Anastomotic Considerations in Diverticulitis

Abstract

Diverticulitis is a common indication for colorectal surgery, both in the acute and the elective setting. The anastomosis between the colon and rectum is a critical component of colectomy for diverticular disease and should be approached thoughtfully. This article reviews important surgical considerations when creating a colorectal anastomosis in the setting of diverticular disease, whether following the reversal of an end colostomy, during an acute episode of diverticulitis, or electively for chronic or complicated disease. Timing of surgery and preoperative assessment, minimally invasive approaches, and intraoperative maneuvers and considerations are discussed.

Prior to anastomotic creation, attention should be paid to the extent of colon resection. This is important for both anastomotic healing and decreasing the risk of recurrence of diverticulitis. Proximally, the site of resection should be at soft, pliable colon that is free from inflammation or scarring. Studies have shown that inadequate length of resection, both proximally in leaving macroscopically inflamed colon, as well as distally, increases the rates of recurrence. ¹ It is not necessary to resect all proximal diverticula. ² The distal resection margin should be at healthy rectum, as anastomosis to the distal sigmoid is associated with higher risk of recurrent diverticulitis. ^{2 3} Intraoperatively, the proximal rectum can be differentiated from the colon by identification of the splaying of the tinea and loss of epiploic appendages, with the rectum starting at the loss of the tinea ( Fig. 1 ). Endoscopic evaluation can be useful to ensure that there are no diverticula distal to the rectal resection margin.

Fig. 1.

Intraoperative view of the rectum, demonstrating splaying of the tinea and loss of epiploic appendages. Drawing by Wali Johnson, MD.

For anastomoses, judged to have an unacceptably high-leak risk, proximal diversion can be considered. Reasons for this can include a wide range of factors that parallel poor wound healing in general. These can include steroids, sepsis, intraoperative blood loss, smoking, and many others. The two main options for diversion can include a loop ileostomy or a loop colostomy. ⁴ Two meta-analyses suggest equivalent functional outcomes, as well as strengths and weaknesses, to each approach. ^{5 6} The weaknesses of a loop ileostomy are the potential for acute kidney injury secondary to high output and adhesive disease, resulting in small bowel obstruction. Weaknesses of a diverting colostomy are the rate of parastomal hernia formation and prolapse. A sigmoid loop colostomy decreases the options available for future low pelvic anastomoses. A loop transverse colostomy is very difficult to manage compared with an ileostomy and should be avoided. Finally, closure of a loop colostomy can be a major procedure requiring laparotomy, something that is infrequent when closing a loop ileostomy.

In the acute setting, the performance of a primary anastomosis with diversion is becoming more popular. ⁷ Surgeon preference is the main factor in choosing an ostomy for perforated diverticulitis. ⁷ Diversion not only allows the healing of a higher risk anastomosis but also increases the likelihood of restoring bowel continuity. Stoma reversal rates are observed to be much lower in Hartmann's patients (30–57%) compared with a diverting ileostomy (90–95%). Those with primary anastomosis and diversion had decreased serious complications, shorter operating time, shorter hospital stay, and lower costs. ⁸ In the DIVERTI trial, while no significant difference in the morbidity and mortality was observed at the initial operation, the morbidity of the reversal operation was 21% for Hartmann's reversal versus 12% for a loop ileostomy. ⁹

