Complications: Early Anastomotic Complications—Leak, Abscess, and Bleeding

Abstract

Colorectal surgery has experienced significant improvements in surgical techniques, minimally invasive approaches, antibiotic prophylaxis, and perioperative care. In spite of these advancements, anastomotic complications continue to occur and often contribute substantially to postoperative morbidity and mortality. Surgeons must be prepared to identify anastomotic complications early in order to manage them thoughtfully. Infectious complications related to an anastomotic leak may lead to local or generalized peritonitis. Management options range from observation and antibiotic treatment, to image guided drainage and reoperation, based on the location of the leak and the patient’s clinical condition. Hemorrhage from the anastomosis will often resolve without intervention, but in situations where there is a major bleed, management strategies should follow the same principles as any other patient presenting with a lower gastrointestinal bleed.

Intra-abdominal Anastomotic Leak

Refer to Algorithm in Fig. 85.1

• A. A high index of suspicion is necessary when evaluating patients following creation of a colorectal anastomosis. Anastomotic leak occurs in 2–19% of patients, depending on the location of the anastomosis. Operative factors such as intestinal blood supply, tension at the anastomosis, and prolonged operative duration can increase the risk for developing anastomotic complications. In addition, patient and disease factors such as elevated body mass index, old age, tobacco use history, inflammatory bowel disease, cancer stage, preoperative radiotherapy, and immunosuppression increase the risk of developing post-operative complications. Major leaks are more easily identified as they present with classic signs including generalized peritonitis or septic physiology. However, soft signs such as a low-grade fever or a prolonged post-operative ileus may be the only clinical manifestations of more insidious leaks. Research related to improved detection and earlier interventions to reduce the morbidity associated with leaks is ongoing. Some groups now advocate the use of standardized post-operative surveillance protocols to more effectively diagnosis and treat anastomotic leaks. Others are targeting biomarkers that may identify leaks. Serum cytokines, lipopolysaccharides, and C-reactive protein have been reported to potentially help identify patients with post-operative complications.

• B. Management strategies for pelvic anastomoses (e.g. extraperitoneal anastomoses) will be discussed in Fig. 85.2. Key differences are present when evaluating patients with pelvic leaks. Given the lack of an innervated peritoneal surface, patients suffering from pelvic leaks will often not develop peritoneal signs which are common in the setting of intra-abdominal anastomotic leaks. However, if a localized abscess is to rupture into the peritoneal cavity, generalized abdominal pain may develop.

• C. Patients may present with subclinical leaks . Subclinical leaks may be identified on studies evaluating the anastomosis (i.e., when planning for the reversal of a proximal diverting ostomy) or by the presence of enteric contents in a drain. In these patients who do not have any clinical evidence of leak, management can be performed in an expectant manner. If the fluid collection is accessible, it may be worthwhile to percutaneously drain and initiate a short course of antibiotics. Interval imaging after 6–8 weeks is performed to evaluate if the leak has resolved. This cycle of antibiotics and imaging can continue for months, with most subclinical leaks eventually resolving and allowing for a safe second stage operation to reverse the proximal stoma. In the event of a persistent leak, the second stage operation will require revision of the anastomosis.

• D. Oftentimes, patients will present with non-specific findings such as localized peritonitis. In the correct clinical setting, this may imply an anastomotic leak. So long as a patient’s clinical status is stable, further evaluation with imaging studies should be performed. When performing these studies, a thoughtful approach and clear communication with the radiologist can make a significant impact. When assessing an anastomosis, the most sensitive study to identify a leak is considered a CT scan with oral, intravenous (IV), and rectal contrast (Fig. 85.3). If coordinated appropriately, cross-sectional imaging can serve as a diagnostic and therapeutic strategy, potentially identifying leaks that are amenable to drainage or aspiration. Though not as sensitive for identifying leaks, water-soluble contrast enemas can also be performed.

