CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2013;11:15321537

PERSPECTIVES IN CLINICAL
GASTROENTEROLOGY AND HEPATOLOGY
Diverticular Disease: Reconsidering Conventional Wisdom
ANNE F. PEERY and ROBERT S. SANDLER
Division of Gastroenterology and Hepatology, Department of Medicine, University of North Carolina School of Medicine, Chapel Hill, North Carolina

Podcast interview: www.gastro.org/cghpodcast.

Also available on iTunes; see similar articles on
pages 1609, 1614, 1622, and 1631 in this issue
of Clinical Gastroenterology and Hepatology.

Colonic diverticula are common in developed countries and

complications of colonic diverticulosis are responsible for a
signicant burden of disease. Several recent publications
have called into question long-held beliefs about diverticular
disease. Contrary to conventional wisdom, studies have not
shown that a high-ber diet protects against asymptomatic
diverticulosis. The risk of developing diverticulitis among
individuals with diverticulosis is lower than the 10% to 25%
proportion that commonly is quoted, and may be as low as
1% over 11 years. Nuts and seeds do not increase the risk of
diverticulitis or diverticular bleeding. It is unclear whether
diverticulosis, absent diverticulitis, or overt colitis is responsible for chronic gastrointestinal symptoms or worse
quality of life. The role of antibiotics in acute diverticulitis
has been challenged by a large randomized trial that showed
no benet in selected patients. The decision to perform
elective surgery should be made on a case-by-case basis and
not routinely after a second episode of diverticulitis, when
there has been a complication, or in young people. A colonoscopy should be performed to exclude colon cancer after
an attack of acute diverticulitis but may not alter outcomes
among individuals who have had a colonoscopy before the
attack. Given these surprising ndings, it is time to reconsider conventional wisdom about diverticular disease.
Keywords: Colonic Diverticula; Diverticular Disease.

olonic diverticulosis is extremely common in developed

countries. In a consecutive series of 2000 barium enemas at
the Massachusetts General Hospital, approximately two thirds of
adults older than the age of 85 were found to have colonic
diverticula.1 More recently, diverticulosis was found in 71% of
colonoscopies in individuals older than the age of 80.2 Complications of colonic diverticula are responsible for a signicant
burden of disease. In 2009, diverticular disease was the sixth most
frequent outpatient gastrointestinal diagnosis, with 2.6 million
clinic visits.3 In that same year, diverticular disease was the most
common in-patient gastrointestinal diagnosis in the United
States with 283,355 hospitalizations at a cost of 2.7 billion dollars.3 Almost a quarter (22%) of patients admitted with diverticulitis underwent urgent or elective surgery for that condition.4

Diverticula of the large bowel are out-pouchings in the wall

of the colon at weak points in the circular muscle where blood
vessels penetrate to supply the mucosa.5 Diverticula may be
found throughout the colon, but are most common in the
sigmoid colon where abnormalities include thickening and
elastosis of the teniae, shortening of the bowel, and thickening
6,7
and folding of the circular muscle. Diverticulosis most often
is uncomplicated and asymptomatic. Complications may occur,
including bleeding, abscess, perforation, stula, stricture, or
colitis. The term diverticular disease generally is used to denote
diverticulosis with a complication.
Recent reports have challenged long-accepted dogma. For
example, it is widely believed that diverticulosis is the result of a
low-ber diet.8 According to conventional wisdom, 10% to
25% of patients with colonic diverticula will develop
diverticulitis during their lifetime.8 Diverticulosis is thought to
be associated with chronic gastrointestinal symptoms in the
absence of diverticulitis or overt colitis.9 Most physicians
believe that diverticulitis is an infection and must be treated
with antibi10
otics. Elective surgical resection of the affected colon has been
recommended after a second attack of diverticulitis and after a
rst attack in young adults to reduce the risk of recurrence.8
Colonoscopy is suggested after an episode of acute diverticulitis to exclude colon cancer.8
The goal of this article is to review recent publications that
challenge conventional wisdom about diverticular disease.

Fiber
The hypothesis that diverticulosis is a deciency disease
of Western civilization was made popular by Painter and Burkitt
based on their observation that diverticulosis was rare in rural
Africa, but increasingly common in economically developed
countries.11,12 They attributed the difference in disease prevalence to differences in dietary ber. They presumed that the
rural African diet was high in dietary ber and that economically developed countries consumed a low-ber diet. They proposed that this deciency of ber predisposed the population to
diverticulosis. Neither diet nor diverticulosis was actually
measured in their studies and they did not account for important potentially confounding variables such as age and sex.

Abbreviation used in this paper: IBS, irritable bowel syndrome.

