Introduction

Despite new available effective methods of treatment, acute pancreatitis (AP) is a socially significant problem for public health and one of the most common causes of hospitalization of patients with gastrointestinal diseases [1 — 3]. Moreover, acute pancreatitis is characterized by high mortality (62-65% for infected pancreatic necrosis) [3 — 5].

In 40% of cases, AP is complicated by abdominal fluid collections, which resolve spontaneously in most patients [6]. However, patients with moderate-to-severe AP are more prone to development of fluid collections > 6 cm [7].

Ultrasound is currently the gold standard for diagnosis of fluid collections, especially in early stages of disease [8]. The role of timely ultrasonic verification of acute pancreatitis cannot be overestimated, since favorable prognosis directly depends on correct diagnosis and timely initiation of effective therapy.

Along with development of intraluminal endoscopy, endoscopic ultrasonography (EUS) became the most important element in assessing the volume of pancreatic lesions and subsequent drainage. This method allows accurate analysis of pancreatic parenchyma, ductal system and differential diagnosis between different fluid collections. These measures affect further treatment strategy [5].

Endoscopic transluminal drainage of fluid collections in patients with acute pancreatitis have been actively introduced into clinical practice over the past 5-7 years [4, 5, 9]. Obvious advantage of this approach is minimal trauma, lower mortality (8 — 15%) and shorter postoperative hospital-stay (13 — 19 days).

Despite multiple reports devoted to diagnosis and treatment of fluid collections, there is currently no generally accepted strategy for transluminal treatment depending on the type of fluid collection.

Objective of the study was to improve treatment outcomes in patients with acute pancreatitis using a differentiated approach to transluminal drainage.

Material and methods

There were 1074 patients with acute pancreatitis between January 2018 and December 2021 at the Sklifosovsky Research Institute for Emergency Care.

All patients underwent transabdominal ultrasound. CT of the abdomen and retroperitoneal space was performed in patients with fluid collections to clarify their characteristics. Of these, EUS of pancreatobiliary zone was carried out in 136 patients. EUS was used as a final diagnostic method to determine localization, dimensions of fluid collections, shape and contours, to assess the features of content, presence or absence of a capsule and connection with pancreatic ductal system, possibility of EUS-assisted intraluminal drainage.

We used the classification of fluid collections mentioned in the Atlanta classification of acute pancreatitis (2012). This classification is a result of multiple-center work of international group of experts. Moreover, this grading system has received worldwide approval and practical application [10].

The study group included 63 patients who underwent transluminal drainage of fluid collections. There were 46 (73%) men and 17 (27%) women. Mean age of patients was 57 [48; 69] (35; 86) years.

The APACHE II scoring system was used to assess severity of disease. Mean score was 24 [15; 46] (11; 43). Mild acute pancreatitis was found in 18 patients, moderate pancreatitis — in 23, severe pancreatitis — in 22 cases. The period between onset of disease and admission to the hospital was 34 [25; 46] (12; 63) days.

The indications for transluminal drainage were the distance between the hollow organ wall and fluid collection < 1 cm, no great vessels on puncture trajectory (> 3 mm) and fluid collection > 5 cm.

For endoscopic drainage, we used the Exera III ultrasonic endoscopic system and the Olympus GFUCT160 convex ultrasound endoscope, the GE Healthcare OEC Elite X-ray equipment, the COOK cystotome, the Olympus 0.35 guidewire, special self-expanding stents for pancreatic duct drainage and double pigtail plastic stents.

Transluminal drainage was performed under general anesthesia in patient supine position. We determined the safest site for intraluminal drainage using EUS. We drained the cyst through the intended trajectory using a cystotome and aspirated content for subsequent microbiological examination and analysis of sensitivity to antibacterial drugs.

After that, we injected contrast agent to the cyst to determine its true size, homogeneity of content, tightness and connection with pancreatic ductal system. In case of connection with pancreatic ducts, we performed additional stenting of the main pancreatic duct using plastic stent 5 cm 5 Fr.

After that, we inserted a stent in the canal between the cavity of the stomach or duodenum and the cyst.

In case of homogeneous hypo- and anechoic cavity with clear even contours and aspiration of serous or serous-hemorrhagic fluid, we installed a plastic stent with rounded ends. The indication for installation of a fully covered self-expanding stent was inhomogeneous anechoic fluid collection with hyperechoic inclusions (sequesters) and purulent content.

Patients with necrotic forms required cystonasal drainage (7 Fr) under X-ray control for debridement of the cyst with a 0.05% aqueous solution of chlorhexidine. Lavage has been carried out for 2 days after drainage of purulent cavity. Endoscopic sequestrectomy was performed every 24—48 hours depending on diagnostic data and severity of intoxication. To do this, a video gastroscope with a diameter of no more than 8.8 mm was inserted through the stent. We examined the cavity, assessed regeneration processes and performed sequestrectomy through this approach. Necrotic tissues were removed using endoscopic forceps, polypectomy loop and soft distal cap for better visualization. Sanitation of the cyst was performed using a 0.05% aqueous solution of chlorhexidine.

Effectiveness of endoscopic treatment was analyzed considering clinical (fever, pain syndrome) and laboratory data (leukocytosis, CRP increase), cyst volume changes according to ultrasound and CT of the abdomen, amount and nature of discharge through the drainage tube.

CT data on persistent cyst with ineffective drainage were an indication for additional percutaneous drainage.

Stent was removed in 6 (for encapsulated peripancreatic fluid collections) or 1 month (for other types of fluid collections) after discharge from the hospital.

Major EUS criteria that guided us in characterizing the fluid collections are presented in Table 1.

Table 1. EUS criteria for various types of fluid collections

Statistical analysis was carried out using the Statistica for Windows v. 10.0 software package (StatSoftInc., USA). Distribution normality was tested using the Shapiro-Wilk test.

