Utility of screening for diabetes in a burn center: hemoglobin A1c, Diabetes Risk Test, or simple history?

OBJECTIVE

Rates of diabetes mellitus (DM) are increasing. Early identification and treatment of hyperglycemia in the critical care setting can decrease morbidity and mortality. Many burn centers measure hemoglobin A1c (A1c). This study evaluates the prevalence of pre-existing DM and the utility of using A1c for identifying DM compared with a non-invasive risk assessment.

METHODS

Adult patients admitted to our regional ABA-verified burn center from July 2008 to July 2009 had A1c levels evaluated and were asked to complete the American Diabetes Association Diabetes Risk Test (DRT).

RESULTS

Forty-one patients consented to participate: 24 patients with burn (19 male) and 17 patients with non-burns (10 male). Non-burn patients had greater BMIs (BMI 32 vs. 28, p=0.093) and had a higher rate of DM prior to admission (35% vs. 17%, p=0.159) than the burn patients. These differences were not statistically significant. Most patients (23/41) were at high risk for developing DM based on the DRT. Patients with pre-existing DM were significantly more likely to have elevated A1c levels (>6.5%) compared with those without pre-existing DM (60% vs. 0%, p<0.001). Compared with history of DM, DRT had a poor positive predictive value of 36% and 50% (burn and non-burn respectively) but a 100% negative predictive value for DM for both groups.

CONCLUSION

DM and obesity were more common in non-burn patients than in burn patients. A history of DM provides a simple, accurate method for identifying patients with DM. Use of A1c in the ICU provides little additional data for diagnosis of DM and does not impact patient management.