Continuous glucose monitoring reveals improved hyperglycemia and altered hypoglycemia target attainment after gastrectomy in patients with type 2 diabetes Authors: Ji Won Seo, Ki Bum Park, Kyong-Hwa Jun Published: 2026 ABSTRACT This study aimed to investigate glycemic variability and biochemical profiles before and after gastrectomy using continuous glucose monitoring (CGM) and to evaluate whether postoperative improvements occur in blood glucose levels and biochemical parameters. We analyzed data from patients who underwent surgery for gastric cancer. Glucose profiles were assessed using a CGM device before gastrectomy and after discharge. All patients discontinued antidiabetic medications one day before surgery and resumed them only after reassessment at outpatient follow-up. CGM-derived metrics were evaluated according to international consensus recommendations, including time above range (TAR), time below range (TBR), time in range (TIR), CV (coefficient of variation), and MAGE (mean amplitude of glycemic excursions). A total of 33 patients were included. The mean duration of diabetes was 16.9 ± 10.4 years, and 12.1% were receiving insulin therapy. Mean glucose levels and TAR significantly decreased after gastrectomy while TIR significantly increased. Although TBR did not increase significantly, adherence to hypoglycemia targets worsened postoperatively. Glucose variability, assessed by CV, remained stable. Also, most patients maintained CV values below the threshold for glycemic instability. MAGE significantly decreased after surgery, indicating a reduction in large-amplitude glucose excursions. Among 16 patients whose postoperative medication changes were evaluated, seven (43.8%) reduced their antidiabetic medication after gastrectomy. In conclusion, despite the temporary discontinuation of antidiabetic medications, hyperglycemia improved without aggravation of glycemic variability, as reflected by increased TIR, stable CV and reduced MAGE. However, the achievement of hypoglycemia targets was reduced. Therefore, instead of routinely resuming preoperative antidiabetic regimens, postoperative glucose management should be individualized based on CGM-derived glycemic profiles.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-55271-9. Introduction Diabetes mellitus (DM) is a common metabolic disease, and its prevalence continues to rise, with approximately 537 million adults affected worldwide in 20211. DM is known to increase the risk of microvascular and macrovascular complications2 and may adversely influence postoperative outcomes. A previous meta-analysis demonstrated that DM increases the risk of postoperative complications, prolonged hospitalization, and mortality after non-cardiac surgery3. The increased risk of postoperative mortality among patients with DM is attributed to increased susceptibility to acute renal failure, deep venous thrombosis, surgical site infection, urinary tract infection, and hemorrhage4. In the context of gastrectomy, several studies have reported that DM is significantly associated with various postoperative complications, including postoperative pneumonia5. In addition to chronic hyperglycemia, diabetic patients frequently experience fluctuations in blood glucose levels, referred to as glycemic variability, which is associated with DM-related complications6. Although self-Monitoring of Blood Glucose (SMBG) is commonly used in clinical practice, it has inherent limitations, particularly in detecting asymptomatic and nocturnal hypoglycemia, due to factors such as intermittent measurements, patient discomfort, and missed testing during sleep. In contrast, continuous glucose monitoring (CGM) provides real-time, continuous assessment of interstitial glucose levels and offers reliable metrics of glycemic variability, such as the coefficient of variation (CV). Several studies have demonstrated the clinical efficacy and safety of CGM in surgical patients with DM 7,8. Accordingly, the use of CGM enables a more detailed understanding of perioperative glycemic patterns in patients undergoing gastrectomy. However, no research to date has analyzed CGM-derived glycemic patterns before and after gastrectomy for gastric cancer. Recently, the adoption of Enhanced Recovery After Surgery protocols for gastric cancer surgery has contributed to shorter hospital stays9. As a result, adequate postoperative glycemic management may not be fully ensured after discharge. This has raised concerns regarding the optimal timing for resuming antidiabetic medications after surgery and the potential risk of hypoglycemia if medications