Involvement of capillary endothelial dysfunction and decreased plasma calcium concentration in persistent severe hypotension during pancreaticoduodenectomy Authors: Yinan Chen, Fan Yu, Shi Shen, Xingjun Wang, Nana Li, Rong Ma, Zhaochu Sun, Youli Hu Published: 2026 ABSTRACT BackgroundIntraoperative persistent severe hypotension (PSH) is a common complication following pancreaticoduodenectomy (PD) and may affect the prognosis of patients. The aim of this trial was to explore the ultrastructural alterations of capillary endothelia cells and relevant factors that is associated with PSH and postoperative outcomes in patients undergoing PD.MethodsTotal 103 patients were divided into PSH or non-PSH group during PD. The basic clinical features, hemodynamic monitoring, electrolyte biochemical indexes, postoperative inflammatory biomarkers and adverse outcomes were analyzed. Particularly, the great omentum samples were collected and the morphological changes of capillary endothelial cells were revealed by electronic microscopy.ResultsPSH occurred in forty-four patients (42.7%) during PD. The PSH patients experienced greater extensive surgery scope, more bleeding, longer surgery time, increased fluid infusion and blood glucose. Hemodynamic monitoring revealed that the PSH patients primarily showed reduced vascular resistance with enhanced cardiac index. The postoperative inflammation biomarkers and incidence of postoperative complications in PSH group were significantly higher than those in non-PSH group (P < 0.05). Notably, the arterial plasma Ca2+ decreased significantly in PSH patients as that in non-PSH group (P < 0.001). Electron microscopy revealed in PSH patients ultrastructural evidence of capillary endothelial injury, including cellular swelling and increased pinocytotic activity.ConclusionEndothelial cell permeability alteration and endothelial dysfunction might potentially reflect intensive surgical stress-inflammation-endocrine responses during PD. Endothelial edema formation and decreased plasma Ca2+ are probably associated with PSH and postoperative adverse outcomes of PD patients.Our study highlights the key point to keep the functionality and integrity of endothelia cells and timely supplement calcium to prevent PSH during pancreaticoduodenectomy and postoperative outcomes.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12871-026-04002-8. Introduction Intraoperative persistent severe hypotension (PSH, mean arterial pressure MAP<65mmHg) occurs quite often during/after major surgery and is associated with post-operative complications [1, 2]. Hypovolaemia (blood loss and insufficient fluid supplement), decreased cardiac function due to intravenous and volatile anaesthetic drugs and surgical manipulations are well-proved major reasons for the intraoperative hypotension. Adjustment of anesthetic drugs for proper anesthesia depth, goal-directed fluid therapy (GDT) and vasopressor treatment are generally considered as standardized paradigm to keep hemodynamic stable. However, such sort of strategies cannot eliminate hemodynamic dysfunction in some circumstances especially during major surgery [3]. Endothelia cells lining the blood vessels sense the surgical-related mechanical and chemical signals and play key role in modulating endothelia permeability and controlling vascular tone, therefore endothelia dysfunction potentially leading to intra-operative hypovolemia and decreased peripheral vascular resistance. In fact, tissue ischemia, surgical procedures and following inflammatory response inevitably cause post-operative capillary leakage [4, 5]. However, intra-operative endothelia dysfunction evident from morphological alterations and associated postoperative adverse outcomes during major surgery are less described. The capillary endothelium creates a semi-permeable barrier to regulate tissue fluid hemostasis through controlling the distribution of macromolecules and fluid between the circulation and tissues. Disruption of endothelial transport either via transcelluar (for albumin, water et ac) and paracellular (for ion, urea and glucose) pathways results in abnormal endothelia permeability to plasma proteins and liquid [6, 7]. Excessive oxidant stress that overproduces reactive oxygen species, ischemia and inflammation can trigger the influx of extracellular Ca2+ into endothelia cells, leading to aberrant intracellular Ca2+ signaling for capillary leakage and the loss of control on vascular tone [8]. Indeed, we recently report that in a large retrospective cohort study, lower plasma Ca2+ act as one of risk factors associated with intraoperative persistent hypotension during pancreaticoduodenectomy [9]. It has been described that acute pancreatitis and pancreaticoduodenectomy (PD) cause excessive inflammatory response and microcirculation dysfunction [10, 11]. As thus, endothelia permeability disturbance and disruption of Ca2+ distribution due to PD-related surgical insults might form the cellular and molecular basis for PSH during and after major surgery. Consequently, we compared patients with and without PSH during PD on hemodynamic parameters, plasma Ca2+ concentration, and performed electron microscopic analysis on the changes of endothelial structure to identify the association of capillary endothelial dysfunction and decreased plasma Ca2+ concentration with PSH and postoperative adverse outcomes. Materials and methods Ethics This was a prospective cohort study conducted in The First Affiliated Hospital of Nanjing Medical University from January 2020 to June 2023 (chictr.org.cn ChiCTR2000029922; February 17, 2020). The study was approved by the ethics committee (The First Affiliated Hospital of Nanjing Medical University/Jiangsu Province Hospital, China, July 2, 2019) and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participating patients before inclusion in the study. An independent data and safety monitoring board oversaw the study conduct and reviewed blinded safety data. Study design Patients and grouping We enrolled the patients