Effectiveness of a low-glycaemic-index formula on post-gastrectomy hypoglycaemia in patients with gastric cancer: randomized crossover study

Abstract

Background

Patients undergoing gastrectomy often experience postprandial hypoglycaemia, late dumping syndrome, and night-time hypoglycaemia. However, countermeasures for post-gastrectomy hypoglycaemia rely on the patients’ own efforts. We sought to investigate how post-gastrectomy hypoglycaemia could be nutritionally improved in patients with gastric cancer.

Method

Single-centre prospective, open-labelled, randomized crossover study including patients aged 20–80 years diagnosed with gastric adenocarcinoma, which have undergone total or distal gastrectomy 1–5 years before the study. The patients consumed 100 ml of a low-carbohydrate/high-monounsaturated fatty acid formula orally 30 min after meals and before sleep (400 kcal/day) during the first or second half of a 14-day glucose-monitoring period. The effects of the low-carbohydrate/high-monounsaturated fatty acid formula on the time below range, that is, the percentage of time during which the glucose concentration was <70 mg/dl, and the coefficient of variation (CV) of the glucose concentration when the ideal time below range and CV were set at <5% and ≤36% respectively were assessed. Dumping symptoms were investigated before and after the study.

Results

Thirty-eight patients were included in this study. In patients who had undergone total gastrectomy, the (median) daytime time below range, daytime CV, and night-time time below range remained high at 7.6% ((range) 0.0–45.0), 35.6% ((range) 9.5–50.5), and 10.8% ((range) 0.0–56.3) respectively, even after a long postoperative period. The (median) night-time time below range in patients who had undergone distal gastrectomy and total gastrectomy improved from 3.5% ((range) 0.0–47.9) to 1.4% ((range) 0.0–26.6) (P < 0.001, effect size 0.58) and 10.8% ((range) 0.0–56.3) to 9.4% ((range) 0.0–39.9) (P = 0.078, effect size 0.45) respectively. However, the daytime time below range and CV, as indicators of late dumping syndrome, did not change.

Conclusion

The low-carbohydrate/high-monounsaturated fatty acid formula improved post-gastrectomy night-time hypoglycaemia, but not daytime glycaemic variability or hypoglycaemia. Thus, further investigation of nutritional optimization is required.

Clinical trial registration

Japan Registry of Clinical Trials, jRCT https://jrct.niph.go.jp/, identifier jRCTs s051210200.

Introduction

Gastric cancer is the fifth most common type of cancer and the third leading cause of cancer-related deaths worldwide¹. Despite major advances in medical treatment, including chemotherapy and immunotherapy, surgical resection is the mainstay of curative treatment for gastric cancer. However, many patients experience post-gastrectomy sequelae due to alterations in the gastrointestinal anatomy, recognized as post-gastrectomy syndrome. One of them, late dumping syndrome, usually appears 1–3 h after a meal with general fatigue, cold sweats, and tremors. It is characterized by postprandial hypoglycaemia-related symptoms associated with neuroglycopenia and autonomic/adrenergic reactivity². Therefore, such patients are advised to consume sugar or other sugar-containing products to rapidly counteract hypoglycaemia when hypoglycaemic symptoms appear.

