Gamma-Glutamyl Transferase as a prognostic risk factor in complex coronary disease: a serendipitous (re)discovery of machine learning in the SYNTAXES?

Abstract Background/Introduction Over the last decade, epidemiology and pathology studies have suggested that gamma-glutamyl transferase (GGT) has an independent role in the pathogenesis and clinical evolution of cardiovascular diseases associated with atherosclerosis. Recently, machine learning (ML) algorithms have identified pre-procedural GGT as a significant predictor of long-term mortality after coronary revascularization in the SYNTAX trial. Purpose The aim of this study is to investigate the impact of pre-procedural GGT on 10-year all-cause mortality in patients with three-vessel disease (3VD) and/or left main coronary artery disease (LMCAD) following revascularization. Methods The SYNTAX trial was a randomized trial comparing percutaneous coronary intervention (PCI) with coronary artery bypass grafting (CABG) in 1,800 patients with complex CAD. The present report is a post-hoc sub-analysis of the SYNTAXES study, an investigator-driven extended 10-year follow-up of the SYNTAX trial. Before revascularization, blood samples were taken for GGT as well as other biological markers and analyzed by an independent central chemistry laboratory. The primary endpoint was all-cause mortality at 10-years. The changes in hazard for all-cause mortality in GGT over time were modelled using restricted cubic spline functions in a Cox regression model adjusted by age and sex. Subsequently, we compared survival in four groups defined by gender specific GGT thresholds of <32 IU and ≥32 IU for females and <64 IU and ≥64 IU for males, in accordance with prior studies. Results Out of 1,800 patients enrolled in the SYNTAX study, 167 (9.3%) did not have a GGT available at baseline and 1,633 (90.7%) patients were included in the present study. The mean values of GGT for males and females were 43.5±48.5 and 36.4±46.1 U/L, respectively. In the overall population a 10 U/L increase in GGT was significantly associated with 10-year all-cause death (adjusted hazard ratio [HR] 95% confidence interval [CI]: 1.03 [1.02–1.04], p <0.001, Figure A). For Kaplan–Meier analysis, patients were stratified into four groups according to previously reported sex related GGT thresholds. Ten-year mortality was numerically higher in males (adjusted HR 1.31, 95% CI: 0.90–1.91), and significantly higher in females (adjusted HR 1.88, 95% CI: 1.16–3.05) among patients with higher versus lower GGT (Figure B). Furthermore, GGT was significantly associated with CRP, a surrogate-marker of inflammation, as well as established cardiovascular risk biomarkers such as TG, total cholesterol, and glucose. Conclusions Preprocedural GGT is an independent predictor of 10-year mortality following coronary revascularization in patients with 3VD and/or LMCAD. This long-term association has been identified using ML algorithms and should be corroborated using data from large biobanks.Figure AFigure B