who were diagnosed with pancreatic tumor and planned to undergo PD surgery under general anesthesia. Participants were screened, and recruited by study staff, who evaluated patient eligibility, obtained informed consent, and enrolled the participants. Exclusion criteria included poor general condition, with severe cardiovascular and pulmonary disease, preoperative severe liver and kidney dysfunction, or patient refusal, and with extensive abdominal metastasis. In this study, PSH defined as MAP < 65 mmHg or the reduction of MAP more than 30% of the base value, and the accumulated time of hypotension was still more than 30 min after corresponding hypertensive treatment, or required continuous noradrenaline infusion during the surgery [1, 12]. Perioperative management All patients received elective surgical treatment and admitted to hospital in advance to complete relevant examinations which were performed by the same team of surgeons. In keeping with institutional protocol, patients fasted 8 h before surgery. General anesthesia was administered according to standardized procedure. ProAQT (PULSION, Germany) were applied to constantly monitor hemodynamic parameters for GDT. Anaesthesia was induced with midazolam (0.05 mg·kg− 1), propofol (1.5-2.0 mg·kg− 1), fentanyl (3–5 µg·kg− 1) and cisatracurium (0.2 mg·kg− 1) that facilitated tracheal intubation. Sevoflurane, propofol and remifentanil were used for anesthesia maintenance, cisatracurium for muscle relaxation, and fentanyl for pain relief. The proper depth of anesthesia was monitored by BIS (Medtronic, USA) around 40–60. Monitoring included ECG, pulse oximetry, invasive arterial blood pressure, central venous pressure, nasopharyngeal temperature, cardiac and vascular function assessment by pulse contour analysis (Nexfin® (BMEYE, )) and ProAQT monitoring. A urinary catheter and nasogastric tube were inserted. Normothermia was maintained by active warming (Air system (3 M, USA)) and heating fluid . The surgical procedure was a standardized PD including resection of the pancreatic head, duodenum, gall bladder, common bile duct and the gastric antrum. Reconstruction was performed with a pancreaticojejunostomy, hepaticojejunostomy and gastrojejunostomy on the same jejunal loop. The pancreaticojejunostomy was performed as duct-to-mucosa procedure or as an invaginated end-to-side anastomosis depending on the calibre of the pancreatic duct. In case of porto-mesenteric vein resection, venous reconstruction as performed as an end-to-end anastomosis, or less commonly, with a bridging graft. Vasopressors (Phenylephrine/ Norepinephrine) were initiated after anaesthesia induction to maintain MAP at or above 65 mmHg. Fluid therapy consisted of a background infusion of ringers lactate 5 ml·kg− 1·h− 1 throughout the procedures and GDT by a standardized stroke volume variation optimization algorithm: initial assessment and optimization began after anaesthesia induction and prior to surgical incision with the patient in the supine position for 5 min. A bolus of 250 ml Hydroxyethyl starch 130/0.4 (HES) was administered and if SVV > 13%, a new bolus of 250 ml HES or human albumin (HA) 5% was administered until SVV ≤ 13%. Maximum total HA 5% dosage was 1000 ml and additional SVV optimization included fresh frozen plasma (FFP). Packed red blood cells were transfused when the hemoglobin level was less than 8 g/dL in patients without cardiovascular disease [13, 14]. Variables and outcomes The primary outcome was the hemodynamic change, including mean arterial pressure, cardiac index, stroke volume index, systemic vascular resistance index, left ventricular contractility, stroke volume variation, pulse pressure variability. These variables were monitored continuously by ProAQT and recorded at the following timepoints: 5 min before anesthetic induction (T1), 5 min after anesthetic induction (T2), 5 min after surgery began (T3), the lowest blood pressure after lesion resection or before the infusion of noradrenaline (T4), at the end of the surgery (T5). In addition, the incidence of PSH was recorded. Secondary outcomes included blood glucose, electrolyte levels and biochemical results of the patients perioperatively. Furthermore, the electron microscopy analysis on the capillary endothelial cells of the great omentum at the timepoint when the abdomen was opened and before closed were compared between the PSH and non-PSH groups. The postoperative complications included postoperative pancreatic fistula, postoperative pulmonary complications, intra-abdominal infections, delayed gastric emptying, and acute kidney injury. Pancreatic fistula was defined as any measurable volume of drain fluid on or after postoperative day 3 with amylase level > 3 times the upper limit of normal amylase for each specific institution [15]. Postoperative Pulmonary Complications includes: Respiratory infection, Respiratory failure, Pleural effusion, Atelectasis, Pneumothorax, Bronchospasm, and Aspiration pneumonitis [16]. Intra-abdominal infection describes a diverse set of diseases. It is broadly defined as peritoneal inflammation in response to microorganisms, resulting in purulence in the peritoneal cavity [17]. Delayed gastric emptying was defined and graded according to the consensus criteria established by the International Study Group of Pancreatic Surgery (ISGPS). It was defined as the inability to return to a standard diet by the end of the first postoperative week, necessitating prolonged nasogastric tube (NGT) intubation, with mechanical causes of gastric outlet obstruction excluded by endoscopy or contrast studies. Three. grades of severity were defined based on the duration of NGT requirement and the postoperative day (POD) by which the patient could tolerate solid food: Grade A, requiring NGT for 4–7 days or reinsertion after POD 3 with inability to tolerate solid food by POD 7; Grade B, requiring NGT for 8–14 days or reinsertion after POD 7 with inability to tolerate solid food by POD 14; and Grade C, requiring NGT for more than 14 days or