In a previous study, the authors measured interstitial glucose concentration trends using continuous glucose monitoring (CGM) in patients who had undergone gastrectomy. They identified large glycaemic variability and frequent hypoglycaemia 1 month after gastrectomy³. Furthermore, 12 months after gastrectomy, the hypoglycaemia frequency remained high and glycaemic variability was exacerbated⁴. Hypoglycaemia unawareness including night-time hypoglycaemia has also been reported^4,5; therefore, patients are advised to continue diet therapy such as split and small meals regardless of the presence of symptoms. However, no action has been taken against the occurrence of post-gastrectomy hypoglycaemia and glycaemic variability, although these are risk factors for cardiovascular diseases and dementia in older people^6–8. A recent report has revealed the effectiveness of a low-glycaemic-index diet in patients with symptomatic postprandial hypoglycaemia following oesophagogastric cancer surgery using CGM⁵; however, this was a pilot study of only four patients, and no details about dietary intervention and CGM parameters were mentioned. Therefore, the optimal oral nutritional supplement (ONS) type, dosage, and timing for glycaemic control remain unclear. Diabetes-specific nutritional formulas with a low glycaemic index and a low-carbohydrate/high-monounsaturated fatty acid (LC/HM) content (Glucerna®-REX) have also been used to prevent hypoglycaemia and postinfusion hyperglycaemia in patients without diabetes who are fed via a jejunostomy tube⁹. The LC/HM formula can reduce daytime glycaemic variability compared with the standard nutrient formula for tube-fed patients¹⁰. In addition, because it increases glycaemia slowly and maintains blood glucose levels for a relatively long time, it may prevent post-gastrectomy night-time hypoglycaemia. Although the pilot study revealed improvements in postprandial hypoglycaemia with a low-glycaemic-index diet⁵, no prospective interventional studies have been conducted on the influence of an LC/HM formula on post-gastrectomy hypoglycaemia in patients with gastric cancer.

The aim of this study was to examine the preventive effect of an LC/HM formula on post-gastrectomy daytime and night-time hypoglycaemia as well as glycaemic variability, by regulating the intake amount and timing of the LC/HM formula.

Methods

Study design and patients

This prospective, open-labelled, randomized crossover study was conducted in accordance with the principles of the Declaration of Helsinki and approved by the Institutional Committee on Human Studies of Kyoto Prefectural University of Medicine. The clinical and pathological stages of malignancies were determined based on the 14th edition of the Japanese Classification of Gastric Carcinoma¹¹. All participants provided written informed consent, and the study was registered in the Japan Registry of Clinical Trials (jRCT (https://jrct.niph.go.jp/); identifier: jRCTs s051210200). Eligible subjects were patients aged 20–80 years, diagnosed with gastric adenocarcinoma, who had undergone total gastrectomy (TG) or distal gastrectomy (DG) 1–5 years before the study, had a stable and independent daily dietary intake, understood the purpose of this study, and could self-regulate ONS formula intake. After enrolment, patients were randomly allocated using the random function of Microsoft Excel (Microsoft Office® 365, WA, USA). Regarding the reconstruction method, after TG, Roux-en-Y (R-Y) reconstruction without a pouch was performed in all cases. After DG, Billroth I (B-I) reconstruction was the first choice; if the residual stomach was small, R-Y reconstruction was performed. The length of the Roux limb during R-Y reconstruction was 35 cm for DG and 45 cm for TG. The exclusion criteria were as follows: other malignancies or laparotomy within 3 months, a history of recurrence and recurrent treatment, other oral/enteral nutrition supplementation, galactosaemia, a history of hypersensitivity to or diarrhoea due to a low-glycaemic-index formula, a history of antidiabetic medication, glycated haemoglobin A1c (HbA1c) level of >6.5% and/or a fasting blood sugar concentration of >126 mg/dl¹², or refusal to participate in this study.

Oral nutritional supplement

The LC/HM formula was based on a low-glycaemic-index formula (Glucerna®-REX; Abbott Japan, Tokyo, Japan). Unlike standard ONS, the carbohydrate content of the LC/HM formula was limited to 33% and its fat content was enriched to 50% (Table S1), resulting in the blood glucose level increasing slowly.

As patients often report that ONS is difficult to consume because of its unpleasant odour, the LC/HM formula was administered in small, divided doses, and patients were provided with lidded containers with straws to avoid odour inhalation. During the randomly assigned first or second half of the 14-day study period, patients were instructed to consume 100 ml of the LC/HM formula 30 min after each meal and before sleep (400 kcal/day in total). Their daily diet intake rate (subjective percentage of dietary intake based on the standard daily dietary intake before the study), single ONS intake, and intake time were recorded on a predetermined survey sheet, which was returned along with the CGM sensor after the study.

