ABSTRACT Although dexamethasone reduces mortality in patients with COVID‐19, plasma glucose (PG) levels increase upon initiation. In a multicenter observational cohort of 530 adults, we estimated glycemic variability based on baseline HbA1c among patients with normoglycemia ( N = 238), prediabetes ( N = 159), unknown ( N = 63), and known diabetes ( N = 159). Glycemic variability, diabetic‐ and hyperglycemic events (≥ 11.1 and ≥ 16 mmol/L) were analyzed using a linear mixed model and competing risks analysis adjusted for confounders. Before dexamethasone, mean PG levels were similar in those with normoglycemia (6.5 mmol/L) and prediabetes (6.6 mmol/L), but higher in unknown (8.5 mmol/L) and known diabetes (9.9 mmol/L). After treatment, PG increased across all groups. Prediabetes showed a larger increase (1.5 mmol/L) than normoglycemia (0.7 mmol/L, p = 0.002), and known diabetes had the highest increase (2.4 mmol/L, p < 0.001), reaching an average of 12.6 mmol/L. All groups except prediabetes returned to baseline after dexamethasone. The cumulative incidence of diabetic events was 98% in known diabetes, 67% in unknown diabetes, 31% in prediabetes, and 8% in normoglycemia, with significant differences between groups ( p < 0.001). We conclude that dexamethasone treatment increased average PG and caused frequent hyperglycemic events in patients with prediabetes, unknown, and known diabetes, while persistent PG elevation post‐treatment occurred in prediabetes.
Clausen et al. (2026) studied this question.