Diverticulitis: A Progressive Disease?

Abstract

Introduction:

Our understanding of complicated diverticulitis is based on outdated literature. Antecedent episodes of diverticulitis are felt to increase the risk of developing complicated diverticulitis, as well as its subsequent morbidity and mortality. Practice parameters recommend elective resection after 2 episodes of diverticulitis to reduce this morbidity and mortality.

Methods:

A total of 150 patients with prior episodes of diverticulitis who were hospitalized with complicated diverticulitis were retrospectively analyzed. Statistical analysis was conducted using χ² and Fisher exact test tests.

Results:

Patients were separated into 2 groups for analysis: group A = those with 1 or 2 prior diverticulitis episodes (n = 118) versus group B = patients with more than 2 prior episodes (n = 32). Characteristics of the groups were similar for age and preexistent comorbid conditions. The majority of patients presented with pericolonic abscess and inflammatory phlegmon. Perforated diverticulitis occurred more often in group A compared with patients with >2 episodes of diverticulitis. Because of the higher rate of perforation, patients in group A underwent surgical diversion more often than group B patients. No significant differences in operative complications, morbidity, or mortality rates were identified between the groups.

Conclusion:

Patients with multiple (>2) episodes of diverticulitis are not at increased risk for poor outcomes if they develop complicated diverticulitis. Morbidity and mortality rates are not significantly different between patients with multiple episodes of diverticulitis compared with those with 1 or 2 prior attacks. Reevaluation of the practice of elective resection as a strategy for reducing the mortality and morbidity from complicated diverticulitis is needed.

Reports suggest that patients with >2 recurrent episodes of diverticulitis are more likely to have poor outcomes and higher mortality if complicated diverticulitis develops. In a retrospective review of patients with complicated diverticulitis, patients with >2 prior attacks had lower mortality rates, similar morbidity rates, lower incidence of perforation, and reduced diversion compared with patients with fewer prior episodes.

Colonic diverticulosis is among the most common diseases of developed Western countries. While most people with diverticular disease remain asymptomatic, between 10% and 25% of patients with diverticulosis will ultimately develop diverticulitis. Of those patients who develop diverticulitis, 15% to 20% will develop significant complications.^1–3 Complicated diverticular disease is defined as diverticulitis associated with abscess, fistula, obstruction, phlegmon, bleeding, or perforation.^4–10

At present, elective sigmoid resection is recommended after 2 episodes of uncomplicated diverticulitis to prevent the serious complications of recurrent colonic diverticulitis.4,7,8,10,11 The Ad Hoc Practice Parameters Committee of the American College of Gastroenterology, states that “…recurrent attacks are less likely to respond to medical therapy and have a higher mortality rate, therefore most authorities agree that elective resection is indicated after 2 attacks of uncomplicated diverticulitis.…”¹¹ This current guideline has been based on the assumption that recurrent episodes (2 or more) of diverticulitis will lead to complicated diverticulitis and higher mortality. The data used to support this assumption are based on only a few small studies that reported recurrence rates of >40% after one episode of diverticulitis with complications occurring in 30% to 60% of patients.1–2,8,12,13 Furthermore, the majority of these studies were published over 20 years ago.

Advances in diagnostic modalities, medical therapy, and surgical techniques over the past 2 decades have changed both the management and outcomes of diverticulitis. It is clear that data on diverticular disease have not kept pace with advances in the surgical management of the disease. The aim of this study was to examine the impact of antecedent episodes of uncomplicated diverticulitis on the prognosis of patients who subsequently develop complicated diverticulitis.

MATERIALS AND METHODS

Between 1990 and 2003, the charts of all patients hospitalized for the diagnosis of diverticulitis associated with perforation, abscess, obstruction, phlegmon, fistula, or bleeding at the Mayo Clinic, Rochester, MN were retrospectively reviewed. Patients were excluded from the study for the following criteria: diagnosis of diverticulitis changed on review of the operative or pathology reports, diagnosis of uncomplicated diverticulitis, patient refusal to participate in research, and patient chart with incomplete data such as missing pathology or operative report. A total of 996 patients had the diagnosis of diverticulitis at our institution during the 13-year time period. After the exclusion criteria were applied, 330 patients with complicated diverticulitis were identified. Of the 330 patients identified with the diagnosis of complicated diverticulitis, 150 patients had an antecedent episode of diverticulitis. This group of patients were further reviewed and analyzed.

