Prediction of antidiabetic effect after gastrectomy with Roux-en-Y reconstruction in patients with gastric cancer and type 2 diabetes

Abstract

This study investigated the antidiabetic outcomes after gastrectomy with long-limb RY reconstruction (LRYR) and the prognostic factors for remission after 1 year in patients with type 2 diabetes (T2DM) and gastric cancer. In 25 Koreans with T2DM and gastric cancer, plasma glucose and insulin levels were measured during a 75 g oral glucose tolerance test, before and 1 week after gastrectomy with LRYR. Patients were examined after 1 year and we defined glycemic control as “remission” when the HbA1c level after 1 year was <6.0% without medication. One year after surgery, 12 patients achieved HbA1c < 6.0% without medication. Among the preoperative indices, the duration of diabetes was shorter in the remission group than that in the non-remission group (median 2.0 [0–6.5] years vs 7.0 [4.5–10.0] years, P = .023). At 1 week after surgery, significant improvements in fasting, 30 minutes, 60 minutes, 90 minutes stimulated glucose levels and insulin resistance (HOMA-IR and Matsuda index) were found only in the remission group. The multivariable logistic regression analysis results showed that higher 30 minutes stimulated glucose level and HOMA-IR index at 1 week after surgery were independent factors for lower odds of 1-year diabetes remission. Shorter duration of diabetes and early postoperative improvements in 30 minutes stimulated glucose level and HOMA-IR were important determinants of long-term antidiabetic outcomes after gastrectomy with LRYR in patients with T2DM and gastric cancer.

1. Introduction

Roux-en-Y (RY) gastric bypass (RYGB) surgery is a useful treatment option for uncontrolled type 2 diabetes mellitus (T2DM) accompanied by severe obesity.^[1] Various hypotheses have been proposed to explain the antidiabetic effects of RYGB surgery. The antidiabetic effect is considered to be mainly because of the weight loss caused by caloric restriction that occurs after surgery.^[2,3] However, T2DM remission often occurs within a few weeks before significant weight loss after surgery, and some studies suggest that antidiabetic effects may be related to factors other than weight loss after surgery.^[1] RYGB surgery induces a decrease in ghrelin, which is produced by the stomach and duodenum, resulting in appetite reduction and antidiabetic effects.^[4] The lower intestinal hypothesis is also suggested to be a possible mechanism. When unabsorbed nutrients reach the distal intestine, it can lead to increased secretion of glucagon-like peptide-1 (GLP-1) and improvement of glucose metabolism.^[5]

The method of gastric cancer surgery is similar to that of gastric bypass surgery, and the metabolic effects of gastric cancer surgery have also been investigated. The benefits of T2DM improvement after gastric cancer surgery were observed in patients without severe obesity (body mass index [BMI] < 40 kg/m²).^[6] Improvement of T2DM in patients undergoing gastrectomy with RY reconstruction or Billroth II reconstruction was also observed in patients with gastric cancer with BMI < 30 kg/m².^[7] However, not all patients who undergo gastrectomy with RY reconstruction achieve complete T2DM remission.^[8,9] The preoperative duration of diabetes was one of the important factors affecting postoperative T2DM remission in patients with gastric cancer undergoing gastrectomy with RY, Billroth I, or Billroth II reconstruction.^[7] Kim et al reported that the BMI reduction ratio was the most important factor for antidiabetic effects in patients with gastric cancer undergoing gastrectomy with RY, Billroth I, or Billroth II reconstruction.^[9]

The underlying mechanisms of the antidiabetic effect have not been clearly elucidated, and there are still many unknowns about the characteristics of the patient group that can result in remission of T2DM after surgery. In addition, only a few studies have examined the relation between short-term outcome and long-term outcome in patients with gastric cancer and T2DM undergoing gastrectomy with RY reconstruction. Therefore, we investigated the short- and long-term antidiabetic outcomes and prognostic factors for long-term antidiabetic outcomes after gastrectomy with long-limb RY reconstruction (LRYR) in patients with gastric cancer and T2DM.

