Approximately 40% to 70% of patients with Crohn’s disease require intestinal resection within 10 years of diagnosis.1,2 Anastomotic leakage rates are increased in patients with Crohn’s disease compared with elective resections in noninflammatory conditions, ranging from 1.2% to 16.7%.3 Anastomotic leaks contribute to significant patient morbidity, increased mortality, prolonged hospital stay, and increased cost.4 Risk factors for anastomotic leak have been described previously. They include low serum albumin, preoperative steroid usage, and intra-abdominal abscess or fistula at the time of surgery.5–7
Of those patients who undergo resection for Crohn’s disease, ≈25% to 45% need repeated intestinal resection over the following 10 years.2,8,9 Looking out to 25 years after primary resection, the rate of repeat resection in Crohn’s disease further increases to 65%.10 Approximately one third of patients who undergo repeat intestinal resection require multiple repeat resections.8,10 It is unclear what effect repeat resections have on postoperative anastomotic leak. The current study investigated the hypothesis that patients with Crohn’s disease requiring repeat intestinal resection are at a higher risk of anastomotic leak than those undergoing initial resection.
PATIENTS AND METHODS
Patients were identified from a retrospective review of a prospectively maintained departmental database at a tertiary care facility from July 1, 2007, through March 31, 2016. The data set included all of the patients who underwent a surgical procedure at the Lahey Hospital & Medical Center Department of Colon and Rectal Surgery. The database included consecutive cases of operative encounters performed within our department. The data were collected prospectively at the time of surgery, and 30-day outcomes were collected by chart review by a trained impartial nurse or research assistant.
Inclusion criteria for the study were all of the patients with the diagnosis of Crohn’s disease who underwent intestinal resection and anastomosis. Patients were excluded if they underwent anastomosis with proximal fecal diversion, because proximal diversion has been shown previously to delay presentation of anastomotic leak beyond 30 days after surgery.11 Patients with Crohn’s disease undergoing loop ileostomy closure were also not included, because a resection must have been performed. The cohort was divided into 2 groups for comparison, including patients who had undergone previous intestinal resection and those who had no previous intestinal resection. The definition of previous intestinal resection did not include patients who had previous appendectomy, strictureplasty, or perianal surgery. The selection of the patient cohort is demonstrated in Figure 1. The study was reviewed and approved by Lahey Hospital & Medical Center Institutional Review Board (No. DR13-1638).
Variables evaluated included demographics (age and sex), as well as variables previously described that are related to the risk of anastomotic leak.5,7,12 These variables included preoperative serum albumin, intra-abdominal abscess or fistula at the time of surgery, ASA score, preoperative use of immune-altering medications, type of resection performed, duration of surgery, estimated blood loss, type of anastomosis, and surgeon. Preoperative serum albumin must have been measured within 30 days before surgery or was excluded. Immune-altering medications included systemic steroids, aminosalicylates (5-aminosalicylic acid or sulfasalazine), immunomodulators (6-mercaptopurine, azathioprine, or methotrexate), and biologic agents (antitumor necrosis factor antibodies) and were considered present if the medication was used within 12 weeks before surgery. The total duration of medication use was not collected because patients were often treated at multiple outside institutions before coming to our hospital. Patients were determined to have preoperative abscess or fistula if this was documented in the operative findings or if documented by preoperative cross-sectional imaging. For patients with multiple previous resections, the number of previous resections was determined, as well as the time between resections. The type of anastomosis was classified as stapled with GI anastomosis stapler, stapled with end-to-end anastomosis stapler, or handsewn.
Intestinal resections in patients with Crohn’s disease followed the basic principle of removing all macroscopic disease. Intraoperative microscopic evidence of the resection margins was not performed. Ten board-certified colon and rectal surgeons contributed to the data collected over the study period at a single institution. All of the surgeons maintained the same core principles, and primary anastomosis was only performed when there was acceptable condition of the bowel, the abdomen, and the patient. To evaluate for surgeon-specific factors, the surgeon for each patient was identified and then assigned a number (1–10) to allow comparison between groups.
