Distinct colonoscopy findings of microscopic colitis: Not so microscopic after all?

Abstract

Microscopic colitis (MC) is considered an “umbrella term”, comprising two subtypes, i.e., collagenous colitis (CC) and lymphocytic colitis (LC). They are classically associated with normal or unremarkable colonoscopy. In the last few years, reports have been published revealing findings that are thought to be characteristic or pathognomonic of MC, especially CC. A systematic electronic and manual search of PubMed and EMBASE (to December 2010), for publications on distinct endoscopic findings in MC, resulted in 42 relevant reports for inclusion in this review. Eighty eight patients with collagenous colitis were presented. Only one publication describing a distinct endoscopic pattern in LC was found. Typical findings in CC are alteration of the vascular mucosal pattern, mucosal nodularity, a sequence of change from mucosal defects to mucosal cicatricial lesions, and perhaps (although of doubtful relevance) mucosal pseudomembranes. A causal connection of mucosal defects with the use of lansoprazole seems to exist. Adoption of the proposed lesion description herein is recommended in order to improve homogeneity of future reports.

INTRODUCTION

Microscopic colitis (MC), regarded as a rare entity in the early 80s (and certainly overlooked), has now emerged as an increasingly common cause of chronic, non-bloody/watery diarrhea[1].

MC is an “umbrella term”, comprising two entities/subtypes, i.e., collagenous colitis (CC) and lymphocytic colitis (LC)[2]. The two entities are characterized by a variable, yet apparently benign, clinical course of protracted, non-bloody diarrhea and classically normal or unremarkable colonic mucosa on endoscopy[3]. In 1984, Gledhill[4] established that thickening of the colonic acellular basement membrane by > 15 μm is invariably associated with diarrhea.

The histological abnormalities in MC are discontinuous, subtle and often unequally located in the colon, making it necessary to take multiple biopsies from various colonic regions for identification of the pathognomonic microscopy, i.e., thickened sub-epithelial collagen band and increased intraepithelial lymphocytes[5] (Figure 1).

Figure 1.

Van Gieson Stain, sub-epithelial collagen table.

However, there are occasions where endoscopy reveals findings that are thought to be characteristic or pathognomonic of MC, and especially CC. Although the estimated prevalence of MC is up to 10% in patients with chronic diarrhea[5], there are few reports of macroscopic findings in MC. This review attempts to describe the known characteristic endoscopy findings in MC and to categorize them in different types.

PATHOPHYSIOLOGICAL BACKGROUND

CC was first described in 1976, independently in Sweden by Lindström[6] and in Canada by Freeman[7], while LC was first described by Lazenby et al[8] in 1989. An increase in their incidence has been recently reported, but this is most likely an artifact secondary to increased awareness and prompt diagnosis[9]. In the absence of persistent endoscopic findings, diagnosis is based mainly on specific histological criteria[9].

It is not clear whether CC and LC are separate entities or part of the spectrum of a single disease[2]. With regard to pathogenesis, several hypotheses have been suggested, including inflammation secondary to medication, smoking, immune dysfunction, autoimmunity, and/or infection.

Studies of collagen typing in patients with CC have produced conflicting results. Electron microscopy findings have suggested that the collagen in CC appears similar to that found in granulation tissue, supporting the hypothesis that its presence would suggest a reparative response to injury[10]. In fact, it is plausible to assume that overproduced, multiple, and different collagen types may deposit in the sub-epithelial layer of the colon and manifest clinically as CC[11]. Günther et al[12] showed that increased connective tissue growth factor expression might be the final mediator of local fibrosis in CC.

