eAppendix. Search Terms and Strategy
Regenbogen SE, Hardiman KM, Hendren S, Morris AM. Surgery for Diverticulitis in the 21st CenturyA Systematic Review. JAMA Surg. 2014;149(3):292-303. doi:10.1001/jamasurg.2013.5477
Diverticulitis of the sigmoid colon is an increasingly common disease. Patterns of care and management guidelines have significantly evolved in recent years.
To review and classify the primary data published since 2000 that are guiding decision making, technical considerations, and the outcomes of surgery for sigmoid diverticulitis.
We searched the National Guideline Clearinghouse, PubMed, and Cochrane databases for studies pertaining to the diagnosis and management of chronic and recurrent diverticulitis from January 1, 2000, to March 31, 2013. We supplemented this automated search with references drawn from included studies and PubMed. We rated the level of evidence according to American College of Cardiology/American Heart Association guidelines.
We identified 68 studies meeting inclusion criteria for final review. The studies were almost exclusively observational and had limited certainty of treatment effect. We found that complicated recurrence after recovery from an uncomplicated episode of diverticulitis is rare (<5%) and that age at onset younger than 50 years and 2 or more recurrences do not increase the risk of complications. Chronic symptoms may persist even after resection in 5% to 22% of patients. Prophylactic surgery is generally not recommended for average-risk patients with diverticulitis, irrespective of the number of episodes of acute, noncomplicated disease. Decisions to proceed with colon resection should be based instead on the patient-reported frequency and severity of diverticulitis symptoms.
Conclusions and Relevance
The prior standard for proceeding with elective colectomy following 2 episodes of diverticulitis is no longer accepted. Decisions to proceed with colectomy should be made based on consideration of the risks of recurrent diverticulitis, the morbidity of surgery, ongoing symptoms, the complexity of disease, and operative risk. Laparoscopic surgery is preferred to open approaches. Recent evidence suggests that existing guidelines should be updated.
Diverticulitis of the sigmoid colon is a common condition, responsible for substantial morbidity and ongoing uncertainty about the optimal strategies for surgical management. The incidence of diverticulitis has increased in the last decade,1,2 accounting for nearly 300 000 US hospital admissions3 and $1.8 billion of annual direct medical costs.4 Also in the last decade, treatment algorithms have evolved in response to improved diagnostic imaging,5,6 greater understanding of the natural history and disease prognosis,7,8 and increased application of nonoperative management including percutaneous abscess drainage.9 Urgent operations for acute diverticulitis are now less common,2,9 and national guidelines have begun to question the traditional indications for elective resection as well.10 Still, per capita rates of elective surgery for diverticulitis have continued to increase, especially in patients younger than 65 years.1,11
Treatment goals for diverticulitis, whether for the hospitalized patient with an acute episode or a stable patient being treated as an outpatient, are to alleviate symptoms, prevent recurrence, perform elective rather than emergency operations when possible, and maintain intestinal continuity. The high incidence of operative complications, long-term morbidity,12,13 and persistent symptoms such as pain, bloating, and altered defecatory function after colectomy14- 16 all have imparted new uncertainty about optimal strategies for surgical management.
We performed a systematic review of recent primary reports on the decision making, technical aspects, and outcomes of surgery for acute, recurrent, and chronic sigmoid diverticulitis. The specific research questions were the following: (1) What are the indications for surgical resection? (2) Should a specific surgical approach or technique be recommended? and (3) What are the outcomes of surgical treatment? We consider this evidence in the context of published clinical practice guidelines10 and current patterns of care.2,11 A review of the nonoperative aspects of diverticulitis are concurrently published in JAMA.17
We searched the National Guideline Clearinghouse and PubMed between January 1, 2000, and March 31, 2013, for existing clinical practice guidelines addressing surgical management of sigmoid diverticulitis. The only relevant guideline identified was the practice parameter published by the American Society of Colon and Rectal Surgeons in 2006 (which was based on literature up to 2005).10 With this guideline as a starting point, we conducted a focused literature search using the MEDLINE and Cochrane databases for studies pertaining to the diagnosis and management of chronic and recurrent diverticulitis and published from January 1, 2000, to March 31, 2013. Our search terms and algorithm are provided in the eAppendix in the Supplement. We supplemented this automated search by reviewing additional references identified from included studies and from cited references searches in the Web of Science Citation Index.
