Introduction

Pylorus-preserving gastrectomy (PPG) has been quite widely performed for early gastric cancer located in the middle of stomach [1,2,3,4]. Although PPG has advantages of reducing bile reflux, dumping syndrome, gastritis, gallbladder stones formation, malnutrition, and suppression of immune function, and it has similar oncological safety compared to distal gastrectomy [5,6,7,8], postoperative gastric stasis or delayed-gastric emptying (DGE) is the most common and challenging complication of this surgery, occurring in 7–23% of cases [9,10,11,12,13].

Irrespective of the cause of pyloric stenosis, gastric stasis or DGE after PPG not only delays the recovery of patients, but it sometimes requires invasive treatment. Therefore, methods, such as intraoperative manual dilatation of the pylorus, botulinum toxin injection, postoperative gastric balloon dilatation, and/or temporary stent insertion have been introduced in previous studies to reduce the incidence of pylorus dysfunction [14,15,16]. In Korea, endoscopy-guided balloon dilatation and/or stent insertion are performed by gastroenterologist, and fluoroscopy-guided balloon dilation and/or stent insertion are performed by radiologist. While intraoperative endoscopy and balloon dilatation may also be effective alternatives, considering the Korean medical system, it is not an easier and cost-effective way to perform. Therefore, it was thought that if an absorbable stent is placed intraoperatively in the pylorus to expand the pyloric sphincter muscle even for a short time immediately after surgery, it would reduce the occurrence of DGE. In addition, if the stent dissolves on its own, it would preclude a secondary removal procedure and potentially be effective for decreasing the patient morbidity caused by migration or the extraction procedure.

The idea of materials needed to make a biodegradable stent came from two different kinds of synthetic absorbable sutures that are commercially available: the poly(glycolide-co-caprolactone) (PGCL) stent (Vicryl®; Ethicon, Inc.) with a braided construction and the poly-p-dioxanone (PPDO) stent (PDS II®; Ethicon, Inc.) made of monofilament [17]. Although a monofilament suture carries less risk of infection than a braided suture, it dissolves slower than a braided one [18]. However, when the two materials are made in the form of a stent, it is unknown whether biodegradation is feasible and which material would be more suitable for use as a pyloric stent, so each stent was made of each material. These first-generation novel stents were made using the three-dimensional (3D) printing technique (a patent was filed for both stent types).

The aim of this study was to investigate the feasibility and safety of inserting the novel self-biodegradable pyloric stents (PGCL and PPDO) in porcine models by evaluating the duration of biodegradation and assessing the local mucosal reaction and/or systemic adverse events.

Methods

Design and manufacture of the pyloric stents

The two different types of biodegradable scaffold stents were constructed using the same material as commercially available synthetic absorbable sutures, PGCL and PPDO (Fig. 1a). The stents were designed in the shape of a dumbbell with both ends open by 3D printing technology to avoid migration. The sizes of the stents were 30 mm in length, 17 mm in diameter for the middle part, and 22 mm in diameter for both open ends. A radiographically detectable tag was wrapped around both ends. Each sterile stent was packed separately for the animal experiments (Fig. 1a).

Fig. 1
figure 1

Animal experiment. a Design and type of the biodegradable pylorus stent. b Surgery for stent insertion and different number of tagging suture

Full size image

Animals and the study design

Ten domestic male pigs were divided into two groups used for implantation of the biodegradable PGCL stent (n = 4) and PPDO stent (n = 4), and one group without stent implantation (shame, n = 2) (Supplementary Table 1). Pigs weighing about 50–60 kg were chosen as the experimental model because their stomach anatomy and the size of pylorus are similar to that of humans. The number of pigs was decided taking into account our research funding limits. Pigs were placed into the experimental animal facility of the Seoul National University Hospital Biomedical Research Institute and fed 550 g of a formulated feed mixture twice a day (at 9 am and 4 pm). Free access to water was provided all day. After 1 week of an adaptation period, we proceeded with operation for stent insertion.

In terms of both potential scientific contribution and ethics in animal protection, the study was approved by the Institutional Animal Care and Use Committee at Seoul National University Hospital (IACUC no. 19–0186-S1A1).

