Obesity Surgery

, Volume 21, Issue 9, pp 1444–1450

Reversible Gastric Restriction Implant: Safety and Efficacy in a Canine Model


  • Xiaomei Guo
    • Department of Biomedical EngineeringIndiana University–Purdue University Indianapolis
  • Hai Zheng
    • Department of Biomedical EngineeringIndiana University–Purdue University Indianapolis
  • Samer G. Mattar
    • Department of SurgeryIUPUI
  • Xiao Lu
    • Department of Biomedical EngineeringIndiana University–Purdue University Indianapolis
  • George Sandusky
    • Department of PathologyIUPUI
  • Jose A. Navia
    • Department of SurgeryAustral University
    • Department of Biomedical EngineeringIndiana University–Purdue University Indianapolis
    • Department of SurgeryIUPUI
    • Cellular and Integrative PhysiologyIUPUI
Animal Research

DOI: 10.1007/s11695-010-0299-4

Cite this article as:
Guo, X., Zheng, H., Mattar, S.G. et al. OBES SURG (2011) 21: 1444. doi:10.1007/s11695-010-0299-4



Gastric restrictive procedures are considered effective weight loss treatment for severe obesity. The aim of the study was to evaluate the efficacy and safety of a reversible implant that renders a partial restriction of stomach in a canine model.


The device was comprised of two longitudinal parallel non-compressive plates with two C-rings to create a small gastric pouch that opposed gastric distension. Three groups of non-obese mongrel dogs were included: group I (n = 6) underwent surgical implant for 6 weeks, group II (n = 6) underwent surgical implant for 6 weeks, followed by surgical removal of the implant and additional monitoring for 5 weeks, and group III (n = 5) served as sham-operated controls for groups I and II. Food intake and body weight were monitored, and the stomachs were examined histologically postmortem.


The average food intake was significantly decreased by 38.2% in group I as compared to group III throughout the 6 weeks of surgical implant (P < 0.05). The implanted dogs showed a progressive weight loss as compared to sham, which reached 21% by the end of 6 weeks. In group II, after 5 weeks of implant removal, the body weights recovered to approximately 96% of baseline. Histological evidence of the implant site at the gastric walls revealed no significant structural changes, tissue ischemia, hemorrhage, or necrosis.


Our results validate the feasibility of a reversible gastric restriction implant in a non-obese canine model, with the potential for achieving significant weight loss within 6 weeks and with no injury to the gastric wall.


ObesityRestrictive procedureImplantFood intakeBody weight


Obesity is considered a major worldwide health problem [1, 2]. The incidence of obesity is steadily growing, and it has been projected that 40% of the US population will be obese by the year 2025 [3]. Obesity is not only a physiological metabolic dysfunction in humans with environmental, genetic, and hormonal causes [4] but is also a chronic disease that is characteristically refractory to medical attempts at weight loss [5, 6].

Over the last half century, numerous bariatric operations have been proposed and performed in an attempt to achieve safe and effective weight loss [7, 8]. Although there has been major progress, the field remains one of flux and evolution, characterized by the regular emergence of new techniques and derivations of more established operations. Currently, there is no consensus on the superior operation or the ideal patient. However, there is a growing body of evidence that the best surgical outcomes occur in those patients whose risk status has been optimized through preoperative weight loss [9, 10].

Gastric restrictive procedures, such as vertical sleeve gastrectomy (VSG) or vertical banded gastroplasty (VBG) are effective weight loss operations [3, 1113]. The VSG is created by the longitudinal resection of about 85% of the stomach to leave a long banana-like pouch. Initial reports demonstrated that the VSG was a viable option in bridge patients, who were instructed to return for more definitive operations such as gastric bypass or duodenal switch [14]. However, it rapidly became evident that VSG resulted in dramatic and durable weight loss and has eventually become widely accepted as a definitive operation on its own right [3, 13, 15]. The VBG uses a combination of staples and a band to create a small pouch in the upper stomach, designed to restrict food intake and prolong satiety. Several studies have shown that the VBG provides good weight loss results, with an acceptable safety profile in a short-term follow-up [12, 16, 17]. Despite the efficacy of these purely restrictive procedures, there have been several limitations including the risk of staple line leaks and late complications, such as weight regain and gastric sleeve or pouch dilatation [1821]. In addition, the removal of stomach in the VSG is not reversible, a not insignificant concern to patients.

