The putative manifestations of colonic diverticular disease are common, ranking eighth for general practice and outpatient visits in the United States, with an estimated economic burden of 4 billion USD per year (1). Diverticulosis is defined by the presence of diverticula, formed by the herniation of mucosa through the muscularis mucosa at sites of vascular penetration (2). The highest prevalence rates exist in North America, Europe, and Australia, whereas it is rarer in rural Africa (2). Diverticulosis prevalence increases with age and affecting up to 60% of people older than 60 years, and is rare in those younger than 40 years (3). Moreover, it is expected to increase, given the overall aging population worldwide.
The vast majority with diverticulosis is asymptomatic, but diverticulosis per se can also be associated with specific symptoms. We previously reported diverticulosis to be associated with diarrhea symptoms independent of age. Furthermore, although the prevalence of irritable bowel syndrome (IBS) and gastrointestinal (GI) symptoms decreased overall with age, individuals older than 60 years and older with diverticulosis were noted to report more abdominal pain and diarrhea predominant IBS compared with those without diverticulosis in the same age group. However, given the low number of observations, caution in interpretation is warranted (3). Diverticular disease may be classified based on symptoms and pathology, from symptomatic uncomplicated diverticular disease (SUDD) to an acute inflammatory presentation with diverticulitis, ranging from segmental colitis associated with diverticulitis, complicated diverticulitis with perforation, abscess, strictures or fistulas, or uncomplicated diverticulitis with colonic wall thickening (4). Although acute diverticulitis, defined as overt inflammation of diverticula, is only believed to occur in 4% in those with diverticulosis (5), a common and chronic presentation with recurrent abdominal pain and change in bowel habits plus reduced quality of life is now recognized (4,6).
The pathogenesis of diverticular disease is multifactorial, complex, and still poorly understood (7). Chronic low-grade inflammation has been postulated to play a role in the pathogenesis and specifically in symptomatic diverticular disease (8–12). Few studies have been published, with conflicting results on the coexistence of symptomatic diverticulosis, diverticulosis per se, and mucosal inflammation (9,13,14). Recently, though, a study of 619 outpatients undergoing screening colonoscopy found no evidence of mucosal inflammation measured as messenger ribonucleic acid (mRNA) expression of tumor necrosis factor alpha (TNF-α), CD4+ cells, CD8+ cells, and CD57+ cells in sigmoid biopsies of normal-appearing mucosa in participants with diverticulosis and/or abdominal pain and IBS compared with controls without diverticulosis (15). However, the presence of active inflammation was measured only by cytokine levels and immunohistochemistry, and only in the sigmoid colon, leaving the remaining colon unexplored. Assessing the whole colon is of value because it allows exclusion of early or unknown colitides of other etiologies such as microscopic colitis or IBD. Furthermore, an early necropsy study comparing colons with and without diverticulosis found an increased number of colonic lymphoid nodules and lymphoid–glandular complexes in colonic segments, not affected by diverticula correlating with possible weak points in the bowel wall and thus suggestive to play a role in formation of diverticula (16). In addition, in the recent study by Peery et al., selection or referral bias may have impacted the results and possible underrepresentation of symptomatic diverticulosis, given the small sample size of symptomatic participants and symptoms limited to only chronic abdominal pain and IBS. We aimed at investigating colon histopathology in asymptomatic and symptomatic diverticulosis by performing a nested case–control study from the well-characterized Swedish cross-sectional population-based colonoscopy study (PopCol), which has the advantage of minimized selection and referral biases. We hypothesized that individuals with diverticulosis, particularly symptomatic diverticulosis, would have low-grade colonic inflammation on histology, not limited to the sigmoid colon compared with individuals without diverticulosis.