2. Materials and Methods

By reviewing patient documents using electronic medical records at Inha University Hospital, we investigated patients aged ≥ 19 years with T2DM who had undergone gastrectomy with LRYR surgery for primary gastric cancer between January 2010 and November 2019. The patients underwent LRYR with 80 to 100 cm of the biliopancreatic limb and 80 to 100 cm of the Roux limb. T2DM was defined according to the American Diabetes Association definition.^[10] Patients were excluded if they fulfilled any one of the following criteria: 1) were <19 years of age, 2) had type 1 diabetes, 3) did not have 1 year of follow-up data, 4) received chemotherapy after the surgery, or 5) did not fully satisfy the inclusion criteria. This retrospective review of medical records was approved by the Institutional Review Board of Inha University Hospital (2020-12-020-000), and the requirement for informed consent was waived.

Age, sex, weight, height, duration of diabetes, and medications before surgery were recorded. BMI was calculated as the weight divided by height squared (kg/m²). Following an overnight fast (≥8 hours), blood samples were collected before (0 minute, designated as fasting) and after (30, 60, 90, and 120 minutes, designated as stimulated) ingestion of 75 g glucose (Diasol-S 75 soln; Taejoon Pharm co, Yongsan-gu, Seoul, Republic of Korea) to measure hemoglobin A1c (HbA1c), basal and stimulated glucose/insulin/C-peptide, and other chemistry profiles. We calculated the homeostasis model assessment of insulin resistance (HOMA-IR)^[11] and the Matsuda index^[12] to assess insulin sensitivity, and the insulinogenic index (IGI) and homeostasis model assessment of β-cell function (HOMA-β) to assess insulin secretion before and 1 week after the surgery. HOMA-IR was calculated using the following formula: HOMA-IR = fasting glucose (mmol/L) × fasting insulin (μU/mL)/22.5. The Matsuda index was calculated using the following formula: Matsuda index = $\text{1}0,000/\sqrt{(\text{fasting insulin}\times \text{fasting glucose})\times (\text{Mean insulin}\times \text{Mean glucose})}$ (insulin expressed as μU/mL, glucose expressed as mg/dL). Elevated values of HOMA-IR and reduced values of the Matsuda index indicate the presence of insulin resistance.^[13–15] HOMA-β was calculated as (20× fasting insulin [μU/mL])/(fasting glucose [mmol/L] −3.5)%.^[11] The IGI was calculated as follows: (30 minutes insulin [μU/mL] − fasting insulin [μU/mL])/(30 minutes glucose [mg/dL] − fasting glucose [mg/dL]), and was used to estimate early phase insulin secretion.^[16,17] We defined the remission of T2DM as HbA1c levels <6.0% without drug treatment, and patients satisfying this definition were classified into the remission group.^[18,19]

Data were presented as medians (interquartile range) for continuous variables and as numbers or percentages for categorical variables. For comparison between groups, the χ² test was used for categorical variables, and Mann–Whitney U test was used for continuous variables. A mixed model was used that considered the group as a fixed effect and time as a random effect for group comparison of the 1-week changes. Multiple logistic regression analysis was performed to adjust for duration of diabetes. The adjusted odds ratios (aORs) with its 95% confidence intervals were used as measures of association. All statistical analyses were performed using SPSS ver. 26 (SPSS Inc., Chicago, IL, USA). The results were considered statistically significant if the P-value was <.05.

3. Results

We investigated 25 patients (21 male and 4 female) with T2DM who had undergone gastrectomy with LRYR for gastric cancer. The baseline characteristics are shown in Table 1. One year after surgery, 12 patients (48%) achieved HbA1c <6.0% without medication (diabetes remission group). When we compared the baseline characteristics between the remission and non-remission groups, the duration of diabetes in the remission group was shorter than that in the non-remission group (2.0 [0–6.5] years vs 7.0 [4.5–10.0] years, P = .023). The 2 groups showed no differences in HbA1c or HOMA-IR. There were no statistically significant differences in other preoperative factors, including age, BMI, medication history before surgery, current status of alcohol and smoking, fasting plasma glucose levels, and stimulated plasma glucose levels at each time point (Table 1 and Figure 1A). However, the beta cell function preservation calculated by the IGI (IGI > 0.095 [median value of the study population]) was significantly higher in the diabetes remission group than in the non-remission group (72.7% vs 30.8%, P = .041, Table 1).

