Special Considerations of Anastomotic Leaks in Crohn's Disease

Abstract

Rates of anastomotic leak following intestinal resections in the setting of inflammatory bowel disease are significantly influenced by clinical characteristics. While the literature can be contradictory due to significant heterogeneity in the published data, several common themes appear to consistently arise. With respect to Crohn's disease, low serum albumin, preoperative abscess, reoperative abdominal surgery, and steroid use are associated with an increased risk of postoperative intra-abdominal septic complications. On the contrary, biologic therapy, immunomodulator use, and method of anastomosis appear not to confer increased anastomotic-related complications. Undoubtedly, a low rate of anastomotic leakage is inherent to procedures within colorectal surgery but diligent attention must be paid to identify, optimize, and, therefore, reduce known risks.

Anastomotic leak (AL) leading to intra-abdominal sepsis is a highly morbid and potentially fatal complication of gastrointestinal surgery. Rates of AL are widely variable, ranging from 1 to 30% and are influenced by several clinical factors including procedure types and patient characteristics. ¹ Special consideration of AL in the setting of Crohn's disease (CD) is warranted as despite significant advances in medical therapy, up to 46% of patients will require at least one intestinal resection within 10 years of diagnosis. ²

Intra-abdominal septic complications (IASCs) can be viewed as a spectrum of disease and are generally inclusive of abdominal abscesses, ALs, and enterocutaneous fistulae. Risk factors of AL have been extensively studied but remain a topic of significant controversy in the literature, likely due to patient heterogeneity and variation in surgical procedures and/or technique. The published data are often conflicting for many of the studied clinical variables, thereby complicating the landscape. Specifically examining causes of ALs, a recent meta-analysis of observational studies identified and screened 813 publications to identify risk factors for anastomotic dehiscence after surgery in CD. ³ Inclusion criteria for the studies included CD patients undergoing intestinal resection with primary anastomosis and/or strictureplasty; diverting stomas without anastomosis was excluded. After a rigorous analysis of the 15 studies included in the meta-analysis, the study authors concluded the following risk factors contribute to IASCs: low serum albumin, preoperative abscess, previous abdominal surgery, and steroid use ( Table 1 ). ³ Of note, the authors additionally concluded that neither biologic therapy, immunomodulator use, nor method of anastomosis (handsewn vs. stapled) contribute to AL in the setting of intra-abdominal surgery for CD ( Table 2 ).

Table 1. Clinical variables with association of postoperative IASCs in Crohn's disease.

Measured variable	No. of studies/patients	Pooled OR (95% CI)	Quality of evidence
Low serum albumin	8/2,127	1.93 (1.36–2.75)	Low
Preoperative abscess	12/3,132	1.94 (1.26–3.00)	Low
Prior abdominal surgery	9/2,737	1.50 (1.15–1.97)	Moderate
Steroid use	13/3,502	1.99 (1.54–2.57)	Moderate

Abbreviations: CI, confidence interval; IASC, intra-abdominal septic complications; OR, odds ratio.

Note: Adapted from Huang et al. ³

Table 2. Clinical variables without association of postoperative IASCs in Crohn's disease.

Measured variable	No. of studies/patients	Pooled OR (95% CI)	Quality of evidence
Biologic therapy	6/1,833	1.29 (0.79–2.11)	Low
Immunomodulator use	6/2,146	1.07 (0.66–1.73)	Low
Method of anastomosis (handsewn vs. stapled)	8/2,532	0.94 (0.58–1.53)	Low

Abbreviations: CI, confidence interval; IASC, intra-abdominal septic complications; OR, odds ratio.

