INTRODUCTION

Intestinal anastomoses have been routinely performed for over a century to restore bowel continuity after resective procedures. Despite the remarkable advances in surgical techniques with a corresponding improvement in clinical outcomes, anastomotic failure remains the most feared complication after intestinal surgery, with a significant impact on patients’ short- and long-term outcomes. In the continued effort to minimize the incidence of anastomotic-related complications, a variety of different methods to secure the ends of the bowel have been described, including a wide range of suturing techniques as well as stapling devices; nonetheless, anastomotic leak (AL) remains a relatively common occurrence in colorectal surgery, with rates reported between 1% and 19%, depending on the site of the anastomosis. The lack of uniformity in the definition of AL accounts, at least in part, for the wide range of incidences reported in the literature. Based on the published data, more distal anastomoses are more prone to developing a leak, with the lowest rates reported for small bowel anastomoses (1–2%), the highest for colorectal and coloanal anastomoses (5–19%), and ileoanal pouch anastomoses somewhere in between (4–7%) ( Table 26.1 ). ALs can manifest themselves at variable time intervals from the index procedure; while the majority are diagnosed within 2 weeks from surgery, up to 12% are detected after 30 days, and more than one-third after the patient has already left the hospital. Furthermore, the magnitude of the leak and its clinical implications can vary from incidental findings on cross-sectional imaging performed for other indications to contained abscesses with minimal systemic signs, and even diffuse peritonitis with overt sepsis in the case of a major anastomotic dehiscence. Although signs and symptoms such as fever, leukocytosis, emesis, and purulent discharge from the wound or drains can be suggestive of a leak in an acute setting, the diagnosis is usually confirmed by imaging, with computed tomography (CT) representing the modality of choice. Once a diagnosis is established, the treatment largely depends on the severity of the leak, its location, and the overall patient’s conditions. Traditionally, the management of AL has been operative, involving resection of the anastomosis and the creation of an enterostomy, however, over the last few decades, there has been a paradigmatic shift toward less invasive approaches, mostly thanks to earlier detection and the refinement of percutaneous image-guided drainage modalities. This chapter will discuss the definition, risk factors, diagnosis, and current treatment options of acute AL after colorectal surgery, with a particular focus on interventional radiology (IR) and surgical approaches; the management of chronic leaks as well as endoscopic techniques for the management of AL are addressed in separate chapters.

ETIOLOGY, CLINICAL PRESENTATION, AND DIAGNOSIS OF ANASTOMOTIC LEAKS

Restoration of normal anatomy after any general surgery procedure almost invariably involves the creation of an anastomosis to restore continuity between two distant hollow anatomical structures. Abnormal healing of such connections represents the major source of morbidity for patients in the postoperative period after vascular, hepatobiliary, and intestinal resective operations. Each anatomical site has a specific risk profile and poses peculiar challenges: esophageal ALs have a reported incidence above 50% in some series, with different implications based on the cervical versus thoracic location; pancreatic leak rates after pancreaticodudenectomy are as high as 20% even in the most recent series, with an associated 20% to 40% postoperative mortality, mostly due to the detrimental effects of the leaking pancreatic enzymes on the surrounding intra-abdominal structures. ^, In a retrospective study from the University of Vermont, Hyman et al. reported an overall 2.7% leak rate after colorectal resections. The incidence varied based on the surgical site: enteroenteric 0.9%, colocolic 2.5%, ileocolic 2.4%, colorectal 2.5%, ileorectal 23.3%, ileoanal 3.5%, and coloanal 0%. No statistically significant difference in the incidence of the leak was noted across sites except for a markedly increased risk with ileorectal anastomoses, although these results could have been affected by the relatively low number of events. What all the sites have in common is a general lack of uniformity in the definition of AL, as demonstrated in a systematic review from 2001 where a total of 56 different definitions were identified across 97 studies; specifically, 13 definitions were reported for upper gastrointestinal anastomoses, 14 for hepatopancreaticobiliary, and 29 for lower gastrointestinal. This observation was further substantiated by a survey conducted in the United Kingdom among 738 colorectal surgeons: while the vast majority (>90%) agreed that “extravasation of contrast on enema” and “fecal material seen in drains/from the wound” constituted an AL, “intra-abdominal sepsis requiring a laparotomy” defined a leak for only 70% of the participants, and only half agreed that collections identified on imaging are representative of a leak, no matter if treated with antibiotics or drainage by IR. Although the presence of an infectious process (abscess, peritonitis, and sepsis) is considered the distinguishing factor between a clinical and sub-clinical leak, whether a surgical site infection (SSI) represents the manifestation of an AL or an independent process often remains controversial and difficult to determine. As reported in a retrospective review of data collected by the Upstate New York Surgical Quality Initiative from 2010 to 2011, the rate of organ space SSI after colectomy was 3% versus a reported 4% AL rate; interestingly, only a quarter of the ALs were also coded as SSI, further highlighting the intrinsic difficulty in unequivocally categorizing the two entities. Colorectal resections have the highest reported SSI rates among elective general surgery procedures, ranging between 3% and 30%, which is explained by the clean-contaminated nature of the wound. ^, In a prospective study by Tang et al. on nearly 3000 patients undergoing colorectal resections with anastomosis, aimed at identifying risk factors associated with SSI, the authors concluded that it is important to distinguish SSI with or without an associated AL, as they have different risk factors, however, the only criteria to define a leak was the proximity of the abscess to the anastomosis.

