Evaluation of the SafeHeal Colovac+ Anastomosis Protection Device: A Preclinical Study

Introduction

The protective ostomy is the current standard of care to protect a low colorectal anastomosis from leakage, but exposes patients to bothersome complications including peristomal skin breakdown, dehydration, prolapse, and parastomal hernia in addition to a significant psychological burden.^1-5 The ostomy is typically reversed a few months later to allow the accomplishment of adjuvant therapy. ⁶ The reversal also carries a risk of complications such as postoperative wound infection, leakage of the ileo-ileal anastomosis, and incisional hernia at the ostomy site.^7,8 Moreover, approximately 20% of these so-called “temporary” ostomies are never reversed. ⁹ The lifelong maintenance and cost of a permanent ostomy have a significant impact on quality of life. ¹⁰

Therefore, these patients would benefit from a device that temporarily protects the anastomosis during the healing period, avoiding the need for an ostomy and the associated drawbacks. ⁶ The SafeHeal Colovac Anastomosis Protection Device (SafeHeal SAS, Paris, France) is designed to shield the anastomosis from fecal content, preventing the clinical outcomes of anastomotic leakage (AL). It consists of a flexible bypass sheath attached to a vacuum stent anchoring the device in the colon proximal to the anastomosis. Preliminary studies with the initial Colovac design have already demonstrated the device’s potential value, but migration from the anchoring site was shown to be an important issue that remains to be resolved.^11,12 In response to these findings, the Colovac design was adjusted in order to reduce the migration risk during the implantation period.

The objective of the present study is to evaluate the preliminary efficacy, safety, and technical feasibility of the upgraded design (Colovac+) in a porcine model.

Methods

Study Design

This prospective preclinical study was conducted in a living porcine model. All procedures were performed by an experienced colorectal surgeon (CK) at the Institut Mutualiste Montsouris Recherche (IMMR) in Paris, France. The study was performed in accordance with FDA Good Laboratory Practice (GLP) regulations, ISO Standard 10993 (Biological evaluation of medical devices), EU Directive 2010/63/EU for animal experiments, and the internal IMMR Quality System.

The preliminary efficacy, safety, and technical feasibility of the Colovac+ were evaluated through the following endpoints: prevention of contact between the anastomosis and fecal content, device migration, feasibility of the implantation and retrieval procedure, collateral damage to the colonic wall, colon healing after device retrieval, and systemic toxicity related to the device. The methods used to assess the study endpoints are described in Table 1.

OPEN IN VIEWER

Table 1.

Methods used to assess the study endpoints.

Study Endpoints	Method
Efficacy endpoints
Prevention of contact between the anastomosis and fecal content	Daily visual check of the absence of feces on the external surface of the sheath Endoscopic evaluation of the absence of feces between the sheath and colonic wall before the retrieval procedure on day 10
Device migration	Measurement of the distance between the anal margin and distal (external) end of the sheath after implantation on day 0 and daily during follow-up
Safety endpoints
Collateral damage to colonic wall	Endoscopic and histopathological evaluation of the anchoring site after device retrieval on day 10
Systemic toxicity	Clinical examinations, measurement of body weight and temperature, and laboratory investigations during follow-up Histopathological evaluation of internal organs (heart, lungs, liver, kidneys, and lymph nodes at the implantation site, adrenal glands, spleen, and brain)
Colon healing after device retrieval	Endoscopic and histopathological evaluation of the anchoring site 28 days after device retrieval
Technical feasibility endpoints
Implantation procedure	Rating on a scale of 1 (very easy) to 5 (impossible) of the technical ease of the procedure
Retrieval procedure	Rating on a scale of 1 (very easy) to 5 (impossible) of the technical ease of the procedure

The SafeHeal Colovac+ Anastomosis Protection Device

The Colovac+ is a temporary, single-use, disposable device composed of 2 elements: an anchor made of a covered double stent delimiting a vacuum chamber connected to 2 vacuum tubes; and a flexible cylindrical sheath attached to the anchor, covering the anastomosis with appropriate length so that it protrudes about 5 cm outside the pig’s anus (Figure 1). The negative pressure in the vacuum chamber is generated by Redon vials connected to the tubes and sucks the colonic wall toward the mesh of the stent in order to anchor the device and avoid the passage of feces between the stent and colonic wall (Figure 2).

OPEN IN VIEWER

Figure 1.

The SafeHeal Colovac+ Anastomosis Protection Device. (A) Real life picture depicting the vacuum stent (1), flexible sheath (2) and vacuum tubes (3). (B) Detailed structure.

OPEN IN VIEWER

Figure 2.

Colovac+ anchoring.

Postoperative Course

The animals were divided into 2 follow-up groups through randomization (Figure 3). Six pigs were to be sacrificed on day 10 immediately after device retrieval (group A). The other 6 pigs were to be sacrificed on day 38 (group B). Analgesics were administered to all animals, and a uniform feeding program was followed.

OPEN IN VIEWER

Figure 3.

Postoperative course.