Results

According to EUS data in 136 patients with fluid collections, endoscopic transluminal drainage was performed in 63 (46%) patients. Of these, fluid collections accompanying interstitial edematous pancreatitis were detected in 39 (62%) patients, necrotizing pancreatitis — in 24 (38%) patients (Table 2).

Table 2. Results of EUS and TLD in patients with acute pancreatitis, n=136

Acute necrotic accumulations prevailed among all fluid collections (74.3% of necrotic forms of acute pancreatitis). Encapsulated peripancreatic fluid collections prevailed among interstitial forms (63.6%).

In 39 (61.9%) patients with interstitial forms of edematous pancreatitis, we performed TLD using plastic stent with rounded ends (“double pig tail”). Of these, 2 (3.2%) patients with acute peripancreatic fluid collections demonstrated homogeneous accumulations of contrast agent with fuzzy uneven contours and multiple spurs (Fig. 1).

Fig. 1. Peripancreatic fluid collections.

a — X-ray image; b — EUS scan.

Contrast enhancement of encapsulated peripancreatic fluid collections in all 37 (58.7%) patients revealed homogeneous delimited round or oval accumulations with clear even contours (Fig. 2).

Fig. 2. Encapsulated peripancreatic fluid collections.

a — X-ray image; b — EUS scan.

Among patients with necrotic forms of AP, 24 (38.1%) ones underwent installation of a fully covered self-expanding stent. Of these, 18 (28.6%) patients with acute necrotic accumulations had inhomogeneous irregularly shaped cavities with fuzzy uneven contours and multiple spurs (Fig. 3).

Fig. 3. Acute necrotic fluid collection.

a — X-ray image; b — EUS scan.

Six (9.5%) patients with delimited necrosis had inhomogeneous round or oval accumulations of contrast agent with clear even contours (Fig. 4).

Fig. 4. Delimited necrosis.

a — X-ray image; b — EUS scan.

Among 63 patients undergoing TLD, connection with pancreatic ductal system was found in 5 (7.9%) patients: 3 (4.8%) ones with acute necrotic collections and 2 (3.1%) patients with encapsulated peripancreatic accumulations. They underwent pancreaticoduodenal stenting (Fig. 5).

Fig. 5. Communication of fluid collection with pancreatic duct.

a — contrast agent leakage beyond the contours of the duct; b — stent in the main pancreatic duct.

Elective sequestrectomy after stenting of necrotic fluid collections with fully covered self-expanding stents was performed in 22 patients. Mean number of procedures was 7 [6; 9] (2; 24).

Two patients did not undergo debridement of purulent cavity due to a lethal outcome. Additional percutaneous drainage was required in 11 (45.8%) out of 24 patients. Mean hospital-stay was 9.5 [6; 22] (3; 30) days.

Complicated postoperative period was observed in 4 (6.3%) patients with acute necrotic fluid collections. There were bleedings from the area of pancreatogenic destruction (Clavien-Dindo grade IIIb). In this regard, 2 (3.1%) patients required angiography and endovascular embolization of a. gastroduodenalis. In 1 (1.6%) case, endoscopic hemostasis was achieved by filling the cavity with Hemoblock hemostatic solution 4 ml. Combination of both methods was necessary in 1 (1.6%) patient. There was no evidence of recurrent bleeding in these patients.

Four (6.3%) patients died. Autopsy revealed regression of purulent-inflammatory process in all cases. Three (4.7%) patients died from multiple organ failure, 1 (1.6%) patient died from severe nosocomial pneumonia developed in 32 days after drainage.

Discussion

We found similar incidence of fluid collections corresponding to interstitial edematous and necrotizing pancreatitis (48.6% vs. 51.4%). Among them, acute necrotic accumulations (38.2%) and encapsulated peripancreatic collections (30.9%) prevail. These findings are consistent with the world literature data [11, 12].

Development of intraluminal endoscopy made it possible to perform endoscopic drainage under ultrasound control that reduced the incidence of postoperative complications (6.3%) and mortality (6.3%). There are literature data on safe drainage of fluid collections > 10 cm. According to our data, transluminal drainage is advisable for cysts > 6 cm.

Analyzing the world literature and own data, we propose the following algorithm for endoscopic intraluminal drainage of fluid collections in patients with acute pancreatitis (Fig. 6).

Fig. 6. Algorithm for endoscopic intraluminal drainage of fluid collections in patients with acute pancreatitis.

Such a differentiated approach to drainage of fluid collections improves the treatment outcomes in patients with acute pancreatitis.

According to our data, intraluminal surgery is advisable not only for homogeneous delimited fluid collections (encapsulated peripancreatic accumulations), but also for advanced lesions including infected ones (acute necrotic accumulation, delimited necrosis). Installation of a self-expanding stent makes it possible to perform full-fledged endoscopic sanitation and sequestrectomy of pancreatogenic destruction cavities. This approach allows us to consider endoscopic intraluminal drainage as a final minimally invasive method of surgical treatment of pancreatic necrosis with effectiveness of up to 45.8%.

If fluid collections are not adjacent to hollow organ lumen, their percutaneous ultrasound-assisted drainage is required.

Conclusions

1. Endoscopic ultrasonography is not only the most important method for diagnosis of fluid collections in patients with acute pancreatitis. This method also ensures determining further strategy of intraluminal drainage.

2. EUS-guided transluminal drainage of encapsulated peripancreatic fluid collections is highly effective method of treatment.

3. Endoscopic intraluminal drainage of fluid collections is the final minimally invasive method of surgical treatment of large-focal pancreatic necrosis in 45.8% of cases.

The authors declare no conflicts of interest.