Procedures

Anesthesia

Pigs were fasted for 24 h before surgery with free access to water. Alfaxalone (2 mg/kg) and xylazine (2.5 mg/kg) were injected intramuscularly 30 min before anesthesia induction. After endotracheal intubation, anesthesia was maintained with inhaled 60% oxygen and 2% isoflurane until the end of the surgical procedure. Hartmann’s solution (60 ml/h) was perfused intravenously during the operation after bolus injection of a prophylactic antibiotic (cefazoline, 1 g) and analgesic (tramadol, 100 mg).

Operative procedures

With pigs under general anesthesia, preoperative body weight, body temperature, complete blood cell count, and X-ray findings were collected. A sterile drape was used to isolate the abdomen for surgery. Through 15-cm upper midline abdominal incision, a 7-cm vertical gastrotomy was made on the low body and anterior wall of the stomach. Each stent was inserted in the pylorus so that the distal end of the stent was placed in the duodenum, the middle part was placed in the pylorus, and the proximal end was placed in the stomach. For each group of animals, a different number of tagging sutures was performed with 3–0 Vicryl thread to fix the proximal end of the stent to the stomach wall: zero tagging sutures (t0), one tagging suture on the anterior wall (t1), two tagging sutures on the anterior wall and posterior wall (t2), and four tagging sutures at 3, 6, 9, and 12 o’clock (t4). For the shame group, gastrotomy was performed the same way without any stent insertion (Fig. 1b). The gastrotomy was closed with a continuous interlocking double layer suture. The abdominal wall was closed layer by layer.

Postoperative follow-up

Feeding of the pigs resumed on POD 1. The same doses of the antibiotic (cefazoline, 1 g/day) and analgesic (tramadol, 100 mg/day) were injected intramuscularly until POD 3. The body weight, body temperature, complete blood cell count, and X-ray abdomen were collected on PODs 1, 3, 6, 8, 10, and 13 under monitored anesthesia care (intramuscular injection of 2 mg/kg alfaxalone and 2.5 mg/kg xylazine). On the POD 13, distal gastrectomy without lymph node dissection was performed for subsequent pathological examination.

Euthanasia

On the scheduled day of sacrifice (POD 13), pigs were fasted for 12 h before surgery with free access to water. The animals were anesthetized with the same anesthetic drugs and procedures used for the initial operation for stent insertion. After distal gastrectomy, the pigs were euthanized with an intravenous injection of potassium chloride after the surgery.

Pathological examination

The intestinal metaplasia was evaluated to prove that no pig was affected by chronic gastritis from the beginning and to determine whether the stent insertion causes chronic gastritis and/or atrophic gastritis. Inflammatory cell infiltration in the pylorus and duodenum was counted to evaluate for gastritis or duodenitis. Because there is no definite standard for how many lymphocytes are normally in a pig’s gastrointestinal (GI) mucosa, a comparative evaluation was performed based on the degree of inflammatory cell infiltration of the shame group. The infiltration of neutrophils into the glands was considered to be active inflammation, as routinely assessed in human endoscopic biopsy evaluation.

Statistics analysis

The data were recorded during each experiment. SPSS 26.0 software (IBM Corp.) was used to perform data analysis. P < 0.05 was considered to indicate a significant difference, and two-way repeated measures analysis of variance (ANOVA) was used in each analysis to detect the effects of the different stents.

Results

A total of ten pigs underwent successful operation with or without stent insertion. There was no case of intraoperative or immediate postoperative mortality (within POD 7).

Timing of the stent degradation

The PGCL stent self-degradation was observed on POD 3 as the middle part of the stent, which receives the strongest radial force by pylorus, was broken (Fig. 2). In all PGCL stents with more than one tagging sutures, the middle part of the PGCL stent was broken and separated into two pieces: the PGCL-t1 and PGCL-t2 stents on POD 3, and PGCL-t4 stent on POD 6. On POD 6, the shape of stent was deformed as the stent self-degradation progressed. The shape deformities of the proximal parts of the X-ray detectable tag were observed: the PGCL-t1 and PGCL-t2 stents on POD 6, and PGCL-t4 stent on POD 6 and POD 8. (Fig. 2). All PGCL stents were completely dissolved within POD 8 to 10. No stent remained in the digestive tract in all four pigs when the whole bowel was inspected after euthanasia.