This study was part of the Swedish cross-sectional population-based colonoscopy study (PopCol) with an aim to survey colonic pathology and GI symptoms in a normal adult population. The endoscopies were performed for research purpose alone and not with other indications or intentions. Consequently, the endoscopists were instructed of this aim and thus to report all findings per se. A detailed description of the sampling procedure and study population including evidence that the population is a reasonable representation of the adult Swedish population has been previously described (17). The symptoms in this cohort have been previously described in our recent publication in which diverticulosis was associated with loose stools (3). The study was approved by the Ethics Committee (no 394/01, Forskningskommitte Syd). All participants provided written informed consent.
A validated questionnaire on abdominal symptoms (the Abdominal Symptoms Questionnaire) was sent to a random sample of 3,556 individuals (aged 18–70 years) from 2 adjacent urban districts in Stockholm, Sweden. Of the total 2,293 responders, 1,673 were reached by telephone, of which 1,244 were scheduled for a gastroenterologist consultation. At the time of consultation, in addition to the completion of the Abdominal Symptoms Questionnaire and the Rome II modular questionnaire, a gastroenterology consultation including review of medical history and physical examination was conducted. A total of 745 participants (426 women) accepted to undergo a colonoscopy, whereof 130 (17.4%) had diverticulosis, and none had diverticulitis or a previous history of diverticulitis (Figure 1).
When comparing those undergoing colonoscopy vs those who did not, colonoscopy participants were older (51.7 vs 44.3 years), P < 0.001) and reported more abdominal pain (29.1% vs 23.5%, P = 0.04) and alteration of bowel habits namely, high-frequency defecation (P = 0.03), lumpy stools (P = 0.001), bloating (P = 0.03), urgency (P = 0.004), straining (P < 0.001), sensation of incomplete bowel movement (P < 0.001), passing mucous (P = 0.004) and fecal incontinence (P = 0.03), but were otherwise similar in terms of sex (57.2% vs 57.5% women, P = 0.91). While colonoscopy participants reported significantly more GI symptoms compared to non-colonoscopy participants, 31.9 % were still symptom free (17).
Cases and controls
Cases were defined as those participants with diverticulosis present on colonoscopy. Controls were sex and age (±2 years) matched to cases from participants without diverticulosis. Participants with inflammatory bowel disease or microscopic colitis were excluded.
Colonoscopy and biopsy
Preparation included 45 mL oral Phosphoral (CCS Healthcare AB, Borlänge, Sweden) taken twice with a 4-hour interval, and colonoscopies (Olympus Endoscopes, Tokyo, Japan) were performed by experienced endoscopists at Ersta Hospital, Stockholm, from June 2002 to October 2006. Sedation and analgesics were given on demand as per Swedish clinical practice. All colonic segments (cecum, transverse, sigmoid and rectum) were biopsied with standard endoscopic forceps (Boston Scientific, Boston, MA) in all participants. Diverticula were not specifically targeted. The endoscopists were blinded to participants' medical history and questionnaire results but aware of the study's general aims and protocol.
Routinely processed formalin-fixed paraffin-embedded tissue blocks were sectioned and stained with hematoxylin and eosin (H&E) and independently analyzed by 2 pathologists blinded to the case–control status and symptoms. To evaluate reproducibility or interobserver variability, both pathologists shared and reviewed the slides at the beginning of the project. Both conferred on concordance of the results for the first 20 slides, which were totally in accord, and any discrepancies or queries thereafter were resolved by consultation. Most cases and controls were in fact normal. The H&E-stained biopsy slides were assessed for the presence or absence of markers of histological inflammation—intact surface epithelium, mucin depletion, Paneth cells, cryptitis or crypt abscesses, apoptosis of the epithelium, normal architecture/crypt branching, intact chronic inflammatory gradient (no expansion of chronic inflammatory cells from base to surface), basal plasmacytosis, and granulomas. The number of lymphoid aggregates, follicles, and neutrophils was counted in the lamina propria, and intraepithelial lymphocytes were counted/100 colonocytes (18). A comment was made of any other pathology noted.
