Non Specific Ulcerations On Capsule Endoscopy Biology Essay

Published:

Due to the length of the small bowel, it is often difficult to image. Traditional techniques are often not ideal; endoscopic examinations such as ileoscopy and push enteroscopy are not only invasive, but also rarely achieve complete visualisation. Ileoscopy can realistically only extend to the terminal ileum and whilst often superior to traditional endoscopes, push enteroscopy still has a limited range [5, 11]. Recently, balloon enteroscopy has been developed (double and single balloon enteroscopy); this has improved the extent of the small bowel that can be visualised using a scope by incorporating overtubes and balloons into the scope. Through a series of steps involving inflating and deflating the balloons in advancement cycles, scoping by this method has allowed a much larger proportion of the small bowel to be examined via both the anterograde and retrograde routes, in some instances offering more than twice the scoping ability in terms of length when compared to traditional push enteroscopy or ileoscopy. Balloon enteroscopy is however an incredibly time consuming, invasive process requiring sedation and is relatively high risk, with a reported complication rate of 0.8%, rising to nearly 4% in instances where interventions are also administered through the scope [2, 38, 41].

Lady using a tablet
Lady using a tablet

Professional

Essay Writers

Lady Using Tablet

Get your grade
or your money back

using our Essay Writing Service!

Essay Writing Service

Contrast studies such as barium small bowel follow through combat the problem of invasiveness and complete visualisation, however are not ideal, as they are poor at detecting mucosal lesions and subtle inflammatory lesions that need direct line of sight [4,5,11]. They also expose the patient to high doses of radiation [13]. As such, magnetic resonance (MR) enterography is now beginning to replace barium small bowel follow through as a safe and effective alternative to image the small bowel. Whilst it does offer a potentially safer way to image the small bowel and has benefits over barium small bowel follow through (and indeed traditional endoscopic examinations) in its ability to detect extraluminal pathology, there still remains the drawback that it can miss subtle lesions [4, 39, 40].

After being cleared for use by the FDA in 2001, CE has swiftly been recognised as an imaging modality that is not only well tolerated and minimally invasive, but also able to image the entire small intestine [1, 2]. This is helped by the high (0.1mm) resolution of the lens incorporated into the capsule, which can detect pathologies missed by other imaging techniques, for example denuded and ulcerated lesions along with many others [4]. Indeed, a recent pooled data analysis by Lewis et al. demonstrated that CE had a miss rate of only 0.5% for ulcers [24]. Several studies have also suggested that CE has a significant incremental yield in the diagnosis of obscure gastrointestinal bleeding (a fourfold chance of a positive finding when compared to other imaging modalities [25]) and Crohn's Disease (CD) compared to traditional imaging techniques [25] and it is therefore now widely used to evaluate patients ranging from obscure GI bleeding to malabsorption syndromes [1]. It has also been used to study the adverse effects of NSAID use on the small intestine [26].

Use in CD

Until the advent of CE, the diagnosis of CD was made on the basis of traditional imaging techniques (eg. radiology), with histological diagnosis reserved for only a minority of these patients. Image based diagnosis, combined with history and physical examination was therefore the usual procedure as there is no definitive gold standard [4, 5]. Many cases of CD are in the small intestine (and 30-40% are in isolation here), although it can manifest anywhere in the gastrointestinal tract [2]. Whilst CE has indeed opened up the avenue of fully imaging the small bowel lumen, the diagnostic yield of CE is low when performed in patients with abdominal pain as the only referring complaint. The yield is slightly higher in patients with abdominal pain and diarrhoea [27, 37, 5]. However, if other criteria are present in addition to this (such as raised inflammatory serum markers eg. ESR, CRP) prior to the decision to perform CE, this diagnostic yield increases further [5, 27]. It has also been noted in a meta analysis that in patients who are suspected to have CD but in whom all other investigations have been negative, CE does not have a significantly different yield. Its forte seems to be confirming disease activity in those patients with diagnosed CD [2].

Whilst several pieces of literature do point to the yield of CE being higher with certain referral criteria, there still remains the problem that the visual findings that CE detects can be shared between pathologies. For example, mucosal breaks and other lesions that have been attributed to CD can also be found in patients treated with NSAIDs, and in addition to this, approximately 14% of healthy subjects have been found to have mucosal breaks and erosions in the small bowel. This non-specificity with regards to findings is a large hurdle to overcome when trying to accurately diagnose CD on the basis of imaging [6, 29].