Non-steroidal anti-inflammatory drugs (NSAIDs) have been implicated as causative factors, through their ability to inhibit prostaglandin synthesis from the colonic mucosa. More recently, several reports have been published incriminating proton pump inhibitors (PPIs), especially lansoprazole, in the induction of CC. Most of the findings to support this came from the temporal relationship of resolution of symptoms with cessation of NSAID or PPI therapy. PPI-induced conformational changes in the cytoskeleton of epithelial cells may result in alterations in the function of the tight junction, leading to increased paracellular permeability. Keszthelyi et al[13] postulated that this could allow the luminal contents to easily penetrate the lamina propria causing an immune and/or inflammatory reaction. On this basis, and in light of some recent reports[14], which incriminate lansoprazole as the main cause of linear mucosal defects in CC, it may be plausible to suggest that CC is a syndrome with various causes and perhaps graded histopathology. The purpose of this review is to present the published experience on distinct endoscopic findings in MC and suggest a unifying lexicon for the reported lesions.

SEARCH STRATEGY

We conducted a PubMed and EMBASE computer search (to December 2010) in order to identify articles on microscopic colitis and endoscopic findings. Our search strategy for PubMed was [“Colitis, Microscopic” (MeSH) or “Colitis, Collagenous” (MeSH) or “Colitis, Lymphocytic” (MeSH)] and [“Endoscopy” (MeSH) or “colonoscopy” (MeSH) or “intestinal mucosa” (MeSH)]. We confined our search to articles in humans but we did not apply any language restriction. In order to search EMBASE we used the following key words: “collagenous colitis”, “microscopic colitis” or “lymphocytic colitis”, “endoscopy” or “colonoscopy”. A further search of electronic journals was undertaken.

Duplicate articles identified in PubMed or EMBASE were manually deleted. The first selection, based on the title and/or abstract was carried out by one of the authors (AK). From the outset, we agreed not to include for further review reports or studies on endoscopic technology, e.g., confocal laser endomicroscopy, which is not yet widely available or restricted to a small number of tertiary institutions. The full paper of each potentially relevant report was then obtained. Thereafter, the two authors independently assessed publications for inclusion in the review. In addition, the reference lists of relevant reports and review papers were cross-searched, in order to identify papers that our initial computer search may have missed.

The following data were extracted from each included publication: year of publication and first author, country of origin, number of cases reported, gender and age of the cases, described endoscopic findings, histopathological diagnosis, post-endoscopy/clinical complications and any important clinical associations (Table 1).

Table 1.

