Changes in basal insulin secretion capacity after pancreatectomy: a single-center retrospective study

DOI: https://doi.org/10.21203/rs.3.rs-1966286/v1

Abstract

Background

This study evaluated the changes in basal insulin secretion capacity (BISC) after pancreatectomy across two surgical procedures. We also investigated the association between decreased BISC and the introduction of postoperative insulin therapy.

Methods

The data on 110 patients who underwent pancreatectomy during October 2018–February 2022 at our hospital were retrospectively registered and analyzed. We focused on the C-peptide index (CPI) as an index for BISC. A decreased postoperative BISC was defined as a postoperative CPI (post-CPI) < 1.0, which was in turn defined as the event occurrence in our study. The receiver operating characteristic curve for the event occurrence was plotted for factors related to preoperative glucose tolerance. Of the 73 preoperative non-diabetic patients, 44 and 29 who underwent pancreaticoduodenectomy (PD) and distal pancreatectomy (DP), respectively, were examined.

Results

A post-CPI of < 1.0 was observed in 29 (39%) patients. Although the proportion of remnant pancreatic volume was significantly smaller in patients with PD than in patients with DP (38% vs. 68%, p < 0.0001), no significant difference was observed in the event rates (39% vs. 41%). In patients with PD, preoperative CPI (area under the curve: 0.75) was used for predicting post-CPI < 1.0. Multivariate analysis revealed that preoperative CPI < 1.65 (odds ratio: 7.69, 95% confidence interval: 1.87‒31.5) was an independent predictor of decreased postoperative BISC. However, in patients with DP, no significant predictors were identified. Induction of insulin therapy was significantly lower in preoperative non-diabetic patients (n = 73) after undergoing a pancreatectomy (1.4% vs. 37.5%, p < 0.0001) than in preoperative medically treated patients (n = 16). There was no significant difference in event (post-CPI < 1.0) rates (39.7% vs. 56.2%, p = 0.23).

Conclusions

Although pancreatectomy reduces BISC after surgery, the coexistence of insulin resistance, which is a pathophysiology of type 2 diabetes mellitus, may play a role in whether postoperative glucose tolerance is reduced to the point where insulin therapy is necessary.

Background

Pancreatectomy is essential in several neoplastic diseases, such as cholangiocarcinoma, cancer of the duodenal papilla, intraductal papillary mucinous tumors, and neuroendocrine tumors, besides pancreatic cancer. Although some postoperative complications, such as pancreatic fistulas, remain to be resolved, surgical techniques and perioperative management have improved the short-term outcomes, such as perioperative mortality [1, 2]. Medium- to-long-term complications after pancreatectomy include disease-free survival and deterioration of exocrine and endocrine functions.

Pancreatectomy has an effect on pancreatic endocrine functions; it reduces the capacity for insulin secretion and may lead to the development of new-onset diabetes mellitus (NODM) [3]. Insulin is secreted by β-cells of the islets of Langerhans in the pancreas. Many studies have indicated that the islets of Langerhans are predominantly distributed in the pancreatic tail [46]. Although some studies suggest that pancreatic tail resection is likely to cause NODM [7], others have reported that the incidence of NODM does not differ between pancreaticoduodenectomy (PD) and distal pancreatectomy (DP) [8, 9]. As noted above, some studies examining factors contributing to the occurrence of NODM after pancreatectomy have reported contradictory findings [79]. However, only a few studies have evaluated the changes in endocrine function, particularly insulin secretion capacity, before and after pancreatectomy [10, 11].

There are two known types of insulin secretion: basal insulin secretion, which remains constant regardless of food intake, and additional insulin secretion, which is in response to increasing blood glucose levels post food intake. Decreased basal insulin secretion reduces the blood glucose uptake by the muscles and other tissues and weakens the ability of insulin to inhibit gluconeogenesis in the liver during fasting, leading to increased fasting blood glucose levels, which is a diagnostic criterion for diabetes mellitus (DM). However, the factors affecting changes in the basal insulin secretory capacity (BISC) after pancreatectomy remain unclear. Therefore, in preoperative non-diabetic patients, this study aimed to evaluate changes in BISC after pancreatectomy under two surgical procedures (PD and DP) and identify predictive factors for the postoperative decline in BISC. Further, changes in DM treatment after pancreatectomy were tracked for patients who had received only preoperative oral DM therapy.

