Preoperative Carbohydrate Load and Intraoperative Goal-Directed Fluid Therapy for Elderly Patients Undergoing Open Gastrointestinal Operation: A Prospective Randomized Controlled Trial

Background: At present it remains uncertain as to whether carbohydrate (200 ml) loaded 2 hours before anesthesia induction combined with intraoperative goal-directed uid therapy (GDFT) is benecial to elderly patients undergoing gastrointestinal operations. As such, a randomized controlled trial was designed to evaluate the relative impact of perioperative uid optimisation versus conventional uid therapy (CFT) on clinical outcomes in elderly patients following gastrointestinal surgery. Methods: A total of 120 elderly patients undergoing gastrointestinal surgery were randomized into a CFT group (n = 60) with traditional methods of fasting and water-deprivation, and a GDFT group (n = 60) with carbohydrate (200 ml) load 2 hours before anesthesia induction. The CFT group underwent routine monitoring during surgery, however, the GDFT group was conducted by a Vigileo/FloTrac monitor with cardiac index (CI), stroke volume variation (SVV), and mean arterial pressure (MAP). For all patients, demographic information, intraoperative parameters and postoperative outcomes were recorded. Results: Patients in the GDFT group were administered less crystalloid uid (1111 ± 442.9 ml vs 1411 ± 412.6 ml; p = 0.000) and produced less urine output (200 ml [150-300] vs 400 ml [290-500]; p = 0.000) relative to patients in the CFT group. Moreover, GDFT was associated with a shorter average time to rst atus (55 ± 13.9 hours vs 65 ± 22.6 hours; p = 0.004) and oral intake (72 ± 17.4 hours vs 85 ± 27.5 hours; p = 0.002), as well as a reduction in the rate of postoperative complications (14 (25.5%) vs 27(47.4%) patients; p = 0.016). However, postoperative hospitalization or hospitalization expenses were similar between groups. Conclusions: Focused on elderly patients undergoing open gastrointestinal surgery, we found perioperative uid optimisation may be associated with improvement of bowel function and a lower incidence of postoperative complications. studies regarding whether preoperative carbohydrate load combined with intraoperative GDFT is similarly benecial in elderly patient populations. As such, in the present study we conducted uid optimisation in an elderly patient population via the use of hemodynamic indicators (CI, SVV, MAP) and vasoactive drugs, as necessary to. We hypothesized that preoperative carbohydrate (200 ml) load and intraoperative GDFT based upon SVV and CI as determined by analyzing the radial artery pulse waveform and MAP may lower postoperative hospitalization and postoperative complications in elderly patients undergoing gastrointestinal surgery.


