Fancourt and Hill (Investigation for iron deficiency anaemia. N Z Med J. 2005;118(1211). URL: http://www.nzma.org.nz/journal/118-1211/1362) have highlighted crucial points about iron deficiency anaemia in their letter.

Firstly, they have shown that faecal occult blood testing is probably worse than useless with so many false negatives and positives. Secondly, a quarter of these anaemic patients will have bowel cancer. Colonoscopy is therefore mandatory. However a disappointing feature is that fewer than half of their patients underwent duodenal biopsy. I know that I have missed the diagnosis of coeliac disease in the past. This condition is surprisingly common, being present in 1–2% of our population. The symptoms are legion and variable, but iron deficiency anaemia is certainly a presentation.

I now routinely biopsy the antrum and duodenum in every gastroscopy, and consequently make this diagnosis in four or five adult patients a year. Coeliac disease may well have been missed somewhere in this study’s cohort of 85.