Preoperative carbohydrate loading in diabetic patients undergoing gastrointestinal surgery: A pilot randomized trial


 BackgroundPreoperative carbohydrate loading is used to improve patients’ comfort and recovery, but evidence remains limited in diabetic patients. We tested the feasibility of a preoperative carbohydrate drink with supplemental insulin in diabetic patients for gastrointestinal surgery.MethodsAdult patients with type 2 diabetes mellitus who were scheduled for major gastrointestinal surgery were randomized to carbohydrate group (carbohydrate drink with supplement insulin selectively) or control group (clinical routine management). The primary outcome was the time to first flatus after surgery. Among secondary outcomes, subjective feelings of thirsty, hunger and fatigue were assessed with the Visual Analogue Scale (scores range from 0 to 100, where 0 indicate no discomfort and 100 the most severe discomfort) before and after surgery.ResultsA total of 63 patients were randomized. Time to first flatus did not differ between groups (median [95% CI], 40 hours [30, 50] in control group vs. 43 hours [37, 48] in carbohydrate group, hazard ratio [95% CI], 1.24 [0.74, 2.07], P = 0.411). Both pre- and postoperative subjective feelings of discomfort were all significantly lower in carbohydrate group than in control group (median difference from − 50 to 0; all P < 0.05). Patients with carbohydrate drink developed less intraoperative hypotension (40.6% [13/30] vs. 16.1% [5/31], P = 0.031) and postoperative nausea and vomiting within 24 hours (31.3% [10/32] vs. 9.7% [3/31], P = 0.034).ConclusionIn diabetic patients undergoing gastrointestinal surgery, preoperative carbohydrate loading with supplemental insulin selectively did not affect gastrointestinal recovery; but it improved perioperative comfort, and reduced intraoperative hypotension and postoperative nausea and vomiting.Trial RegistrationClinicalTrial.gov, NCT03204344; Registered July 2, 2017, https://clinicaltrials.gov/ct2/show/NCT03204344.


Conclusion
In diabetic patients undergoing gastrointestinal surgery, preoperative carbohydrate loading with supplemental insulin selectively did not affect gastrointestinal recovery; but it improved perioperative comfort, and reduced intraoperative hypotension and postoperative nausea and vomiting.

Background
Previous studies proved bene ts of preoperative carbohydrate loading, which not only improved perioperative well-being but also decreased insulin resistance after elective surgery (Bilku, et al., 2014;Awad, et al., 2013). Some studies also reported that preoperative oral carbohydrate promoted the recovery of gastrointestinal function (ÖZdemİR, et al., 2011; and shortened the length of hospital stay without increasing the risk of aspiration (Gianotti, et al., 2018;Smith, et al., 2014). Several guidelines recommend that carbohydrate drinks should be allowed in patients without diabetes undergoing elective surgery until 2 hours before initiating anesthesia (Gustafsson, et al., 2018;Feldheiser, et al., 2016).
Currently, available evidences regarding the risks and bene ts of preoperative carbohydrate loading mainly come from patients without diabetes. Data on diabetic patients are limited and there are no related recommendations until now.
In a preliminary study of patients with well-controlled type 2 diabetes, Gustafsson et al. (Gustafsson, et al., 2008) found there was no delayed gastric emptying after oral intake of carbohydrate drink, and they suggested it could be administered until 3 hours before anesthesia. Other studies also showed that preoperative carbohydrate loading was safe in type 2 diabetic patients; however, it did not affect postoperative insulin resistance or help to preserve lean body mass after surgery (Azagury, et al., 2015;Breuer, et al., 2006). To be noted, additional insulin was not provided after oral carbohydrate intake in the above studies; and hyperglycemia after surgery might be associated with worse outcomes, such as delayed intestinal recovery, more complications, and longer hospital stay (Hu, et al., 2016;Kiran, et al., 2013;Kwon, et al., 2013).
This pilot trial was designed to test the feasibility of preoperative carbohydrate drink with supplemental insulin in type 2 diabetic patients undergoing gastrointestinal surgery. We hypothesized it would facilitate the return of gastrointestinal function, and thus promote recovery after surgery without increasing hyperglycemia.

Study design
This pilot, observer-blinded, randomized trial was approved by the Biomedical Research Ethics Committee of Peking University First Hospital (2017 − 1362) and registered with clinicaltrials.gov (NCT03204344) on July 2, 2017. Written informed consent was obtained from each patient before enrollment.