Background
Intraoperative appropriate capacity is essential in the eld of anesthetic management and is of particular concern in the treatment of older adults [1]. Individuals ≥ 65 years old are termed elderly [2]. Elderly patients typically present with a range of physical, pharmacological, and psychological comorbidities that must be carefully considered by clinicians and anesthesiologists prior to any major surgical operation [3]. Indeed, more advanced age remains a key risk factor associated with higher rates of postoperative morbidity and mortality [4].
Elderly patients scheduled to undergo gastrointestinal surgery are more likely to present to the operating room in a dehydrated or hypovolemic state as a consequence of having followed the necessary fasting and bowel preparation procedures. Intraoperative hypovolemia can, in turn, result in a range of serious postoperative complications including hypotension and severe arrhythmia, whereas hypervolemia can also cause serious problems such as anastomotic leaks, infection, pulmonary edema, or even death [5][6][7][8][9]. As such, it is clear that both insu cient and excess uid infusion can cause harm in patients [10].
Owing to age-related declines in organ function and greater di culty adjusting uid preloading, elderly patients are thus at a substantially higher risk of postoperative mortality. It is essential that caution be taken when infusing uids into elderly patients.
Preoperative uid therapy mainly aims to present the patients from a dehydrated or hypovolemic state before anesthesia induction. Therefore, multiple ERAS (Enhanced Recovery After Surgery) guidelines include the oral intake of carbohydrate load (200 ml) 2 hours before surgery, which may help to decrease the postoperative complications, such as postoperative nausea and vomiting, wound infection. Besides, the use of routine hemodynamic measurements (such as arterial blood pressure or heart rate) is often a relatively imprecise means of monitoring for changes in blood volume [11,12]. In contrast, goal-directed uid therapy approaches rely upon monitoring more advanced hemodynamic variables such as SVV and pulse pressure variability, which are more sensitive to hypovolemia and therefore allow for optimal preloading by enabling clinicians to appropriate titrate uids and inotropic substances [13,14]. A number of GDFT devices have been employed in prior trials. The Vigileo/FloTrac system commonly utilized owing to its performance being similar to that of the PiCCO system when predicting uid responsiveness based upon [15]. In addition, this system can improve cardiovascular function and maintain an adequate intraoperative oxygen supply [16]. This procedure is also minimally invasive and simple to use, and as such this advanced hemodynamic monitor is recommended for measuring uid responsiveness in patients undergoing high-risk major surgery [17].
GDFT has been shown to reduce the duration of hospitalization and to decrease the incidence of postoperative complications by 20-50% in previous systematic reviews [18][19][20]. However, there have been few studies regarding whether preoperative carbohydrate load combined with intraoperative GDFT is similarly bene cial in elderly patient populations. As such, in the present study we conducted uid optimisation in an elderly patient population via the use of hemodynamic indicators (CI, SVV, MAP) and vasoactive drugs, as necessary to. We hypothesized that preoperative carbohydrate (200 ml) load and intraoperative GDFT based upon SVV and CI as determined by analyzing the radial artery pulse waveform and MAP may lower postoperative hospitalization and postoperative complications in elderly patients undergoing gastrointestinal surgery.  [21], and as such patients were excluded from this study if they exhibited severe arrhythmia, or if they had a history of gastrointestinal surgery, peripheral artery disease, or with high risk of re ux and aspiration (pyloric obstruction or achalasia of the cardia). All patients provided written informed consent upon recruitment.

Randomization and blinding
The randomization of patients was conducted by using a series of computer-generated random numbers. The blinding of assignment for study group or control group was ensured by the opaque envelopes.
However, only the patients were blind to group assignment because the attending anesthesiologist may regulate the speed of uid infusion based on the parameters of the cardiac index trending monitor.

Intraoperative management
All patients were monitored for pulse oximetry, ve-lead-electrocardiogram, blood pressure, heart rate, temperature and end-tidal carbon dioxide concentrations. Central venous pressure (CVP) monitoring was conducted in some patients as decided by the attending anesthesiologist. Patients in both groups received a minimum of one peripheral intravenous access (18G), with a central venous catheter being inserted via the internal jugular vein and with radial arterial catheters being introduced via the nondominant forearm under low-dose dexmedetomidine for sedation in the operating room at 7:30 AM on the day of surgery. For each patient, the following was recorded: baseline characteristics, surgery type, and basic hemodynamics variables (blood pressure, MAP, HR). An initial arterial blood gas analysis was also conducted at this time. For patients in the GDFT group, the radial arterial lines were attached to the fourth generation Vigileo/FloTrac monitor (Edwards Life Sciences, CA, USA). After making sure that the patient was not in high risk of re ux aspiration using the technique of point-of-care gastric ultrasound, anesthesia was then induced by using propofol (2-2.5 mg.kg − 1 ) or etomidate (0.2 mg.kg − 1 ) in combination with sufentanil (15-25 ug) and cisatracurium (0.15-0.2 mg.kg − 1 ). Anesthesia was maintained with remifentanil (0.2-0.25 ug.kg − 1 min − 1 ), dexmedetomidine (0.4-0.7 ug.kg − 1 h − 1 ) and sevo urane at a MAC (minimum alveolar concentration) of 1-1.3 in order to achieve a BIS (bispectral index) value of 40-60. Each patient was ventilated with a tidal volume of 8 ml/kg without PEEP (positive end-expiratory pressure) at a respiratory rate su cient to achieve an end-tidal carbon dioxide value of 35-45 mmHg. Flow rates were set to 2-3 L/min, and 60% oxygen was administered for the operative duration.