Laparoscopic Approach

Numerous studies have compared laparoscopic and open surgery for management of diverticulitis since the first laparoscopic colon resections were described in 1991. ¹⁰ A study of over 124,000 patients using the Nationwide Inpatient Sample from 2002 to 2007 compared laparoscopic and open elective surgery for diverticulitis and determined that laparoscopic surgery was associated with a lower rate of complications and less morbidity and mortality, supporting laparoscopic surgery as the preferred approach. ¹¹ The SIGMA randomized controlled trial compared laparoscopic and open sigmoid resection for diverticulitis and reported that while the laparoscopic approach incurred a longer operation, it was associated with a 15.4% reduction in major complications, improved quality of life and shorter hospital length of stay in short-term analysis, ¹² and a 27% reduction in major morbidity on follow-up assessment at 6 months. ¹³ Contrary to these studies, a Cochrane review of three trials comparing the efficacy of laparoscopic and open surgery concludes that there is insufficient evidence to declare laparoscopic surgery, the preferred option for management of acute sigmoid diverticulitis. ¹⁴ Considering the superiority of a laparoscopic approach, in regard to wound healing, length of stay, narcotic use, and hernia risk, this approach should be considered when expertise is available. ¹⁵

Robotic Approach

Interest in robotic surgery for the treatment of diverticulitis continues to grow. Proponents of the robotic platform report increased freedom of motion with articulating instruments, three-dimensional visualization with a surgeon-controlled camera, and improved ergonomics. ^{16 17} A retrospective, propensity-matched analysis of patients with diverticular disease who underwent robotic, laparoscopic, and open surgery concluded that the robotic approach was favorable to laparoscopic in regard to conversion to open (7.9 vs. 12.5%, p  = 0.0002). Hospital length of stay was noted to be 0.3 days shorter in the robotic group compared with the laparoscopic group ( p  < 0.0001), and complication rates were lower in the robotic group (9.5 vs. 12.7%, p  = 0.02), as were rates of ileus (6.6 vs. 9.9%, p  = 0.004). ¹⁸ A single institution study compared laparoscopic and robotic surgery for diverticular disease and determined equivalent return of bowel function, length of stay, and postoperative pain, ¹⁹ while two population-based studies have suggested a decreased rate of conversion to open surgery and shorter hospital length of stay with the robotic approach. ^{20 21} Another single-institution study demonstrated significantly decreased surgeon stress during pelvic dissection in complicated diverticulitis with use of the robotic platform. ²²

The benefit of the robotic platform may be especially appreciated in surgery for diverticulitis complicated by fistula, as this pathology may require deeper dissection into the pelvis to arrive at healthy tissues. These operations are challenging due to the degree of inflammation and altered tissue planes between structures in the pelvis and are associated with a higher rate of conversion from laparoscopic to open surgery. ²³ A retrospective review of robotic and laparoscopic surgery for diverticulitis with colovesicle fistula at a single institution demonstrated increased rate of conversion to open in the laparoscopic group compared with the robotic group (14.55 vs. 0%, p  = 0.001), with a significantly longer total operating room time noted in the robotic group. ²⁴ In contrast, a retrospective study comparing the outcomes between robotic and laparoscopic surgery for diverticulitis with a variety of fistulae reported a higher conversion rate in the robotic group (18 vs. 0%, p  = 0.048), as well as a longer length of stay (7 vs. 4 days, p  = 0.007). ²⁵ In addition, the use of robotic surgery for sigmoid colon resection with intracorporeal anastomosis was researched in both benign and malignant disease suggested a benefit to the robotic approach, with decreased rates of conversion to open and decreased incidence of incisional hernia. ²⁶

Drains and Coating

Anastomotic leak is a devastating complication of colorectal surgery, and practices are born out of an effort to reduce this risk. After ensuring that healthy, well-perfused tissue is anastomosed without tension and a negative leak test is confirmed, the role of placing drains and/or topical agents near the anastomosis is considered.

