Externalized Ileocolic Anastomosis: Case Report

Introduction

The practice of externalizing the colon and performing a colostomy to facilitate repair of colonic injuries was first reported around the time of the Second World War.¹ Toward the end of this period, James Mason, a military surgeon, introduced the technique of primary anastomosis of the unprepared colon and exteriorization of the affected bowel segment outside the peritoneal cavity.² Since then, exteriorization of an anastomosis has gained widespread acceptance, and a number of reports advocate its use.^3–5 Advantages of exteriorization of an anastomosis include decreased infection rates, decreased morbidity, and shorter periods of hospitalization.^3–5

In small animals, rates of dehiscence at anastomosis sites or following small intestinal enterotomy are between 7% and 15.7%.^6,7 Sepsis is reportedly more common following surgery for intestinal foreign bodies than for other intestinal surgeries.⁸ Mortality rates in animals having dehiscence of enterotomies or anastomoses are between 50% and 74%.^7,9 In another study, 46% of animals that were presented with generalized peritonitis had a history of abdominal surgery in the preceding 14 days.¹⁰ Half of the dogs included in that study had dehiscence of a surgical wound.¹⁰

When peritonitis develops, proteolytic activity leads to the degradation of collagen and extracellular matrix. These conditions are unfavorable for the healing of intestinal surgical repairs and predispose animals to dehiscence.¹¹ A retrospective study of 115 animals identified three risk factors associated with leakage of an intestinal anastomosis: preoperative peritonitis, serum albumin levels <2.5 g/L, and an intestinal foreign body as the indication for surgery. The study showed that 84% of animals with either two or all three of these factors experienced leakage from the anastomosis site.⁸ Externalization of an ileocolic anastomosis, a novel alternative for the management of intestinal anastomoses in a septic environment in small animal practice, is reported herein.

Case Report

A 6-year-old, spayed female Labrador retriever was presented with abdominal pain 48 hours after an intestinal resection and anastomosis had been performed for management of a small intestinal foreign body (a corn cob). On presentation, the dog’s heart rate was 190 beats per minute (BPM), rectal temperature was 40.2°C, capillary refill time was <1 second, and the dog was panting. Femoral pulses were weak, and the abdomen was painful on palpation. On admission, the packed cell volume was 44% (reference range 37% to 55%), blood glucose was 5.2 mmol/L (reference range 4.11 to 7.94 mmol/L), and the total protein was 46 g/L as determined by a refractometer (reference range 52 to 82 g/L). Initial stabilization included intravenous (IV) fluid therapy (Hartmann’s 5 mL/kg per hour) and buprenorphine (0.01 mg/kg IV).

Perfusion parameters improved over the next 2 hours. The heart rate decreased to 120 BPM, capillary refill time increased to 1 to 2 seconds, and the rectal temperature decreased to 38.5°C. Eight hours after admission, the dog was depressed and showed signs of hyperdynamic shock. The dog displayed tachypnea (60 breaths per minute), tachycardia (160 BPM), hyperemic mucous membranes, a capillary refill time <1 second, and pyrexia (39°C). A second complete blood count revealed a packed cell volume of 44%, total protein of 46 g/L, and hypoalbuminemia (17 g/L, reference range 23 to 40 g/L). A blood smear identified a left shift with >30% of the neutrophils in band form and some displaying prominent granulation, vacuolation of the cytoplasm, and Dohle bodies—consistent with toxic degeneration.

The dog’s abdominal discomfort progressed, and a constant-rate infusion (CRI) of morphine (0.24 mg/kg per hour) and ketamine (0.24 mg/kg per hour) was initiated. An abdominal ultrasound confirmed the presence of peritoneal effusion. Abdominocentesis was performed, and 5 mL of fluid was collected. Analysis of the fluid revealed a total protein of 32 g/L and numerous neutrophils (>20 per oil-immersion field, 1000× magnification). A large number of these neutrophils displayed degenerate changes and intra-cellular cocci. Some macrophages and red blood cells were also present. Based on these findings, septic peritonitis was diagnosed. Enrofloxacin (5 mg/kg IV q 24 hours), metronidazole (15 mg/kg IV q 12 hours), and cephazolin (22 mg/kg IV q 8 hours) were initiated. Chlorpheniramine (11 mg subcutaneously) was administered, followed by 1 unit (200 mL) of fresh-frozen plasma (Caniplas;^a 10 g/L canine gamma globulins). The plasma was administered at 2.5 mL/kg per hour for 15 minutes, and then it was increased to 10 mL/kg per hour in the absence of systemic signs of a hypersensitivity reaction. Blood testing following transfusion revealed a packed cell volume of 33%, a total protein of 44 g/L, and hypoalbuminemia (16 g/L).