Table 1.

Baseline characteristics of the patients in the remission and non-remission groups.

	Non-remission (n = 13)	Remission (n = 12)	P-value
Age, years	60.0 (53.0–64.0)	60.5 (58.0–64.0)	.763
Female sex, n (%)	3 (23.1)	1 (8.3)	.315
BMI, kg/m2	26.0 (24.0–27.5)	26.0 (24.3–28.8)	.603
Duration of diabetes (years)	7.0 (4.5–10.0)	2.0 (0.0–6.5)	.023*
Medication history before surgery	10 (76.9)	9 (75.0)	.910
Alcohol, current	7 (53.8)	4 (33.3)	.302
Smoking, current	8 (61.5)	3 (25.0)	.066
Glucose, fasting, mg/dL	109.0 (78.0–161.5)	108.0 (81.8–124.5)	.849
Glucose, stimulated (30 minutes), mg/dL	176.0 (159.5–211.0)	185.0 (161.3–223.8)	.550
Glucose, stimulated (60 minutes), mg/dL	251.0 (210.0–279.5)	232.0 (211.0–289.3)	.913
Glucose, stimulated (90 minutes), mg/dL	267.0 (251.5–285.0)	255.0 (237.5–286.5)	.341
Glucose, stimulated (120 minutes), mg/dL	275.0 (241.0–321.5)	241.0 (178.0–296.3)	.142
Insulin, fasting, μIU/mL	9.3 (5.9–14.2)	9.1 (7.3–15.3)	.744
Insulin, stimulated (30 minutes), μIU/mL	14.7 (11.2–37.5)	21.4 (13.7–32.9)	.434
Insulin, stimulated (60 minutes), μIU/mL	21.4 (12.7–48.8)	30.2 (17.6–48.3)	.577
Insulin, stimulated (90 minutes), μIU/mL	23.9 (17.5–52.8)	40.1 (21.6–66.7)	.231
Insulin, stimulated (120 minutes), μIU/mL	21.9 (15.4–54.9)	29.3 (22.4–43.6)	.586
HbA1c, %	6.9 (6.5–8.8)	6.7 (6.3–7.7)	.384
HOMA-IR	2.2 (1.2–4.3)	2.7 (1.7–3.7)	.744
HOMA-IR ≥ 2.49 (M), n (%)	6 (46.2)	7 (58.3)	.543
Matsuda index	4.7 (2.8–8.8)	4.0 (2.9–5.7)	.642
Matsuda index ≥ 4.0 (M), n (%)	7 (53.8)	5 (50.0)	.855
HOMA-β	70.3 (47.7–212.6)	65.0 (47.4–144.7)	.773
HOMA-β ≥ 70 (M), n (%)	6 (50.0)	6 (50.0)	>.99
Insulinogenic index	0.08 (0.03–0.11)	0.18 (0.07–0.28)	.117
Insulinogenic index ≥ 0.095 (M), n (%)	4 (30.8)	8 (72.7)	.041*

Values are presented as medians (interquartile range) or numbers (%).

Remission: HbA1c level after 1 year was <6.0% without medication.

BMI = body mass index, HbA1c = hemoglobin A1c, HOMA-β = homeostasis model assessment of β-cell function, HOMA-IR = homeostasis model assessment: estimated insulin resistance, (M) = median value.

^* Statistically significant remission vs non-remission by Mann–Whitney U test.

Figure 1.

Effect of gastrectomy with long-limb RY reconstruction on glycemic control. Plasma glucose levels during the oral glucose tolerance test before and 1 week after surgery. Empty square: non-remission group, before surgery; filled square: non-remission group, after 1 week; empty circle: remission group, before surgery; filled circle: remission group, after 1 week. (A) Comparison of glucose levels between the 2 groups during the oral glucose tolerance test before surgery. (B) Comparison of glucose levels before and after 1 week in the non-remission group. (C) Comparison of glucose levels between the 2 groups during the oral glucose tolerance test after 1 week. (D) Comparison of glucose levels before vs after 1 week in the remission group. RY, Roux-en-Y