We were interested in studies addressing surgical decision making, technical considerations, and outcomes for sigmoid diverticulitis. We included English-language studies reporting primary data on surgical treatment of diverticulitis of the descending and/or sigmoid colon in human adults. Three evidence domains were considered: (1) the indications for resection; (2) elements of surgical approach and technique; and (3) outcomes of surgical treatment. Review articles only reporting information present in other articles that we reviewed as well as case series with fewer than 30 patients were excluded. We also excluded studies with data accrued exclusively before 2000 because clinical practice guidelines18 and trends in surgical management of diverticulitis changed significantly at that time,2,11 and we sought to evaluate contemporary data derived primarily from patients treated after these changes.
Studies selected for inclusion were reviewed according to guidelines from the Meta-analysis of Observational Studies in Epidemiology Group.19 We graded the level of evidence and classified recommendations by size of treatment effect according to the guidelines process from the American College of Cardiology/American Heart Association Task Force on Practice Guidelines.20 We then compared the resulting recommendations with those of the most recent clinical practice guidelines on management of diverticulitis from the American Society of Colon and Rectal Surgeons.10
The database search elicited 2123 published articles, of which 740 were duplicates, leaving 1383 unique abstracts. Our manual review of references identified an additional 30 nonduplicate studies for abstract evaluation. Of these 1413 abstracts, we eliminated 1265 during abstract review and an additional 85 after full-text review based on the defined exclusion criteria described earlier. The remaining 63 studies serve as the basis for the systematic review. The selection process, based on the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) schema,21 is detailed in the Figure. The included studies and key findings are presented sequentially for each section in Table 1, Table 2, and Table 3.
We identified 6 studies that met the inclusion criteria and specifically reported data relevant to surgical decision making regarding urgent colectomy for diverticulitis (Table 1).5,22- 24,26,27
In patients presenting with acute diverticulitis, urgent surgery is performed for those with sepsis and diffuse peritonitis or those who fail to improve despite medical therapy and/or percutaneous drainage.10 Recognizing the substantial morbidity associated with urgent colectomy for complicated diverticulitis, however, there is a trend in favor of medical therapy, even in severe acute diverticulitis, whenever it can be achieved successfully.2,11 In the absence of severe clinical signs of sepsis, 2 single-institution studies suggest the potential for nonoperative treatment even for patients with complicated, perforated acute diverticulitis. Dharmarajan et al24 reviewed 136 patients with extraluminal air, fluid, and/or abscess and found that only 5 required urgent operation and another 12 failed expectant management. However, 28% underwent percutaneous drainage and 27% received parenteral nutrition. In a series reported by Costi et al,23 7% required surgery and an additional 18% needed percutaneous drainage.23,24 Nevertheless, in observational series, about 15% to 20% of patients with acute diverticulitis still undergo urgent colectomy during their index hospitalization.2,26,30,38
Patients with complicated acute diverticulitis—that is, with extraluminal abscess or pneumoperitoneum—are more likely to undergo urgent operation than those with colonic inflammation alone in reported case series.5,26 They are also more likely to experience complications and perforation after nonoperative management.22,26,27 In a series of more than 500 patients treated in a single academic medical center, Kaiser et al26 found that patients with pelvic abscess, for example, had more than 3 times greater risk of recurrence after nonoperative treatment than after operation (41% vs 13%, respectively). These findings were corroborated in another single-institution series of 465 patients22 in which 51% of patients with mesocolic abscesses and 71% of those with pelvic abscesses eventually required surgery, either acutely or in follow-up. To codify this additional risk, Ambrosetti et al6 defined a severity grading system (Table 4), based on the presence of an abscess or extraluminal air or on contrast on computed tomography, that provides useful guidance on the likelihood of acute resolution and freedom from recurrence. Patients who recover from acute inflammation with nonoperative management in the presence of perforation are typically encouraged to undergo elective resection because of a higher incidence of late complications.5,10,22,26 However, some authors have questioned the need for surgery even in this setting.25,30 In a retrospective, single-institution study, Gaertner et al25 described 36 patients managed nonoperatively after percutaneous drainage of colonic diverticular abscess, of whom fewer than half had recurrent diverticulitis, all managed without resection. Broderick-Villa et al30 likewise found that patients with percutaneously drained abscesses were no more likely than those with simple acute diverticulitis to have future recurrence or complication.
