Neil Hyman
Presentation
A 64-year-old man with hypertension and mild chronic obstructive pulmonary disease is now 6 days after an elective sigmoid colectomy for recurrent diverticulitis. His initial postoperative course was unremarkable, but he developed confusion and agitation on the evening of postoperative day 4. His abdomen became more distended, but he did not have evidence of peritonitis. His morphine was stopped and a nasogastric tube was inserted. He seemed to improve somewhat initially, but then developed progressive tachypnea and a low-grade fever over the next 24 hours. He remained hemodynamically stable but required two fluid boluses to maintain adequate urine output. On postoperative day 6, his clinical status became acutely worse. At this point, he developed respiratory distress and tachycardia with a heart rate of 130. His lungs were clear with decreased breath sounds at the bases. He had marked abdominal distention and was tender across his lower abdomen with diffuse peritoneal signs.
Differential Diagnosis
Anastomotic leaks are perhaps the most dreaded complication of bowel resection. The reported incidence varies greatly based on definitions, indication for surgery, and anastomotic site. For a sigmoid colon resection, the anastomotic leak rate should be approximately 5%. Early postoperative leaks typically present in a dramatic fashion with severe abdominal pain, tachycardia, high fevers, and a rigid abdomen, often with hemodynamic instability. But leaks manifesting further along the postoperative course usually present far more insidiously, often with low-grade fever, ileus, and failure to progress. This latter presentation often mimics other infectious complications, such as pneumonia or urinary tract infection, or an early postoperative small bowel obstruction. Mental status changes and tachypnea are often the earliest signs of anastomotic leak, but these signs are nonspecific and could also suggest pulmonary embolism, pneumonia, atelectasis or an adverse drug reaction. In our patient scenario, the patient presents with symptoms that were initially nonspecific (fever, tachypnea, and tachycardia) and the differential diagnosis would include leak as well as other infectious complications.
But the patient eventually progresses to having a distended abdomen with peritonitis, which makes anastomotic leak the clear working diagnosis.
Workup
Patients with an early postoperative leak often have signs and symptoms of peritonitis and sepsis. In many circumstances, the diagnosis is clinically evident, and prompt return to the operating room is warranted. Too often, radiologic studies are wishfully obtained, when the indication for reexploration is clear. These studies should be limited to cases where there is some clinical uncertainty about the diagnosis. A water-soluble contrast enema will often rapidly confirm the presence of a leak in equivocal cases in the first few days after surgery (Figure 1). However, later in the postoperative period when the leaks that manifest themselves tend to be smaller and contained, this study becomes less reliable and a CT scan is preferable. In our experience, contrast enema failed to show a proven leak 60% of the time.
FIGURE 1 • Anastomotic leak: Contrast enema. Arrow indicates extravasation of water-soluble contrast from the colorectal anastomosis.
As described above, most leaks occurring later in the postoperative course are associated with a nonspecific presentation, and the diagnosis can actually be quite difficult to make. It must be recognized that a clouded sensorium or respiratory symptoms are often the presenting signs of a leak.
A chest x-ray and PE protocol chest CT scan may demonstrate pneumonia, lobar collapse, or pulmonary embolism. However, caution must be exercised in ascribing a downhill clinical course to subtle or minor abnormalities on these studies. Laboratory evaluation is generally nonspecific and often indicative of early multiorgan dysfunction. Plain abdominal films are not usually specific enough to make the diagnosis, but a large or an increasing amount of free intraperitoneal air is suggestive of a leak. The most helpful study in the setting of an occult leak is usually a CT scan of the abdomen and the pelvis with rectal contrast (Figure 2). We have found that CT demonstrates the diagnosis almost 90% of the time; but it must be acknowledged that there is broad overlap in CT findings in postoperative patients with or without a leak. Free air may be present up to 10 days in patients without a leak and loculated air up to 30 days after surgery.
FIGURE 2 • Anastomotic leak: CT scan. Arrow denotes the interface between extraluminal air and extravasated rectal contrast.