Fig. 2
figure 2

Degradation and migration of PGCL stent with different number of tagging suture: follow-up X-ray results of PGCL stents

Full size image

The shapes of the PPDO stents were maintained for 13 days on radiography. Self-degradation of the PPDO stent also proceeded, but slower (Fig. 3a). Even though the PPDO-t1 stent was found in the colon with an intact shape during euthanasia, the middle part of the stent was immediately broken with even light force and divided into two pieces (Fig. 3b). The PPDO-t4 stent remained in the pylorus until the day of euthanasia. It was more fragile than PPOD-t1 stent that it was crushed into small pieces with very light force. On gross specimen, ulcerative lesion was noticed with weakened PPDO-t4 stent in the duodenum (Fig. 3c).

Fig. 3
figure 3

Degradation and migration of PPDO stent with different number of tagging suture. a Follow-up X-ray results of PPDO stents. b PPDO-t1, which remained in the large intestine at the time of euthanasia. c PPDO-t4, which remained in the duodenum at the time of euthanasia

Full size image

Stent migration

Both PGCL and PPDO stents without tagging suture (t0) were remained in pylorus until the day after surgery. Migration of both stents was started before POD 3 without any shape deformation. Both stents were already excreted from the GI tract by POD 6 (Fig. 4). Migration of both stents was delayed by applying different number of tagging suture. In both PGCL and PPDO stents, one (t1) or two (t2) tagging suture held stents at the pylorus until POD 6. In both stents with four tagging sutures (t4), stent migration was not observed during the stent self-degradation (Figs. 2, 3a).

Fig. 4
figure 4

Migration of PGCL and PPDO stent without tagging suture

Full size image

In the PPDO stent model, because there was no stent break or stent shape deformity due to self-degradation within the study period, as the number of tagging sutures increased, migration of the stent was delayed: the PPDO-t1 and PPDO-t2 stents started migration on POD 6, whereas the PPDO-t4 stent did not migrate until POD 13 (Fig. 3a).

The difference of PGCL stent was that because self-degradation proceeded faster in PGCL than PPOD, the middle part of the PGCL stent was broken first on POD 3. Therefore, t1 or t2 tagging sutures held only the proximal part of the stent until POD 6. As the self-degradation of PGCL stent progressed, the shape deformation of X-ray detectable tag was observed in the proximal side of stent. In PGCL-t4 stent, even though stent break was not observed on POD 3, only the X-ray detectable tag of proximal side of stent was observed in the pylorus on POD 6 and 8, along with the shape deformation (Fig. 2).

Body weight, body temperature, and white blood cell

Loss of body weight was observed in all pigs with a PPDO stent, except the PPDO-t0, which migrated by POD 3. In the contrast, there was rather weight gain in all pigs with PGCL (Supplementary Table 2). However, according to two-way repeated measures ANOVA, there was no significant differences in the body weight, body temperature, and white blood cell (WBC) (neutrophil) counts between groups (shame versus [vs.] PGCL vs. PPDO). There was no significant change in the body weight, body temperature, and WBC (neutrophil) counts over time between the three groups (Table 1).

Table 1 Repeated measures ANOVA for weight, body temperature, white blood cell and neutrophil
Full size table

Adverse event

For both PGCL and PPDO stent, no severe adverse event related to the stent autolysis, such as local and/or systemic inflammation, was observed during the 13-day follow-up. However, unexpected death was observed in the pig with a PPDO-t2 stent. The pig suddenly died on POD 10 with vomiting and seizure which started 2 h prior to death. Stent migration to the left lower quadrant (LLQ) area was observed on POD 6. However, the stent was completely dissolved and only the radiographically detectable tag was observed on POD 8 in LLQ area (Fig. 3a). The pig’s body temperature was 39.3 °C (normal range: 38.7 ± 1 °C). However, WBC (neutrophil) count increased up to 42.0 × 10E3/µL (73.9%) on POD 8 (Supplementary Fig. 1). On autopsy, focal narrowing was observed in the duodenum, and there were ischemic changes in the whole intestine below the duodenum. On gross specimen, a duodenal stricture was observed, and acute volvulus was suspected as the cause of death after the autopsy (Fig. 5).

Fig. 5
figure 5

Autopsy and pathology result of PPDO-t2

Full size image

Pathology

On histopathological examination, intestinal metaplasia was not observed in all ten pigs. Inflammatory cell infiltration was slightly higher in the PPDO group compared to the PGCL group. However, inflammatory cell infiltration in the PPDO group was within a similar range when compared to that in the shame group. No active inflammation was observed, except in the PPDO-t4 stent case with a duodenal ulcer caused by physical pressure from the stent edge (Table 2).