The decision was made to explore the abdomen surgically. No further diagnostic tests were performed prior to anesthesia. The dog was premedicated with acepromazine (0.01 mg/kg IV). Anesthesia was induced with alfaxalone (1 mg/kg IV) and maintained with 2% isoflurane. The morphine/ketamine CRI was continued throughout surgery at the rate described previously. Anesthetic monitoring included pulse oximetry, capnography, noninvasive blood pressure monitoring, and electrocardiography. A routine ventral mid-line celiotomy was performed through the previous surgical incision, extending from pubis to xiphoid. Two liters of san-guinopurulent material were suctioned from the abdomen. The anastomosis was readily identified 4 cm orad to the cecum and was found to be leaking intestinal contents from the mesenteric aspect of the anastomosis [Figure 1⇓].

The intestine was resected at a level of the colon and ileum that was considered viable based on gross appearance. The colon was resected approximately 4 cm aborad to the cecum, necessitating removal of the cecum and ileocolic valve. An end-to-end, appositional ileocolic anastomosis was performed with three end-to-end sections of simple continuous 4-0 polydioxanone (PDS) sutures placed in a triangular pattern. The remainder of the abdomen was inspected and appeared to be grossly normal.

The abdomen was lavaged with 4 L of warmed 0.9% sodium chloride (NaCl). A 10-cm, paramedian incision was made through the abdominal musculature to the left of the midline incision [Figure 2⇓], and the anastomosis was externalized from the abdomen into a subcutaneous pocket [Figure 3⇓]. Three horizontal mattress sutures of 0 PDS were placed along the paramedian incision and passed through the mesentery of the intestine to maintain the intestinal loop and the anastomosis in a position external to the peritoneal cavity. Careful attention to suture tension ensured the anastomosis was maintained external to the abdomen, strangulation of the blood supply to the intestine was avoided, and passage of food through the intestine was possible.

A fenestrated Jackson-Pratt^b peritonostomy tube was placed through the abdominal wall 5 cm lateral to the umbilicus to allow drainage of fluid from the peritoneal cavity and serial cytological evaluation of the abdominal fluid. The midline abdominal musculature was closed with 0 PDS in a simple continuous pattern. A laparotomy sponge was soaked in 0.9% NaCl and placed over the anastomosis to act as a wet-to-dry bandage and to protect the repair. The skin was then partially closed over the laparotomy sponge with a simple continuous suture pattern using 0 PDS to create an open subcutaneous pouch, allowing drainage and facilitating daily changing of the wet-to-dry bandage.

Recovery from anesthesia was unremarkable. Postoperative analgesia was administered via a fentanyl patch (50 μg per hour, delivering 1.6 μg/kg per hour), and the morphine/ketamine CRI was decreased incrementally over the first 24 hours postoperatively following placement of the patch. The peritonostomy tube was aspirated every 12 hours until negative pressure was achieved. Repeat cytological analysis was performed on the fluid, and over the 4 days the neutrophils decreased in number and progressively appeared less degenerate.

Following the exploratory abdominal surgery, the dog was hypoalbuminemic (11 g/L). A second unit of fresh-frozen plasma was administered (as described above). Following the transfusion, oncotic pressure was supported with synthetic colloids (6% hydroxyethyl starch in 0.9% NaCl) administered as a 5 mL/kg bolus, followed by a 20 mL/kg per day CRI. Serum albumin decreased to 8 g/L; however, the dog was eating readily (Hills I/D^c) within 12 hours of recovering from surgery, and no evidence of edema formation was seen. No further colloid support was administered. Serum albumin increased to 13 g/L and 15 g/L by 24 and 48 hours postoperatively, respectively.

A general anesthetic was administered daily (induction: alfaxalone [1 mg/kg IV]; maintenance: 2% isoflurane) for 4 days to allow assessment of the anastomosis site [Figure 4⇓]. For the first 3 days postsurgery, the wet-to-dry bandage (i.e., the 0.9% NaCl-soaked laparotomy sponge) was changed, and progressive healing of the anastomosis site was noted. On day 4 postoperatively, no evidence of a continued peritonitis was seen based on cytological examination of the peritoneal fluid obtained from the peritonostomy tube. The mattress sutures used to maintain the intestinal loop external to the abdomen were cut, and the anastomosis was returned to the abdomen. The paramedian incision in the abdominal wall was closed with a simple continuous suture pattern using 0 PDS. The subcutaneous tissue was closed with a simple continuous suture pattern using 2-0 PDS, and the skin was closed with a cruciate pattern using 2-0 nylon.