reinsertion after POD 14 with inability to maintain oral intake by POD 21 [18]. Acute kidney injury was defined using the serum creatinine (SCr) criteria of the kidney disease: Improving Global Outcomes (KDIGO) guidelines: an increase in SCr by ≥ 0.3 mg/dL (≥ 26.5 µmol/L) within 48 h or a ≥ 1.5-fold increase from baseline within 7 days or urine volume < 0.5 ml/kg/hour for 6 h [19]. Transmission electron microscopy analysis on ultrastructural alterations. When the abdomen was opened and before closed, small pieces of the great omentums were collected and cut into 1cm3 to be fixed in 2.5% glutaraldehyde (Sigma, G5882) in 0.1 M phosphate buffer (pH 7.4) for 2 h, postfixed in 1% osmium tetroxide in the same buffer at 4˚C for 2 h, incubated in 2% uranyl acetate at 4˚C for 2 h, dehydrated in a graded series of ethanol (50%, 70%, 90%, 95%, and 100%), embedded in Epon812 resin (SPI-CHEM, GS02660). After polymerizing for 48 h at 60 °C, ultrathin sections were cut on an ultramicrotome (Leica, UC7), and stained with uranyl acetate and lead citrate. Image acquisition was performed by TEM (FEI, Tecnai Spirit G2) at an acceleration voltage of 120 kV. Ultrastructural changes of capillary endothelial morphology, nucleus, pinocytotic and coated vesicles, mitochondria, intracellular filaments and intercellular junctions were analyzed. Specimen identification was managed exclusively by a separate investigator using patient name and hospital admission number without disclosure of group assignment, and all TEM imaging was performed by a single investigator who was blinded to group allocation, thereby ensuring single-blind conditions throughout the imaging process. A total of 9 capillary profiles were examined in the non-hypotension group and 8 in the hypotension group. For each specimen, all identifiable capillary profiles present on the section were systematically surveyed at low magnification, and representative fields were selected at higher magnification to document predominant ultrastructural features; image selection was based on consistent and reproducible morphological characteristics observed across multiple fields within each group. All images meeting these predefined criteria were included regardless of whether findings were favorable or unfavorable to the study hypothesis. Statistical analysis Statistical analyses were performed using IBM SPSS Statistics (Version 20, IBM SPSS, Chicago, IL, USA) and GraphPad Prism (Version 8, GraphPad Software, La Jolla, CA, USA). Continuous data were expressed as a mean ± (SD) or median (range) according to normality. The quantitative data were expressed as mean ± standard deviation, and one-way Anova test or independent sampled t-test was used to compare the variables between groups. The Chi-square test or Fisher’s exact test was used for comparison of the qualitative data between the groups. A P value less than < 0.05 was considered significant. Results A total of 120 patients were approached for participation in the study (Supplementary Fig. 1). Recruitment ended when the number of included subjects reached the calculated required sample size. 17 patients were not enrolled or randomised because of a variety of reasons including severe cardiovascular disease (n = 2), refusal to participate (n = 6), and pancreaticoduodenectomy was not performed (n = 9). 103 subjects were enrolled and divided into PSH group or non-PSH group based on the presence or absence of intraoperative hypotension, the HR and MAP during PD in the two groups are shown in Supplementary Fig. 2. Data from all 103 subjects were analyzed with no subjects lost during follow-up. The clinical characteristics and surgery-related variables of the two groups were shown in Table 1. The results demonstrated that compared with the non-PSH group, the patients in the PSH group suffered a larger operation scope, longer surgery time and more fluid loss (P < 0.001). Definitely, these patients also received more fluid therapy accordingly (P < 0.001). However, the length of ICU stay (P = 0.064) and postoperative stay (P = 0.211) were not significantly different between the two groups. Table 1Clinical characteristics and surgery-related variables of patientsNon-PSH group(n = 59)PSH group(n = 44)P-valueAge (yr)62.3 (9.8)63.2 (8.5)0.648Male (n; %)31 (52.5%)23 (52.3%)0.978BMI (kg/m2)22.6 (2.6)23.9 (2.7)0.015ASA physical status (n; %)0.291 I3 (5.1%)5 (11.4%) II46 (78.0%)35 (79.5%) III10 (16.9%)4 (9.1%)Comorbidity (n; %) Hypertension26 (44.1%)27 (61.4%)0.181 Diabetes10 (16.9%)14 (31.8%)0.077 Coronary heart disease7 (11.9%)4 (9.1%)0.898Protopathy (n; %)0.468 PDAC39 (66.1%)30 (68.2%) IPMN6 (10.2%)8 (18.2%) SCA5 (8.5%)2 (4.5%) SPTP3 (5.1%)2 (4.5%) PNET2 (3.4%)2 (4.5%)Duodenal Tumor4 (6.8%)0Scope of surgery (n; %) Heidelberg triangle8 (13.6%)16 (36.4%)0.007 Superior mesenteric artery20 (33.9%)29 (65.9%)0.001 Glycosylated hemoglobin (mmol/L)6.4 (1.5)6.4 (1.2)0.912 Surgery time (min)256.4 (62.8)329.4 (69.5)< 0.001Total infusion volume (m)2700.0 [2300.0, 3100.0]3237.5 [2800.0, 3932.5]< 0.001Crystalloid solution (ml)1600.0 [1500.0, 2100.0]2000.0 [1600.0, 2100.0]0.061Colloid solution (ml)700.0 [500.0, 1000.0]1000.0 [775.0, 1200.0]0.0075% Albumin (ml)745.8 (439.2)795.5 (408.0)0.560Blood transfusion (n; %)8 (13.6%)19 (43.2%)0.001Total loss volume (ml)600 [460.0, 750.0]900.0 [725.0, 1100.0]< 0.001Urine volume (ml)300.0 [300.0, 600.0]500.0 [362.5, 737.5]0.090Blood loss (ml)200.0 [150.0, 300.0]400.0 [262.5, 737.5]< 0.001BIS value46.4 (6.3)47.4 (6.0)0.397Duration of ICU stay (d)0 [0, 0]0 [0, 2]0.064Duration of postoperative stay (d)14.0 [11.0, 19.0]16.0 [13.0, 21.0]0.211Abbreviations BMI Body mass index, ASA American Society of Anesthesiologists, BIS Bispectral index, PDAC Pancreatic ductal adenocarcinoma, IPMN Intraductal papillary mucinous neoplasm, SCA Serous cystadenoma, SPTP Solid pseudopapillary tumor, PNET Pancreatic neuroendocrine tumor Table 1Clinical characteristics and surgery-related variables of patientsNon-PSH group(n = 59)PSH group(n = 44)P-valueAge (yr)62.3 (9.8)63.2 (8.5)0.648Male (n; %)31 (52.5%)23 (52.3%)0.978BMI (kg/m2)22.6 (2.6)23.9 (2.7)0.015ASA physical status (n; %)0.291 I3 (5.1%)5 (11.4%) II46 (78.0%)35 (79.5%) III10 (16.9%)4 (9.1%)Comorbidity (n; %) Hypertension26 (44.1%)27 (61.4%)0.181 Diabetes10 (16.9%)14 (31.8%)0.077 Coronary heart disease7 (11.9%)4 (9.1%)0.898Protopathy (n; %)0.468 PDAC39 (66.1%)30 (68.2%) IPMN6 (10.2%)8 (18.2%) SCA5 (8.5%)2 (4.5%) SPTP3 (5.1%)2 (4.5%) PNET2 (3.4%)2 (4.5%)Duodenal Tumor4 (6.8%)0Scope of surgery (n; %) Heidelberg triangle8 (13.6%)16 (36.4%)0.007 Superior mesenteric artery20 (33.9%)29 (65.9%)0.001 Glycosylated hemoglobin (mmol/L)6.4 (1.5)6.4 (1.2)0.912 Surgery time (min)256.4 (62.8)329.4 (69.5)< 0.001Total infusion volume (m)2700.0 [2300.0, 3100.0]3237.5 [2800.0, 3932.5]< 0.001Crystalloid solution (ml)1600.0 [1500.0, 2100.0]2000.0 [1600.0, 2100.0]0.061Colloid solution (ml)700.0 [500.0, 1000.0]1000.0 [775.0, 1200.0]0.0075% Albumin (ml)745.8 (439.2)795.5 (408.0)0.560Blood transfusion (n; %)8 (13.6%)19 (43.2%)0.001Total loss volume (ml)600 [460.0, 750.0]900.0 [725.0, 1100.0]< 0.001Urine volume (ml)300.0 [300.0, 600.0]500.0 [362.5, 737.5]0.090Blood loss (ml)200.0 [150.0, 300.0]400.0 [262.5, 737.5]< 0.001BIS value46.4 (6.3)47.4 (6.0)0.397Duration of ICU stay (d)0 [0, 0]0 [0, 2]0.064Duration of postoperative stay (d)14.0 [11.0, 19.0]16.0 [13.0, 21.0]0.211Abbreviations BMI Body mass index, ASA American Society of Anesthesiologists, BIS Bispectral index, PDAC Pancreatic ductal adenocarcinoma, IPMN Intraductal papillary mucinous neoplasm, SCA Serous cystadenoma, SPTP Solid pseudopapillary tumor, PNET Pancreatic neuroendocrine tumor Clinical characteristics and surgery-related variables of patients Abbreviations BMI Body mass index, ASA American Society of Anesthesiologists, BIS Bispectral index, PDAC Pancreatic ductal adenocarcinoma, IPMN Intraductal papillary mucinous neoplasm, SCA Serous cystadenoma, SPTP Solid pseudopapillary tumor, PNET Pancreatic neuroendocrine tumor As seen in Fig. 1, compared with the non-PSH group, the PSH patients primarily present decreased systemic vascular resistance with a significant reduction of SVRI (P < 0.001), but there was no significant difference in myocardial contractility as shown in dPmx (P = 0.051). Notably, the cardiac index (P = 0.002) and stroke volume index (P = 0.028) were significantly higher in the PSH group than that in the non-PSH group, reminiscent of the consequence of the decreased vascular resistance. Fig. 1Perioperative variables of hemodynamic parameters. The hemodynamic parameters of the two groups at 5 time points during PD were shown: T1, 5 min before anesthetic induction; T2, 5 min after anesthetic induction; T3, 5 min after surgery began; T4, the lowest blood pressure after lesion resection or before the infusion of noradrenaline; T5, at the end of the surgery. *P < 0.05, **P < 0.01, PSH group versus non-PSH group at certain timepoints. Abbreviations: CI, cardiac index; SVI, stroke volume index; SVRI, systemic vascular resistance index; dPmx, left ventricular contractility; SVV, stroke volume variation; PPV, pulse pressure variability Perioperative variables of hemodynamic parameters. The hemodynamic parameters of the two groups at 5 time points during PD were shown: T1, 5 min before anesthetic induction; T2, 5 min after anesthetic induction; T3, 5 min after surgery began; T4, the lowest blood pressure after lesion resection or before the infusion of noradrenaline; T5, at the end of the surgery. *P < 0.05, **P < 0.01, PSH group versus non-PSH group at certain timepoints. Abbreviations: CI, cardiac index; SVI, stroke volume index; SVRI, systemic vascular resistance index; dPmx, left ventricular contractility; SVV, stroke volume variation; PPV, pulse pressure variability We further analyzed the difference in electrolyte and biochemical levels between the two groups that might potentially be associated with persistent hypotension (Table 2). In PSH patients, blood glucose (P = 0.025) and plasma lactic acid (P = 0.007) were significantly higher with much lower arterial calcium concentration (P < 0.001) during PD. The postoperative albumin level in the PSH group was significantly lower than that in the non-PSH group (P = 0.007). Preoperative serum Ca²⁺ levels at T1 were comparable between groups (P > 0.05). Table 2Changes of biochemical indexes and internal environmental levelsTime-pointsNon-PSH group(n = 59)PSH group(n = 44)P-valuePH valueaT17.45 (0.03)7.45 (0.02)0.280T47.39 (0.06)7.39 (0.05)0.675Blood glucose (mmol/L)aT16.65 (2.25)6.76 (1.98)0.786T48.75 (2.74)9.96 (2.59)0.025Hemoglobin(g/dL)aT113.09 (1.72)13.31 (1.46)0.487T411.32 (1.53)11.41 (1.53)0.786Ca2+ (mmol/L)aT11.12 (0.05)1.11 (0.05)0.428T41.07 (0.05)1.00 (0.05)< 0.001K+ (mmol/L)aT13.58 (0.33)3.58 (0.35)0.965T43.73 (0.46)3.70 (0.43)0.758Plasma lactic acidaT11.02 (0.50)0.98 (0.36)0.60T41.21 (0.73)1.65 (0.91)0.007Albumin (g/L)bPre-operative39.45 (2.98)40.13 (2.78)0.240Postoperative36.23 (4.04)34.05 (3.91)0.007Creatinine(µmol/L)bPre-operative64.02 (18.25)62.95 (14.06)0.749Postoperative61.78 (19.50)65.60 (15.94)0.290Blood urea nitrogen(mmol/L)bPre-operative5.23 (1.61)5.78 (1.89)0.117Postoperative7.54 (2.44)8.05 (2.16)0.279aThe data were obtained by arterial blood gas analysis intraoperativelybThe data were collected from serum biochemical tests by the day on admission and the first day after surgery Table 2Changes of biochemical indexes and internal environmental levelsTime-pointsNon-PSH group(n = 59)PSH group(n = 44)P-valuePH valueaT17.45 (0.03)7.45 (0.02)0.280T47.39 (0.06)7.39 (0.05)0.675Blood glucose (mmol/L)aT16.65 (2.25)6.76 (1.98)0.786T48.75 (2.74)9.96 (2.59)0.025Hemoglobin(g/dL)aT113.09 (1.72)13.31 (1.46)0.487T411.32 (1.53)11.41 (1.53)0.786Ca2+ (mmol/L)aT11.12 (0.05)1.11 (0.05)0.428T41.07 (0.05)1.00 (0.05)< 0.001K+ (mmol/L)aT13.58 (0.33)3.58 (0.35)0.965T43.73 (0.46)3.70 (0.43)0.758Plasma lactic acidaT11.02 (0.50)0.98 (0.36)0.60T41.21 (0.73)1.65 (0.91)0.007Albumin (g/L)bPre-operative39.45 (2.98)40.13 (2.78)0.240Postoperative36.23 (4.04)34.05 (3.91)0.007Creatinine(µmol/L)bPre-operative64.02 (18.25)62.95 (14.06)0.749Postoperative61.78 (19.50)65.60 (15.94)0.290Blood urea nitrogen(mmol/L)bPre-operative5.23 (1.61)5.78 (1.89)0.117Postoperative7.54 (2.44)8.05 (2.16)0.279aThe data were obtained by arterial blood gas analysis intraoperativelybThe data were collected from serum biochemical tests by the day on admission and the first day after surgery Changes of biochemical indexes and internal environmental levels Hemoglobin (g/dL)a Creatinine (µmol/L)b Blood urea nitrogen (mmol/L)b aThe data were obtained by arterial blood gas analysis intraoperatively bThe data were collected from serum biochemical tests by the day on admission and the first day after surgery To further identify cellular basis underlying PSH during PD, we performed electronic microscopy to investigate the ultrastructural alterations of capillary endothelial cells for the submicroscopic characteristics of capillary permeability and endothelia vascular tone. We studied on the capillary vessels of the great omentum, focusing on endothelia morphology, modifications of vesicular and vacuolar transendothelial transport, the alterations of endothelial junction and basement membrane. As seen in Fig. 2, the capillary endothelia cells showed smooth and regular morphology with flattened nucleus, mitochondrial, transport vesicles and intact intercellular tight junctions at the timepoint when the abdomen was opened (Fig. 2A) and before closed in non-PSH patients (Fig. 2B and Supplementary Fig. 3). By contrast, endothelia cells in the great omentum turned to be swelling at the time before the abdomen was closed in PSH patients (Fig. 2C and D). Endothelia nucleus in PSH patients became round or project into the capillary lumen, making it moderately obstructed (Fig. 2C), while normal elongated or oval nucleus was observed in non-PSH patients. Prominence of nucleus is considered as a sign of endothelial swelling and edema. Consistently, the basement membrane became enlarged and rarefied in PSH patients, similar observations being reported in human cerebral edema [20]. In addition, transendothelial transport pathway was comparatively activated including pinocytosis and coated vesicles derived from the invaginations of the luminal membrane and large vacuoles formation in PSH patient. It is known that over-activation of transendothelial exchange of materials between the capillary lumen and the surrounding tissue is related to the severe edema. The swelling features were even more severe in the second PSH case with dramatically narrowed lumen, leading to one blood cell being stuck inside the capillary vessel (Fig. 2D). There were several pore formations and discontinuous points of the luminal membrane, further supporting the damage of capillary endothelial vessels for the edema and fluid leakage. Fig. 2Ultrastructure alterations revealed by the high-resolution transmission electron microscopy on capillary endothelial cells of the great omentum during PD at the timepoint when the abdomen was opened (A) and before closed (B, C and D) in non-PSH (B) and PSH patients including two cases (C for case 1 and D for case 2). A. When the abdomen was opened, the endothelia cells showed the normal morphology with flattened nucleus, mitochondrial, transport vesicles, interacted cytoskeletal filaments and intact endothelial junctions (white arrow). B. In non-PSH patients before the abdomen was closed, the endothelia cells showed the normal morphology but with slightly more pinocytotic and coated vesicles (black arrow). C. Two representative images from the first PSH patients. Endothelial nucleus became round (upper panel) or prominent towards lumen (lower panel). There were numerous pinocytotic and coated vesicles (black arrow) and formation of large vacuoles (V). A bit more mitochondrial appeared (upper panel) and the basement membrane became enlarged and rarefied. The endothelial junction was intact and closed (white arrow). D. Two representative images from the second PSH patients. Both endothelial cells became severe edematous with the random-oriented and dispersed filaments. In the lower panel, the lumen turned to be dramatically narrowed and one blood cell was stuck inside the capillary vessel. There were several pore formations and discontinuous points of the luminal wall (arrow head), implicating the damage of endothelial vessels. Left hand side: Lower power image; right hand side: higher power image, corresponding to that in the lower power image indicated in white square frame. L: lumen Ultrastructure alterations revealed by the high-resolution transmission electron microscopy on capillary endothelial cells of the great omentum during PD at the timepoint when the abdomen was opened (A) and before closed (B, C and D) in non-PSH (B) and PSH patients including two cases (C for case 1 and D for case 2). A. When the abdomen was opened, the endothelia cells showed the normal morphology with flattened nucleus, mitochondrial, transport vesicles, interacted cytoskeletal filaments and intact endothelial junctions (white arrow). B. In non-PSH patients before the abdomen was closed, the endothelia cells showed the normal morphology but with slightly more pinocytotic and coated vesicles (black arrow). C. Two representative images from the first PSH patients. Endothelial nucleus became round (upper panel) or prominent towards lumen (lower panel). There were numerous pinocytotic and coated vesicles (black arrow) and formation of large vacuoles (V). A bit more mitochondrial appeared (upper panel) and the basement membrane became enlarged and rarefied. The endothelial junction was intact and closed (white arrow). D. Two representative images from the second PSH patients. Both endothelial cells became severe edematous with the random-oriented and dispersed filaments. In the lower panel, the lumen turned to be dramatically narrowed and one blood cell was stuck inside the capillary vessel. There were several pore formations and discontinuous points of the luminal wall (arrow head), implicating the damage of