Clinical judgment alone has been shown to be inaccurate at predicting patients at higher risk for AL, hence, many studies have been published aimed at identifying more objective risk factors in order to prevent the associated increase in length of hospital stay and mortality. ^, In a multicentric prospective national study from Europe on 3193 patients after oncological colon resection with primary anastomosis without diversion, the following patient, surgery, and hospital-related risk factors for AL were identified on multivariable analysis: body mass index (BMI), preoperative serum total proteins, male sex, ongoing anticoagulation, intraoperative complications, and the number of hospital beds. Another prospective study on 1598 patients undergoing 1639 procedures involving the creation of an anastomosis for benign or malignant conditions demonstrated that extraperitoneal anastomoses (within 10 cm from the anal canal) were at significantly higher risk of AL compared to intraperitoneal anastomoses (6.6% vs. 1.5%, respectively). Although never proven to reduce the incidence of leaks, the proximal diversion does decrease the magnitude and mitigates the consequences of ALs, and it is routinely performed in the setting of a high-risk anastomoses, particularly when below the peritoneal reflection. For colorectal anastomoses, transanal tube decompression has been often advocated but never widely adopted as a simple alternative to fecal diversion, despite its efficacy in reducing AL rates, as reported in several series as well as meta-analyses. ^, Across the literature, additional well-established risk factors for AL include inadequate blood supply, an American Society of Anesthesiologists (ASA) score of Grade III–V, smoking history, increased operative time, and emergent presentation. Conversely, there is a lack of conclusive data to support an association between leaks and any of these factors: neoadjuvant treatments, perioperative steroids and nonsteroidal anti-inflammatory drugs (NSAIDs), anastomotic technique (hand-sewn vs. stapled), oversewing of stapled anastomoses, minimally invasive approach, and surgical drains. Interestingly, in a recent, well-conducted study on a nationwide large sample of patients who underwent a right colectomy for colon cancer, the staple technique was associated with a 2-fold increase in ALs when compared to hand sewing (leak rate 5.4% vs. 2.4%, respectively; P = .004). In 2011, Dekker et al. proposed their colon leakage score (CLS) to aid surgeons in selecting patients who would benefit from fecal diversion, assessing 11 preoperative and intraoperative factors: a score of 11 out of 43 was identified as predictive of AL in left-sided colectomies, by adopting a 3% leak rate as a cut-off to justify a defunctioning stoma. Interestingly, two separate large retrospective studies in 2016 and 2017 identified the individual surgeon as an independent predictor of anastomotic failure, adding emphasis to the importance of the technical aspects involved in the creation of the anastomosis. ^, Furthermore, reoperative surgery per se represents a risk factor for anastomotic complications, as demonstrated in a retrospective study on 83 patients with CD, where a 5% leak rate was observed in patients with no previous bowel resection versus 17% in those with a history of previous enterectomy ( P < .005). With regard to preoperative interventions to minimize the risk of AL, a study by Kiran et al., subsequently replicated by the European Society of Coloproctology with comparable findings, demonstrated that mechanical bowel preparation with oral antibiotics halves the risk of leak after colorectal resections when compared to no bowel preparation. ^, It is common practice for colorectal surgeons to perform an air leak test via intraoperative endoscopy to check the integrity of the anastomosis after left-sided resections; however, the yield of this modality in predicting the fate of the anastomosis has been shown to be poor. Nonetheless, intraoperative endoscopy might represent a valuable tool in identifying high-risk anastomosis; any endoscopic findings suggesting vascular compromise should prompt revision or re-do of the anastomosis. In recent years, real-time fluorescence angiography to assess the perfusion of the bowel edges after intestinal resections has been proposed as a tool to decrease the incidence of anastomotic failure, with encouraging results in several published studies. It is reasonable to expect that this technology will gain popularity as surgeons become familiar with the technique, and larger studies will help define the significance of its findings and the corresponding management algorithm. Lastly, over the past decade, numerous studies have been published reporting experimental data suggesting a prominent role of the intestinal microbiome in the development of ALs; clinical data on this topic remains limited at this time, but the gut-microbiota interaction certainly represents an exciting field of investigation in intestinal surgery. Risk factors for AL are discussed in more detail in Chapters 4 and 7 .

