Elevated glucose levels at 24 hours predict mortality in cardiogenic shock

Hyperglycemia is associated with poor outcomes in critically ill patients. However, its role as a predictor of outcomes in cardiogenic shock (CS) remains unclear due to conflicting evidence in the literature.

Our study aimed to assess whether blood glucose levels 24 hours after Cardiac Intensive Care Unit (CICU) admission could predict 30-day all-cause mortality in patients with CS.

All patients enrolled in the multicentre prospective AltShock-2 registry between March 2020 and November 2023 with recorded blood glucose levels 24 hours post-CICU admission were included. The relationship between 24-hour blood glucose levels and 30-day all-cause mortality was analyzed. Optimal 24-hour blood glucose at cut-off values for outcome prediction were identified across the cohort. Patients were categorized as follows: Group A (BGL < 140 mg/dL), Group B (BGL 140–210 mg/dL), and Group C (BGL > 210 mg/dL).

In total, 408 patients with CS (mean age 64 ± 15 years, 76% males) were included. At 24 hours post-CICU admission, blood glucose levels were < 140 in 211 patients (52%), 140-210 in 153 (37%), and > 210 mg/dl in 44 (11%). A previous diagnosis of diabetes mellitus (DM) was more common in groups B and C (p<0.01). Elevated 24-hour blood glucose was independently associated with increased 30-day all cause mortality (p=0,04). Patients with blood glucose >210 mg/dL had significantly higher 30-day mortality (aOR 3.2, 95% CI 1.2–8.9, p=0.02) compared to lower glucose groups. Intriguing, at multivariable logistic regression analysis, adjusting for DM status, the effect remained significant (aOR 2.94, 95% CI 1.07–8.02, P = 0.03). The 24-hour glucose measurement showed higher predictive accuracy for mortality than other time points (AUC 0.6 vs. 0.5). The optimal glucose threshold for mortality prediction was 155 mg/dl (aOR 2.0, IC 1.1-4.0, p=0.03). Notably, mechanical circulatory support use in this cohort was protective (aOR 0.44, 95% CI 0.2-0.9, p=0,04), while etiology of CS had no impact on outcomes.