NSAID Enteropathy

Lady using a tablet
Lady using a tablet

Comprehensive

Writing Services

Lady Using Tablet

Plagiarism-free
Always on Time

Marked to Standard

Order Now

Whilst NSAIDs have been a highly successful class of drugs, the major limitation on their use is their association with serious gastrointestinal side effects such as dyspepsia, ulceration, bleeding and perforation [7]. These side effects have previously been thought to manifest mainly in the gastroduodenal area, with only indirect evidence of involvement throughout the gastrointestinal tract and case reports linking NSAIDs to lesions encountered such as "bleeding, protein loss, strictures and increased intestinal permeability" [7]. Other lesions include denuded areas, petechiae and mucosal breaks [9]. With advances in imaging techniques (such as CE), more has become known about the side effects of NSAIDs, such that NSAID enteropathy is now clearly established, with the small bowel being cited as more common than the stomach with regards to side effects of NSAIDs [7,8]. Indeed, studies using CE have demonstrated that ingestion of NSAIDs can lead to small bowel injury [30]; small bowel inflammation has been found in approximately 60% of long term NSAID users [8], and visible small bowel ulcers have been found in more than 25% of patients on NSAIDs [8, look up 5, 6].

The risk of small bowel injury has also been found to extend further than traditional NSAIDs, to the class of drugs known as COX-2 selective agents. These were designed to reduce the gastrointestinal side effects of non selective NSAIDs by reducing inhibition of COX-1 (which is constitutively active in many tissues and has a role in mucosal defence and repair [32]). When taken for a short period of time these agents have been found to cause very little damage [31] however, when these drugs are taken long term, they have been associated with damage almost comparable to that of traditional NSAIDs [7].

Though the general risk of small bowel injury in chronic NSAID users is relatively high, some lesions are more common than others. Maiden reported the most common lesions in patients on long term NSAIDs as denuded areas (39%), mucosal breaks (29%), petechiae/red spots (34%) and reddened folds (13%). Strictures, which are considered to be the hallmark of NSAID enteropathy (characteristic concentric 2-4mm septa, ranging from 3-70 in each patient which tend to be localised to the mid small intestine however can be found in the caecum in patients on certain NSAID preparations [8]), were found in 2% of subjects on long term NSAIDs. These are speculated to arise after many sustained cycles of inflammation, granulation and fibrosis of circumferential mucosal folds [7]. Surprisingly, the findings in patients on COX-2 inhibitors were only slightly different; the most common lesions were petechiae/red spots (28%), mucosal breaks (22%), denuded areas (18%) and reddened folds (8%) [15]. In addition to these findings by Maiden, Morris et al. reported that 7 of 15 patients taking long term NSAIDs had jejunal or ileal ulceration upon evaluation with sonde enteroscopy [10].

As with CD however, NSAID enteropathy is hard to diagnose merely on the basis of CE.

Non Specific Ulceration

Non specific ulceration has been recognised as a separate entity, characterised by multiple small, histologically nonspecific intestinal ulcers and chronic, persistent blood and protein loss [33]. Although hard to distinguish from NSAID enteropathy, non specific ulceration (chronic nonspecific multiple ulcers of the small intestine, CNSU) has been reported to have differences in the sites and patterns of occurrence along with differences in ulcer activity [10].

A common clinical problem is distinguishing between differential diagnoses, as findings are often shared between diagnoses; for example as mentioned earlier, ulcers could be due to active CD, NSAID enteropathy, or could even be incidental findings in an otherwise healthy patient. Ulceration is a particular problem, with imaging often suggesting CD when this may not be the case [8]. This is particularly important given that often the diagnosis cannot be confirmed histologically, and NSAID use remains undisclosed amongst many patients; a recent study by Sidhu et al. highlighted the finding that 13.6% of patients do not mention that they are taking NSAIDs even when asked [9]. Whilst histological examination can provide clues regarding drug induced aetiology, CD and NSAID enteropathy are still hard to distinguish [34]. Non specific ulceration, which is also now recognised as a separate clinical entity is also hard to distinguish.