Background variables and gastrointestinal symptoms
Body mass index was calculated as weight in kilograms divided by the square of height in meters. Smoking was reported as nonsmoker or current smoker. Recent antibiotic use was defined as any report of antibiotic exposure within the last 3 months before colonoscopy.
Gastrointestinal bowel symptoms were derived from the Rome II modular questionnaire (19) filled out during the preparation hospital visit. The Rome II modular questionnaire, in use since 1996, includes 38 questions describing the severity and frequency of recorded symptoms presented per organ (see Supplementary document, Supplemental Digital Content, http://links.lww.com/AJG/A47). The Rome II modular questionnaire questions begin with the sentence “In the last 3 months, did you often have….” “Often” was defined as present at least 1 day per week during 3 weeks for the past 3 months. The GI bowel symptoms included in the present study were abdominal discomfort or pain, mushy stools (loose, mushy, or watery stools), high-frequency defecation (more than 3 bowel movements a day), lumpy stools (hard or lumpy stools), infrequent defecation (1 bowel movement less than every 3 days), bloating (abdominal fullness, bloating, or swelling), urgency (having to rush to the toilet to have a bowel movement), straining (a sensation that the stool cannot be passed (i.e., blocked) when having a bowel movement), sensation of incomplete bowel movement, passing mucous (slime) during a bowel movement, and fecal incontinence (“In the last year, when you had constipation or diarrhea, did you accidentally leak or pass stool for more than 1 occasion in a month”). We defined diarrhea to be positive if there was presence of high-frequency stools (more than 3 bowel movements a day), mushy stools (loose, mushy, or watery stools), or both.
Proportions were calculated for categorical data, mean and s.d. for continuous data, and medians for data with skewed distribution. To assess associations between diverticulosis with or without symptoms and mucosal markers, logistic and linear regression models were used when appropriate. Multivariable models were adjusted for age and sex. All analyses were performed using Stata (13.1 StataCorp, Texas, TX). Results were considered significant when the 2-sided P value was <0.05.
Missing data left 127 of 130 cases with known localization of diverticulosis in the analysis. Participants with diverticulosis (cases) and those without diverticulosis (controls) were similar in age [60.8 years (7.5) vs 60.7 years (7.4), P = 0.92], and sex distribution was equal (45.7% males, P = 1.00). Table 1 presents participants' characteristics and descriptive statistics by diverticulosis status. All cases had diverticula in the sigmoid colon, but only 8 (6.3%) had both the right and left colons affected. Body mass index (P = 0.71), smoking (P = 0.34), and recent antibiotic exposure (P = 0.68) were similar between cases and controls. The characteristics in this nested case–control population is congruent with the previously reported results, with cases reporting more mushy and high-frequency stools and more abdominal pain (P = 0.01, P = 0.03, and P = 0.04, respectively) than controls (3). Gastrointestinal symptoms were otherwise similar between the groups.
Diverticulosis and inflammation
The median C-reactive protein (CRP) levels were similar among cases and controls [1.05 mg/L (0.3, 2.7) vs 0.8 (0.4, 2.2), P = 0.53]. The results of histopathology assessment of inflammatory cells by colon site are depicted in Table 2. The surface epithelium and chronic inflammatory gradient were intact in all participants, and basal plasmacytosis and granulomas were absent in all samples and are thus not presented. There was a trend of increased numbers of cecal lymphoid aggregates in cases vs controls (P = 0.07), but no other associations between diverticulosis and inflammatory markers were found (Figures 2 and 3).
Symptomatic diverticulosis and inflammation
Overall, 45 cases (38%) and 30 controls (25%) reported diarrhea out of 241 participants with data on GI symptoms. Cases and controls, with or without diarrhea, were similar between the groups in all background variables except for recent antibiotic use, which was 2-fold more likely to be reported by cases with diarrhea in comparison with asymptomatic controls [odds ratio (OR) 2.50, 95% confidence interval (CI), 1.01–6.15, P = 0.047]. There were no significant differences in median CRP levels (P = 0.23) or mucosal inflammatory markers comparing symptomatic cases, asymptomatic cases, and symptomatic controls to asymptomatic controls (Tables 3 and 4).
