Overuse of Proximal Fecal Diversion in Colorectal Surgery

Abstract

Many surgeons tend to overuse proximal fecal diversion in the setting of colonic surgery. The decision to proximally divert an anastomosis should be made with careful consideration of the risks and benefits of proximal diversion. Proximal diversion does not decrease the rate of anastomotic leak, but it does decrease the severity of leaks. Anastomotic height for low pelvic anastomoses, hemodynamic instability, steroid use, male sex, obesity, malnutrition, smoking, and alcohol abuse increase the rate of anastomotic leak. Biologics, most immunosuppressive agents, unprepped colons, and radiation for rectal cancer do not contribute to increased rates of anastomotic leak.

Proximal fecal diversion creates additional potential morbidity, higher rates of readmission, and need for a subsequent hospitalization and operation for reversal. Additionally, diverted patients have higher rates of anastomotic stricture and delayed recognition of chronic leaks. These downsides to diversion must be weighed with a patient's perceived ability to handle the physiologic stress and consequences of a severe leak if reoperation is required. When trying to determine which patients can handle a leak, the modified frailty index can help to objectively determine a patient's risk for increased rate of morbidity and failure to rescue in the event of a leak.

While proximal diversion is still warranted in many cases, we find that certain clinical scenarios often lead to overuse of proximal diversion. The old surgical adage “If you are considering diverting, you should probably do it” should be tempered by an understanding of the risk and benefits of diversion.

The decision to divert a colonic anastomosis is not simple—it is a multifactorial decision simultaneously weighing risks and benefits of proximal diversion with perceived individualized rate of anastomotic leak. No matter how objective we try to make the decision to divert, some degree of subjectivity (clinical judgment) will always come into play during this decision-making process. In this article, we will discuss the rationale behind diversion, the consequences of a leak, the downside of diversion, risk factors for leak, tools to help determine a patient's ability to handle a leak, and specific scenarios where surgeons tend to overuse proximal diversion. We hope to provide a logical framework to consider proximal diversion while also hopefully decreasing unnecessary diversion.

Why do we Proximally Divert?

The overwhelming body of evidence suggests that proximal fecal diversion does not significantly decrease the rate of anastomotic leak, but it does decrease the severity of leaks and the subsequent complications in those patients who leak. The one exception is coloanal anastomoses, where the rate of leak is significantly lower in patients who are diverted. ¹ The treatment of anastomotic leaks varies based on the severity of their presentation. A minor clinical leak is managed by interventions other than reoperation including antibiotics and percutaneous drainage, while a major clinical leak requires reoperation. ² When patients leak with proximal diversion, they are much less likely to develop a major clinical leak requiring operative intervention and septic complications compared with patients who are not diverted. ³

What are the Long-Term Consequences of a Leak?

The short-term consequences of a leak are often the focus of outcome studies; however, the long-term consequences of anastomotic leak are often underappreciated. Patients with a colorectal leak have worse long-term function, higher rates of low anterior resection syndrome, and decreased quality of life. ^{4 5} Not unexpectedly these long-term complications result in higher rates of permanent stomas. Additionally, long-term oncologic outcomes for potentially curable disease are worse for patients who leak. ⁶

Why not Divert Everyone?

If diversion decreases the rate of severe leaks requiring reoperation, why don't surgeons divert everyone? For those who have done any significant amount of colorectal surgery, it comes as no surprise that diversion adds a significant amount of potential morbidity to a patient and an almost certain additional surgery and hospitalization (unless the diverting stoma becomes permanent). Among the highest drivers for readmission rates following colorectal surgery are issues related to management of ostomies. ⁷ Diverting ileostomies are preferred by most (including our practice) over diverting colostomies due to the decreased technical difficulty of the subsequent stoma takedown as well as decreased risk for injury to the blood supply of the conduit at takedown. However, ileostomies have higher rates of dehydration, electrolyte derangement, and pouching issues. ⁸ These issues lead to higher rates of readmission for diverted patients. The need for an additional operation and higher rates of ostomy-related complications lead to higher overall cost for patients and systems compared with a primary anastomosis (assuming no leak). ⁹

Two additional more subtle downsides to fecal diversion in colorectal surgery are higher rates of anastomotic stricture and higher rates of delayed recognition and treatment of leaks. ^{3 10} Anastomotic strictures will often respond to gentle dilation either digitally or with an endoscopic balloon, but they generally require additional procedures under sedation. Delayed recognition of a leak is a much more difficult issue to address. We previously mentioned that diversion does not necessarily decrease the rate of leak but decreases the severity of a leak. If a leak is found at the time of proposed reversal, this is often at least 2 months after the initial operation depending on the individual practice of the surgeon either on a contrast enema or after reversal. By this time, a chronic sinus has had time to form which makes treatment of the leak and potential chronic infection much more difficult to treat than if the leak was initially recognized and redone or repaired. ¹¹

Who Leaks (and Who Doesn't Leak)?