To mimic the benefits of the gastric restrictive procedures, while avoiding some of their shortcomings, we developed a device, referred to as the reversible gastric restriction (RGR) implant. The device uses two longitudinal parallel non-compressive plates to form the line of restriction instead of the typical surgical staples and/or gastrectomy used in the restrictive procedures. Two C-shape rings attached to the two plates provide the mechanical force that opposes distension of the stomach. Thus, it can restrict gastric motility and capacity, but preserves the normal anatomical configuration of the stomach. The main objective of this study was to evaluate the feasibility (safety and efficiency) of the RGR implant for weight loss in a non-obese canine model.



Seventeen healthy mongrel dogs (22–29 kg) of either gender were used in this study. The dogs were randomly divided into three groups. Group I (n = 6) was subjected to gastric clamping for a period of 6 weeks and was sacrificed at the end of 6 weeks. Group II (n = 6) was subjected to the same treatment as group I but was allowed to recover for another 5 weeks after implant removal at the end of 6 weeks. Group III (n = 5) served as sham-operated controls for group I and II. The sham group underwent an identical surgical procedure but without actual implantation and were also sacrificed at the end of 11 weeks. All animal experiments were performed in accordance with national and local ethical guidelines, including the Institute of Laboratory Animal Research guidelines, Public Health Service Policy, the Animal Welfare Act, and were approved by Institutional Animal Care and Use Committee at University of Indiana–Purdue University, Indianapolis.

Surgical Procedure

All the dogs were allowed an 8 h daily access to unlimited measured quantity of dry commercial canine food (LabDiet 5006, PMI Nutrition International, St. Louis, MO) following 7 days of acclimatization. The access to water was ad libitum. The leftover food was measured to calculate the amount consumed daily. The study was performed only when the food intake was stable and reproducible, prior to implantation. Each dog was weighed immediately before the surgical procedure.

After a 12-h overnight fast, the dog underwent general anesthesia induced with propofol (4–8 mg/kg) and maintained with isoflurane (2–3%). A laparotomy (about 5 in.) was performed on the ventral midline. The RGR implant (weight 15 g) consisted of two longitudinal parallel plates (size, 12 × 1.2 × 0.3 cm), held together by two C-shape rings (approximately 2.5 cm in length, 2.5 cm in width, or 6.25 cm2 in area) as shown in Fig. 1. The material of the plates is high-density polyethylene, and the C-shape rings are stainless steel. The stomach was lifted out gently. The two plates with C-rings were then longitudinally placed on the outside of front and back wall of stomach along lesser curvature to form a narrow longitudinal gastric pouch, whose diameter was 2.5 cm (distance from plates to lesser curvature), and volume was about 15–20 ml (Fig. 1). The space between the two plates (1–2 mm larger than the thickness of the two walls of the stomach) imposed no compression on the gastric wall. The proximal ends of the anterior and posterior plates were loosely tightened together by a 2–0 silk suture at the superior border of the gastric fundus, adjacent to the angle of His. Two sutures of the anterior plate were used to fix the device on the serosa layer of gastric walls. Two additional sutures were then used to restrain the lateral portion of the stomach to the C-ring. These serosal sutures served to prevent migration of the device along the surface of the stomach wall. For the sham-operated group (group III), only the two plates without C-shape rings were sutured on the surface of the stomach at a similar position as the experimental group.
Fig. 1

A schematic of RGR implant mounted on dog stomach. The two plates (one anterior and one posterior) “sandwiched” the gastric wall to create a stomach pouch. The C-rings between the two plates served to maintain a distance slightly larger than the thickness of both wall to prevent compression of the stomach. The pouch diameter created by implant was around 2.5 mm (distance from plates to lesser curvature)

Terminal Studies

The animals (group I) were terminated by administration of an overdose of pentobarbital at the end of 6 weeks. The stomachs were excised immediately and fixed for later histological evaluation.