Abdominal pain was reported in 34 cases (29%) and 21 controls (17%). In terms of background variables, controls with abdominal pain were 5-fold more likely to report the recent antibiotic use than asymptomatic controls (OR 5.03, 95% CI, 1.60–15.86, P = 0.006), and cases with abdominal pain had a trend to be active smokers (OR 2.51, 95% CI, 0.93–6.72, P = 0.07) in comparison with asymptomatic controls, respectively. No other differences in background variables were noted between cases and asymptomatic controls.
Cases with abdominal pain did not differ in median CRP levels (P = 0.19) or in mucosal inflammatory markers vs asymptomatic controls (Tables 5 and 6). The presence of cryptitis, albeit rare, was similar within the colonic segments among cases and controls except for in the transverse colon. Two of the 7 individuals with cryptitis present in the transverse colon were controls with abdominal pain and were 16 times more likely to have cryptitis present in the transverse colon than asymptomatic controls (OR 16.12, 95% CI, 1.31–197.96, P = 0.03); an analysis adjusted for smoking and recent antibiotic exposure provided similar results (OR 17.38, 95% CI, 1.29–233.24, P = 0.03), but the analysis did not converge when the recent antibiotic users were excluded. Although cases with abdominal pain had significantly less rectal lymphoid aggregates present than asymptomatic controls (P = 0.022), asymptomatic cases had significantly more cecal lymphoid aggregates present vs asymptomatic controls (P = 0.045).
In this large Swedish population-based cohort, we found no association between diverticulosis and serological or mucosal inflammation, in any colonic segment, regardless of the location of diverticula. Moreover, there was no association between symptomatic diverticulosis with abdominal pain or diarrhea and serological or mucosal inflammation.
Conflicting literature exists on the role of inflammation in diverticulosis with or without symptoms. Several studies support the presence of chronic inflammation in symptomatic diverticular disease. Kealy et al. (16) reported the density of microscopic lymph follicles and aggregates to be increased in necropsied colons of participants with diverticular disease compared with those from participants without diverticular disease, suggesting that lymphoid follicles may be weak points in the mucosa, and diverticula could develop at these points. Chronic low-grade inflammation has also been reported within and around diverticula in participants who underwent sigmoid resection for SUDD but not diverticulitis (20) and as evident by overexpression of TNF-α in participants with a history of diverticulitis and with diverticular disease (13). However, these studies were not population based and included patient populations with SUDD, history of diverticulitis, or significant symptoms requiring surgical resection as symptomatic treatment, raising concern for selection bias. Furthermore, pathologists were not blinded to the symptom status in all the studies.
Our study complements the recent findings by Peery et al. (15) demonstrating the absence of colonic mucosal inflammation in diverticulosis or symptomatic diverticulosis. Although both studies were colonoscopy based and included study populations with the absence of past diverticulitis, the present study was population based and assessed the whole colon for inflammation, regardless of localization of diverticula, whereas Peery et al. assessed only the sigmoid colon. Furthermore, although Peery et al. measured the presence of active immune response by TNF-α and cytokine levels, we evaluated the actual presence of inflammatory cells, changes, and gradients.
Similarly, in the symptomatic diverticulosis subanalysis, we found no association between symptomatic diverticulosis with abdominal pain or diarrhea and serological or mucosal inflammation throughout the colon. These findings extend both Peery et al. and Elli et al.'s recent findings of an absence of colonic mucosal inflammation in the sigmoid, as measured by levels of serological immune markers and histological cytokine levels in participants with symptomatic diverticulosis (IBS or abdominal pain) and SUDD compared with controls without diverticulosis (14,15).
