The Gut Microbiome in Diverticulitis and Diverticulosis

Diverticulosis affects 60% of people over 60 and roughly 10-25% of individuals will have symptoms, ranging from acute diverticulitis requiring emergent intervention to symptomatic diverticular disease with persistent lower abdominal pain and altered bowel habits. Up to 5% of patients with acute diverticulitis will die while hospitalized. In spite of the prevalence and severity of its manifestations, fundamental questions remain about the nature of diverticulitis and diverticulosis.

The human colon contains 100 trillion bacteria, including hundreds of species and 10 times as many genes as the human genome. While alterations in the microbiome have been implicated in disease states such as Irritable Bowel Syndrome (IBS) or Inflammatory Bowel Disease (IBD), the microbiome has not been adequately described in relation to colonic diverticula. As King's College Hospital has the only known dedicated Diverticular Disease clinic, we have a uniquely large patient population in which to learn about this condition.

When colonic diverticula become inflamed or perforate, patients have diverticulitis. Inciting factors for episodes of acute diverticulitis are not well characterized, but include age, lifestyle factors and certain medications. Clinicians have no way to predict which patients will get diverticulitis, nor how severe its presentation will be. In some cases of pericolonic diverticulitis (known as Hinchey I), inflammation may resolve even without antibiotics, while in other cases, perforated diverticulitis will include feculent peritonitis (Hinchey IV) and require emergent colon resection. Diverticulitis may be caused by alterations in the microbiome, but this has not been sufficiently addressed.

The majority of patients with asymptomatic diverticulosis, that is, the presence of colonic diverticula in the absence of symptoms, will never experience an episode of diverticulitis. There has been, however, some limited evidence that the microbiome in diverticulosis is altered, and similar to that in acute diverticulitis. If that is so, then predicting which patients will develop acute diverticulitis will remain impossible. It will also mean that research which did not assess or control for the presence of diverticula in patient populations will need to be revisited.

Currently, clinicians have no basis to predict (1) which patients will develop severe acute diverticulitis requiring surgery or (2) which patients will develop acute diverticulitis which will resolve with antibiotics. We believe that further insights into the gut microbiome will allow for better diagnostics and treatments in the future.

There have been a few, limited studies of the microbiome in diverticulitis, but not full pyrosequencing. This is because research costs have fallen 5 orders of magnitude in the past decade while computational capacity has increased. Studies have looked at the individual strains of bacteria within stool or mucosal samples, which does not suffice to understand the complex microbiome.

As the gut microbiome has not been thoroughly investigated in diverticular disease, this is an exploratory study. However, we do wish to test the hypotheses that patients with asymptomatic diverticulosis have a microbiome more similar to patients with acute diverticulitis, than to that of normal controls.

14. Methodology Setting: King's College Hospital, a tertiary referral hospital in south London. King's is a referral centre for the work up of two-week wait for colorectal cancer (2ww) assessments, with 900 new patient referrals assessed per year. Also, the Accident & Emergency Department at King's admits up to 250 patients per year with acute diverticulitis.

Patients: 4 sets of patients will be recruited. 25 patients per group will be recruited:

Groups A and B. Patients with acute diverticulitis will be recruited from the A&E department at King's, prior to commencement of antibiotics. Rectal swabs will be obtained from patients with a chief complaint of acute abdominal pain upon initial assessment, prior to initial antibiotics. All specimens will be initially frozen, but only patients specimens with an admitting diagnosis of acute diverticulitis and colonic imaging (CT imaging or visualization at surgery) will be analysed. Patients with acute diverticulitis will be assessed in 2 groups, those patients with perforated diverticulitis (i.e., Hinchey III or IV) and those patients with unperforated diverticulitis (Hinchey I or II).

Group C. Patients with asymptomatic diverticulosis will be recruited from the 2-week wait (2ww) colorectal cancer pathway. Patients presenting with a chief complaint of fresh rectal bleeding, with no other complaints, will be recruited prior to diagnostic imaging, either by CT, CT colonography, or endoscopy. They will provide a stool sample or rectal swab prior to bowel cleansing, if required for their evaluation, as that may alter the microbiome. Those patient samples with diverticulosis (and or haemorrhoids, as the other most common cause of fresh PR bleeding) will be analysed.

Group D. Patients with or without haemorrhoids, and no other colonic pathology will be recruited from the 2ww colorectal cancer pathway. These will be the normal controls. Patients presenting with a chief complaint of fresh PR bleeding, with no other complaints, will be recruited prior to diagnostic imaging, either by CT, CT colonography, or endoscopy. They will provide a stool sample or rectal swab prior to bowel cleansing, if required for their evaluation, as that may alter the microbiome. Those patient samples with diverticulosis (and or haemorrhoids, as the other most common cause of fresh PR bleeding) will be analysed.

Stool analysis: Stool samples will be analysed for faecal microbiome composition by 16S ribosomal RNA gene pyrosequencing, through a cooperative agreement with the Bruce lab at King's College London.

In brief, gene targeted 16S rRNA sequencing examines a structural component of the bacterial ribosome through targeted primers. Constant portions of the ribosomal RNA are amplified, which allows for identification of the variable portions. Sequences are then clustered in to OTUs, operational taxonomic units and may be assigned to lineages from genus all the way to phyla.

Planned statistical analyses: Microbiome composition will be assessed in several different ways. Alpha diversity will be measured, which is a measure of the average diversity within a sample. Multiple assessments of alpha diversity will be made, including Chaol richness estimate, OTU richness, and Shannon index. Alpha diversity has been shown to be reduced in patients with active IBD as compared to patients with quiescent disease, but has not been described in diverticular disease. Beta diversity is a comparison of diversity between samples (that is, the separation of the phylogenetic structure of the OTUs in one sample as compared to all other samples) will be calculated through Unifrac distances.

Power calculation: Metabonomic studies of the gut microbiome at this time are necessarily exploratory. As such power calculations have not been performed in the field. However, in this instance analysing 25 samples per group should be sufficient to discriminate between groups and assess their diversity.

Existing infrastructure: All infrastructure needed is currently in place. The study team (Professor Bjarnason and Mr Papagrigoriadis) currently manage the 2ww clinics, as well as attend to acute admissions in A&E at King's College Hospital. A materials transfer agreement is in place with the Bruce lab at King's College London for transfer of other faecal samples, and would only need to be amended.