Implant Removal

Anesthesia, sterility, and preparation of the animals (group II and III) were the same as described above. A laparotomy (about 5 in.) was performed on the ventral midline. The stomach was lifted out gently and the two C-rings were carefully exposed and cut with a wire cutter. The two plates held together by C-rings were immediately separated. The stomach was then placed back into the abdomen, and the two plates were left on the outside of stomach wall. For the sham-operated group (group III), the stomach was only lifted out gently and then placed back into the abdomen. The groups II and III were sacrificed at 11 weeks after 5 weeks of implant removal. The surgical procedure was identical to that described for group I.

Food Intake and Weight Change

After the first 12 h of NPO following surgery, the animals were fed canned food for 48 h, and then allowed 8 h of daily access to standard food chow and water. Food intake was measured daily and summed over each week. The body weight was measured weekly after surgery. The percent food intake reduction rate was defined as: (weekly food intake − initial food intake)/initial food intake × 100. The percent weight loss rate was defined as: (weekly weight − initial weight)/initial weight × 100.

Histological Evaluation

Necropsies were performed on all animals to visually examine the position of implants and their potential impact on the gastric wall. The gastric tissue immediately below the implant as well as the tissue adjacent to the implant for both experimental and sham groups was harvested and fixed with 10% formalin in phosphate buffer for at least 10 h. The sample was then embedded in paraffin and cross-sectioned at 3 μm thickness. The tissue sample were stained with hematoxylin and eosin (HE) and examined under light microscopy.

Statistical Analysis

Data were shown as mean ± SD, and significance of the differences between two groups was evaluated by one-way ANOVA or T test. The results were considered statistically significant when p < 0.05 (two-tailed).


All animals subjected to surgical implant and implant removal survived until the termination date. The dogs were playful and responsive with no sign of distress, as determined by the monitoring veterinary staff.

Changes in Food Intake

There were no significant differences between group I (surgical implant) and group II (surgical implant plus implant removal), so the data on both food intake and body weight of group I and II has been combined. Figure 2a shows the reduction of food intake after surgical implant and implant removal in experimental and sham group. For the experimental group, the food intake in the first postoperative week was sharply reduced by 64.3%, compared to preoperative baseline. Subsequently, this significant decrease of food intake was sustained at 38.2% in weeks 4 to 6. For the sham group, the food intake was initially reduced by 21%, compared to preoperative baseline, in the first postoperative week, and then recovered in the second week so that the food intake reduction rate was significantly higher in the experimental group as compared to the sham throughout the 6 weeks of surgical implant (P < 0.01). In the seventh week after implant removal, the food intake in experimental group was clearly increased and recovered to the pre-implant level at week 8. There was no significant difference in food intake between experimental and sham group after implant removal.
Fig. 2

Food intake reduction rate (a) and weight loss rate (b) in experimental and sham-operated dogs after implant surgery and implant removal. The food intake reduction rate and weight loss rate were significantly higher in the experimental as compared to the sham group throughout the 6 weeks of surgical implant (P < 0.01). After implant removal at week 6 (arrow designated), the food intake in experimental group was recovered to the pre-implant level at week 8, and body weight was recovered to about 96% of their pre-implant levels at week 11

Changes in Body Weight

Baseline mean body weight was comparable in experimental (group I and II) and sham group (group III) before treatment. The mean body weight was 26.6 ± 3.0 kg in experimental and 26.5 ± 1.8 kg in sham-operated, respectively (P = 0.86). Figure 2b shows the body weight loss after surgical implant and implant removal in experimental and sham group. In line with the reduced food intake, the weight loss rate in experimental group was significantly higher than sham-operated group during the 6 weeks of implant (P < 0.01). A progressive loss in weight was observed in the experimental group, reaching 21% by the end of week 6. For the sham group, the body weight was decreased by 3.2% in the first postoperative week, and then returned to baseline in the second week. By the end of 6 weeks, the mean body weight was 20.7 ± 2.5 kg in experimental and 27.1 ± 2.2 kg in sham-operated group. After implant removal, a progressive increase in weight was observed in experimental group and recovered to about 96% of their pre-implant levels in 5 weeks. The change in weight observed in sham group was not statistically significant after implant removal relative to the initial level.

