Objectives To assess the association between fasting plasma glucose (FPG) and all-cause mortality across the spectrum of coronary artery disease (CAD).
Methods There were 18999 patients in this study. The primary end points were in-hospital and follow-up all-cause mortality. According to the quartiles of FPG levels, patients were categorised into 4 groups: Q1, <5.1 mmol/L (to convert to mg/dL, multiply by 18); Q2, 5.1 - 5.9 mmol/L; Q3, 5.9 - 7.5 mmol/L; and Q4, ≥7.5 mmol/L. Q1 was recognised as the lower glycemic group, Q2/Q3 as the normoglycemic groups, and Q4 as the higher glycemic group.
Results In patients with acute myocardial infarction (AMI), all-cause mortality for the dysglycemic groups were higher than normoglycemic groups (in-hospital mortality for Q1 to Q4 was 1.0%, 0.9%, 0.2%, and 1.5%, respectively, P value = 0.001; follow-up mortality for Q1 to Q4 was 1.7%, 0.9%, 0.3%, and 1.8%, respectively, P value <0.001). In patients with stable coronary artery disease (SCAD), there were no significant differences in mortality among groups. While in patients with unstable angina pectoris (UAP), the normoglycemic groups were associated with lower follow-up mortality and roughly equal in-hospital mortality compared with the dysglycemic groups. After adjusting for confounding factors, this observation persisted.
Conclusions The association between lower glycemia and mortality differed across the spectrum of CAD. In patients with AMI, there was a U-shaped relationship. In patients with SCAD or UAP, mildly to moderately decreasing FPG was associated with neither higher nor lower all-cause mortality.