• E. In patients with localized peritonitis and stable vital signs , it’s unlikely that a free intra-abdominal leak is identified on imaging. However, if such a situation is encountered, management strategies should focus on a similar approach as described for patients with severe leaks (Fig. 85.1, Sect. K). Management strategies for patients presenting with contained leaks are related to the complexity of the fluid collections.

• F. For contained leaks larger than 3 cm, that contain multiple abscesses, or have imaging characteristics of a multi-loculated collection, the clinical condition of the patient will guide management. Percutaneous drainage may be used to manage many of these findings. Operative intervention should be sought for patients who have large burdens of infection, abscesses in inaccessible anatomic sites, or have persistent sepsis or deteriorating clinical status. Preoperative evaluation of the anastomosis with water-soluble contrast enema or CT scans with 3D reformatting can help to develop an operative plan—minimizing manipulation of a potentially salvageable anastomosis. However, if these studies are not available or difficult to interpret, careful intraoperative assessment of the anastomosis is still possible. At the time of laparotomy, the anastomosis is especially fragile secondary to local inflammation and infection. The viability of the anastomosis can be assessed in a variety of ways if it is unclear whether or not it is intact. Air-leak tests, instillation of intraluminal fluids (i.e., methylene blue with saline), or concurrent endoscopic evaluation can help determine if the anastomosis is perfused, and if it is amenable to repair and salvage.

• G. For small, contained abscesses measuring less than 3 cm, broad-spectrum antibiotics should be initiated and source control obtained via percutaneous drainage, if possible. If the cavity is large enough, a drain should be left in place to allow for further drainage. Some have advocated the use of saline irrigation and fibrinolytic agents to maintain patency and improve drainage. In the event that the cavity cannot be accessed, broad-spectrum antibiotics will be adequate for the majority of patients.

• H. Anastomotic complications that present with generalized peritonitis or severe sepsis require urgent interventions, as delay can contribute to morbidity and mortality. In addition, any patient who is managed conservatively with non-operative measures and fails to improve will require surgical intervention.

• I. For patients who are critically ill, standardized guidelines for the management of sepsis should be followed. Initial resuscitation to restore tissue perfusion is imperative. Measuring biomarkers, obtaining blood cultures, initiating broad spectrum antibiotics and administering fluids should be done expediently. If the patient does not respond to initial fluid administration and is persistently hypotensive, vasopressors should be applied to maintain a mean arterial pressure greater than 65 mmHg. Upon completing these bundles of care, operative management for the anastomotic leak is necessary for source control.

• J. No true consensus has been established for defining the severity of anastomotic leaks preoperatively. The International Anastomotic Leak Study group has identified major leaks both anatomically (i.e., defects greater than 1 cm or greater than one-third the circumference of the anastomosis) and clinically (i.e., severity of sepsis, physical examination findings). Intraoperatively, a variety of situations may be encountered and are discussed below. For any patient undergoing a laparotomy, a decision must be made regarding the original anastomosis. The options include resection of the anastomosis with formation of an end stoma and Hartmann’s type distal closure, exteriorization of both ends of the anastomosis, resection and repeat anastomosis with proximal diversion, primary repair of the anastomosis with or without proximal diversion, and proximal diversion alone. Irrespective of procedure, the goal is for source control. Decisions regarding the use of drains should err on the side of caution, with a low threshold to have wide drainage (at least initially). In cases where there is severe soiling of the abdominal cavity, damage control techniques such as open management with delayed abdominal closure may be considered, allowing for reevaluation of the abdomen when the patient’s physiologic parameters are more favorable for a prolonged operation. Patients should be advised that regardless of the final outcome related to diverting stomas, future operations for the sake of reestablishing gastrointestinal continuity will nearly always be delayed 3–6 months to allow resolution of acute inflammatory adhesions and reduce the risk of operative complications.