There are other problems with using CE to diagnose CD. One such problem is the lack of standardised terminology to describe lesions. Such systems have been compiled; however they have not necessarily been implemented [1, 4]. Another problem is the lack of a universally accepted scoring system for reporting these findings taking into account extent and severity, and the lack of threshold diagnostic criteria/scores to support diagnoses. One of the most widely used diagnostic criterion for CD is the presence of more than 3 ulcerations (as long as NSAID use has been excluded) as proposed by Mow et al. [28]. Fidder et al defined a positive diagnosis for CD as "four or more ulcers, erosions or a region with clear exudate and mucosal hyperaemia and oedema" [1]. The first index, known as the Lewis score, was later created using five endoscopic parameters derived from structured terminology that has been developed for CE: "erythema, oedema, nodularity, ulcer and stenosis" [1, 42]. Gralnek et al. modified this index and the final proposed index (Figure 1) included "villous oedema, ulcer and stenosis" [1]. This is the index that is included within the software on PillCam SB CE workstations (Given Imaging Ltd., Yoqneam, Israel). It works on the basis of splitting the small bowel into three sections; scoring villous appearance/ulceration for each section and scoring global stenosis at the end. This generates a score ranging from 8 to 4800 points (comprising of the worst section score plus the stenosis score) with <135 being normal or clinically insignificant mucosal inflammatory change, 135-790 indicating mild CD and ≥790 indicating moderate to severe CD [1].

Lady using a tablet
Lady using a tablet

This Essay is

a Student's Work

Lady Using Tablet

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Examples of our work

With this in mind, this project aims to shed some light on whether the currently used parameter for multiple ulcers (≥8 ulcers) in the Lewis score that was modified by Gralnek et al. (Figure 1) is able to distinguish between the final diagnoses of NSAID enteropathy and CD, and whether the presence or absence of inflammation is also another distinguishing factor.

Hypothesis

Patients with ≥8 ulcers on CE are more likely to have CD than NSAID enteropathy

Patients with inflammation on CE are more likely to have CD than NSAID enteropathy

Methods

Data was collected from a possible cohort of 425 patients referred to Charing Cross Hospital, London for capsule endoscopies between the dates of January 2008 to March 2011. Initially, CE reports for all the patients reported on during these dates were studied in order to identify those patients who had the presence of ulceration and to record both the indication for the procedure and the diagnosis if one was made using only the evidence from the study. After identifying these patients, their CE studies were examined on at least two separate occasions in order to accurately determine the number and size of ulcers (an apthous ulcer was classed as 1-4 villi visible within the boundary of the ulcer, a small ulcer as 5-10 villi visible and a large ulcer as >10 villi visible (Figure 2)), along with any other associated findings and inflammation. Any images of questionable pathology that were found were verified via a second opinion from an experienced endoscopist.

After this initial stage of the study was performed further information regarding the final diagnosis was acquired. This involved looking at the pathology reports (using the computerised ICE system) and correspondence for each patient and requesting patient notes in the instance where the former two sources did not yield a final diagnosis.

Statistics

Having acquired the data for each patient, the data was sorted into discrete groups in order to test the hypotheses. Statistical tests were carried out using the GraphPad Prism 5 program. In order to evaluate whether the number of ulcers differed between the CD and NSAID group, the Mann-Whitney U test was used which compared the medians between the two groups. This is because the data was judged to be non-parametric, and as such a Student's t-test could not be used. To compare the data between groups, contingency tables were constructed comparing the Crohn's and NSAID groups with respect to ulceration (≤7 and ≥8 ulcers) and inflammation (presence or absence). The two-tailed Fisher Exact Test was used to evaluate the significance of the findings; due to the fact that the groups were small, it would have been overly optimistic and unrealistic to use a Chi-squared test.

Results

CE reports for 425 patients who had been referred to Charing Cross Hospital, London were studied. 82 patients (19%) were reported as having the presence of ulceration on CE, and were included in the cohort for analysis. Of these 82 patients, 39 patients (48%) were diagnosed with either NSAID enteropathy or CD. The remainder were either normal (15 patients), had inconclusive diagnoses (13 patients), were lost to follow up (4 patients) or had other diagnoses eg. Angiodysplasia, Coeliac disease (11 patients) (Table 1).