Cytological examination of the peritoneal fluid, performed every 12 hours for an additional 48 hours after returning the anastomosis to the abdominal cavity, showed no evidence of peritonitis. Clinically, the dog continued to improve and was eating well and toileting normally. The dog was discharged from the hospital 6 days after the repeated anastomosis. Ten days following return of the anastomosis to the abdomen, an 8-cm swelling over the site of bowel exteriorization appeared. Ultrasonography of the swelling revealed a subcutaneous, homogenous, anechoic pocket that was superficial to the abdominal musculature and consistent with a seroma. Needle centesis was not performed. No specific treatment was instituted for the swelling, and it resolved without complication within 10 days. The dog subsequently made a full recovery. Oral antibiotics (amoxicillin/clavulanic acid [12.5 mg/kg q 12 hours] and enrofloxacin [5 mg/kg q 24 hours] for 7 days) and an oral nonsteroidal antiinflammatory medication (carprofen [2 mg/kg q 24 hours] for 5 days) were prescribed upon discharge.

Discussion

In a paper by Ralphs et al (2003), three risk factors for leakage of intestinal anastomosis were identified: preoperative peritonitis, hypoalbuminemia, and an intestinal foreign body as the indication for surgery.⁸ The dog described herein had all three of these risk factors. Thus, primary repair of the leaking anastomosis with routine intraabdominal resection and anastomosis was considered a significant risk for dehiscence.

Treatment of peritonitis includes correction of fluid and electrolyte abnormalities, appropriate antimicrobial therapy, and exploratory surgery to investigate and correct the underlying cause of the peritonitis. Minimizing contamination by thorough lavage to remove bacteria, exudates, foreign material, and toxic products from the abdomen is indicated. Depending on the severity of contamination and effective correction of the underlying cause of peritonitis, peritoneal drainage may be appropriate.^9,12 Peritoneal drainage techniques have their limitations and may not be appropriate or required in all cases.

Omentalization and serosal patching are two surgical techniques that have been advocated to help prevent leakage following intestinal surgery in the presence of peritonitis.¹³ Experimental studies have shown that the omentum can adhere to the affected intestinal site and help prevent perforation and leakage.^14,15 One of these studies reported a devascularized, small intestinal anastomosis wrapped in omentum. This prevented leakage and development of peritonitis in all experimental animals (n=6). The animals in that study did not have a preexisting peritonitis.¹⁴ A second study evaluated dogs that had an intestinal anastomosis with an adjacent avascular segment. In dogs that had omental wrapping, 11 out of 14 survived to 3 weeks compared to two of 17 dogs that had the omentum sutured to the stomach (i.e., the control group).¹⁵ Nonetheless, some authors report that the omentum will not prevent leakage and that serosal patching may instead be required.¹³ One author has described serosal patching as a surgical parachute; that is, “It is rarely needed, but when the occasion arises nothing else will take its place.”¹⁶ While serosal patching and omentalization are accepted, externalizing an intestinal repair may offer surgeons an alternative to those techniques, especially when risk factors for wound dehiscence are present.

A study of extraperitoneal anastomosis in rats demonstrated that the bursting pressure of jejunal anastomoses was less for extraperitoneal anastomoses compared to intraperitoneal anastomoses at both 3 and 7 days postoperatively.¹⁷ In these rats, no peritonitis was present in the intraperitoneal group. The bursting pressure of the anastomoses 7 days postoperatively was 104 mm Hg for the extraperitoneal group compared to 200 mm Hg for the intraperitoneal group. The normal physiological pressure created by canine jejunum in vivo is about 38 to 40 mm Hg,¹⁸ which is well below the bursting pressure of the anastomoses recorded in rats. This suggests that the decrease in strength of the externalized anastomosis may not be clinically relevant. Pierie et al (1999) recommended that extraperitoneal repair should only be considered when risk factors, such as peritonitis and low serum albumin, are present. If these risk factors are present, then the bursting pressure for intraabdominal anastomoses may be decreased.

Interestingly, in a separate study by Pierie et al (2000), omentalizing an extraperitoneal anastomosis increased the bursting strength of the anastomoses at day 3 postoperatively.¹⁹ While the externalized repair could be omentalized, assessment of the healing of the anastomosis may be impeded with omentalization.

