Risk Factors for Anastomotic Leak, Consideration for Proximal Diversion, and Appropriate Use of Drains

Abstract

Anastomotic leaks are a major source of morbidity after colorectal surgery. There is a myriad of risk factors that may contribute to anastomotic leaks. These risk factors can be categorized as modifiable, nonmodifiable, and intraoperative factors. Identification of these risk factors allows for preoperative optimization that may minimize the risk of anastomotic leak. Knowledge of such high-risk features may also affect intraoperative decision-making regarding the creation of an anastomosis, consideration for proximal diversion, or placement of a drain. A thorough understanding of the interplay between risk factors, indications for proximal diversion, and utility of drain placement is imperative for colorectal surgeons.

Despite optimizations in surgical technique and patient care over the years, the reported anastomotic leak rates in colorectal surgery remain between 0.5 and 30%. ^{1 2 3 4} Anastomotic leaks have potentially grave clinical implications for patient with both short- and long-term consequences. There are a variety of risk factors that contribute to anastomotic leaks which have been researched extensively in the literature. Identifying patient and intraoperative risk factors may affect surgical outcomes and prevent devastating consequences. These risk factors can be divided into nonmodifiable, modifiable, and intraoperative risk factors. Nonmodifiable risk factors cannot be altered, but an awareness of these elements may influence intraoperative decision-making. Conversely, modifiable risk factors offer an opportunity to intervene prior to surgery to improve clinical outcomes. Lastly, intraoperative factors, such as significant blood loss and fecal contamination, can also increase the risk of anastomotic complications which may not be avoidable.

Nonmodifiable Risk Factors

Nonmodifiable risk factors can be determined during the first clinical encounter with the patient. Based on history and physical examination alone, several of these risk factors can be ascertained.

Male Sex

One significant nonmodifiable risk factor associated with increased anastomotic leaks is male gender. ^{2 5} Several studies have demonstrated that males have higher rates of anastomotic leak compared with females, particularly after low anterior resection. The underlying theory for increased leak rate in men is the technical challenges that are encountered when operating in the narrower male pelvis. ⁶

Comorbid Conditions

A thorough evaluation of patient comorbid conditions must be performed, and frequently many of these conditions will not be amenable to further optimization. Comorbidities, such as chronic pulmonary obstructive disease, ischemic heart disease, diabetes mellitus, and preexisting vascular and renal disease are important factors to consider due to their association with increased leak rates. ^{3 7 8 9} The American Society of Anesthesiologist (ASA) physical status classification system which is used to assess a patient's fitness based on medical comorbidities prior to surgery is also an important consideration. ⁷ Patients with higher ASA scores are at increased risk for perioperative complications. In a study by Choi el at, higher anastomotic leak rates were noted for patients with ASA scores III to V compared with patients with ASA scores I and II. ¹⁰

Radiation Therapy

Radiation therapy, which can either present in the form of a patient having received therapy for another malignancy (cervical or prostate) or in the neoadjuvant setting in rectal cancer, should be taken into consideration when evaluating patients for low pelvic anastomoses. Neoadjuvant chemoradiotherapy is the preferred treatment for locally advanced rectal cancer and has been shown to decrease local recurrence and downstage tumors to allow for an increased rate of sphincter sparing surgery. ¹¹ There are conflicting reports regarding the risk of anastomotic leak in patients receiving neoadjuvant chemoradiotherapy or radiotherapy. Several retrospective studies involving patients given preoperative radiotherapy with or without chemotherapy demonstrated an association with increased anastomotic leak rates, particularly in patients undergoing low anterior resection. ^{12 13 14 15} However, larger retrospective studies, including the Medical Research Council (MRC) CR07 and Dutch Total Mesorectal Excision (TME) trials, have reported no association with neoadjuvant radiotherapy or chemotherapy and anastomotic leak. ^{16 17 18} While radiation therapy has not been confirmed to be a risk factor for leak, it is generally ill advised to perform an anastomosis between two radiated segments of bowel. When one segment of bowel has been radiated, it is argued that proximal diversion should be considered. Patients who have a history of pelvic radiation for prostate or cervical cancer may also have increased anastomotic leak rates and typically these patients have received radiation doses much higher than that used to treat rectal cancer. ¹⁹

Level of Anastomosis

The most consistently reported risk factor for anastomotic leak reported in the literature is the tumor or disease location and level of the anastomosis. ^{20 21} Distal anastomoses, between 5 and 7 cm proximal to the anal verge, are at the highest risk for anastomotic leaks. ^{20 21 22} The proposed explanations provided for the increased leak rate include lack of peritoneal covering in distal rectum, decreased vascularity to distal rectal stump, and increased bacterial burden. In a study by Rullier et al, anastomoses less than 5 cm from the anal verge had 6.5 times higher anastomotic leak rates when compared with those above 5 cm. ²⁰ Furthermore, Buchs et al reported that colorectal anastomoses were seven times more likely to leak compared with ileocolic anastomoses. ⁴