While the placement of therapeutic drains in the event of a leak is appropriate, the routine use of drains following creation of the colorectal anastomosis should be avoided. ¹⁵ A Cochrane review of six randomized controlled trials compared the use of prophylactic drains after creation of colorectal anastomosis in elective surgery and declared insufficient evidence to support the practice of routine drainage to prevent anastomotic complications, ²⁷ a finding echoed by a systematic review. ²⁸

External coating of the colorectal anastomosis is a proposed technique to reduce the risk of leak. The coating materials researched are myriad and include both endogenous and exogenous materials such as amniotic membrane, collagen matrix-bound coagulation factors, fibrin sealant, hyaluronic acid, omentum, peritoneum, platelet rich plasma, synthetic absorbable and nonabsorbable meshes, and small intestinal submucosa and stem cells. Most of these coating materials have been studied in animals only, and often with contradictory results. Despite the lack of positive results, this is an ongoing area of research. ²⁹

The use of an omental pedicle flap has likewise not been shown to reduce the incidence of anastomotic leak. ^{30 31}

Anastomotic Considerations for Acute Disease

The primary goal of sigmoid colectomy in the acute setting is resection of the diseased colon to therefore obtain source control. As mentioned above, the proximal resection margin should be descending colon that is free of inflammation and scarring. Residual diverticular disease need not be resected if the colon is otherwise normal. The distal rection margin should be upper rectum to reduce the risk of recurrence.

Once the colon is resected, there are two main options: creation of an end colostomy (Hartmann's procedure) or creation of a primary anastomosis with or without diversion. Two randomized controlled trials offer guidance in making this decision. The first trial randomized 62 patients to receive either a Hartmann's procedure or a primary anastomosis with diverting ileostomy. ⁸ Overall complications were comparable, but they observed decreased serious complications, operating room time, hospital stay, and cost in the primary anastomosis group. There were improved ostomy reversal rates with primary anastomosis (Hartmann's: 57% vs. primary anastomosis: 90%). Finally, a safety analysis reported significantly more serious complications with stoma reversal after Hartmann's than primary anastomosis (20 vs. 0%). The DIVERTI trial was a multicenter randomized trial that enrolled 102 patients and assigned them to either a Hartmann's procedure or primary anastomosis with diverting ileostomy. ³² The study observed comparable mortality and morbidity in both groups, but an improved 18-month stoma reversal rate in the primary anastomosis group (Hartmann's procedure: 65% vs. primary anastomosis: 95%). These data suggest that in the setting of perforated diverticulitis, primary anastomosis with diverting ileostomy has similar mortality and morbidity with improved restoration of intestinal continuity.

There are important clinical scenarios where an anastomosis should not be performed. These include a friable, inflamed rectal stump, a hemodynamically unstable patient, and any other scenario where an anastomosis is not technically feasible or the added time of an anastomosis would be detrimental to the patient.

Staged Reconstruction: Hartmann's and Hartmann's Reversal Procedures

Sigmoid colectomy is advised in the setting of acute diverticulitis with diffuse peritonitis. The decision to create an anastomosis versus an end colostomy is dependent on several factors, and there are numerous studies exploring this topic.

Acute diverticulitis with diffuse peritonitis is frequently managed with an initial sigmoidectomy with end colostomy or Hartmann's procedure, with later reestablishment of continuity via colostomy takedown and creation of a colorectal anastomosis. The goals of the initial operation are removal of the diseased colon and creation of the end colostomy, leaving a rectal stump. It is advised to avoid any unnecessary pelvic dissection in this initial operation and to maintain length of the rectal stump or “Hartmann's pouch” to facilitate later identification for colorectal anastomosis.

Colostomy Takedown Workup and Timing, Intraoperative Considerations

Reversal of the end colostomy after acute diverticulitis remains a complicated procedure with considerable morbidity. Colostomy reversal rates vary in the literature from under 25 to 50%, reflecting that for many patients, the stoma is permanent. ³³

Colostomy takedown following acute, perforated diverticulitis can be a technically challenging operation, and it is important to plan accordingly before returning to the operating room. Review of the index operative report for surgical planning is advised. Details regarding extent of splenic flexure and transverse colon mobilization, proximal point of resection, and inadvertent injury during the operation are important to note. The management of the rectal stump should be noted regarding suture tagging or tethering to the abdominal wall with a suture pexy.