Continuous glucose monitoring

We used a CGM device, FreeStyle Libre® Flash Glucose Monitoring System (Abbott Diabetes Care Inc., Alameda, CA, USA), to document post-gastrectomy glycaemic profiles. A Flash CGM sensor measured the interstitial glucose concentration continuously every 15 min for up to 14 days to record each patient’s glucose trends and patterns. The sensor was placed subcutaneously in the patient’s upper arm immediately after registration. After the 14-day measurement period, the patients removed their CGM sensors and returned them to our department. Data were downloaded and analysed using standard measures of amplitude and timing, including the mean; standard deviation; percentage of time below range (TBR), during which the glucose concentration was <70 mg/dl; time in range (TIR), during which glucose concentration was 70–180 mg/dl; and time above range (TAR), during which glucose concentration was >180 mg/dl. Glycaemic variability was evaluated using the coefficient of variation (CV; SD/mean × 100%) of glucose concentration. The data recorded over the first 10 h after placement of the CGM sensor until the data became stable were deleted. The evaluation criteria of the measures obtained using CGM were based on the consensus report of the American Diabetes Association (ADA) in 2019¹³. According to the ADA Standards of Medical Care in Diabetes in 2022, the CGM targets of TIR, TAR, TBR, and CV were set at >70%, <25%, <5%, and ≤36% respectively¹⁴.

Evaluation of post-gastrectomy dumping symptoms

To evaluate post-gastrectomy syndrome and LC/HM formula side effects, we used the Post-Gastrectomy Syndrome Assessment Scale 37-item (PGSAS-37) questionnaire developed by the Japanese Post-Gastrectomy Syndrome Working Party¹⁵. PGSAS-37 symptom outcomes were recorded on seven symptom scales (oesophageal reflux, abdominal pain, meal-related distress, indigestion, diarrhoea, constipation, and dumping). The dumping symptom score was calculated as the mean value of early dumping abdominal symptoms, early dumping general symptoms, and late dumping symptoms. A questionnaire including the information on body weight (BW) was distributed to all participants, who were instructed to complete it before and after the study and return it to the department along with the CGM sensor.

Sample size calculation

From our previous data⁴, the mean SD of TBR for 2 weeks and 1 year after gastrectomy was 14.04 h for the first half-week and 18.03 h for the second half-week. Therefore, the number of cases in which a difference could be detected between the two groups using a paired t-test with an alpha error of 0.05, beta error of 0.8, 10% difference, and an individual difference index of 1 was 12. The sample size to maintain a statistical power of 0.8 was 24 cases because of a two-sided test in case of a carryover effect. Considering dropouts and data errors of approximately 20%, the minimum number of required cases was 30 (including eight cases of TG as an allocation adjustment factor).

Statistical analysis

Continuous variables are presented as medians with interquartile ranges. All statistical analyses were performed using JMP software, version 13 (SAS Institute, Cary, NC, USA). Outcome measures were performed using the Wilcoxon signed-rank test, with significant differences set at P < 0.05. The Cohen’s d effect size was also calculated. Cohen’s d value reflects the impact of each causal variable. Values between 0.2 and 0.5 denote a small but clinically meaningful effect due to the intervention; values between 0.5 and 0.8 denote a medium effect; and values ≥0.8 indicate a major effect.

Results

Patient characteristics

Figure 1 presents the patient flowchart. From March 2022 to March 2024, patients in the outpatient clinic were consecutively assessed for eligibility, and 72 patients were found to be eligible. Among these, 22 did not provide consent to participate in the study, which resulted in 50 patients being enrolled. Four patients who could not consume the LC/HM formula because of diarrhoea or personal preference, four whose CGM sensor was removed, three with data containing errors, and one who withdrew consent were excluded; therefore, 38 patients (21 in Sequence 1 and 17 in Sequence 2) were included in the analysis. Patients who had received adjuvant chemotherapy were included, although all cases were completed within 1 year after surgery. No patients received chemotherapy during the study period.

