Article Text

Download PDFPDF

Candidal infection is uncommon in acute oesophagitis: evidence from a non-selected DGH population
  1. L J Baldwin1,
  2. D N Poller1
  1. 1Department of Pathology, Queen Alexandra Hospital, Cosham, Portsmouth PO6 3LY, UK

    Statistics from

    Mucosal candida infection of the lower oesophagus is unusual except in certain groups of patients who are either immunosuppressed or who have other recognised causes of candidal infection.1 Invasive oesophageal candidiasis occurs most often in the immunosuppressed, occurring in 10–20% of patients with myeloproliferative disorders or leukaemia and in up to 74% of patients with AIDS, with an increased frequency of infection in patients with endocrine disorders such as hypoparathyroidism.1 Previous studies have shown that the incidence of candidal infection in 22 000 consecutive hospital admissions was 0.1% (27 cases), whereas it was found in less than 5% of a general population presenting with gastrointestinal complaints.2 It is important to exclude invasive candidiasis as this is a major risk factor in the development of candidal septicaemia, which can result as a direct effect of visceral wall invasion. Complications of oesophageal candidiasis also include oesophageal stenosis and perforation, which may occur in the acute phase of infection, and which can be life threatening. Less commonly, pseudodiverticulosis may also result.1

    Candidal oesophagitis is caused most commonly by C albicans, C tropicalis, and C krusei.3 The diagnosis of invasive candidiasis requires a combination of clinical suspicion and laboratory investigation. Difficulties in diagnosis arise where the clinical presentation is non-specific or serology produces false positive results because of vulvovaginal candidiasis. Oesophageal brushings are useful for identifying mucosal surface candidal colonisation but are poor at detecting fungal invasion.4,5

    Here we present the results of a retrospective review of 60 cases of acute oesophagitis, with reassessment for candida by D-PAS staining.

    Sixty consecutive oesophageal biopsies coded on the laboratory computer SNOMED database as acute inflammation with or without ulceration were examined from the period 1997–1998. In our hospital few specimens are sent for mycological culture, and brush cytology is not performed for the diagnosis of candida.

    Candidal staining—All cases had remaining tissue available. Additional sections were cut if no spare unstained sections were available and then the sections were stained with PAS-D. Sections were pretreated with fresh diastase solution (0.5 cm3 diastase in 50 ml deionised water) followed by treatment with periodic acid Schiff reagent and Carrazzi's haematoxylin, with tapwater washes in between each stage. The original H&E sections were re-examined to identify the presence of squamous mucosa or gastric mucosa and the nature of inflammation present (acute or chronic). The sections were also re-examined for evidence of ulceration. The minimum criterion required for oesophageal ulceration was the presence of squamous mucosa with epithelium and inflammatory granulation tissue representing the ulcer bed, and/or squamous mucosa with fibrinopurulent and necrotic exudate representing the surface of an ulcer crater. The PAS-D stained sections were examined by both of us for evidence of candidal hyphae and spores, looking for mycelial forms of the fungi and for pseudohyphae, which were considered the minimum criteria for diagnosis of invasive candidiasis.3

    Results—Nine of the 60 patients (15%) showed acute ulceration of the oesophagus; the remainder showed acute inflammation with Barrett's change in many cases but no evidence of acute ulceration. Of the 60 oesophageal biopsies examined, no candidal hyphae or spores were identified, either on initial haematoxylin and eosin examination or during this study, after restaining of all the cases with D-PAS. Two of the biopsies did contain some D-PAS positive material, but this material did not have the appearances of fungal spores and we both considered it to be artefact. No mycelial forms or hyphae were present in the sections of any of the cases examined.

    Comment—Assessment of endoscopic biopsies for fungi using only haematoxylin and eosin staining may be difficult. The differential diagnosis of eosinophilic bodies or hyphae-like tissue in areas of acute oesophageal ulceration may include food debris, degenerate material from the base of an ulcer, herpetic ulceration, reflux oesophagitis, glycogenic acanthosis, other artefacts, and carcinoma. It is widely recognised that the use of D-PAS or Grocott stains may assist in the recognition of fungal hyphae and spores. More elaborate methods have also been used for detection of candida, such as immunostaining for C albicans6 and polymerase chain reaction.7

    The aim of our study was to examine the method of routine reporting currently used in our laboratory when dealing with inflamed oesophageal biopsies—that is, without routine additional fungal stain—and to examine the efficacy of detecting candida on routine haematoxylin and eosin stained slides with an additional PAS or PAS-D stain. We showed that the addition of a routine fungal stain for endoscopic biopsies is unlikely to be of benefit in the detection of candida if the patient is not in an at risk group. If the patient is at risk of candida then combined endoscopic biopsy and brush cytology is likely to be much more sensitive than biopsy alone for the detection of the organism.5

    Causes of candidal infection of the lower oesophagus include use of antibiotics, which may allow overgrowth or colonisation by Candida species, inhaled corticosteroids, reduced gastric acid output from H2 receptor blockers, proton pump inhibitors, or prior vagotomy. Alcoholism, diabetes, malnutrition, advanced age, and abnormal oesophageal motility associated with conditions such as scleroderma and achalasia have also been associated with candida oesophagitis. Strictures, obstructing tumours, or diverticula which cause stasis are also associated with fungal oesophagitis, as is also chronic mucocutaneous candidiasis syndrome. Nine patients (15%) in our study had D-PAS stains performed at the time of biopsy by the reporting pathologist and the majority of these additional stains were requested when there was evidence of acute ulceration, not just inflammation.

    The use of routine stains for candida is not justified unless the patient is immunosuppressed or falls into an at risk group. Routine investigation for oesophageal candidal infection should also include brush cytology.


    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.