Anastomotic leaks can have an insidious onset with nonspecific signs and symptoms characterized by low-grade fever, tachycardia, prolonged ileus, and worsening pain along with failure to thrive, making the differential diagnosis with postoperative abscesses uncertain and often arbitrary. Given the potential detrimental effects of ALs on patients’ short- and long-term outcomes, every effort should be made to identify early and promptly institute the appropriate treatments. With this goal in mind, in 2009 den Dulk et al. proposed a leakage score (DUtch LeaKage, DULK), subsequently simplified in 2013 to include fewer elements thus improving applicability in clinical practice (overall clinical deterioration, abdominal pain notlocalized at the wound, C-reactive protein level >250 µg/ml and respiratory rate ≥20/min); when three or more criteria are satisfied, the score provides a 20% sensitivity and 99% specificity, with a positive and negative predictive value of 56% and 93%, respectively. ^, In the aforementioned survey, the vast majority of surgeons (75%) selected serial clinical examination as the most sensitive clinical feature associated with AL; nonetheless, 80% of the respondents routinely confirm the diagnosis with a radiographic study: a CT with rectal contrast was the study of choice in 43% of the cases for a left-sided leak, while for a right-sided/small bowel anastomosis, a CT with oral contrast and a CT with intravenous contrast were the preferred modalities for 45% and 43% of the providers, respectively. Cross-sectional imaging appears to be superior to conventional contrast enema, which fails to identify ALs in up to 60% of the cases; however, contrast enema has been shown to identify leaks in sporadic instances where the suspicion for AL is high despite a negative CT, thus suggesting that the modalities should be considered complementary rather than an alternative. The most commonly accepted radiographic criteria for attributing an intra-abdominal abscess to an underlying AL include extravasation of enteric contrast, the presence of an obvious communication between the abscess cavity and the anastomosis, and “more than expected” perianastomotic gas. In a study by Power et al. comparing the CT findings in 73 patients undergoing imaging for workup of postoperative sepsis after bowel surgery with intestinal anastomosis, the only finding significantly associated with a clinically important AL (defined as a nonhealing anastomosis requiring IR or surgical intervention or extended antibiotic therapy) was peri-anastomotic loculated fluid containing air, while the presence of free air, distant collections, and free fluid did not show any association. In June 2020, a consensus-based definition of AL was published following the recommendations of 23 international experts; the proposed criteria are summarized in Table 26.2 . The widespread adoption of minimally invasive techniques, along with the implementation of enhanced recovery strategies, has significantly shortened the length of hospital stay after bowel resection, raising some concerns about the potential deleterious effects related to a delayed diagnosis of AL. In a study by Telem et al. published in 2010, while presentation after discharge caused a significant delay in diagnosis (12 vs. 6 days; P < .001), there were no differences in terms of interventions and outcomes, including long-term stoma rate.