Other findings in the present study are also important to highlight. Although symptomatic controls with abdominal pain were 16-fold more likely to have cryptitis in the transverse colon than asymptomatic controls (OR 16.12, 95% CI, 1.31–197.96), this correlation was based on only 7 individuals and likely due by chance. Thus, results need repeating before confirmation. Furthermore, when excluding all recent antibiotic users, the analysis did not converge. Second, in line with our previous results (3), we found that diverticulosis per se is associated with GI symptoms supporting the concept that diverticulosis is associated with a higher prevalence of GI symptoms. Moreover, we found that recent antibiotic use was 2-fold more common in cases with diarrhea (P = 0.047) and 5-fold more likely in symptomatic controls with abdominal pain (P = 0.006) compared with asymptomatic controls, respectively. Lastly, there was a trend that cases with abdominal pain were active smokers (P = 0.068) in comparison with asymptomatic controls. Antibiotic use per se is associated with GI symptoms, usually diarrhea (21) and smoking is associated with abdominal pain, visceral sensitivity (22) and symptomatic diverticular disease (23), while changes in the microflora have been postulated as a symptomatic trigger in both scenarios (21,24). To evaluate confounding by smoking and recent antibiotic use, we performed a sensitivity analysis, but the results were essentially unchanged. Hence, although causality cannot be established and caution in interpretation is needed, given the small sample size and wide CIs in the subanalysis of those with abdominal pain and diarrhea, our findings suggest an alternative pathophysiological mechanism to inflammation in diverticulosis.
This study has several strengths. The population-based study design minimizes selection bias and improves the generalizability. Gastrointestinal symptoms were well characterized using modified Rome II questionnaires and the sex and GI symptom similarities between included and excluded participants' support that the study population was a good representation of the overall population. The diagnostic method for identifying diverticulosis was colonoscopy. Biopsies were obtained in standard uniform fashion from each colonic segment and not limited to the location of diverticula. Histology analyses were performed by 2 independent pathologists with good interconcordance reliability, blinded to both diverticulosis and symptom status. Confounding variables were measured and accounted for in the analyses. Similarly, rifaximin, a poorly absorbable antibiotic used in the IBS, was not available in Sweden until 2013, long after the study completion, thus further minimizing possible confounding.
Some important limitations need to be mentioned. Biopsies from the colonic segments were randomly taken and not targeted to within or around the diverticula. None of the participants in our study population had previous history of diverticulitis; we cannot exclude a severe disease subgroup, which has recurrent episodes of acute diverticulitis where colonic inflammation plays a role, although randomized trials with mesalazine in prevention of recurrent diverticulitis have been disappointing (25,26).
Another potential limitation is the lack of fecal calprotectin measurement as a correlating inflammatory marker. However, fecal calprotectin was not readily available in Sweden at the time of the study, and moreover, histology remains the gold standard for evaluating mucosal inflammation. Similarly, the presence of macrophages, which are reportedly increased in participants with SUDD and diverticulosis per se, was not assessed in the present study because immunohistochemistry was not performed (27). Furthermore, the new Diverticular Inflammation and Complication Assessement (DICA), in which the extent of diverticular disease has been purported to be a predictor of severity of disease, was not available at the time of the study (28). Although the DICA classification was not applied in our study, most of our participants had limited disease (mainly sigmoid colon) corresponding to a DICA score of 1, with only 8 participants (6.3%) having extensive disease, suggesting that limited diverticulosis without previous history of diverticulitis is not associated with inflammation throughout the colon. Missing data secondary to insufficient material for histopathology evaluation is another limitation. However, given the random effect with no significant difference in included and excluded cases in terms of age, sex, or symptoms, it is unlikely to have affected the results. The lack of information on individuals' diet and previous knowledge of diverticulosis might have been a limitation because diet and possibly bowel habits may have been altered in participants with diverticulosis. Last, underreporting of over-the-counter Non-Steroidal Anti-Inflammatory Drug use, possibly masking histological inflammation in the samples, is possible, but detailed questioning of all medication use occurred twice on 2 different occasions, during the initial gastroenterology visit and at the time of endoscopy, to minimize recall bias. Hence, we feel confident that medications did not impact our results.