Participants
The inclusion criteria were: (1) age of 18 years and beyond; (2) diagnosed as type 2 diabetes with fasting blood glucose < 10 mmol/L before surgery; (3) scheduled to undergo elective gastrointestinal surgery with anticipated duration ≥ 2 hours. Patients who met any of the following criteria were excluded: (1) comorbid with diaphragmatic hernia or gastric esophageal re ux, or with pregnancy; (2) previous history of total or partial gastrectomy; (3) preoperative pyloric or intestinal obstruction; or (4) preoperative New York Heart Association class IV, renal failure, severe hepatic disease, or American Society of Anesthesiologists class IV or higher.

Randomization and intervention
Random numbers were generated using the SAS statistical package version 9.3 (SAS Institute, Cary, NC, USA) in a 1:1 ratio with a block size of 4. Randomization was strati ed according to the approach of surgery, i.e., laparotomy or laparoscopic surgery. The randomization results were sealed in sequentially numbered envelops until shortly before the enrollment. On the afternoon of the day before surgery, an investigator (XL) opened the envelop and managed the intervention according to the allocation.
Patients randomly assigned to the control group were managed according to routine practice, i.e., they were fasted from solid foods from 22:00 p.m. the night before surgery, but clear water was allowed until 06:00 a.m. on the surgery day. For patients whose surgery was scheduled after 12:00 a.m., 5% glucose (500-1000 ml) through intravenous infusion was prescribed with a glucose-to-insulin ratio of 4-6 to 1 and adjusted according to blood glucose level. Electrolytes were added when considered necessary.
Patients randomly assigned to carbohydrate group were provided with a carbohydrate-rich drink (Outfast [355 ml/bottle, 14.2% carbohydrate, osmotic pressure 280-300 mmol/L, pH 3.8-4.3, and contains maltodextrin, fructose, glucose, arginine, taurine, vitamin B, and electrolyte], Yichang Human-well Pharmaceutical Co., Ltd., China). They were instructed to ingest the carbohydrate drink at 22:00 p.m. the night before surgery, at 06:00 a.m. the morning of surgery. No other clear uid was allowed. Insulin aspart was injected subcutaneously selectively 10 minutes before giving carbohydrate drink. The volume of carbohydrate drink and the dose of insulin were prescribed by an endocrinologist according to the situation of individual patients (Appendix 1).

Anesthesia and perioperative management
No anesthetic premedication was administered. General anesthesia was performed in all patients with routine monitoring. Regional anesthesia including epidural or trunk block could be combined with general anesthesia at the discretion of anesthesiologists. Patient-controlled analgesic pump with uni ed drugs was provided after surgery.
Preoperative mechanical bowel preparation, nasogastric tube insertion and surgical approaches were decided by surgeons according to patients' condition. Parenteral nutrition was prescribed for all patients from the rst postoperative day. The time to start water drinking and liquid diet were decided by surgeons. Patients were discharged to home when they met the following criteria: (1) do not require intravenous nutrition; (2) tolerate pain with/without oral analgesics; (3) able to ambulate without assistance; and (4) no severe complications requiring therapy (Fujikuni, et al., 2016; Surgery and Anesthesiology, 2018).

Data collection and outcome assessment
Baseline data including demographics, comorbidities, laboratory test results and other necessary information were collected. The Charlson comorbidity index was also calculated (Quan, et al., 2011).
Intraoperative data about details of anesthesia and surgery were documented.
Subjective well-being feelings including the severities of thirsty, hungry, and fatigue were assessed by investigators who didn't know the allocation (XQL and YTL) using the Visual Analogue Scale (VAS, scores range from 0 to 100, where 0 indicate no discomfort and 100 the most severe discomfort) at 30 minutes before anesthesia induction and 4-6 hours after the surgery. Patients were followed up daily (09:00 a.m.) by XQL and YTL during the rst 5 postoperative days and weekly until hospital discharge. On postoperative day 30, patients were followed-up by telephone interview. Postoperative complications (including re-hospitalization) were de ned as newly occurred medical conditions that required therapeutic intervention (class 2 or higher in Clavien-Dindo classi cation) (Katayama, et al., 2016).
The primary outcome was the time to rst atus. Secondary outcomes included subjective well-being feelings before and after surgery, time to recovery ( rst ambulation, drink, defecation, and liquid diet, as well as walking distance) after surgery, blood glucose level, occurrence of postoperative complications within 30 days, length of stay in hospital after surgery, and all-cause 30-day mortality. As a post-hoc analysis, we also compared maximal blood glucose uctuation between groups. Adverse events were monitored from 22:00 p.m. (i.e., the beginning of carbohydrate drink in the carbohydrate group) until 24:00 p.m. on the rst postoperative day. Hypoglycemia was de ned as blood glucose < 3.9 mmol/L, with or without symptom. Hyperglycemia was de ned as blood glucose > 13.3 mmol/L (Fayfman, et al., 2019;Pasquel, et al., 2017). Aspiration was de ned as gastric content appeared in the mouth and entered the trachea or lungs. Hypotension was de ned as systolic blood pressure < 90 mmHg or a decrease of > 30% from baseline (average value in the ward). Occurrence of postoperative nausea and vomiting (PONV) within the rst 24 postoperative hours was noted. All adverse events were managed according to routine practice.

