Optimizing ischemic preconditioning for esophagogastric anastomosis: A standardized rat model to evaluate histological changes

INTRODUCTION

Anastomotic dehiscence and leakage have been reported in 10% of patients after esophageal resection for cancer and represent a serious complication resulting in significant morbidity and mortality [1]. Depending on the site of the tumor, cervical or intrathoracic anastomoses are usually constructed using gastric conduit [2]. Although esophagogastric anastomosis is more fragile compared to all other anastomoses of the gastrointestinal tract, especially when cervical anastomosis for higher reconstruction is indicated, many surgeons prefer to use cervical anastomosis rather than intrathoracic [3]. Many factors are linked to the failure of esophagogastric anastomosis. One of the most delicate aspects is the ischemia of the gastric conduit. Ischemia is a major cause of anastomotic dehiscence leading to leakage and a sequence of fatal aberrances. Hence, the elimination of anastomotic leakages is essential for improving the morbidity and mortality after esophagectomy [4].

The stomach is supplied by the left gastric artery (LGA), right gastric artery (RGA), and gastroepiploic arteries. Three out of four main branches are closed by ligation during gastric mobilization, and the right gastroepiploic artery (RGEA) remains the principal supplier of the gastric conduit [5]. The upper part of the gastric tube is vascularized only by the intramural network and is thus the most vulnerable part for ischemia [6,7]. An alteration of venous drainage of the gastric tube may also increase the possibility of ischemia of this special conduit. The degree of ischemia could be amplified by forced manipulation with tension and strangulation during surgery [4].

A shortage in blood flow to any organ or tissue causes cellular dysfunction and death [8]. After a certain period of ischemia, the gastrointestinal tissue becomes vulnerable to damage due to reperfusion, in which the restored oxygen supply produces free radical species that lead to further tissue injury. Ischemic reperfusion damage after clamping the celiac artery induces gross damage to gastric mucosa, resulting in the formation of petechiae and ulcers deep down in the mucosa [9]. It is believed that ischemia after LGA, RGA, and left gastroepiploic artery (LGEA) ligation is the dominant factor resulting in future anastomotic leakage [10]. Therefore, numerous experimental and clinical studies have focused on the prevention of ischemia. Gastric ischemic conditioning prior to surgery was verified as a method to improve the vascularization of the gastric conduit [4,10−16]. However, available meta-analyses have confirmed conflicting results because of search strategy discrepancies, methodological incongruities due to inclusion of single arm studies and natural heterogeneity of the studies. A lack of a randomized study confirming the benefit of gastric ischemic conditioning (GIC) is also noticeable. On the other hand, there is still no consensus on the optimal timing of GIC [11,12,16].

Hence, this study aimed to establish a rat gastric ischemic conditioning model, to observe histological changes following gastric ischemia, and to standardize the evaluation of selected immunocompetent cell counts. The use of a reproducible model is necessary for a future investigation of GIC effect on esophagogastric anastomosis viability. Recording histological changes following GIC is assumed as an important factor in predicting anastomosis failure. Furthermore, sophisticated understanding of gastric conduit ischemia from the point of view of histological changes observation and immunocompetent cell counts may be useful in optimal timing for esophagogastric anastomosis creation.

METHODS

Ethical statement:

The experiment was approved on May 17th 2021 by the Ethics Committee of the Faculty of Medicine of Pavol Jozef Šafárik University and by the State Veterinary and Food Administration of Slovakia (48935/2021-220).

Animal model:

Male Sprague-Dawley rats (n=24) weighing 500–600 g were obtained from the Animal Facility of P. J. Šafárik University and used for the experiment. The animals were individually housed in plexiglass cages under standard conditions (55±5% humidity, 22±2 °C, 12/12 h light-dark cycle) with free access to standard laboratory diet and tap water. All rats were fasted for 12 hours (with free access to water) before the experiment. The rats were randomly divided into four groups: three ischemic groups and one control group. The ischemic groups were divided according to the time when samples were collected for evaluation (I1H – ischemic group, samples were collected one hour after ischemia (n=6); I1D – ischemic group, samples were collected one day after ischemia (n=6); I7D – ischemic group, samples were collected seven days after ischemia (n=6)). A sham laparotomy with no vessel ligation was done in the control group (n=6), and samples were collected on the seventh day.