Table 2 shows the number of ulcers recorded in each patient, along with the presence or absence of inflammation. Reason for referral and final diagnosis is also included for each patient. Of the entire cohort, the most common reasons for referral (Table 3) were Iron Deficiency Anaemia (40%) and suspected Crohn's disease (13%).

The 39 patients with CD or NSAID enteropathy are the patients who were used in the analysis for the hypotheses. The 26 patients with CD had ulcers ranging from 1 to 42 in number, whilst the 13 patients with NSAID induced ulceration had ulcers ranging from 1 to 23 in number. Compared to patients with Crohn's, those with NSAID injury had a smaller number of ulcers (median 5.0 vs 10.5, p=0.188) (Figure 3). 8 out of 13 (62%) patients who were diagnosed with NSAID induced small bowel injury had been referred for Iron Deficiency Anaemia, compared to 7 out of 26 (27%) in the Crohn's group.

The patients were further subcategorised into groups according to whether they had ≤7 ulcers or ≥8 ulcers and whether they had presence or absence of inflammation (Table 1). In the Crohn's group, 58% of the patients had ≥8 ulcers on CE. In contrast, in the NSAID group, 31% had ≥8 ulcers. 85% of the Crohn's group had the presence of inflammation on CE, whereas 77% of patients in the NSAID group had the presence of inflammation on CE. The Fisher Exact test yielded p values of 0.176 for the number of ulcers and 0.666 for the presence or absence of inflammation (Figures 4 and 5)

Discussion

Summary Of Results

The difference between the median number of ulcers in the two groups being studied was considered to be insignificant, though the medians did lie on either side of the predicted boundary of 7/8 ulcers (Figure 3). The Fisher Exact Test regarding the hypothesis that those patients who had NSAID induced ulceration would have less than 8 ulcers, and those patients who had Crohn's induced ulceration would have more than or equal to 8 ulcers yielded a p value of 0.176. This was not significant enough to reject the null hypothesis on the basis of this study. Likewise, the comparison between the two groups regarding presence or absence of inflammation yielded a p value of 0.666 that was not significant enough to reject that null hypothesis.

Analysis and explanation of results

Based upon these results, the parameter of number of ulcers within the Lewis score is unable to distinguish between CD and NSAID induced ulceration. Whilst there was a difference in the median number of ulcers (and indeed the median number of ulcers for those patients diagnosed with NSAID enteropathy was under 8 ulcers and the mean number of ulcers for those patients diagnosed with CD was over 8 ulcers), the Fisher Exact Test did not show that the cut off was an appropriate one to diagnose one or the other. Six patients in the CD cohort had been referred for assessment of established CD (Table 2). It is therefore conceivable that these patients had less ulcers than they would have if they were not undergoing treatment; in order to test this, another Fisher exact test was performed after removing these patients from the sample. The p value for this test (Figure 6) was 0.284, thus the result is still statistically insignificant.

The result of this study does correlate with other literature regarding using CE to diagnose small bowel pathology, in that several articles cite scoring indices as only being markers of whether the small bowel mucosa is in a disease state, rather than being diagnostic tools [17]. Indeed, Gralnek et al. even stated this in the development of his CE scoring index [1]. In addition to this, Bjarnason et al. carried out a study in which patients with CD and NSAID induced small bowel damage were assessed via CE by four blinded readers. Whilst this used certain descriptors such as reddened folds, denuded areas etc. rather than a scoring system or a system of purely counting ulceration, it was still trying to distinguish between the two diseases. It was found that the median number of mucosal breaks differed between the two diseases (as shown in this study), however the general range of pathology was the same, and patients were often misdiagnosed by the blinded readers as either having CD when they had NSAID induced small bowel injury or vice versa, highlighting the similarities between the two disease processes [19]. Voderholzer et al. performed another similar study which also found that there was a high rate of misdiagnosis between the two diseases, going hand in hand with the fact that the two diseases share characteristics [22].

Since up to 14% of asymptomatic individuals can have mucosal breaks and other minor lesions of the small bowel detected by CE [29], and lesions such as this can be shared by numerous other causes such as coeliac disease and autoimmune pathologies as well as CD and NSAID enteropathy [4], it is not surprising that it is hard to distinguish between diseases merely on the basis of a cut off for number of ulcers.

