Anastomotic Leak

Anastomotic complications are among the most feared and difficult problems that colorectal surgeons commonly encounter in clinical practice.[1] The diagnosis of an anastomotic leak in its many varied forms and presentations is often quite delayed. In a review of 1223 patients undergoing an intestinal resection with anastomosis, the leak rate was 2.7%. Fourteen of 33 leaks (42%) were only diagnosed upon readmission to the hospital and 12% were identified more than 30 days after surgery. The positive predictive value of CT scan was 89.5% versus 40% for contrast enema (Figure 2.1). However, these studies were used in somewhat different clinical settings and the CT scans were often thought to be suggestive of a leak, rather than truly definitive (Figure 2.2).[2]

Figure 2.1

Figure 2.2

Categorizing CT scans dichotomously into “positive” or “negative” for leak is difficult because there is broad overlap in radiologic findings between patients who have an uncomplicated postoperative course and those ultimately diagnosed with a leak. There are also clinical similarities in presentation between a leak and other common postoperative complications. The fact is that surgeons often worry or even agonize when things turn out to be fine and are commonly led astray by “reassuring” clinical data when patients have actually suffered an anastomotic leak. One may consider the judicious use of rectal contrast at the time of a CT scan, particularly if the patient does not have a diverting ostomy, because it can help guide therapy if positive for rectal contrast extravasation.

Many factors need to be considered when deciding on the most appropriate management option for a patient with an anastomotic leak.[3] These include patient-specific factors such as the degree of hemodynamic derangement, physiologic reserve, nutritional status, comorbid complications, initial surgical indications/goals, and the potential need for additional treatments (e.g., chemotherapy for a malignant diagnosis). Similarly, features of the leak such as location (e.g., intraperitoneal, extraperitoneal), size of the defect, and the presence of concomitant tissue ischemia also play a major role in the surgeon’s decision-making process. Perhaps the most useful classification in outlining the principles of management is early versus late presentation. Patients with an early leak classically present in the first week after surgery with signs and symptoms of localized pain or peritonitis, end organ dysfunction associated with sepsis, and tachycardia. In this clinical setting with a profoundly sick patient, the diagnosis is generally quite evident and prompt return to the operating room is usually required. Radiologic studies are often unnecessary and may provide a false sense of reassurance.

However, even in the early postoperative period, patients with an anastomotic leak will often present with signs and symptoms that lead the surgeon astray and suggest other serious postoperative complications such as a pulmonary embolism, cerebrovascular event, or acute coronary syndrome. This is because patients with a leak will often appear short of breath or develop mental status changes, and the basic acute work-up will commonly reveal an abnormal chest roentgenogram or electrocardiogram. The surgical team must maintain a high index of suspicion for a leak in this setting and remain wary of alternative diagnoses.

If an anastomotic leak presents in the first few days after the initial surgery, most patients will require operative exploration if they do not already have fecal diversion. Intravenous antibiotics and close observation may be appropriate in selected patients with small, contained leaks that otherwise appear reasonably well, particularly if they already have a proximal diversion. Otherwise, at re-operative surgery, the peritoneal cavity is thoroughly explored and irrigated as the anastomosis is interrogated.

Compared to small bowel or colo-colonic anastomoses, anastomotic leaks following a low anterior resection pose additional anatomic challenges. This is because the location in the pelvis limits the ability to safely resect, repair, or revise a pelvic anastomosis. As with abdominal leaks, patients presenting with generalized peritonitis and signs of sepsis will require an emergent operation. The exact repair depends on the level of the rectal anastomosis and whether the patient was diverted at the time of the original surgery. Options include diversion with a proximal loop ileostomy combined with pelvic drainage. In other circumstances such as a major anastomotic disruption, the only option may be anastomotic disconnection and end colostomy. The distal rectal stump can often not be closed and the pelvis should be irrigated and drained.

Patients who are not terribly ill may be imaged with CT scan to delineate the location and size of fluid collections and their relationship with the anastomosis and bowel lumen. Low rectal anastomoses are often protected by a diverting ileostomy, making the clinical presentation more subtle. Low pelvic anastomotic leaks commonly present as a presacral fluid collection. Antibiotics are used in most patients and can be typically used alone in patients with phlegmonous changes in the pelvis or for those with small fluid collections. Larger abscesses can be managed via image-guided percutaneous approaches (Figure 2.3). For low-lying anastomoses, examination under anesthesia can provide another option for transanal drainage. During anoscopy or proctoscopy, a mushroom-tipped catheter can be placed through the anastomotic defect and within the abscess cavity and secured to the proximal bowel at the site of the defect (Figure 2.4). Enlarging the dehiscence bluntly with a finger or instrument can facilitate drainage. Once the cavity has drained and regressed in size, the catheter can be removed to allow for the anastomosis to heal.

Figure 2.3

Figure 2.4

Patients with late presentations of anastomotic leak are often managed with antibiotics, percutaneous drainage, and patience. Even in the presence of a demonstrable leak, percutaneous drainage alone may allow for complete resolution of the local sepsis and ultimate healing of the anastomosis. Unfortunately, this is commonly a slow process requiring patience, serial imaging, and repeat percutaneous interventions. Both covered stents and vacuum-assisted devices have been used with anecdotal success.[4],[5] Nutritional support, using the enteral route whenever possible, should not be neglected. Although patients are commonly restricted to clear liquids or nothing by mouth for prolonged intervals based on surgical custom, it is not at all clear that this enables healing of the anastomosis and may often exacerbate patient discomfort (physical and psychological) and diminish their ability to tolerate a prolonged recovery with repeated imaging studies and invasive interventions.