The primary outcome evaluated was clinical anastomotic leak, defined in accordance with recommendations from a systematic review and the Surgical Infection Study Group.13,14 Clinical anastomotic leak was defined as intestinal luminal contents outside of the bowel coming from a surgical junction. This was characterized by luminal contents in a wound or drain or as a perianastomotic abscess on cross-sectional imaging in the presence of systemic inflammation (ie, leukocytosis, fever, or persistent tachycardia). To be a clinical leak, radiographic findings had to correlate with clinical signs of inflammation. No attempt was made to detect subclinical anastomotic leaks with routine radiologic investigation. Patients were monitored for anastomotic leak for 30 days after surgery.
Data were collected and analyzed using SAS version 9.4 (SAS Institute Inc, Cary, NC). Data were analyzed to compare patients who underwent previous resection with those patients who did not have any previous resection. Categorical variables were compared using χ2 test. Continuous variables were compared with Student t test. Correlation was evaluated with Pearson correlation coefficient. Significance level was determined as p < 0.05. Multivariate analysis was not performed because of size constraints of the study group.
Over the study period, 523 patients with Crohn’s disease underwent intestinal resection. Almost half (n = 237; 45%) were excluded because intestinal anastomosis was not performed, including anorectal procedures. Of the remaining patients, 37 patients had proximal fecal diversion with a loop ileostomy and 43 patients had a stoma takedown with no intestinal resection. The remaining 206 patients were divided into 2 groups based on previous resection (Fig. 1). Of the patients with Crohn’s disease who underwent resection and anastomosis without proximal diversion, 123 (60%) had no previous intestinal resection, whereas 83 (40%) had ≥1 previous intestinal resection. Patient demographics of the 2 groups are shown in Table 1. The groups were similar in terms of age, sex, preoperative serum albumin, the presence of preoperative fistula or abscess, ASA score, procedure performed, estimated blood loss, operating time, and surgeon. Ileocolic resection was the most common procedure performed in both groups. Preoperative serum albumin levels were collected on 66 patients (54%) with no previous resection and 42 patients (51%) with previous resection. Although patients with no previous resection were more frequently prescribed aminosalicylate medication, the use of steroids, immunomodulating medication, and biologic therapy was similar between the 2 groups.
The overall clinic anastomotic leak was 10% (n = 20/206). There were significantly more clinically apparent anastomotic leaks in patients undergoing repeat resection. There were 14 leaks (17%) detected in patients with previous resection compared with 6 leaks (5%) detected in patients without previous resection (p = 0.004). The OR of anastomotic leak in patients with Crohn’s disease with previous resection compared with no previous resection was 3.5 (95% CI, 1.3–9.4).
In the group of patients with previous resection, 30 patients (36%) had ≥2 repeat intestinal resections, whereas 53 patients (64%) had only 1 previous resection. The most common previous resection was ileocolic resection (83%) followed by small-bowel resection (10%). Time between resections was 15.2 ± 11.1 years (median, 12.0 y) and was similar between patients who had a leak after resection and those who did not have a leak (p = 0.73). Patients with 1 previous resection had a 13% leak rate (n = 7/53), whereas patients with ≥2 previous resections had a 23% leak rate (n = 7/30). The number of previous resections correlated with increasing risk for clinical anastomotic leak (correlation coefficient = 0.998; Fig. 2).
Anastomotic leaks most commonly presented with abdominal pain and fever and were detected from 4 to 25 days after surgery (mean = 9.8 d). Patients were managed with resection and end stoma formation (35%), primary repair and proximal fecal diversion (25%), bowel rest and antibiotics (15%), percutaneous drainage (10%), repeat bowel resection with anastomosis (10%), and abscess drainage with proximal fecal diversion (5%). Operative management depended on the degree of fecal contamination and the timing of presentation. Patients with feculent peritonitis were treated with a Hartmann procedure, whereas the patients with contained abscess or minimal feculent contamination were more commonly treated with primary repair and proximal diversion. Patients managed nonoperatively with either percutaneous drain or bowel rest with antibiotics presented later (12–25 days postoperatively) with contained leaks. Stoma formation was similar between the 2 groups (50% in no previous resection versus 71% in previous resection group; p = 0.36; Table 2).
Patients undergoing repeat intestinal resection for Crohn’s disease represent a particularly high-risk group of patients for risk of anastomotic leakage. Patients with Crohn’s disease with repeat resection had a >3-fold increase in leak rates compared with similar patients undergoing initial resection. Moreover, there appeared to be a direct correlation between the number of repeat resections and the rate of clinically apparent anastomotic leak. This dramatically increased leak rate in patients undergoing repeat resection was apparent despite similar patient characteristics, medication use, and operative factors as patients undergoing initial resection.
