Investigating Glycemic Control in Patients Undergoing Lower Extremity Bypass within an Enhanced Recovery Pathway at a Single Institution
Charles Adam Banks1, Adam W Beck1, Marvi Tariq1, Zdenek Novak1, Benjamin J Pearce1, Mark A Patterson1, Danielle C Sutzko1, Miles Morgan2, Marc Passman1, Emily L Spangler1
1University of Alabama at Birmingham, Birmingham, AL;2Medical College of Georgia, Augusta, GA
Introduction: Enhanced recovery pathways (ERP) have been utilized with success in several surgical subspecialties including colorectal surgery and cardiovascular surgery. Vascular surgery ERPs for lower extremity bypass have also been associated with improved patient outcomes regarding length of stay after index operation. The primary goal of ERPs is to reduce perioperative physiological stress. This is achieved in a multimodal fashion by limiting preoperative fasting, optimizing perioperative pain control, early initiation of postoperative nutrition, and early involvement of physical rehabilitation. In ERPs, as part of a limited duration preoperative fast, patients are also recommended to have a preoperative glucose load via clear carbohydrate drink up to 2-hours prior to surgery unless contraindicated by comorbidities. Preoperative glucose loading has been shown to decrease metabolic stress and is associated with early return of bowel function, improved pain control, and decrease in hospital length of stay after surgery. However, there is a paucity of research regarding the management of diabetic patients within vascular ERPs and glycemic control postoperatively. Poor glycemic control in the perioperative period is associated with increased rate of surgical site infection and worse overall outcomes. Given the high prevalence of diabetes among vascular patients, we sought to determine the effects of preoperative glucose loading on perioperative glycemic control and operative outcomes.
Methods: We performed a retrospective review of patients undergoing infrainguinal lower extremity bypass at our institution from January-2016 to July-2022; our lower extremity ERP was implemented in May 2018. Patient data regarding ERP implementation has been prospectively collected since this time. Patients were stratified by diabetes diagnosis and hemoglobin A1C (HbA1C) levels. Perioperative glycemic control was then compared between the glucose loading (GL) and non-glucose loading (NGL) patients within groups. Early perioperative hyperglycemia (EPH) was designated as blood glucose (BG) of > 180 mg/dL within the first 24 hours of surgery. Average patient glucose values on each postoperative day were compared between GL and NGL-groups. Additionally, patients were stratified by indication for revascularization (claudication, rest pain, or tissue loss) and perioperative glycemic control was compared. Finally, perioperative outcomes such as surgical site infection, readmission, reinterventions, and complications were compared between the groups. Chi-square tests were utilized for categorical variable comparison and two-sided T-tests for continuous variables.
Results: The cohort consisted of 393 patients identified within the Electronic Health Record with 161 patients in the pre-ERP group and 232 patients in the ERP group. There were no demographic differences between the ERP and pre-ERP groups. Overall, 42% of the cohort were diabetics with 44.5% of these patients undergoing preoperative glucose loading. All glucose loading (GL) patients were within the vascular ERP. Regarding demographics within the NGL-patients and GL-patients, the GL-group contained more non-white patients (44.8% vs 34.7%; p=0.041) and had a lower number of diabetic patients (36.9% vs. 48.9%; p=0.016). Otherwise, patient demographics were similar with average age of 64.7 ±9.4 in the NGL-group and 63.5 ± 9.7 in the GL-group.
Evaluating glucose loading within diabetic patients revealed 72.6% of GL-diabetic patients experienced EPH compared to 44.0% of NGL (CI 95%; p=<0.001). While a trend towards greater rates of EPH were seen in all glucose-loaded groups, when stratified by HbA1C, a significantly higher rate of EPH was seen only in the HbA1c >8 groups, with 90.3% in the GL-patients compared to 59.3% of NGL-patients (CI 95%; p=0.006) (Table 1).
Looking beyond the immediate perioperative period, the median BG levels from each postoperative day were averaged among patients in different groups and compared. The trends in hyperglycemia of BG>180 across hospital days are shown in Figure 1. There is a statistically significant difference observed on postoperative day (POD) 1 with 66.1% of GL-diabetic patient average blood glucose values above the
threshold of 180 mg/dL compared to 29.3% of NGL-diabetic patients (p=<0.001) (Figure 1). This significant difference dissipates at POD3. Average BG values for GL-diabetic patients on POD1, 2, and 3 were 206 ± 65 mg/dL, 226 ± 89 mg/dL, and 194 ± 72 mg/dL compared to 176 ± 56 mg/dL, 166 ± 61, and 170 ± 58 mg/dL, respectively. To assess for sensitivity of using average of daily median BG levels, we repeated this analysis using the average of maximum daily values. This yielded similar results within the first 24-hours postoperatively as well as throughout hospital stay.
Hypoglycemic events (BG<70 mg/dL) among the cohorts were also explored. This revealed only one patient within the GL-diabetic patients and one patient within the GL-non-diabetic patients experiencing hypoglycemic events within the first 24-hours postoperatively. Beyond the first postoperative 24-hours, there were less than 3 events per day of hypoglycemia within the cohort. Interestingly, all patients experiencing hypoglycemia were within the GL group.
Patients were further stratified by claudication, rest pain, or tissue loss as indication for revascularization to assess if procedural indication, particularly wound presence, was a confounding factor for EPH. GL-diabetic patients follow a similar postoperative trend demonstrated in Figure 1 regardless of surgical indication. Within the first 24-hours postoperatively, 69.2% of GL-diabetic patients with claudication experienced BG > 180 mg/dL compared to 87% with rest pain, and 64.9% with tissue loss (p=0.168). In NGL-diabetic patients, 35% with claudication experienced BG > 180 mg/dL compared to 45% with rest pain, and 48% with tissue loss (p=0.633). Overall, surgical indication did not significantly affect perioperative glycemic control.
Despite differences in glycemic control, there were no significant differences between diabetic GL-patients and NGL-patients regarding surgical site infection, readmission, reintervention, or complications at 30-days. Interestingly, a lower percentage of GL-diabetic patients experience surgical site infection (9.3% vs 14% p=0.474) and overall complications (26.7% vs 34.4% p=0.316) when compared to NGL-diabetic patients, respectively. This is likely due to the fact that GL-diabetic patients exclusively within the ERP and potentially benefited from additional enhanced recovery components compared to the routine care within the patients not in the ERP.
Conclusions:Overall, preoperative glucose loading in diabetic patients is associated with poor glycemic control perioperatively. Patients with severe, uncontrolled diabetes with HbA1C>8 appear to be at increased risk of hyperglycemia within 24 hours after surgery. Non-diabetic GL and NGL-patients, however, had comparable low rates of perioperative hyperglycemia. Although early postoperative hyperglycemia and perioperative hyperglycemia are associated with preoperative glucose loading clinical events of 30-day complications and surgical site infections were not associated with glucose loading in our cohort. While the early postoperative hyperglycemia may be more directly attributed to presence or absence of glucose loading, ERP principles of early return to postoperative oral intake and differences in postoperative glycemic control regimens hold the potential for confounding associations with glycemic control throughout the later phases of the perioperative course during index hospitalization. However, given the current study, it may benefit poorly controlled diabetic patients to withhold preoperative glucose loading within vascular ERPs. 
Back to 2023 Abstracts