Table 1 shows the clinical background characteristics of the 38 patients. Initially, the patients’ (median) BMI was 20.1 kg/m² (i.q.r 16.5–27.1), and (median) HbA1c was 6.0% (i.q.r. 5.3–6.7). Eight had undergone TG and 30 DG respectively. The number of patients with pathological stage I/II/III/IV was 23/6/9/0, and 12 of the 15 patients with pathological stage II/III received adjuvant chemotherapy for 1 year. The median period from surgery to the study was 3.6 years (i.q.r. 1.0–5.0).

Compliance with LC/HM formula intake

Our dietary intervention involving the intake of the LC/HM formula included: a small amount per intake; timing of administration; providing patients with a lidded container for drinking; and monitoring the effect of the LC/HM formula on daily diet using a predetermined survey sheet. After registration and random assignment, patients with an LC/HM formula intake rate of <50% due to preference (n = 2) and diarrhoea and abdominal pain (n = 2) were excluded as ‘Not accepting the LC/HM formula’. The reason for one patient withdrawing consent was also diarrhoea. In the study cohort (n = 38), adverse events caused by the LC/HM formula included diarrhoea and abdominal pain (n = 3), nausea (n = 1), and abdominal fullness (n = 3), although the seven patients (18.4%) who experienced these events could continue the study. The mean(s.d.) LC/HM formula intake rate was 91.2(13.4)%, and the daily diet intake rate during the intervention period was 96.7(12.7)%. Thus, compliance with the LC/HM formula was favourable, and the impact on daily diet was minimal. No adverse events due to CGM were observed.

Effect of LC/HM formula on post-gastrectomy hypoglycaemia and glycaemic variability

The mean(s.d.) CGM data usage rate was 95.1(4.4)% (excluding the first 10 h of measurement of every patient). As no carry-over effect was observed in the CGM parameters in Sequence 1 (first half of the LC/HM formula intervention group) compared with Sequence 2 data, the analysis was a two-group comparison between the intervention of the LC/HM formula and the non-intervention of the daily diet only (Fig. S1).

When each parameter of the CGM in patients who had undergone DG was analysed, night-time TBR showed a significant improvement (P < 0.001) with a medium effect size (Cohen’s d 0.58). Accordingly, the TIR increased (P < 0.001, Cohen’s d 0.60). By contrast, the mean night-time glucose concentration increased as a result of the intervention (P = 0.035), but the effect size was small (Cohen’s d 0.19). However, the daytime TBR and CV were below the target line and did not change as a result of the intervention (Table 2). In the comparison of B-I and R-Y procedures in patients who underwent DG, the night-time TBR was higher in R-Y than that in B-I reconstruction, and the improvement effect with the intervention was equivalent and medium (B-I: 2.3% to 0.0% [median], P = 0.002, Cohen’s d 0.58; R-Y: 10.5% to 3.8% [median], P = 0.070, Cohen’s d 0.61; Tables S2, S3).

In patients who had undergone TG, the night-time TBR without the intervention was higher than that in patients who had undergone DG, and it showed a non-significant improvement with the intervention (P = 0.078), with a medium effect size (Cohen’s d 0.45). Although the night-time CV in patients who had undergone TG was not very high initially, a significant decrease was observed (P = 0.039, Cohen’s d 0.45). The daytime TBR without the intervention was 7.6% (median), which was higher than that of patients who had undergone DG, but no improvement was observed as a result of the intervention. Although the daytime CV without the intervention was high, no improvement was observed as a result of the intervention (Table 3).

Effect of LC/HM formula on post-gastrectomy dumping symptoms

The authors also sought to evaluate dumping symptoms before and after the study. Five patients of 30 (16.7%) had undergone DG and 2 of 8 (25%) had undergone TG, with a dumping score of the PGSAS-37 of >3 before the study, but their scores did not improve after the study. The median dumping score of the 38 patients was 1.7 (i.q.r. 1.0–2.3) before the study, and it did not change after the study. Dumping score represents the average scores for the three component items—early dumping general symptoms, early dumping abdominal symptoms, and late dumping symptoms—each of which further consists of early and late specific symptoms. However, no changes were observed before and after the study in any of the specific items. Regarding LC/HM formula-related adverse events, no increase in diarrhoea and abdominal pain was observed. No oesophageal reflux due to LC/HM formula intake before sleep was observed after the study. Given the limited study period of 2 weeks, no changes in BW were observed between before and after the study (Table 4).

