OBJECTIVETo investigate the effect of hyperglycemia within postoperative 48 hours on gastrointestinal (GI) fistula patients without preoperative diagnosis of diabetes undergoing selective GI reconstruction.

METHODSClinical data of GI fistula patients with age of 18 to 70 years and without diffuse peritonitis and systemic infection undergoing definitive GI reconstruction at Intestinal Fistula Center of Jinling Hospital from September 2012 to December 2015 were retrospectively analyzed. According to the highest blood glucose (BG) value detected within postoperative 48 hours, patients were divided into normoglycemia (BG<6.9 mmol/L), mild hyperglycemia (6.9 to 11.4 mmol/L) and severe hyperglycemia (BG≥11.5 mmol/L) groups. Clinical manifestations were compared among three groups. Effects of postoperative hyperglycemia on associated parameters, including postoperative surgical site infection, anastomotic leakage, intestinal nutrition recovery, hospital stay and hospitalization cost were investigated.

RESULTSA total of 314 patients were enrolled, of whom postoperative gastric fistula occurred in 6 cases, small intestinal fistula in 95 cases, ileocolonic anastomotic fistula in 116, and colorectal fistula in 97 cases. One hundred and ninety-three (61.5%) patients experienced hyperglycemia, including 148 cases of mild hyperglycemia group and 45 cases of severe hyperglycemia group, the other 121 cases were of normoglycemia group. There were no significant differences in gender, BMI, ratio of smoking, ratio of alcohol user and primary diseases among 3 groups (all P>0.05). Older patients were vulnerable to postoperative hyperglycemia and patients who developed hyperglycemia were also prone to have increased ASA score (all P=0.000). Hyperglycemia patients had significantly higher ratio of postoperative ileocolonic anastomotic fistula (mild hyperglycemia group: 40.5%, 60/148; severe hyperglycemia group: 44.4%, 20/45) than normoglycemia cases (29.8%,36/121). Compared to normoglycemia group, ratio of intra-operative transfusion case was higher (P=0.001), operative time was longer (P=0.026), ratio of number of anastomosis >2 was higher (P=0.001), ratio of receiving laparoscopic-assisted operation was lower (P=0.005), ratio of postoperative surgical site infection was higher (P=0.006), incidence of anastomotic leakage was higher (P=0.004), ratio of re-operation was higher (P=0.004), intestinal nutrition recovery time was longer (P=0.001), ICU stay was longer (P=0.001), total hospitalization time was longer (P=0.000) and hospitalization cost was more expensive (P=0.000) in both two hyperglycemia groups. Multivariate regression analysis showed that mild hyperglycemia and severe hyperglycemia were independent risk factors to predict surgical site infection (OR=1.99, 95%CI: 1.12 to 3.54, P=0.019; OR=3.02, 95%CI: 1.36 to 6.70, P=0.007) and anastomotic leakage (OR=7.59, 95%CI: 1.68 to 34.34, P=0.009; OR=13.4, 95%CI: 2.50 to 71.65, P=0.002). Multivariate linear regression analysis indicated that intestinal recovery time of normoglycemia group was 2 days shorter and 3 days shorter, and hospitalization time of normoglycemia group was 2 days shorter and 10 days shorter as compared with mild hyperglycemia and severe hyperglycemia group, respectively.

CONCLUSIONSElevated postoperative BG is common in GI fistula patients receiving selective GI reconstruction. Postoperative hyperglycemia is significantly associated with surgical site infection, anastomotic leakage and prolonged intestinal recovery. BG control treatment should be recommended for those patients with postoperative hyperglycemia.