Studies such as those by Goldstein et al. [18, look up 7] have compared NSAID damage using a method of counting mucosal breaks per tertile to evaluate the damage, however this was in patients who were given one of two drugs in a controlled environment. Likewise, Graham et al. [18, look up 4] assessed small bowel injury in known chronic NSAID users, scoring lesions descriptively as "normal, red spots, small erosions, large erosions or ulcers". Once again however, this was assessing injury in a setting where the entire cohort was taking a drug, and the mere extent of injury was being quantified.

Counting the number of ulcers may therefore not be the best way to distinguish NSAID use from CD, however some hope may lie in the scoring system as a whole. Endo et al. recently analysed the ability of the currently used CE scoring index in quantifying low dose aspirin associated small bowel injury using a CE scoring index [18]. Although they were once again assessing mucosal damage in relation to controlled groups of patients who were taking a drug and not attempting to distinguish between two disease processes which manifest with similar findings, Endo et al. did find a benefit in using the CE scoring system over merely using descriptive findings such as mucosal breaks, red spots etc. with regards to comparing mucosal inflammation. They also discussed the possibility of further using the CE scoring index by splitting up the score by tertiles and found that the score in particular tertiles corroborated with previous studies which measured the distribution of lesions within the small bowel [18].

Therefore, whilst using a parameter such as threshold number of ulcers on its own may not be able to distinguish between disease states, using a scoring system such as the Lewis score to initially grade the severity of the small bowel inflammation, in addition to applying the current knowledge regarding distribution of lesions with different disease states/types of small bowel injury (for example Lengeling et al. reported that 83% of patients with ulcerative ileitis during routine ileoscopy were taking NSAIDs [36], and Matsumoto et al. reported differences in intestinal areas involved when comparing NSAID induced injury with nonspecific ulceration [10]) may be able to help to distinguish between diseases such as NSAID induced ulcers and CD. As a score <135 is considered to be clinically insignificant under the Lewis score, mucosal damage such as that caused by NSAID enteropathy may be classed as insignificant (as shown in this study, even those with minimal ulceration can have NSAID enteropathy), thus the classification of severity would need to be adapted, or used as a continuous range.

Another possible way to improve diagnosis may in fact come from the finding in studies that COX 2 inhibitors, when taken long term, had a comparable prevalence of small bowel damage compared to conventional NSAIDs. Earlier papers have suggested that COX-1 inhibition was mainly to blame for small bowel damage [26,29], however the finding that COX-2 inhibitors also cause small bowel damage forced a rethink of the mechanisms by which NSAIDs cause small bowel damage, and concurrently a hypothesis was formulated which implicated not only indirect COX mediated mechanisms, but also direct mechanisms by which small bowel mucosal damage is caused. The current theory is that small intestinal damage can be caused by a combination of COX-1 inhibition (affecting the consitutively active enzyme, leading to alterations in for example mucosal blood flow), COX-2 inhibition via an "unknown immunological mechanism" and the fact that NSAIDs themselves, with their acidic properties, are able to cause direct topical damage [12]. Bjarnason et al. proposed that this was due to the fact that NSAIDs solubulise lipids of phosopholipids on the mucosal surface, causing damage to epithelial mitochondria. The subsequent formation of free radicals interrupts intercellular junctions and increases intestinal epithelial permeability. Agents such as bile acids and bacteria with their degredation products are then able to directly damage the small bowel mucosa [12, look up 23], manifesting as small bowel erosions and ulcers [9, look up 12, 21].

This theory, especially the theory regarding increased intestinal permeability, leads onto the fact that current documentation and experiments regarding NSAID enteropathy and its diagnosis are largely based upon the measurement of small intestinal permeability and the assay of surrogate markers of inflammation such as fecal calprotectin [35, 7, look up 40]. The assay of surrogate markers of inflammation is not disease specific (only specific for inflammation) and raised levels are consistently found in inflammatory bowel disease and NSAID enteropathy. Measures of intestinal permeability however, are slightly more promising, and although this measure can also be abnormal in other conditions (LOOK UP 12, 17), they are able to measure the effect of NSAIDs on the small intestine [12]. It is possible that these laboratory based tests (especially permeability tests), when used in conjunction with CE findings using a score of mucosal damage would be able to effectively diagnose NSAID enteropathy. A major drawback however, is their availability in hospitals. Secondly, there may not be a correlation between fecal calprotectin and the number of ulcers in the small bowel; it is however possible that the test may show correlation with the scoring system as a whole, which gives a global picture of the small bowel injury [26].