In conclusion, we found no evidence of low-grade colonic inflammation in participants with diverticulosis, regardless of GI symptom status, challenging the reasoning of using of anti-inflammatory agents in symptomatic diverticulosis. Our results further highlight the importance of future studies evaluating other explanations for symptomatic diverticulosis, especially dysbiosis in diverticulosis and symptomatic diverticulosis.
CONFLICTS OF INTEREST
Guarantor of the article: Maria Ellionore Järbrink-Sehgal, MD.
Specific author contributions: M.E.J.-S.: study hypothesis, analysis and interpretation of data, statistical analysis, and drafting of the manuscript. L.R.: analysis and interpretation of data. A.J.: analysis and interpretation of data. M.M.W.: analysis and interpretation of data and critical revision of the manuscript. N.J.T.: study idea and critical revision of the manuscript. L.A.: study idea and critical revision of the manuscript. A.A.: study hypothesis, statistical analyses, and critical revision of the manuscript. P.T.S.: study idea, analysis and interpretation of data, and critical revision of the manuscript. We thank Dr. Lars Kjellström for data and sample collection.
Financial support: This work was supported by Magtarmförbundet, Bengt Ihre Stiftelsen, Stockholm County Council (ALF project), and Signe och Olof Wallenius Stiftelse, Ersta Hospital, Astra Zeneca, Stockholm, Sweden, Ruth and Richard Julin foundation, Nanna Svartz foundation.
Potential competing interests: None.
WHAT IS KNOWN
- ✓ Diverticular disease is age prevalent, common, and costly.
- ✓ Pathogenesis is multifactorial and poorly understood.
WHAT IS NEW HERE
- ✓ Diverticulosis, regardless of symptom status, is not associated with mucosal or serological inflammation in a general community sample.
1. Peery AF, Crockett SD, Barritt AS, et al. Burden of gastrointestinal, liver, and pancreatic diseases in the United States. Gastroenterology 2015;149:1731–41.e3.
2. Stollman NH, Raskin JB. Diagnosis and management of diverticular disease of the colon in adults. Ad hoc practice parameters committee of the American College of Gastroenterology. Am J Gastroenterol 1999;94:3110–21.
3. Järbrink-Sehgal ME, Andreasson A, Talley NJ, et al. Symptomatic diverticulosis is characterized by loose stools. Clin Gastroenterol Hepatol 2016;14:1763–70.e1.
4. Strate LL, Modi R, Cohen E, et al. Diverticular disease as a chronic illness: Evolving epidemiologic and clinical insights. Am J Gastroenterol 2012;107:1486–93.
5. Shahedi K, Fuller G, Bolus R, et al. Long-term risk of acute diverticulitis among patients with incidental diverticulosis found during colonoscopy. Clin Gastroenterol Hepatol 2013;11:1609–13.
6. Elisei W, Tursi A. Recent advances in the treatment of colonic diverticular disease and prevention of acute diverticulitis. Ann Gastroenterol 2016;29:24–32.
7. Walker MM, Harris AK. Pathogenesis of diverticulosis and diverticular disease. Minerva Gastroenterol Dietol 2017;63:99–109.
8. Spiller R. How inflammation changes neuromuscular function and its relevance to symptoms in diverticular disease. J Clin Gastroenterol 2006;40(Suppl 3):S117–20.
9. Humes DJ, Simpson J, Smith J, et al. Visceral hypersensitivity in symptomatic diverticular disease and the role of neuropeptides and low grade inflammation. Neurogastroenterol Motil 2012;24:318–e163.