If we are trying to decrease the severity of leaks with diversion, it makes sense to divert patients who are at the highest risk for leak. More specifically, what are the risk factors for developing a leak? Each individual risk factor has been extensively studied, and this is not an all-inclusive list, rather a broad overview to keep in mind for practical use. Specific situations and certain medications will always require individualized care.

Perhaps the most well-published risk factor for leak is anastomotic distance from the anal verge. While studies have varied over the years, it seems that less than 8 cm is higher risk and less than 5 cm is highest risk for leak. ^{12 13} Coloanal anastomoses are highest risk and should be almost uniformly diverted in our opinion.

Two additional well-published risk factors are male sex and obesity. ¹² These are grouped together because the actual increased risk is likely a surrogate of technical difficulty. The male pelvis (or any narrow pelvis) can significantly increase the technical difficulty of a proctectomy. One way that we preoperatively anticipate a difficult dissection is to measure the distance between the narrowest part of the bony pelvis on cross-sectional imaging. In our experience, if the measured diameter is less than 10 cm, many surgeons will have a difficult time with visualization or palpation due to limitations of hand size. Obesity can often lead to more difficult visualization of anatomy and proper planes, leading to similar technical difficulties.

Primary anastomoses performed during hemodynamic instability are more likely to leak, as evidenced by trauma and acute care surgery literature. ¹⁴ However, in the setting of hemodynamic instability, often the safest option for unstable patients is an end ostomy (as opposed to anastomosis and diversion) to decrease time under general anesthesia.

Steroid use, malnutrition, smoking, and alcohol abuse are all risk factors known to increase anastomotic leak rate. ^{2 15} Steroid use in the perioperative period has been shown to increase overall complications, infectious complications, and rates of anastomotic leak in the setting of colorectal anastomoses. ¹⁶ Albumin levels below a relatively high threshold of 3.0 to 3.5 g/dL have been associated with increased rates of anastomotic leak. ^{17 18} Active smoking or alcohol abuse have similarly shown increased rates of anastomotic leak. ¹⁵

Contrary to popular belief, evidence would suggest that use of biologics, most immunosuppressive agents (aside from corticosteroids), unprepped colons, and radiation for rectal cancer do not increase the rate of anastomotic leak. ¹⁵ While these scenarios have often mandated diversion for many surgeons over the years, the most recent literature does not show an increased rate of anastomotic leak with these individual risk factors.

Who can Handle a Leak?

This is perhaps the most significant question for surgeons to ask themselves aside from the known risk factors for leak described previously. If a patient leaks, the degree of clinical significance will often correlate with the required treatment. Not all leaks require reoperation, and minor leaks are treated with antibiotics, percutaneous drains, time, or more novel therapies such as endosponge therapy. However, for major leaks, a reoperation will be necessary with either an end ostomy or revision and diversion.

If the worst possible outcome of a leak leaves a patient with an ostomy, why not wait and see who leaks before giving a patient an ostomy? Some patients are certainly able to handle the physiologic stress of a leak better than others, but identifying those patients preoperatively is more difficult. Extreme scenarios often make the decision to divert easy. For example, if a patient is in septic shock on high-dose norepinephrine, a primary anastomosis is out of the question. The additional septic complication of a leak could prove to be fatal.

In all but the most obvious clinical scenarios, clinical judgment comes into play when deciding whether to divert, and the “eyeball-test” will vary from surgeon to surgeon. Clinical judgment is a nice way to describe a subjective assessment, but there are adjunct objective data points to consider which could assist the surgeon with this critical decision making.