Another way in which NSAID enteropathy may be distinguished from CD in conjunction with CE findings is via histological examination of mucosal biopsies. This can provide clues to drug induced aetiology [9, look up 22,23], although once again unless histology is at the extreme end of NSAID induced damage, exhibiting features such as......., it can be difficult to ascertain the cause, especially in patients who are on many medications.

It was thought that theoretically this should not occur, however there have been many speculated reasons as to why this does occur, such as the fact that COX-2 selective agents do still have a degree of COX-1 inhibitory activity, and the fact that although COX-2 is considered to be the inducible form, it is still found constitutively in some tissues [7]. It is also worth noting that it has been speculated that whilst enteric coated aspirin was designed to spare gastric injury, it may injure the small bowel through a topical irritant effect [12].

There are various limitations to this study. Unfortunately the power of this study was only 43%; it is therefore possible that a type II error was made. If this study was to include a greater sample size, it may therefore yield a result that is more accurate, indeed it may actually yield a result that is significant. Due to the retrospective nature of the study, even within the sample that was available, a total of 17 patients (21% of the entire sample) were either lost to follow up or had an inconclusive diagnosis. If the study had been prospective in nature, it may have been possible to ensure that the patients were correctly followed up and that a diagnosis was sure to be made using for example surrogate markers of inflammation to diagnose NSAID enteropathy, or making a histological diagnosis for CD. This is often not deemed necessary by the clinician, and as such it is a major drawback as many of the patients will only have the CE performed, and not have any other follow up tests. In many of these patients the indication for the procedure was anaemia/occult GI bleeding. In those patients for whom the indication for the procedure resolves after CE, often no further investigations are performed.

Another limitation of the study is that it is looking at the specific finding of ulceration. If taking the Lewis score as an example, ulceration is only one part of the global assessment of the small bowel. It may be the case therefore, as mentioned in this discussion, that the Lewis score in its entirety is able to distinguish between NSAID induced ulceration and Crohn's induced ulceration. A study which used the Lewis score, combined with patients who were suspected to have NSAID enteropathy/CD (clinical suspicion based upon a patient history, examination and laboratory tests) may be able to distinguish between the two much better. Unfortunately designing a study in this manner may be difficult; CE is mainly used for obscure GI bleeding, which by its nature, the cause for is unknown. In such cases, tests for specific pathology would likely not be done as there are many differential diagnoses; to do so would be a significant waste of resources, and as such, the clinical significance of designing a study in such a fashion may be very small.

NSAID induced small bowel injury is a surprisingly common disease process. Although more is now known about its progression and the types of injury it causes, it is still hard to definitively diagnose. Further work is needed in order to determine diagnostic criteria and algorithms for diagnosis. This will not only have an impact on the diagnosis of NSAID induced small bowel damage, but may also affect the diagnosis of diseases such as Crohn's. It is also important to remember that CE is still a relatively new imaging technology, and one that has enabled us to image an area that was very inaccessible. With such an advance, it might be possible that many of the lesions which are being found are actually variants of normal. Currently, if all small bowel ulcers are classed as abnormal, CE has a low positive predictive value for CD (31%). If suitable diagnostic criteria can be established, this may improve. Given that the diagnosis of Crohn's disease can not only lead to a patient being prescribed (sometimes unnecessary) toxic medications, and can lead to the patient being psychologically affected by their diagnosis, it is important to get it right [27].