10. Horgan AF, McConnell EJ, Wolff BG, et al. Atypical diverticular disease: Surgical results. Dis Colon Rectum 2001;44:1315–8.
11. Böttner M, Wedel T. Abnormalities of neuromuscular anatomy in diverticular disease. Dig Dis 2012;30:19–23.
12. Golder M, Burleigh DE, Belai A, et al. Smooth muscle cholinergic denervation hypersensitivity in diverticular disease. Lancet 2003;361:1945–51.
13. Tursi A, Elisei W, Brandimarte G, et al. Musosal tumour necrosis factor α in diverticular disease of the colon is overexpressed with disease severity. Colorectal Dis 2012;14:e258–63.
14. Elli L, Roncoroni L, Bardella MT, et al. Absence of mucosal inflammation in uncomplicated diverticular disease. Dig Dis Sci 2011;56:2098–103.
15. Peery AF, Keku TO, Addamo C, et al. Colonic diverticula are not associated with mucosal inflammation or chronic gastrointestinal symptoms. Clin Gastroenterol Hepatol 2018;16:884–91.e1.
16. Kealy WF. Lymphoid tissue and lymphoid-glandular complexes of the colon: Relation to diverticulosis. J Clin Pathol 1976;29:245–9.
17. Kjellström L, Molinder H, Agréus L, et al. A randomly selected population sample undergoing colonoscopy: Prevalence of the irritable bowel syndrome and the impact of selection factors. Eur J Gastroenterol Hepatol 2014;26:268–75.
18. Feakins RM, British Society of Gastroenterology. Inflammatory bowel disease biopsies: Updated British society of gastroenterology reporting guidelines. J Clin Pathol 2013;66:1005–26.
19. Drossman DA. Rome II. The Functional Gastrointestinal Disorders. 2nd edn. McLean, VA: Degnon Associates; 2000.
20. Goldstein NS, Ahmad E. Histology of the mucosa in sigmoid colon specimens with diverticular disease: Observations for the interpretation of sigmoid colonoscopic biopsy specimens. Am J Clin Pathol 1997;107:438–44.
21. Högenauer C, Hammer HF, Krejs GJ, et al. Mechanisms and management of antibiotic-associated diarrhea. Clin Infect Dis 1998;27:702–10.
22. Zvolensky MJ, Bakhshaie J, Norton PJ, et al. Visceral sensitivity, anxiety, and smoking among treatment-seeking smokers. Addict Behav 2017;75:1–6.
23. Järbrink-Sehgal ME, Schmidt PT, Sköldberg F, et al. Lifestyle factors in late adolescence associate with later development of diverticular disease requiring hospitalization. Clin Gastroenterol Hepatol 2018;16:1474–80.e1.
24. Opstelten JL, Plassais J, van Mil SW, et al. Gut microbial diversity is reduced in smokers with crohn's disease. Inflamm Bowel Dis 2016;22:2070–7.
25. Kruis W, Kardalinos V, Eisenbach T, et al. Randomised clinical trial: Mesalazine versus placebo in the prevention of diverticulitis recurrence. Aliment Pharmacol Ther 2017;46:282–91.
26. Urushidani S, Kuriyama A, Matsumura M. 5-aminosalicylic acid agents for prevention of recurrent diverticulitis: A systematic review and meta-analysis. J Gastroenterol Hepatol 2018;33:12–9.
27. Barbara G, Scaioli E, Barbaro MR, et al. Gut microbiota, metabolome and immune signatures in patients with uncomplicated diverticular disease. Gut 2017;66:1252–61.
28. Tursi A, Brandimarte G, Di Mario F, et al. Predictive value of the Diverticular Inflammation and Complication Assessment (DICA) endoscopic classification on the outcome of diverticular disease of the colon: An international study. United Eur Gastroenterol J 2016;4:604–13.
Supplemental Digital Content
© The American College of Gastroenterology 2019. All Rights Reserved.