With improvements in surgical technique and equipment over the years, more sphincter sparing surgeries are being performed for rectal cancer. In a meta-analysis of 13 studies by Gu and Wu, the use of protective stomas is favored to reduce anastomotic leak rates seen in patients undergoing low anterior resections with ultralow colorectal or coloanal anastomoses. ²³ Restorative proctocolectomy with ileal pouch anal anastomosis (IPAA) is the procedure of choice for ulcerative colitis and familial adenomatous polyposis patients. ²⁴ IPAA surgery is at risk for pelvic sepsis which includes anastomotic leak, abscess, and fistula which can negatively affect pouch function. ²⁵ These data suggest that distal anastomotic location is an important consideration when deciding whether an anastomosis should be performed altogether or if the patient should be counseled regarding the need for temporary diversion.

Modifiable Risk Factors

Modifiable risk factors must be considered in patients undergoing colorectal surgery. In the elective setting, identifying and providing modifications to these factors may help mitigate the risk of an anastomotic leak.

Smoking and Alcohol Consumption

Excess alcohol consumption and smoking are both associated with increased risk. In a study by Sørensen et al, patients who abused alcohol (>35 drinks per week) or smoked tobacco had a relative risk of 7.18 (95% confidence interval [CI]: 1.2–43.01) and 3.18 (95% CI: 1.44–7.00), respectively, when compared with patients who abstained from alcohol or smoking. ²⁶

Obesity and Malnutrition

Obesity is a known risk factor for postoperative wound infection; however, the data regarding obesity and anastomotic leaks are conflicting. ^{7 27 28} Several studies have demonstrated that a body mass index (BMI) greater than 30 kg/m ² results in increased anastomotic leak rates in patients undergoing colorectal surgery, while other studies show no difference based on weight. ^{27 29 30 31} In a retrospective review of obese patients undergoing low colorectal and coloanal anastomoses, there was a leak rate of 33% compared with 15% in nonobese patients ( p  = 0.03). ²⁰ Malnutrition, weight loss greater than 10% in last 6 months and serum albumin level of <3.5 g/dL have a role in anastomotic leak rates. ^{3 32 33} Mäkelä et al reported that patients with significant weight loss and malnutrition had increased leak rates especially when undergoing left-sided anastomoses. ³³ In the elective setting, the above risk factors may be mitigated with reduction in alcohol consumption, smoking cessation, and weight loss programs. Preoperative maximization of nutrition, preferably with enteral nutrition, for malnourished patients is important to reduce risk. Unless patients are severely malnourished, there is no evidence to support the use of preoperative parenteral nutrition. ³⁴

Immunosuppressive Medications

There are some medications, such as corticosteroids and immunosuppressive agents, that may be associated with increased leak rates. Numerous studies have demonstrated an association with chronic steroid use and anastomotic leaks. ^{3 5 35 36} However, other studies involving patients with inflammatory bowel disease on chronic steroids within 8 weeks of surgery, immunosuppressants (azathioprine or 6-mercaptopurine), and biologic agents (infliximab or adalimumab) have demonstrated no increased risk. ^{37 38} The evidence for antitumor necrosis factor (TNF) agents, such as infliximab, with regard to anastomotic leaks is conflicting. ^{39 40} Immunosuppressive drugs, such as cyclosporine, tacrolimus, and mycophenolate mofetil, may have an effect on wound healing and therefore anastomotic healing. ^{41 42 43} Avastin (bevacizumab), a humanized monoclonal antibody targeting vascular endothelial growth factor receptor approved for the treatment of advanced colorectal cancer, is associated with increased anastomotic leak rates. ⁴⁴ Evidence for tapering of corticosteroids and discontinuation of bevacizumab 6 weeks before surgery is supported, though cessation of anti-TNF therapy is not routinely performed. ^{36 45} Typically, a wash-out period of 4 weeks is allowed, and surgical resection is performed accordingly.

Intraoperative Risk Factors

There are several intraoperative factors that can contribute to the development of an anastomotic leak, and frequently these factors are not predicable. When they do occur, thought should be given to the safety of creating an anastomosis, consideration for proximal diversion, and the complexity of future stoma reversal. Excessive intraoperative blood loss is a predictor of anastomotic leak. Several studies have demonstrated that blood loss greater than 100 mL and/or patients who received multiple transfusions had increased leak rates. ^{33 46 47} Patients operated on under emergent conditions have higher leak rates, as well as increased risk of death, from anastomotic leaks. ⁴⁸ Long operative times, greater than 3 hours, and fecal contamination have increased leak rates. ^{4 35} Unexpected intraoperative events, such as stapling device malfunction, ischemia to proximal colon, tumor perforation, and injury to other organs or blood vessels, can also contribute. ^{14 49} Trencheva et al reported that patients that had intraoperative complications were 4.1 times more likely to have postoperative leaks. ⁴⁹ These intraoperative considerations may be unavoidable but should certainly be considered when creating an anastomosis.