2013 by the AGA Institute
1542-3565/$36.00
http://dx.doi.org/10.1 016/j.cgh.2013.04.048

December 2013

Painter13 proposed that the deciency of ber led to constipation and high-pressure segmentation of the colon that
resulted in mucosal herniation through weak sections of the
colon wall. To support their hypothesis they conducted motility
studies that compared intracolonic pressures in patients with
diverticulosis with intracolonic pressures in controls. Although
they reported that patients with diverticulosis had higher
colonic pressures, the investigators only reported pressure
measurements for select cases and there were no statistical analyses. Motility studies of the colon have not consistently shown
that patients with diverticulosis have increased colonic pressures.1419 Along the same lines, Burkitt et al20 conducted
colonic transit studies and found that a population consuming
a Western diet had longer mean colonic transit times and lower
mean stool weights compared with an African population.
Studies in populations with colonic diverticula, on the other
hand, have shown shorter colonic transit times compared with
controls.21,22
The ber hypothesis is extremely popular. The concept of
forceful contractions of the colon leading to herniation makes
sense. However, the hypothesis has persisted for 4 decades largely
without proof. Historically it has been a challenge to refute the
hypothesis because a proper study would require a structural
examination (eg, barium enema) in asymptomatic individuals
to document the presence of diverticula. With the widespread
use of screening colonoscopy we now have the opportunity to
study large numbers of people who are undergoing a structural
examination of their colon in the absence of symptoms.
We recently published a colonoscopy-based, cross-sectional
study of dietary risk factors for diverticulosis.23 Each of the
2104 subjects had a colonoscopy to the cecum between 1998
and 2010. Participants completed a telephone interview that
included a comprehensive semiquantitative food frequency
questionnaire as well as questions about bowel frequency and
physical activity. Contrary to expectation, we found that a highber diet was associated with a higher (not lower) prevalence of
diverticula.23 The association with dietary ber intake was dosedependent and stronger when limited to cases with multiple
diverticula. We also found that constipation was not a risk factor
for diverticulosis. Instead, we found that participants who had
regular bowel movements (7 bowel movements/wk) had a 34%
higher risk of diverticulosis compared with participants who had
less frequent bowel movements (<7 bowel movements/wk).
Study participants were interviewed after their colonoscopy
and were aware, in most cases, that diverticulosis had been
found. In response to the concern that the study may have been
susceptible to response bias and reverse causality from the
subjects knowledge of their diagnoses,24 we performed a second
cross-sectional study in a different population in which we
limited the analysis to participants with no knowledge of their
diverticulosis status. The second study conrmed the results of
the initial work (unpublished data). Similar results with respect
to ber were found in 2 colonoscopy-based studies in nonWestern populations, although diverticula in Asia are found
predominantly in the right colon and may have a different
etiology.25,26
Although recent studies have suggested that a high-ber diet
does not protect against the development of diverticulosis23
there is some evidence that a high-ber diet may protect
against diverticular disease. Crowe et al27 studied 47,033 men
and women in England and Scotland. Individuals who reported

DIVERTICULAR DISEASE 1533

consuming a vegetarian diet had a lower risk of admission to the

hospital or death from diverticular disease. There was also an
inverse association for ber intake. Those in the highest quintile
of ber consumption were 41% less likely to have a complication
from diverticular disease. Similar results were seen in a prospective study of US male health professionals.28
It is important to recognize that the research to date was
designed to improve our understanding of risk factors for
diverticulosis and its complications and not to make dietary
recommendations. There are advantages to eating a high-ber,
plant-based diet with respect to constipation, heart disease,
and cancer.29,30