Note: Adapted from Huang et al. ³

Hypoalbuminemia

Hypoalbuminemia is consistently associated with poor postoperative outcomes after gastrointestinal surgery. ⁴ Although low serum albumin is not necessarily a direct marker of nutritional status, an association with poor nutritional reserve is supported in the literature. ⁵ Additional causes of low serum albumin specific to CD include active inflammation and/or sepsis, each an independent risk factor of AL. In an early study to determine the effects of protein-calorie malnutrition, a cohort of rats were randomized to a normal laboratory diet or a low-protein diet before and after left-sided colonic anastomoses. The protein-depleted rats had both significantly lower serum albumin concentrations, as well as lower colonic burst pressures, suggestive of diminished anastomotic integrity. ⁶ Furthermore, in a recent study of 3,849 colorectal cancer patients, not only was hypoalbuminemia (serum albumin <35 g/L) associated with a two-fold increase in anastomotic complication rate, the 5-year overall survival was significantly reduced as well. ⁷ Specifically, addressing the role of low serum albumin in IASCs following intestinal resection and/or strictureplasty for CD, a meta-analysis by Huang et al of eight studies and over 2,100 patients identified a significantly increased rate of anastomotic dehiscence (odds ratio [OR] = 1.93, 95% confidence interval [CI]: 1.36–2.75). ³

Of note, the rates of malnutrition in patients with CD range from 20 to 85% with an estimated prevalence of 34% in acute inpatients. ⁸ Consequently, preoperative optimization of nutritional status remains an area of intense focus in gastrointestinal surgery. A landmark study in 1991 by The Veterans Affairs Total Parenteral Nutrition Cooperative Study Group examined 395 malnourished patients undergoing laparotomy or noncardiac thoracotomy. The study participants were randomized into one arm receiving oral diet as clinically indicated and into a second arm receiving total parenteral nutrition (TPN) 7 to 15 days prior to and 3 days after surgery. Severely malnourished patients, as defined by the Subjective Global Assessment or the Nutritional Risk Index, who had received perioperative TPN had fewer complications, including intra-abdominal abscesses, than controls (5 vs. 43%, p  = 0.03). ⁹ In a smaller study but specific to CD, a series of 15 consecutive patients given preoperative TPN for 18 to 90 days prior to undergoing bowel resection were compared with historical controls. In the patients on TPN, nutritional status and inflammatory markers all significantly improved and no IASCs occurred 30-day postoperatively. ¹⁰ The small size of the study clearly limits conclusions but, nonetheless, is evidence-favoring nutritional optimization of CD patients undergoing intestinal resection.

While the exact mechanism is poorly understood, exclusive enteral nutrition (EEN) has been successfully utilized to induce clinical remission in pediatric CD, likely due to a combination of anti-inflammatory effects, as well as alterations to the gut microbiota. ^{11 12} To study the effects of EEN, a retrospective review of 123 CD patients undergoing bowel resections for enterocutaneous fistulas was conducted. Interestingly, in the cohort of patients receiving 3 months of EEN, a decreased risk of IASCs was measured (3.6 vs. 17.6%, p  < 0.05). ¹³ Regardless of the method of nutritional support, decades of surgical literature clearly advocates for the assessment and subsequent optimization of nutritional status in the perioperative period. The benefits derived clearly go beyond anastomotic healing to include a reduction in both infectious and noninfectious complications. ⁵

Preoperative Abscess

One presentation of penetrating CD is the presence of an intra-abdominal abscess or fistula formation to adjacent organs. In fact, intra-abdominal abscesses occur in 7 to 28% of patients with CD. ¹⁴ A retrospective study reviewed 279 patients undergoing ileocolic resections for CD from 2000 to 2010 to identify risk factors for anastomotic-associated complications. One hundred and seventy-three patients underwent primary anastomosis and 11% experienced an anastomotic complication. The presence of an undrained abscess identified on preoperative computed tomography (CT) or during surgery was an independent predictor of AL (OR = 3.4, 95% CI: 1.2–9.8). ¹⁵ Tzivanakis et al hypothesized that presence of an undrained abscess may cause anastomotic complications through direct effect on healing or may simply reflect more advanced disease activity. Similarly, a comprehensive meta-analysis of 12 observational studies evaluating preoperative abscess found an increased risk of IASCs in postoperative CD patients (OR = 1.94, 95% CI: 1.2–63.0). ³

When suitable, percutaneous drainage of an intra-abdominal abscess from penetrating CD is a logical first option to control sepsis. Following abscess drainage, some management algorithms will favor definitive surgery, while others will advocate for nonoperative management if feasible. ¹⁶ A single-institution series of perforated CD treated with percutaneous drainage of the abscess in conjunction with up to 7 days of intravenous (IV) antibiotics, high-dose steroids, and parenteral nutrition found 16 of 19 patients were able to be safely managed with ileocolic resection and primary anastomosis without upstream diversion. ¹⁴ Resection with primary anastomosis in the setting of penetrating disease, therefore appears safe in carefully selected patients.