Study protocol
The bowel preparation was performed by the surgeons' instruction 1 day before surgery. The GDFT group was fasted for 8 hours and drunk 200 ml carbohydrate (Shuneng; Yi Chang Ren Fu Pharmaceutical Co Ltd, China) before anesthesia induction. The CFT group was fasted for 8 hours and prohibited drinking for 4 hours before surgery. The CFT group was initially infused with a balanced crystalloid solution at 2 ml/kg/h for baseline uid therapy and 6 ml/kg/h during surgery (from incision to suturing). In addition, additional bolus (crystalloid or colloids) could be infused at the discretion of the attending anesthesiologist, or vasoactive drugs could be administered to ensure that routine monitoring variables remained within the normal range (heart rate < 100 bpm, MAP > 65 mmHg, urine output > 0.5 ml/kg/h). Patients in the GDFT treatment group were infused based on the Vigileo/FloTrac instrument. If CI was ≥ 2.5 l/min/m 2 , MAP was > 65 mmHg, and SVV was < 12%, the administration rate was maintained at 6 ml/kg/h for the entire surgical operation. When CI reached ≤ 2.5 l/min/m 2 with SVV > 12%, additional colloid 250 ml boluses were administered within 10 minutes. When CI was ≤ 2.5 l/min/m 2 and SVV was < 12%, dobutamine was injected to achieve a CI of ≥ 2.5 l/min/m 2 . Ephedrine boluses of 3-18 mg or norepinephrine infusions were administered when CI was ≥ 2.5 l/min/m 2 and MAP was < 65 mmHg ( Fig. 2). All hemodynamic variables were reassessed after 10 minutes.

Data collection
We recorded HR and MAP values at 5 min prior to the induction of anesthesia (T1), 30 min after the operation was initiated (T2), 60 min after the operation was initiated (T3), 90 min after the operation was initiated (T4), 120 min after the operation was initiated (T5), and at the end of surgery (T6). We additionally measured arterial lactate concentrations at T1, T3, and T6. All intraoperative data were collected upon the completion of the operation, including operative duration, duration of anesthetization, types and amounts of uid infusion, estimated blood loss, urine output, and vasoactive agent use. Following intensive care unit (ICU) or ward admission, the management of all patients was conducted by medical staff blinded to patient group assignments. Time to rst atus and time to rst oral intake was recorded. Following patient discharge, we recorded the length of hospitalization as well as postoperative complications. The prede ned criteria for complications in the study protocol were described in Additional 1. At rst, we determined the length of hospitalization as the primary outcome and the rate of postoperative complications as secondary outcome. However, during analysis of some experimental data, we noted that the reduction in the length of hospitalization was lower than that we had been predicted in the original study design. Moreover, we found that the interventions may have more effect on the rate of postoperative complications than the length of hospitalization. Therefore, we changed the rate of postoperative complications for primary outcome. Moreover, we calculated the sample size once again based on the new primary outcome.

Statistical analysis
Sample size calculation was according to a prospective analysis of postoperative hospital stay and the rate of postoperative complications for patients undergoing major bowel surgery [22]. The rate of postoperative complication in the study was 59.3% in the control group and 37.5% in the treatment group.
For a reduction in the postoperative complication from 60 to 32% with type I error of 0.05 and a power of 80%, a sample size of 50 patients was calculated by our researchers for each group. Owing to 20% patients with changes in surgical procedure, our study nally included 60 patients per group. SPSS 26.0 (IBM, IL, USA) was used for all statistical testing. Data were given as mean ± SD, absolute values (percentages) or medians (25th -75th percentiles). Data were compared via Student's t-tests (for normally distributed variables), Wilcoxon rank-sum tests (for non-normally distributed variables), and chi-squared or Fisher's exact tests (for categorical variables). Normality was assessed via the Kolmogorov-Smirnov test. Hemodynamic variables (MAP or HR) that were normally compared via repeated measures one-way ANOVAs, whereas they were compared via the Friedman test when non-normally distributed. p < 0.05 was the signi cance threshold.