are restarted prematurely. Therefore, we conducted this study to characterize perioperative glycemic variability and biochemical changes in gastric cancer patients undergoing gastrectomy using CGM. Furthermore, we aimed to determine whether postoperative glycemic patterns, including mean glucose levels and variability indices, improve after gastrectomy. Results Baseline characteristics A total of 38 participants were initially enrolled, but five were excluded due to poor compliance (n = 4) and re-operation (n = 1). Consequently, 33 patients who underwent surgery for gastric cancer from August 2022 to August 2024 were included in the final analysis. Baseline characteristics of the patients who underwent gastrectomy are presented in Table 1. The mean age at the time of surgery was 63.7 ± 11.7 years, and 63.6% of the patients were male. The mean BMI was 25.6 ± 3.3 kg/m², and the average duration of DM was 16.9 ± 10.4 years. Overall, 12.1% of patients received insulin therapy. Five patients were taking only one type of oral antidiabetic medication, 12 patients were on two different medications, and 16 patients were taking three or more medications. Among the surgical procedures, 84.8% were performed laparoscopically, 9.1% robotically, and 6.1% via open surgery. Distal gastrectomy was performed in 87.9% of patients. Regarding reconstruction methods, 12.1% underwent B-I, 63.7% B-II, and 12.1% R-Y reconstruction, while 12.1% underwent total gastrectomy. The distribution of pathologic TNM stages was as follows: stage I, 78.8%; stage II, 15.2%; stage III, 3.0%; and stage IV, 3.0%. The mean length of postoperative hospital stay was 6.5 ± 1.4 days. Table 1Baseline characteristics of the patients.Variablesn = 33Age63.7 ± 11.7Sex (male, %)21 (63.6%)BMI25.6 ± 3.3Diabetes duration (yr)16.9 ± 10.4Insulin therapy4 (12.1%)Total cholesterol, mg/dL157.6 ± 37.1Triglyceride, mg/dL113.6 ± 59.3HDL-C, mg/dL48.7 ± 16.2LDL-C, mg/dL86.1 ± 27.5Hypertension29 (87.9%)Dyslipidemia14 (42.4%) Approach Open2 (6.1%)Laparoscopy28 (84.8%)Robot3 (9.1%) Extent of resection Total gastrectomy4 (12.1%)Distal gastrectomy29 (87.9%) Reconstruction method Billroth-I4 (12.1%)Billroth-II21 (63.7%)Distal gastrectomy c Roux-en-Y4 (12.1%)Total gastrectomy c Roux-en-Y4 (12.1%) LN dissection D1+1 (3.0%)D232 (97.0%)Operation time (min)269.7 ± 67.1Estimated blood loss (ml)30.5 ± 52.9 Pathologic TNM stage I26 (78.8%)II5 (15.2%)III1 (3.0%)IV1 (3.0%)Postoperative hospital days6.5 ± 1.4Variables are presented as mean ± standard deviation or number (%).BMI = body mass index; HDL-C, high-density lipoprotein-cholesterol; LDL-C, low-density lipoprotein- cholesterol; LN, lymph node; TNM, tumor-node-metastasis. Table 1Baseline characteristics of the patients.Variablesn = 33Age63.7 ± 11.7Sex (male, %)21 (63.6%)BMI25.6 ± 3.3Diabetes duration (yr)16.9 ± 10.4Insulin therapy4 (12.1%)Total cholesterol, mg/dL157.6 ± 37.1Triglyceride, mg/dL113.6 ± 59.3HDL-C, mg/dL48.7 ± 16.2LDL-C, mg/dL86.1 ± 27.5Hypertension29 (87.9%)Dyslipidemia14 (42.4%) Approach Open2 (6.1%)Laparoscopy28 (84.8%)Robot3 (9.1%) Extent of resection Total gastrectomy4 (12.1%)Distal gastrectomy29 (87.9%) Reconstruction method Billroth-I4 (12.1%)Billroth-II21 (63.7%)Distal gastrectomy c Roux-en-Y4 (12.1%)Total gastrectomy c Roux-en-Y4 (12.1%) LN dissection D1+1 (3.0%)D232 (97.0%)Operation time (min)269.7 ± 67.1Estimated blood loss (ml)30.5 ± 52.9 Pathologic TNM stage I26 (78.8%)II5 (15.2%)III1 (3.0%)IV1 (3.0%)Postoperative hospital days6.5 ± 1.4Variables are presented as mean ± standard deviation or number (%).BMI = body mass index; HDL-C, high-density lipoprotein-cholesterol; LDL-C, low-density lipoprotein- cholesterol; LN, lymph node; TNM, tumor-node-metastasis. Baseline characteristics of the patients. Variables are presented as mean ± standard deviation or number (%). BMI = body mass index; HDL-C, high-density lipoprotein-cholesterol; LDL-C, low-density lipoprotein- cholesterol; LN, lymph node; TNM, tumor-node-metastasis. CGM-measured glycemic profiles Figure 1 illustrates the daily glucose profiles before and after gastrectomy, summarized using median glucose values with the interquartile range (IQR). Despite daily fluctuations, median glucose levels demonstrated an overall decreasing trend after surgery. A detailed comparison of CGM metrics is shown in Table 2. The mean glucose level significantly decreased after gastrectomy compared with preoperative values (157.6 ± 38.8 mg/dL vs. 133.3 ± 41.5 mg/dL, P < 0.001). Time above range > 180 mg/dL (TAR1) was significantly reduced (27.4 ± 22.9% vs. 17.3 ± 23.8%, P = 0.015), whereas TAR > 250 mg/dL (TAR2) also showed a decreasing trend after surgery. However, this change did not reach statistical significance (P = 0.252). Time in range (TIR) significantly increased after gastrectomy (70.6 ± 22.0% vs. 78.2 ± 22.5%, P = 0.047). Table 2CGM profiles before and after gastrectomy.VariablesBefore surgeryAfter surgeryMean difference (95% CI)P valueMean glucose, mg/dL157.6 ± 38.8133.3 ± 41.5-24.3 (-39.2 to -9.3)< 0.001Maximum of glucose level (mg/dL)301.5 ± 78.6274.9 ± 95.8-26.6 (-59.6 to 6.2)0.080Minimum of glucose level (mg/dL)73.6 ± 21.267.0 ± 22.8-6.7 (-18.9 to 5.5)0.127TAR > 250 mg/dL, %8.2 ± 12.15.3 ± 12.9-3.2 (-8.7 to 2.4)0.252TAR > 180 mg/dL, %27.4 ± 22.917.3 ± 23.8-10.9 (-19.6 to -2.2)0.015TIR 70–180 mg/dL, %70.6 ± 22.078.2 ± 22.58.4 (0.1 to 16.6)0.047TBR < 70 mg/dL, %2.0 ± 5.24.5 ± 8.22.6 (-0.9 to 6.1)0.145TBR < 54 mg/dL, %0.7 ± 2.40.3 ± 0.5-0.4 (-1.3 to 0.5)0.421CV, %27.6 ± 6.929.0 ± 6.61.2 (-2.0 to 4.4)0.449MAGE (mg/dL)91.2 ± 35.177.9 ± 27.6-15.1 (-24.8 to -5.3)0.004Variables are presented as mean ± standard deviation or number (%).CGM, continuous glucose monitoring; CI, confidence interval; TAR, time above range; TIR, time in range; TBR, time below range; CV, coefficient of variation; MAGE, mean amplitude of glycemic excursion. Table 2CGM profiles before and after gastrectomy.VariablesBefore surgeryAfter surgeryMean difference (95% CI)P valueMean glucose, mg/dL157.6 ± 38.8133.3 ± 41.5-24.3 (-39.2 to -9.3)< 0.001Maximum of glucose level (mg/dL)301.5 ± 78.6274.9 ± 95.8-26.6 (-59.6 to 6.2)0.080Minimum of glucose level (mg/dL)73.6 ± 21.267.0 ± 22.8-6.7 (-18.9 to 5.5)0.127TAR > 250 mg/dL, %8.2 ± 12.15.3 ± 12.9-3.2 (-8.7 to 2.4)0.252TAR > 180 mg/dL, %27.4 ± 22.917.3 ± 23.8-10.9 (-19.6 to -2.2)0.015TIR 70–180 mg/dL, %70.6 ± 22.078.2 ± 22.58.4 (0.1 to 16.6)0.047TBR < 70 mg/dL, %2.0 ± 5.24.5 ± 8.22.6 (-0.9 to 6.1)0.145TBR < 54 mg/dL, %0.7 ± 2.40.3 ± 0.5-0.4 (-1.3 to 0.5)0.421CV, %27.6 ± 6.929.0 ± 6.61.2 (-2.0 to 4.4)0.449MAGE (mg/dL)91.2 ± 35.177.9 ± 27.6-15.1 (-24.8 to -5.3)0.004Variables are presented as mean ± standard deviation or number (%).CGM, continuous glucose monitoring; CI, confidence interval; TAR, time above range; TIR, time in range; TBR, time below range; CV, coefficient of variation; MAGE, mean amplitude of glycemic excursion. CGM profiles before and after gastrectomy. Variables are presented as mean ± standard deviation or number (%). CGM, continuous glucose monitoring; CI, confidence interval; TAR, time above range; TIR, time in range; TBR, time below range; CV, coefficient of variation; MAGE, mean amplitude of glycemic excursion. Time below range < 70 mg/dL (TBR1) numerically increased postoperatively, however, the difference was not statistically significant (2.0 ± 5.2% vs. 4.5 ± 8.2%, P = 0.145). TBR < 54 mg/dL (TBR2) remained unchanged. Glucose variability, assessed as CV, did not differ significantly between the preoperative and postoperative periods (27.6 ± 6.9% vs. 29.0 ± 6.6%, P = 0.449). In contrast, mean amplitude of glycemic excursions (MAGE) significantly decreased after gastrectomy, from 87.8 ± 33.5 mg/dL preoperatively to 74.9 ± 23.8 mg/dL postoperatively (P = 0.004). Changes in the proportion of patients who achieved glycemic targets before and after gastrectomy are summarized in Table 3. The proportion of patients achieving TAR1 < 25% increased significantly from 57.6% preoperatively to 81.8% postoperatively (P = 0.021). The proportion of patients achieving TAR2 < 5% increased from 63.6% to 78.8%, although this trend did not reach statistical significance (P = 0.062). The proportion of patients maintaining TBR1 < 4% significantly decreased after gastrectomy (93.9% vs. 69.7%, P = 0.008), while the proportion achieving TBR2 < 1% showed no significant difference. The proportion of patients with TIR ≥ 70% increased from 63.6% to 78.8% after surgery, but this difference was not statistically significant (P = 0.180). Similarly, the proportion of patients maintaining CV < 36% slightly decreased from 90.9% to 81.8% postoperatively, without statistical significance. Individual changes in glycemic profiles before and after gastrectomy are presented in Fig. 2. Table 3Changes in the proportion of patients achieving glycemic targets before and after gastrectomy.VariablesBefore surgeryAfter surgeryP valueTAR1† < 25%57.6%81.8%0.021TAR2‡ <5%63.6%78.8%0.062TBR1§ < 4%93.9%69.7%0.008TBR2∥ <1%84.8%81.8%1.000TIR * ≥ 70%63.6%78.8%0.180CV < 36%90.9%81.8%0.508Variables are presented as mean ± standard deviation or number (%).TAR, time above range; TBR, time below range; TIR, time in range; CV, coefficient of variation.* Glucose concentration 70–180 mg/dL.