endothelial vessels. Left hand side: Lower power image; right hand side: higher power image, corresponding to that in the lower power image indicated in white square frame. L: lumen Furthermore, the intra-endothelia filaments were intact and closely connected to each other inside normal capillary endothelia cells (Fig. 2A and B), while the random-oriented and dispersed filaments were seen in the edematous endothelium of PSH patients (Fig. 2D), suggesting that the endothelia tone and contraction activity might be significantly reduced following the major surgery. However, the endothelial junctions were sealed well by intact tight junctions even in the cases of swelling endothelia cells with enlarged basement membrane and enhanced transendothelial transports (Fig. 2C). Therefore, these observations clearly indicate the capillary endothelia and pericellular edema occurred in PSH patients. the constant leakage of fluid and solutes into endocellular and extracellular space implicates the insufficient intracellular circulation volume and diminished vessel tone, leading to the persist severe hypotension during and after PD. To evaluate diverse inflammatory responses relevant with PSH, we analyzed various pre-operative and one-day postoperative inflammatory biomarkers (Table 3). There was no significant difference among pre-operative biomarkers, but postoperative biomarkers in PSH patients showed significant decreased white blood cell count and obvious increase in C-reactive protein and procalcitonin (P < 0.05). Table 3Perioperative inflammatory index analysesTime-pointsNon-PSH group(n = 59)PSH group(n = 44)P-valueWBC count (109 /L)Pre-operative5.59 (1.71)5.58 (1.47)0.983Postoperative day14.95 (5.49)12.59 (3.47)0.014Neutrophil ratio(%)Pre-operative60.83 (10.65)59.53 (7.90)0.497Postoperative day86.82 (4.44)85.83 (5.08)0.296Lymphocyte ratio (%)Pre-operative28.26 (9.09)29.70 (7.53)0.395Postoperative day6.30 (3.07)7.51 (3.78)0.077CRP (g/L)aPre-operative2.85 [1.24, 5.00]2.14 [1.39, 4.73]0.796Postoperative day49.40 [35.93, 57.18]61.30 [41.58, 75.33]0.007PCT (ng/mL)bPre-operative0.02 [0.02, 0.04]0.03 [0.02, 0.08]0.801Postoperative day0.91 [0.36, 1.45]1.15 [0.45, 2.98]0.043Abbreviations: WBC White blood cell, CRP C-reactive protein, PCT ProcalcitoninaAssessed for 40 patients in PSH group and 56 in non-PSH group. bRegistered in 41 patients in PSH group and 52 in non-PSH group Table 3Perioperative inflammatory index analysesTime-pointsNon-PSH group(n = 59)PSH group(n = 44)P-valueWBC count (109 /L)Pre-operative5.59 (1.71)5.58 (1.47)0.983Postoperative day14.95 (5.49)12.59 (3.47)0.014Neutrophil ratio(%)Pre-operative60.83 (10.65)59.53 (7.90)0.497Postoperative day86.82 (4.44)85.83 (5.08)0.296Lymphocyte ratio (%)Pre-operative28.26 (9.09)29.70 (7.53)0.395Postoperative day6.30 (3.07)7.51 (3.78)0.077CRP (g/L)aPre-operative2.85 [1.24, 5.00]2.14 [1.39, 4.73]0.796Postoperative day49.40 [35.93, 57.18]61.30 [41.58, 75.33]0.007PCT (ng/mL)bPre-operative0.02 [0.02, 0.04]0.03 [0.02, 0.08]0.801Postoperative day0.91 [0.36, 1.45]1.15 [0.45, 2.98]0.043Abbreviations: WBC White blood cell, CRP C-reactive protein, PCT ProcalcitoninaAssessed for 40 patients in PSH group and 56 in non-PSH group. bRegistered in 41 patients in PSH group and 52 in non-PSH group Perioperative inflammatory index analyses Neutrophil ratio (%) Abbreviations: WBC White blood cell, CRP C-reactive protein, PCT Procalcitonin aAssessed for 40 patients in PSH group and 56 in non-PSH group. bRegistered in 41 patients in PSH group and 52 in non-PSH group We further analyzed postoperative complications between two groups. As show in Table 4, the composite adverse events occurred in 21 of the 59 patients (35.6%) in the non-PSH group and in 26 of the 44 patients (59.1%) in the PSH group (P = 0.018). There was no significant difference in the occurrence of postoperative pancreatic fistula and acute kidney injury. However, the incidence of postoperative pulmonary complications (P = 0.023), intra-abdominal infections (P = 0.025) and delayed gastric emptying (P = 0.001) were much higher in PSH patients. Table 4Postoperative complications analysesComposite adverse events (n; %)Non-PSH group(n = 59)(n = 59)PSH group(n = 44)P-valueTotal21 (35.6%)26 (59.1%)0.018Postoperative pancreatic fistula3 (5.1%)5 (11.4%)0.420Postoperative pulmonary complications12 (20.3%)18 (40.9%)0.023Intra-abdominal infections7 (11.9%)13 (29.5%)0.025Acute kidney injury02 (4.5%)0.180Delayed gastric emptying13 (22.0%)33 (75.0%)0.001grade A6 (10.2%)16 (36.4%)0.001grade B3 (5.1%)9 (20.4%)0.036grade C4 (6.8%)8 (18.2%)0.141The components of the composite adverse events (within the first 30 postoperative days after the PD surgery) Table 4Postoperative complications analysesComposite adverse events (n; %)Non-PSH group(n = 59)(n = 59)PSH group(n = 44)P-valueTotal21 (35.6%)26 (59.1%)0.018Postoperative pancreatic fistula3 (5.1%)5 (11.4%)0.420Postoperative pulmonary complications12 (20.3%)18 (40.9%)0.023Intra-abdominal infections7 (11.9%)13 (29.5%)0.025Acute kidney injury02 (4.5%)0.180Delayed gastric emptying13 (22.0%)33 (75.0%)0.001grade A6 (10.2%)16 (36.4%)0.001grade B3 (5.1%)9 (20.4%)0.036grade C4 (6.8%)8 (18.2%)0.141The components of the composite adverse events (within the first 30 postoperative days after the PD surgery) Postoperative complications analyses The components of the composite adverse events (within the first 30 postoperative days after the PD surgery) Discussion In this prospective controlled cohort study, we found that approximately 43% patients experienced persistent hypotension with the requirement of continuous vasopressor infusion to keep MAP over 65mmHg during PD, similar to the observations by Taylor Pitter and colleagues [1]. Following the standardized GDT with PPV and SVV assessment, more total infusion volume including colloid fluid and blood transfusion is needed in PSH patients. Such enhanced fluid infusion requirement, along with the occurrence of decreased systemic vascular resistance during PD, implicate disrupted vascular activity of