Risk of Diverticulitis
It is generally reported that 10% to 25% of patients with
diverticulosis will develop diverticulitis during their lifetime.31
This estimate is based on a widely cited review of the natural
history of diverticular disease published by Parks32 in 1975. The
largest case series, published almost 3 decades earlier in 1947,
included 47,000 roentgenologic examinations of the colon.33
Diverticulosis was diagnosed in 8.5% of the examinations. Of
the patients with diverticulosis, 15% were diagnosed with
diverticulitis. Because the study did not include any formal
description of the methods, the indication for the original
roentgenologic examination and the criteria for a diagnosis of
diverticulitis are unknown. A 1958 case series described the
natural history of diverticulosis in 300 patients diagnosed with
diverticulosis by double-contrast barium enema.34 A diagnosis
of diverticulitis was made for patients who subsequently presented with acute constipation or diarrhea, abdominal cramping, localized tenderness, fever, and leukocytosis. In that study,
10% of patients with diverticulosis developed diverticulitis over a
follow-up period of 1 to 5 years. Among those followed up for
6 to 10 years, 25% developed diverticulitis. Smaller case series
also have reported estimates of the risk of developing diverticular disease in those with diverticulosis.35
To accurately calculate the cumulative incidence of diverticulitis it would be necessary to enroll a population of patients
with diverticulosis and no history of diverticulitis. These individuals then would be observed over time for the development
of diverticulitis. The widely quoted estimates are based on
studies that were performed more than 50 years ago and
potentially were biased. The studies began with patients who
sought medical care and subsequently received a diagnosis of
diverticulosis. This approach selects a population more likely to
have diverticular disease and thus likely overestimates the risk
of diverticulitis. Furthermore, these studies only included patients who were followed up in subsequent visits. Patients with
diverticulosis without any symptoms were less likely to attend a
follow-up visit.
Recent data suggest that the lifetime risk of diverticulitis is
lower than commonly cited. Shahedi et al36 estimated the incidence of acute diverticulitis in a retrospective cohort of veterans
with diverticulosis incidentally found on colonoscopy. The risk
of diverticulitis conrmed by computerized tomography (CT)
scan or surgery was 1% over 11 years of follow-up evaluation.
With the widespread use of colonoscopy for screening for
colorectal cancer, many patients are being given a diagnosis of
diverticulosis and then warned that their risk of a complication
is 1 in 4. This prediction appears to be exaggerated.

1534

PEERY AND SANDLER

Nuts and Seeds

For more than 50 years, patients with diverticulosis were
advised to avoid foods that may leave coarse particulate in the
stool (ie, nuts, corn, and seeds).10,34,37 This recommendation
was based on the hypothesis that these particles might obstruct
a narrow-necked diverticulum or physically abrade a diverticulum, thereby initiating a cascade of events resulting in diverticulitis or diverticular hemorrhage. In a landmark study, Strate
et al38 found that dietary nuts, corn, and seeds were not associated with an increased risk of diverticulitis or diverticular
bleeding in a prospective cohort of male health professionals.

Chronic Symptoms
A recent taxonomy of diverticular disease terms distinguishes several types of symptomatic disease.9 The taxonomy
includes 2 types of chronic diverticulitis: chronic recurrent
diverticulitis and segmental colitis associated with diverticulosis.9 Chronic recurrent diverticulitis may begin early after the
initial episode, and may be consistent with failure of the index
episode to settle.39 Segmental colitis associated with diverticulosis is dened as peridiverticular colitis that spares the
rectum.9 It is not surprising that patients with chronic recurrent
diverticulitis and segmental colitis associated with diverticulosis
have symptoms, given their measurable colonic inammation.
Another category of chronic disease has been termed symptomatic
uncomplicated diverticular disease and is dened as diverticulosis
associated with chronic gastrointestinal symptoms in the
absence of diverticulitis or overt colitis.9 The criteria for a
diagnosis of symptomatic uncomplicated diverticular disease do
not include a history of acute diverticulitis.
The literature on symptomatic uncomplicated diverticular
disease is limited and confusing. For example, a small Italian
study administered a general quality-of-life survey (Short-Form
36) to 58 outpatients who met the criteria for symptomatic
uncomplicated diverticular disease and had symptoms of
abdominal pain/discomfort, bloating, tenesmus, diarrhea,
abdominal tenderness, fever, or dysuria.40 Not surprisingly, the
quality-of-life scores were lower than a normative Italian population at baseline. It is not possible to conclude that either the
symptoms or the decrement in quality of life were the result of
diverticulosis. At 6 months, quality-of-life scores had improved,
suggesting that the initial low scores were caused by diverticulitis
or other acute illness at baseline. In another study, 12 patients
with symptomatic uncomplicated diverticular disease were
compared with 13 controls. The diverticular disease patients
were found to have visceral hypersensitivity based on rectal
barostat studies, perhaps mediated by ongoing inammation
and up-regulation of tachykinins based on colonic histology.41
In contrast, a study of 784 subjects who had a structural examination of their colon found that the frequency of abdominal
pain, diarrhea, constipation, and irritable bowel was similar in
patients with and without diverticulosis.42 The study used Rome
criteria for irritable bowel syndrome (IBS) to classify patients and
the response rates were high. Subjects completed questionnaires
before structural examinations. This study casts doubt on
whether diverticulosis is associated with chronic symptoms.
Symptomatic uncomplicated diverticular disease simply may
be irritable bowel syndrome in patients who are found to have
diverticulosis because they undergo colonoscopy. A population-