Regardless of the exact technique employed, one benefit of draining fluid from the abdomen in cases of septic peritonitis includes removal of infectious agents, inflammatory mediators, and foreign material. Also, fluid within an infected body cavity has been shown to significantly impair humoral and cell-mediated immunity.²⁰ The problems associated with tube drainage of the abdomen are related to the inability to drain the entire abdomen, because tubes can be rapidly blocked by omentum and debris.²¹ Further, bacterial migration into the abdomen can occur as early as 24 hours after drain placement.^21,22 Open peritoneal drainage has been well documented in the literature. Benefits of this procedure include increased efficiency of removal of bacteria, inflammatory mediators, and foreign material, as well as the creation of an environment that is unsuitable for anaerobic bacteria. Disadvantages include the loss of fluid and protein from the abdomen, development of nosocomial infections, increased nursing care, and increased costs.^9,12,22–24 Interestingly, in prospective clinical trials conducted in humans, no advantages have been demonstrated, and significantly higher complication rates have been reported for patients receiving open peritoneal drainage.^25,26

In one study of 28 dogs treated surgically for septic peritonitis without peritoneal drainage, a 46% mortality rate was reported.⁹ This study suggested that if the cause of the peritonitis is corrected, appropriate peritoneal lavage is utilized, and postoperative medical management is initiated, then closure of the abdomen is acceptable. In the case described herein, the above criteria were met, and the decision was made to close the abdomen. A study by Staatz et al (2002) compared open peritoneal drainage versus primary closure of the abdomen in cases of septic peritonitis. No difference in survival between the two groups of animals was noted, and an overall mortality rate of 29% was reported.²⁷

With externalized loop anastomosis, the anastomosis is returned to the abdomen once the peritonitis has resolved. Placement of a peritonostomy tube can aid in the assessment of the abdominal environment. In one study, needle paracentesis was shown to have an overall accuracy of 43% compared to 82.9% for catheter paracentesis and 94.6% for diagnostic peritoneal lavage.²⁸ To our knowledge, the accuracy of a peritonostomy tube for evaluating peritonitis has not been determined. The peritonostomy tube possibly can become walled off by the omentum, creating a small pocket of fluid isolated from the remainder of the peritoneum. If this were to occur, the fluid obtained from the peritonostomy tube would not likely be representative of the fluid in the remainder of the abdomen. In the case described in this report, the fluid collected by the peritonostomy tube was evaluated, and clinical assessment of the dog (i.e., voluntary eating, reduction in abdominal discomfort, absence of vomiting) was made. Findings from both evaluations were used to evaluate resolution of septic peritonitis. While abdominal ultrasonography could be used to assess for peritoneal fluid accumulation at a site isolated from the peritonostomy tube, this was not used in this case.

In humans, the time for return of the colonic anastomosis to the abdomen is between the fifth and 14th day postoperatively.⁵ One paper on human externalized colonic anastomosis reported that 76% of patients had their colonic anastomosis dropped back into the abdomen an average of 5 days postoperatively.²⁹ Asfar et al (2007) advocate early return of the anastomosis, between the fifth and seventh days postoperatively, for the following reasons: the risk of developing serositis from prolonged exposure of the colon to atmospheric air is reduced; a leak in the anastomosis would be expected to show itself within this timeframe; the integrity of the repair would have been confirmed by passage of fecal material; the presence of postoperative peritonitis would be expected to declare itself by this time; and finally, less sharp dissection would be required to break down the fibrinous adhesions between the intestinal section and abdominal wall prior to them becoming fibrous.⁵

In the present case, a laparotomy sponge soaked in 0.9% NaCl acted as a wet-to-dry medium. This may have helped prevent a serositis in the externalized intestinal segment. The integrity of the repair was assessed by direct visualization of the anastomosis and by the observation that the dog was defecating prior to return of the anastomosis to the abdomen with no leakage of intestinal contents. Serial cytology of fluid collected from the peritonostomy tube, in addition to the clinical assessment of the dog, helped to confirm the resolution of the peritonitis prior to return of the anastomosis to the abdomen 4 days postoperatively.

Major complications commonly reported with peritoneal drainage include hypoproteinemia and hypoalbuminemia.²³ Loss of protein into the inflammatory effusion is likely to occur irrespective of the drainage technique. The study by Hosgood (1988) found that the peritoneal fluid from normal dogs experimentally treated with open peritoneal lavage contained between 24 and 51 g/L of protein. Between 3 and 5 mL of fluid was retrieved from the peritonostomy tube at each aspiration before negative pressure was obtained. In our opinion, this small volume of fluid is unlikely to contribute significantly to overall protein loss. As discussed previously in this report, the omentum possibly can wall off the peritonostomy tube, thereby explaining why only small volumes of fluid are obtained from peritonostomy tubes. If this had been the situation, accumulation of protein-rich peritoneal effusion in the abdomen, not effectively drained by the peritonostomy tube, may have contributed to the dog’s hypoalbuminemia.