Methods

Patients

We enrolled 110 patients who underwent pancreatectomy (PD and DP) at Japanese Red Cross Kumamoto Hospital between October 2018 and February 2022. Pre- and postoperative clinical data were retrospectively collected.

Patients were excluded from the analysis if they had undergone central pancreatectomy (n = 1), partial resection (n = 1), or total pancreatectomy (including total remnant pancreatic resection [n = 2]). Moreover, patients with missing data (inadequate preoperative glucose tolerance assessments) were also excluded (n = 7). Patients on preoperative insulin therapy were also excluded (n = 10). Finally, 89 patients were enrolled (Fig. 1). The following main analyses were performed on preoperative non-diabetic patients (n = 73), whereas preoperative type 2 diabetic patients on medication therapy (n = 16) were primarily evaluated for the presence of postoperative induction of insulin therapy.

Index of basal insulin secretion

Among the various indices of BISC, we used the C-peptide index (CPI), a useful indicator for requirement of insulin therapy [12]. The CPI can be calculated using the following simple formula: fasting C-peptide immunoreactivity (ng/mL)/fasting plasma glucose (mg/dL)×100. For normal glucose tolerance, the CPI is approximately 1.60 ± 0.57 [12]. A CPI of < 0.8–1.0 is indicated as the criterion for insulin therapy requirement for type II DM [12, 13]. We defined a decrease in BISC as a postoperative CPI (post-CPI) of < 1.0, and the changes in CPI pre- and post-surgery were compared according to the operative procedure.

Perioperative factors

The surgical procedures of PD and DP were analyzed. Data on operative time, blood loss, blood transfusion, pancreatic firmness in intraoperative findings, postoperative pancreatic fistula (POPF), and delayed gastric emptying (DGE) as postoperative complications were extracted as factors. The POPF was evaluated using the revised classification of the International Study Group of Pancreatic Fistula [14], and DGE was evaluated using the International Study Group of Pancreatic Surgery classification in 2007 [15].

Preoperative and perioperative factors related to diabetes mellitus and glucose tolerance

We considered patient characteristics, surgery, and DM-related factors as possible predictors of the occurrence of the event of interest (post-CPI < 1.0). Patient factors included age, sex, and body mass index; surgical factors included the type of surgical procedure and proportion of remnant pancreatic volume (PRPV) after surgery based on computed tomography (CT) volumetric results; and DM factors included the preoperative hemoglobin A1c (HbA1c) level, BISC (CPI), serum insulin level, C-peptide level, and index of insulin resistance (homeostasis model assessment of insulin resistance).

In Japan, prior to liver surgery, hepatic function is estimated based on the K indocyanine green (K-ICG) value, which is obtained by calculating the extinction rate of intravenously injected ICG. The future remnant hepatic function is estimated by multiplying the K-ICG value by the expected proportion of remnant liver volume [16, 17]. However, we introduced a new index, CPV, calculated using the following formula: preoperative CPI (pre-CPI) ×PRPV measured by CT volumetry.

Similarly, using the homeostasis model assessment of β-cell function (HOMA-β), which is one index of basal insulin secretion capacity, we generated a new index of HPV, obtained by multiplying preoperative HOMA-β by PRPV, as a predictor of event occurrence (post-CPI < 1.0).

Blood tests were performed in the fasting state, and the values were measured preoperatively and within 1 week to 1 month postoperatively. CT volumetry was performed using ZIOstation2 (Ziosoft, Inc., Tokyo, Japan); the pancreatic parenchyma contour was plotted on a 1-mm-width slice from the preoperative and first postoperative CT. The plot area of each slice was calculated, and the pancreatic volume was obtained by integrating the areas.