Data on the mode of presentation, location of disease, physical findings, radiographic, and laboratory findings, preoperative comorbidities, prior history of diverticulitis, type of complicated diverticulitis, details of operation, postoperative complications and death, duration of hospitalization, and ostomy closure were collected for all patients by comprehensive chart review. The type of complicated diverticulitis was determined by the following criteria: phlegmon was a symptomatic inflammatory mass separate from and not associated with a purulent fluid collection; pericolic abscess was a collection of purulent material localized near the colon or in the pelvis; obstruction was characterized by obstructive symptoms and a stricture or stenosis diagnosed by radiographic studies or pathologic evaluation; fistulas included colovaginal and colovesicular connections; perforation was either a collection of contained air outside the bowel wall or a large amount of free intraabdominal air; and patients who required a transfusion for a lower gastrointestinal hemorrhage associated with symptoms and signs of diverticulitis were placed in the bleeding category.4,5,7,10 The Hinchey classification system was not used to define patients. Peritonitis was not used to designate a patient as “perforated.”

For this study, the 150 patients with antecedent episodes of diverticulitis were separated into 2 groups for analysis and statistical comparison: Group A included patients with 1 or 2 prior episodes of diverticulitis. Group B included patients who had more than 2 prior episodes. The demographic characteristics, baseline clinical parameters, and outcomes of patients in these 2 groups were summarized and compared statistically. Parameters comprised of discrete, categorical variables were analyzed using χ² tests or Fisher exact test tests when necessary and appropriate, and continuous variables were analyzed using Wilcoxon rank sum tests. Logistic regression was used to examine the association of the group with perforation after adjusting for steroid use. All statistical tests were 2-sided, and the threshold of statistical significance was set at P < 0.05.

RESULTS

Characteristics of Patients

Of the 150 patients hospitalized for complicated diverticulitis, 118 patients had 1 or 2 antecedent episodes of uncomplicated diverticulitis managed by medical therapy (group A). Thirty-two patients had a prior history of more than 2 episodes of diverticulitis (group B). The mean age of patients in group A was 65.6 ± 12.9 years and the mean age in group B was 61.5 ± 11.7 years. There were 19 patients ≤50 years old in group A and 7 patients 50 years or younger in group B. However, there was no statistical difference between the groups in terms of age distribution. In the group with ≤2 episodes, 48% of the patients were male and 52% were female. In the group with >2 prior diverticulitis attacks, 47% of the patients were male and 53% were female. No significant difference in gender was identified between the groups (Table 1).

TABLE 1. Demographics and Preexisting Conditions

There were no identifiable differences in preexistent comorbid conditions, such as diabetes, cardiovascular disease, or pulmonary disease. The most common preexistent comorbidity in this cohort was cardiovascular disease such as hypertension. There were 4 patients in group A who had collagen vascular disease and none in group B. Seventeen percent of patients in group A were taking steroids at the time of the presentation of their complicated diverticulitis compared with 3% in group B (Table 1).

Presentation

Complications of diverticular disease present on admission included acute phlegmon (28% in group A vs. 38% in group B, P = 0.31); paracolic abscess (41% in group A vs. 41% in group B, P = 0.97); obstruction (25% in group A vs. 34% in group B, P = 0.27), fistula (21% in group A vs. 9% in group B, P = 0.15); bleeding (3% in group A vs. 3% in group B, P > 0.99); and perforation (17% in group A vs. 0% in group B, P = 0.001). Overall, the majority of patients with antecedent episodes of diverticulitis most commonly presented with pericolonic abscess and inflammatory phlegmon (Table 2). Sixty percent of the patients (91 of the 150) in the study presented with only one type of complicated diverticulitis. Twenty-nine percent of patients had 2 forms of diverticular complications such as a phlegmon and an obstruction. Eleven percent (16 patients) were identified as suffering from more than 2 types of complicated diverticulitis. Group A and group B patients were not found to have a significant difference in the number of complications of diverticular disease on presentation. The site of complicated diverticulitis was the sigmoid colon in 97% of group A patients and 97% in group B patients.

