Effect of glycemic state on hospital mortality in critically ill surgical patients.

Intensive insulin therapy can reduce mortality. Hypoglycemia related to intensive therapy may worsen outcomes. This study compared risk adjusted mortality for different glycemic states. A retrospective review of patients admitted to a surgical intensive care unit over 4 years was performed. Patients were divided into glycemic groups: HYPER (≥1 episode > 180 mg/dL, any <60), HYPO (≥1 episode < 60 mg/dL, any >180), BOTH (≥1 episode < 60 and ≥1 episode > 180 mg/dL), NORMO (all episodes 60-180 mg/dL), HYPER-Only (≥1 episode > 180, none <60 mg/dL), and HYPO-Only (≥1 episode < 60, none >180 mg/dL). Observed to expected Acute Physiology and Chronic Health Evaluation (APACHE) III mortality ratios (O/E) were studied. Number of adverse glycemic events was compared with mortality. Hypoglycemia and hyperglycemia occurred in 18 per cent and 50 per cent of patients. Mortality was 12.4 per cent (O/E = 0.88). BOTH had the highest O/E ratio (1.43) with HYPO the second highest (1.30). Groups excluding hypoglycemia (NORMO and HYPER-only) had the lowest O/E ratios: 0.56 and 0.88. Increasing number of hypoglycemic events were associated with increasing O/E ratio: 0.69 O/E for no events, 1.19 for 1-3 events, 1.35 for 4-6 events, 1.9 for 7-9 events, and 3.13 for ≥ 10 events. Ten or more hyperglycemic events were needed to significantly associate with worse mortality (O/E 1.53). Hyper- and hypoglycemia increase mortality compared with APACHE III expected mortality, with highest mortality risk if both are present. Hypoglycemia is associated with worse risk. Glucose control may need to be loosened to prevent hypoglycemia and reduce glucose variability.