Introduction

Various methods of surgical treatment of infected pancreatic necrosis lead to the need for choosing the most effective surgical approach [1–10]. Currently, there are trends towards active implementation of minimally invasive technologies because traditional open surgical debridement of purulent-necrotic parapancreatitis do not significantly improve the outcomes and survival [2, 3, 6, 11, 12]. Various minimally invasive techniques have been developed: interventions with “Mini-assistant” equipment for small-access surgery [13, 14], laparoscopic debridement of omental bursa and retroperitoneal space [15–17], transluminal [18] and puncture-drainage procedures [2, 3, 19]. The last ones are perspective due to minimal surgical trauma, short-term anesthesia and low number of perioperative complications [3, 20, 21]. However, many surgeons use ultrasound-guided minimally invasive interventions only at the first stage of treatment of pancreatic necrosis for evacuation of purulent exudate. Puncture-drainage technologies for necrectomy at subsequent stages cause great difficulties that prevents widespread application of this method and causes a critical attitude to its effectiveness [22].

The purpose of the study was to develop an effective minimally invasive method for the treatment of infected pancreatic necrosis.

Material and methods

There were 168 patients with infected pancreatic necrosis who underwent minimally invasive treatment at the Regional Clinical Hospital No. 2 between 2011 and 2018. Age of patients ranged from 43 to 84 years. There were 107 (63.7%) men and 61 (36.3%) women. Elderly (62.9%) and middle-aged (26.1%) patients prevailed; 147 (87.5%) patients were delivered from regional hospitals in 12.3±2.6 days after onset of disease.

Primary drainage with large double-lumen drains 28-32 Fr was performed in 87 (51.8%) patients. The developed technique of transfistulous endoscopic sequestrectomy was applied to remove necrotic tissue [23].

Percutaneous interventions followed by transfistulous endoscopic sequestrectomy were performed in 23 (26.4%) patients with local and 64 (73.6%) patients with widespread purulent-necrotic parapancreatitis. Severity of disease and multiple organ failure was assessed using the integral scales Apache II, Ranson, SOFA. Fifty-nine (67.8%) patients had concomitant diseases.

Transfistulous endoscopic sequestrectomy was performed as follows. After preliminary visualization of exudative and necrotic lesion of omental bursa and retroperitoneal space using ultrasound and pre-performed contrast-enhanced computed tomography of abdominal cavity and retroperitoneal space, we planned surgical access. The main requirements for accesses were as follows: no great arteries and veins along the intended trajectory, as well as no hollow and parenchymal internal organs. Omental bursa was punctured through the gastrocolic ligament and retroperitoneal space (paracolon) on the right and/or on the left along the middle or posterior axillary line using a needle for primary access under ultrasound control. Then, we inserted 2-3 rigid guidewires into omental bursa and 1-2 ones into retroperitoneal space through this needle. After that, the canal was dilated under ultrasound control, and hollow pipes were installed. Then, drainage tubes 28–32 Fr were inserted through the pipes.

After 7–10 days, drainage tubes were removed in operating theatre. After that, flexible endoscope, laparoscopic clamp or dissector and silicone drainage tube connected to electric aspirator (if there was a third access) were inserted through the channels.

Saline solution was supplied to surgical field through the endoscope to facilitate visualization of sequesters and washing the omental bursa. A clamp was used to dissect and remove necrotic tissue (Fig. 1). Intervention was continued until all sequesters were removed as much as possible (Fig. 2, 3).

Fig. 1. Intraoperative endoscopic image. Stage of sequestrectomy. Capture and removal of sequester.

Fig. 2. Intraoperative endoscopic image after removal of sequesters.

Fig. 3. Postoperative image. Macroscopic specimen. Removed sequesters after transfistulous endoscopic sequestrectomy.

Surgical intervention was finished by drainage of omental bursa and retroperitoneal space with double-lumen tubes under endoscopic control. In postoperative period, we performed fractional washing of cavities with aqueous solutions of antiseptics (chlorhexidine, sodium hypochlorite, etc.).

Progression of endogenous intoxication required repeated endoscopic sequestrectomy. In case of positive dynamics, we replaced drainage tubes every 7 days under X-ray control with contrast enhancement (to prevent their occlusion with large sequesters) and continued fractional washing every 6 hours.

All patients received standard therapy.

Statistical analysis was performed using Statistica 10 software. Differences of means were assessed using parametric Student's t-test. Differences were significant at p<0.05.

Results

Approaches to omental bursa and retroperitoneal space is the first stage of minimally invasive treatment implying subsequent removal of necrotic tissue via transfistulous endoscopic sequestrectomy. A total of 352 minimally invasive interventions were performed: drainage of omental bursa — 23, drainage of omental bursa and retroperitoneal space — 64, replacement and correction of drainage tubes — 265. There were 3 accesses to omental bursa in 69 (79.3%) patients and 2 accesses in 18 (20.7%) patients. There were 2 accesses to retroperitoneal space in 54 (62.1%) patients and 1 access in 33 (37.9%) patients.

