Management of symptomatic colonic diverticular disease

Colonic diverticular disease is primarily a disease of the Western world. Management of diverticular disease can be challenging, and current best practice continues to evolve. Advances in intensive care, antibiotics and application of interventional radiology have reduced morbidity and mortality rates in patients undergoing treatment for diverticular disease.

Deciding the timing and place of surgery can be difficult. Controversies surround the indications for elective surgery, the role of age as an indicator for elective surgery, and the role of multistage operations in treating complicated diverticular disease. These issues will be reviewed here to help guide clinicians in managing patients with symptomatic diverticular disease.

In Western countries, about 60% of people over the age of 60 years will develop colonic diverticula. The increasing incidence in developed countries is directly related to reduced fibre intake following the introduction of steel rolling mills. Men and women are affected equally. However, only 10%–25% of those affected will have symptoms, with 15% of symptomatic patients suffering significant complications. In patients requiring urgent surgical intervention, overall mortality is between 12% and 36%.1

Colonic diverticula are caused by raised intraluminal pressure and segmentation of the colon from abnormal motility. This results in herniation of the mucosa through the muscle layer at weak points produced by penetration of vessels between the taeniae coli. These diverticula are pseudodiverticula, as most do not contain a muscle layer.

The sigmoid colon is the most common site for formation of diverticula, as it has the smallest diameter in the colon and a different motility pattern and is less compliant than the ascending colon.2 Intraluminal pressures in the colon have been shown to be nine times higher in patients with diverticular disease than in control patients. Right-sided colonic diverticula, though rare, occur more commonly among Asian people, perhaps due to a genetic predisposition.

Low dietary fibre intake reduces stool bulk and prolongs transit time, resulting in increased intraluminal pressure.2 Recent observations suggest diverticular disease may be related to long-term alterations in diet and to the patient’s microecology. A low-fibre diet alters the host’s microflora, resulting in changes to the colonic immune system.3 Slower transit time leads to increased exposure to intraluminal antigens.

A proportion of patients suffer disabling recurrent abdominal pain and altered bowel habit similar to irritable bowel syndrome. They may have changes in gut motility and sensitivity as a result of mucosal inflammation. It has been suggested that diverticular disease may be a form of chronic inflammatory bowel disease and that its recurrent nature may be controlled by altering the inflammatory response.3

Non-steroidal anti-inflammatory drugs have been implicated as a possible factor contributing to bleeding and perforation from diverticular disease.1 These drugs reduce gut prostaglandin synthesis, which affects the maintenance of the mucosal barrier and results in mucosal damage, increased colonic permeability and bacterial translocation.

Diverticulosis usually causes no symptoms. Its acute presentation is due to an obstructing faecalith, which results in micro-perforation of the diverticulum and secondary infection of the colon. The inflammation is extracolonic, with minimal intraluminal evidence of diverticulitis at colonoscopy during an acute attack. The extent of perforation will determine the presentation, which can range in severity from diverticulitis to faecal peritonitis (Box 1). This can be rapidly assessed with computed tomography (CT) scanning, which is the imaging modality of choice.5 It is very sensitive and has superseded contrast enemas, which carry a risk of perforation and barium leaks.

Most symptomatic patients (85%) can be managed non-operatively. About a third of patients will have a recurrent attack.6 Uncomplicated diverticulitis can be treated with bowel rest, broad-spectrum antibiotics and slow progression to a high-fibre diet. With appropriate antibiotic therapy, small abscesses will regress, possibly due to internal drainage. Abscesses less than 2 cm in diameter can be safely managed in an outpatient setting with oral antibiotics and sports drink supplementation.7

Pericolic abscesses larger than 5 cm in diameter can be drained percutaneously with CT guidance.5 This permits rapid control of sepsis, with success rates of up to 80%, and allows for subsequent single-stage resection. Following drainage, some patients may have a persistent symptomatic fistula caused by a faecalith obstructing the diverticular neck. Drainage alone is insufficient to control sepsis. If faecal drainage continues, surgical intervention is usually required.

The goals of surgery are to remove the diseased bowel segment, control sepsis, and restore bowel continuity with minimal morbidity and mortality. Deciding when to operate and which operation to perform can be difficult.

Mayo first introduced surgical resection for diverticulitis in 1907. Proximal transverse colostomy without resection of the diseased colon was a common method in the early 1900s.8 In 1942, Smithwick recommended a three-stage approach, as he found that the inflammatory process did not resolve without resection. This involved a proximal colostomy initially, followed by sigmoid resection, and finally, closure of the diverting stoma.9 The three-stage resection later fell out of favour, as mortality in patients with peritonitis was high if the diseased sigmoid was left behind.10 A two-stage procedure subsequently evolved. In the 1950s, the Hartmann procedure, involving sigmoid colon resection and end colostomy, was frequently performed for complicated diverticular disease. Operative mortality was almost halved.10

A trend towards single-stage resection was advocated in the early 1980s. The use of a one-stage procedure had earlier been limited by concerns about anastomotic leakage in the presence of peritoneal contamination, colonic inflammation and unprepared bowel. The risk of anastomotic leakage after resection for diverticular disease is a concern for several reasons, which include the presence of inflammation, shortening of the colon due to inflammatory fibrosis, muscular abnormalities, and the presence of diverticula at the site of anastomosis. In recent years, a number of authors have reported successes in treatment with primary anastomosis without colonic lavage.11-13 Good anastomotic integrity is influenced by a good blood supply and tension-free technique.11 A single-stage resection benefits the patient by avoiding a second operation for bowel restoration and its associated complications. The offset is a longer operative time and increased risk of anastomotic leak.

