To analyze the risk factors for early stress hyperglycemia in patients with severe burn and the relationship between the risk factors and the prognosis.

Seventy-eight patients with severe burn who were admitted to Department of Orthopedics, Yan′an People′s Hospital during May 2014 to March 2017 were selected as the subjects. The mean amplitude of glycemic excursions(MAGE) was calculated according to the blood glucose in 72 h after admission. The patients were divided into low value group (MAGE<3.3 mmol/L, n=18), median value group (3.3 mmol/L≤MAGE<5.5 mmol/L, n=37) and high value group (MAGE≥5.5 mmol/L, n=23) according to the calculated value. The general data, such as gender, age, weight, causes of burn, Glasgow coma scale (GCS), respiratory tract burn, total area of burn and acute physiology and chronic health evaluation (APACHE) Ⅱscore were compered. Biochemical indexes, such as procalcitonin, C-reactive protein (CRP) and arterial partial pressure of CO₂(PaCO₂) of the three groups were compared, too. The number of death cases and the incidence of complications were statistically analyzed. The enumeration data in multi-group were analyzed by chi-square test, while the measurement data in multi-group were analyzed by analysis of variance. The LSD method were used in the comparison of two groups. The inducing factors of stress hyperglycemia were analyzed by multivariate logistic aggression analysis, and the survival curve was analyzed by Log-rank test to judge the relationship between MAGE and the prognosis.

The PCT, CRP and the total area of burn in the high value group [(9.2±1.8) ng/mL, (58.3±10.6) mg/L, (76.7±15.3) % total body surface area(TBSA)]were significantly higher than those in the median value group [(6.7±1.3) ng/mL, (42.5±8.7) mg/L, (65.4±13.5)% TBSA] and low value group [(5.3±1.2) ng/mL, (37.4±7.9) mg/L, (62.7±12.4)% TBSA], the differences were statistically significant (t_{median value group}=6.238, 6.286, 2.995, t _{low value group}=7.909, 6.978, 3.153; with P values below 0.05). The APACHEⅡ score and GCS of the high value group [(22.1±4.3) and (13.5±2.7) points] were significantly higher than those of the low value group [(17.5±3.4) and (9.4±2.1) points], the differences were statistically significant (t=3.716, 5.304; with P values below 0.05). CRP, PCT and the total area of burn were independent risk factors of early stress hyperglycemia (with P values below 0.05). The 90 d survival rates of the low value group, the median value group and the high value group were 94.4% (17/18), 86.5% (32/37) and 69.6% (16/23), respectively. The area under receiver operating characteristic(ROC) curve of MAGE in predicting the prognosis of patients was 0.825, and the optimal threshold was 4.8 mmol/L. The sensitivity and specificity in predicting death were 86.8% and 87.9%, respectively. The incidence rates of infection and acute stress ulcer in the high value group (82.6%, 78.3%) were significantly higher than those in the low value group (22.2%, 33.3%) and the median value group (51.4%, 45.9%) ( χ _infection²=14.951, 5.967, χ _{acute stress ulcer}²=8.398, 6.064; with P values below 0.017).

The level of stress hyperglycemia in patients with severe burn is related to inflammatory factors and the area of burn. The higher the early stress glucose is, the worse the prognosis is. Early stress hyperglycemia can be used as an indicator for evaluating the prognosis.