Eleven studies met inclusion criteria and provided data informing decisions about elective surgery for recurrent or chronic diverticulitis (Table 1).28- 38
Elective resection was recommended after 2 episodes of uncomplicated diverticulitis (or a single episode in young patients).83 This practice was based on the idea that such patients demonstrated, by virtue of recurrence and age at onset, a more “virulent” syndrome at greater risk for rerecurrence. Thus, elective intervention was deemed necessary to prevent future complicated episodes, emergency operation, and/or colostomy.84 These assumptions have been recently challenged.
The incidence of recurrent diverticulitis may be lower than was previously assumed. In 6 observational studies evaluating recurrence rates after a first episode of uncomplicated diverticulitis, the recurrence rates ranged from 10% to 35%,5,30,32,33,38,39 meaning that most patients never had another episode and would have no opportunity to benefit from surgery. We identified 11 studies eligible for review that addressed recurrence rates by age, but there was incomplete consensus. Some retrospective studies reported higher recurrence rates in younger patients.30,39,42,43,47 However, in most case series, the likelihood of recurrence was affected not by age at onset but rather by the severity of the initial episode.31,35- 37,39,83,84
Second, after 2 bouts of diverticulitis, the likelihood of rerecurrence may be modestly higher,30 but the severity of attacks generally does not increase.28,31 Instead, most patients’ recurrent episodes will mimic their earlier, uncomplicated ones. For example, in retrospective single-institution reviews with relatively long follow-up, only 3% to 5% of patients experienced a complicated recurrence after nonoperative management of an episode of acute diverticulitis.32,33 About 5% of patients managed conservatively in a single-institution series from Shaikh and Krukowski38 and a statewide cohort reported by Anaya and Flum39 went on to require emergency surgery. Further, the likelihood of free perforation seems to be far greater in patients with no history of diverticulitis than those who have recovered from 1 or more episodes.28,35- 37 In a retrospective cohort of more than 1300 patients, Ritz et al37 found that free perforation occurred in 25% of patients presenting with their first episode of acute diverticulitis, but only 12% with their second, 6% with their third, and 1% thereafter.
We included 14 studies evaluating patient-specific risk factors for recurrence and complications (Table 1).39- 52 Age at onset had traditionally been a point of emphasis, but the importance of age is increasingly downplayed, as discussed earlier. Observational studies have identified family history, retroperitoneal abscess, and length of the colon segment involved as factors associated with increased risk of recurrence,33 but it is not clear how well these factors can identify those with enough increased risk to justify prophylactic surgery. There is increased concern for complications of diverticulitis among patients with substantial comorbidity. Patients with immunosuppression, collagen vascular disease, glucocorticoid use, and malnutrition45,52 are at increased risk for recurrence and perforation with recurrent episodes, in some41,44 (though not all41) series, and might therefore have a lower threshold to consider elective resection. On the other hand, their operative risk may be substantially greater as well.71,74,85
The collective evidence supports an individualized approach to consideration of elective resection, which takes into account the severity of prior episodes, patient-specific risk factors, ongoing symptoms, and patients’ preferences. The number of episodes and age at onset should be considered secondary to these other factors. For those with recurrent diverticulitis who experience complete recovery between episodes, colectomy is intended to prevent future episodes of diverticulitis of similar severity as those they have already experienced. The data suggest that prophylactic surgery to prevent severe septic complications in asymptomatic patients with a history of diverticulitis is not necessary in most cases. Rather, patients may choose elective operation if the frequency and severity of their episodes is sufficient to justify the burden of surgery.