At 1 week after surgery, fasting plasma glucose levels were improved in the remission group (108 [82–125] mg/dL to 89 [82–100] mg/dL, P = .017). In the remission group, stimulated (30 minutes, 60 minutes, 90 minutes, and 120 minutes after oral glucose tolerance test) glucose levels were also significantly decreased (all P < .05). In the non-remission group, there was no statistically significant difference in the fasting plasma glucose levels before and 1 week after surgery (109.0 [78.0–161.5] mg/dL to 117.0 [93.0–144.5] mg/dL, P = .889). While there were no significant decreases in the 30 minutes, 60 minutes, and 90 minutes stimulated glucose levels, significant improvement was observed in the 120-minute stimulated glucose level (275.0 [241.0–321.5] mg/dL to 223.0 [189.5–276.5] mg/dL, P = .003). Difference between groups was significant only in 30 minutes stimulated glucose level (P = .01, Table 2). In the comparison between the groups, the 90-minute and 120-minute stimulated glucose levels after surgery in the remission group were significantly lower than those in the non-remission group (Figure 1C). Overall, there was a significant improvement in the plasma glucose levels over the entire test period in the remission group, but not in the non-remission group. (Figure 1B and 1D).

Table 2.

Biochemical data before and at 1 week after surgery in the remission and non-remission groups.

	Non-remission (n = 13)		P-value	Remission (n = 12)		P-value	Difference between groups
	Before	After 1 week		Before	After 1 week		P-value
Glucose, fasting, mg/dL	109.0 (78.0–161.5)	117.0 (93.0–144.5)	.889	108.0 (81.8–124.5)	89.0 (81.8–99.5)	.017*	.208
Glucose, stimulated (30 minutes), mg/dL	176.0 (159.5–211.0)	206.0 (173.5–231.0)	.133	185.0 (161.3–223.8)	164.5 (137.0–179.5)	.005*	.010
Glucose, stimulated (60 minutes), mg/dL	251.0 (210.0–279.5)	254.0 (210.5–281.0)	.861	232.0 (211.0–289.3)	220.5 (152.0–236.3)	.015*	.070
Glucose, stimulated (90 minutes), mg/dL	267.0 (251.5–285.0)	252.0 (216.5–297.5)	.345	255.0 (237.5–286.5)	211.0 (167.3–242.8)	.008*	.285
Glucose, stimulated (120 minutes), mg/dL	275.0 (241.0–321.5)	223.0 (189.5–276.5)	.003*	241.0 (178.0–296.3)	178.0 (146.8–237.3)	.015*	.888
Insulin, fasting, μIU/mL	9.3 (5.9–14.2)	5.8 (4.8–7.4)	.019*	9.1 (7.3–15.3)	4.8 (2.5–6.0)	.002*	.969
Insulin, stimulated (30 minutes), μIU/mL	14.7 (11.2–37.5)	9.4 (7.7–17.6)	.071	21.4 (13.7–32.9)	12.1 (7.9–16.8)	.016*	.972
Insulin, stimulated (60 minutes), μIU/mL	21.4 (12.7–48.8)	18.0 (10.5–24.6)	.279	30.2 (17.6–48.3)	13.0 (7.5–35.0)	.594	.748
Insulin, stimulated (90 minutes), μIU/mL	23.9 (17.5–52.8)	28.7 (12.3–35.0)	.311	40.1 (21.6–66.7)	19.7 (16.6–30.8)	.346	.943
Insulin, stimulated (120 minutes), μIU/mL	21.9 (15.4–54.9)	17.3 (9.7–36.1)	.173	29.3 (22.4–43.6)	15.1 (11.6–36.8)	.158	.945
HOMA-IR	2.2 (1.2–4.3)	1.8 (1.0–2.5)	.184	2.7 (1.7–3.7)	1.1 (0.5–1.3)	.002*	.793
Matsuda index	4.7 (2.8–8.8)	6.8 (4.6–10.7)	.530	4.0 (2.9–5.7)	7.9 (7.1–14.9)	.008*	.055
HOMA-β	70.3 (47.7–212.6)	36.2 (22.5–66.6)	.023*	65.0 (47.4–144.7)	62.2 (36.5–111.6)	.308	.143
Insulinogenic index	0.08 (0.03–0.11)	0.07 (0.03–0.13)	.529	0.18 (0.07–0.28)	0.13 (0.09–0.18)	.533	.425
Body weight, kg (n = 12)	71.0 (64.8–82.1)	68.1 (60.6–80.2)	.004*	71.6 (64.3–78.5)	68.8 (60.1–75.5)	.008*	.657
BMI (n = 12)	26.2 (24.0–28.0)	24.5 (22.4–27.0)	.004*	26.0 (24.4–28.4)	25.0 (22.8–27.6)	.006*	.675