microcirculation and capillary endothelial dysfunction, leading to insufficient circulating volume and loss of proper function of vascular contraction. Importantly, the fluid loading may itself serve as a mechanical and chemical trigger that further exacerbates endothelial swelling and barrier dysfunction [21]. This proposition is supported by electron microscopy analysis to reveal the ultrastructural alterations of capillary endothelial cells, showing swelling and edematous morphology with capillary leakage of fluids and solutes into interstitial space. Intraoperative higher blood glucose and much lower arterial plasma Ca2+ concentration occurred in PSH patients as compared with non-PSH patients, suggesting the involvement of hyperglycemia and aberrant intracellular /extracellular Ca2+ balance in the alteration of endothelial permeability. Consistently, higher incidence of pleural effusion as major sign of postoperative pulmonary complications implicates much severer intro-operative damage on the integrity and function of capillary endothelial cells in PSH patients.Procalcitonin with the edema-inducing effects by disrupting endothelial cell function has been assumed to release from monocytes when adherent to the activated endothelium [22]. Endothelial permeability to plasma proteins and liquid is increased in inflammation manifested by protein-rich edema. The inflammatory response due to the surgical trauma generally upregulate the expression of cytokines, ORS and C-reactive protein that result in the disruption of capillary vessel barrier and endothelial leakage [5, 7]. Notably, despite markedly elevated CRP and PCT levels, PSH patients paradoxically exhibited lower peripheral WBC counts compared with non-PSH patients. This seemingly contradictory pattern likely reflects concurrent pathophysiological mechanisms operating simultaneously during the perioperative period. First, leukocyte consumption driven by the intense systemic inflammatory response associated with the extensive surgical trauma of PD. The significantly elevated CRP and PCT observed in PSH patients reflect a massive inflammatory burden that implies sustained activation and peripheral depletion of circulating leukocytes through accelerated turnover and tissue recruitment [23]. Second, leukocyte margination, whereby activated leukocytes adhere to the injured endothelium and are sequestered from the circulating pool. This mechanism is directly corroborated by our TEM findings demonstrating endothelial swelling and ultrastructural barrier disruption in PSH patients; consistent with this, endothelial activation and upregulation of adhesion molecules, might promote leukocyte adhesion, rolling, and transmigration, effectively reducing the peripheral WBC count despite an ongoing and intensifying inflammatory state [24]. Beyond leukocyte dynamics, surgical stress normally induces the blood glucose elevated, especially causing dramatic rise in the most extensive abdominal surgical procedures with a massive inflammatory response as seen in PD. It is well-described that acute and chronic hyperglycemia is associate with the disruption on endothelial permeability and vascular function [25, 26]. Hyperglycemia drives endothelial dysfunction through a multifaceted cascade of oxidative stress and inflammation. Intracellular glucose overload saturates metabolic pathways [27], triggering massive mitochondrial superoxide overproduction. In the present study, the blood glucose increased higher in PSH patients, raising the possibility of increased blood sugar involved in the endothelia swelling and protein-rich edema. Furthermore, abnormal Ca2+ metabolism or distribution including intracellular Ca2+ overload or insufficient elevation may cause endothelial leakage and therefore severely compromise cardiovascular homeostasis [28]. Oxidative stress, hyperglycemia and inflammatory mediators are known to cause an increase in intracellular Ca2+ concentration by increasing Ca2+ release from intracellular stores and Ca2+ influx through store-dependent and store-independent channels [29, 30]. We observed much lower plasma Ca2+ in PSH patients, this finding may reflect an alteration in Ca²⁺ homeostasis, potentially involving redistribution or enhanced influx into endothelial cells; however, as plasma Ca²⁺ levels do not directly reflect intracellular Ca²⁺ concentrations, intracellular measurements would be required to confirm this hypothesis. Taken together, the interplay of surgical stress-induced inflammation, hyperglycemia, oxidative stress, fluid overload, and aberrant Ca²⁺ redistribution collectively contribute to capillary endothelial dysfunction and then could hamper the cardiovascular system function during and after PD. Water traverses the endothelial barrier by two transport modes of paracellular and transcellular routes either via passive diffusion or through water-transporting membrane channels, the aquaporins [31, 32]. The swelling phenotype of capillary endothelium, a nonspecific response of the endothelial cells to a great variety of insults, also appeared in PD, suggesting the enhanced trans-endothelial water transport via passive diffusion and aquaporin-mediated transcellular pathway following the major surgery. In addition, more caveolae (the vesicular carriers of the transcellular pathway), coated vesiculas and vacuoles formation, which is for plasma protein transcellular transport [33], are observed inside capillary endothelial cells of PSH patients. It suggests that PSH patients experienced activated albumin transport into interstitials space to form paracellular edema. In the present study with GDT, more crystal fluid and colloid fluid (albumin and plasma) infusion along with vasopressors in PSH patients are required to maintain proper macrocirculation hemostasis. However, the cellular and interstitial edema might be exacerbated by excessive leakage of water and albumin supplement and then potentially further impede oxygen and nutrition supply to the tissues. Higher plasma lactic acid seen in PSH patients