† Glucose concentration > 180 mg/dL.‡ Glucose concentration > 250 mg/dL.§ Glucose concentration < 70 mg/dL.∥ Glucose concentration < 54 mg/dL. Table 3Changes in the proportion of patients achieving glycemic targets before and after gastrectomy.VariablesBefore surgeryAfter surgeryP valueTAR1† < 25%57.6%81.8%0.021TAR2‡ <5%63.6%78.8%0.062TBR1§ < 4%93.9%69.7%0.008TBR2∥ <1%84.8%81.8%1.000TIR * ≥ 70%63.6%78.8%0.180CV < 36%90.9%81.8%0.508Variables are presented as mean ± standard deviation or number (%).TAR, time above range; TBR, time below range; TIR, time in range; CV, coefficient of variation.* Glucose concentration 70–180 mg/dL.† Glucose concentration > 180 mg/dL.‡ Glucose concentration > 250 mg/dL.§ Glucose concentration < 70 mg/dL.∥ Glucose concentration < 54 mg/dL. Changes in the proportion of patients achieving glycemic targets before and after gastrectomy. Variables are presented as mean ± standard deviation or number (%). TAR, time above range; TBR, time below range; TIR, time in range; CV, coefficient of variation. * Glucose concentration 70–180 mg/dL. † Glucose concentration > 180 mg/dL. ‡ Glucose concentration > 250 mg/dL. § Glucose concentration < 70 mg/dL. ∥ Glucose concentration < 54 mg/dL. Fig. 1Daily mean glucose profiles before and after gastrectomy. Lines indicate the median glucose values, and shaded areas represent the interquartile range. Daily mean glucose profiles before and after gastrectomy. Lines indicate the median glucose values, and shaded areas represent the interquartile range. Fig. 2Individual changes in CGM-derived glycemic metrics before and after gastrectomy (red dashed lines indicate the glycemic target thresholds for each level). CGM, continuous glucose monitoring; TAR, time above range; TBR, time below range; TIR, time in range; CV, coefficient of variation; MAGE, mean amplitude of glycemic excursion. Individual changes in CGM-derived glycemic metrics before and after gastrectomy (red dashed lines indicate the glycemic target thresholds for each level). CGM, continuous glucose monitoring; TAR, time above range; TBR, time below range; TIR, time in range; CV, coefficient of variation; MAGE, mean amplitude of glycemic excursion. Biochemical profiles BMI, HbA1c, and glycoalbumin significantly decreased after gastrectomy (25.6 ± 3.3 vs. 23.4 ± 2.9, P < 0.001; 7.2 ± 1.0% vs. 6.5 ± 0.8%, P < 0.001; 16.5 ± 3.8% vs. 15.1 ± 3.0%, P = 0.003, respectively). However, no significant changes were observed in fasting plasma glucose (FPG), C-peptide, or insulin levels (Table 4). Table 4Biochemical profiles before and after gastrectomy.VariablesBefore surgeryAfter surgeryMean difference (95% CI)P valueBMI25.6 ± 3.323.4 ± 2.92.2 (1.5 to 2.7)< 0.001FPG145.5 ± 44.4127.0 ± 35.918.5 (0.4 to 36.7)0.088HbA1c (%)7.2 ± 1.06.5 ± 0.80.7 (0.4 to 1.0)< 0.001C-peptide, ng/mL2.5 ± 1.42.3 ± 1.80.1 (-0.7 to 0.9)0.623Glycoalbumin (%)16.5 ± 3.815.1 ± 3.01.4 (0.3 to 2.5)0.003Insulin, µU/mL13.7 ± 23.68.1 ± 11.25.6 (-3.7 to 14.8)0.098Values are presented as mean ± standard deviation.CI, confidence interval; BMI = body mass index; FPG, fasting plasma glucose; HbA1c, glycosylated hemoglobin A1. Table 4Biochemical profiles before and after gastrectomy.VariablesBefore surgeryAfter surgeryMean difference (95% CI)P valueBMI25.6 ± 3.323.4 ± 2.92.2 (1.5 to 2.7)< 0.001FPG145.5 ± 44.4127.0 ± 35.918.5 (0.4 to 36.7)0.088HbA1c (%)7.2 ± 1.06.5 ± 0.80.7 (0.4 to 1.0)< 0.001C-peptide, ng/mL2.5 ± 1.42.3 ± 1.80.1 (-0.7 to 0.9)0.623Glycoalbumin (%)16.5 ± 3.815.1 ± 3.01.4 (0.3 to 2.5)0.003Insulin, µU/mL13.7 ± 23.68.1 ± 11.25.6 (-3.7 to 14.8)0.098Values are presented as mean ± standard deviation.CI, confidence interval; BMI = body mass index; FPG, fasting plasma glucose; HbA1c, glycosylated hemoglobin A1. Biochemical profiles before and after gastrectomy. Values are presented as mean ± standard deviation. CI, confidence interval; BMI = body mass index; FPG, fasting plasma glucose; HbA1c, glycosylated hemoglobin A1. Factors associated with postoperative TAR improvement The change in FPG was the only variable independently associated with the change in TAR1 (β = 0.195, 95% CI 0.04 to 0.36, P = 0.018). The changes in BMI and glycoalbumin, R-Y reconstruction, age, and sex were not independently associated with the change in TAR1. For the change in TAR2, no variable reached independent statistical significance. Detailed results are presented in Supplementary Table 1. Changes in antidiabetic medication Among the 33 patients, changes in antidiabetic medication regimens were analyzed in 16 patients (48.5%). After gastrectomy, seven patients (43.8%) reduced their