Discussion

In this prospective crossover study, we demonstrated that the LC/HM formula improved post-gastrectomy night-time hypoglycaemia, although it did not decrease daytime hypoglycaemia or glycaemic variability, the so-called late dumping syndrome. Hypoglycaemia and glycaemic variability are risk factors for cardiovascular events and dementia^6–8. Specifically, it has been reported that night-time hypoglycaemia may cause arrhythmia and coronary artery disease due to increased vagal counteraction after sympathetic neural activation^16,17. Recently, a negative relationship between night-time hypoglycaemia and sleep quality has been suggested¹⁸. Therefore, our results may lead to the prevention of night-time cardiovascular events and improve the quality of life. In a pilot study using a CGM system, Fanning et al. reported that a low-glycaemic-index diet improved postprandial hypoglycaemia in four patients who had undergone surgery for oesophagogastric cancer⁵. To the best of our knowledge, ours is the first prospective interventional trial using ONS that targeted post-gastrectomy hypoglycaemia and glycaemic variability in patients with gastric cancer.

Compliance with ONS is often an issue in clinical studies. The dietary intervention used in this study (dividing the LC/HM formula into smaller volumes and using lidded containers with straws) resulted in almost all patients who had undergone gastrectomy consuming 400 ml/day of the LC/HM formula. Reasons why patients cannot consume ONS include preference, odour, a heavy stomach sensation, and diarrhoea. In a clinical trial that administered 600 kcal ONS/day and obtained negative results, the authors argued that consuming ONS before meals could have suppressed daily dietary intake and resulted in poor compliance despite the instruction provided to the patients. They also suggested that forced administration, such as enteral feeding, may be necessary¹⁹. In addition, in the authors' previous study, in which patients were provided with 600 kcal ONS in bulk and could select their preferred intake method, the intake rate was <20% (unpublished data). Therefore, we divided the LC/HM formula intake into four daily doses, thereby reducing the volume of each ingestion to 100 ml. An RCT reported that higher frequencies of lower ONS volumes increased compliance in patients requiring ONS²⁰. The smaller volume of the LC/HM formula may have thus contributed to increasing compliance. Additionally, the lidded containers with straws helped to eliminate odours. These methods, alongside the timing of intake after rather than before meals, enabled patients to consume the LC/HM formula as scheduled and minimized its impact on their daily diet. This adherence to the study intervention reflected in improvements of night-time TBR but not in the CV and daytime TBR.

BW loss is a common post-gastrectomy syndrome that poses substantial problems to patients who have undergone gastrectomy for gastric cancer. It negatively impacts glycaemic control due to the decrease in pancreatic secretory function associated with continuous pancreatic atrophy²¹. Differences in BMI at the time of the study and the period since surgery may have also affected the glycaemic trend in our cohort. Herein, patients with a low BMI had a significantly higher CV, although no correlations were found between BW loss, the length of the post-gastrectomy period, BMI, and night-time TBR improvement due to the intervention (Fig. S2).