MANAGEMENT OF ACUTE ANASTOMOTIC LEAKS

Once an AL leak is diagnosed, the institution of the appropriate treatments is of paramount importance in order to minimize the magnitude of its consequences. The management is largely dictated by the overall patient’s conditions (most importantly, the presence and severity of systemic sepsis), the anatomic location of the anastomosis (intra- vs. extra-peritoneal), the size and conformation of the potentially associated abscess, and the presence or absence of a fecal diversion at the time of the diagnosis. Based on these factors, patients are treated with a combination of medical therapies and endoscopic, IR, or surgical interventions as indicated; this chapter will focus on the latter 2 modalities, whereas endoscopic approaches will be discussed in a dedicated chapter 16 . The relatively low incidence along with the broad range of clinical presentations explain the lack of high-quality data to support the adoption of specific treatment strategies in each clinical scenario. In an effort to overcome these limitations, in 2008 the International Anastomotic Leak Study Group published an algorithm for the management of AL based on the recommendations from a panel of colorectal surgeons and interventional radiologists, which, to date, represents the best evidence-based reference for clinicians on this topic.

According to the classification proposed by the International Study Group of Rectal Cancer, Als can be graded based on their impact on the patient’s postoperative course: grade A leaks are sub-clinical and do not require treatment; the need for some type of intervention defines grade B leaks, while those requiring a return to the operating room are classified as grade C. When an AL manifests with diffuse peritonitis and signs of sepsis, including hemodynamic abnormalities, altered mental status, and end-organ dysfunction, prompt resuscitation along with intravenous broad-spectrum antibiotics followed by surgical exploration is indicated. The operative interventions are largely dictated by the intraoperative findings, both concerning the control of the intra-abdominal sepsis and the management of the anastomotic failure itself.

After a thorough washout and assessment of the abdominal cavity, control of sepsis is achieved by draining any obvious abscess and breaking any loculations. If phlegmonous changes (rather than a walled-off abscess cavity) are identified at the anastomotic site, exploration of the inflammatory mass can be more detrimental than beneficial, possibly causing significant bleeding and further disruption of a potential small leak, and hence is not recommended. In this case, the abdomen should be lavaged and surgical drains placed around the area, with a plan for a second look at laparotomy in 24 to 48 hours if necessary. Additionally, there is no consensus regarding the need for proximal diversion in these circumstances, especially when it would entail the creation of a very proximal small bowel enterostomy. The surgeon’s assessment of the risks and benefits to the patient as well as the patient’s general conditions determine the decision to divert in these circumstances.

The management of the anastomotic defect depends on its size, the perfusion of the bowel, and the quality of the surrounding tissues. Despite the lack of a standardized definition, most surgeons would categorize as a major defect any dehiscence greater than 1 cm or involving more than one-third of the circumference. If a major leak is encountered, or the ends of the bowel appear ischemic, the anastomosis should be resected with the creation of an end ostomy and a mucus fistula; this approach appears to be the safest and minimizes the chances of additional intra-abdominal complications postoperatively. For left-sided colonic resections or in those cases where the end of the distal bowel would not reach the abdominal wall due to limited length or mobility (i.e., thickened, inflamed mesentery), the stump can be closed (either with a stapler or with sutures) and left in the abdomen as a Hartmann pouch; leaving a surgical drain in its proximity and decompressing the bowel with a rectal tube in the immediate postoperative period can help minimize the chances of stump blowout or mitigate the consequences. For more proximal resections, an acceptable alternative is to exteriorize both ends of the intestine as a double-barrel stoma. Lastly, in selected cases (i.e., normal hemodynamics, ability to achieve healthy bowel ends approximating without tension, no evidence of active inflammatory bowel disease), a resection with reanastomosis can be a safe option to avoid the need for additional major procedures; most surgeons will perform a proximal diversion when this approach is elected.