Surgical procedure:

All procedures were done under aseptic conditions and general anesthesia (Zoletil, Tiletamine, Zolazepam, Virbac, Westlake, Texas), 20 mg/kg i.m., and 5 mg/kg of mabron (Tramadoliumchlorid, Medochemie, Limassol, Cyprus). Each animal was placed in the supine position, and the operation field was prepared using standard methods (shaved abdomen cleaned with Kodan antiseptic tincture, Schülke, Norderstedt, Germany). An upper median laparotomy was performed, and after mobilization of the stomach, ligations of the left gastric artery (LGA) and short gastric vessels (SGV) were done using absorbable sutures (3-0 Polysorb, Covidien). In the control group, rats underwent only mobilization of the stomach without LGA occlusion (shown in Fig. 1). The laparotomy was closed using continuous sutures in anatomical layers (3-0 Polysorb suture Covidien) for the musculofascial layer and peritoneum, Chiraflon 5/0 running suture (Chirmax, Prague, Czech Republic) for skin closure. The animals were allowed free access to water after the operation, and standard laboratory diet was given on the second day after surgery. Postoperative analgesia was performed on the second day with mabron i.m. (2.5 mg/kg Tramadoliumchlorid, Medochemie, Limassol, Cyprus). The animals were euthanized according to the group by an overdose of anesthetics at one hour, one day, and seven days after surgery. Macroscopic photographs of the gastric mucosa were taken, and samples were stained for histological evaluation.

Macroscopic assay:

The distal part of the abdominal esophagus, gastroesophageal junction (GE), and the whole stomach were removed, and dissection was made on the greater curvature. Samples were cleaned using a sterile solution. Photographs of the mucosa were taken using an Olympus E330 digital camera equipped with a digital ED 50 mm f 2.0 macro-objective and a ring set flash SRF-11 (Olympus, Tokyo, Japan). Macroscopic signs of mucosal ischemia and necrosis were evaluated.

Histological Analysis:

A tissue sample was processed for light microscopy using routine techniques including fixation in 4% buffered formaldehyde, dehydration, embedding, cutting, and staining with hematoxylin-eosin. Two sections were prepared from the proximal stomach and GE junction of each sample. An experienced pathologist, who was blinded to the operative protocol, evaluated a section of the cardiac region and esophagogastric junction under a light microscope and performed microscopic analysis. Additionally, thirty microscopic fields per condition were randomly selected, and the number of immunocompetent cells was counted.

Statistical Analysis:

Descriptive statistics were expressed as mean ±SD (standard deviation). The data obtained from cell counting were compared using one-way ANOVA followed by the Student-Newman-Keuls test. A p-value of less than 0.05 was considered statistically significant.

RESULTS

Macroscopic Analysis:

The results of the macroscopic assay are shown in Fig. 2. Limited hyperemia was observed only in group I1D. No erosions, necrotic areas, or significant changes were observed in any of the experimental groups.

Microscopic Analysis:

I1H ischemic group: The dominant finding in this group was passive hyperemia, ischemic denudation with mucosal erosion (Fig. 3), and capillary leukostasis. Eosinophils (45.37%) and lymphocytes (23.53%) were the dominant inflammatory cells. The total number of eosinophils was significantly higher (p<0.05) in groups I1H and I7D compared to group I1D. However, the total lymphocyte count was significantly lower in the I1H group (p<0.05) compared to the I1D and I7D groups (Graph 1).

I1D ischemic group: One day after ischemia with reparative changes and fibroblast proliferation, the fundal mucosa was reduced to one-third of the height compared to the nonischemic group. The inflammatory response in the epithelial-subepithelial junction was reduced, and vascular proliferation, mitotic activity, and partial regeneration of glands were increased at this time after vessel ligation (Fig. 4). The increase in the number of polymorphonuclear cells (23.15%) and histiocytes (16.59%) indicated a non-specific immune response and stimulation of reparation. The total count of polymorphonuclear cells was significantly higher (p<0.05) at this time of healing (group I1D) compared to group I7D, where the samples were collected seven days after ischemia. Furthermore, the number of histiocytes was significantly higher (p<0.05) in the I1D group compared to the I1H group (Graph 1). The percentage of eosinophils (31.56%) was reduced.