oral antidiabetic medications. Two patients discontinued oral antidiabetic therapy entirely, while five patients reduced medication dosage. Details of medication changes are presented in Supplementary Table 2. Eight patients maintained their preoperative regimens, while one patient increased the dosage of antidiabetic agents postoperatively. Discussion Unlike other types of cancer surgery, gastrectomy for gastric cancer has distinct metabolic consequences, including restricted food intake10 and alterations in gut hormones11, which may contribute to reductions in hyperglycemia. Additionally, depending on the type of reconstruction performed after gastrectomy, changes in the passage of food can induce hypoglycemia, a phenomenon known as dumping syndrome. Consequently, the inappropriate continuation of preoperative antidiabetic medications after gastrectomy may further increase the risk of hypoglycemia. For this reason, we analyzed perioperative glycemic profiles using CGM to determine whether antidiabetic medications should be maintained or adjusted following surgery. In this study, gastrectomy was associated with an overall improvement in glycemic exposure with significant reduction in mean glucose levels and TAR, accompanied by a significant increase in TIR. Although the mean proportion of TBR did not change significantly after gastrectomy, attainment of the guideline recommended TBR1 target (< 4%) was significantly reduced, despite the temporary discontinuation of antidiabetic medications for approximately 2 weeks after surgery. These findings suggest that antidiabetic medications should be carefully individualized according to patients’ postoperative conditions rather than being resumed immediately at their preoperative dosages after gastrectomy. Several previous studies have examined the risk of hypoglycemia after gastrectomy and its long-term association with mortality12,13. However, it is important to note that, during the immediate postoperative hospitalization period, hypoglycemia may manifest as adverse events, such as falls or other accidents. Moreover, most previous research on hypoglycemia in the context of gastric cancer surgery has focused primarily on dumping syndrome8,14. In contrast, the present study analyzed overall postoperative glycemic profiles using CGM rather than centering exclusively on dumping syndrome. Postoperative hypoglycemia did not increase uniformly across all patients. Rather, it was observed mainly in a subset of individuals, which led to fewer patients meeting the guideline-recommended TBR1 target. This pattern suggests that some patients became more susceptible to clinically relevant hypoglycemia after gastrectomy, even though the overall average hypoglycemic exposure remained unchanged. Some antidiabetic medications, such as sulfonylureas, are known to cause hypoglycemia depending on their mechanisms of action15. Therefore, continuing preoperative antidiabetic regimens without adjustment after surgery may further elevate the risk of hypoglycemia. HbA1c reflects mean plasma glucose over 2–3 months and remains a standard indicator of long-term glycemic control and chronic diabetic complications16,17. Glycoalbumin, which represents average glycemia over 2–3 weeks and responds more rapidly to therapeutic changes, is useful for assessing short-term treatment effects18. Previous studies evaluating metabolic changes in patients with DM after gastrectomy have reported improvements in metabolic hormones and indices, including reductions in BMI and HbA1c19. Our study consistently demonstrated significant reductions in both HbA1c and glycoalbumin after gastrectomy, indicating that surgery effectively improved glycemic control over both short- and longer-term periods. Moreover, earlier studies have shown that 47–54% of gastric cancer patients who undergo gastrectomy reduce their doses of antidiabetic medications21, which was consistent with our findings. To further explore the metabolic basis of the postoperative improvement in glycemic exposure, we examined which clinical factors were independently associated with the change in TAR after gastrectomy. In multivariate regression analysis, the change in FPG was the only variable independently associated with the change in TAR1, whereas the change in BMI, the change in glycoalbumin and reconstruction method were not. Although the reduction in FPG at the group level did not reach statistical significance, the variation in FPG change at the individual level was significantly associated with the variation in TAR1 change, indicating that patients who experienced larger