Effects of Preoperative Oral Carbohydrate Electrolyte Drinks on Preoperative Hypokalemia Incidence in Patients Scheduled for Laparoscopic Colorectal Resection: A Randomized Clinical Trial

Background: In our previous study, hypokalemia incidence was high in patients scheduled for laparoscopic colorectal resection. The present trial was conducted to verify the effects of preoperative carbohydrate drinks containing potassium in these patients. Methods: Patients were randomly assigned to control, placebo, and case groups. In control group, patients were fasted from midnight. In placebo group, patients were fasted from midnight and received carbohydrate drinks 2–3 h before surgery. In case group, patients were fasted from midnight and received carbohydrate drinks containing potassium supplementation 2–3 h before surgery. The primary outcome was the incidence and severity of preoperative hypokalemia. The other outcomes included postoperative gastrointestinal function, including the time to postoperative rst atus (FFL) and rst feces (FFE) and other complications. Results: The nal analysis included 122 participants. The incidence of preoperative hypokalemia in case group was signicantly lower than that in control and placebo groups (50% vs. 88.1% vs. 77.5%, p < 0.001). The severity of hypokalemia in control and placebo groups was greater than that in case group. No regurgitation or aspiration occurred in the three groups. No signicant differences were observed among the three groups regarding time to FFL and FFE. Conclusions: Preoperative carbohydrate drinks containing potassium may have signicantly reduced the incidence of preoperative hypokalemia and improved preoperative thirst and hunger, but did not improve postoperative well-being. However, as part of the enhanced recovery after surgery protocol, preoperative carbohydrate drinks containing potassium should be recommended, as early as rst admittance to hospital. Trial registration: study the Chinese Clinical Trial www.chictr.org (registration


Introduction
Compared with traditional management, enhanced recovery after surgery (ERAS) has been recommended in recent years. Following the ERAS protocol in elective colorectal surgery could improve the prognosis [1]. Preoperative fasting time and carbohydrate treatment are crucial parts of the ERAS protocol. Intake of clear uids until 2 h before anesthesia is not associated with increased complications such as regurgitation and aspiration [2,3]. Administration of oral carbohydrates 2-3 h before anesthesia induction provides bene cial effects such as improved insulin resistance [4,5]. However, preoperative carbohydrates might not speed up surgical recovery or improve postoperative well-being [6].
Similar to fasting and preoperative uids, bowel preparation is another focus of ERAS. Bowel preparation is recommended for rectal surgery and remains controversial in colonic surgery [1,7]. Hypokalemia is associated with various symptoms, including gastrointestinal disorders, arrhythmia, and muscle weakness. Therefore, the importance of normal perioperative potassium levels might be underestimated.
Nevertheless, no study has been conducted on preoperative potassium management in patients scheduled for laparoscopic colorectal resection, whose preoperative hypokalemia incidence was found to be high in our previous study [8]. Although oral administration is commonly recommended for prophylactic potassium supplement, a previous study demonstrated that use of the oral route is extremely rare [9]. Therefore, adding potassium to preoperative uids might be a useful choice.
The present randomized clinical trial was conducted to verify the effects of preoperative carbohydrate drinks containing potassium in patients scheduled for elective laparoscopic colorectal resection.

