Introduction Oesophageal Candidiasis (OC) is strongly associated with various immunocompromised states. We noted locally that some endoscopists feel confident in their ability to identify the condition visually at OGD, and therefore do not biopsy the oesophagus. The aim of this retrospective study was to analyse the outcomes for a cohort of patients diagnosed with OC at endoscopy to see whether oesophageal biopsy affected outcomes.

Method All OGDs performed at DCH over one year (21/03/13–21/03/14) were reviewed. Patients diagnosed with OC at endoscopy were followed up. Three groups were identified: Group 1, Histology proven OC; Group 2, oesophageal histology suggesting alternative diagnosis; and Group 3, patients not biopsied. Diagnoses and patient outcomes were assessed between 10 and 22 months following the initial OGD.

Results 3375 OGDs were performed. 103 patients (63 female, 40 male), mean age 72.9 years, (range 34–97 yrs) had endoscopic evidence of possible OC. 49 patients (47.5%) had an oesophageal biopsy. Only 61% of these biopsies confirmed OC (Group 1); of these patients one was subsequently diagnosed with HIV, seven took inhaled steroids, seven had cancer, one had intercurrent sepsis, one was reflux related, one had poor diabetic control, two took continuous antibiotics and ten were on a PPI. Mortality in Group 1 was 10% within one month following OGD (all were related to disseminated malignancy). 39% of biopsied patients had histology negative for candida (Group 2) with the following alternative diagnoses made histologically: 16 had reflux oesophagitis, one had HSV oesophagitis, one had eosinophilic oesophagitis and one had bisophosphonate-related oesophagitis. Mortality in Group 2 was 0%. 52.5% of patients with suspected OC did not receive a biopsy (Group 3). Four patients had contraindications to biopsy: two oesophageal varices, two were over anti-coagulated. 57% of these patients had strong risk factors for OC pre-endoscopy: 12 were on long term inhaled/oral steroid, 11 had cancer, three were acutely malnourished, two had decompensated liver disease, and two were on other immunosuppressant medication. Mortality in Group 3 was high: 42.6% in six months (Range: 0–17 months) with varied causes of death that included fungal pneumonia and colonic adenocarcinoma.

Conclusion We suggest all patients with endoscopically suspected OC should be biopsied unless there are compelling reasons not to, as up to 40% of patients will have an alternative diagnosis despite similar endoscopic appearances. Histologically proven OC is often associated with immunocompromise, it is therefore important for endoscopists to advise upon further investigation, particularly in the setting of GP open access endoscopy. Our data suggest a high mortality in this group although treatable diseases may be found.

Disclosure of interest None Declared.