Values are presented as medians (interquartile range).

Remission: HbA1c level after 1 year was <6.0% without medication.

BMI = body mass index, HOMA-IR = homeostasis model assessment: estimated insulin resistance, HOMA-β = homeostasis model assessment of β-cell function.

^* Statistically significant preoperative data vs 1 week after surgery data, as shown by Mann–Whitney U test.

Postoperative (1 week after surgery) weight and BMI showed a statistically significant decrease in both groups. The insulin resistance index (HOMA-IR) and sensitivity index (Matsuda index) significantly improved 1 week after surgery only in the remission group (2.7 [1.7–3.7] to 1.1 [0.5–1.3] for HOMA-IR, P = .002, and 4.0 [2.9–5.7] to 7.9 [7.1–14.9] for the Matsuda index, P = .008), however, there was no statistically significant improvement in the non-remission group. In both groups, insulin secretion indices, including IGI or the HOMA-β, did not show any significant improvement before and after the surgery (Table 2, supplementary figure 1, Supplemental Digital Content, http://links.lww.com/MD/H145).

The association between each baseline characteristics, metabolic parameters at baseline and at 1 week after surgery, and diabetes remission at the 1-year visit is shown in Table 3. In the univariate analysis, duration of diabetes, 90-minute, 120-minute stimulated glucose levels before surgery, and some metabolic parameters 1 week after surgery including fasting, 30-minute, 60-minute, 90-minute stimulated glucose levels, and HOMA-IR were associated with diabetes remission (all P < .05). In the multivariable analysis, duration of diabetes was adjusted. Only higher 30 minutes stimulated glucose level (aOR = 0.96, P = .044) and higher HOMA-IR index (aOR = 0.09, P = .041) at 1 week after surgery were independent factors for lower odds of 1 year diabetes remission. The parameters related to the preoperative examination were no longer significant.

Table 3.

The effect of baseline factors on the probability of diabetes remission at the 1-year visit.

Variables	Univariable Model	P-value	Multivariable Model	P-value
	OR (95% CI)		OR (95% CI)
Duration of diabetes (years)	0.75 (0.58–0.96)	.024	Adjusted*
Age (years)	1.02 (0.90–1.17)	.717	1.00 (0.86–1.16)	.957
Female sex	0.30 (0.03–3.41)	.334	0.37 (0.03–5.20)	.458
Medication history before surgery	0.55 (0.07–4.01)	.551	1.29 (0.11–15.28)	.841
Metabolic parameters, before surgery
HbA1c, %	0.70 (0.37–1.34)	.287	0.81 (0.40–1.64)	.553
BMI, kg/m2	1.10 (0.83–1.46)	.501	1.31 (0.91–1.90)	.147
Glucose, fasting, mg/dL	1.00 (0.98–1.02)	.797	1.00 (0.98–1.02)	.994
Glucose, stimulated (30 minutes), mg/dL	0.99 (0.98–1.00)	.170	1.01 (0.99–1.03)	.519
Glucose, stimulated (60 minutes), mg/dL	0.99 (0.98–1.00)	.157	1.01 (0.99–1.03)	.394
Glucose, stimulated (90 minutes), mg/dL	0.99 (0.98–1.00)	.047	1.00 (0.98–1.02)	.998
Glucose, stimulated (120 minutes), mg/dL	0.99 (0.98–1.00)	.033	1.00 (0.98–1.01)	.500
HOMA-IR	0.94 (0.78–1.13)	.496	0.94 (0.78–1.13)	.511
HOMA-β	1.00 (0.99–1.00)	.474	1.00 (0.99–1.00)	.609
Insulinogenic index ≥ 0.095 (M)	6.00 (1.02–35.37)	.048	3.96 (0.51–30.52)	.187
Parameters of OGTT after 1 week
BMI, kg/m2	1.07 (0.84–1.38)	.574	1.25 (0.90–1.72)	.183
Glucose, fasting, mg/dL	0.96 (0.92–1.00)	.031	0.96 (0.92–1.00)	.075
Glucose, stimulated (30 minutes), mg/dL	0.96 (0.93–1.00)	.026	0.96 (0.92–1.00)	.043
Glucose, stimulated (60 minutes), mg/dL	0.98 (0.96–1.00)	.041	0.98 (0.95–1.00)	.060
Glucose, stimulated (90 minutes), mg/dL	0.98 (0.96–1.00)	.039	0.98 (0.95–1.00)	.057
Glucose, stimulated (120 minutes), mg/dL	0.98 (0.96–1.00)	.054	0.98 (0.96–1.00)	.077
HOMA-IR	0.18 (0.02–0.72)	.021	0.09 (0.01–0.90)	.041
HOMA-β	1.01 (0.99–1.03)	.229	1.01 (0.99–1.02)	.369
Insulinogenic index ≥ 0.1 (M)	4.00 (0.73–21.84)	.109	2.36 (0.34–16.18)	.383