Night-time hypoglycaemia was particularly evident in patients who had undergone TG, even after long postoperative periods, and split, small volumes of the LC/HM formula showed equal preventive effects against night-time hypoglycaemia in patients who had undergone TG and DG. One of the reasons for this successful result is that the patients were able to consume 100 ml (100 kcal) of the LC/HM formula before sleep, despite concerns about oesophageal reflux. The PGSAS-37 showed no increase in oesophageal reflux scores after the study. In addition, because the interval between the intake after dinner and sleep was short, it is possible that the intake after dinner also had a positive effect on night-time hypoglycaemia. Furthermore, the effect on night-time hypoglycaemia may not only be due to the direct glucose-elevating effect of the LC/HM formula but also due to counterregulation through glycogen storage and glycolysis. When caloric intake is increased with an ONS such as the LC/HM formula that contains 30% carbohydrate, 50% fat, and 20% protein, glycogen storage in the liver is usually affected within a few hours. Studies have shown that glycogen levels in the liver peak around 5–6 h after consuming a mixed meal, with approximately 19% of the ingested carbohydrates being stored as glycogen²². Moreover, when an individual resumes a carbohydrate-rich diet after fasting, glycogen levels in the liver increase rapidly and reach very high levels within a day²³. Therefore, upon ingesting the 400 kcal of the LC/HM formula used in this study, glycogen storage in the liver began within a few hours after ingestion, peaked in 5–6 h, and was maintained in response to subsequent energy expenditure and meals. During hypoglycaemia, counterregulation to maintain blood glucose concentration must occur through an autonomous nerve response, followed by glycolysis^24,25. In this study, patients were able to consume approximately 400 kcal of the LC/HM formula per day without reducing their daily dietary intake, which could increase glycogen storage in the liver, resulting in the counterregulation of night-time hypoglycaemia using glycogen storage and glycolysis.

The difference in the reconstruction method deployed after DG (B-I versus R-Y) is an important issue when considering post-gastrectomy syndrome. In the present study, the TBR and CV for B-I and R-Y reconstructions were comparable and below the target line in patients who had undergone DG 1 year or more after surgery and showed no difference due to the intervention. In bariatric surgery, R-Y is considered to have a greater effect than B-I reconstruction on lowering blood glucose, BW loss, and diabetes remission; however, R-Y reconstruction has also been observed to cause greater glycaemic variability^26,27. In our previous study on surgery for gastric cancer, R-Y reconstruction was associated with a longer TBR and larger glycaemic variability than was B-I reconstruction 1 month after DG²⁸; however, the difference in TBR and glycaemic variability disappeared [data not shown]. Differences in glycaemic trends depending on the reconstruction method may not need to be considered in patients who had undergone DG >1 year earlier.

By contrast, although the daytime CV was relatively high in patients who had undergone TG and DG, and the daytime TBR was favourable in patients who had undergone DG but still high in those who had undergone TG, the LC/HM formula could not improve the daytime TBR and CV. Various reasons may underlie this unsuccessful result, such as the small LC/HM formula amount, ingestion timing, short study duration, uncontrolled standard daily diet, and limitations of dietary intervention. The international consensus on the diagnosis and management of late dumping syndrome recommends dietary adjustment as the first treatment step. Additionally, acarbose is effective for late dumping syndrome symptoms, whereas somatostatin analogues are preferred for patients who do not respond to dietary adjustments and acarbose²⁹. With the increasing number of patients undergoing bariatric surgery, the incidence of dumping syndrome has also increased in recent years. Surgeons should recognize the syndrome and understand its management. In early and late dumping syndrome, the incretin hormone glucagon-like peptide 1 (GLP-1) is thought to be a key factor. Acarbose reduces the rapid glucose absorption associated with hyperglycaemia and GLP-1 secretion and subsequently suppresses excessive insulin release. Somatostatin analogues delay gastric emptying and inhibit insulin and GLP-1 secretion as well as postprandial vasodilation. These analogues are, therefore, potentially beneficial in early and late dumping syndrome. This study, however, included patients who had undergone surgery for gastric cancer, and such cases largely differ from patients who undergo bariatric surgery. Patients with gastric cancer are older, and although their BMI and HbA1c are not high, their food intake typically decreases after gastrectomy— thus, BW loss is not the goal but rather the problem. Therefore, the use of acarbose or somatostatin analogues in patients who undergo surgery for gastric cancer involves the risk of hypoglycaemia as an adverse effect^30,31. Furthermore, the use of somatostatin analogues is impractical because it requires multiple injections daily. The authors conclude that CGM should first be used to monitor glycaemic trends when treating late dumping syndrome with hypoglycaemic agents.