The following data were registered daily: symptoms and signs of AL, bowel movements, the appearance of feces on the external surface of the sheath, signs of vacuum loss, and signs of device migration. The vacuum was monitored by checking the Redon vials and patency of the tubes. Device migration was assessed by measuring the distance between the anal margin and distal (external) end of the sheath using printed depth marks, and was defined as a change in length from the anal margin to the distal end of the sheath of greater value than the maximum length measured on day 0. A significant migration was defined as a migration leading to an unprotected anastomosis. This includes a migration where the stent is located at the level of the anastomosis.

On day 10, the Colovac+ was retrieved endoscopically under general anesthesia. Before removal, a colonoscopy was performed to assess the absence of feces between the sheath and colonic wall. Next, the vacuum was terminated and the device was removed endoscopically through the anus by pulling on the retrieval loops located at the proximal and distal ends of the stent. After device retrieval, a second colonoscopy was performed to evaluate the anchoring site and anastomosis. The animals from group B underwent a control colonoscopy on day 24 and 38 to evaluate healing of the anchoring site and anastomosis.

The animals were euthanized with an intravenous injection of sodium pentobarbital (50 mg/kg). After sacrifice, the left colon and internal organs were harvested and sent for histopathological (HP) examination. Samples from the anchoring site and anastomosis were obtained, and subsequently fixed with formaldehyde 4%, embedded in paraffin, cut into 5 μm sections, and stained with hematoxylin, eosin, and saffron (HE&S).

Results

Twelve pigs underwent a partial colectomy with creation of a colorectal anastomosis at a mean distance of 108 (80–140) mm from the anal sphincter. The Colovac+ was implanted proximal to the anastomosis, and the mean distance between the anastomosis and distal end of the stent was 64 (30–120) mm. No adverse events were observed during the implantation procedure, except for slight abdominal bleeding due to transsection of mesenteric vessels in one animal.

Eleven pigs completed the study according to their assigned group. One pig, assigned to group B, demonstrated a diminished health status (lethargy and distended abdomen) on day 2 and underwent an exploratory laparotomy. Despite the surgery and subsequent treatment, the animal died on day 5. The cause of death was determined as uroperitoneum following bladder damage during the surgical procedure (unrelated to the device).

Safety Endpoints

The postretrieval colonoscopy revealed stent imprints and mucosal inflammation at the anchoring site in most animals. In all 10 animals that underwent the retrieval procedure, there were no visible lesions at the anchoring site and the anastomosis was intact. HP evaluation of the anchoring site was performed in the 6 animals from group A to assess collateral damage to the colonic wall. The macroscopic examination corresponded with microscopic observations, revealing multifocal ulcerations in 4 animals (Figure 4). The lesions were more pronounced at the proximal anchoring site as compared to the middle and distal level: the incidence of ulcerations was slightly higher, and the ulcerations were deeper (extending beyond the submucosa, sometimes transmural) and accompanied by marked fibrosis of the colonic wall. Other findings included stent imprints, congestion, mucosal atrophy, and mild inflammation, but these were minor lesions that can be expected for this type of procedure. In all 6 animals from group A, HP evaluation of the anastomosis confirmed anastomotic integrity and only showed mild signs of congestion and inflammation.

OPEN IN VIEWER

Figure 4.

Histopathological examination of the anchoring site. (A) Macroscopic examination showing a stent imprint in the colonic mucosa. (B) Microscopic examination revealing multifocal ulcerations extending to the submucosa. The ulcerations correspond with the points of contact between the colonic mucosa and the mesh of the stent. Mu, mucosa; Sm, submucosa; Ms, muscularis; Se, serosa.

The control colonoscopies, performed in the 5 group B animals on days 24 and 38, revealed adequate healing of the colonic mucosa both at the anchoring site and anastomosis. On HP evaluation of the animals from group B, no ulcerations were noted at the anchoring site (including the proximal level), and all anastomoses were intact. Only minor findings were observed, such as residual stent imprints and mild mucosal inflammation and atrophy at the points of contact between the stent and colonic wall.

In all 11 animals that completed follow-up, HP evaluation of the internal organs revealed no abnormalities indicating systemic toxicity related to the Colovac+. In the 5 animals from group B, body weight gradually increased throughout the study and body temperature remained within the clinically acceptable range.

Discussion

The concept of intraluminal bypass devices as a protective mechanism for colorectal AL has been well described in the literature, following the introduction of the Coloshield device by Ravo and Ger in the 1980s.^11,13 These devices are intended to shield the anastomosis from fecal content and prevent the clinical outcomes of AL. ⁶ The potential benefit is the avoidance of the need for a protective ostomy and the associated drawbacks. ⁶ Previous studies with intraluminal bypass devices have demonstrated proof of concept in terms of efficacy, safety, and technical feasibility. ⁶ However, they have not been widely implemented in clinical practice mainly due to technical reasons, as described earlier by our research group. ⁶

First, early designs required a complex open surgical approach for implantation and were incompatible with modern laparoscopic techniques. ⁶ The Coloshield was anchored proximal to the anastomosis by manually suturing the device to the everted proximal bowel loop. ¹³ Next, the posterior part of the colorectal anastomosis was performed. A rectal tube was then passed transanally, tied to the Coloshield, and subsequently drawn back outside in order to cover the anastomosis. Finally, the anterior part of the anastomosis was completed. In 1997, Chen et al. described the Valtrac-secured Intracolonic Bypass, which consisted of a latex condom attached to a biofragmentable anastomosis ring (BAR). ¹⁴ The BAR was manually inserted into the proximal bowel loop and anchored with a suture around the serosal surface of the colon. After creating the anastomosis with a transanal stapler, the condom is pulled down alongside the anastomosis and outside the anus.