Active removal of sequesters using transfistulous endoscopic sequestrectomy was performed 98 times (once in 78 (86.2%) patients, twice in 7 (8.6%) patients, 3 times in 2 (5.2%) patients). Surgery time in patients with 3 approaches to omental bursa was 62±4.3 min, 2 accesses — 96±8.2 min (p<0.05). Time of redo sequestrectomy in patients with 3 accesses was 47±5.6 min, 2 accesses — 52±4.7 min (p>0.05). Duration of transfistulous endoscopic sequestrectomy was similar regardless the number of accesses to retroperitoneal space.

The following local complications occurred: external intestinal fistula in 1 (1.1%) patient, internal intestinal fistula in 2 (2.3%) patients, multiple organ failure in 5 (5.7%) patients.

No bleeding was recorded. Three (3.4%) patients developed venous bleeding during sequestrectomy that was stopped by temporary occlusion of drains after surgery.

There were the following extra-abdominal complications: pulmonary embolism in 3 (3.4%) patients, acute coronary syndrome in 1 (1.1%), lower lobe left-sided pneumonia in 4 (4.6%) patients, exudative pleuritic in 9 (10.3%) patients.

Laparotomy was required in 2 (3.4%) patients due to progression of sepsis despite minimally invasive interventions combined with endoscopic sequestrectomy. Mortality rate was 12.6%.

Discussion

There were previous multiple attempts of sequestrectomy via percutaneous approaches. MIPN (minimally invasive pancreatic necrosectomy) and MARPN (minimal access retroperitoneal pancreatic necrosectomy) techniques were developed and proposed. They suggested a staged approach to the treatment of infected pancreatic necrosis with primary drainage through the tubes 8 Fr followed by dilation of channel up to 30 Fr before sequestrectomy. Active debridement of necrotic tissue was carried out using a nephroscope inserted through the channel [2, 24, 25].

Importantly, one access to pathological zone and small channel of nephroscope (4 mm) complicate effective sanitation and increase surgery time that can adversely affect severe patients with concomitant diseases.

The main feature of percutaneous interventions in omental bursa and retroperitoneal space is surgical manipulations within a small space (small space surgery). Transfistulous endoscopic sequestrectomy was proposed to increase efficiency and reduce duration of necrectomy. The main directions of this procedure were aimed at improving intraoperative visualization, creating the optimal number of accesses and combining the methods of sequestrectomy.

Improving visualization during sequestrectomy

One of the most important techniques determining the effectiveness of transfistulous endoscopic sequestrectomy is visualization of surgical field and equipment. Various optical instruments were used for visualization during development and testing of technique: laparoscope, nephroscope, broncho- and choledochoscope, gastroscope. Obvious advantages of flexible optical systems were confirmed. Indeed, these systems allow passing through different channels and examine the cavities with irregular configuration. Among flexible optical systems, standard gastroscope with a distal end diameter of 9.4 mm proved to be the most effective device. Gastroscope has a high optical resolution and ensures high-quality imaging.

Optimal number of accesses

In our practice of transfistulous endoscopic sequestrectomy, we made 2–3 accesses to omental bursa sac and 1–2 accesses to retroperitoneal space. Significant relationship between the number of approaches to omental bursa and surgery time was found. Three available accesses can reduce surgery time by 34 minutes. Prolonged sequestrectomy in patients with two accesses is explained by insertion of gastroscope and simultaneous injection of solutions for sanitation through one access and alternate insertion of electric aspirator tip and instruments through the second access. This significantly increases surgery time. No significant difference in time of repeated sequestrectomy was found that seems to be related to creation of necessary workspace during the first procedure. The number of accesses to retroperitoneal space did not significantly affect duration of intervention.

Combination of types of sequestrectomy

One of the features increasing effectiveness of active necrectomy was combination of instrumental and aspiration sequestrectomy. In case of instrumental sequestrectomy, we dissected tissues and adhesions using a laparoscopic clamp or dissector. When manipulating in omental bursa, we created working space after separation of posterior stomach wall from anterior surface of the pancreas. After creating the working space, we washed the cavity with antiseptic solutions. Then, we performed instrumental palpation and removal of free-lying sequesters using a laparoscopic clamp. All stages were accompanied by permanent supply and aspiration of solutions (0.05% aqueous solution of chlorhexidine and 3% solution of hydrogen peroxide). Aspiration-assisted sequestrectomy was another element that increased the efficiency of necrectomy. Alternative use of different plastic tips with a diameter of 8–9 mm and silicone drains with end and side working holes allowed the vacuum aspirator to capture and remove sequesters.

Conclusion

Transfistulous endoscopic sequestrectomy provides adequate minimally invasive removal of parapancreatic necrotic tissue. Several accesses make it possible to carry out a staged sequestrectomy under visual control within a sufficient surgical space. Effective debridement of necrotic focus, especially in severe and debilitated patients, allows us to characterize this procedure as an important component in improving the results of treatment of infected pancreatic necrosis.

Original technique of transfistulous endoscopic sequestrectomy increases efficiency of sanitation of infected parapancreatitis, improves treatment outcomes and reduces mortality to 12.3%. This is due to better visualization during sequestrectomy through the gastroscope, optimal number of accesses to omental bursa and combination of instrumental and aspiration-assisted sequestrectomy.

The authors declare no conflicts of interest.