during PD might provide clues that systemic aerobic metabolism was disrupted due to cellular and interstitial edema with insufficient oxygen supply. On the other hand, our fluid therapy approaches seem to be beneficial for the microcirculation since there was no difference in the incidence of postoperative acute kidney dysfunction between PSH and non-PSH patients. Nevertheless, it cannot be excluded that the potential abnormal capillary circulation may lead to the embolism formation and make local tissue at hypoxia state since the circulating red cell stuck inside the capillary lumen was observed in PSH patient as seen in Fig. 2D. An increase in cytosolic Ca2+ concentration following the stimulation with inflammatory mediators has been established as the initial pivotal signal that precedes endothelial cell shape change and the opening of inter-endothelial junctions that results in barrier dysfunction [34, 35]. Enhanced intracellular Ca2+ may result in increased paracellular permeability via activation of actin-myosin machinery and increased transcellular permeability via increased vesicle fusion leading to fenestrations and transcellular channels [36]. In the present study, the disorganized cytoskeletal filaments and swelling morphology suggest that actin-myosin machinery of capillary endothelium is disrupted, which might make endothelial cells incapable to contract properly. A closed and intact inter-endothelial junction is also visible, further supporting transport via the transcellular route and a sign of abnormal endothelial cell contraction. Furthermore, communication between endothelia cells and vascular smooth muscle cells has been well described [37]. The present identification of capillary endothelial dysfunction could reflect the loss of proper function of endothelial cells in the small-resistance arteries and arterioles. Endothelial edema and dysfunction could impede vascular smooth muscle constriction, therefor leading to the reduced vascular tone and decreased peripheral vascular resistance in PSH patients during PD. We found that treatment with calcium supplement (1–2 gram CaCl2) to increase arterial Ca2+ over 1.05 mmol/L could improve peripheral vascular resistance and hemodynamic stability (data not shown). This finding suggests that sufficient circulating calcium is crucial for maintaining functional vascular activity and normal peripheral vascular resistance, possibly by promoting vascular smooth muscle contraction and balancing the influx and efflux of Ca2+ in vascular endothelial cells. While this study provides novel insights into the pathophysiological basis of PSH during pancreaticoduodenectomy with clinical evidence of the ultrastructural alterations of capillary endothelia cells and decreased plasma Ca2+ there are still several limitations. First, as a single-center analysis with a relatively small sample size (n = 103), our findings may be influenced by specific institutional practices. Importantly, patients with PSH often underwent more complex procedures with longer operative times and greater blood loss. Although our prior studies have identified these as risk factors for PSH [9], the limited number of positive complication events in the current series precluded a robust multivariable logistic regression. Furthermore, the present study may have been underpowered to detect statistically significant between-group differences in certain contractility indices, as evidenced by the borderline result observed for dPmx (P = 0.051), which approached but did not reach conventional statistical significance. A Type II error cannot therefore be excluded, and this trend warrants confirmation in a larger, adequately powered cohort. Future large-scale, multi-center external validation is necessitated to adjust for these confounders and confirm our findings. Second, several limitations of the TEM analysis merit acknowledgment. Specimens were obtained from only three patients per group, yielding 9 and 8 capillary profiles respectively, limiting generalizability. Imaging was performed under single-blind rather than double-blind conditions, introducing potential observer bias. Rigorous morphometric quantification was not feasible due to variable sectioning angles and irregular capillary geometry; findings are therefore reported descriptively. Nonetheless, all presented images represent predominant ultrastructural features consistently observed across profiles, rather than isolated or incidental findings. Furthermore, due to the observational nature of the study design, we can only establish an association rather than a definitive causal relationship between endothelial edema, hypocalcemia, and PSH. It remains to be fully elucidated whether endothelial dysfunction is the primary driver of hemodynamic instability or a secondary consequence of surgical stress and fluid resuscitation. Future prospective studies need to focus on endothelial function analysis and the optimization of hemodynamic management protocols. Conclusions In conclusion, endothelial cell permeability alteration and endothelial dysfunction might potentially reflect surgical stress-inflammation-endocrine responses during major surgery. Our study indicates that extensive surgical trauma during PD exerts the significant impact on the capillary endothelial function, which is associated with severe hypotension and postoperative adverse outcomes. The crucial contribution of endothelial edema formation, decreased plasma Ca2+ and the associated hypotension with compromised perfusion to the adverse outcome of patients render strategies to preserve endothelial function as a key point to improve patient care; such strategies should include monitoring and regulating plasma Ca2+ as well as other methods that improve the functionality and integrity of the vascular barrier, i.e. immune modulatory interventions. Supplementary Information Supplementary Material 1. Supplementary Material 1. Supplementary Material 2. Supplementary Material 2. Supplementary Material 3. Supplementary Material 3.