Consideration for Proximal Diversion

The role of proximal diversion of fecal material in colorectal surgery is controversial, though most surgeons agree that proximal diversion should not be routinely performed for every anastomosis. Temporary diversion is considered for patients with low colorectal anastomoses (<5 cm from the anal verge), including coloanal and ileal pouch anal anastomoses, as well as difficult resections, especially in malnourished or male patients. ^{20 50} Most diversions are typically with a loop ileostomy or transverse colostomy. The best choice for diversion is debated; however, most surgeons favor loop ileostomy due to the lower incidence of prolapse, sepsis, and parastomal hernia along with fewer complications upon reversal. ^{51 52} Proximal diversion will likely not prevent anastomotic leaks but diverted patients have a decrease in adverse effects such as peritonitis, sepsis, need for reoperation, and lower mortality rates. ^{14 53 54 55} In a randomized multicenter trial by Matthiessen et al, patients who had a defunctioning loop stoma had decreased rates of symptomatic anastomotic leakage after low anterior resections for rectal cancer. ⁵⁶ In a meta-analysis by Wu et al, patients undergoing low anterior resections for rectal cancer, proximal diversion decreased anastomotic leaks and the need for reoperation. ⁵⁷ Therefore, proximal diversion should be a consideration when performing low anastomoses in high-risk patients.

Temporary diversion is not without potential complications. Patients may be subjected to stoma related complications, such as retraction, prolapse, peristomal skin problems, as well as electrolyte abnormalities and dehydration. ⁵⁷ In addition, there is morbidity associated with stoma closure. ^{58 59} A multicenter statewide database study identified a 9% major complication rate and 19% minor complication rate following loop ileostomy reversal. ⁵⁸ Similarly, a meta-analysis by Tan et al, of patients undergoing stoma closure after low anterior resection for rectal cancer had complications rates as high as 19.8%. ⁵⁵ The decision for proximal diversion must be carefully considered and decided on a case-by-case basis focusing on the impact of a leak on the patient versus the morbidity of a stoma and its associated complications.

Appropriate Use of Drains

The prophylactic use of drains is also debated and without clear consensus in the literature. The use of drains was first introduced by Theodore Billroth in 1877. ⁶⁰ Historically, drains were thought to reduce the incidence of anastomotic leaks by reducing the bacterial load surrounding the anastomotic site. While some studies have shown that drains may be preventative in leaks, others report the harmful effects of drains. Although most surgeons have abandoned the use of prophylactic drains for intraperitoneal anastomoses, pelvic drains for colorectal or coloanal anastomoses are utilized and may be beneficial. Pelvic drains allow for the drainage of contaminated fluid or blood, potentially preventing the formation of an abscess or hematoma, as well as allowing for early detection of anastomotic leaks. ^{61 62} A meta-analysis by Rondelli et al examining eight studies with 2,277 patients undergoing colorectal anastomoses showed that pelvic drainage reduced anastomotic leak and reoperative rates. ⁶³ Drains were also shown to be protective against anastomotic leaks in another meta-analysis, looking at risk factors after laparoscopic anterior resection for rectal cancer. ⁴⁶ On the other hand, there are many studies which have shown no benefit to drain utilization and potential harmful effects. A Cochrane review by Jesus et al examined three randomized controlled trials with 908 patients, 454 allocated to drainage and 454 to no drainage after colorectal anastomosis. This review showed no significant difference in anastomotic dehiscence, reoperation, or wound infection when drains were used. ⁶⁴ More recently, a meta-analysis by Zhang et al with 11 randomized controlled trials examining 1,803 patients who showed no significant difference in patients with drains versus no drains in terms of anastomotic leakage rates (clinical and radiographic), wound infection, reoperation, and mortality. ⁶⁵ The GRECCAR 5 trial, a prospective randomized multicenter study, demonstrated that pelvic drainage after rectal cancer surgery did not result in a decrease in postoperative pelvis sepsis which was defined as anastomotic leakage, pelvic abscess, or peritonitis. ⁶⁶ Furthermore, some studies suggest that drains not only have no effect on prevention of anastomotic leaks but may lead to an increase in other complications, such as small bowel obstruction. ⁶⁷ Therefore, drains are not beneficial for intraperitoneal anastomoses and should be considered only selectively during creation of low anastomoses to prevent accumulation of fluid or after difficult surgeries at the surgeon's discretion.

Conclusion

Anastomotic leaks are a dreaded consequence of colorectal surgery. A thorough knowledge of the risk factors may be helpful and affect surgical decision-making. While some of the risk factors (smoking and malnutrition) may be modified, others (emergent surgery and level of the anastomosis) are inherent and unavoidable part of colorectal surgery. Understanding and thereby mitigating patient risk factors when possible may decrease or prevent anastomotic leaks. Proximal diversion and drain placement are debated but may be a consideration in the highest risk patients.