Overview of reports of endoscopic findings/appearance of collagenous colitis

Year	Ref.	No. of cases, gender, age	Endoscopic findings	Lesion locationcollagen table thickness	Clinical associations	Complications
1990	Giardiello et al[15]	1, M, 60	Pseudomembranes	S colon	Watery diarrhea, received NSAIDs/antibiotics	None
				50-70 μm
1993	Richieri et al[16]	1, F, 43	Linear mucosal tears/lacerations	R colon	Watery diarrhea, abdominal pain	None
			Absent vascular mucosal pattern	30-40 μm	Successful therapy with steroids, some bloody stools
			6 mo later; many linear cicatricial lesions
1993	Smiley et al[17]	1, F, 53	Carpet-like patch with nodularity (5 cm)	R colon	Watery diarrhea	None
			Similar to sessile villous adenoma	20-40 μm	Therapy with bulking agents
1995	Katanuma et al[18]	1, F, 72	Diminished vascular pattern	Pancolonic	RA on sulindac, diarrhea and wt loss	None
			Edematous/red mucosa	n/s	Treated by discontinuation of NSAID and salazosulphopyridine
1997	Katsinelos et al[19]	1, M, 65	Multiple red mucosal spots	R colon	Watery diarrhea,	None
			Diminished vascular pattern	n/s	Successful therapy with steroids
1997	Yabe et al[20]	1, F, 47	Multiple red mucosal spots	Pancolonic	6 F/U colonoscopies showed no improvement	None
			Diminished vascular pattern	n/s
1998	Sato et al[21]	1, F, 78	Crowded/tortuous vascular pattern	R + T colon	Watery diarrhea	None
			I/C spray: coarse/nodular surface	n/s
1999	Bermejo et al[22]	1, F, n/s	Pseudomembranes and aphthae	n/s	Watery diarrhea, received NSAIDs/antibiotics	None
				n/s
2001	Freeman et al[23]	1, F, 37	Deep, elliptical mucosal defect/ulcer	S colon	Watery diarrhea, acute abdomen	Perforation
				n/s	Diagnostic laparotomy + IV antibiotics
2001	Yagi et al[24]	1, F, 77	Mucous-covered lesions in R colon	R + T colon	Watery diarrhea, 4 colonoscopy linear lesions in rectum, ASA-associated	None
			Ulcer in descending colon	30-60 μm
			2nd look: rectal pseudomembranes
2002	Cruz-Correa et al[25]	2, F, (73/61)	Deep lacerations/tears	R and T colon	All had hypothyroidism	None
		1, M, 62	0.5-4 cm long (to muscularis mucosa)	n/s	Therapy with tetracycline/5-ASA
2003	Kakar et al[26]	8, F, (a. r: 37-91)	Linear ulcers or lacerations (5)	R colon (5)	Aspirin and NSAID-associated CC	None
		1, M, 27	Diminished vascular pattern (2)	S colon (3)	Treated with discontinuation, bismuth
			Aphthae (2), pseudomembranes (1)	n/s	Mesalamine or azathioprine/6-MP
2003	Sato et al[27]	1, F, 78	1st colonoscopy: 3 mm nodule	Pancolonic	Watery diarrhea, wt loss	None
			2nd look: crowded/tortuous vascular pattern	R: 40-70 μm	ASA-associated
			I/C spray: coarse and nodular, uneven surface	L: 20 μm
2003	Byrne et al[28]	1, F, 27	Erythematous mucosa	S colon	Watery diarrhea,	None
			Multiple pseudomembranes	n/s	common variable immunodeficiency (CVID)
2003	Yuan et al[29]	6, F, (a. r: 54-81)	Linear ulcers (1),	T colon	Pseudomembranes in CC, only endoscopic cases included	None
			R colon ulcers (2), inflamed rectum (1)	n/s
2004	Buchman et al[30]	1, F, 58	Hemorrhagic mucosal spots and erythema, granularity/pseudomembranes	R colon	Prednisolone, antibiotics, TPN, PPI, hypoalbuminemia	None
		1, F, 46
2004	Sherman et al[31]	3, F, (a. r: 66-73)	Mucosal tears and fractures	R + T colon	Watery diarrhea, wt loss, hypoalbuminemia	Perforation in 3/4 cases
		1, M, 69	Granularity of mucosa at places	40-50 μm
2006	Wickbom et al[32]	3, F, (a. r: 73-86)	Mucosal tears and fractures (4-5 cm long)	R + T colon	All on aspirin	None
			Mucosal scars on repeat colonoscopy	14-40 μm	ACE/lansoprazole-induced (1 case)
2006	Koulaouzidis et al[33]	1, F, 83	Mucosal tears	Cecum	Iron deficiency anemia	None
				n/s
2007	Poupardin-Moulin et al[34]	1, F, 80	Longitudinal mucosal fractures	R + T colon	No significant clinical associations, diagnosis missed	None
