Change in admission blood glucose from chronic glycemic status in acute heart failure hospitalization and 30-day outcomes: A retrospective analysis.

Background: Admission blood glucose (BG) has demonstrated contradictory association with 30-day mortality in acute heart failure (AHF) hospitalization. To explore these contradictory findings, we aimed to determine if admission BG reflects an acute change from chronic glucose control and investigate the association between the admission and chronic BG change (ΔBG) with 30-day mortality in AHF.

Methods: We analyzed patients (n = 1045) age ≥ 65 with Centers of Medicare Services benefits and known 30-day all-cause mortality hospitalized with AHF at an academic medical center from 2009 to 2016. We included diabetic (n = 736) and non-diabetic (n = 309) patients with recent Hemoglobin A1c (HbA). We defined ΔBG as the difference in the admission and chronic BG, calculated from HbA.

Results: Admission BG was 126 (101, 167) mg/dl and was moderately elevated (≥170 mg/dl) in 25% of admissions. The median (IQR) ΔBG was -7 (-29, 26) mg/dl, with 74% of all admissions (66% diabetic, 92% non-diabetic) presenting within ±50 mg/dl of the chronic BG. Admission BG was not associated with mortality. ΔBG > 100 mg/dl displayed increased 30-day mortality (18.6% vs 6.9%, p < 0.001) compared to 26 to 25 mg/dl. When admission BG was >200 mg/dl (n = 166), a ΔBG > 100 mg/dl was present in 77% of those deceased vs 31% of those alive at 30 days (p = 0.003; positive likelihood ratio = 5.7). ΔBG > 100 mg/dl was strongly associated with 30-day mortality risk (OR 6.4, 95% CI 2.3-18.9; p = 0.0005) after multivariate adjustment.

Conclusions: Admission BG predominantly reflects chronic glycemic status. Increased change in admission from chronic BG was associated with increased 30-day mortality.