Over the past few years, the frailty index has proven to correlate well with postoperative surgical outcomes. ^{19 20} More recently, research within colorectal literature has shown the modified frailty index (m-FI) to correlate well with rate of anastomotic leak and subsequent complications. ²¹ Briefly, the m-FI consists of five items: diabetes, congestive heart failure, hypertension, chronic obstructive pulmonary disease/pneumonia, and partially or totally dependent functional status prior to surgery. Each item contributes one point to the overall m-FI score for a patient. In a recent American College of Surgeons National Surgical Quality Improvement Program database study, increased m-FI correlates with increased rate of anastomotic leak as well as failure to rescue, septic shock, and overall morbidity in patients who leaked. For patients who leaked, an m-FI score of 2 corresponded to an odds ratio of 2.14 for overall morbidity and 2.30 for failure to rescue compared with patients with an m-FI score of 0. An m-FI score of 3 corresponded to an odds ratio of 3.99 for overall morbidity and 6.35 for failure to rescue compared with patients with an m-FI score of 0. ²¹ Our interpretation of these results leads us to strongly consider diversion for patients with an m-FI of 2 and almost certainly divert patients with an m-FI of 3 or higher if anastomotic integrity is in question or additional risk factors are present.

Clinical Scenarios Where Surgeons Overuse Proximal Diversion

Left-Sided and High Rectal Anastomoses

Above 8 cm from anal verge should have a relatively low rate of leak (3–5%). ^{13 15} Radiation for rectal cancer does not mandate diversion, as studies have not shown preoperative radiation for rectal cancer to be an independent risk factor for anastomotic leak. ¹⁵

Right-Sided Anastomosis (Especially in Inflammatory Bowel Disease)

The leak rate for a right-sided (ileocolic) anastomosis should approach 1 to 3% in experienced hands. ¹⁵ Even in patients with increased risk factors for a leak, the functional consequences of a right-sided leak are not the same as a left-sided or pelvic leak. Diversion for an ileocolic anastomosis should be reserved for only the most frail, malnourished, or unstable patients.

Small Bowel Anastomosis

Given the low rate of small bowel anastomotic leak of approximately 1%, if there is ever a consideration for proximal diversion of a small bowel anastomosis, the site of proposed anastomosis should likely be brought up as an ostomy. ¹⁵ Again, this should be reserved only for the most frail, malnourished, or unstable patients.

Large Bowel Obstruction

Many surgeons are concerned about a primary anastomosis in the setting of an unprepped colon (no increased rate of leak) or size mismatch that often results from a large bowel obstruction. Colonic stents and on-table lavage are two potential options to mitigate these concerns, and a size mismatch can often be addressed with a Baker-type (side-to-end) anastomosis. ^{22 23} The decision to divert should depend on tissue integrity and additional risk factors previously described.

Biologic Use

As discussed above, most studies do not show an increased rate of anastomotic leak when operating on patients treated with biologics. ¹⁵ Dr. Kann will discuss additional anastomotic considerations for Crohn's disease in much greater detail within this issue.

Transplant Immunosuppression

While each medication should be examined on a case-by-case basis, the few studies examining outcomes of colorectal surgery following solid organ transplant show no increased rate of anastomotic complications while maintaining immunosuppressive regimens throughout the perioperative period. ^{24 25 26 27} The risk of graft rejection or failure should be weighed against any perceived increased rate of anastomotic leak in consultation with transplant physicians to guide alterations in immunosuppressive regimens.

Diverticulitis

We would suggest proximal diversion in the setting of feculent peritonitis (regardless of additional risk factors). Purulent peritonitis or localized inflammation does not mandate diversion. For anything aside from Hinchey IV diverticular disease, individual patient risk factors should be considered to guide the decision to divert. ²⁸ More detailed anastomotic considerations in the setting of diverticulitis will be addressed by Dr. Hawkins and Dr. McChesney in this issue.

Conclusion

The decision to proximally divert an anastomosis should be made with careful consideration of the risks and benefits of proximal diversion while also considering each patient's individual risk factors for anastomotic leak and physiologic reserve to handle a leak. Anastomotic height for low pelvic anastomoses, hemodynamic instability, steroid use, male sex, obesity, malnutrition, smoking, and alcohol abuse increase the rate of anastomotic leak. Biologics, most immunosuppressive agents, unprepped colons, and radiation for rectal cancer do not contribute to increased rates of anastomotic leak.

The old surgical adage “If you are considering diverting, you should probably do it” should be tempered by an understanding of the risk and benefits of diversion. Proximal fecal diversion creates additional potential morbidity, higher rates of readmission, and need for a subsequent hospitalization and operation for reversal. These downsides to diversion must be weighed with a patient's perceived ability to handle the physiologic stress and consequences of a severe leak if reoperation is required. When trying to determine which patients can handle a leak, the m-FI can help to objectively determine a patient's risk for increased rate of morbidity and failure to rescue in the event of a leak. While proximal diversion is still warranted in many cases, many surgeons tend to overuse proximal diversion in the setting of colonic surgery.