				n/s
2007	Smith et al[35]	1, F, 43	Long, linear mucosal fractures	R colon	Treated with sulfasalazine	Perforation
				n/s		Hemicolectomy
2007	McDonnell et al[36]	3, n/s, n/s	Bright linear marks/parallel corkscrew lesions: “cat scratch” colon	R colon	n/s	none
				n/s
2008	Allende et al[37]	9, F, (a. r: 44-80)	Mucosal fractures to muscularis propria (7)	R colon (6)	2/12 underwent barium enema	Perforation all cases
		1, M, 71	Ragged mucosal defect (1)	T colon (3)		2 during colonoscopy
			Wall induration (1), constriction (1)	L colon (1)
2008	Umeno et al[38]	7, n/s, (a. r: 37-92)	Longitudinal mucosal defects (ulcers/tears)	L colon	Only in the lansoprazole treated group	None
			Longitudinal scar in one case	12.5-50 μm
2008	Hashimoto et al[39]	1, F, 66	Whirling/circling mucosal vessel network	Pancolonic	SLE, treated with mesalamine	None
			Linear (20 cm) ulcer/scar in the descending	n/s	2nd look: normal vessels, smaller scar
2009	Watanabe et al[40]	1, F, 68	Multiple, longitudinal thin ulcers	L colon	Lansoprazole, discontinued and healed	None
				30 μm
2009	Yusuke et al[41]	1, F, 78	Ragged and linear, long mucosal tear	S colon	Abrupt abdominal pain, PR blood	None
			Hypertrophic scar	n/s	Lansoprazole, discontinued
2009	Cuoco et al[42]	1, F, 68	Deep linear ulcer-type defects	R + L colon	Watery diarrhea, abdominal pain	None
			7 cm long in ascending	n/s	4 L PEG for cleansing
2009	Dunzendorfer et al[43]	1, F, 60	3 cm hypertrophic mucosal scar	S colon	Long history of constipation	None
				n/s	Wt loss, combination therapy
2009	Chiba et al[44]	1, F, 70	Distinct diffuse mucosal cloudiness	Pancolonic	On lansoprazole and loxoprofen,	Reoccurred on a further Lansoprazole course
			Indistinct vascular pattern (UC-like pattern)	n/s	treated with sulfasalazine
2009	Sekioka et al[45]	1, F, 82	2 longitudinal mucosal fractures	T colon	Lansoprazole-associated (6 mo)	Peritonitis, pre-endoscopy
			2nd look: A ridge-type cicatricial lesion	n/s	Treated by discontinuation
2010	Couto et al[46]	1, F, 48	Hemorrhagic mucosal tears	T + L colon	OA on nimesulide and lansoprazole, abdominal pain, wt loss (10%)	None
			Longitudinal white ridges/lines	n/s		Colonoscopy halted at T colon
2010	Sawada et al[47]	1, M, 77	Disappearance of vascular network,	L colon	Lansoprazole-associated (6 years)	None
			Red (numerous) mucosal spots	25 μm	Wt loss, treated by discontinuation collagen table reduced on 2nd look
2010	Koulaouzidis et al[48]	1, M, 83	Fine cicatricial line	L colon	n/s	None
		1, F, 45	Hypertrophic mucosal scar	n/s
2010	van Velden et al[49]	1, F, 63	Linear tears	R + S colon	Instrumentation-induced and insufflation-induced mucosal tears	Perforation
			Diminished vascular pattern and edema	20 μm		Treated conservatively
			2nd look colonoscopy: multiple linear scars
2010	Nomura et al[50]	1, F, 67	Linear mucosal defect x 2,	L colon	Lansoprazole-associated	None
			Linear scar in sigmoid, I/C spray	n/s	Improved on discontinuation	Painful left abdomen
2010	Miyagawa et al[51]	1, M, 81	Longitudinal mucosal defect	L colon	Lansoprazole and hemodialysis	None
				n/s
2010	Milestone et al[52]	3, F; 1, M	Long (5-20 cm) linear ulcers,	S colon	Treated with budesonide and/or bismuth subsalicylate	None
		(a. r: 57-75)	non-hemorrhagic with evidence of healing	n/s
2010	Kawamura et al[53]	3, n/s, n/s	Longitudinal mucosal ulcers	L + S colon	Lansoprazole induced	None
				n/s
2010	Fasoulas et al[54]	1, F , 68	“Cat scratch” colon	R colon	n/s	None
				n/s
2010	Cimmino et al[55]	4, F, (a. r: 24-77)	Mosaic pattern (honeycomb image),	Rectum+ S colon	Case control study	None
			I/C spray: for delineation of pattern	n/s	Mosaic pattern had high LR+/spec