Timing

The timing of the colostomy reversal is debated, as the literature is quite varied. While delaying surgery may in theory allow adhesions to soften resulting in a safer surgery, an earlier return to the operating room may allow easier identification of less atrophied rectal stump. ³³ There are studies in support of both early and late reversal, as well as studies demonstrating no difference, ^{34 35 36 37 38} but it is generally accepted to delay surgery to 6 months, a time frame that also allows for patient optimization and prehabilitation. ³⁹

Rectal Stump Evaluation Preoperatively

Evidence promoting evaluation of the rectal stump prior to Hartmann's reversal is lacking, ⁴⁰ and a multicenter retrospective review did not find the length of the rectal stump to correlate with postoperative complications, including anastomotic leak. ⁴¹ Despite this, it is common practice to assess the rectum for length and overall health using contrasted enema and flexible endoscopy. The colon and rectal stump should undergo age-appropriate surveillance endoscopy for colorectal cancer screening.

Identification of Rectal Stump Intraoperatively

Identification of the rectal stump may be the most difficult part of the colostomy reversal surgery. Abscess, chronic inflammation, and previous rectal dissection can alter the planes of dissection for subsequent surgery. Some maneuvers to assist with the identification of the rectum include insertion of sizers into the vaginal cuff in patients with previous hysterectomy or use of a uterine manipulator and insertion of sizers or proctoscope into the rectal cuff. The bladder can be backfilled with saline to assist with defining anatomy, as well as evaluating for injury. Once the rectum is identified, it should be mobilized and then assessed for length and appropriate compliance for anastomosis. The dissection should begin posterior along the presacral plane and anteriorly below the peritoneal reflection as required, depending on the length needed to permit a tension-free anastomosis. Mobilization of the rectum will also allow easier passage of EEA sizers and stapler at the time of anastomosis. Compliance of the rectum can be assessed by palpation of supple and nonfibrotic tissue that expands with endoscopic insufflation.

The stump should then be trimmed to soft and healthy rectum. While during the Hartmann's procedure, it is acceptable to leave a long rectal stump comprised of distal sigmoid colon, at the time of colostomy reversal, it is recommended to excise any remaining sigmoid colon and to perform the anastomosis to the proximal rectum as retention of sigmoid colon can result in a four-fold increased risk of recurrent diverticulitis. ^{3 42} The stump can then be assessed endoscopically to ensure there are no diverticula distal to an air-tight staple line.

Preparation of the Proximal Colon

Once the rectum is determined appropriate for anastomosis, the proximal colon can be prepared by dissection of the proximal colon from the abdominal wall. Thereafter, it should likewise be trimmed back as required to arrive at soft, uninflamed colon with adequate perfusion. This proximal point of resection need not be motivated by removal of all diverticula-containing colon, as has been mentioned previously. However, it is important to ensure that no diverticula are contained within the anastomosis, as this may increase the incidence of anastomotic leak.

Anastomotic Considerations for Chronic Disease

In the setting of chronic diverticular disease, the first decision to make is timing of surgery. The decision to postpone resection following an episode, or flare, is based largely on allowing the acute inflammation to subside. Typically, a 6-week interval is observed. Prior to proceeding to the operating room, a preoperative colonoscopy is advisable, especially in complicated diverticulitis, to rule out an underlying malignancy and confirm to additional colon pathology that may warrant resection. ^{43 44} This likewise should be arranged after resolution of acute inflammation.

A well-accepted indication for resection for diverticulitis is stricture. In the setting of an obstructing or partially obstructing stricture, the surgeon may encounter proximal dilation of the descending colon with associated size mismatch when fashioning an end-to-end colorectal anastomosis. A side-to-end or Baker's anastomosis, performed either stapled or handsewn, can be considered in this scenario ( Fig. 2 ).

Fig. 2.

Creation of Baker's side-to-end anastomosis., Image provided by Aimal Khan, MD.