reductions in fasting plasma glucose also experienced correspondingly larger reductions in time spent above 180 mg/dL. This finding is notable in two respects. First, despite a significant overall reduction in BMI after gastrectomy, the change in BMI itself was not independently associated with the change in TAR1, suggesting that postoperative improvement in glycemic variability was not primarily mediated by weight loss. Second, the concurrent reduction in FPG and TAR1 is consistent with weight-independent mechanisms that improve both fasting and postprandial glucose homeostasis simultaneously, such as enhanced incretin response following upper gastrointestinal anatomical rearrangement11. R-Y reconstruction has long been recognized as an important contributor to postoperative glycemic improvement in gastric cancer surgery19,20. In the present analysis, however, R-Y reconstruction was not a significant independent contributor to TAR improvement, which likely reflects the limited sample size of the R-Y subgroup rather than the absence of a true metabolic effect. HbA1c was not included as a candidate variable in the regression analysis because it reflects average glycemia over the preceding 2–3 months, which is not temporally aligned with the postoperative CGM monitoring period of approximately two weeks. More broadly, improvement in HbA1c does not necessarily indicate optimal glycemic stability after gastrectomy. Gastrectomy is frequently accompanied by anemia, altered iron absorption, and changes in red blood cell turnover, all of which may lead to underestimation of true glycemic exposure when HbA1c is used alone 23. In our additional analyses, we observed postoperative discordance between HbA1c and CGM-derived metrics. Specifically, although 75.8% of patients achieved HbA1c < 7.0% at 3 months after gastrectomy, 36.4% of patients still exhibited CGM-defined hyper or hypoglycemia, defined as TAR1 > 25% and/or TBR1 > 4%. Notably, among patients with postoperative HbA1c < 7.0%, nearly half (48.0%) continued to demonstrate clinically relevant hyperglycemic or hypoglycemic exposure on CGM. When discordance was analyzed separately, hypoglycemia discordance increased substantially after surgery, whereas hyperglycemia discordance showed minimal change (data not shown). These findings suggest that HbA1c may mask clinically important glycemic abnormalities, particularly hypoglycemia, after gastrectomy 8. Therefore, although conventional biochemical markers demonstrate overall improvement after gastrectomy, CGM provides complementary and clinically important information regarding residual glycemic excursions that may not be captured by HbA1c or glycoalbumin alone. However, few studies have examined glycemic variability before and after gastrectomy. Glycemic variability is considered a more critical determinant of diabetic complications than mean glucose levels, as large glucose fluctuations may induce oxidative stress more effectively than sustained hyperglycemia27. Since glycemic variability has emerged as an important therapeutic target beyond conventional markers, such as HbA1c, the present study adds to the evidence by evaluating perioperative glycemic variability in diabetic patients undergoing gastrectomy. In the present study, postoperative CV remained stable and below the threshold of 36% in most patients, indicating preserved glycemic stability after gastrectomy. Importantly, MAGE significantly decreased after surgery, suggesting a reduction in large amplitude glucose excursions. These findings indicate that gastrectomy did not shift patients from a pattern of chronic hyperglycemia toward high glycemic variability, despite reduced attainment of hypoglycemia targets in a subset of patients. Rather, postoperative improvements in hyperglycemia occurred with maintained overall glycemic stability. This observation may be partly explained by the results that baseline TIR and CV values were already within stable ranges in most patients, limiting the extent of further improvement. In addition, postoperative CGM measurements were obtained before resumption of antidiabetic medications, and the observed changes may reflect postoperative dietary restriction and physiological adaptation rather than pharmacologic effects alone. Several limitations of this study should be acknowledged. First, as a single-center study with a limited sample size, there may be inherent selection bias. Second, potential confounding factors affecting glycemic status, aside from antidiabetic medications, were not fully controlled. Third, The CGM device used in this study stores glucose