Perioperative blood glucose control

In cases of patients with poor preoperative blood glucose control, surgery was performed after inpatient fortified insulin therapy (these patients were excluded in this study). After surgery, a sliding scale was used, and the fasting blood glucose level was maintained at < 200 mg/dL. When the fasting blood glucose level was over 200 mg/dL after food intake was stabilized, oral hypoglycemic agents or self-injection of insulin was introduced under the guidance of diabetes specialists.

Statistical analyses

To evaluate the predictive power of each factor for event occurrence (post-CPI < 1.0), receiver operator characteristic (ROC) curves were generated for each factor, and the area under the ROC curve (AUC) was calculated and compared for each factor. Thereafter, the ROC-AUC variable suggestions found to be the most useful were added to the logistic regression model for event occurrence that was adjusted for age, sex, and surgical procedure to verify whether they remained significant after these adjustments.

Further, we conducted stratified fourfold cross-validation to evaluate the generalizability of the factors selected in the above tests. Specifically, the data were divided into two groups (training and testing groups) at a 3:1 ratio; the data were further distributed such that the proportion of the event occurrence (post-CPI < 1.0) was equal between the two groups (training and testing groups). The same division was randomly performed 10 times, and logistic regression was performed to predict the event occurrence (post-CPI < 1.0) in the split data sets: using the previously obtained event predictors for each surgery category as explanatory variables—to predict the event occurrence (post-CPI < 1.0) for each surgery. Logistic regression was performed using the following formula:

where 𝑌=1 is the event occurrence (post-CPI < 1.0), β0 is the intercept, β1 is the coefficient, and X1 is the obtained predictors of the event occurrence (post-CPI < 1.0).

The ROC-AUC was calculated for each split data set and assessed as a mean of the 10 split data sets. Finally, the maximum Youden index (sensitivity- [1-specificity]) was obtained from the ROC curve to calculate the cutoff value of event occurrence predictors for each surgical procedure [18].

Continuous data are expressed as mean ± standard deviation (SD) or median and range. Continuous data were analyzed using Student’s t-test or Wilcoxon’s signed-rank sum test for unpaired data and paired t-test or Wilcoxon’s signed-rank test for paired data, as appropriate. Categorical data were analyzed using either the chi-squared or Fisher’s exact test. All statistical analyses were performed using JMP® Pro version 16.0.0 (SAS Institute Inc., Cary, NC, USA) and R version 4.0.0 (The R Foundation for Statistical Computing, Vienna, Austria). All analyses were two-tailed, and p < 0.05 was considered statistically significant.

Ethical statement

The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted as per the Declaration of Helsinki (as revised in 2013) guidelines. This study was approved by the Ethics Board of the Japanese Red Cross Kumamoto Hospital (approval no.: 474), and the requirement for individual consent for this retrospective analysis was waived. We posted a summary of the trial on our website and instructed eligible patients to inform us if they wished to be excluded from this study. However, none of the participants requested exclusion.

Results

Of the 73 preoperative non-diabetic patients included in the analysis, 41 were men and 32 were women (median age, 67.0 [15‒84] years). The diseases investigated were pancreatic cancer (n=39), biliary tract cancer (n=14), and low malignancy tumor (n=14). The surgical procedures included PD (n=44) and DP (n=29). Four patients developed NODM postoperatively; one of them was introduced to insulin therapy. The patients’ background characteristics are presented in Table 1.

Patient, surgical, and glucose tolerance-related factors were compared according to the operative procedure and presence/absence of the event (post-CPI <1.0) (Table 2). A post-CPI <1.0 was observed in 29 patients (Table 2, 39.7%). Although PD patients had a significantly smaller PRPV than DP patients (Table 2, PD: 38%, DP: 68%, p<0.0001), no significant difference was noted in the incidence of post-CPI <1.0 (Table 2, 38.6% vs. 41.3%, p=0.81). Comparison of the presence of an event (post-CPI <1.0) occurrence revealed significant differences in body mass index (BMI), diagnosis (pancreatic cancer), and glucose tolerance-related factors, such as preoperative HbA1c, C-peptide, CPI, HOMA-β, CPV, and HPV (Table 2).