Results
In total, 225 patients were eligible for inclusion in the present study between May 2018 and November 2019. After excluding patients based on our exclusion criteria and failure to give consent, the remaining 120 patients were randomized into the GDFT (n = 60) and CFT (n = 60) groups. Among these patients, 5 (2 and 3 in the GDFT and CFT groups, respectively) dropped out of the study, 3 (1 and 2 in the GDFT and CFT groups, respectively) were found to be inoperable (Fig. 1). In total, our nal analyses thus included 55 patients in the GDFT group and 57 patients in the CFT group.
With respect to postoperative bowel function, we found that GDFT was associated with a shorter time to rst atus (55 ± 13.9 hours vs 65 ± 22.6 hours; p = 0.004) and to rst oral intake (72 ± 17.4 hours vs 85 ± 27.5 hours; p = 0.002) relative to the CFT group. In addition, there were signi cantly fewer instances of postoperative complications among patients in the GDFT group relative to the CFT group (14 (25.5%) vs 27 (47.4%) patients; p = 0.016). Further analysis of prede ned types of complications found that there was a lower rate of infection complications in in the GDFT group relative to the CFT group (5 (9.1%) vs 13 (22.8%) patients; p = 0.048). However, GDFT was not associated with any signi cant differences in postoperative hospitalization duration, ICU admission, or hospitalization expenses. Although there were no signi cant differences between groups with respect to postoperative mortality, two patients in the CFT group did die on postoperative day 3 as a result of acute massive gastrointestinal hemorrhage and acute myocardial infraction respectively. For further details, see Table 4.