data for up to eight hours. Therefore, data loss occurred when patients did not scan the sensor within this time window, resulting in variability in data completeness among participants. The mean proportion of missing CGM data was 34.5% ± 24.4% in the preoperative period and 39.5% ± 23.8% in the postoperative period. However, when missing data were evaluated at the day level, the extent of missingness was substantially lower, with a mean missing-day proportion of 15.7% ± 17.9%. These findings suggest that, despite intermittent data loss related to the technical limitations of flash CGM systems, most monitoring days retained at least partial glucose information suitable for analysis, reflecting real-world perioperative CGM use. In addition, this study did not include an analysis of dietary intake, which can influence glycemic fluctuations. Although patients received nutritional education before discharge (supplementary data), their postoperative dietary patterns were not directly monitored, and variations in food composition may have affected glucose levels. Fourth, the impact of glycemic variability control on postoperative outcomes was not evaluated. Finally, the impact of various reconstruction methods on postoperative glycemic profiles could not be fully characterized in the present study. For distal gastrectomy, three different reconstruction types were performed according to surgeons’ preferences. Although reconstruction method was included as a variable in the multivariate regression analysis of postoperative TAR change, the small number of R-Y cases limited the statistical power to detect a true effect, and dedicated subgroup analyses for hypoglycemic events by reconstruction type were not feasible. The specific impact of different reconstruction methods on both hyperglycemic exposure and hypoglycemia risk after gastrectomy needs to be confirmed in larger sample size in future studies. In conclusion, gastrectomy for gastric cancer exerts dual metabolic effects. It can stabilize blood glucose levels through metabolic improvements but may also induce hypoglycemia due to complications such as dumping syndrome. Consequently, debate continues regarding optimal adjustment of antidiabetic medications after surgery. In the present study, both blood glucose levels and glycemic variability were assessed before and after gastrectomy. Even after the discontinuation of antidiabetic medications, glycemic variability remained stable, whereas hyperglycemia improved. However, adherence to guideline-recommended hypoglycemia targets was reduced. Therefore, rather than routinely maintaining preoperative antidiabetic regimens after gastrectomy, postoperative glucose management should be individualized based on each patient’s metabolic status and glucose profile. Materials and methods Study design and participants This prospective, observational, single-center study was conducted at the Catholic University of Korea, St. Vincent’s Hospital. The study was approved by the Institutional Review Board of our institution (approval no. VC22DIST0078), and written informed consent was obtained from each participant. The inclusion criteria were as follows: age ≥ 19 years, diagnosis of type II DM, and a confirmed diagnosis of gastric adenocarcinoma prior to surgery, followed by curative gastrectomy with radical lymph node dissection. The exclusion criteria included inability to use a smartphone application compatible with the CGM system, advanced gastric cancer requiring neoadjuvant chemotherapy, undergoing palliative gastrectomy, and postoperative complications classified as Clavien-Dindo grade IIIB or higher. In patients undergoing distal gastrectomy, Billroth-I (B-I), Billroth-II (B-II), or Roux-en-Y (R-Y) reconstruction was performed based on the surgeon’s preference. In total gastrectomy patients, R-Y reconstruction was performed. The extent of lymph node dissection and treatment strategy followed the Korean Gastric Cancer Association guidelines28. Clinical variables, including age, sex, body mass index (BMI), DM duration, insulin therapy, and hypertension, were assessed preoperatively. DM treatment was categorized into insulin injections and oral medications for analysis. Perioperative characteristics, such as surgical approach, extent of lymph node dissection, operation time, estimated blood loss, pathologic tumor-node-metastasis (TNM) stage, and length of postoperative hospital stay, were also evaluated. TNM staging was assigned according to the AJCC Cancer Staging Manual, 8th edition29. Glucose monitoring A flash CGM device, the Freestyle Libre Flash Glucose