Figures

Figure 1 - The modified Lewis Score

Parameters

Number

Longitudinal Extent

Descriptors

First Tertile

Villous appearance

Normal - 0

Short segment - 8

Single - 1

Oedematous - 1

Long segment - 12

Patchy - 14

Whole tertile - 20

Diffuse - 17

Ulcer

None - 0

Short segment - 5

<1/4 - 9

Single - 3

Long segment - 10

1/4-1/2 - 12

Few - 5

Whole tertile - 15

>1/2 - 18

Multiple - 10

Second Tertile

Villous appearance

Normal - 0

Short segment - 8

Single - 1

Oedematous - 1

Long segment - 12

Patchy - 14

Whole tertile - 20

Diffuse - 17

Ulcer

None - 0

Short segment - 5

<1/4 - 9

Single - 3

Long segment - 10

1/4-1/2 - 12

Few - 5

Whole tertile - 15

>1/2 - 18

Multiple - 10

Third Tertile

Villous appearance

Normal - 0

Short segment - 8

Single - 1

Oedematous - 1

Long segment - 12

Patchy - 14

Whole tertile - 20

Diffuse - 17

Ulcer

None - 0

Short segment - 5

<1/4 - 9

Single - 3

Long segment - 10

1/4-1/2 - 12

Few - 5

Whole tertile - 15

>1/2 - 18

Multiple - 10

Stenosis - rated for whole study

Stenosis

None - 0

Ulcerated - 24

Traversed - 7

Single - 14

Non-ulcerated - 2

Not traversed - 10

Multiple - 20

box and whisker 16cm.jpg

Number Of Ulcers.jpg

inflammation.jpg

number of ulcers 2.jpg

Tables

Table 1 - Final Diagnoses And Subgroups

Final Diagnosis

Number Of Patients

Number ≤7 ulcers

Number ≥8 ulcers

Inflammation

No Inflammation

Angiodysplasia

2

2

0

2

0

Ascaris Infection

1

1

0

1

0

Coeliac Disease

3

2

1

1

2

CD

26

11

15

22

4

GvHD

1

0

1

1

0

Inconclusive diagsnosis

13

8

5

12

1

Lost to F/U

4

4

0

2

2

Lymphoma

1

1

0

0

1

Metastatic disease

1

1

0

1

0

Normal

15

14

1

10

5

NSAIDs

13

9

4

10

3

Small bowel vasculitis

1

0

1

0

1

Ulcerative Jejunitis (Coeliac)

1

0

1

0

1

Total

82

53

29

62

20

Table 2 - Reason For Referral, Final Diagnosis, Number Of Ulcers and Presence or Absence of Inflammation For Each Patient

#

Referral

Final Diagnosis

Number

Inflammation

1

N/A

Angiodysplasia

1

Yes

2

Iron Deficiency Anaemia

Angiodysplasia

2

Yes

3

Occult GI Bleeding

Ascaris Infection

3

Yes

4

Coeliac

Coeliac disease

3

No

5

N/A

Coeliac disease

4

No

6

Resistant Coeliac

Coeliac disease

11

Yes

7

?Crohn's

Crohn's

1

Yes

8

?Crohn's/NSAID Enteropathy

Crohn's

1

Yes

9

Iron Deficiency Anaemia

Crohn's

1

Yes

10

Iron Deficiency Anaemia

Crohn's

1

Yes

11

?Lymphoma

Crohn's

2

Yes

12

?Crohn's

Crohn's

3

No

13

Crohn's Assessment

Crohn's

3

No

14

?Crohn's

Crohn's

4

Yes

15

?Crohn's

Crohn's

5

Yes

16

Crohn's Assessment

Crohn's

5

Yes

17

N/A

Crohn's

6

Yes

18

?Crohn's

Crohn's

8

Yes

19

Crohn's Assessment

Crohn's

10

No

20

Crohn's Assessment

Crohn's

11

No

21

Iron Deficiency Anaemia

Crohn's

11

Yes

22

Iron Deficiency Anaemia

Crohn's

12

Yes

23

?Crohn's

Crohn's

14

Yes

24

Anaemia/Abdo Pain

Crohn's

14

Yes

25

Iron Deficiency Anaemia

Crohn's

15

Yes

26

Iron Deficiency Anaemia

Crohn's

15

Yes

27

?Crohn's

Crohn's

16

Yes

28

Iron Deficiency Anaemia

Crohn's

16

Yes

29

?Crohn's

Crohn's

17

Yes

30

Crohn's Assessment

Crohn's

18

Yes

31

Crohn's Assessment

Crohn's

25

Yes

32

?Crohn's

Crohn's

42

Yes

33

Anaemia/PR Bleeding

GvHD

43

Yes

34

Occult GI Bleeding

Inconclusive diagnosis

0

Yes

35

Iron Deficiency Anaemia

Inconclusive diagnosis

1

No

Table 2 - Reason For Referral, Final Diagnosis, Number Of Ulcers and Presence or Absence of Inflammation For Each Patient