n/s: Not stated; M: Male; F: Female; a. r: Age range; I/C spray: Indigo carmine spray; R colon: Right colon; T colon: Transverse colon; L colon: Left colon; S colon: Sigmoid colon; LR: Likelihood ratio; spec: Specificity; UC: Ulcerative colitis; TPN: Total parenteral nutrition; OA: Osteoarthritis; ASA: Acetyl salicylic acid; wt: Weight; PPI: Proton pump inhibitor; 6-MP: 6-mercaptopurine; PEG: Polyethylene glycol; ACE: angiotensin converting enzyme; NSAIDS: Nonsteroidal antiinflammatory drugs; CC: Collagenous colitis; SLE: Systemic lupus erythematosus; RA: Rheumatoid arthritis.

SEARCH FINDINGS

Our initial computational search returned 89 articles in PubMed and 499 in EMBASE. Nineteen and 50 articles, from PubMed and EMBASE respectively, were included for further review. After obtaining the full papers, 35 papers were selected. Another seven publications were identified from references lists and included in the final analysis.

The terms mucosal break, defect, tear, fracture or laceration were used indiscriminately. For the purpose of this review and in order to standardize the terminology, we agreed to use the term “mucosal defect” as a collective one, under which there are two subtypes of lesions: (1) mucosal lacerations/tears which are the longitudinal (superficial or deep) and mainly fresh/hemorrhagic in appearance mucosal breaks (Figure 2); and (2) mucosal fractures describing the deeper (with occasional exposure of the muscularis mucosa) and white-based or more chronic looking mucosal defects (Figure 3).

Figure 2.

Colon lacerations/mucosal breaks in collagenous colitis. A, B and C: Colon lacerations/mucosal breaks; D: Cat-scratch colon.

Figure 3.

Colon mucosal fractures in collagenous colitis.

Although to an extent arbitrary, we believe that this terminology will aid the introduction of a universal lexicon for future reports of similar lesions. It is obvious that in accordance with the above, the “cat scratch colon” belongs to the first category, i.e., mucosal lacerations or tears.

Eighty eight cases [65 females, 10 males, 13 not stated (n/s); median age: 67 years] were reported in 41 publications. Of these, 14 publications were from Japan[18,20,24,27,37-41,44,45,47,50,51], 12 from the United States[15,17,25,26,28-31,35-37,43], three from the United Kingdom[33,48,52], two each from France[16,34], Sweden[21,32], and Greece[19,54], and one each from Argentina[55], Canada[23], Italy[42], the Netherlands[49], Portugal[46], and Spain[22]. Where reported, the submucosal collagen table thickness ranged from 14-70 μm. The only publication reporting endoscopic findings in LC described the presence of a subtle mucosal change in an 85-year-old female[56].

Gardiello et al[15] were the first to report distinct endoscopic findings in CC (i.e., pseudomembranes), but in fact it was Richieri et al[16] who first described the presence of multiple linear mucosal lacerations with sharp edges in the right colon of a 43-year-old female, with sub-epithelial collagen table thickness of 30-40 μm. Eventually, on repeat colonoscopy 6 mo later the lesions had healed, resulting in fine cicatricial lines on an otherwise unremarkable colonic mucosa. Therefore, Richieri et al[16] had effectively pointed to a pattern seen in some of the reports that followed, i.e., the continuum of laceration to cicatricial healing of the mucosa.

Since this report, 53 cases (34 females/6 males/13 n/s; median age: 69 years) of linear, long or shorter and finer (cat-scratch type) mucosal tears, fractures and ulcers have been reported[25,26,29,31-42,45,46,49-54]. Sixteen patients with mucosal defects were on lansoprazole, and in the majority, discontinuation of the medication resulted in symptomatic, endoscopic and histopathological improvement.