Second, these devices may damage the colonic wall or anastomosis. Autopsy findings from an experimental study with the Fecal Diverting Device revealed colonic wall erosions at the fixation area in 8 of the 30 animals. ¹⁵ In the subsequent preliminary clinical study, erosions at the fixation area were noted in 3 cases, and of these patients developed a pericolic abscess. ¹⁶ In their international randomized trial, Bakker et al. found a higher AL rate in patients with the C-seal device compared to the control group (10.4% vs 5.0%). ¹⁷ The authors advocated that traction on the anastomosis may have occurred through the sheath as the device is anchored directly to the anastomotic staple line.

Finally, anchoring anastomosis protection devices represents the main, so far unresolved issue as the digestive tract is not designed to retain foreign objects. ¹¹ Thus, devices can migrate from their original position and may even be expelled prematurely, leading to an unprotected anastomosis. ¹¹ Tsereteli et al. used a plastic-covered stent to cover a colorectal anastomosis in a swine model, but the stents were spontaneously expelled between postoperative days 6 and 9 in all cases, despite being sutured to the colonic wall. ¹⁸ The most critical period for anastomotic healing is during the first postoperative week when collagenase activity is high.^11,19 AL usually becomes apparent at this stage and is uncommon thereafter.^20,21 Therefore, for this study, we aimed to leave the Colovac+ in situ for 10 days.

Studies with the initial Colovac design indicated that migration was an important issue for the Colovac as well. The Colovac was first introduced in an experimental study similar to ours. ¹¹ Early (after 3 to 11 days) device migration occurred in 7 of the first 8 pigs, and was presumed to result from feces bulking at the distal end of the stent leading to excessive pressure and subsequent stent displacement. After diet adaptation to a strict residue-free regimen, a subsequent group of 14 pigs was included. Disconnection of the suction drain from the vacuum chamber occurred in 1 case, leading to device expulsion on day 10. No migration was noted in the other 13 animals. The subsequent first-in-human clinical trial included 15 patients that underwent laparoscopic anterior resection with Colovac implantation. ¹² Four patients experienced AL during the implantation period (14 days), and 3 leaks were correlated with device migration. Two of these migrations were attributed to a vacuum defect, and 1 was due to device misplacement during the implantation procedure. The migrated devices were endoscopically removed, and conversion to protective ileostomy was performed to ensure further protection of the anastomosis.

In response to these results, the Colovac design was adjusted in order to enhance the vacuum anchor and reduce the migration risk during the implantation period. The vacuum was reinforced by using a double stent and an extra Redon vial (instead of a single stent and Redon vial). The vacuum chamber is situated in the area where the stents are overlapping to provide a homogeneous vacuum distribution all around the anchor. Moreover, the connection between the vacuum tubes and Redon vials was strengthened to prevent unintended disconnection and vacuum loss.

In the present study, we examined the preliminary efficacy, safety, and technical feasibility of this upgraded design (Colovac+) in a porcine model. The device met all our endpoints. There was no evidence of contact between the anastomosis and fecal content, none of the animals had symptomatic AL, and there were no significant migrations, demonstrating the efficacy of the Colovac+. The postretrieval examinations revealed ulcerations at the anchoring site in 4 animals indicating mechanical damage caused by the stent. These lesions were more pronounced at the proximal anchoring site, which may be related to the peristaltic movements of the colon at the proximal end of the device. One month after device retrieval, however, no significant lesions were observed at the anchoring site and the colonic mucosa had sufficiently healed, confirming the safety of the Colovac+. In one animal, the device could not be deployed during implantation and a second attempt had to be undertaken. Nevertheless, the device was easily implanted and retrieved in all other cases without technical difficulties or major adverse events.

There are some limitations to the present study. First, this was a small sample-sized series using a swine model, and the results must be confirmed in larger studies with human subjects before any strong conclusions can be made. We chose not to have a control group as no animal would have presented with AL. Moreover, the surgical model did not simulate the vascular supply of the anastomosis obtained after colorectal resections, as there were no ligations of the inferior mesenteric artery. However, the purpose of the device is the same as the protective ileostomy: not to reduce the leakage rates but to prevent the clinical consequences of AL by covering the anastomosis.

Conclusion

In conclusion, this preclinical study demonstrates that the Colovac+ is an efficient, safe, and technically feasible device for the protection of a colorectal anastomosis that may prevent the clinical outcomes associated with AL in a porcine model. The device holds promise for clinical use and warrants further research.