Background: Acute heart failure (AHF) in diabetic patients is strongly associated with increased mor-tality. However, the impact of acute glycemic dys-function on mortality in these patients is unclear. Purpose: To test the glycemic variability as an inde-pendent predictor of in-hospital mortality in AHF. Methods: We retrospectively identified AHF diabetic patients (2013-2015). Glucose variables of interest were glycated haemoglobin (HbA1c), admission glycaemia and mean glycemia index (MGI), defined as mean of all glycemic points during hospitalization. We evaluated also conventional prognostic markers: age, admission and mean systolic/diastolic blood pressure (SBP/DBP), admission and mean heart rate (HR/HRm), respiratory failure (SaO2), mechanical ventilation, ejection fracti-on (EF), atrial/ventricular arrhythmia, troponin I, NT-proBNP, admission renal (eGRF) and hepatic function (AST and ALT), and lactic acid. Outcome was in-hos-pital mortality.
Results: 414 patients were identified, 217 (52%) with diabetes (41-93 years). 36 diabetics (17%) died. Age (70±10 vs 69±11y), NTproBNP (6565±7393 vs 9065±9024 pg/ml), Na+ level, and HbA1c (7,45±1,4 vs 7,42±1,19%) were similar in non survivors vs. sur-vivors. SBP(153±39 vs 112±35 mmHg)/DBP (86±22 vs 70±21 mmHg), LVEF(33±12 vs 24±11%), SaO₂ (88±10 vs 83±16%), eGRF(51±21 vs 39±18 ml/min) were lower, whereas HRm(85±14 vs 97±25 bpm), TpI(1±2 vs 3±9 pg/ml), ALT(50±49 vs 89±131mg/dl), AST(51±75 vs 117±200 mg/dl), and lactic acid(3±1 vs 5±4 mmol/l) were higher in non-survivors (p<0,05 in all pairs). Admission glycemia(325±105 vs 248±108 mg/dl) and MGI (265±112 vs 169±40 mg/dl) were significantly higher in non-surivors (p<0.001) (Table). By logistic regression analysis, the best predictor model for in-hospital mortality included SBP, mechanical ventilati-on, EF, eGRF, admission glycaemia, and MGI (r=0.65, r2=0.4, P<0.001). MGI ≥188 mg/dl (AUC=0.78) was the best independent predictor of mortality.
Conclusion: Glycemic variability is independently associated with mortality in AHF diabetic patients. Mean glycemic index seems to be superior to other conventional parameters for prediction of mortality.