Preservation of the anastomosis is a viable option in selected cases. When a minor leak (<1 cm defect or dehiscence involving less than one-third of the lumen) is encountered in a patient without hemodynamic compromise, a primary sutured repair may be performed, provided that the bowel wall is healthy enough to hold the stitches. Regardless of the quality of the repair, a diverting ostomy is created in this scenario. In a study from Germany, the recurrent leak rate after primary repair was significantly higher in patients who had not received an ileostomy at the time of the reoperation (50% vs. 12.5%). Some experts advocate on-table irrigation of the bowel via the distal limb of the ostomy to minimize the amount of contamination in cases of recurrent or ongoing leakage; however, there is no data in the literature to support the utility of this maneuver. Similarly, the placement of surgical drains near the anastomosis is commonly suggested but has never been demonstrated to be effective in improving outcomes.

Over the past two decades, laparoscopy has gained popularity in bowel surgery, and as surgeons have acquired familiarity with the technique, minimally invasive approaches have been adopted for the management of anastomotic complications. First, laparoscopy can represent a valuable tool in the diagnosis of anastomotic complications when the preoperative workup is inconclusive, thus potentially avoiding an unnecessary laparotomy. In one of the earliest studies on this topic, the authors observed that, when compared to patients undergoing relaparotomy for AL after open surgery, a patient who had a laparoscopic reoperation after an initial minimally invasive colorectal resection experienced a significantly quicker return of bowel function and resumption of diet, with a trend toward less 30-day morbidity, a shorter hospital stay, and a lower incisional hernia rate. This was a small, retrospective study with 15 and 10 patients in each group, respectively. Similar findings were reported in a more recent study comparing outcomes after laparoscopic management (61 patients) versus laparotomy (16 patients) for the treatment of AL after minimally invasive colorectal surgery. Additionally, in this series, the stoma-free survival rate was significantly higher in the laparoscopic group (43.8% vs. 80.5%, P < .001); this is in concordance with the data reported one year later in a study by a group from Denmark. With an observed conversion rate of only 8.2%, it is reasonable to conclude that minimally invasive surgery is safe and effective in selected patients undergoing reoperation for anastomotic complications, with several potential advantages in terms of more favorable short- and long-term outcomes. ^,

In the absence of systemic sepsis with profound hemodynamic changes, patients presenting with focal peritonitis and evidence of single or multiple abscesses can often be managed without returning to the operating room. Small abscesses (<4 cm) tend to respond to intravenous antibiotics alone, with third-generation cephalosporins as the class of choice. If technically feasible, some experts recommend aspiration of the abscess cavity under CT guidance with gentle lavage; additionally, the injection of fibrinolytic agents has been described as effective in breaking up locations, thus improving drainage. ^,

Larger abscesses, especially if >6 cm, are unlikely to respond to antibiosis alone and should be managed with IR drainage if the expertise is available and the anatomy is suitable. Image-guided percutaneous drainage has become the standard of care for the management of localized intra-abdominal infections, both primary and postoperative, with demonstrated safety and efficacy, along with the ability to temporize or avoid surgery in the majority of cases. ^, In fact, percutaneous drainage appears to be more effective when employed for the drainage of postoperative abscesses versus other settings, as reported in a retrospective study by Cinat et al. Overall primary success rates are reported at around 60% to 70%, rising well above 80% with repeated drainage; while the presence of a single abscess, as well as a longer time interval between surgery and the onset of symptoms, is predictive of success, a documented residual collection after the procedure represents a risk factor for failure. Benoist and colleagues, in their retrospective study on 73 patients undergoing CT-guided drainage of postoperative abscess, found that a size <5 cm and the absence of concurrent antibiotic therapy were independent predictors of treatment failure. When compared to open reoperative surgery, percutaneous drainage is associated with significantly less mortality (4.2% vs. 14.6%, P = .0001), as documented in a study by Politano et al. conducted on a cohort of patients undergoing general surgery procedures over 13 years; interestingly, this result held true on multivariable analysis even after controlling for patients’ acuity. In a series of 41 patients who developed an intra-abdominal abscess after bowel resection for Crohn’s disease, IR drainage proved effective in treating 95% of the patients, without evidence of a leak on contrast study and no need for further interventions; no major procedure-related complications were observed. In one study, the outcomes of 27 patients undergoing surgical treatment of AL were compared with those of 103 patients treated with percutaneous drainage; of note, this study included all types of gastrointestinal anastomoses, of which about half were colorectal. With the limitation of the retrospective nature, the percutaneous approach provided better outcomes in terms of significantly shorter hospital stay and decreased overall costs. The presence of an underlying fistula has been associated with failure of percutaneous drainage (defined as the eventual need for an operation); more than a limitation of the technique, this likely represents the natural evolution of the underlying anastomotic defect, and the ability to the defer surgical repair until complete resolution of the concomitant sepsis should not be underestimated. The management of fistulae secondary to ALs will be discussed in a separate chapter 30 .