The association between each baseline characteristics and diabetes remission at the 1-year visit is presented as an OR and its CI estimated using logistic regression analysis.

Significant values (P < .05) are in boldface type.BMI = body mass index, CI = confidence interval, HbA1c = hemoglobin A1c, HOMA-IR = homeostasis model assessment: estimated insulin resistance, HOMA-β = homeostasis model assessment of β-cell function, (M) = median value, OGTT = oral glucose tolerance test, OR = odds ratio.

^* Each variable was adjusted with duration of diabetes in multivariable model.

4. Discussion

In this study, we investigated the short-term (1 week from the surgery) and long-term (1 year from the surgery) glycemic outcomes in patients with gastric cancer and T2DM who underwent gastrectomy with LRYR. Forty-eight percent of our study population achieved HbA1c <6.0% without medication 1 year after surgery (defined as diabetes remission). The preoperative prognostic factors for diabetes remission were a shorter duration of diabetes and preservation of beta cell function, as calculated by the IGI. One week after surgery, only the patients in the diabetes remission group showed rapid significant improvement in fasting and 30-minute, 60-minute, 90-minute stimulated glucose levels. Higher 30-minute stimulated glucose level and HOMA-IR index at 1 week after surgery were independent factors for lower odds of 1-year diabetes remission.

According to a previous study, a longer preoperative duration of diabetes was associated with a lower T2DM remission rate in patients with gastric cancer who underwent gastrectomy with RY, Billroth I, or Billroth II reconstruction.^[7] T2DM is a chronic metabolic disorder, which is characterized by an exacerbation of ongoing beta cell dysfunction,^[20,21] and several studies have demonstrated that beta cell dysfunction aggravates as the duration of overt hyperglycemia increases.^[22,23] Prolonged exposure of beta cells to elevated glucose levels induces beta cell apoptosis, leading to decreased beta cell mass.^[24,25] Therefore, a shorter duration of diabetes and better preservation of beta cell function can be expected to be favorable factors for T2DM remission after weight reduction following gastrectomy with RY reconstruction. However, even in patients with a short duration of T2DM, a severe decrease in beta cell function can often be observed.^[26,27] Considering these individual differences, there is a limit to predicting the long-term outcome of glycemic control only by the duration of diabetes.

Nannipieri et al reported that T2DM remission after RYGB depends on the initial degree of beta cell dysfunction in morbidly obese individuals,^[28] and it is not fully clear whether similar results could be expected in patients with gastric cancer with a relatively lean state. In our study result, the insulin secretion and resistance indicators evaluated by various methods before surgery did not reflect the 1 year diabetes remission probability well. Only the IGI, a valid marker of beta cell function,^[29] showed limitedly significant associations. The progressive deterioration of β-cell function in T2DM consists of an irreversible component (e.g., β-cell apoptosis) and a reversible component due to the detrimental effects of hyperglycemia.^[30] However, several indicators currently used for measuring β-cell function cannot distinguish the relative contributions of these 2 components.