The PGSAS-37 median dumping score was low and remained unchanged after the study, indicating that many patients who had undergone gastrectomy did not report dumping symptoms. However, daytime hypoglycaemia in patients who had undergone TG occurred almost twice as often, and night-time hypoglycaemia 3–4 times more often than the ADA’s recommended target of <5%. These results suggest that patients who had undergone TG were unaware of their hypoglycaemia for >1 year after surgery. As the effect of the LC/HM formula on postprandial late dumping syndrome is unsatisfactory, surgeons need to recognize the presence of post-gastrectomy hypoglycaemia unawareness in patients.

This study had certain limitations. First, it was a single-centre study and the proportion of patients who had undergone TG was low. Second, in this study, patients with diabetes were excluded, although a significant number of patients with glucose intolerance in actual clinical practice often have more problems with perioperative glycaemic control. In the authors' previous study using CGM, patients with diabetes rarely developed hypoglycaemia after gastrectomy (data not shown); therefore, the study criteria excluded such patients. However, this previous result does not mean that patients with glucose intolerance who underwent gastrectomy are safe under glycaemic conditions, rather that glycaemic variability and hyperglycaemia of patients with glucose intolerance were higher than patients without glucose intolerance (data not shown). Thus, countermeasures may be necessary to prevent hyperglycaemia and glycaemic fluctuations. Third, the patients’ daily dietary intake was not controlled during the study period, as older patients with gastric cancer find it challenging to document in detail their daily dietary intake. Instead, to estimate the patients’ daily dietary intake, the authors requested them to report their dietary intake rate during the study period compared with their daily diet before the study on a predetermined survey sheet. This resulted in a subjective and relative evaluation of dietary intake status, and an objective evaluation was not possible. Fourth, there was no washout period in this study. However, no carry-over effects were observed. Fifth, regarding CGM accuracy, interstitial and blood glucose concentrations have been reported to be almost correlated, albeit with a slight time lag³². However, unusual patterns of daily glycaemic trends do sometimes occur (such as obvious missing needle or no fluctuations), and because there is no clear standard for determining sensing errors, the authors excluded such data from their analysis. Sixth, given that the aim of this study was to improve post-gastrectomy hypoglycaemia and glycaemic variability using the LC/HM formula, we did not present data to support the preventive effect of the ONS on cardiovascular events in patients who had undergone gastrectomy. However, our previous study had revealed that glycaemic variability is correlated with cardiovascular risk scores³³. We have thus commenced an observational study following patients who have undergone gastrectomy and who had their glycaemic profile measured using CGM in our studies.

In summary, we demonstrated that the LC/HM formula improved post-gastrectomy night-time hypoglycaemia but not daytime hypoglycaemia and glycaemic variability. Great glycaemic variability and prolonged night-time hypoglycaemia persisted, particularly in patients who had undergone TG, even after long postoperative periods. Further investigation of nutritional support or the development of new treatments for post-gastrectomy hypoglycaemia is required.

Funding

This study was supported in part by the Japan Society for the Promotion of Science (no. 18H02882).

Acknowledgements

The authors thank G. Horiguchi and S. Teramukai (Department of Biostatistics, Kyoto Prefectural University of Medicine) for advice on sample size calculation and formal statistical analysis. They also thank Editage (www.editage.cn) for English language editing.

Supplementary material

Supplementary material is available at BJS Open online.

Disclosure

The authors declare no conflict of interest.

Data availability

Anonymized trial data will be available from the corresponding author on reasonable request.

Author contributions

Takeshi Kubota (Conceptualization, Data curation, Formal analysis, Methodology, Writing—original draft), Takuma Ohashi (Investigation), Keiji Nishibeppu (Investigation), Kazuya Takabatake (Investigation), Hiroyuki Inoue (Investigation), Yudai Nakabayashi (Investigation), Hirotaka Konishi (Conceptualization, Methodology, Writing—review & editing), Atsushi Shiozaki (Conceptualization, Methodology, Project administration), Hitoshi Fujiwara (Data curation, Formal analysis), Emi Ushigome (Data curation, Supervision, Validation, Writing—review & editing), Michiaki Fukui (Supervision), and Eigo Otsuji (Funding acquisition, Project administration, Supervision)

References