Table 2 Pathology result
Full size table

Discussion

Despite the significant advantages of preserving the pylorus, the complication of gastric stasis and DGE are also well known [19,20,21,22]. The incidence of gastric stasis after PPG has been reduced [23], but still immediate postoperative pyloric stenosis is one of the reasons why many surgeons avoid PPG [24]. Even though the exact cause of DGE after PPG is unknown, postoperative pylorus spasm or dysfunction due to thermal tissue damage, nerve injury, and/or reduced vascularity during the surgery may be the causes of gastric stasis or DGE after PPG.

Some patients with DEG after PPG recover quickly in a few days with nasogastric tube insertion and fasting. However, some patients need intervention such as balloon dilatation and/or transient stent insertion up to 2 weeks. In our previous study, gentle manual dilation of the muscular layers of the pylorus with intestinal clamp during the operation resulted in significant reduction of pylorus stenosis by 1.1%, compared to non-manual dilation (8.6%) [14]. Intraoperative manual dilatation and postoperative balloon dilatation are effective procedures to reduce and/or resolve DGE. However, those procedures are one-off treatments. Some of the patients still suffer from DGE even after postoperative balloon dilatation.

For those patients with severe DGE, fully covered retrievable stents have been used to provide temporary pylorus dilatation (approximately 1–2 weeks) at our institute. Although the outcomes are acceptable, this strategy is time consuming and expensive, and the risk of stent migration is still problematic [25]. Moreover, stent insertion requires a separate procedure for removal, since all the commercially available stents for benign stricture are made of non-absorbable covered stents [16, 26, 27]. Based on the experience that the simple routine procedure of intraoperative manual dilatation reduced gastric stasis, and that the severe gastric stasis was resolved by mounting a retrievable stent to keep the pylorus open for a short time immediately after surgery, we developed the idea of a biodegradable stent. Even though it is not appropriate to routinely place the biodegradable stent to all patients receiving PPG, it is expected that the preventive insertion of the biodegradable stent may be helpful in some certain cases that high risk of postoperative DGE is considered based on the surgeon’s experience.

For novel biodegradable stent, we used two materials, PGCL and PPDO, which are commercially available absorbable surgical suture materials [17, 18]. The stent market is mostly developed in the cardiovascular area [28]. Only a few attempts to make a biodegradable stent for esophageal stricture, biliary stricture, and urinary tract stricture with similar materials have been reported [29,30,31,32,33]. However, this is the first time an absorbable pylorus stent was made for gastric stasis after PPG.

We first focused on whether stents could persist against the force of the pylorus sphincter. Gastric stasis, which is caused by transient pylorus dysfunction after PPG, is different from cancer obstruction or benign stricture. The stent should be designed to not only prevent the pylorus from remaining narrowed, but also to prevent the sphincter function from becoming deteriorated because of too much stretch.

Next, we considered how to prevent stent migration, which is one of the biggest problems with currently available stents on the market. We made the dumbbell-shaped stents so that the thin middle part of the stent fits into the pylorus, and the wider open both ends are located in the stomach and duodenum, respectively (Fig. 1a). Because the stents were made with a diameter of 1.7 cm in the middle thin part according to the general pylorus diameter of the human adult pylorus and constructed by a 3D printer with less flexibility and elasticity, the insertion of the stent during the surgery was not smooth; thus, it was expected that the stent would not migrate easily. However, both stents without tagging sutures began to migrate before POD 3, and were excreted from the digestive tract before POD 6 (Fig. 4). This problem was solved by simply adding a tagging suture to the stents. In both stents, the addition of one or two tagging sutures between the stent and the stomach delayed migration until POD 6 (Figs. 2, 3a). However, since this study was not conducted in PPG pig model, further studies on the ideal timing of stent degradation and/or migration in PPG pig models should be performed for this stent to be used in clinical practice.

In PPDO stents with four tagging sutures, the stent remained in the pylorus for 13 days (Fig. 3a). On the other hand, the PGCL stent disintegrated much faster than the PPDO stent (Fig. 2). Even though the PGCL stent also remained in the pylorus with a tagging suture, it broke into two pieces. The PGCL stent would not be able to resist against the tightening force of the pylorus because of its faster degradation. The proximal part of the PGCL stent remained in the stomach area with tagging sutures, whereas the distal part of the PGCL stent migrated downstream of the intestine immediately after the stent broke. The degradation of both parts of the stents was observed on POD 3, and they completely degraded by POD 10.

