Continuous glucose monitoring trajectories in patients with acute coronary syndrome

Continuous glucose monitoring (CGM) captures dysglycaemia and glycaemic variability after acute coronary syndrome (ACS), but patient-level trajectories from early recovery to mid-term follow-up—particularly in people without diabetes—remain insufficiently characterized. In this prospective multicenter observational study (ORACLE program), consecutive high-risk ACS patients wore a FreeStyle Libre 3/3 Plus sensor for 14–15 days near discharge and again at ~ 4 months. We quantified CGM time-in-range metrics (70–180 mg/dL and tight range 70–140 mg/dL), time above/below range, and variability/risk indices, including within-day profiles. Clinically relevant changes were categorized using pre-specified thresholds, and predictors of worsening were explored using multivariable models. Among all patients, 213 patients had analyzable baseline CGM recordings meeting quality criteria. In the early post-ACS period, median IQR time in range 70–180 mg/dL was 96.57% 86.91–98.92, time in tight range 70–140 mg/dL was 85.00% 62.16–93.24, time above range > 180 mg/dL was 1.12% 0.18–9.82, and time below range < 70 mg/dL was 0.43% 0.00–1.84, with a mean glucose of 114.26 mg/dL 106.26–133.96 and a median glucose management indicator of 6.04% 5.85–6.51. CGM demonstrated marked inter-individual heterogeneity and a reproducible late-morning (10:00–12:00) vulnerability window with lower range time and higher hyperglycaemic exposure, consistent across diabetes status and similar on weekdays and weekends; adverse CGM profiles were more prominent in patients with diabetes, older individuals, and women. Although CGM parameters improved modestly during the initial monitoring period, glycaemic control showed a slight but consistent deterioration from baseline to ~ 4 months after ACS, including in patients without diabetes. Tight-range time decreased by 4.5% (p = 0.008) and mean glucose increased by 4.67 mg/dL (p = 0.03), accompanied by a parallel worsening of variability and glycaemic risk indices. In contrast, HbA1c remained stable over follow-up. Across CGM endpoints, ~ 20–40% of patients showed a worsening trajectory (20.2% by broad time-in-range thresholds); higher comorbidity burden clustered with deterioration, with hypertension and COPD independently associated with tight-range worsening. After ACS, CGM reveals substantial inter-individual heterogeneity and a reproducible late-morning vulnerability window. From discharge to mid-term follow-up, deterioration—also affecting patients without diabetes—may be preferentially detected by tight-range and variability/risk metrics that traditional monitoring of blood glucose and static measures such as HbA1c may overlook, supporting CGM-informed phenotyping to refine post-ACS metabolic surveillance.