There are other patients who have atypical, chronic, or so-called smoldering diverticulitis in which symptoms are protracted and subacute.86 These patients may choose surgery as a remedy for their ongoing symptoms, not merely for prophylaxis. In these patients, it is particularly important to consider and exclude other diagnoses such malignant neoplasm and functional bowel disorders, which increase the likelihood of persistent symptoms after resection.15 In 2 reports of nonoverlapping series of patients from the same institution diagnosed as having chronic diverticulitis, resection relieved symptoms in 77% to 89% of cases, with persistent symptoms mostly related to comorbid irritable bowel syndrome.29,34 Thus, the decision to perform elective surgery for chronic sigmoid diverticulitis will be made on a case-by-case basis,10 weighing the severity of the ongoing syndrome and the likelihood of improvement against the short-term morbidity and long-term functional consequences of sigmoid colon resection.
Three randomized trials have compared laparoscopic and open colon resection for diverticular disease. In the Sigma trial, 100 patients with recurrent sigmoid diverticulitis were randomized to laparoscopic vs open sigmoid resection in 5 European centers. Patients randomized to laparoscopy experienced decreased incidence of major complications, reduced hospital stay and postoperative pain,87 and improved subjective quality of life at 6 months after surgery.88 A similarly designed and powered French study mostly replicated the short-term benefits but concluded that the only long-term advantage of laparoscopic resection was cosmesis.89,90 A third trial in Germany found no meaningful differences in short- or longer-term outcomes but was underpowered owing to difficulty with enrollment.91 Indeed, recruiting patients to participate in randomization for diverticulitis has proven difficult, often because of patients’ prejudged preference for laparoscopic surgery.92
Whether open or laparoscopic, options for definitive surgery in acute or chronic diverticulitis include resection of the affected colon segment with or without anastomosis and/or upstream diversion, depending on the health of the remaining bowel, extent of peritoneal contamination, overall condition of the patient, and baseline defecatory function. The data supporting each of these therapies consist predominantly of retrospective comparisons, limiting the conclusions that can be drawn about procedure choice for any given patient.
Because operations to close a colostomy incur substantial rates of major morbidity,52,71 as many as 45% of patients who undergo a Hartmann procedure with end colostomy and rectal stump closure may never attempt colostomy closure.53,54 There is therefore increased recent interest in strategies to maintain intestinal continuity with urgent surgery for diverticulitis. We included 8 studies that addressed this question (Table 2).53- 60
In a series of retrospective reviews comparing the Hartmann procedure against primary colorectal anastomosis with or without diverting loop ileostomy, short-term outcomes were similar and authors concluded that primary anastomosis in carefully selected patients undergoing urgent operations for acute diverticulitis is reasonable and safe.55- 59,93 Conclusions drawn from these retrospective trials are limited importantly, however, by selection bias in the choice of operation. In a small, underpowered trial comparing these approaches, Oberkofler et al60 randomized 62 patients to primary anastomosis with diverting ileostomy vs Hartmann procedure. Rates of mortality (9% vs 13%, respectively) and complication (75% vs 67%, respectively) were high but not different between groups. However, the likelihood of stoma closure was far greater with primary anastomosis than with the Hartmann procedure (90% vs 58%, respectively) and major complication rates for patients who underwent stoma closure were lower (0% vs 20%, respectively). These data suggest some preference for primary anastomosis with proximal diversion in the acute setting, when clinical conditions allow. However, the clinical factors to inform decision making in this setting are not well defined, leaving surgeons to make case-by-case decisions according to the suitability of the rectum for anastomosis, the clinical condition of the patient in the operating room, and the degree of pelvic and peritoneal inflammation around the intended anastomosis.
Regardless of the reconstruction or diversion method chosen, the margins of resection must be the same, as they appear to be the most important contributor to the likelihood of recurrent diverticulitis after resection. Specifically, in the presence of a colocolonic anastomosis with retained distal sigmoid colon, the odds of recurrent diverticulitis increased 4-fold compared with creation of a colorectal anastomosis.94 The distal resection margin therefore should extend below the rectosigmoid junction, beyond the point at which the taeniae coli coalesce. Proximally, the resection should include the thickened and chronically inflamed or fibrotic colon segment but need not remove all of the colonic diverticula. Thus, when the inflammation and fibrosis are limited to the sigmoid colon, an anastomosis from descending colon to rectum is adequate, whereas involvement up to the proximal descending colon would necessitate extended left colectomy.
