- Poster presentation
- Open Access
Hemoglobin A1c predicts outcome of sepsis in patients with diabetes
© BioMed Central Ltd 2006
- Published: 21 March 2006
- Diabetic Patient
- Hospital Mortality
- Glucose Control
- Multiple Regression Model
- Hospital Length
Glycated hemoglobin (HbA1c) is a widely used marker of long-term glycoregulation in diabetes since it reflects levels of glucose in the 120 days prior to measurement. A positive impact of strict glucose control on the outcome of diabetic patients with sepsis has been well established, but the influence of glycoregulation prior to the occurrence of sepsis has not been investigated in relation to outcome. Nonenzymatic glycation and formation of advanced glycation end products (AGEs) are associated with chronic hyperglycemia in diabetes and have many biochemical and cellular effects, including changes in immune and inflammatory responses. Our hypothesis was that such changes could influence the outcome of sepsis.
A prospective, single-center, observational study included adult patients admitted with sepsis and a history of diabetes during a 1-year period. HbA1c was measured on the first hospital day and the result did not influence treatment. APACHE II and SOFA scores were calculated for all patients at admission. Complications of diabetes (renal, vascular and ocular) were noted. WBC, CRP and blood culture results were recorded. The hospital mortality and hospital length of stay (LOS) were used as measures of outcome. Nonparametric tests, multiple regression and logistic regression were used in statistical analyses.
The study included 286 patients admitted to medical wards or the medical ICU. Hospital mortality for all patients with sepsis was 21.7% with a median LOS of 6 days (95% CI 6–7 days). Patients who survived had significantly (P < 0.001) lower HbA1c (median 8.2%; 95% CI 7.8–8.6%) than patients who died (median 9.75%; 95% CI 8.7–10.6%). A correlation was found between the HbA1c values and patient's LOS (r = 0.289, P < 0.001). In a logistic regression model HbA1c was found to be related to lethal outcome (OR 1.36; P < 0.001), together with APACHE II score (OR 1.08; P = 0.014) and SOFA score (OR 1.27; P < 0.001) and female sex (OR 2.223; P = 0.032). In a multiple regression model HbA1c was found to relate to LOS (P < 0.001) together with APACHE II (P = 0.015) and SOFA (P < 0.001) scores, age (P = 0.032), female sex (P = 0.033) and WBC (P = 0.048).
HbA1c was shown to be predictive of mortality and hospital LOS of patients with sepsis and a history of diabetes. Chronic hyperglycemia and consequent increased AGE formation may influence inflammatory and immune responses, and thus be responsible. Proper glycoregulation in diabetic patients could reduce the risks in the event of infection.