One hundred and eighty-seven consecutive patients were identified for study inclusion. Four-thousand and eighteen BG values were collected and the mean number of values for each patient was 21.5 (SD: 28.8, median: 12). On average each patient had at least 2 BG values obtained daily.
Population demographics and injury characteristics are described in . Orthopaedic injuries included 62 upper extremity injuries, 36 pelvic or acetabular fractures, 86 femur fractures, 56 tibia fractures, and 20 foot injuries. The most frequent co-morbidity was hypertension (56/187, 30.0%); and the mean number of co-morbidities per patient was 0.7 (SD: 0.9, median: 0). Average BG at hospital admission was 140.5 ± 45.0 mg/dL and 33/187 patients (17.6%) had 2 or more BG values ≥200 mg/dL during the entire hospital stay. Mean HGI was 1.2±1.2 and 41/187 patients (22.0%) had an HGI>1.72.
Baseline Demographics and Injury Characteristics
Of the entire study population, there were 51 peri-operative infectious complications in 43 patients, consisting of 22 pneumonia, 7 UTI, 6 sepsis, and 16 surgical-site infections, for an overall infection rate of 23.0%. Five patients with a surgical-site infection had wound cultures which were positive for methicillin-resistant Staphylococcus aureus. Patients with an infectious complication were older (55.4, SD: 21.8 years vs. 45.4, SD: 23.1 years; p=0.006) and received greater amounts of blood (9.1, SD: 9.4 units vs. 4.7, SD: 8.0 units; p=0.0001). Mean ISS was greater, but not clinically significant, in patients with an infection (10.6, SD: 3.7 vs. 9.5, SD: 2.8; p=0.03). There were no differences in admission BG (p=0.35), 2 BG values ≥ 200 mg/dL (p=0.10), or HGI (p=0.08) and the presence of an infection. However, when surgical-site infections were considered alone there was a significant difference in HGI (2.1, SD: 1.7 vs. 1.2, SD: 1.1; p=0.006). There was no difference in admission BG (p=0.65) or 2 BG ≥200 mg/dL (p=0.21) and surgical-site infection. When evaluated separately from surgical-site infections, there was no significant difference between HGI and other infectious complications (1.2±0.9 vs. 1.3±1.2, p=0.67).
The overall rate of surgical-site infection was 16/187 (8.6%). No patient with a surgical-site infection had a prior infection (i.e. pneumonia, UTI, or bacteremia). There were no differences in age, race, history of tobacco use, or ISS and the presence of a surgical-site infection (). Male patients were significantly more likely to have a surgical-site infection; however, male gender was not associated with HGI (1.2±1.3 vs. 1.3±1.4, p=0.61). Eighty-three patients had open fractures: 9 Type I, 13 Type II, 37 Type IIIA, 9 Type IIIB, and 18 Type IIIC. Open fractures were not more likely to develop a surgical-site infection (8/83, 9.6% vs. 8/104, 7.7%; p=0.64) and all infections in open fractures occurred in Type III fractures: 3/34, 8.8% Type IIIA; 1/9, 11.1% Type IIIB; and 4/18, 22.2% Type IIIC). Of the 41 patients with an HGI >1.72, eight patients (19.5%) developed a surgical-site infection (p=0.005). The unadjusted risk for a surgical-site infection in patients with an HGI >1.72 was 4 times greater than in patients with an HGI ≤1.72 (OR: 4.2, 95% CI: 1.5–12.0). Patients with an HGI >1.72 were significantly older (57.8, SD: 22.3 years vs. 44.8, SD: 22.7 years; p=0.008), however, there were no differences in sex, number of co-morbidities, ISS, or hospital and ICU LOS ().
Variables by Surgical-Site Infection
Demographic and Injury Variables by Hyperglycemic Index >1.72
A multivariable logistic regression model was constructed () adjusting for potential confounding variables, to test the association between HGI and surgical-site infection. Variables significant by univariate analysis were considered for potential inclusion in the multivariable model. Due to the rather limited number of surgical-site infections (n=16), clinically and statistically significant continuous variables (HGI and total units of blood transfused) were inserted in the model. After adjusting for units of blood transfusion (OR: 1.1, 95% CI: 1.0–1.2), HGI remained a significant independent risk factor for surgical-site infections (OR: 1.8, 95% CI: 1.3–2.5).
Multivariable Logistic Regression for Risk Factors of Surgical-Site Infection