Diagnosis and Treatment
Leaks occurring within the first few days after surgery almost invariably require operative exploration. Patients should be properly prepared for surgery with fluid resuscitation, administration of broadspectrum antibiotics, and adequate intravenous access. A Foley catheter should be inserted as well as appropriate monitors such as a central line and an arterial line as needed. Of particular importance, patients should usually be marked for an intestinal stoma as this will often be required. Many ostomies created in the setting of a leak turn out to be permanent. Poor stoma siting (e.g., in a skin crease) can be a nightmare for the patient and be the difference in whether the patient can reassume self-care once they recover from surgery.
Contained leaks manifesting later on in the postoperative period in stable patients can often be managed nonoperatively with patience, antibiotics, and percutaneous drainage. Reoperation beyond 7 to 10 days after the initial procedure can be hazardous and may make things worse. Minor anastomotic disruptions may heal over time and obviate the need for any further operative treatment and an intestinal stoma.
Surgical Approach
The value and the importance of surgical planning when it comes to the management of an anastomotic leak cannot be overemphasized. Both the patient’s physiologic condition and the nature of the leak are crucial to medical decision making. As a rule of thumb, the longer the interval between reoperation and the initial surgery, the more difficult the procedure is likely to be. Surrounding viscera will attempt to seal the leak off, incorporating the anastomosis into an inflammatory mass surrounded by friable bowel with an associated serositis. There is usually an associated ileus making the bowel distended and tense. As such, mobilizing the anastomosis without tearing the bowel or extending the anastomotic defect can be a challenge. In this setting, precision of purpose (e.g., having a clearly defined plan) and technical efficiency (e.g., avoiding unnecessary dissection) are the key elements to success.
The patient is usually positioned supine, but the lithotomy position is preferred in patients who have a colorectal anastomosis, so it can be tested with betadine or inspected endoscopically. The previous incision is utilized and extended as necessary. Using primarily gentle blunt maneuvers, the bowel is pinched off the abdominal wall to facilitate access and exposure. The peritoneal cavity is irrigated and suctioned to clear as needed, and a specimen obtained for Gram stain and culture. The anastomosis is gently exposed and inspected. In a hemodynamically stable patient, right-sided anastomoses (e.g., ileocolic) can be resected and redone as long as the ends are not ischemic. In patients with extensive local sepsis and hemodynamic instability, an end-loop stoma should be created whenever possible (Figure 3). Although the surgeon’s primary goal is control of sepsis, exteriorizing the distal end of the bowel will save the patient another major laparotomy down the line to restore gastrointestinal continuity. Resection with anastomosis and proximal loop ileostomy is another alternative.
FIGURE 3 • End-loop stoma. The ileum and the proximal colon are exteriorized through the same hole to facilitate later closure without the need for laparotomy. A: The stapled ileum and the colon are delivered through the stoma site. B: A standard Brooke ileostomy is created. The antimesenteric edge of the colonic staple line is excised and sutured to the skin. C: The end-loop stoma is completed.
Left-sided anastomotic leaks usually require fecal diversion. Again, preoperative stoma marking is vital, especially in obese patients. The anastomosis is inspected and the patient’s hemodynamic status reviewed. Stable patients with a very small leak in an otherwise intact anastomosis, no ischemia, and minimal local sepsis can be treated with repair, omentoplasty, and loop ileostomy. Otherwise, ischemic anastomoses, major disruptions, or those associated with systemic sepsis should usually have the anastomosis resected and a Hartmann procedure performed. Resection with anastomosis and loop ileostomy is another option for stable patients. Most patients with leaking low colorectal anastomoses are best treated with fecal diversion and drainage. The anastomosis is usually deep in the pelvis and attempts at exposure will usually worsen the defect. Endoscopic visualization to assure the anastomosis is not ischemic and to visualize the defect can provide the needed information for decision making without worsening the problem. A presacral drain is placed and the omentum is mobilized for placement over the anastomosis (or at least into the pelvis).
