We analyzed patients who underwent EUS-guided transmural drainage in our center from June 2019 to May 2023. All patients with symptomatic PFC were collected and analyzed in the study. The inclusion criteria were: 1. Symptomatic PFC mainly includes abdominal distension, abdominal pain, nausea, jaundice, etc.; 2. PFC that was resistant to conservative treatment; 3. All patients had a disease course of more than 4 weeks; 4. There is no strict age limit. Patients under the age of 18 must obtain informed consent from their guardians. The exclusion criteria were: 1. Enhanced CT observed a patient with a high density in the cyst cavity suspected of hemorrhage; 3. Suspicion of malignancy. All patients with platelet abnormalities, coagulation abnormalities, and conditions with cirrhosis or esophageal varices that could affect postoperative bleeding were excluded from the study. A total of 181 patients were finally enrolled in the study.

This was a retrospective study in which we obtained informed consent and signature from patients before collecting their data. The study was approved by the Ethics Committee of our center (2021-R158). I confirm that all methods are carried out in accordance with the relevant guidelines and regulations

Potential risk factors that may affect postoperative bleeding were collected as follows. The factors related to patients included gender, age, etiology, cyst diameter, cyst type, and cyst wall thickness. The factors associated with EUS drainage mainly included postoperative infection and stent type, location of puncture, Number of puncture channels. Firstly, univariate regression analysis of risk factors was performed, and related factors with P < 0.05 were included in multivariate regression analysis, and independent risk factors affecting bleeding were identified.

All procedures were performed by experienced endoscopists under intravenous anesthesia. Endoscopic ultrasound (ME2, OLYMPUS, JAPAN) was used to perform standard scanning of the pancreas (shown in Figure 1), and bleeding and solid nodules will be excluded if detected. The cyst was then punctured with puncture needle (ECHO-19, COOK, United States) at a suitable location (stomach or duodenum) avoiding blood vessels, A 0.035-inch guidewire (Jagwire, Boston Scientific, United States) was inserted through the needle lumen into the cyst. A cystotome (10 Fr, COOK, United States) was used to dilate the puncture tract. Plastic or self-expanding metal stent (SEMS Boston Scientific, United States) and nasal cyst tubes (for postoperative irrigation) were then placed depending on the size of the cyst and inside necrosis (shown in Figures 2, 3). All patients with metal stents were fitted with fully coated self-expanding metal stents, and all patients with plastic stents were fitted with two 7Fr diameter double pigtail plastic stents. Nasal cyst tubes were routinely inserted in all patients to facilitate postoperative observation of complications such as bleeding. To prevent infection, routine prophylactic treatment with antibiotics (ceftriaxone, 2g, once daily, intravenously) was given 3 days before surgery. Abdominal CT was reexamined at 7 and 30 days after surgery, and then once a month. The stent was removed when the patient’s symptoms had disappeared for at least 2 weeks and CT confirmed the cyst had disappeared.

Technical success was defined as access to the PFC by placement of a stent and drainage. Clinical success is defined as the disappearance of symptoms, a reduction in PPC size (more than 50%), or complete regression (2). If the fever exceeds 38 degrees Celsius for more than 24 h after EUS drainage, it is defined as an infection. After the operation, rinse with the nasal cyst tube. A decrease of 2 g/L in hemoglobin after surgery, or the occurrence of hematemesis and melena, is defined as bleeding (5). The classification of adverse events is divided into six levels. A:No treatment is needed; B: have minor therapy or consequence, including overnight admission; C: that requires major therapy or hospitalization (24–48 h); D:major therapy, need unplanned increase in level of care, or hospitalization > 48 h; E:result in permanent adverse sequelae; F: death (5). Re-intervention was defined as the need for repeat surgery or endoscopy owing to persistent symptoms in association with a residual pseudocyst that was not less than 50% of the original size on follow-up imaging (6). All patients were divided into two groups according to the QNI system: the group with lower QNI stratification (≤ 2 quadrants and necrosis rate ≤ 30%) and the group with higher stratification (≥ 3 quadrants, necrosis rate ≥ 30% in 2 quadrants or necrosis and infection rate > 60% in 1 quadrant) (7). After the occurrence of bleeding symptoms, red blood cells were transfused in time, hemostatic drugs were applied, and endoscopic hemostasis or vascular interventional embolization was performed if necessary. If infection or bleeding is difficult to control, embolismor surgical intervention may be considered. The type of cyst was defined by the amount of solid necrotic material in the cyst. The procedure was performed by experienced endoscopists and imaging specialists. The size of the cyst and necrotic material was measured by ultrasound images and nuclear magnetic images. According to Atlanta classification (8), if the solid debris in the cyst was greater than 50%, it was identified as walled off necrosis (WON), and if it was less than 50%, it was identified as pancreatic pseudocyst (PPC).