TABLE 2. Presentation of Diverticular Complication and Operative Management

For this cohort, 52% of the patients who were taking steroids presented with perforation. Although steroids were significantly associated with perforation (P = 0.03), statistical analysis determined that steroid use was not the reason that group A patients presented with more perforations than group B patients. The probability of presenting with perforation was higher in group A patients than group B patients (odds ratio, 5.9; 95% confidence interval [CI], 1.7–20.8, P = 0.005). Adjusting for steroid use only slightly diminished the strength of this association (odds ratio, 5.3; 95% CI, 1.5–18.9, P = 0.009). With the use of a logistic regression model containing steroid use and the groups, steroid use, although having a large odds ratio, was not statistically significant (odds ratio, 2.2; 95% CI, 0.85–5.7, P = 0.11). Therefore, the variable of having 1 or 2 antecedent episodes of uncomplicated diverticulitis had the most significant association with the development of perforation in our cohort.

The average white blood cell count was 9.6 ± 4.5 in group A patients and 8.5 ± 4.2 (P = 0.26). Eighty-eight percent of the patients in group A had normal vital signs at presentation, and fever was seen in only 19%. All patients in group B had normal vitals on presentation to the hospital and only 3% had a fever.

Operative Management

The majority of patients in both groups (63% in group A and 97% in group B) were managed with segmental colon resection and primary anastomosis. Perforated diverticulitis was mostly treated with segmental colectomy and colostomy with a Hartmann's pouch. There was a significantly higher diversion rate with ostomy in group A (37%) patients compared with patients with >2 episodes (3%) of diverticulitis (P < 0.001). This finding is likely due to an increased occurrence of perforated diverticulitis in group A patients compared with group B patients. In the patients without perforation, 21% of the patients with 1 or 2 prior uncomplicated episodes had a diversion procedure compared with 3% of the patients with more than 2 prior episodes (P = 0.03). Abscess, inflammatory phlegmon, obstruction, fistula, and bleeding were managed primarily with resection and primary anastomosis (70%).

Mortality and Morbidity

The overall mortality rate for the cohort was 2.67% (n = 4 patients). There was no significant difference in mortality between patients with 1 to 2 antecedent episodes and patients with >2 prior episodes of diverticulitis. In the group A patients, there were 2 deaths. One patient was a 72-year-old man with one prior episode of uncomplicated diverticulitis on steroids for rheumatoid arthritis who presented with perforated diverticulitis and a pelvic abscess. The other patient in group A was an 80-year-old woman who had 2 antecedent episodes of diverticulitis and presented with perforated diverticulitis. She developed sepsis and multiorgan failure. There were 2 deaths in group B. Both deaths were a result of anastomotic leak. The first patient was a 33-year-old man who had 5 prior episodes of uncomplicated diverticulitis and presented with a stricture causing obstructive symptoms. He underwent a segmental resection with primary anastomosis and had an anastomotic leak on postoperative day 4, which led to sepsis. The second patient was a 67-year-old man who had had 3 prior diverticulitis episodes and developed an abscess with his fourth episode. He underwent a resection and primary anastomosis and was found to have an anastomotic leak on the 10th postoperative day.

The overall morbidity rate for the cohort was 40%. Sixteen patients (11%) had a complication postoperatively that required return to the operating room or need for a surgical debridement. The anastomotic leak rate was 1.3%. Twelve patients in group A (10%) and 4 patients in group B (13%) developed an operative complication. There was no significant difference in operative complication rates between the 2 groups (P = 0.75). Ileus was the most common cause of postoperative morbidity in both groups (15% in group A vs. 19% in group B). There were no significant differences identified between group A and group B for postoperative complications (Table 3). Group A patients had an overall rate of 40% for the development of nonoperative morbidity compared with group B, which had a 25% rate (P = 0.12).

TABLE 3. Morbidity and Mortality

DISCUSSION

Diverticulitis is a common condition with an estimated incidence of 25%.^1,3,14 After 1 episode of diverticulitis, one third of patients have recurrent symptoms. After a second episode, a further third will develop a subsequent episode.¹⁴ Based on early studies, these recurrent episodes were thought to be associated with the development of complicated diverticulitis and an increased risk of mortality. Thus, a recommendation of elective sigmoid resection after 2 episodes of uncomplicated diverticulitis has been elaborated.