I7D ischemic group: Nearly total architectural regeneration of the fundal mucosa and capillary hyperplasia were documented at this time. Inflammatory response was reduced at this time after ischemization. The total count of polymorphonuclear cells (16.65%) was decreased (Fig. 5).

DISCUSSION

As far as we know, this is the first study to summarize both histological changes and inflammatory cell counts following gastric ischemia. Various animal models have been used to assess the changes that occur in the gastric remnant and subsequent esophagogastric anastomosis after partial gastric devascularization [7,9,13]. Measurements in animals have shown that gastric perfusion drops substantially immediately after vessel ligation, with continuous increase over time. Additional benefit in the conduit blood flow was shown when the short gastric arteries were ligated in addition to the left gastric artery [17]. In our study, we used laparotomy with stomach mobilization and ligation of the left gastric artery and short gastric vessels. Monitoring of histological changes, especially neovascularization and capillary hyperplasia, are essential for assessment of adequate tissue perfusion after gastric conduit creation. On the other hand, presence of necrosis and apoptosis in histological view are believed to be a negative factor for future anastomosis healing. Here we present nearly total histological recovery with regeneration of fundal mucosa, capillary hyperplasia, and absence of necrotic tissue 7 days after ischemia with a total count of polymorphonuclear cells (PMNL) significantly lower (p<0.05) compared to the 1 day after ischemia group.

Neutrophils/polymorphonuclear cells are the predominant leukocyte population in human blood and among the first cells recruited to an inflammatory site. During sterile tissue injuries, neutrophils are also essential to participate in the clearance of cellular debris, returning tissue to homeostasis [18,19]. In our study, acute inflammatory response was reduced in the group I7D, presented by a significantly lower total count of PMNL (p<0.05) at this time of healing (group I7D) compared to the group I1D. From this point of view the decreased number of PMNL 7days after GIC could be interpreted as a positive prognostic factor for future anastomosis healing.

Reavis et al. [20] ligated left, distal right, and short gastric vascular pedicles in an opossum model, resulting in an immediate decrease of gastric fundus blood flow by 73%. The distance of full thickness muscularis propria atrophy was 0.53 mm in the immediate group compared to 0.10 mm in the delay group (4 weeks after ischemia). In a study with Sprague-Dawley rats, Lamas et al. [7] verified higher rates of apoptosis and necrosis after 24 hours of partial gastric ischemia. Similar findings are reported in our present study with ischemic denudation and mucosal erosions as the dominant finding one hour after ischemia. Additionally, eosinophils (45.37%) were the major inflammatory cells at this time of healing, and the total count of lymphocytes was significantly lower compared to other groups. Histological changes in a mean of mucosal erosions and ischemic denudation presented one hour after ischemia belong to negative factors, increasing the complication rate of anastomosis healing.

Mingol-Navarro et al. [4] described the anatomic and physiopathologic background of ischemic preconditioning in their complex review. Kechagias et al. [21] analyzed all currently available information dealing with ischemic conditioning of the stomach in the prevention of esophagogastric anastomotic leakage after esophagectomy. Benefits of ischemic preconditioning were also described by several experimental and clinical studies [13−17]. Significantly increased neovascularity with decreased inflammation was observed by Perry et al. 30 days after gastric ischemic preconditioning in their animal study [13]. Therefore, we report an easily reproducible animal model of GIC, following histological changes in different time intervals from ischemia and measuring total counts of immunocompetent cells. Existence of this model is important for future esophagogastric anastomosis studies and to help optimize the time interval between GIC and anastomosis creation. On the one hand presence of neovascularization and capillary hyperplasia with architectonic regeneration of fundal mucosa 7 days after ischemia promote healing, on the other hand verifying mucosal alterations and denudation one hour after ischemia impairs the process of future anastomosis healing.