Introduction

According to the world literature [1—7], in the structure of acute surgical diseases of the abdominal organs, acute pancreatitis (AP) occupies the 3^rd place, ranging from 4.5 to 10%, and does not tend to decrease. Of these, necrotizing forms occur in about 15-30% of patients. Most often they are complicated by parapancreatic infiltrate or abscess, peritonitis or phlegmon of retroperitoneal tissue, which requires adequate drainage with subsequent sanation interventions.

The traditional approach to the treatment of infected necrotizing pancreatitis is necrosequestrectomy from the laparotomic approach in order to completely remove the infected necrotic tissue [2, 3]. However, recent studies have shown that a large amount of surgery is accompanied by high postoperative mortality and increases the risk of postoperative complications [1—3]. Minimization of surgical trauma is one of the key areas of modern surgery for pancreatic necrosis. Among surgical interventions, minimally invasive techniques are currently actively used [8—10]. Due to the rapid development of flexible endoscopy, the arsenal of minimally invasive surgical technologies for pancreatic necrosis has expanded significantly due to laparoscopic access and ultrasound-guided endoscopic interventions, which create the basis for minimizing surgical aggression through transgastric access [11]. However, a small number of publications require a clearer definition of indications for endoscopic drainage and a comprehensive assessment of its effectiveness [4].

The aim of the study was to evaluate the effectiveness of intraluminal drainage of acute fluid accumulations in necrotizing pancreatitis.

Material and methods

For the period from January 2018 to December 2020 at the Sklifosovsky Research Institute for Emergency Medicine 848 patients diagnosed with AP were treated. Interstitial edematous form, according to the Atlanta classification of AP, was observed in 616 (72.6%) patients, necrotizing pancreatitis — in 232 (27.4%) [11]. Among necrotic forms, pancreatic parenchymal necrosis was detected in 56 (24.1%) patients, its combination with peripancreatic necrosis — in 176 (75.9%).

Endoscopic transluminal drainage was performed in 22 (12 (55%) men and 10 (45%) women) patients with necrotizing pancreatitis, in whom fluid accumulations were adjacent to the stomach or duodenum. The average age of patients was 48.5 [39; 56] (35; 88) years.

All patients underwent transabdominal ultrasound, CT scan of the abdominal cavity and retroperitoneal space, esophagogastroduodenoscopy, pancreatobiliary endo-ultrasonography.

The size and localization of acute necrotic accumulations were determined according to ultrasound and CT data.

Based on the results of CT, the prevalence of peripancreatic changes was also determined, which, depending on the localization, were divided into 3 types: central (omental sac, projection of the gastro-transverse-colic ligament, root of the mesentery of the small and large intestine), left (paracolon on the left, pararenal and paranephrium on the left, hilum and vascular pedicle of the spleen), right (paraduodenal, paracolon on the right, pararenal and paranephrium on the right).

During esophagogastroduodenoscopy, not only the state of the mucous membrane of the esophagus, stomach and duodenum was visually assessed, but also the bulging of the posterior wall of the hollow organ was determined due to compression from the outside (Fig. 1).

Fig. 1. External compression and deformation of the stomach.

Endoultrasoundography was used to determine the trajectory for safe placement of the stent, based on the minimum distance between the stomach wall and the area of pancreatogenic destruction and the absence of large arterial vessels >1–2 mm in diameter in the intervention area (Fig. 2).

Fig. 2. Endo-ultrasonography (choice of the puncture point).

For endoscopic drainage, an Exera II ultrasonic endoscopic system and an Olympus GFUCT140 convex echoendoscope, a ZiehmImaging X-ray unit, a COOK cystotome, a 0.35 inch diameter string from Olympus, and special self-expanding stents for drainage of pancreatic cysts were used.

Interventions were performed in the X-ray operating room under general anesthesia with the patient in the supine position. A gastro- or duodenocystostomy was performed through the intended point, after which the contents from the purulent focus were aspirated and sent for bacteriological examination to identify the pathogen and adjust antibiotic therapy.

Next, the cavity was contrasted through the cystotome channel to determine its true size, tightness, and connection with the pancreatic ductal system (Fig. 3).

Fig. 3. Contrast enhancement of pancreatogenic destruction zone.

When the contrast agent leaked into the pancreatic duct, endoscopic retrograde pancreaticography, wirsungotomy, and stenting of the main pancreatic duct were performed (Olympus plastic stent, 6 cm long, 5 Fr in diameter) (Fig. 4).