Much of the current knowledge of the natural history of diverticular disease and the foundation of practice parameters in managing the disease arise from the pioneering work of T.G. Parks. In 1969, Parks published a prospective study of 455 patients that were followed between 1951 to 1965 and from 1 to 16 years.¹⁵ A total of 317 patients were admitted with the diagnosis of diverticulitis. Of these 317 patients, 25% were admitted with a second episode, 4% with a third episode, and 1.6% with a fourth episode. The mortality rate increased from 4.7% during the first admission to 7.8% with subsequent admissions. Thus, Parks concluded that medical treatment of recurrent disease was less rewarding with each admission.^14,15 However, this study was conducted prior to the application of computed tomography, and many, including Parks himself, have questioned the accuracy of the diagnosis of diverticulitis in his series. Other studies that have reported an association between recurrent diverticulitis and higher mortality rates, as well as increased risk of complications from diverticulitis, have been retrospective, with smaller modest numbers of patients.^4,7,13

In contrast, several newer studies have noted no increased risk of complicated diverticulitis, or unacceptably high mortality. In a study by Makela et al,¹⁶ 336 patients with diverticular disease were followed for a mean of 10 years. Eighty-four percent were admitted once for uncomplicated diverticulitis and treated medically; 8% twice, 5% 3 times, and 3% were admitted 4 times. During follow-up, no deaths were due to recurrent diverticulitis, and recurrence was not found to be associated with an increased rate of complicated diverticular disease. Chautems et al¹⁷ followed 118 patients after a first acute diverticulitis episode for a median of 9.5 years. Seventy-one percent of the patients had no recurrent episodes. None of the patients with subsequent episodes of diverticulitis died of complicated diverticulitis or required emergency surgery. Moreover, Haglund et al¹⁸ noted that only 25% of 392 patients followed over a 12-year period with diverticular disease developed recurrent episodes. In this 25%, there were no perforations, and medical management was successful in all cases. In contrast to other older published results, the risk of recurrence in the first year was 10% but fell to 3% per year thereafter. Furthermore, the majority of complicated diverticulitis, such as perforation, was seen as the initial presentation, and it was observed that, after the first episode, the disease appeared to run a benign course. A recent population-based study¹⁹ demonstrated that only a small percentage (5.5%) of patients who recovered from an initial episode of uncomplicated diverticulitis required subsequent emergency colectomy or colostomy.

Our study confirms this last work that suggests that multiple episodes of diverticulitis are not associated with increased mortality or increased risk of complicated diverticulitis. Of 330 patients who presented to the Mayo Clinic over the past 13 years with complicated diverticulitis, only 45.4% had an antecedent history of uncomplicated diverticulitis. However, the overall mortality rate for those patients with a prior history of diverticulitis was not increased but was lower than patients who present with complicated diverticulitis as an initial presentation (2.5% in patients with prior history vs. 10% with no prior history).²⁰

This significant difference in mortality rates between patients with no history of diverticulitis compared with those with recurrent episodes is most likely due to the well-known finding that perforated diverticulitis is more common as an initial presentation of diverticulitis. In a case-controlled study, Hart et al²¹ found that 78% of patients with perforated diverticulitis had no prior history of diverticulitis.

In our initial cohort study of 330 patients with complicated diverticular disease, 54.6% presented with complicated diverticulitis as the first manifestation of their disease. Sixty-eight percent of those who presented with perforation presented with this as the initial event of the disease.²⁰ In comparison, patients with a history of diverticulitis tended to present with pericolonic abscess or inflammatory phlegmon rather than perforation. In our current study, patients with fewer prior episodes of diverticulitis (1 or 2) presented more often with perforation compared with patients with >2 episodes of diverticulitis. Patients who had only 1 or 2 episodes of uncomplicated diverticulitis were at a higher risk for perforation than patients with >2 episodes.

It is interesting to note that mortality in group A patients was from sepsis secondary to perforation. The mortality in group B patients was due to complications from the operation itself (anastomotic leak), rather than to the complications of the disease per se. Morbidity is a significant problem after the treatment of complicated diverticulitis. The overall morbidity rate was 40% for this cohort. However, there was no increased risk of postoperative complications in patients who had multiple recurrent episodes of diverticulitis. As shown in our recent overall analysis of the 330 patients with complicated diverticular disease, preexistent comorbid conditions and steroid use are associated with increased risk of postoperative morbidity for both patients who present with no prior history of diverticulitis as well as patients with multiple antecedent episodes.²⁰

Moreover, prophylactic colectomy is not without its attendant mortality and morbidity rates. Elective sigmoid resection is associated with risks of mortality and colostomy as high as 2.3% and 14.2%, respectively.^22,23 Furthermore, the risk of recurrent diverticulitis is not eliminated after colectomy. Rates of recurrence after colectomy have been reported as between 2.6% to 10.4%.²⁴ In a recently published decision analysis, using a Markov model, Salem et al²⁵ determined that performing colectomy after the fourth episode of diverticulitis, rather than the second episode of diverticulitis, in patients older than 50 resulted in 0.5% fewer deaths, 0.7% fewer colostomies, and a reduction in cost per patient.