Monitoring System (Abbott Diabetes Care Inc., Alameda, CA, USA), was used to record each patient’s interstitial glucose levels. The CGM sensor was placed subcutaneously on the left or right upper arm 10–14 days before surgery. On the day of discharge, a new CGM sensor was placed and used for an additional 2 weeks. The device measured and recorded glucose concentrations every 15 min in interstitial fluid using a small filament inserted just beneath the skin, and glucose data were displayed when scanned. Stored data were subsequently downloaded for analysis. Metrics including mean, median, standard deviation (SD), maximum, minimum, and percentage of time within specific glucose ranges (< 54 mg/dL, < 70 mg/dL, 70–180 mg/dL, > 180 mg/dL, and > 250 mg/dL) were evaluated. Glucose levels were assessed by comparing CGM readings obtained before gastrectomy with those recorded 3 weeks postoperatively. Perioperative Glucose Management. Beginning the day before surgery, antidiabetic medications were discontinued due to initiation of nothing by mouth status. Sips of water were permitted on postoperative day (POD) 1, followed by a liquid diet on POD 2. On POD 3, a soft diet was initiated for distal gastrectomy patients. For total gastrectomy patients, sips of water began on POD 3, followed by liquid diet on POD 4 and soft diet on POD 5. During hospitalization, insulin was administered according to a Humalog sliding-scale protocol. At discharge, patients were advised not to immediately resume their preoperative antidiabetic medications. Instead, they were instructed to visit their outpatient clinic for individualized reassessment and adjustment of their antidiabetic regimen. Outcomes and measurements The primary outcome of this study was that of the difference in glucose patterns before and after gastrectomy. Secondary outcomes included mean differences in TAR, TIR, TBR, and biochemical markers, such as HbA1c, C-peptide, glycoalbumin, and insulin. TBR was defined as glucose concentration < 70 mg/dL (Level 1, TBR1) and < 54 mg/dL (Level 2, TBR2), while TAR was defined as glucose concentration > 180 mg/dL (Level 1, TAR1) and > 250 mg/dL (Level 2, TAR2). TIR was defined as glucose concentration between 70 and 180 mg/dL. The CV was calculated as 100 × (SD / mean glucose) and reported as %CV. MAGE was calculated by taking the arithmetic mean of the absolute glucose increases or decreases between consecutive nadirs and peaks (or vice versa), when both ascending and descending excursions exceeded one standard deviation of the mean glucose value for the corresponding period30. CGM-derived metrics (CV, TIR, TBR, and TAR) were calculated for both preoperative and postoperative periods. Target thresholds for these metrics followed international consensus recommendations31: CV < 36%, TIR ≥ 70%, TAR1 < 25%, TAR2 < 5%, TBR1 < 4%, and TBR2 < 1%. The proportion of patients meeting each target was compared between the pre and postoperative periods. Biochemical markers, including C-peptide, glycoalbumin, and insulin, were measured 2 weeks before surgery and 3 weeks after surgery, whereas HbA1c was assessed 3 months after gastrectomy. Plasma glucose was measured using the hexokinase method, and HbA1c and glycoalbumin (Lucica® GA-L, Tokyo, Japan) were measured via high-performance liquid chromatography. Serum insulin (DIAsource ImmunoAssays, Nivelles, Belgium) and C-peptide (Izotop, Budapest, Hungary) levels were assessed using radioimmunoassay techniques. Statistical analysis All data were analyzed using IBM SPSS Statistics version 25.0 (IBM Corp., Armonk, NY, USA) and R software (version 4.3.0; Vienna, Austria). Baseline characteristics are presented as means and standard deviations for continuous variables and as frequencies and proportions for categorical variables. Continuous variables were analyzed using paired t-tests or the Wilcoxon signed-rank test, and categorical variables were compared using McNemar’s test. To identify clinical factors associated with postoperative TAR improvement, multivariable linear regression analyses were performed using the postoperative change in TAR1 and TAR2 (calculated as postoperative minus preoperative values in percentage points) as continuous dependent variables. Pre-specified candidate variables selected on the basis of clinical relevance were entered into the multivariable model together with baseline TAR, which was included as a covariate to adjust for regression to the mean. All statistical tests were two-sided, and a P value of < 0.05 was considered statistically significant. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1