Methods
This randomized prospective parallel study was performed in accordance with the Declaration of All patients who were scheduled for elective laparoscopic colorectal resection from March 1, 2017 to February 29, 2018 were considered for inclusion. The inclusion criteria were as follows: (i) age 18-80 years, (ii) American Society of Anesthesiologists Physical Status I to III, (iii) provided informed written consent. Patients who met any of the following criteria were excluded: (i) comorbid intestinal obstruction and gastric emptying disorder; (ii) diabetes; (iii) body mass index 30 kg/m 2 ; (iv) severe renal insu ciency or long-term use of potassium-sparing diuretics; (v) history of neurological or mental illness; (vi) patients with hyperkalemia or hypokalemia. Patients who were transferred to open surgery or underwent colostomy during the operation were excluded from the nal analysis.
Patients were randomly assigned to control, placebo, and case groups according to computer-generated random numbers. In the control group, patients were fasted from midnight; in the placebo group, patients were fasted from midnight and received carbohydrate drinks (5-6 mL/kg containing 14.2% carbohydrates) 2-3 h before surgery; and in the case group, patients were fasted from midnight and received carbohydrate drinks containing potassium supplementation (5-6 mL/kg containing 14.2% carbohydrates and 1 g of potassium) 2-3 h before surgery. All the drinks were prepared in bottles prepared by the hospital pharmacy with the same appearance and were distributed by the same trial assistant who was not involved in the follow-up or assessment. The random number was sealed in an envelope until the morning of surgery. All patients were fasted from solid food from midnight before surgery. To maintain blinding, the anesthetist was not involved in the follow-up or assessment. Patients received standardized care during the perioperative period.

Sample size
This study was a randomized clinical trial with three parallel groups. The main outcomes were the incidence rates of preoperative hypokalemia. The following formula was used to calculate the sample sizes: According to our pilot trial, which resulted in p A = 0.88, p B = 0.77, and p C = 0.41, 40 participants were required in each group for a power of 80% and a two-tailed p-value of 0.05 was considered as statistically signi cant. A nal sample size of n = 44 was determined based on an assumed dropout rate of approximately 10%; therefore, 44 patients were enrolled in each group, and the total number of patients was 132.
Potassium was measured by obtaining venous blood before anesthesia induction. If the blood potassium was less than 2.8mmol/L, potassium would be supplemented intravenously before induction of anesthesia.
Standardized analgesic strategy was used for all patients. Fifteen minutes before the end of surgery, patients were intravenously infused with urbiprofen axetil (1 mg/kg) as an analgesic and tropisetron (5 mg) to prevent vomiting. Postoperatively, patient-controlled intravenous analgesia devices containing sufentanil (2 μg/kg) and tropisetron (10 mg) were applied to all patients for 2 days.

Outcomes
The primary outcome was the incidence and severity of preoperative hypokalemia.
Normal potassium levels were de ned as blood potassium levels between 3.5 and 5.5 mmol/L. Potassium levels between 3.0 and 3.5 mmol/L, more than 2.5 and less than 3.0 mmol/L, and less than 2.5 mmol/L were considered as slight, moderate, and severe hypokalemia, respectively [8].
The other outcomes included postoperative gastrointestinal function, the time to postoperative rst atus (FFL) and rst feces (FFE) and other complications. Postoperative hospitalization stays and hospitalization costs were also recorded.

Statistical analysis
The Statistical Package for the Social Sciences 21.0 (IBM Corporation, Armonk, NY, USA) was used to perform statistical analyses. Quantitative data were expressed as the mean ± standard deviation, and qualitative data and ordinal data were expressed as absolute frequencies. The one-sample Kolmogorov-Smirnov test was used to test the normality of the quantitative data. According to the distribution of quantitative data, a one-way analysis of variance or non-parametric test was used. The least signi cant difference method was used for further comparison among the groups. The Kruskal-Wallis test was used to compare quantitative data which was not normal distribution. For qualitative data, the chi-square test or Fisher's exact probability test was used to compare differences. Differences were considered signi cant when two-tailed p-values were < 0.05. A justi ed p-value was used for multiple comparisons (the signi cance of p value should be less than 0.05/n, n means comparison frequency).