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 11, No. 12

based, cross-sectional study found an increased risk for colonic

diverticulosis among patients with irritable bowel syndrome
43

compared with those without. Whether IBS and diverticular

disease share an underlying pathophysiology or whether patients with IBS symptoms are simply more likely to undergo
structural examinations that reveal diverticulosis is uncertain.
Some patients appear to have chronic abdominal pain after
a bout of diverticulitis. This postdiverticulitis IBS may be
akin to the IBS that develops after an acute bowel infection
(postinfectious IBS), and would seem more appropriately
considered a form of IBS rather than a type of chronic diverticular disease.
In summary, it is unclear whether diverticulosis, in the
absence of diverticulitis or overt inammation, is responsible for
chronic gastrointestinal symptoms or decreased quality of life.
The prevalence of diverticulosis approaches 80% in 80 year olds.
The vast majority of those individuals do not have symptoms.
We need more research to identify the subset that has symptoms
and to determine why.

The Role of Antibiotics in Diverticulitis

A 2007 New England Journal of Medicine review on
diverticulitis recommended that mild attacks of acute diverticulitis should be treated with 7 to 10 days of oral broad-spectrum
antibiotics.10 The review noted that the pathogenesis of diverticulitis is unknown but restated the common hypothesis that
diverticulitis is the result of an obstruction at the neck of the
diverticulum, mucosal abrasion, low-grade inammation, and
bacterial translocation. Mild diverticulitis is thought to be
associated with microperforation and more severe disease, gross
perforation. In short, all manifestations of diverticulitis have
been presumed to be caused by infection that should be treated
with antibiotics.
A recent multicenter randomized controlled trial in Sweden
and Iceland recruited 623 patients who had typical symptoms,
uncomplicated left-sided diverticulitis documented by CT scan,
and a temperature higher than 38 C.44 Subjects were excluded
if they had an abscess, stula, or free air on CT, or if they had
high fever, peritonitis, or sepsis. Subjects were randomized to
broad-spectrum antibiotics or intravenous uids without antibiotics. Despite randomization, the groups were not completely
balanced, with more episodes of prior diverticulitis in the antibiotic group (44.8%) compared with the nonantibiotic group
(35.6%) (P .02). During the rst 5 days after admission there
were no differences in abdominal pain, temperature, or
abdominal tenderness. There was no difference in mean hospital
stay, complications (perforation or abscess), or recurrent diverticulitis necessitating re-admission to the hospital during the
rst year. In the group randomized to no antibiotics, only 3 (1%)
patients developed an abscess compared with no cases in the
antibiotic group (P .08) The investigators concluded that
antibiotics should be used only in cases of complicated diverticulitis.44 Danish national guidelines for the treatment of
diverticular disease do not recommend antibiotics for the
treatment of uncomplicated diverticulitis (grade A).45 The
guidelines support the use of antibiotics for the treatment of
uncomplicated diverticulitis if the patient is septic, pregnant, or
immunosuppressed based on limited data (grade C).45
The strong evidence from a large randomized controlled trial
showing no benet of antibiotics in uncomplicated diverticulitis

December 2013

raises important questions about etiology and management of

diverticulitis.46 Acute diverticulitis may be an inammatory
rather than an infectious condition.47 Recently, a form of
segmental colitis associated with diverticular disease that
mimics chronic idiopathic inammatory bowel disease histologically has been recognized.48 Although this entity is not
common,49 it might explain some cases of diverticulitis that
improve without antibiotics.

Elective Surgery
Guidelines for the diagnosis and management of
diverticular disease of the colon in adults published in 1999
state that elective (prophylactic) surgery may be reasonable in
patients with recurrent attacks of diverticulitis.8 The recommendation was based on the fact that the risk of recurrent
symptoms after an attack of acute diverticulitis ranged from 7%
to 62%, and because recurrent attacks were less likely to respond
to medical therapy and have a higher mortality rate. The
guidelines suggested that the approach should be individualized
based on the severity and responsiveness of the attack, general
health of the patient, and the risk of surgery compared with the
risk of a future attack.
More recent studies of the natural history of acute diverticulitis, on the other hand, suggest that medically managed acute
diverticulitis has a low recurrence rate and rarely progresses to
complications. In one retrospective cohort study, the risk of
recurrent acute diverticulitis after an initial episode of medically
managed acute diverticulitis was 13% over 9 years.50 A second
retrospective cohort found that the risk of recurrent acute
diverticulitis was 19% over 16 years.51 A study published in
2010 found that 23% of patients had a recurrence.39 Most had a
single recurrence, with only 4.7% having more than 2 episodes of
diverticulitis. After an initial episode of uncomplicated acute
diverticulitis, the risk of complicated disease was 5% over 8
years.39 The risk of recurrence was no greater in complicated
disease (abscess, stricture, stula) than uncomplicated disease,
suggesting the elective surgery should not be recommended
routinely for complicated disease.
Janes et al52 calculated that the risk of an individual requiring
urgent surgery was 1 in 2000 patient-years of follow-up evaluation. They further noted that there was a high complication
rate in surgery for diverticular disease and that 27% to 33% of
patients had ongoing symptoms after bowel resection, not
necessarily attributed to recurrent acute diverticulitis.
The more widespread use of laparoscopic resection might be
expected to decrease the surgical risk for elective colectomy. We
still must balance the risk of surgery with the risk of a
complicated future attack of diverticulitis. The American Society
of Colon and Rectal Surgeons has appropriately recommended
consideration of elective sigmoid colectomy after recovery from
acute diverticulitis on a case-by-case basis, with the decision
based on the age, comorbid disease, the frequency and severity
of the attacks, and whether symptoms persist after the acute
episode.53