The changes in pre- and post-operative CPI are shown in Fig 2. The CPI decreased postoperatively in 85% of the preoperative non-diabetic patients; post-CPI was significantly lower than pre-CPI in patients with these two operative procedures (Fig 2A, PD: p<0.001) (Fig 2B, DP: p<0.001).

To assess the ability of each factor to predict the occurrence of post-CPI <1.0, we used the ROC-AUC. Preoperative CPI (pre-CPI) was the best predictor of the occurrence of post-CPI <1.0 in PD patients (Table 3, AUC: 0.75, 95% confidence interval [CI]: 0.59‒0.90). Similarly, pre-BMI was the most useful predictor in DP patients (Table 3; AUC: 0.65, 95% CI: 0.51‒0.85). Consequently, logistic regression analysis after adjustment for age and sex was performed for pre-CPI in PD patients and pre-BMI in DP patients. Pre-CPI for PD patients was significantly associated with the event occurrence (post-CPI <1.0) in the adjusted multivariate analysis (Table 4, p=0.038), while in DP, pre-BMI could not be associated with an independent predictor of the event occurrence (Table 4, p=0.082).

The stratified cross-validation results are shown in Table 5. As a predictive marker for the post-CPI <1.0 event occurrence, pre-CPI in PD demonstrated reasonable generalization performance.

The cutoff value for predicting post-CPI <1.0 was 1.63 (sensitivity: 0.81, specificity: 0.71) for pre-CPI in preoperative non-diabetic patients who underwent PD (Table 6).

Finally, pre- and post-operative glucose tolerance and post-operative insulin induction rate were compared between preoperative non-diabetic patients and patients on preoperative diabetic medication (Table 7). Preoperative non-diabetic patients had a significantly better preoperative HbA1c level (Table 7, mean [SD]: 5.75 [0.48] vs. 6.60 [0.87], p<0.0001) and preoperative CPI (Table 7, 1.92 [0.76] vs. 1.42 [0.99], p<0.0001) than in patients on preoperative diabetic medication. Although the postoperative CPI and event occurrence (post-CPI <1.0) rates were not significantly different (Table 7, 1.27 [0.58] vs. 1.18 [0.90], p=0.60; 39.7% vs. 56.2%, p=0.23), the postoperative insulin induction rate was significantly lower in the preoperative non-diabetic patients than in patients on preoperative diabetic medication (Table 7, 1.4% vs. 37.5%, p<0.0001).

Discussion

We used CPI as an index of BISC, and a postoperative CPI <1.0 was defined as a decrease in BISC that marked the occurrence of the event of interest. Postoperative CPI was significantly lower than preoperative CPI in approximately 85% of the patients, regardless of surgical procedure. The proportion of remnant pancreas volume by surgical procedure was significantly greater in DP than in PD, but there were no significant differences in the event rates. We identified preoperative CPI in PD patients as a predictor of post-CPI <1.0, whereas no significant factors were observed for DP patients. Preoperative non-diabetic patients had a significantly lower rate of insulin therapy induction after pancreatectomy than preoperative medically treated patients. In a comparison between preoperative non-diabetic patients and patients on diabetic medication, there were no significant differences in the postoperative CPI and event occurrence rates. However, the postoperative insulin induction rates were significantly lower in preoperative non-diabetic patients than in patients on diabetic medication.

A previous report suggested preoperative 24-h urinary C-peptide excretion measurement and multiplication of this value by the predicted pancreatic volume as a predictor of postoperative endocrine function [11]. However, 24-h urine storage is a complicated and burdensome test for patients and is considered less versatile. However, CPI is a simple indicator that can be measured using blood tests and is considered more versatile.

The results of this study suggest that in terms of predictors of postoperative decreased BISC, the factors that are significant in PD are not significant in DP, suggesting that the predictors differ by surgical procedure. This difference might be due to the heterogeneous distribution of the islets of Langerhans in the pancreas. Generally, the islets of Langerhans are largely distributed in the pancreatic tail [4-6]. In PD, the pancreatic head, which has a large volume, is removed, whereas the pancreatic tail, which is thought to contain many islets of Langerhans, is preserved. Therefore, it was suggested that BISC after PD strongly reflects the index of preoperative BISC, which is represented by preoperative CPI. Since the residual pancreatic volume was large in DP cases, we evaluated the index considering the PRPV but did not obtain significant results. In a previous report, PRPV was not a risk factor for NODM after DP, and obese patients were significantly more likely to develop NODM [19]. Preoperative BMI was a candidate predictor of event occurrence in DP patients of this study, but it was not significant on multivariate analysis.