A large volume of fluid (an estimated 2 L per day) exuded from the open wound containing the externalized intestinal segment in the Labrador retriever described in this case. It is not clear if this fluid was peritoneal fluid exiting the abdomen through the paramedian incision around the intestine or if this fluid was produced primarily by the externalized segment of intestine. If the fluid was draining from the abdomen, this could also explain the relatively low volumes of fluid obtained from the peritonostomy tube. The total protein of the fluid from the externalized intestinal segment was not measured, but this fluid loss would have contributed in some degree to the dog’s hypoalbuminemia. Other factors that may have contributed to the dog’s hypoalbuminemia include the effect of dilution from IV fluid therapy, decreased albumin production, and the nutritional status of the dog (as the dog had not eaten for 3 days prior to the externalization of the anastomosis). The dog began eating 12 hours following externalization of the anastomosis and continued to eat when food was offered during hospitalization.

In humans, to maintain the colon in a location external to the peritoneum, a semirigid drain or piece of chest tube is passed via the mesenteric border of the repair. Each end of the tube is passed subcutaneously and brought through the skin 3 to 4 cm away from the externalized repair.⁵ The colonic anastomosis is therefore maintained on the outside of the skin, protected with a colostomy bag.⁵

In the canine case described here, the anastomosis was maintained subcutaneously in an effort to protect it from trauma and atmospheric air. The decision was made to use a horizontal mattress suture pattern rather than a semirigid tube to maintain the anastomosis external to the peritoneum. Excessive tension on these sutures was thought to potentially impede the passage of ingesta through the externalized anastomosis (by creating a stenosis where the intestine exits the abdominal wall) or to compromise blood flow to the anastomosis. In the case of inadequate suture tension, however, the anastomosis may not be effectively externalized from the peritoneum, and the risk of herniation of abdominal viscera through the incision is increased. The number of mattress sutures used should be appropriate to the size of the incision made to externalize the repair. In this case, three horizontal mattress sutures with 5- to 10-mm bites were placed at 10- to 15-mm intervals (for a 10-cm incision), leaving a 25- to 30-mm gap at each end of the paramedian incision for the passage of the intestine. In the study by Asfar (2007), only one of 103 human cases had a prolapse of the proximal colonic section through the abdominal incision.⁵

In the dog described in this report, the site of the leaking anastomosis was 4 cm orad to the cecum. The ileum and colon adjacent to the anastomosis were resected at a level of tissue that was considered healthy, based on gross appearance and bleeding from the cut surface. In order to achieve this, the cecum and ileocolic valve were sacrificed. To our knowledge, no complications associated with typhlectomy have been reported. The ileocolic valve has been reported to minimize access of colonic bacteria to the small intestine, and removal of the ileocolic valve may result in bacterial overgrowth in the small intestine, deconjugation of bile salts, and steatorrhoea.³⁰

Fresh-frozen plasma was administered in this case on two occasions. The dog was hypoproteinemic prior to surgery. The amount of plasma required to increase serum albumin by 10 g/L has been estimated to be approximately 45 mL/kg.³¹ Thus, the volume of each transfusion (200 mL) in this case was unlikely to have significantly increased the serum albumin. Nonetheless, along with factors that support intravascular oncotic pressure, plasma also contains clotting factors and alpha 2 macroglobulins. These products may well have been beneficial in this case.

Maintaining the anastomosis in a subcutaneous pouch on the ventral abdomen appeared to be well tolerated in this case, even with the dog in sternal recumbency. Potential limitations associated with externalization of an anastomosis include the morbidity associated with daily general anesthesia and the need for 24-hour care to monitor for complications such as self-trauma to the externalized anastomosis. It is not clear based on this case if the morbidity, complications, nursing care, or expense would be improved if other recognized treatment options were used for septic peritonitis. The expense, complications, and high level of nursing care need to be balanced against the high mortality rate reported with septic peritonitis following intestinal surgery and the expense associated with alternative treatments.

Conclusion

Externalizing an intestinal repair permits healing away from the detrimental environment created by peritonitis and helps prevent subsequent breakdown of the surgical anastomosis. Externalizing an anastomosis allows for direct visualization of the anastomosis to assess wound healing. Placement of a peritonostomy tube enables frequent cytology to ensure peritonitis has resolved prior to return of the repair into the abdomen, and it may mitigate the need for open peritoneal drainage. Externalization of an anastomosis also offers an alternative to omentalization and serosal patching in the face of peritonitis. This technique should be considered if risk factors for healing of intestinal anastomosis are present. Controlled clinical trials are needed to compare externalization of an anastomosis to serosal patching and/or omentalization and to assess the safety and efficacy of this procedure in a larger number of animals.