Diverticulitis in Younger Patients

Diverticulitis is thought to be more virulent in
younger patients, with 25% to 80% reportedly requiring urgent
surgery during their initial attack.28,5458 The recurrence rate

DIVERTICULAR DISEASE 1535

also has been reported to be higher in younger patients in

older studies.5962 American College of Gastroenterology
practice guidelines for diverticular disease state that elective
(prophylactic) surgery after one episode of uncomplicated
diverticulitis may be reasonably considered in younger patients
based on both the low surgical risk of an elective procedure in
a healthy young person and the many years of future risk of
recurrence.8
More recent studies have challenged the view that diverticulitis is more virulent in younger patients and the recommendation for surgical intervention after an initial event.51
Anaya and Flum51 found that although there was an
increased relative risk for emergency colectomy or colostomy in
younger patients, the absolute risk was low. These investigators
found that 73% of younger patients had no recurrence
and most resolved without surgery. Their work showed that a
policy of routine elective colectomy in a younger population
would require 13 elective surgeries to prevent 1 emergency
colectomy.

Colonoscopy
The American College of Gastroenterology guidelines
recommend colonic evaluation after resolution of clinically
diagnosed diverticulitis to exclude other diagnostic considerations, particularly cancer.8 Because of the potential for perforation as a result of the microabscess that presumably caused
acute diverticulitis, the examination typically is postponed for
at least 6 weeks.
There is an increased risk of colon cancer in the rst year after
a diagnosis of diverticular disease. A population-based, casecontrol study of 41,037 patients with colon cancer found an
increased odds ratio of 25 (95% condence interval, 1738) of a
colon cancer diagnosis within 6 months of an admission for
diverticular disease. There was no association with a colon
cancer diagnosis 12 months after the admission for diverticular
disease. The increased risk of colon cancer within 12 months of
an admission for diverticular disease was attributed to surveillance bias and misclassication.63
If a patient has had a recent colonoscopy before developing
acute diverticulitis, whether there is any use in a repeat endoscopic examination is unknown. Lau et al64 found a number of
cancers and other signicant lesions when they performed a
colonoscopy after an acute attack of diverticulitis. However, they
excluded patients who had a colonoscopy within a year, and
therefore recommended only performing colonoscopic examinations for individuals who had not had a recent radiologic or
endoscopic colonic examination.
A colonoscopy should be performed to exclude colon cancer
after an initial episode of suspected diverticulitis. If a patient
has had a recent colonoscopy before developing acute diverticulitis, the value of a repeat colonoscopy to exclude cancer is
unknown.

Conclusions
Recent work in diverticulosis has created uncertainty
among both researchers and clinicians. The theories of prior
generations have been proven to be questionable and in
some cases unsupportable. A high-ber diet may not protect
against asymptomatic diverticulosis. The risk of developing

1536

PEERY AND SANDLER

diverticulitis is likely a fraction of the 10% to 25% commonly

quoted. Nuts and seeds do not increase the risk of diverticulitis
or diverticular bleeding. The decision to perform elective surgery
after surgical management of acute diverticulitis should be
made on a case-by-case basis. A colonoscopy should be performed to exclude colon cancer after an attack of acute diverticulitis but may not alter outcomes among individuals who
have had a colonoscopy before the attack.
A number of important questions remain. What is the
mechanism of diverticular disease? Is there an important
component of infection, or is diverticulitis something else
entirely? Are there high-risk patients who should undergo early
elective surgery? Are there diets that truly prevent diverticular
disease? We know so little, but at least we are starting to ask the
right questions.

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 11, No. 12