The sample size of our study is estimated based on the number of independent variables (greater than 15 times the number of independent variables). Data were analyzed using SPSS 27.0 (SPSS Inc., Chicago, Illinois, United States). Results are expressed as medians and means and ranges. A P-value below 0.05 was considered statistically significant. Univariate analysis for screening purposes for risk factors. In addition, multivariate analysis calculating the hazard ratio using logistic binary regression was added on the factors identified by univariate analysis as independently significant for bleeding after EUS—guided transmural drainage.

A total of 181 patients were enrolled in the study, including 111 male and 70 female with a mean age of 49 years (range 8–85 years). Their baseline characteristics were shown in Table 1. The causes of PFC were acute pancreatitis in 150 cases (82.9%), chronic pancreatitis in 17 cases (26%), surgery in 6 cases (3%) and trauma in 8 cases (4%). The median diameter of the cyst was 10.9 cm (range 4.1–25.3 cm). Among these patients, 21 (11.6%) developed symptoms of infection (fever or elevated WBC count). In our process, 179 cases were punctured through the stomach and 2 cases were punctured through the duodenum. The average thickness of the cyst wall was 4.6 mm (1—10 mm). 161 patients were identified as PPC and 20 patients were identified as WON. Five patients were treated with dual-channel puncture (puncture at two locations with stent or nasal cyst tube inserted) and 176 patients were treated with single-channel puncture (puncture at only one location with stent or nasal cyst tube inserted).

All patients were successfully drained by EUS and the stent (plastic or metal) was successfully removed after the cyst was identified and symptoms resolved postoperatively. The median follow-up period was 11.5 months, and 9 (4.9%) patients recurred, 7 patients were managed by re-endoscopic intervention, and 2 patient was managed by surgery. In our cohort, 22.6% of patients developed symptoms of infection. Among them, 4 patients underwent repeated endoscopic removal of necrotic tissue, the remaining 38 patients with symptoms of infection disappeared after antibiotics and nasal cyst tube irrigation. Among the classification of adverse infection events, 2 patients were classified as grade C and 40 patients as grade D. Among these patients, 14 (7.7%) had bleeding after EUS drainage. Seven patients were treated with conservative therapy for hemostasis, two patients were treated with endoscopic self-expandable partially covered metal stent for hemostasis, three patients were treated with endoscopic hemostasis clamp, and two patients was treated with vascular embolism for hemostasis. All bleeding patients were infused with suspended red blood cells and rehydration to improve clinical symptoms. All patients with bleeding were classified as grade D according to the classification of adverse events. No serious adverse events such as piercings and death occurred.

The treatment methods of endoscope hemostasis include metal stent placement, hemostatic clamp clamping, balloon packing and compression. In our patients with hemorrhage, two cases were caused by vascular variation in the cyst, and the hemostasis was successfully stopped by vascular interventional embolization (Figure 4). There were 5 patients with bleeding at the puncture site and through the puncture channel, and the bleeding was successfully stopped by the implantation of the self-expandable partially covered metal stent and application of the hemostatic clamp (Figure 5). In seven patients, the cause of bleeding was unclear, but hemostasis was eventually achieved through conservative treatment. All bleeding occurred 24 h after the operation, the minimum time was 24 h, and no bleeding occurred during the operation. The longest bleeding time was 7 days after surgery. Tailed characteristics of all bleeding patients are shown in Table 2.

Univariate and multivariate analyses were performed for possible factors related to bleeding after EUS-guided transmural drainage, and Univariate analysis (shown in Table 3) showed that cyst type and cyst size were risk factors for bleeding after EUS guided transmural drainage (P = 0.04 and P = 0.004). Multivariate analysis (shown in Table 4) showed that the cyst size was statistically significant (P = 0.006, OR = 2.722). Cyst size influence bleeding after EUS transmural drainage, and cyst size was an independent risk factor. The larger the cyst, the higher the risk of bleeding after endoscopic drainage.

According to the QNI system, 18 patients were assigned to the high QNI group and 163 patients were assigned to the low QNI group. There was no significant difference in the incidence of adverse events such as bleeding and infection between the two groups of patients (P > 0.05) (Table 5).