• K. At times, inflammation and adhesions preclude a safe evaluation of the anastomosis. Further dissection in the area of interest may convert a minor leak into a major one. In these scenarios, damage control measures should be undertaken. Thorough irrigation and drain placement can help decrease the burden of infection. Next, identification of the proximal segment of bowel can allow for diversion upstream of the anastomosis. If the proximal bowel is mobilized, a looped ostomy will allow for the safest operation. Unfortunately, these cases often present with significant inflammation of the bowel near the anastomosis. As such, diversion may require the utilization of proximal ileum or jejunum, making the post-operative management even more challenging, as many patients can require supplemental nutrition via parenteral access.

• L. Major anastomotic defects have large infectious burdens. In addition, the viability of the bowel near the anastomosis is typically poor. In these situations, resection of the anastomosis with end stoma formation may be necessary. If there is adequate mobilization and length to the distal bowel segment, some advocate exteriorizing both ends of bowel to reduce potential complications related to a distal stump blowout. A surgeon should exercise caution if considering the formation of a new anastomosis. In these cases, the benefits of a simplified second stage operation to reestablish gastrointestinal continuity must be weighed against the potential for another anastomotic leak. As such, this intervention should be avoided in patients with physiologic evidence of severe sepsis or underlying risk factors that contributed to their first leak.

• M. Management strategies for minor leaks (i.e., size less than 1 cm or encompassing less than one-third of the anastomotic circumference) are based on the patient’s clinical status and the quality of the tissues at the original anastomosis. If the tissues at the area of anastomotic dehiscence are of good quality, a primary repair may be attempted. Generally, it is advisable to leave drains near the anastomosis and perform a proximal diversion with a loop ostomy. If the tissues are not of good quality, then management strategies are similar to those described for major leaks (Fig. 85.1, Sect. L).

Fig. 85.1

Algorithm for management of intra-abdominal anastomotic leaks. Adapted from Phitayakorn et al.

Fig. 85.2

Algorithm for management of pelvic anastomotic leaks. Adapted from Phitayakorn et al.

Fig. 85.3

CT scan to detect leak

Pelvic Anastomotic Leaks

Refer to Algorithm in Fig. 85.2

• A. Anastomotic leaks within the pelvis are managed with many of the same principles as those for patients with intra-abdominal leaks. However, low rectal anastomoses may pose additional anatomic challenges, as there may be cases in which it is not possible to safely resect, repair, or revise a pelvic anastomosis. Patients presenting with generalized peritonitis and signs of sepsis will require an emergent operation, but upstream diversion may be the only viable intervention (Fig. 85.1, Sect. H). When resecting the anastomosis, one may encounter a very inflamed and friable distal rectal stump which is impossible to suture or staple closed. In these scenarios, one should irrigate the pelvis and rectum and leave pelvic and transanal drains. Patients presenting with localized peritonitis will require similar management approaches as described for patients with intra-abdominal anastomoses (Fig. 85.1, Sect. D). Cross sectional imaging with CT scans with IV and rectal contrast will help delineate the location of fluid collections and determine if it communicates with the lumen, respectively. Given the friable nature of the anastomosis, rectal contrast should be administered by a member of the surgical team or an experienced radiologist. Soft rubber catheters and careful instillation of contrast can reduce the chance of further disrupting the anastomosis. Endoscopic or proctoscopic evaluation can be pursued to further identify the location of the anastomotic dehiscence.

• B. As mentioned before, pelvic anastomotic leaks may present differently than intra-abdominal leaks secondary to a lack of a large, exposed peritoneal surface. In addition, low rectal anastomoses are often protected by a diverting ileostomy, making the clinical presentation sometimes more subtle. Paying attention to subtle signs such as tachycardia, oliguria, prolonged post-operative ileus, or changes in mental status may be the first clues that a patient is suffering from pelvic sepsis. If not identified early in the course of infection, pelvic abscesses may increase in size and rupture into the abdomen (causing peritonitis) or drain into the intestinal lumen (providing source control for the infection in some cases).