Based on previous study, more than half of patients diagnosed with Crohn’s disease will require resection within 10 years of diagnosis.1 After resection, approximately half will have postoperative recurrence of Crohn’s disease, and many will require repeat resection. The need for repeat intestinal resection in our cohort was 36%, which is similar to other studies showing a 25% to 45% rate of repeat resection in Crohn’s disease over 10 years.2,8,9 Therefore, patients with Crohn’s disease are frequently faced with bowel resection and repeat resection. Each intestinal resection poses a risk for anastomotic leak, a dreaded complication that affects patient morbidity and mortality. Surgeon decisions to alter the procedure to minimize risk of anastomotic leakage often are influenced by known risk factors associated with leakage. Repeat resection has not traditionally been considered a risk factor for leak.
The role of previous resection as a risk factor for anastomotic leakage in Crohn’s disease has been evaluated previously by other groups with mixed results. Huang et al15 recently published a meta-analysis of previous observational studies in Crohn’s disease and found that previous resection was a significant risk factor for anastomotic leakage with a pooled OR of 1.50 (95% CI, 1.15–1.97). However, each of the 9 observation studies that were included in the meta-analysis did not individually demonstrate previous surgery as a risk factor, thus complicating the interpretation of the pooled analysis. Yamamoto et al16 described repeat resection as a risk factor for anastomotic leak in their article evaluating the role of immunosuppressive medications on anastomotic leak. The increased rate of leak was significant on multivariate analysis (OR = 2.87 (95% CI, 1.01–8.18)) but was not significant on univariate analysis. Previous resection was an incidental finding, and the study concluded that more studies were necessary to corroborate the result. The findings of our study fit in the context of the existing literature and further support the role of previous resection as a risk factor for anastomotic leak in patients with Crohn’s disease.
Patients undergoing initial resection for Crohn’s disease had a 5% incidence of anastomotic leak, similar to departmental anastomotic leak rates published previously.17,18 Patients undergoing their first repeat resection had a 13% leak rate, whereas patients having multiple repeat resections had a 23% leak rate. A 1:4 leak rate is substantially higher than expected and demonstrates the need for a different approach in redo resections in Crohn’s disease. Repeat resection may be a surrogate marker for a more aggressive form of Crohn’s disease, a more complex surgery, or altered vascular supply. Time between intestinal resections was similar among patients who had an anastomotic leak and those who did not, suggesting that the aggressiveness of disease is either not an issue or is not well measured by time between surgical resections. Repeat resection may indicate a prolonged dissection, which would increase the risk of inadvertent enterotomy and potential devascularization of the bowel if the planes of dissection are atypical. In this retrospective study, it is likely that the surgeons would have recognized threatened bowel and therefore would not have restored intestinal continuity or would have performed proximal diversion. As a result, these patients would not have been included in the study cohort. Given that the study only included patients who underwent primary anastomosis, the reason for an increased rate of anastomotic leakage in repeat resection is likely subclinical, not readily apparent at the time of resection, and may be related to compromised tissue integrity in patients requiring repeat resection. Although the recommendation for routine use of proximal fecal diversion in patients with Crohn’s disease undergoing repeat resection seems extreme, the current study highlights a need for novel approaches to intestinal anastomosis in this patient population to attempt to decrease the rate of anastomotic leak. For example, if repeat resection is indicated, additional effort should be given to correct any modifiable risk factor for that patient, such as nutrition, anemia, or medication use.
The findings of the current study must be viewed within certain limitations. First, it is a retrospective study and therefore represents a highly selected group of patients. Patients were only included if anastomosis was performed, indicating that the operating surgeon perceived that the patient would do well at the time of anastomosis. Secondly, multivariate analysis was not performed because of the limited number of anastomotic leaks. The study was designed to compare patients with Crohn’s disease who were undergoing intestinal resection on the basis of repeat resection. Secondary analysis of the data comparing patients based on the presence of leak was not performed in an effort to limit type I error. Furthermore, the study does not account for the variability in the Crohn’s disease burden in each patient, the incidence of smoking, or the duration of immune-suppressing medication use.