Regulation of insulin secretion is through stimulation by hyperglycemia or effects of various endocrine hormones. Incretin is a general term for gastrointestinal-derived hormones that act on pancreatic β-cells to promote insulin secretion [20,21]. Glucagon-like peptide-1 (GLP-1) and gastric inhibitory peptide are known as incretins, and their mechanisms of enhancing incretin effects and inhibiting metabolism have been elucidated and clinically applied as therapeutic targets for DM [20,21]. Muscogiuri et al. showed that duodenal resection in PD resulted in an increase in GLP-1 concentration, which was expected to increase insulin secretory capacity, but in fact, they reported a decrease in insulin secretory capacity [22]. This decrease is hypothesized to be the result of the suppression of insulin secretion by increased glucagon levels [22]. Although we did not measure glucagon or incretin levels before and after surgery in this study, we believe that it is necessary to understand the behavior of hormones involved in glucose tolerance before and after surgery to elucidate the pathogenesis of changes in glucose tolerance due to pancreatectomy, and this is a topic for future studies.

In the present study, most non-diabetic patients with post-CPI <1.0 did not progress to diabetes that would require insulin therapy postoperatively. It is hypothesized that factors other than a short-term decrease in BISC are involved in the development of NODM. Mezza et al. stated that even if there is forced removal of β-cells by pancreatectomy, the function of the remaining β-cells will not lead to insulin deficiency if there is no insulin resistance [23]. However, if there is insulin resistance at the time of pancreatectomy, the residual β-cells overload postoperatively, leading to insulin deficiency and NODM. In other words, the true determinant of the development of NODM is not forced removal of β-cells by pancreatectomy but the presence or absence of a pre-existing diabetic milieu, which may lead to type 2 DM, a condition accelerated by surgery [23]. Similar to the previous report, in this study, preoperative non-diabetic patients had a significantly lower rate of postoperative insulin introduction than patients on preoperative diabetic medication. These results suggest that the coexistence of pancreatectomy induced BISC reduction and insulin resistance, which is considered one of the pathophysiological features of type 2 DM, may promote postoperative insulin deficiency, resulting in the need for insulin therapy.

Although the etiology of insulin resistance is not fully understood, obesity has been identified as an acquired risk [23]. Patients who are expected to have decreased BISC after pancreatectomy may reduce the development of NODM by weight control based on appropriate diet and exercise therapy, with appropriate intervention through follow-ups being essential in the future.

This study has some limitations, including the single-center, retrospective design. Using the three explanatory variables of age, sex, and predictors of event occurrence in multivariate analysis to determine the minimum sample size, a minimum of 30 cases of event occurrence were required per surgical procedure, and assuming an event occurrence rate of 40%, the total number of cases should be at least 75 each per surgical procedure. Since the overall numbers of cases and events were limited in this study, further analysis with a larger number of cases is crucial. Furthermore, medium- to long-term follow-up is necessary to clarify how early postoperative decline in BISC changes over the medium to long term and whether it leads to the development of NODM.

Conclusions

Preoperative CPI was determined to be the most useful factor for predicting a decrease in BISC after pancreatectomy in patients who underwent PD. Predicting postoperative BISC decline and simultaneously confirming the presence of insulin resistance can be useful in predicting the development of NODM after pancreatectomy.