Recognizing that resectional approaches to urgent colectomy incur substantial associated morbidity, there is increasing discussion of nonresectional operations for acute diverticulitis. Laparoscopic peritoneal lavage has been proposed as a damage-control operation to contain contamination and give patients with acute perforation and purulent peritonitis a bridge to elective resection with primary anastomosis. We found 6 studies reporting results of this procedure that met inclusion criteria (Table 2).61- 66 They were uniformly observational, small, single-institution studies that demonstrated fewer complications in patients who underwent lavage compared with primary resection. However, recognizing the substantial selection bias as patients converted from lavage to resection were typically analyzed with the resection group, it is difficult to draw firm conclusions from these comparative data. In addition, skeptics have argued that many of the patients “successfully” managed with lavage may have otherwise been amenable to nonoperative medical therapy with or without percutaneous drainage anyway, as many had localized abscesses and lacked substantial comorbidity.95 Thus, until results of expected randomized trials96,97 become known, there is insufficient evidence to recommend laparoscopic lavage as an alternative to resection in patients who fail nonoperative management.
We identified 8 recent articles that fit the inclusion criteria and specifically discussed 30-day postoperative outcomes (Table 3).67- 74
Most studies reported a postoperative infection rate of about 10% to 20%. Some reported substantially higher postoperative complication rates,72,73 likely related to the relative proportion of urgent rather than elective operations evaluated. Mortality rates tended to be less than 5% and were associated strongly with patients’ comorbid conditions and severity of presentation.68,71,74 Historical recommendations for management of acute and recurrent diverticulitis were based on the poor medical and surgical outcomes of the time. Today, the benefit of an elective operation to avoid the risk of future sepsis must be weighed against the risk of perioperative complications, especially among elderly patients and those with extensive comorbid disease.71,74,85
Studies examining long-term outcomes after colectomy for diverticulitis addressed 1 or both of the following: (1) risk of recurrent diverticulitis, and/or (2) patient-reported outcomes, including persistence or recurrence of symptoms. For the former outcome, 3 articles fulfilled our selection criteria (Table 3).75- 77 In these studies, imaging-confirmed postoperative recurrence rates ranged from 5.8% to 8.7%. Time to recurrence tends to be prolonged—29 months was the median and 2 of 5 recurrences were diagnosed more than 5 years after the index diagnosis. Thaler et al77 evaluated 155 patients who underwent elective laparoscopic or open sigmoid colectomy and reported that postoperative recurrence was not related to the urgency of the initial operation or to surgical approach, a finding corroborated in a slightly larger survey by Andeweg et al75 in which there was no association between an index emergency procedure and subsequent recurrent disease.
Six survey studies that examined patient-reported outcomes consistently noted that most patients had some relief of symptoms.16,78- 82 However, a substantial proportion of patients also experienced chronic or persistent symptoms of abdominal pain78,81,82 and, in 1 cohort study, poor postoperative continence.80 In the study by Andeweg et al,75 22% of patients reported ongoing diffuse, chronic abdominal pain that persisted after the operation. Overall, 2% to 20% of respondents reported no real long-term relief of their symptoms postoperatively and between 5% and 10% actually reported worse abdominal pain or gastrointestinal function.
Based on the findings of this systematic review, we compared our interpretation of the evidence and consequent recommendations with those in the most recent practice parameters published by the American Society of Colon and Rectal Surgeons.10 These recommendations and their class of treatment effect20 are presented in Table 5 for comparison.