Sample size estimation
Based on a preliminary survey in our patients, time to rst atus after gastrointestinal surgery in patients without preoperative carbohydrate loading was 40 ±12.5 hours. We assumed that preoperative carbohydrate loading could decrease the time interval by 9.6 hours (Smith, et al., 2014). With the signi cance level set at 0.05 and power at 80%, the sample size required to detect differences was 56 patients (28 in each group). Taking into account a drop-out rate of approximately 10%, we planned to enroll 62 patients. Sample size calculation was performed with the PASS 15.0 software (Stata Corp. LP, College Station, TX).

Outcome analyses
The balance of baseline data between the two groups was assessed with the absolute standardized differences, which were calculated as the absolute difference in means, medians, or proportions divided by the pooled standard deviation (Austin, 2009 Outcome analyses were performed in the intention-to-treat population. For the primary endpoint, a perprotocol analysis was also performed. A two-sided P < 0.05 was considered statistically signi cant. Statistical analysis was performed on SPSS 25.0 software package (IBM SPSS, Chicago, IL).

Results
From August 1, 2017 to May 7, 2018, 63 patients were enrolled and randomly assigned to either the carbohydrate group (n = 31) or the control group (n = 32). All enrolled patients received the allocated intervention, but three patients in carbohydrate group had protocol deviation as they took carbohydrate drink only once at 22:00 p.m. the night before surgery. No patients were lost to follow-up during hospitalization and at day 30 postoperative (Fig. 1).
Among baseline data, glycosylated hemoglobin, hemoglobin and Barthel index were higher in carbohydrate group than in control group; other variables were comparable between the two groups ( Table 1). Intra-and postoperative management were comparable between the two groups (Tables 2).  ASD, absolute standardized difference (an ASD of ≥ 0.493 is considered imbalanced between the two groups); OAD, oral antidiabetic drug; w or w/o, with or without; COPD, chronic obstructive pulmonary disease; ASA, American Society of Anesthesiologists; NYHA, New York Heart Association. a Serum aspartate aminotransferase and/or alanine aminotransferase level higher than upper normal limit.
b According to the KDIGO (Kidney Disease Improving Global Outcomes) criteria.
c Score ranges from 0 to 100, with higher scores indicating better daily life activity.  (Table 3). b Assessed with visual analogue scale; scale ranges from 0 to 100 where 0 = no uncomfortable feeling and 100 = the most severe uncomfortable feeling. One patient in control group did not provide subject feeling in 4-6 hours after surgery because of mechanical ventilation with moderate sedation.
c Calculated as maximal minus minimal blood glucose concentration from the night before surgery to the night of the rst day after surgery. from the night before surgery to the night of the rst day after surgery was higher in carbohydrate group than in control group (P = 0.012, Fig. 2). Other secondary outcomes including time to recovery ( rst ambulation, drink, defecation, and liquid diet), walking distance, length of hospital stay after surgery, complications within 30 days and all-cause 30-day mortality did not differ between groups. Maximal blood glucose uctuation was also similar between groups (  PONV, postoperative nausea and vomiting.
a De ned as blood glucose < 3.9 mmol/L, with or without symptom from the night before surgery to the night of the rst postoperative day.
b De ned as blood glucose > 13.3 mmol/L from the night before surgery to the night of the rst postoperative day.
c De ned as gastric content appeared in the mouth and entered the trachea or lungs. d De ned as systolic blood pressure < 90 mmHg or a decrease of more than 30% from baseline (average value in the ward).