Our study confirms other reports in which multiple episodes of diverticulitis are not associated with increased risk of mortality or poor outcomes from complicated diverticulitis. The patients in group B of our study (>2 episodes of uncomplicated diverticulitis) are the defined group for the practice parameter of elective sigmoid resection to prevent the complications themselves, and the higher mortality rate of complicated diverticulitis. We found that these patients have a much lower mortality rate than previously reported, no difference in morbidity rates compared with patients with fewer or no prior episodes, and a lower incidence of perforation with a reduced need for diversion. Thus, it does not seem that diverticulitis is often a progressive disease; it is not predictable.

The current recommendation of prophylactic resection after 2 episodes of uncomplicated diverticulitis comes too late for many patients and may be unnecessary in other patients. Our study is limited by relatively few patients in group B and the retrospective design of the study. This most certainly leads to a concern of statistical power. Our analysis does not shed light on which patients with diverticulitis are at risk for having recurrent episodes or why some patients with recurrent episodes of diverticulitis progress to complicated diverticulitis and some do not. Our study suggests that recurrent episodes of diverticulitis are not associated with an increased risk of mortality or a less favorable outcome if complications of diverticular disease develop. Current guidelines are based on studies largely published before the widespread use of CT scanning for accurate diagnosis, the advent of percutaneous drainage, improved antibiotics, and critical care. Based on our data, reevaluation of the practice of elective sigmoid resection after 2 episodes of diverticulitis as a strategy to reduce morbidity and mortality from complicated diverticulitis seems appropriate, and prospective investigation into the natural history of diverticulitis is warranted.

Discussions

Dr. Susan Galandiuk (Louisville, Kentucky): Up until the last several years, as Dr. Wolff pointed out, there have been few recent publications in the field and many of our practices were, in fact, based on publications and data generated before the routine use of CT-guided abscess drainage. The authors conclude that multiple episodes of recurrent diverticulitis do not necessarily predict poor outcomes. Since they have already selected a population with complicated diverticulitis, in other words, a patient population in which abscesses, phlegmons, fistulas, perforation, or bleeding are present, adverse outcomes in this population would include increased morbidity, mortality, and an increased rate of colostomy.

I have several questions for the authors. Since the study is retrospective, how were the number of episodes of diverticulitis defined? Were these episodes that required hospital admission, such as was the case in studies by Broderick-Villa and the recent study by Flum, or did these include patients with episodes of diverticulitis that were treated as an outpatient with oral antibiotics? The former classification is a much more valid parameter of severity. Similarly, how were these episodes of diverticulitis documented? Were these patients Olmstead County patients in whom data were available, or were these out-of-town patients in whom data on antecedent episodes were not available?

Two major recent similar studies on the frequency of adverse outcomes after recurrent episodes of diverticular disease that required hospitalization have both reported significant correlations with respect to patient age in that younger patients are much more likely to develop recurrent disease. Although the groups here were comparable with respect to age overall, were the group outcomes similar when the groups were stratified within groups with respect to patient age?

When comparing the group of 150 patients that were reported on here that were hospitalized for complicated diverticulitis and who had prior episodes of diverticulitis with that of the 180 patients that were hospitalized for complicated diverticulitis without prior episodes of diverticulitis, were there age differences between these 2 populations? Were there outcome differences?

Despite Dr. Wolff's comments about the deaths in group B being due to anastomotic leaks, the mortality differences between these groups, while they still were not statistically different, there was still a 3-fold difference. Even if you exclude those 2 deaths to the anastomotic leak, with the disparate group sizes, with one group, group A, being 118 patients and group B being 32 patients, was the study powered to detect a difference if it were there?

You report that there was a difference in steroid use between both groups. Did you look at other immunosuppressive drugs? Diverticulitis in the transplant population is a feared complication. Did any of these patients fall into that category? If these patients are excluded, are your study results still valid?

Lastly, a thought-provoking question. I have started treating many of my diverticulitis patients, even those who have undergone CT-guided abscess drainage, conservatively with fiber products, with anti-spasmodics, Tegaserod if needed. What do you think the next several years will bring if you perhaps in 15 years as president of the American Society of Colon and Rectal Surgeons had to write practice parameters for diverticular disease? What will the future bring?