Results
A total of 181 patients were considered for inclusion. After the exclusion of those who met the exclusion criteria, 132 patients were included and randomly assigned to three groups. Ten patients were excluded from the nal analysis, including nine who transferred to open surgery and one who dropped out. The procedure of inclusion is shown in ow diagram. The nal analysis included 122 participants.
Demographics including age and sex distribution and some surgical aspects showed no signi cant differences among the three groups ( Table 1).
The incidence of preoperative hypokalemia in the case group was signi cantly lower than that in the control and placebo groups (50% vs. 88.1% vs. 77.5%, p < 0.001, Table 2). The severity of hypokalemia in the control and placebo groups was greater than that in the case group (Table 2).
Regarding the avor of the uid, no signi cant differences were observed between the placebo and case groups (7.60 ± 1.75 vs. 6.94 ± 1.83, p = 0.130, Table 2).
Pre-anesthesia thirst was more severe in the control group than in the placebo and case groups (4.00 ± 3.24 vs. 1.16 ± 1.46 vs. 2.18 ± 2.37, p < 0.001, Table 2), without signi cant differences between the placebo and case groups (1.16 ± 1.46 vs. 2.18 ± 2.37, p = 0.063, Table 2). Pre-anesthesia hunger was also more severe in the control group than in the placebo and case groups, without signi cant differences between the placebo and case groups ( Table 2). No regurgitation or aspiration occurred in any of the three groups (Table 2).
No signi cant differences were observed among the three groups regarding time to FFL and FFE and hospital stays (Table 3). No signi cant differences were observed among the three groups considering hypokalemia on Postoperative Day 1 and no patient had hypokalemia on Postoperative Day 2 (Table 3).