On the other hand, only 11 (10 females/1 male) cases of mucosal cicatricial lesions have been reported to date, identified either during the index colonoscopy that revealed the mucosal defects, or at follow-up colonoscopic examinations[16,32,38,41,43,45,46,48-50]. The lesions ranged from hypertrophic (celoid-type mucosal scars)[32,38,41,43,45,46,48-50] to fine, cicatricial lines[16,48] (Figure 4).

Figure 4.

Cicatricial mucosal lesions in collagenous colitis (A and B).

We did not manage to establish an association of any of these lesions either with the collagen table thickness or with symptom severity in the review cohort.

The right colon (for the purpose of this review defined as the area from the cecum to the hepatic flexure), irrespective of the type of findings, was affected in 32 cases, the transverse colon in 16 and the left (descending, sigmoid and rectum) colon in 32. Five reports presented cases with pancolonic mucosal involvement[18,20,27,39,44].

Although the sign of a mosaic pattern or mucosa nodularity (“honeycomb mucosa”) was noted first by Smiley et al[17] in 1993 in the ascending colon of a 53-year-old woman, a retrospective case-control study was only published in 2010[55]. In the appropriate clinical context of watery diarrhea, the “honeycomb pattern”had an odds ratio of 19.4 with a specificity of > 99% for diagnosis of CC. The authors though pointed out that, due to both the retrospective nature of the study and the high possibility of under-reporting, this may be an overestimation.

Dye spray (indigocarmine), for improved delineation of the identified lesions, was utilized in four reports[21,27,50,55], and seems helpful in the context of subtle mucosal changes and/or disturbed vascular architecture. However, this should be balanced against the greater resource implications and procedure time.

With regard to complications, there were 17 recorded perforations/peritonitis in the review cohort[23,31,35,37,45,49]. As expected, these were all associated with cases where mucosal defects (tears or fractures) were evident on colonoscopy[52,57].

WHAT IS CURRENTLY KNOWN

We found four broad categories of distinct endoscopic findings in CC: (1) pseudomembranes[15,22,24,26,28,30]; (2) mucosal vascular pattern alteration which includes an indistinct appearance of the blood vessels and a variable degree of pruning of the mucosal vasculature, or a crowded, dilated and tortuous capillary network[16,18-21,26,27,39,44,47]; (3) mucosal abnormalities such as red spots and some mucosal nodularity or textural alteration, evident with or without chromoendoscopy[17,19-21,27,30,31,55]; and (4) a continuum of mucosal breaks/defects, i.e., mucosal lacerations/tears, including the so-called “cat scratch colon” pattern, or fractures usually along the long axis of the colonic wall[16,25,26,29,31-42,45,46,49-54]to the fine linear cicatricial lines or thick scar-like ridges of the mucosal surface (effects of the mucosal healing process of mucosal defects)[16,32,38,41,43,45,46,48-50]. There was only one publication describing a characteristic endoscopic pattern in LC[56].

Hemorrhagic mucosal breaks have an appearance that could be liberally described as “colon craquelé”[16]. The term mucosal fracture was introduced by Sherman et al[31] in 2004 and it is admittedly a successful descriptive one. Thickened and abnormal sub-epithelial collagen table leads, at some areas, to loss of attachment with the epithelial component, and this in turn causes stretching of the mucosa over the deeper wall layers, and eventually tearing of the detached mucosal surface (in a “zip” fashion, hence the longitudinal lesions). The sharply demarcated margin of these mucosal defects, as if the mucosa has been slashed with a sharp knife, helps to differentiate them from ischemic colitis[50].