Repeat Intestinal Resection

CD is associated with a high likelihood of requiring intestinal resection during a patient's lifetime. ² In fact, of those needing an initial resection, 64% would go on to require repeat resection. ¹⁷ A single-institution retrospective review of 206 patients undergoing repeat intestinal resection for CD was performed to evaluate the risk of anastomotic complications. Of the 83 patients with prior intestinal resections, 14 ALs were detected. Johnston et al found that Crohn's patients with a previous resection were at increased risk of AL as compared with no previous resection (OR = 3.5, 95% CI: 1.3–9.4). ¹⁸ The study authors further concluded the number of previous resections correlated with increasing risk for clinical AL (correlation coefficient = 0.998). Additionally, a recent meta-analysis of nine studies similarly concluded that previous abdominal surgery was associated with a higher risk of IASCs (OR = 1.50, 95% CI: 1.15–1.97). ³ Unfortunately, prior abdominal surgery is a nonmodifiable risk factor but will serve to more accurately inform the patient and surgeon of expected risks.

Steroids

Glucocorticoid therapy has been a mainstay of treatment for acute flares in CD due to its potent immunosuppressive properties. However, the poorly tolerated side effects of long-term use have led to avoidance as a maintenance therapy. Within the surgical community, the negative effect on tissue integrity and the subsequent impact on anastomotic healing after surgical resections are reasons for concern. In early work in an animal model, rats pretreated with high-dose corticosteroids demonstrated a decreased bursting pressure in surgically created small-bowel anastomoses between the postoperative days 4 and 7, implying poor anastomotic integrity. ¹⁹ A prospective single-center study in 2012 identified steroid usage > 10 mg of prednisolone (or equivalent) for > 4 weeks before surgery in CD as conferring an increased risk of anastomotic-associated complications (OR = 2.67, 95% CI: 1.0–7.2). ¹⁵ The study authors who practice in a specialist inflammatory bowel disease (IBD) center were reasonably cautious of performing a primary anastomosis, particularly if a preoperative abscess was also present, as the risk of anastomotic-associated complication rose to 11%. Furthermore, a comprehensive meta-analysis of 13 studies and over 3,500 patients found steroid use was associated with a higher risk of IASCs after intestinal resection and/or strictureplasty in CD (OR = 1.99, 95% CI: 1.54–2.57). ³ In examining the dose-responsive effect, a retrospective cohort study of 159 patients undergoing elective surgery for IBD identified corticosteroid use > 40 mg prednisone daily as having the highest rate of major infectious complications, including intra-abdominal abscess. ²⁰ Clearly, while steroid use alone does not preclude primary anastomosis, diversion should be considered in the presence of additional risk factors, including malnutrition and local-tissue sepsis.

Biologic Therapy

A topic of ongoing debate in the IBD community is the role of biologic therapy in postoperative infectious complications, specifically AL. While the majority of clinical experience has been in the use of anti–tumor necrosis factor (TNF) agents, namely, infliximab, pharmacologic advances have expanded these therapies to include anti-interleukins (i.e., ustekinumab) and anti-integrins (i.e., vedolizumab). Based on a recent meta-analysis of six studies with over 1,800 patients included, preoperative biologic therapy is not a risk factor for IASCs in postoperative Crohn's patients (OR = 2.9, 95% CI: 0.79–2.11). ³ However, data to the contrary is seemingly abundant as well and must be considered in the context of each study's strengths and weaknesses.

The largest study demonstrating an increased risk of AL with anti-TNF therapy in CD was published in 2018. The work of Brouquet et al was a prospective cohort study conducted in France examining 592 patients who underwent surgery for ileocolonic CD with 24.1% patients receiving anti-TNF <3 months prior to surgery. ²¹ Characteristics of the surgical procedures included 73% laparoscopic approach, 72% ileocecal resection, and 10% emergency surgery with handsewn preferred to stapled anastomoses ( n  = 276 vs. 189). After propensity score matching, anti-TNF therapy <3 months was associated with an increased risk of IASCs (OR = 2.22, CI: 95% 1.22–4.04). ²¹ Septic intra-abdominal morbidity rates were not statistically significant in patients who received anti-TNF therapy more or less than 5 weeks prior to surgery, implying serum drug levels are not a direct cause. However, caution must be used when drawing definitive conclusions as the use of biologic therapy can be seen a surrogate marker for disease severity and, thus the increased risk of IASCs may not be a direct effect of the serum levels of anti-TNF.

