inexperienced angiographic operators who have been trained using the recommended online Syntax Scoring tutorial. This improved a little with operator experience; however, even experienced operators were unable to consistently score patients into the same tertiles of risk as described in the Syntax trial. Any departments considering using the Syntax score as an integral part of their clinical decision making processes should be aware of the wide variation in scores obtained in this real world sample of cardiac MDT diagnostic angiograms.