Reviewing the published literature on surgical management of diverticulitis since the turn of the millennium, we found substantial changes in contemporary evidence that have reduced the use of urgent surgery for acute diverticulitis,1,2 restricted the indications for prophylactic surgery in recurrent diverticulitis,10,31,98 and led a movement toward generally less invasive and morbid approaches to this disease.99 Because of the rapid pace of recent changes in practice patterns, we focused this review on studies with data predominantly since 2000. At that time, the practice parameters of the American Society of Colon and Rectal Surgeons continued to recommend elective surgery after 2 uncomplicated episodes of diverticulitis, until the 2006 practice guideline that acknowledged the need to individualize timing of surgery.10,18,83
In this review, we found further evidence to support the practice of individualizing decisions about surgery based on particular characteristics of the patient and his or her presentation. Patients may pursue an operation out of fear of emergency colectomy resulting in colostomy. However, the likelihood of such an event after an uncomplicated episode of diverticulitis in average-risk patients is quite low.31,33,36,37 Thus, each patient must consider the frequency and severity of his or her recurrent episodes and the degree to which they limit quality of life. Patients who experience complications of diverticular disease, including perforation, fistula, and stricture, often do require surgical intervention. Thus, rather than relying on a number of episodes to guide surgical decision making, clear, patient-specific risk factors are needed to identify those at highest risk for morbidity with expectant management of recurrent diverticulitis. Outcomes registries have improved our recognition of patient factors contributing to operative morbidity and may help with the evaluation of surgical risk, but prognostic tools to predict the course of diverticular disease and the likelihood of important complications are lacking.
For postacute patients who recover with nonoperative therapy for an initial episode of diverticulitis, many authors have recommended endoscopic and/or fluoroscopic evaluation to confirm the diagnosis of diverticulitis and exclude other common causes of segmental colitis such as neoplasm, Crohn disease, and ischemic colitis.30,100,101 This recommendation remains controversial, however, as the yield and cost-effectiveness of colonoscopy have not been established in this setting. Malignant neoplasms are detected in 0% to 3% of patients with presumed diverticulitis6,102- 105 but perhaps as many as 13% of those with a protracted or otherwise atypical subacute clinical course.101 Some have argued that all of the patients with important endoscopic findings would have had other clinical indications for colonoscopy.105 However, practice parameters from surgical societies endorse routine flexible endoscopy to confirm the diagnosis of diverticulitis for any patient who has not recently undergone appropriate colorectal cancer screening.10,106
Another important component of these decisions is the patient-reported long-term functional outcomes of colectomy for diverticulitis. Most patients assume that surgery will be curative when they evaluate their options. We did find low rates of recurrent diverticulitis after resection with colorectal anastomosis.77,94 However, persistent gastrointestinal symptoms were common, perhaps suggesting overlap between chronic or recurrent diverticulitis, inflammatory bowel disease, and irritable bowel and other functional gastrointestinal syndromes that are not improved with resection.16,86,107 A challenge for future research is differentiating the symptoms of irritable bowel syndrome and inflammatory bowel disease from those of chronically symptomatic diverticulitis—this is particularly difficult because these conditions are so common in Western cultures.
The primary limitation of this study is that the overall quality of the evidence was limited. Retrospective observational trials and epidemiologic studies account for nearly the entire evidence base from which current practice standards are derived. Trials comparing modes of therapy were largely observational and subject to selection bias. Wide variability in surgical technique, diagnostic criteria, and completeness, duration, and method of follow-up made comparisons between studies difficult.
Diverticulitis remains a common, morbid, and costly condition for which optimal surgical management remains controversial across a number of domains. However, recent evidence suggests the safety of avoiding elective colectomy for most patients with uncomplicated disease and opens the door to modern approaches such as selective anastomosis with proximal diversion in the acute setting and laparoscopic colectomy in the elective setting. Ongoing prospective assessments of these emerging strategies are needed to continue to improve surgical management of diverticular disease in the 21st century.
Corresponding Author: Arden M. Morris, MD, MPH, Division of Colorectal Surgery, Department of Surgery, University of Michigan, 1500 E Medical Center Dr, TC-2124, Ann Arbor, MI 48109 (firstname.lastname@example.org).
Accepted for Publication: November 18, 2013.
Published Online: January 15, 2014. doi:10.1001/jamasurg.2013.5477.
Author Contributions: Dr Morris had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Regenbogen, Hardiman, Morris.
Acquisition of data: Hardiman, Morris.
Analysis and interpretation of data: All authors.
Drafting of the manuscript: Regenbogen, Hardiman, Morris.
Critical revision of the manuscript for important intellectual content: All authors.
Administrative, technical, and material support: Hardiman, Morris.
Study supervision: Regenbogen, Morris.
Conflict of Interest Disclosures: None reported.