Dr. Jennifer Chapman (Rochester, Minnesota): To answer your first question, both hospital admissions and outpatient treated episodes of diverticulitis were included in this study in the recording of prior history of uncomplicated diverticulitis. These episodes were determined by CT scans, written reports, and medical records recording the diagnosis as well as the treatment of diverticulitis. These standards of making the diagnosis of diverticulitis were used for both inpatient admissions as well as outpatient care of patients.

The majority of the patients in our study were from Olmsted County. However, there was a minority of patients, approximately 20% of the patients, who were not. Most of those patients who were not from Olmsted County brought their medical records for review when they received care at the Mayo Clinic. And those medical records that are brought from outside health institutions are kept in the Mayo Clinic medical records, so we had access to those records.

I believe your second question was regarding age and age stratification. We did not analyze our data set in the first paper, nor in this paper by age stratification, so I cannot give you an accurate assessment of that issue. However, we noted that there were no significant age differences identified between the patients without prior episodes of diverticulitis compared with patients with prior episodes. There have been many studies on patients under age 50 who develop diverticulitis; and in our entire patient population of 330 patients, there were 20 patients in the group that had no prior episodes who were under 50 and 28 patients in the group that had at least one prior episode before age 50, and that was not found to be significantly different.

You asked about the power of the study. You are correct; we think that power might be limited because of the small number of patients, particularly in group B. Since our endpoint was the development of complicated diverticulitis, and since the majority of the patients in our study had less than 2 prior episodes of diverticulitis prior to their complicated episode, we did not find a large number of patients for group B since the disease itself tended to eliminate many candidates from this category. However, it is interesting to see that multiple episodes did not have an increased mortality rate when compared to the patients who had no prior episodes of uncomplicated diverticulitis. That is over 150, compared to 180 patients, and power is likely sufficient in that regard.

As far as steroid usage, yes, there was a difference between steroid use between group A and group B. However, we also performed a logistic regression model to analyze this further and we found that the number of episodes in group A was actually a stronger factor and more significant than steroid use for the development of perforation. So we felt that the number of prior episodes was more of a factor than the steroid use itself. We did not have patients in either group A or group B who were receiving other types of immunosuppression, such as chemotherapy or transplant medication, so this did not cause an effect in the analysis.

In regard to your last question on future practice parameters Dr. Wolff and the co-authors have discussed this thoroughly. We think that, especially for the first paper, it is essential that we start analyzing patients who are at increased risk for perforation and the poor outcomes of complicated diverticulitis. Those are the patients that you mentioned, the transplant patients, immunosuppressed patients, elderly patients, patients with malignancy, and perhaps we should start focusing on those patients; if they have had one episode of acute diverticulitis, then maybe we should consider performing elective sigmoid resection at that point.

As far as further practice parameters, we need to have prospective data. Currently, we are not basing our practice parameters on valid or accurate data, and future changes in our surgical practice with that prospective data will make matters much clearer.

Dr. David N. Herndon (Galveston, Texas): This article has high impact as it challenges traditional surgical guidelines in regard to treatment of complicated diverticulitis. The authors’ guidelines advise elective resection of involved segments of colon after 2 episodes of prior diverticulitis. That seems to be standard in the surgical community.

The authors here convincingly show that patients admitted after 2 such episodes with additional complications have no greater morbidity, lesser perforation, and less mortality than those admitted with less than 2 prior episodes of complicated diverticulitis.

However, if we are to extrapolate from these findings to the logical conclusion that we should not resect after 2 episodes of complicated diverticulitis, we need to analyze further risk. Risk is cumulative. How many episodes of diverticulitis would the authors have individual practitioners subject patients to before they perform elective resection? The comparative risk of elective resection and potential incidence of diverticulitis and its albeit lessened morbidity are germane issues which direct when to resect.

The 40% complication rate with resection of complicated diverticulitis is far greater, obviously, than the complication rate of elective resection. The morbidity of recurrent attacks is serious. Do the authors have any insight into which of their patients are going to go on to have free perforation? Which of those patients will not and when would they have us resect electively?

Dr. Jennifer Chapman (Rochester, Minnesota): Based on Flum's decision analysis published in 2004, Flum and colleagues felt that elective colectomy had the lowest mortality and morbidity and was most cost-effective to the patient after the fourth episode instead of the second episode. I feel our study somewhat supports that conclusion. However, we can't clearly state when is the right time to operate; that was not the objective in this study. We are suggesting that we re-think the practice parameters and base them on more sound evidence.