Mucosal defects are more likely to be found in the right colon as a result of a colonic insult, i.e., instrumentation or air insufflation due to the abundant presence of a thicker and denser (hence dysfunctional) collagen type III table, in association with increased colon diameter on that side[25,37,58]. The right colon thinner wall and its expansion to a greater diameter during fecal storage and transit, produce greater relative wall tension (Laplace’s law, i.e., tension on the wall of a cylinder is proportional to the radius). Therefore, a competent ileocecal valve and a deformed sigmoid are sufficient to cause colonic air entrapment in a closed space[59], and eventually “cracking”of the brittle colonic mucosa[11,25,31,32,48,58]. Although the colon can not be seen as a simple cylinder[37], we suggest that these breaks can occur spontaneously, and postulate that increased intra-colonic pressure during peristalsis and defecation leads to mucosal stretching and defects that will heal with time leaving behind various types of cicatricial lesions[33,48].

McDonnell et al[36] coined the term “cat scratch colon”to describe the red linear marks in the cecum or ascending colon seen in 21 of 8277 patients undergoing colonoscopy. They reported a 14% prevalence of CC of in their cohort. They also postulated that these marks were due to barotrauma from insufflation[36,59-61]. However, it is unclear whether biopsies were taken in all patients undergoing the test for diarrhea, other than in those that had the “cat scratch” appearance. Furthermore, endoscopic findings are non-specific for CC and have been described in the normal colon (attributed to barotrauma from excessive insufflation during colonoscopy), in diversion colitis, and even in chronic cholestasis[54,61,62].

The true prevalence of mucosal tears is unknown due to the rarity of reported cases, but it is estimated to be around 1%. Under the assumption that not all of the relevant cases have been reported, the true prevalence may be much higher. However, based on the type of publications included in this review, i.e., case reports or series, it is not possible to estimate prevalence. In addition, practices vary worldwide and up until recently flexible sigmoidoscopy was considered sufficient to diagnose MC (it is believed that left-sided biopsies probably miss less that 5% of MC cases, due to its patchy nature), and as lesion awareness rises, the incidence of macroscopic findings will increase[63]. On the other hand, the increased frequency of reports published during the last decade show that there is an increased awareness of the distinct endoscopic appearances in MC, and perhaps endoscopist enthusiasm may result in over-diagnosis (as mucosal tears/scratches have been described in the normal colon, diversion colitis and in lansoprazole colitis[36,64,65]) of an entity whose main hallmark remains histological confirmation.

It is also now known that mucosal defects in CC represent a marker of increased risk of colonic perforation[52,54]. A recent review found 21 cases of perforation in CC. The majority of these were either colonoscopy-associated (15 cases) or barium enema-associated (four cases), while the rest seem to have occurred spontaneously[57].

There are several reports of remission, including disappearance of the collagen layer on follow-up. This would indicate that an environmental factor such as medication may be responsible in susceptible individuals. NSAIDs or PPIs have been implicated. It has also been suggested that collagen plate thickness is greater with lansoprazole[38]. The pathophysiologic mechanism by which lansoprazole induces microscopic colitis and mucosal defects is not well understood. Although a clear temporal correlation exists, it should be remembered that, due to the fluctuating nature of CC[66], it might simply represent a coincidence, as PPIs are one of the most commonly prescribed drug categories worldwide.

It has been postulated that this may be due to higher concentrations of drugs such as NSAIDs in the right colon[26]. However, it is possible that more right sided biopsies are taken because of endoscopic abnormalities, more likely to be observed in the right colon, as mentioned above. More case control studies and multivariate analysis may provide the answer[14].

In conclusion, the endoscopic appearances of CC are becoming more familiar amongst the endoscopic community. We recommend adoption of the proposed lesion description herein in order to improve homogeneity of future reports.

ACKNOWLEDGMENTS

We feel indebted to Ms Sarah Douglas, Senior Scientist and Dr. JN Plevris, Consultant and Reader in Gastroenterology, The Royal Infirmary of Edinburgh for their invaluable suggestions in proofreading. We are grateful to Dr. T Sekioka for allowing us to reproduce image (Figure 3) from Ref. 45. Figures 1, 2A and B from Koulaouzidis et al[33], Endoscopy 2006, with permission. Figure 4A and B from Koulaouzidis[48], GIE 2010, with permission.