#

Referral

Final Diagnosis

Number

Inflammation

36

Loe Fe/B12/Folate

Inconclusive diagnosis

2

Yes

37

Occult GI Bleeding

Inconclusive diagnosis

2

Yes

38

?Crohn's

Inconclusive diagnosis

3

Yes

39

Iron Deficiency Anaemia

Inconclusive diagnosis

5

Yes

40

Iron Deficiency Anaemia

Inconclusive diagnosis

7

Yes

41

Iron Deficiency Anaemia

Inconclusive diagnosis

7

Yes

42

Iron Deficiency Anaemia

Inconclusive diagnosis

9

Yes

43

Iron Deficiency Anaemia

Inconclusive diagnosis

10

Yes

44

Occult GI Bleeding

Inconclusive diagnosis

10

Yes

45

Iron Deficiency Anaemia

Inconclusive diagnosis

16

Yes

46

?SB Carcinoid

Inconclusive diagnosis

21

Yes

47

Iron Deficiency Anaemia

Lost to F/U

1

No

48

Iron Deficiency Anaemia

Lost to F/U

2

Yes

49

?SB Pathology

Lost to F/U

4

No

50

Weight Loss

Lost to F/U

5

Yes

51

Occult GI Bleeding

Lymphoma

1

No

52

Iron Deficiency Anaemia

Metastatic disease

0

Yes

53

Anaemia

Normal

0

No

54

Iron Deficiency Anaemia

Normal

0

No

55

Iron Deficiency Anaemia

Normal

0

No

56

Anaemia

Normal

1

Yes

57

Diarrhoea/Iron Deficiency Anaemia

Normal

1

No

58

Diarrhoea/Steatorrhoea

Normal

1

No

59

Iron Deficiency Anaemia

Normal

1

No

60

Iron Deficiency Anaemia

Normal

1

No

61

Iron Deficiency Anaemia

Normal

1

No

62

Microcytic Anaemia

Normal

1

Yes

63

Iron Deficiency Anaemia

Normal

2

No

64

Iron Deficiency Anaemia

Normal

2

Yes

65

Anaemia

Normal

3

No

66

Occult GI Bleeding

Normal

6

Yes

67

Iron Deficiency Anaemia

Normal

9

Yes

68

Anaemia

NSAIDs

1

Yes

69

Iron Deficiency Anaemia

NSAIDs

1

Yes

70

Iron Deficiency Anaemia

NSAIDs

1

Yes

Table 2 - Reason For Referral, Final Diagnosis, Number Of Ulcers and Presence or Absence of Inflammation For Each Patient

#

Referral

Final Diagnosis

Number

Inflammation

71

Iron Deficiency Anaemia

NSAIDs

2

No

72

Iron Deficiency Anaemia

NSAIDs

3

No

73

Occult GI Bleeding

NSAIDs

3

Yes

74

Iron Deficiency Anaemia

NSAIDs

5

No

75

Iron Deficiency Anaemia

NSAIDs

6

Yes

76

Occult GI Bleeding

NSAIDs

7

Yes

77

Intermittent GI Bleeding

NSAIDs

10

Yes

78

Abdominal Pain

NSAIDs

11

Yes

79

?Crohn's

NSAIDs

16

Yes

80

Iron Deficiency Anaemia

NSAIDs

23

Yes

81

SLE ?Small bowel vasculitis

Small bowel vasculitis

17

Yes

82

?Coeliac

Ulcerative Jejunitis (Coeliac)

10

Yes

Table 3 - Summary Of Reasons For CE Referral

Reasons For Referral

Number

Anaemia

5

Microcytic Anaemia

1

Iron Deficiency Anaemia

33

?Crohn's

11

Weight Loss

1

Occult GI Bleeding

8

?SB Vasculitis

1

Anaemia/Abdominal Pain

1

?Lymphoma

1

Coeliac

1

Diarrhoea/Iron Deficiency Anaemia

1

Abdominal Pain

1

Diarrhoea/Steatorrhoea

1

Crohn's Assessment

6

Intermittent GI Bleeding

1

?SB Pathology

1

Low B12/Fe/Folate

1

?Coeliac

1

Resistant Coeliac

1

N/A

3

?SB Carcinoid

1

?Crohn's/NSAID

1

Total

82