Repeat intestinal resection in patients with Crohn’s disease is a high-risk procedure for anastomotic leak. Previous resection should be viewed as an additional risk factor for anastomotic leak and should be included in the surgeon’s algorithm for patient management. Although it is unclear why these patients who required repeat resection are at increased risk for anastomotic leak, it does not appear to be related to nutrition, immune suppression, or procedural factors. Patients with Crohn’s disease who are undergoing repeat resection should be counseled that they are at high risk for anastomotic leak, all modifiable risk factors should be corrected if possible, and consideration should be given to liberal use of fecal diversion.
1. Peyrin-Biroulet L, Loftus EV Jr, Colombel JF, Sandborn WJThe natural history of adult Crohn’s disease
in population-based cohorts. Am J Gastroenterol. 2010;105:289–297.
2. Bernell O, Lapidus A, Hellers GRisk factors for surgery and postoperative recurrence in Crohn’s disease
. Ann Surg. 2000;231:38–45.
3. Post S, Betzler M, von Ditfurth B, Schürmann G, Küppers P, Herfarth CRisks of intestinal anastomoses in Crohn’s disease
. Ann Surg. 1991;213:37–42.
4. Choi HK, Law WL, Ho JWLeakage after resection and intraperitoneal anastomosis for colorectal malignancy: analysis of risk factors. Dis Colon Rectum. 2006;49:1719–1725.
5. Yamamoto T, Allan RN, Keighley MRRisk factors for intra-abdominal sepsis after surgery in Crohn’s disease
. Dis Colon Rectum. 2000;43:1141–1145.
6. Crowell KT, Messaris ERisk factors and implications of anastomotic complications after surgery for Crohn’s disease
. World J Gastrointest Surg. 2015;7:237–242.
7. Telem DA, Chin EH, Nguyen SQ, Divino CMRisk factors for anastomotic leak
following colorectal surgery: a case-control study. Arch Surg. 2010;145:371–376.
8. Ramadas AV, Gunesh S, Thomas GA, Williams GT, Hawthorne ABNatural history of Crohn’s disease
in a population-based cohort from Cardiff (1986-2003): a study of changes in medical treatment and surgical resection rates. Gut. 2010;59:1200–1206.
9. Dhillon SL, Loftus EV Jr, Tremaine WJ, et alThe natural history of surgery for Crohn’s disease
in a population-based cohort from Olmsted County, Minnesota. Am J Gastroenterol. 2005;100:S305.
10. Landsend E, Johnson E, Johannessen HO, Carlsen ELong-term outcome after intestinal resection for Crohn’s disease
. Scand J Gastroenterol. 2006;41:1204–1208.
11. Leahy J, Schoetz D, Marcello P, et alWhat is the risk of clinical anastomotic leak
in the diverted colorectal anastomosis? J Gastrointest Surg. 2014;18:1812–1816.
12. Resegotti A, Astegiano M, Farina EC, et alSide-to-side stapled anastomosis strongly reduces anastomotic leak
rates in Crohn’s disease
surgery. Dis Colon Rectum. 2005;48:464–468.
13. Peel AL, Taylor EWProposed definitions for the audit of postoperative infection: a discussion paper–Surgical Infection Study Group. Ann R Coll Surg Engl. 1991;73:385–388.
14. Bruce J, Krukowski ZH, Al-Khairy G, Russell EM, Park KGSystematic review of the definition and measurement of anastomotic leak
after gastrointestinal surgery. Br J Surg. 2001;88:1157–1168.
15. Huang W, Tang Y, Nong L, Sun YRisk factors for postoperative intra-abdominal septic complications after surgery in Crohn’s disease
: a meta-analysis of observational studies. J Crohns Colitis. 2015;9:293–301.
16. Yamamoto T, Spinelli A, Suzuki Y, et alRisk factors for complications after ileocolonic resection for Crohn’s disease
with a major focus on the impact of preoperative immunosuppressive and biologic therapy: a retrospective international multicentre study. United European Gastroenterol J. 2016;4:784–793.
17. Waters JA, Francone T, Marcello PW, et alQuality improvement initiatives in colorectal surgery: value of physician feedback. Dis Colon Rectum. 2017;60:213–218.
18. Ricciardi R, Roberts PL, Marcello PW, Hall JF, Read TE, Schoetz DJAnastomotic leak testing after colorectal resection: what are the data? Arch Surg. 2009;144:407–411.