Fig. 4. Stent inside the main pancreatic duct.

A coated self-expanding stent 3.0 cm long and 1.4 cm in diameter was installed in the formed channel between the stomach cavity and the zone of pancreatogenic destruction (Fig. 5).

Fig. 5. Coated self-expanding stent between destruction zone and gastric lumen.

Under X-ray control, a cystonasal drainage with a 7 Fr diameter “pig tail” type was installed through the lumen of the stent into the purulent cavity for daily sanitation of the cavity with 0.05% chlorhexidine solution (Fig. 6).

Fig. 6. Drainage inside the pancreatogenic destruction zone.

The presence of a self-expanding stent made it possible to perform program endoscopic sequestrectomy every 48 hours. For this purpose, a soft distal cap was used to improve visualization, and endoscopic forceps of the “rat tooth” and “crocodile” type, as well as a polypectomy loop were used to remove necrotic tissues (Fig. 7).

Fig. 7. Endoscopic sequestrectomy.

All patients were randomized according to the received basic therapy, including infusion therapy with the predominant use of isotonic polyionic solutions with an electrolyte concentration adapted to the concentration of electrolytes in the blood plasma, 500 ml 2 times a day for 3 days. Pain was relieved by the opioid analgesic tramadol 100 mg or the narcotic analgesic trimeperidine 10 mg intramuscularly. In order to prevent venous thromboembolic complications in the postoperative period, direct-acting anticoagulants were used — enoxaparin 20 mg once a day for 7 days. To reduce the risk of developing erosive and ulcerative lesions of the mucous membrane of the upper gastrointestinal tract, the proton pump inhibitor omeprazole was used at a dose of 20 mg 2 times a day for 7 days. In the postoperative period, all patients received empiric antibiotic therapy. We used the third generation cephalosporin antibiotic for parenteral use ceftriaxone sodium at a dose of 1 g 2 times a day and the antiprotozoal and antimicrobial drug metronidazole at a dose of 500 mg 2 times a day for 7 days intravenously. To prevent fungal infections, oral antifungal therapy with fluconazole 100 mg once was performed. To prevent spasm of smooth muscles of internal organs, spasmolytic therapy with drotaverine 40 mg 2 times a day intramuscularly for 7 days was used.

The effectiveness of endoscopic treatment was assessed by the clinical picture, the endoscopic picture (the absence of sequesters and purulent discharge in the cavity and its size reduction due to the growth of granulation tissue) (Fig. 8) and the dynamics of the decrease in the volume of the cavity according to ultrasound, CT.

Fig. 8. Pancreatogenic destruction zone after endoscopic treatment.

Statistical processing was carried out using the Statisticafor Windows v. 10.0 ("StatSoftInc", USA). The normality of distribution was tested using the Shapiro–Wilk test. When comparing groups by qualitative characteristics, Pearson's χ² test was used. When comparing the quantitative characteristics of the two groups, the Mann-Whitney test was used. Differences were considered statistically significant at p<0.05.

Results

Bulging of the posterior wall of the hollow organ due to external compression, detected during EGDM, was found in 5 (22.7%) patients.

According to the results of CT, pancreatic parenchymal necrosis was observed in 3 (13.6%) patients, of which 2 (9.1%) had an isolated lesion of the pancreatic head, and 1 (4.5%) had an isolated lesion of the body of the pancreas. Subtotal damage to the pancreas in combination with peripancreatic necrosis was detected in 19 (86.4%) patients. Changes spread along the central type in 1 (5%) patient, along the left — in 4 (18%), along the right — in 3 (14%), and in a combination of central and left types in 14 (63%). The volume of lesions according to CT was 344 [158; 510] (10; 800) cm³.

All 22 (100%) patients had acute necrotic accumulations characterized by the absence of a capsule and inhomogeneous content [11, 12].

Connection with the ductal system of the pancreas was found in 3 (13.6%) patients, which required pancreaticoduodenal stenting.

The frequency and duration of program endoscopic sanitation are shown in Fig. 9.

Fig. 9. Elective endoscopic sanitation.

1 patient had 2 sessions, 1 had 4, 1 had 5, 3 had 6, 2 had 7, 1 had 8, 2 had 9, 1 had 12, 1 had 19, and 1 had 21. The rest of the patients did not undergo program sanitation due to death or percutaneous drainage. On average, 7 [6; 9] (2, 21) sanitation.

