Routinely diagnosed low-grade dysplasia in Barrett's oesophagus: a population-based study of natural history.

AIMS: To examine the natural history of columnar-lined oesophagus with routinely diagnosed low-grade dysplasia and ascertain the risk of oesophageal adenocarcinoma development. METHODS AND
RESULTS: A multicentre retrospective cohort study of 283 patients with low-grade dysplasia. Follow-up data were obtained from examination of hospital records. One hundred and forty-four patients had biopsies prior to low-grade dysplasia diagnosis and 217 had follow-up biopsies after index low-grade dysplasia diagnosis. In these patients the incidence of high-grade dysplasia and adenocarcinoma combined was 4.6% per annum and of adenocarcinoma alone was 2.7% per annum. At most recent follow-up, 43 (19.8%) had persistent low-grade dysplasia, 37 (17.1%) had changes indefinite for dysplasia and 108 (49.8%) had non-dysplastic columnar-lined oesophagus. When prevalent cases were excluded (those occurring within 1 year of index low-grade dysplasia diagnosis), the annual incidence of high-grade dysplasia and adenocarcinoma combined was 2.2% and of adenocarcinoma alone was 1.4%. The relative risk for adenocarcinoma development in low-grade dysplasia compared with non-dysplastic columnar-lined oesophagus was 2.871 (P = 0.002).
CONCLUSIONS: Low-grade dysplasia has a threefold increased risk of progression to cancer compared with non-dysplastic epithelium, but in the majority of patients dysplasia is not subsequently detected.