Despite delaying surgery to allow resolution of inflammation, the surgeon may encounter residual inflammation of the colon and rectum. The inflammation may extend distally beyond the proximal rectum where resection and anastomosis is usually performed, and even involve the anterior cul-de-sac. This may be especially noted in colovaginal or colovesicle fistula with associated abscess. In this scenario, it is wise to continue dissection of the rectum beyond the anterior peritoneal reflection to arrive at healthy, uninflamed rectum. To allow reach deeper into the pelvis, the splenic flexure may be fully mobilized to allow tension-free anastomosis.

At times, it is appropriate to consider diverting loop ileostomy to protect a distal colorectal anastomosis while it heals, as described previously.

Ureteral Stents

Ureteral stents are utilized to aid in identification of the ureters when difficult dissection is anticipated due to extensive inflammation or fistula formation between pelvic structures ( Fig. 3 ). While this practice has not decreased the incidence of ureteral injury, the stents facilitate the identification of the injury and allow early repair which is associated with improved outcomes. The surgeon must also consider the increased time and cost of ureteral stent insertion, as well as the associated risk of obstruction, injury to the ureter and infection, ⁴⁵ and selected patients accordingly.

Fig. 3.

Intraoperative identification of bilateral ureters using indocyanine green. Image provided by Aimal Khan, MD.

Multifocal Disease

Multifocal diverticulitis is very rare, described only in case reports. ^{46 47} General principles regarding treatment of uncomplicated and complicated diverticulitis should be applied and tailored to the specific patient. If an individual requires resections of two discrete segments of the colon, it is imperative that adequate perfusion to both anastomoses is maintained.

Fistulae

A colovaginal fistula occurs most frequently in women with diverticulitis with a history of hysterectomy. In addition to a history of feculent vaginal discharge, the fistula may be visualized on speculum examination, on cross-sectional imaging or contrasted enema. Endoscopy is less likely to diagnose a fistula but is important to exclude malignancy. ⁴⁸ Sigmoid colon resection without vaginal repair is sufficient for the majority of these fistulae, but if there is a visible tract, debridement with closure using absorbable suture is appropriate. ⁴⁹ Enlisting a gynecologic surgeon's assistance should be considered.

Symptoms of colovesicle fistula secondary to diverticulitis include pneumaturia, fecaluria, and recurrent urinary tract infection. Diagnosis may be achieved with cross-sectional imaging and cystoscopy ( Fig. 4 ). Removal of the sigmoid colon is indicated, while repair or resection of involved bladder is dependent on the character of the fistula. Cystoscopy with insertion of ureteral stents will assist not only with identification of the ureters during colon mobilization but also with identification of the ureteral orifice relative to the fistula. Should the bladder require repair, absorbable suture is required and backfilling of the bladder with saline and methylene blue is prudent to test the repair. ^{50 51} The assistance of a urologic surgeon should be considered.

Fig. 4.

CT scan of colovesicle fistula. Image provided by Benjamin Hopkins, MD. CT, computed tomography.

Conclusion

Diverticulitis remains a common indication for colorectal surgery in both the acute and elective settings. While the Hartmann's procedure is often employed in the setting of perforated diverticulitis, a primary anastomosis with a diverting loop ileostomy can be performed with similar morbidity and increased rates of ostomy reversal. Creation of the colorectal anastomosis in diverticulitis should follow the tenets of ensuring healthy, well-perfused, and uninflamed tissue in a tension-free fashion. It is appropriate to endoscopically assess the colon and rectum prior to surgery, and to time the surgery to allow resolution of acute inflammation. It is crucial to confirm appropriate proximal and distal points of resection to prevent recurrent disease and when feasible, a minimally invasive approach should be considered.

Funding Statement

Funding A.T.H.'s work on this manuscript was supported by the National Institute of Diabetes and Digestive and Kidney Disease of the National Institutes of Health under award number: K23DK118192.