List Of Abbreviations

95% CI 95% confidence interval

AUC area under the receiver operator characteristic curve

BISC basal insulin secretion capacity

BMI body mass index

CPI C-peptide index

CT computed tomography

DGE delayed gastric emptying

DM diabetes mellitus

DP distal pancreatectomy

HbA1C hemoglobin A1c

HOMA-β homeostasis model assessment of β-cell function

K-ICG K indocyanine green

NODM new-onset diabetes mellitus

PD pancreaticoduodenectomy

POPF postoperative pancreatic fistula

ROC receiver operator characteristic

SD standard deviation

Declarations

Ethics approval and consent to participate

The study was conducted as per the Declaration of Helsinki (as revised in 2013) guidelines. The study was approved by the ethics board of the Japanese Red Cross Kumamoto Hospital (permission no. 474), and the requirement for informed consent was waived due to the retrospective nature of this study. We posted a summary of the trial on our website and asked eligible patients to inform us if they wished to be excluded from the study; none of the participants requested exclusion.

Consent for publication

Not applicable.

Availability of data and materials

The datasets analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests

All authors declare no competing interests.

Funding

No funding was obtained to conduct the research presented in this manuscript.

Authors’ contributions

YK, a hepatobiliary pancreatic surgeon, performed all the surgeries, designed the study, collected the data, performed the statistical analysis, and drafted the manuscript. KM, a statistician, evaluated the collected data and guided the statistical analysis. HY, a senior surgeon, advised on the research and approved the final manuscript. All authors read and approved the final manuscript.

Acknowledgements

The authors acknowledge the support of all medical and surgical staff who cared for the patients in this study.