When I was a resident in surgery, I was taught that we operate to prevent complicated diverticulitis from developing after 2 acute episodes of diverticulitis. I remember hearing my attendings tell the patient, “If you want to prevent an ostomy, you should have this procedure done.” I think we need to re-think this parameter, not only because of the data presented in our study and Flum's but also because of other recent studies which question this practice.

The other fact that we do not address in this study is what we are trying to accomplish with prophylactic colectomy. We are not addressing the issue of patients who have recurrent pain and suffering with multiple episodes, and those patients may well benefit from resection after one or 2 episodes of acute diverticulitis, even though they do not develop complicated diverticulitis.

To respond to your final comment, you are right about the high overall morbidity rate in our study of 40%. However, that rate is based on the patients who developed complications. Most of those patients who developed complications were patients who perforated. In our original cohort of 330 patients, 50% of the patients had no morbidity whatsoever. The 40% total morbidity rate comes from very ill patients who suffered multiple complications, and that accounts for the increased rate.

The patients who seem to be at greater risk for perforation are immunosuppressed, elderly, and transplant patients. I think we should be targeting those patients for prophylactic colectomy. The other patients, who have had multiple attacks, do not seem to have as high a mortality and as significant a morbidity.

Dr. Thomas R. Russell (Chicago, Illinois): The guideline that was established a number of years ago stating that if you had 2 attacks of diverticulitis you should be operated on was truly established in a different era. Also, this guideline was not strictly adhered to by many surgeons. Many attempted to individualize recommendations for surgical care dependent on the type of attacks.

Not only is the literature on this topic old, but a great amount of new technology has been added. Undoubtedly, some of the patients in the past were incorrectly diagnosed with diverticulitis when in reality they may have had irritable bowel syndrome or some other unassociated disease. We are also now entering into a new era of laparoscopic colectomy where removal of the left colon will be seemingly less invasive. Clearly, we experienced this same phenomenon with laparoscopic cholecystectomy with subsequent increase in the number of procedures done in this country and perhaps extension of the number of indications for removal of the gallbladder. Thus, it will be very important that we look at this guideline of 2 attacks given perhaps the greater ease of doing colectomies laparoscopically.

With respect to a previous attack of diverticulitis, many have observed, as I have personally, that an attack of diverticulitis may be somewhat protective. A phlegmon or a mass may occur with a localized perforation, which basically walls off with the omentum or small bowel the area of perforation. This may very well serve as a protection for any subsequent attacks.

This series is a relatively small group of patients; and clearly if the guidelines need to be changed, other groups will need to be involved such as the gastroenterologists. Are there efforts to bring some of this information to those other societies or groups?

Dr. Jennifer Chapman (Rochester, Minnesota): That is an excellent question and goes directly to what we are trying to suggest with this paper. We are not urging changes in practice parameters at this time on the basis of this data, but we are suggesting that some parameters are based on outdated and invalid literature. I agree with you wholeheartedly; we truly need to have a prospective natural history study to better understand the evolution of diverticulitis to complicated diverticulitis and its associated outcomes.

Our current literature and understanding of diverticulitis are based on one of the largest natural history studies on diverticulitis by Parks in 1969. However, in later publications, Parks himself suggests that the diagnosis of diverticulitis in the 1969 study may be incorrect and that a portion of the patients were actually suffering from irritable bowel syndrome, because at the time of this natural history study CT scan was not in use and IBS was an unknown entity. During that time period, they also admitted patients frequently, so there are many patients who did not receive the same medical care as outpatients did with diverticulitis.

We need to develop a large prospective study of patients who have uncomplicated diverticulitis, and follow these patients to determine which patients of that group will go on to complicated diverticulitis and if there are indicators. Outcomes in terms of morbidity and mortality will be more accurate in this type of prospective study.

Approximately 55% of the patients from our cohort had no prior symptoms or signs of diverticulitis but presented with perforation or significant pelvic abscess causing sepsis as their initial manifestation of diverticulitis. Those are the patients on whom we need to focus and aggressively treat. These patients have increased morbidity and mortality, and account for the poor outcomes in complicated diverticulitis, especially perforation. Yet, we have no ability through our current practice guidelines to identify these patients specifically and prevent these very serious attacks.