The Hartmann procedure still remains a safe option for managing patients who have complicated diverticular disease or are in poor condition. However, this procedure carries a significant risk of morbidity (23%–95%) and mortality (7%–28%).21 The presence of premorbid conditions, such as old age, collagen disease and cardiorespiratory disease, significantly increases the risk of postsurgical complications.18 The optimal timing for reversal of the Hartmann procedure is unclear. One study with a cut-off period of 3 months reported no difference in morbidity with early or delayed reversal.22 An earlier study reported better results after early closure (at 1 month) than after later closure (at 4 months). They noted more dense adhesions and less accessible rectal stumps in the delayed-reversal group.23 Sometimes the reversal may never eventuate, owing to poor medical health. About a third of patients retain their stoma for life.

A single-stage operation is advantageous because of the higher cumulative morbidity and mortality associated with multiple-stage resections. Primary anastomosis performed with or without bowel preparation has been shown to be safe in an emergency setting. Intraoperative colonic lavage before anastomosis may be unnecessary. It prolongs operative time and potentially increases the risk of faecal contamination.11 Schilling and colleagues reported no anastomotic leakages in their management of 13 patients with perforated sigmoid diverticulitis.13 Colonic lavage was not performed during the procedures. Morbidity and mortality in this group were similar to morbidity and mortality in another group who underwent a two-stage procedure.

In a systematic review of 98 published studies, Salem and Flum reported that mortality and morbidity in patients with diverticular peritonitis who underwent a primary anastomosis were no higher than in patients who underwent a Hartmann procedure.24 In 50 studies reporting 569 cases of primary anastomosis, they found that mortality was 9.9%, with an anastomotic leak rate of 13.9%. These rates were lower in patients who additionally had either a protective stoma or an intraoperative colonic lavage. In comparison, among 1051 patients who underwent a Hartmann procedure, mortality was 19.6%, with 10.3% having stoma complications and 4.3% having anastomotic leaks. The authors concluded that primary anastomosis is a safe alternative in selected patients.

Another recent systematic review by Constantinides and colleagues found no significant differences between mortality after primary resection with anastomosis and mortality after the Hartmann procedure in trials matched for degree of severity of peritonitis.25 The authors believed that the retrospective nature of these studies allowed for considerable selection bias, which limited their conclusions. They did not identify any beneficial effect of intracolonic lavage or faecal diversion in reducing postoperative mortality. Neither was there any significant difference in operative time between the two operations. It is possible that if surgeons are performing many primary anastomoses safely, they may be operating too often on patients with mild pathology that would have settled without operative intervention.

Several groups have found that laparoscopic drainage of a diverticular abscess can be effective and may obviate subsequent resection. Mutter and colleagues advocated an alternative two-stage laparoscopic approach to treating generalised peritonitis secondary to perforated diverticulitis: a limited primary laparoscopic lavage and drainage were performed to control intra-abdominal sepsis;26 any adhesions were untouched, to avoid disturbing a sealed perforation; and a passive drain was left exteriorising through the left iliac fossa. This converted a generalised purulent peritonitis into a localised diverticulitis, which could then be managed with antibiotic therapy. Once the inflammation had settled, a delayed laparoscopic resection could be performed in a non-emergency setting.

A similar approach using fibrin glue has been proposed. Faranda and colleagues treated 18 patients with laparoscopic peritoneal lavage, then applied fibrin glue over the infected sigmoid colon.27 Four patients required suturing of a perforation and six patients required an omentoplasty. A drain was inserted and antibiotic treatment was continued postoperatively. Thirteen patients later underwent an elective laparoscopic colectomy. The initial perforation was not found in all pathological specimens, and no patients developed an intraperitoneal abscess.

Such techniques are advantageous in that a laparotomy and stoma can be avoided.

Most patients with acute diverticulitis suffer an uncomplicated course. Non-operative management of uncomplicated diverticulitis remains the preferred treatment. Patients can be treated effectively with bowel rest, antibiotics and slow progression to a high-fibre diet. Current indications for elective colonic resection depend on individual circumstances after an attack of diverticulitis or after non-operative management of a complicated attack. Factors to be considered include the severity and frequency of diverticulitis, response to non-operative therapy, immunosuppression, medical comorbidities and patient choice.

In managing complicated diverticulitis, the Hartmann procedure is still considered a safe and quick operation for patients with faecal peritonitis or those who would not cope well physically with an anastomotic leak. Use of single-stage or laparoscopic procedures for managing complicated diverticulitis is gaining favour as a safe alternative to more radical and multistage procedures in selected patients. A single-stage procedure in experienced hands can be equally successful in an emergency setting in selected patients. Where possible, an operation should be converted from an emergency to a semi-elective one by using techniques such as radiologically or laparoscopically guided drainage of collections.

2 Barium enema showing a colovaginal fistula (arrow)

3 Resected specimen of the colovaginal fistula, with forceps entering the fistula tract (arrow)