Discussion
Our results showed that, for patients with type 2 diabetes undergoing elective gastrointestinal surgery, preoperative carbohydrate loading with supplemental insulin selectively did not promote the recovery of gastrointestinal function; however, it improved patients' comfort and decreased intraoperative hypotension and PONV.
Return of gastrointestinal function is crucial for postoperative recovery after gastrointestinal surgery (Hedrick, et al., 2018). Time to rst atus/defecation are widely used variables indicating the return of gastrointestinal function (Short, et al., 2015). In the present study, the median time to rst atus and defecation were 40 and 93 hours in control group, which were consist with previous results (Dulskas, 2015;Liang, et al., 2011). It was reported that preoperative carbohydrate loading improved the recovery of gastrointestinal function in various surgeries (ÖZdemİR, et al., 2011;Noblett, et al., 2006). However, up to now, sample sizes of available studies were small and conclusions couldn't be drawn regarding gastrointestinal function. For the rst time, we investigated the effect of preoperative carbohydrate loading on gastrointestinal function in diabetic patients but did not nd any difference between the two groups.
As for secondary outcomes, we found that preoperative carbohydrate loading signi cantly reduced patients' discomfort and the occurrence of PONV, which improved patients' well-beings during the perioperative period. Our results were consistent with previous studies (Cakar, et al., 2017;Kaska, et al., 2010). In the present study, another interesting nding was that preoperative carbohydrate loading facilitated intraoperative volume and hemodynamic maintenance, evidenced by less intraoperative hypotension. We also found time to rst ambulation tended to be earlier in patients with carbohydrate drinks. Other studies reported that preoperative carbohydrate loading was helpful in preserving muscle strength which might promote early ambulation after surgery (Liu, et  For many years, gastric emptying was thought to be decreased in diabetic patients and they were excluded from trials investigating preoperative carbohydrate loading for aspiration concerns. As a matter of fact, gastric emptying might be accelerated in diabetic patients (Gustafsson, et al., 2008;Mihai, et al., 2018). Studies reported that rapid gastric emptying occurred in those early type 2 diabetic patients (Schwartz, et al., 1996;Phillips, et al., 1992). Similar result was also found in patients with long diabetic course (Weytjens, et al., 1998). In accord with these, none of our patients in carbohydrate group developed aspiration. Another worry when administering carbohydrate loading in diabetic patients is hyperglycemia [9], which may worse perioperative outcomes (Kwon, et al., 2013). To avoid signi cant blood-glucose uctuation, we designed a protocol of prescribing volume of carbohydrate drink and dose of insulin according to the individual situation of patients. Our results showed that, although carbohydrate drink increased blood-glucose level, the maximal blood-glucose uctuation and the rate of hyperglycemia were similar between the two groups, indicating the feasibility of our carbohydrate loading regimen in diabetic patients.
Our study had several limitations. First, we did not measure the residual gastric volume before anesthesia induction, thus direct evidence of su cient gastric emptying was lacking. Second, carbohydrate drink used in this trial has a high glycemic index. Intermittent blood-glucose monitoring might have missed some higher values. A continuous glucose monitoring device may be more helpful in the future. Thirdly, we did not detect insulin resistance on the rst day after surgery.

Conclusions
Preoperative carbohydrate drinks with supplemental insulin selectively was feasible in diabetic patients undergoing gastrointestinal surgery. Although it did not promote the recovery of gastrointestinal function, it improved patients' comfort and decreased intraoperative hypotension and PONV. Availability of data and materials The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.   Blood-glucose level was higher in carbohydrate group than in control group. T1, at 00:00 a.m. on the day of surgery; T2, at 06:00 a.m. on the day of surgery; T3, immediately after anesthesia induction; T4, at the end of surgery; T5, at 21:00 p.m. on the day of surgery; T6, at 06:00 a.m. on the rst day after surgery; T7, at 13:00 p.m. on the rst day after surgery; T8, at 21:00 p.m. on the rst day after surgery. The box and whiskers plots show medians, interquartile ranges and outer ranges, and individual points mean mild outliers (○, which are outside 1.5 times of interquartile range) and extreme outliers (*, which are outside 3 times of interquartile range).