References

  1. McMillan MT, Vollmer CM Jr, Asbun HJ, Ball CG, Bassi C, Beane JD, et al. The characterization and prediction of ISGPF Grade C fistulas following pancreatoduodenectomy. J Gastrointest Surg. 2016;20:262–76. Doi: 10.1007/s11605-015-2884-2.
  2. Hirono S, Shimokawa T, Nagakawa Y, Shyr YM, Kawai M, Matsumoto I, et al. Risk factors for pancreatic fistula grade C after pancreatoduodenectomy: A large prospective, multicenter Japan-Taiwan collaboration study. J Hepatobiliary Pancreat Sci. 2020;27:622–31. Doi: 10.1002/jhbp.799.
  3. Franklin BE, Crisler SC, Shappley R, Armour MM, McCommon DT, Ferry RJ. Real-time support of pediatric diabetes self-care by a transport team. Diabetes Care. 2014;37:81–7. Doi: 10.2337/dc13-1041.
  4. Gersell DJ, Gingerich RL, Greider MH. Regional distribution and concentration of pancreatic polypeptide in the human and canine pancreas. Diabetes. 1979;28:11–5. Doi: 10.2337/diabetes.28.1.11.
  5. Henquin JC, Ibrahim MM, Rahier J. Insulin, glucagon and somatostatin stores in the pancreas of subjects with type-2 diabetes and their lean and obese non-diabetic controls. Sci Rep. 2017;7:11015. Doi: 10.1038/s41598-017-10296-z.
  6. Wang X, Misawa R, Zielinski MC, Cowen P, Jo J, Periwal V, et al. Regional differences in islet distribution in the human pancreas—preferential beta-cell loss in the head region in patients with type 2 diabetes. PLOS ONE. 2013;8:e67454. Doi: 10.1371/journal.pone.0067454.
  7. Wu L, Nahm CB, Jamieson NB, Samra J, Clifton-Bligh R, Mittal A, et al. Risk factors for development of diabetes mellitus (Type 3c) after partial pancreatectomy: a systematic review. Clin Endocrinol (Oxf). 2020;92:396–406. Doi: 10.1111/cen.14168.
  8. Nguyen A, Demirjian A, Yamamoto M, Hollenbach K, Imagawa DK. Development of postoperative diabetes mellitus in patients undergoing distal pancreatectomy versus Whipple procedure. Am Surg. 2017;83:1050–3. Doi: 10.1177/000313481708301007.
  9. Maignan A, Ouaïssi M, Turrini O, Regenet N, Loundou A, Louis G, et al. Risk factors of exocrine and endocrine pancreatic insufficiency after pancreatic resection: A multi-center prospective study. J Visc Surg. 2018;155:173–81. Doi: 10.1016/j.jviscsurg.2017.10.007.
  10. Gotoh M, Monden M, Okamura J, Mori T, Shima K. Insulin and glucagon secretion after pancreatectomies. Correlation of secretion and hormonal contents of remaining pancreas. Diabetes. 1989;38:861–7. Doi: 10.2337/diab.38.7.861.
  11. Yamada D, Takahashi H, Asukai K, Hasegawa S, Wada H, Matsuda C, et al. Investigation of the influence of pancreatic surgery on new-onset and persistent diabetes mellitus. Ann Gastroenterol Surg. 2021;5:575–84. Doi: 10.1002/ags3.12435.
  12. Iwata M, Matsushita Y, Fukuda K, Wakura T, Okabe K, Koshimizu Y, et al. Secretory units of islets in transplantation index is a useful predictor of insulin requirement in Japanese type 2 diabetic patients. J Diabetes Investig. 2014;5:570–80. Doi: 10.1111/jdi.12181.
  13. Funakoshi S, Fujimoto S, Hamasaki A, Fujiwara H, Fujita Y, Ikeda K, et al. Utility of indices using C-peptide levels for indication of insulin therapy to achieve good glycemic control in Japanese patients with type 2 diabetes. J Diabetes Investig. 2011;2:297–303. Doi: 10.1111/j.2040-1124.2010.00096.x.
  14. Bassi C, Marchegiani G, Dervenis C, Sarr M, Abu Hilal M, Adham M, et al. The 2016 update of the International Study Group (ISGPS) definition and grading of postoperative pancreatic fistula: 11 years After. Surgery. 2017;161:584–91. Doi: 10.1016/j.surg.2016.11.014.
  15. Wente MN, Bassi C, Dervenis C, Fingerhut A, Gouma DJ, Izbicki JR, et al. Delayed gastric emptying (DGE) after pancreatic surgery: a suggested definition by the International Study Group of Pancreatic Surgery (ISGPS). Surgery. 2007;142:761–8. Doi: 10.1016/j.surg.2007.05.005.
  16. Nagino M, Kamiya J, Nishio H, Ebata T, Arai T, Nimura Y. Two hundred forty consecutive portal vein embolizations before extended hepatectomy for biliary cancer: surgical outcome and long-term follow-up. Ann Surg. 2006;243:364–72. Doi: 10.1097/01.sla.0000201482.11876.14.
  17. Okochi O, Kaneko T, Sugimoto H, Inoue S, Takeda S, Nakao A. ICG pulse spectrophotometry for perioperative liver function in hepatectomy. J Surg Res. 2002;103:109–13. Doi: 10.1006/jsre.2001.6328.
  18. Youden WJ. Index for rating diagnostic tests. Cancer. 1950;3:32 – 5. Doi: 10.1002/1097-0142(1950)3:1<32::aid-cncr2820030106>3.0.co;2-3.
  19. Kim KJ, Jeong CY, Jeong SH, Ju YT, Jung EJ, Lee YJ, et al. Pancreatic diabetes after distal pancreatectomy: incidence rate and risk factors. Korean J Hepatobiliary Pancreat Surg. 2011;15:123–7. Doi: 10.14701/kjhbps.2011.15.2.123.
  20. Doyle ME, Egan JM. Mechanisms of action of glucagon-like peptide 1 in the pancreas. Pharmacol Ther. 2007;113:546–93. Doi: 10.1016/j.pharmthera.2006.11.007.
  21. Seino Y, Yabe D. Glucose-dependent insulinotropic polypeptide and glucagon‐like peptide‐1: incretin actions beyond the pancreas. J Diabetes Investig. 2013;4:108–30. Doi: 10.1111/jdi.12065.
  22. Muscogiuri G, Mezza T, Prioletta A, Sorice GP, Clemente G, Sarno G, et al. Removal of duodenum elicits GLP-1 secretion. Diabetes Care. 2013;36:1641–6. Doi: 10.2337/dc12-0811.
  23. Mezza T, Cefalo CMA, Cinti F, Quero G, Pontecorvi A, Alfieri S, et al. Endocrine and metabolic insights from pancreatic surgery. Trends Endocrinol Metab. 2020;31: 760 – 72. Doi: 10.1016/j.tem.2020.07.003.

Tables

Tables 1 to 7 are available in the Supplementary Files section.