Postoperative Complications

After surgical implant, vomiting was observed once or twice per week after meals, for the first 2 weeks in four dogs in the experimental group, after which time, the vomiting resolved spontaneously. No other esophageal reflux symptoms were observed throughout the 11 weeks of surgical implant and implant removal in all groups of animals. In two experimental dogs, the pouch was dilated significantly with a threefold larger size than normal. These two dogs lost significant body weight (8–10%) in the first 3 weeks after surgery and then regained some of their weight (3–5%) by the end of 6 weeks. Additional two dogs developed mild proximal pouch dilation of about 1–1.5 times larger than normal. A significant and sustained weight loss (15–17%) was still observed in these two dogs following the 6 weeks of surgical implant.

Histological Findings

At the time of termination, the device and stomach of experimental and sham groups were carefully examined. We found a typical fibrotic capsule (~2 mm in thickness) formed around the plates, which could be easily excised to allow removal of the plates (Fig. 3a). In all groups of animals, no gastric wall erosion at the plate and suturing sites, no gross changes in the submucosa or mucosa, and no tissue ischemia or necrosis were seen (Fig. 3b). The histopathological examination of gastric wall confirmed that no lesion was seen in the mucosa, submucosa, and muscularis in sham or experimental groups. The implant caused mild or moderate focal lymphoid follicles at basement membrane. Minimal focal chronic inflammation, angiogenesis, or fibroplasia in adventitia was observed in experimental group as shown in Fig. 4a. Some of the sham animals and all the experimental animals exhibited localized, focal granulation tissue on the serosa surface of the stomach (Fig. 4). There were no major histological difference between the sham and experimental groups.
Fig. 3

a The gastric wall showing normal tissue with no inflammation or necrosis after an incision of the fibrotic capsule around the plates was made to remove the implant. b The gastric mucosa showing normal tissue with no inflammation, adhesion, or necrosis

Fig. 4

Histological images of gastric wall in response to implant in experimental (a) and sham-operated group (b). Minimal focal lymphoid follicles at basement membrane was seen in sham and experimental group. The experimental animal showed localized, focal granulation tissue on the serosa surface of the stomach. No other histological abnormities were seen in the mucosa, submucosa, and the muscularis (hematoxylin and eosin staining; original magnification, ×100)


The surgical treatment for morbid obesity must be safe and effective. Therefore, the ideal operation must have minimal risks of short- and long-term morbidity and mortality, produce significant and durable weight loss, and maintain normal gastrointestinal function. Gastric restrictive operations have been used widely for the surgical treatment of morbid obesity in humans and are deemed safe and have relatively short operation time [3, 1114, 21]. The anatomical result of the operation is the creation of a narrow gastric tube or channel. Although the exact mechanism of action of restrictive operations is unknown, there is agreement that postoperative patients experience both anorexia and food restriction. In the present study, a non-banded gastric restriction (RGR) implant was designed to mimic the effects of the restrictive procedures by creating a partial narrow gastric tube that reduces gastric capacity and induces an early sensation of satiety that leads to diminished food intake. We chose a canine model as our experimental subject, since there are physiological and anatomical similarities between the foreguts of dogs and humans.

After the device was implanted, the animals achieved a significant decrease in food intake and body weight as compared to the sham-operated group. In general, food intake and body weights decreased in the first postoperative week in all the animals, probably due to surgical trauma. For the duration of the study, the average food intake was reduced by 38.2% per week in experimental animals, when compared to sham-operated animals (Fig. 2a). The body weight was decreased by 21% over the 6-week period (3.4% per week, Fig. 2b).

Unlike the VSG, the present device can be easily reversed by removal. After implant removal, the food intake in the experimental group increased to the pre-implant level in the first postoperative week at week 8. In accordance with the increase of food intake, a progressive weight regain was observed, and the weight recovered to approximately 96% of their pre-implant levels 5 weeks after implant removal (Fig. 2b). This is strong evidence for the cause and effect relation between device and weight loss and it suggests satiety, given the 8 h feeding regimen. Recently, our group reported that a similar device significantly decreases food intake and body weight gain in Zucker fatty and healthy Wistar rats [22]. Our current results suggest that the RGR implant is feasible with an effective weight loss, and it can be reversed if desired.