There was one case of sudden death with the PPDO stent (PPDO-t2) during the follow-up (Fig. 5). After autopsy, the cause of death and whether this event was related to the stent were discussed with the veterinarian and pathologist. Acute volvulus was suspected as the cause of death. Since the migration of the stent had already started on POD 6 and was observed in the LLQ area with an intact shape, this adverse event would not have been directly related to the systemic side effects of the stent. Additionally, the PPDO-t4 stent remained in the pylorus until POD 13, and systemic side effects were not observed. Anesthesia was proceeded every follow-up day for radiography and blood sampling. Therefore, acute volvulus may have been caused by early feeding (free access to water) without full recovery from the aesthesia. Except for this case, there was no severe systemic inflammation caused by stent placement.

Only two stents (PPDO-t1 and PPDO-t4) out of ten pigs remained in the GI tract during euthanasia. The migration of PPDO-t1 started on POD 6 and this stent was found in the colon with an intact shape during euthanasia; however, it broke into two pieces with very light force (Fig. 3b). Local inflammation of the pylorus and duodenum were not detected on the pathological examination. The PPDO-t4 stent remained in the pylorus until the day of euthanasia. The shape was intact; however, it was very fragile (Fig. 3c). Therefore, it was broken into many pieces during the specimen extraction. Grossly, the ulcer was observed on the duodenum. After pathological examination, it was confirmed to be a duodenal ulcer, which is caused by continuous mechanical force from the open end of the stent. There was no local inflammatory change related to the stent material.

Combining the above results, it seems that maximal time of both stents can be placed on the pylorus without any other systemic complication seems to be up to 3 to 6 days after surgery. PGCL stent could be stayed up to POD 3 by adding four tagging sutures because self-degradation of the PGCL stent is faster. In the case of PPDO, four tagging sutures can be mounted in pylorus for up to 13 days, but local duodenal ulcer may occur due to continuous scratch from stent edge. One or two tagging sutures mounted PPDO stent in pylorus until POD 6.

The changes of body weight seemed slightly higher in the PGCL group than in the shame and PPDO groups (Supplementary Table 2). During the 13 days of experimental period, PGCL stent only existed up to 8 days in GI tract, and even that was mostly in the degraded form. Therefore, it seemed that even though PGCL pigs experienced diarrhea for a few days immediately after surgery with stent insertion, they were able to recover in good condition quickly. On the other hand, PPDO stent persisted in pylorus with intact shape until POD 13. Therefore, it is thought that the relatively higher body weight loss was observed in PPDO pigs due to some discomfort in the GI tract caused by the stent. These results suggest that the placing stent in the pylorus may cause mild GI symptoms such as diarrhea or poor oral intake. Even though two-way repeated measures ANOVA showed that insertion of both stents did not cause any significant changes over time in the body weight, body temperatures, and WBC (neutrophil) counts within and between the three groups (Table 1), further studies are needed to investigate whether the stents actually cause GI symptoms.

The purpose of this experiment was to prove that the concept of the biodegradable pylorus stent is possible and safe. Therefore, this preliminary study has some limitations. First, the pigs were not representative of a clinical pyloric stenosis model after PPG. If pyloric stenosis is present, radial force may be higher than that in a normal pylorus. Further studies should evaluate how effective these stents are in actual pyloric stenosis after PPG. Second, because the stent was constructed with 3D printing and had a rigid dumbbell shape, the insertion of the stent in the pylorus was not easy. Since totally laparoscopic PPG is performed widely in practice these days [13, 34], we are developing self-expanding biodegradable stent with a flexibility and elasticity based on the previous studies [29, 31]. These stents may also be mounted endoscopically or in totally laparoscopic PPG. Lastly, the dumbbell shaped did not prevent stent migration as much as we expected. Therefore, we plan to develop a mechanism that can prevent stent migration more effectively.

In conclusion, the concept of biodegradable stents made of absorbable suture material seems feasible in a porcine model. Although biodegradable stents are considered safe in terms of local and systemic inflammation caused by foreign body reactions, since there has been an unexpected death case due to volvulus, early absorbing PGCL would be preferable material for this absorbable stent compared to PPDO. However, in order for these stents to be used in clinical practice, further study is planned to evaluate the efficacy and safety of the next-generation stent which has a self-expandable function and a migration-preventing mechanism.