Subclinical (or grade A) ALs, defined as those identified during routine testing or imaging performed for unrelated reasons, in general, do not require any treatment and have no significant impact on overall patients’ outcomes. ^, These are most commonly identified later on in the postoperative course during contrast enema prior to ostomy reversal, and the management will be addressed in a dedicated chapter of this book.

Some special considerations should be taken for the management of extraperitoneal anastomoses (low colorectal, coloanal, and ileoanal). While the same principles outlined above apply to the management of patients with overt sepsis and diffused peritonitis, it has been shown that among patients requiring reoperation for an AL, those with an extraperitoneal anastomosis are more likely to resolve with diversion and drainage, while intraperitoneal leaks more frequently require resection of the anastomosis. Extraperitoneal anastomotic complications presenting with an isolated pelvic abscess can be considered for transperineal or transanal approaches ( Fig. 26.1 ). Posterior leaks are more common in this location, and although treatment modalities are largely similar, anterior leaks are more prone to be associated with intraperitoneal contamination, and their treatment carries a higher risk of injuries to surrounding structures. Similarly to the intra-abdominal locations, smaller abscesses (<3–4 cm) are likely to resolve with antibiotic therapy alone. Larger abscesses should be approached with CT-guided drainage with a transgluteal or transabdominal route; if that is not technically feasible, the presence of direct communication between the anastomosis and the abscess cavity should be assessed with a contrast study. Abscesses that are in communication with the intestinal lumen via a leak in a low anastomosis are often amenable to examination under anesthesia with transanal drainage. While this approach is routinely used by colorectal surgeons, there is a substantial lack of uniformity with respect to the technical approach. Based on a recent survey, some experts favor creating a wide opening in the anastomosis to allow free drainage, while others recommend keeping the defect small, through which the head of a mushroom catheter is introduced in the abscess catheter with the end of the tube exteriorized through the anus and secured at this level. Before drain removal, imaging should be performed to document the resolution of the abscess and maintain the patency of the anastomosis. The detractors of this approach raise the concern that the drain itself might stent the leak open, thus preventing its closure. In a retrospective study from Canada on 37 patients with symptomatic AL after low anterior resection for cancer, 16 were managed with transanal drainage, 12 underwent a transabdominal procedure and, 9 received medical treatment only; the transanal approach was associated with the highest stoma-free survival (93%), with the most common complication being anastomotic stricture (33%). Similar results were observed in a small series from Japan, where an ostomy was avoided in 83.3% of the cases after transanal drainage, with the indication for diverting ileostomy in the remaining cases being fistula formation with adjacent organs. When both percutaneous and transanal drainage are viable options, the approaches appear to be equally effective, as reported by Kirat et al., with the latter potentially avoiding the risk of the development of an extra-sphincteric fistula, although rare. Alternative endoluminal approaches, including stenting and polyurethane vacuum sponge, will be discussed in a different section of this book. The above recommendations are summarized in a diagrammatic form in Fig. 26.2 .

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related posts:

Endoscopy-Associated Complications in Colorectal Diseases: Clinical Presentation and Diagnostic Evaluation Endoscopic Therapy of Fistulas from Chronic Anastomotic Leaks Endoscopic Management of Diverted or Bypassed Bowel Complications Corrective Surgery for Surgery-Associated Vaginal Fistula J-Pouch and Redo J-Pouch Stoma Complications: Conservative Management

Stay updated, free articles. Join our Telegram channel

Join