In vivo diagnosis of collagenous colitis by confocal endomicroscopy

Collagenous colitis is a form of microscopic colitis which has recently been recognised as an entity of its own, characterised by chronic watery diarrhoea of unknown aetiology. The diagnosis of collagenous colitis relies on histopathological examination of biopsy specimens from colorectal mucosa, which is usually of normal macroscopic appearance. The typical histological feature is diffuse thickening of the subepithelial collagen layer beneath the basement membrane and an unspecific chronic inflammatory infiltrate of the lamina propria.¹,²,³

Recently, a confocal laser endomicroscope has been developed that is integrated into the distal tip of a conventional video endoscope. This confocal laser microscope (EC‐3870CIFK; Pentax, Tokyo, Japan) was designed to enable subsurface imaging of living tissue during ongoing endoscopy and allows confocal microscopy in addition to standard video endoscopy. Images are generated by intravenously administered fluorescein sodium as the fluorescent contrast agent and an argon ion laser integrated into the system that generates an excitation wavelength of 488 nm.⁴,⁵ For the first time, we used fluorescein aided confocal laser endoscopy for the immediate diagnosis of collagenous colitis.

A 67 year old woman was admitted to the first Medical Clinic, Johannes Gutenberg University Mainz, on July 2004, with a long history of watery diarrhoea with postprandial pain and bloating. Colonoscopy was performed using the confocal laser endoscope, enabling video endoscopy and endomicroscopy. We administered intravenously 10% fluorescein 5 ml, as a prerequisite contrast stain for endomicroscopy to enable confocal endomicroscopy of the colonic mucosa. Confocal image data were then collected at a scan rate of 0.8 frames per second (1024×1024 pixels) using an optical slice thickness of 7 μm. Confocal images were generated simultaneously with endoscopic images. Total examination time was 36 minutes (including 12 minutes for endomicroscopy). Using this approach, cellular and subcellular structures of the colonic epithelium (surface epithelium and crypts), connective tissue, and vasculature could be seen in this patient with a lateral and axial resolution of less than 1 μm (see fig 1). Penetration depth of the laser was approximately 250 μm, and this depth allowed us to examine almost all parts of the colonic mucosa.

Figure 1 Collagenous colitis diagnosed in vivo by confocal laser endomicroscopy. (A) Endomicroscopy of the surface of the mucosal layer showing crypt deformation. Four crypts with different shapes were aggregated (arrow). Note that the black dots within the crypts represent mucin in goblet cells. (B) Subepithelial collagenous bands were readily visible in the upper third of the affected mucosa (imaging depth ∼150 μm). The collagenous bands surround single crypts (arrows). (C) In deeper parts of the mucosa (imaging plane depth ∼200 μm) the collagenous bands were displayed as dark bands within the lamina propria (arrows). The inhomogeneous distribution of the bands was clearly visible at high resolution (lateral resolution less than 1 μm). The scale bar at the right upper corner represents 100 μm. The blue line measures the collagenous band (31 μm). (D) Normal colonic mucosa with regular distribution of crypts (arrow) without cryptal damage or tissue changes in the lamina propria. (E) Histological specimen after haematoxylin‐eosin staining. The subepithelial bands were identified beneath the basement membrane (arrow). (F) van Gieson staining highlighted the collagenous bands. The inhomogeneous distribution corresponds well with the endomicroscopic image (see C).

Subepithelial thickened collagenous bands could be readily identified by performing endomicroscopy. All parts of the colon were examined by placing the endomicroscopic window at the distal tip of the endoscope onto the mucosal layer. Endomicroscopic images were obtained every 15 cm within the colon. A total of 322 confocal images from six different sites were graded for the presence of collagenous bands and targeted biopsies were performed.

Biopsy specimens were fixed in 4% formalin and embedded in paraffin. Sections (4 µm thick) were stained with haematoxylin‐eosin. For assessment of the collagen band, van Gieson staining was used. Typical microscopic changes were identified endomicroscopically in four of six sites (right sided colon and transverse colon) which was confirmed by ex vivo histology in all cases.

It is well known that collagenous colitis can demonstrate a discontinuous pattern. Endomicroscopy may be well suited to target and identify affected areas. However, true specificity cannot be given in our limited series.

The patient received 9 mg of oral budesonide daily for six weeks, resulting in reduced postprandial complaints and diarrhoea.

In conclusion, endomicroscopy allows localisation and measurement of the amount of collagenous bands in the mucosal layer. Thus endomicroscopy offers the possibility of targeted biopsies, which is a new approach in collagenous colitis where randomised biopsies, preferably in the right colon, are recommended. The distribution of the collagenous bands is patchy and segmental in the colon. Confocal endomicroscopy helps to differentiate between affected and normal sites. This initial experience was proven in four additional patients. In all patients, collagenous colitis was precisely predicted and the amount of collagenous bands was measured. However, this new diagnostic possibility and its sensitivity and specificity must now be evaluated in prospective studies.