In the course of designing this experiment, we were conscious of two common concerns: the possibility of plate migration and the potential for excessive compression on tissue that may render the possibility for gastric fistulization. In the present study, no leakage and/or tissue laceration or device migration were found in any of the experimental animals. The premise underlying the novel device is that the implant does not compress or interact with the gastric tissue (hence no possible necrosis or fibrosis). The plates mainly prevent distension in the presence of food bolus at which time the force is spread out over the surface area of plates rather than concentrated at the staple line. Hence, this device is largely not “active” in the absence of bolus and accordingly does not compromise the blood supply or alter the integrity of the gastric wall. This was confirmed by histological examination, which showed no evidence of hemorrhage, necrosis, or thrombosis in the gastric wall. No histopathologic abnormality was observed in gastric submucosa or mucosa below and adjacent to the implant (Figs. 3 and 4). This study confirmed that the RGR implant maintains the histological architecture and anatomy of stomach in the canine model.

In addition, the RGR implant is truly light weight (only 15 g) as compared to the average weight of the dog stomach (310 g). We found that four sutures were sufficient to fix the device to the gastric serosal layer. We did not observe device migration ventrally or caudally along the surface of the stomach or gastric wall erosion induced by the sutures probably due to the following: (1) the anchoring stitches were made only through the serosa and only partially in the muscular layer; (2) the C-rings did not allow the plates to move laterally because they impeded the over-distension of the pouch; and (3) the formation of a fibrotic capsule around the device. The zero incidence of device migration in this study may be related to the short duration of follow-up. Whether migration of RGR implant occurs over a longer period remains to be determined in future studies.

One common complication of vertical restrictive operations is gastroesophageal reflux with regurgitation of gastric contents, probably caused when the narrow tube is excessively stretched by food particles [23, 24]. In the present study, four dogs experienced occasional vomiting (maximum twice per week) in the first 2 weeks following surgical implant. Although all the dogs were allowed an 8-h access to unlimited food, we observed that this vomiting only occurred in the first 2 or 3 h after feeding, when these dogs had consumed 70–80% of their daily food intake. However, this mild vomiting recovered spontaneously without any intervention. Beyond the first 2 weeks of implant, no vomiting was observed. It is likely that these dogs “learned” how to moderate the food intake to avoid vomiting. Three of the four dogs that had vomiting also developed gastric tube dilation. It is possible that consuming a large amount of food in a relatively short time may induce the enlargement of gastric tube. In restrictive procedures, patients are strongly encouraged to follow dietary guidelines to reduce the incidence of vomiting [25, 26]. In animals, the post implant feeding pattern is probably an important factor that influences vomiting, esophageal reflux, or tube dilation.

The exact mechanism(s) that lead to sustained weight loss after bariatric operations remain to be elucidated. Gut hormones have been implicated to play an important role in both weight loss and diabetes improvement after weight loss surgery [27, 28]. Ghrelin, recently described as a hunger-regulating peptide hormone mainly produced in the fundus of the stomach [29], is reported to be significantly increased in diet-induced weight loss [27]. The VSG eliminates the fundus, a portion of the stomach that contributes to ghrelin, which may be important for reduction in appetite and weight loss [27, 30]. However, in obese patients subjected to VBG or adjustable gastric banding, fasting ghrelin levels have been shown to increase or remain at baseline level [27, 31, 32]. Since the RGR implant maintains the integrity of the stomach, it is likely that the ghrelin level remains stable or has a compensatory rise following weight loss. This assumption needs to be verified in a future study.

The present study mainly focused on the feasibility (safety and efficacy) of the surgical implant and implant removal of the RGR. Our results show that the RGR implant can achieve effective weight loss with no trauma to the stomach. The present findings establish the merit of this approach and underscore the potential clinical translation of this device in the treatment of obesity. The translation of RGR implant to the clinic will require a laparoscopic device and delivery. The sponsor company (GRest, Inc) is currently developing a laparoscopic prototype for future animal testing and later for clinical trial.


This research was funded by GRest, Inc. and NIH SBIR Phase I grant R43DK083889. We wish to thank Mr. Mitchell L Keel and Mark Svendsen for excellent technical assistance.


GSK is founder of GRest, Inc. There are no other conflicts of interest.

Copyright information

© Springer Science + Business Media, LLC 2010