Discussion
Although preoperative carbohydrate drinks and prophylactic potassium are recommended for patients, the prevalence of perioperative hypokalemia remains high. In the present study, preoperative carbohydrate drinks containing potassium signi cantly reduced the incidence of preoperative hypokalemia. However, the postoperative well-being and complications showed no signi cant differences among patients with and without preoperative carbohydrate drinks with and without potassium.
A previous study demonstrated that uid and electrolyte management with oral rehydration therapy before surgery is superior to intravenous therapy [11]. Patients with normal uid and electrolyte balance are likely to experience stable induction of anesthesia [12,13]. Preoperative oral carbohydrate intake also improves patient-reported outcomes such as hunger before the induction of anesthesia [14,15]. In line with previous studies, the present study also showed that patients treated with preoperative oral carbohydrates had lower thirst and hunger scores. However, carbohydrates with potassium were not superior to carbohydrates without potassium in terms of prognosis, although they signi cantly reduced the incidence of preoperative hypokalemia without changing the avor of the uid.
Hypokalemia is common in hospitalized patients; approximately 20% of inpatients experience hypokalemia [16]. Perioperative hypokalemia is even higher in patients scheduled for laparoscopic colorectal resection. [8] Bowel preparation combined with systemic antibiotics is effective in reducing the incidence of surgical site infection in colorectal surgery [17]. Therefore, bowel preparation is recommended in the ERAS protocol for patients scheduled for rectal resection. [1] For colonic surgery, bowel preparation is not recommended for routine use and remains controversial [1,18] In the present study, all patients received bowel preparation with polyethylene glycol electrolyte. Although bowel preparations based on polyethylene glycol minimize the loss of potassium, the incidence of hypokalemia increases ve-to nine-fold after bowel preparation even for patients without underlying diseases [7]. In our previous study, the hypokalemia incidence was as high as 70% [8].
The clinical symptoms depend on the severity of hypokalemia. Severe hypokalemia would lead to disorders of various systems including lethal heart arrhythmias, respiratory failure, constipation or intestinal paralysis [19]. Slight or mild hypokalemia typically presents with no clinical symptoms.
Hypokalemia remains a risk factor for postoperative complications and delayed recovery [20]. A previous study recommended prophylactic potassium administration for patients scheduled for gastrointestinal surgery [21]. Potassium can be administrated intravenously or orally. Intravenous potassium is suitable for severe hypokalemia and must be monitored under electrocardiography. The injection of potassium is also a high-alert medication requiring careful monitoring. Furthermore, the injection of potassium can cause vascular irritation and discomfort. Oral potassium supplements are recommended for patients with mild to moderate hypokalemia and those receiving prophylactic potassium. However, the administration of potassium by the oral route is extremely rare and research on the reason for this is scarce [9]. Several factors might contribute to this situation; for example, the avor of potassium is not well tolerated and the importance of oral potassium might not be well known. Adding potassium to preoperative uids might be a useful option. In the present study, preoperative carbohydrate drinks containing potassium were well tolerated and the avor did not differ from that of carbohydrates. Most importantly, oral carbohydrate drinks containing 1 g of potassium administered 2 h before anesthesia signi cantly reduced the incidence of preoperative hypokalemia in patients scheduled for colorectal resection compared with those who drank only carbohydrate uids.
Although preoperative oral carbohydrate drinks and prophylactic potassium have been recommended, their effects on postoperative outcomes remain controversial. Our retrospective study demonstrated that preoperative hypokalemia prolonged postoperative time to FFE but not the time to FFL [8]. Several randomized controlled clinical trials have demonstrated that preoperative oral carbohydrates and oral rehydration solutions do not improve the quality of recovery in either minimally invasive body surface surgery or major abdominal surgery [22][23][24]. The present study showed that oral carbohydrate drinks with and without 1 g of potassium administered 2-3 h before anesthesia induction did not improve the postoperative outcomes. Several factors might account for it. First, most of the hypokalemia caused by fasting or bowel preparations was mild to moderate [7]. Second, optimal preoperative carbohydrate drinks and prophylactic potassium administration are only a part of the ERAS protocol. Third, all the patients with hypokalemia were treated intraoperatively or postoperatively in the present study. No patients suffered hypokalemia on Postoperative Day 1.
In addition, several limitations of the study should be disclosed. First, only 1 g of potassium was administered 2-3 h before surgery, which resulted in half of patients still having hypokalemia. All patients were fasted overnight which was also a limitation. Second, bias may have been present in the patients' reporting of some of the parameters used in the present study. Unlike scintigraphy, some of the parameters, including the time to postoperative FFL and the time to FFE used to determine the gastrointestinal motility, are not gold standards. Third, short hospitalization is one of the endpoints of the ERAS protocol; some patients are reluctant to accept short hospitalization because of limited education level or other unclear reasons [25]. The hospitalization stay in the present study was long despite our best efforts to provide consulting services about the bene cial effects of ERAS. Therefore, the effect of hypokalemia on postoperative well-being should be explored in the near future. Finally, this was a single center study the results of which must be con rmed in a future multi-center study with a large population.

Conclusions
In summary, preoperative carbohydrate drinks containing potassium signi cantly reduced the incidence of preoperative hypokalemia and improved preoperative thirst and hunger, but did not improve postoperative outcomes. However, as part of the ERAS protocol, preoperative carbohydrate drinks containing potassium should be recommended, as early as rst admittance to hospital. Availability of data and materials: All data generated or analysed during this study are included in this published article.
Competing interests: The authors have no con icts of interest. Authors' contributions: YD and FT helped in the following stages of the study: inclusions of participants, data collection, data analysis, and study registration. JH helped in the following stages of the study: inclusions of participants, and prepare the drinks. CG helped in the following stages of the study: data collection. QZ helped in the following stages of the study: study design, draft and critical revision of the manuscript. SZ helped in the following stages of the study: study design, submission to IRB approval, and critical revision of the manuscript. All authors read and approved the nal manuscript.   Notes: Data were presented by mean ± SD and percentages. The one-sample Kolmogorov-Smirnov Test was used to test the normality of the distribution of quantitative data. Normally distributed variables were compared using one-way analysis of variance or non-parametric test, non-normally distributed variables using Kruskal-Wallis test, and categorical data using the chi-squared or Fisher's exact tests; P-value <0.05 was considered signi cant.