[The correlation analysis between glucose level and its variability and prognosis in traumatic patients].

Objective: To investigate the effect of glucose level, variability on the prognosis of traumatic patients.

Methods: A retrospective study involving 300 traumatic patients admitted to intensive care unit (ICU) was performed. The average glucose (GluAve), glucose standard deviation (GluSD) and glucose coefficient of variation (GluCV) during the first 72 hours were calculated. Patients were divided into survivor group (n=249) and non-survivor group (n=51) based on outcomes. The GluAve, GluSD and GluCV were compared between the two groups. Patients were allocated into five subgroups based on GluAve (3.9-5.5, 5.6-6.6, 6.7-7.7, 7.8-9.9, ≥10.0 mmol/L) as well as four subgroups on GluCV (<15%, 15%-30%, 30%-50%, >50%). The mortality in hospital was compared among the different subgroups and the different GluCV in the same level of GluAve subgroups, respectively. Multifactor logistic regression was used to determine the risk factor of hospital death.

Results: The levels of GluAve, GIuSD, GluCV of non-survivor group were higher than those of survivor group [11.31±4.38 mmol/L vs. 8.50±3.40 mmol/L, 2.85±1.94 mmol/L vs. 1.87±1.67 mmol/L, (28.30±23.08)% vs. (20.90±13.70)%, all P<0.05]. With the gradual increment of GluAve and GluCV level, the mortality was raised accordingly (χ (2)(1)=26.332, P=0.000; χ (2)(2)=65.522, P=0.000). In the subgroup of GluAve 7.8-9.9 mmol/L, the mortality was 9.09% (3/33) with GluCV <15% versus 46.15% (6/13) with GluCV >50% (P<0.01) respectively, and in the subgroup of GluAve ≥10.0 mmol/L, the mortality corresponding rates were 21.05% (4/19) with GluCV < 15% and 61.54% (8/13) with GluCV > 50% (P<0.05). The multivariable logistic regression analysis demonstrated that GluAve and GluCV were risk factors of mortality[GluAve odds ratio (OR)=1.150, 95% confidence interval (95%CI) was 1.042 to 1.270, P=0.006; GluCV OR=1.022, 95%CI was 0.999 to 1.040, P=0.040], GluSD had no effect on mortality.

Conclusions: The increase in GluAve and GluCV in traumatic patients are significantly correlated with mortality. Control the level and the variability of blood glucose might be an important aspect of the multiple trauma death reduction.