According to the results of bacteriological culture, Klebsiella pneumoniae was detected in 7 (31.9%) patients, Staphylococcus aureus in 5 (22.7%), Escherichia coli in 5 (22.7%), and Escherichia coli in 4 (18.2%) — Pseudomonas aeruginosa, in 1 (4.5%) — Staphylococcus coagulasenegative. All these bacteria turned out to be sensitive to III generation cephalosporins, which proves the effectiveness of the antibiotic therapy used.

In all patients, after intraluminal drainage, free gas was detected in the abdominal cavity. Despite this, the patients continued conservative treatment, which included the placement of a nasogastric tube, correction of fluid and electrolyte disorders, and adequate pain relief. On the 3^rd day, during the control X-ray examination, free gas in the abdominal cavity was not determined.

The pain syndrome was stopped on the 3^rd day, the body temperature returned to normal on the 14^th—52^nd day.

A complicated postoperative period was recorded in 4 (18.1%) cases. Four patients were diagnosed with grade III complications according to the Clavien-Dindo classification, IIIa — in 1 (4.5%), IIIb — in 3 (13.6%).

The early postoperative period was complicated by bleeding from the zone of pancreatogenic destruction in 4 (18.1%) patients, which manifested by the flow of scarlet blood through the drainage. In 2 (9.1%) cases, this required angiography and endovascular embolization a. gastroduodenalis, in 1 (4.5%) — endoscopic hemostasis by filling the cavity with 4 ml of hemostatic solution Hemoblock. The combination of both methods was used in 1 (4.5%) patient. There was no evidence of recurrent bleeding in these patients.

Leaks with suspension and sequesters that do not drain into the lumen of the stomach in the form of irregularly shaped hypoechoic zones with fuzzy, uneven contours with suspension and sequesters, spreading to the gastrocolic ligament, the root of the mesentery of the small intestine, mesocolon during control ultrasound were noted in 11 (50%). The identified accumulations required additional percutaneous drainage under ultrasound guidance.

In 11 (50%) patients, endoscopic drainage of fluid collections was the final method of surgical treatment of necrotizing pancreatitis. The total duration of inpatient treatment was 45 [40; 57] (36; 77) days, the period of stay in the intensive care unit was 11.5 [7; 17] (1; 27) bed-days.

The lethal outcome occurred in 4 (18.1%) patients after the resolution of the purulent-inflammatory process, which was confirmed in the section. Multiple organ failure was the cause of death in 3 (13.6%) patients. In 1 (4.5%) patient, the cause of death was severe nosocomial pneumonia that developed on the 32^nd day after drainage.

Thus, out of 22 patients, only 11 (50%) patients required additional surgical intervention in the form of percutaneous drainage.

Discussion

Despite the fact that at present the arsenal of surgical treatment of necrotizing pancreatitis is quite wide, mortality in patients of this group remains quite high and reaches 65% [5, 7, 13—16]. The development and implementation of methods of mini-invasive surgery and intraluminal endoscopy are among the promising areas to improve the results of treatment of patients with purulent-inflammatory diseases of the pancreas. Percutaneous and transluminal drainage interventions have found wide application. According to the work performed, transluminal drainage of pancreatogenic destruction zones in necrotizing pancreatitis can be considered as a potentially final method of treating patients in 50% of cases, which is comparable to the world literature data [7, 16]. The high efficiency of this technique directly depends on the timeliness of the start of endoscopic treatment and further sanitation of the purulent focus adequate to the extent of the lesion, as well as on the correctly selected antibiotic therapy prescribed based on the results of antibiotic sensitivity. Additional percutaneous drainage was required in 50% of cases, which is comparable with the results of treatment in other leading international clinics [4—6].

Postoperative complications developed only in 18.1% of cases, and all were stopped by a minimally invasive method (X-ray endovascular embolization of a bleeding vessel), which makes this technique even safer in terms of predicting the outcome of the disease. The presence of free gas in the abdominal cavity after intraluminal drainage, in our opinion, is not an absolute indication for surgical treatment, since with proper conservative treatment it stops on the 3^rd day.

Conclusion

The study proves that endoscopic drainage of pancreatogenic destruction zones is a promising minimally invasive method in the complex treatment of necrotizing pancreatitis.

The authors declare no conflicts of interest.