Common pitfalls are doing too much or too little at surgery. As noted above, the dissection should be restricted and focused to only what is needed to examine the anastomosis and washout the contaminated fluids. On the other hand, the surgeon must acknowledge the nature of the problem and treat the leak adequately. Anastomotic leaks exact a major emotional toll on the patient, their family, and the operating surgeon. It can be tempting just to suture the hole closed and hope, in lieu of creating an intestinal stoma (“perfuming the pig”). But if the patient had a leak under “ideal” or elective conditions, it is even more likely they will develop another leak in an emergency situation with local and systemic signs of sepsis. Patients who have already suffered a leak often cannot readily tolerate another septic insult. When in doubt, it is a good time to “phone a friend” and speak with a trusted and experienced colleague (Table 1).
TABLE 1. Key Technical Steps and Potential Pitfalls to Surgical Management of an Anastomotic Leak
Special Intraoperative Considerations
Special mention is made of the obese patient with a leak. Although the principles of management are the same, there are a few additional considerations. First, most obese patients have a much thinner upper than lower abdomen wall, where most of the pannus resides. Bringing out an ostomy in the lower abdomen without undue tension or ischemia may be difficult or impossible. As such, strong consideration should be given to placing the stoma in the right (or left) upper quadrant. Further, a loop ileostomy is possible in almost any patient, whereas creating a colostomy in an obese patient with a leak may be a nightmare. The bowel is friable; there is marked distension from the associated ileus; and the mesentery may be rigid, inflamed, and unyielding. This is one scenario where either repair with diverting loop ileostomy or resection with anastomosis and proximal loop ileostomy may be the only option.
Postoperative Management
An anastomotic leak is associated with a mortality rate of 10% to 15%. The first step in postoperative management is to support the patient through the sepsis as needed with inotropes, ventilatory support, and modern intensive care. Unfortunately, patients who leak commonly required prolonged hospitalization, further reoperations, and aggressive rehabilitation. Antibiotics should be administered in a goal-directed manner (e.g., until afebrile with normal white blood cell count and ileus resolves) instead of continued indefinitely. Nutritional support, good enterostomal therapy, and careful wound management are usually cornerstones of supportive care. Patients who have been reoperated for a leak are at high risk for further complications such as wound infection or intra-abdominal abscess and often require postoperative imaging studies and percutaneous drainage of residual infected collections.
Case Conclusion
In our patient, he was taken to the operating room without radiographic studies, since he had a clear clinical presentation of an anastomotic leak and was severely ill. He was given additional intravenous fluid and broad-spectrum antibiotics while readying an operating room. He was placed in lithotomy position and his midline incision was reopened. After exposing the pelvis and irrigating the abdomen, a small disruption in the lateral colorectal anastomosis was found. The colon and the rectum appeared healthy and viable. The area of anastomotic dehiscence was reinforced with lembert sutures, and omentum was placed over the reinforced area. A diverting loop ileostomy was performed to allow healing of the anastomosis. The patient recovered well and was discharged home after 10 days. At 2 months, a contrast enema showed healing of his anastomosis without leak or stricture. His ileostomy was taken down electively 3 months after his initial surgery.
TAKE HOME POINTS
· Anastomotic leaks are a devastating complication of intestinal surgery.
· Many early leaks are readily diagnosed clinically (not radiographically) and require prompt reoperation and treatment.
· Most leaks actually occur 6 days or more after surgery, and the diagnosis can be very challenging as signs and symptoms are nonspecific.
· CT scan is usually the imaging modality of choice for late leaks.
· Reoperation requires thoughtful planning and a goal-directed approach.
SUGGESTED READINGS
Bruce J, Krukowski ZH, Al-Khairy G, et al. Systematic review of the definition and management of anastomotic leak after gastrointestinal surgery. Br J Surg. 2001;88:1157–1168.
Hyman N, Manchester TL, Osler T, et al. Anastomotic leaks after intestinal anastomosis: its later than you think. Ann Surg. 2007;245:254–258.
Hyman N, Osler T, Cataldo P, et al. Anastomotic leaks after bowel resection: what does peer review teach us about the relationship to postoperative mortality? J Am Coll Surg. 2009:208:48–52.
Power N, Atri M, Ryan S, et al. CT assessment of anastomotic bowel leak. Clin Radiol. 2007;62:37–42.