Discussion
In this prospective randomized study, we compared outcomes in elderly patients undergoing open gastrointestinal surgery that were treated via traditional standard of preoperative fasting and CFT or via perioperative uid optimisation including preoperative carbohydrate load and intraoperative GDFT. This analysis revealed that the latter was associated with more rapid restoration of bowel function and with a lower rate of postoperative complications.
For the present study, we speci cally analyzed outcomes in elderly patients undergoing open gastrointestinal surgery for two reasons. First, elderly patients with gastric or colorectal cancer have high rates of comorbidities that are associated with increased surgical risk [23]. Second, unlike laparoscopic surgery, open gastrointestinal surgery is associated with large changes in intraoperative uid levels [24], and as such poor organ perfusion or uid overload have the potential to lead to poorer surgical outcomes. Further study of optimal care techniques for elderly patients undergoing major abdominal surgery is needed, and as such in the present study we sought to determine whether the reduction of the time for preoperative fasting and GDFT was able to improve clinical outcomes in these patients.
For this study, we opted to utilize CI rather than the oxygen delivery index as the key target variable for our GDFT protocol as it can be readily and continuously monitored via radial artery pulse waveform analysis. When arterial oxygen saturation and hemoglobin levels are adequate, CI can serve as an effective means of evaluating oxygen supply within tissues and organs [25]. As the use of CI and SVV with the Vigileo/FloTrac monitor has been found to be potentially unreliable in patients with irregular heart rhythms [21,26] and in patients with poorly controlled intraoperative ventilation [27], we did not include patients with severe arrhythmia in the present study, and all patients were ventilated using a tidal volume of 8 ml/kg.
Excess uid transfusion can result in a range of problems including increased rates of postoperative cardiac morbidity, pneumonia, respiratory failure, delayed wound healing, and anastomotic leak as a consequence of intestinal edema in patients undergoing colorectal surgery [8]. As shown in some prior trials [16,25,28], our study also found that the GDFT protocol is associated with reductions in the rate of postoperative complications. In addition, the lower incidence of postoperative infection complications in the GDFT group may also be related to preoperative oral carbohydrate loading, which may reduce the metabolic and in ammatory response after surgery and improved surgical clinical outcomes [29]. However, the rates of postoperative complications in some previous studies were higher than in our present analysis. There are two potential reasons for this discrepancy. For one, Mayer et al. [25] selected high-risk patients with at least two risk factors according to risk index of Lee [30] as their experimental subjects. Furthermore, Benes et al. [28] recruited patients that had to meet both operation-related and patient-related high-risk criteria. In contrast, all patients we included were over the age of 65, some of whom may be in the low-to-moderate risk category. In addition, we only monitored the incidence of postoperative complications that occurred during hospitalization, whereas Benes et al. monitored patients for 30 days [28]. Together, our results suggest that perioperative uid optimisation reduces the incidence of postoperative complications in elderly patients, rather than only in high-risk patients.
The relationship between GDFT and the duration of postoperative hospitalization remains controversial.
Early studies found GDFT to be associated with a reduction in the duration of postoperative hospitalization [25], but this nding was not repeated in more recent studies [31], particularly in studies performed in an ERAS (enhanced recovery after surgery) setting [2,32]. The increased application of ERAS protocols has led to improvements in the quality of perioperative care in control groups over the past 15 years, potentially in uencing these study results [10]. However, hospitalization duration can also be in uenced by a range of other factors such as patients' wishes, preoperative physical condition, health care system requirements, and institution-speci c differences in treatment regimens. All of these factors thus have the potential to in uence the relationship between GDFT and postoperative hospitalization duration.
Given that it is a sensitive indicator of organ perfusion, arterial lactate levels are closely linked to tissue hypoxia and blood volume insu ciency [33]. In the context of suboptimal uid management, postoperative restoration of gastrointestinal motility and oral food intake may be delayed due to higher lactate levels. While at baseline in the present study we found that lactate levels were higher in the GDFT group relative to the CFT group, we did not observe any signi cant differences between these values at analyzed time points. However, we did nd that lactate values in the CFT group increased at the end of surgery relative to baseline values. The slower recovery of bowel function observed in the CFT group may thus be associated with increased lactate production.
Urine output is typically utilized to approximate renal function and blood volume. In this study, we found that patients in the GDFT treatment group required reduced crystalloid administration and exhibited reduced urine output relative to patients in the CFT group. Postoperative renal failure occurred in one patient in the GDFT group and two patients in the CFT group. This outcome may be explained by work conducted by Kheterpal et al [34], as these authors suggested that urine is unreliable when used as a marker of blood volume and renal function, given that it can be in uenced by neurohormonal signaling in response to operative stress. The administration of diuretics and vasoactive agents has also been linked to the incidence of acute renal failure.
There are multiple limitations to the present analysis. For one, in an effort to avoid potential compounds we did not utilize the CI and SVV trending monitor for patients in the CFT group, so we were not able to compare these parameters between groups. Moreover, we were not able to blind investigators intraoperatively to patient treatment strategies. Additionally, intraoperative attending anesthesiologists were the primary focus of this study. The professional level of the attending anesthesiologists and physicians for postoperative treatment in the ICU and hospital wards should be equivalent between groups. Otherwise, it is possible that higher rates of postoperative complications may result from poorer postoperative care in the CFT group. Lastly, we did not pre-de ne standardized discharge criteria in the present study, as discharge may be in uenced by a range of factors including patient demands, health care system capacity, and speci c treatment regimens used. Thus, the duration of postoperative hospitalization could be also biased.

Conclusion
The results of this study indicate that preoperative carbohydrate load and intraoperative uid optimization guided by CI, SVV, and MAP may be associated with more rapid improvements in bowel function and a decreased incidence of postoperative complications in elderly patients undergoing open gastrointestinal surgery. However, this GDFT treatment strategy was not associated with any decreases in postoperative hospitalization.