Currently, a multicenter prospective study is being conducted through the Crohn's and Colitis Foundation Clinical Research Alliance by Cohen et al to examine if exposure to anti-TNF is an independent risk factor for infection after abdominal surgery for IBD. Patient recruitment occurred between 2014 and 2017 and included 955 patients. Exposure to biologics was defined as anti-TNF use within 12 weeks of surgery and serum for determination of biologic drug levels was collected at the time. Preliminary data presented at the Digestive Disease Week 2019 demonstrated that neither current anti-TNF administered within 3 months preoperatively (OR = 1.078, 95% CI: 0.744–1.560) nor detectable anti-TNF levels (OR = 1.448, 95% CI: 0.832–2.510) were associated with any infection within 30-day postoperatively. ²² Furthermore, increasing serum infliximab (OR = 0.992, 95% CI: 0.962–1.021) and adalimumab (OR = 0.994, 95% CI: 0.955–1.029) levels were also not associated with any infection. While the manuscript is pending, the study authors' concluded the preoperative use of anti-TNF drugs, as determined by history or by drug levels, was not an independent risk factor for postoperative infections.

As mentioned earlier, clinical data with newer classes of biologic agents, namely, vedolizumab and ustekinumab, are less robust but nonetheless must be considered due to rapid adoption. A retrospective review conducted at an IBD referral center evaluated 712 patients with CD receiving biologic therapy within 12 weeks of abdominal surgery with 127 receiving vedolizumab and 38 receiving ustekinumab. On multivariate analysis, while biologic therapy within the previous 12 weeks was a significant predictor of increased IASCs ( p  = 0.005), the particular class of biologic therapy was not independently associated. ²³ The rates of IASCs for treatment with anti-TNF, vedolizumab, and ustekinumab were 8.1, 11.5, and 9.5%, respectively, as compared with 3.6% on no therapy. As the half-lives of the several biologic therapies indicated in CD vary, with anti-integrins (i.e., vedolizumab) persisting the longest in the bloodstream, the optimal management of biologic therapy in the perioperative period will likely remain a highly contentious matter in the IBD community.

Discussion

Within the subset of patients undergoing intestinal resection and/or strictureplasty for CD exists special considerations unique to this disease process. A comprehensive meta-analysis of published studies has resulted in the identification of four risk factors for AL as follows: (1) low serum albumin, (2) preoperative abscess, (3) prior abdominal surgery, and (4) steroid use. ³ The most significant limitation of this meta-analysis by Huang et al is the observational nature of the included studies, thereby limiting the quality of evidence to moderate and low categories within the Grading of Recommendations, Assessment, Development and Evaluations approach.

With the exception of reoperative surgery, the above risk factors are potentially modifiable. In the severely malnourished patient with hypoalbuminemia, clinical circumstances permitting, optimization of nutritional reserves will profoundly decrease not only infectious complications but noninfectious complications as well. ^{10 13} Similarly, preoperative drainage of intra-abdominal abscesses and use of antibiotic therapy to control sepsis can function to decrease the need for emergency surgery, as well as the use of protective stomas, favoring primary anastomoses. ¹⁴ Lastly, the association of medical therapy for IBD (i.e., steroids, immunomodulators, and biologic therapy) with AL after intestinal resection is an area of significant contention. The largest meta-analysis of patients with CD implicates only steroid use as conferring an increased risk of IASCs. ³

Limitations and Conclusion

Clearly, the limitations of each study, particularly small sample sizes and a heterogeneous patient population with variable disease presentation, prevent the drawing of definitive conclusions. However, with that said, the IBD community eagerly awaits the published results of the PUCCINI study, evaluating the role of anti-TNF on postoperative infectious complications. ²²

Funding Statement

Funding This work was supported by the Peter and Marshia Carlino Early Career Professorship Award for Inflammatory Bowel Disease Research.