Little is known about the relationship between immediate changes in glucose parameters after surgery and long-term glycemic prognosis. In the current study, both the remission and non-remission groups showed significant reductions in body weight and the BMI 1 week after surgery. This may be caused by the difficulty in caloric intake immediately after surgery.^[2,3] Weight reduction in patients with T2DM improves insulin resistance^[31] and can lead to the improvement of hyperglycemia. However, a significant improvement in fasting and 30-minute, 60-minute, 90-minute stimulated glucose levels and HOMA-IR were observed only in the diabetes remission group in this study. In the early stages of T2DM with less permanent damage to the pancreatic beta cells, the deterioration of beta cells might be reversible,^[32] and less deterioration of insulin secretion function can be led to the more effective glucose improvement after weight reduction.^[28] There was no significant change in the IGI before and 1 week after surgery in the non-remission group, but only the HOMA-β value was decreased. β-cells with severely impaired function secrete intact proinsulin, which contributes to the glucose-lowering effect.^[33] However, this is not adequately reflected in the HOMA-β value.^[11] Therefore, there is a possibility that these results may be due to the limitations of HOMA-β values and severely impaired β-cell function in the non-remission group.

The stimulated (30 minutes, 60 minutes, and 90 minutes) glucose level during oral glucose tolerance test were considered to be important indicators of insulin secretion function and insulin resistance.^[34] Hirakawa et al reported that 30-minute stimulated glucose level can be used to predict the risk of future T2DM.^[35] As reported by Kramer et al, decrease in HOMA-IR was a key determinant of improvement of insulin secretion function in response to short-term intensive insulin therapy.^[30] This suggests a fundamental contribution of insulin resistance to the reversible component of β-cell function. Therefore, early postoperative improvements in 30-minute stimulated glucose level and HOMA-IR reflecting reversibility of β-cell function could be useful markers for the prediction of long-term glycemic control outcomes.

This study has several limitations. First, several factors, such as comorbidities of the patients that may affect the final outcome have not been fully addressed because of incomplete documentation as a result of the retrospective design of this study. Second, the number of participants in this study was small leading to a limitation of significant statistical power. Although gender was not an important factor influencing the remission rate after bypass surgery in previous studies,^[36] the biased male to female ratio in the study population may affect the outcome of this study. In addition, the 1-year follow-up period may be short for observing long-term antidiabetic effects. However, this study has several strengths. We evaluated several glycemic indices before and immediately (1 week) after the surgery. Through this, we found that confirmation of early improvement in hyperglycemia and measurement of HOMA-IR could contribute to predicting the remission of T2DM. In addition, 92% of the patients were non-obese with a BMI <30 kg/m² in the current study. It shows that metabolic/oncologic surgery has the potential to be effective in the remission of T2DM even in patients with a BMI of <30.

In conclusion, gastrectomy with LRYR leads to weight loss and insulin resistance improvement in patients with gastric cancer and T2DM and improves glycemic control. The important determinants in the long-term antidiabetic effect were the short duration of diabetes and early postoperative improvement in 30 minutes stimulated glucose level and HOMA-IR index.

Author contributions

Conceptualization: So Hun Kim, Yoon Seok Heo, Yongin Cho and Seong Ha Seo.

Data curation: Seong Ha Seo and Yongin Cho.

Formal analysis: Seong Ha Seo and Yongin Cho.

Funding acquisition: So Hun Kim.

Investigation: Da Hea Seo, Seong Hee Ahn, and Seong Bin Hong.

Methodology: Seong Bin Hong and Young Ju Suh.

Project administration: So Hun Kim and Yoon Seok Heo.

Resources: Yoon Seok Heo.

Software: Da Hea Seo.

Supervision: So Hun Kim and Yoon Seok Heo.

Validation: So Hun Kim and Yoon Seok Heo.

Visualization: Seong Hee Ahn.

Writing—original draft: Seong Ha Seo and Yongin Cho.

Writing—review and editing: So Hun Kim.