• C. For high rectal abscesses near the peritoneal reflection, one should proceed with a similar management strategy as described for leaks encountered in the abdomen (Fig. 85.1, Sect. D).

• D. Low pelvic abscesses identified on cross-sectional imaging are characterized by their anatomic location relative to the rectum (anterior vs. posterior) and whether or not they communicate with the lumen. These leaks more often present posteriorly, which is fortunate as anterior leaks can be more treacherous to treat because of proximity to surrounding organs. Antibiotics are used in most patients, usually in combination with other techniques for larger fluid collections. Abscess that are not in continuity with the lumen can be managed via percutaneous, transrectal, transanal, or transvaginal approaches with a low risk of developing a fistula. In certain cases, the abscess is unable to be accessed via these first-line approaches and alternatives such as transgluteal drains must be utilized. Pain from sciatic nerve inflammation and spread of infection into the gluteal region have been described. For low-lying anastomoses, examination under anesthesia can provide another option for transrectal drainage. During proctoscopy, a mushroom tipped catheter can be placed within the abscess cavity and sutured externally. This allows for continuous drainage of the cavity as well as access for fluoroscopic evaluation during follow up.

• E. For abscesses that are in continuity with the lumen, drainage can be performed through the site of dehiscence to reduce the potential for fistula formation. Enlarging the dehiscence bluntly with a finger or instrument can be performed. Again, utilizing catheters to maintain patency of the communication and allow for drainage of the cavity is recommended. Once the cavity has drained and regressed in size, the catheter can be removed to allow for the anastomosis to heal. The use of endoscopically placed covered stents have recently been described but are not commonly employed. A key element to this strategy is percutaneous drainage of the cavity. The stents are known to migrate and are even expelled by the patient prompting some to use endoclips to better secure the stent following deployment.

Anastomotic Bleeding

Refer to Algorithm in Fig. 85.4

Bleeding from the anastomosis can have a wide spectrum of presentations and should be managed in a similar manner to lower gastrointestinal bleeding that presents from other causes. Minor bleeds are relatively common, yet rarely reported as a complication. They are identified by the passage of blood for the first few bowel movements following operation. For the majority of patients, minor bleeding is a self-limited problem that requires no intervention. However, some patients may progress to a major bleed which result in hemodynamic instability, the need for blood transfusion, or necessitate a procedure to obtain hemostasis. These bleeds typically originate from a trapped mesenteric vessel within the anastomosis or a perforating vessel near the anastomosis. Given the etiology, some advocate the use of endoscopic evaluation during initial operation to ensure anastomotic hemostasis. As with all cases of bleeding, coordination of care is imperative. Adequate intravenous access and blood component resuscitation is necessary. Following initial resuscitation of the patient, hemodynamic stability should be assessed to guide the next steps in management.

It is rare for anastomotic bleeding to result in shock. However, if resuscitation attempts do not normalize the patient’s hemodynamic status, emergent endoscopy and/or abdominal exploration should be performed. During the operation, one should work to quickly identify the anastomosis, evaluate for bleeding, and either over-sew a bleeding vessel or resecting the previous anastomosis. If the patient is stable, a new anastomosis may be created at this time. If not, the patient may require an end ostomy and have a second stage operation for reestablishing gastrointestinal continuity.

The majority of patients requiring intervention for an anastomotic bleed can be managed endoscopically. Techniques including washout using saline, electrocoagulation, injection of epinephrine, and clipping of bleeding vessels have all been successfully utilized to obtain hemostasis. If visualization of the lumen is challenging secondary to clot burden, it may be necessary to convert to laparotomy or obtain angiographic evaluation with subsequent embolization of the bleeding vessel. The latter approach should be exercised with caution as it can place the anastomosis in danger of ischemia which can result in leak acutely or stenosis at long-term follow up. If bleeding persists following non-operative interventions, then laparotomy to evaluate the anastomosis will be needed (Fig. 85.4).

Fig. 85.4

Algorithm for management of anastomotic bleed