Background: Barrett's oesophagus is present in 8-10% of patients with gastro-oesophageal reflux disease (GERD).
Aim: We performed a cohort study to determine the incidence of Barrett's oesophagus in patients with chronic heartburn symptoms.
Methods: We identified patients, with GERD and/or Barrett's oesophagus between 1998 and 2004 by primary or secondary International Classification of Diseases (ICD-9) codes of 530.81 and/or 530.2, who had two or more oesophagogastroduodenoscopies performed at least 6 months apart.
Results: We screened 11 040 patients (41 390 random data entries by ICD-9 code) and enrolled 515 (4.6%) GERD patients and 169 (1.5%) Barrett's oesophagus patients. The mean (+/-s.d.) number of oesophagogastroduodenoscopies in the GERD cohort was 3.2 +/- 1.8 (range: 2-15) over 3.4 +/- 2.2 (range: 0.5-11) years. None of the 412 (80%) GERD patients with non-erosive disease developed Barrett's oesophagus over a mean follow-up time of 3.4 +/- 2.2 years (95% CI: 0-0.9%). Five (1%) of the 103 GERD patients with erosive oesophagitis developed subsequent Barrett's oesophagus. Fifty-seven per cent of the GERD patients were on PPI therapy at the time of index endoscopy. None of the 169 Barrett's oesophagus patients had normal index oesophagogastroduodenoscopies within a mean retrospective time period of 4.5 +/- 2.8 years (95% CI: 0-2%). Using the ICD-9 code of 530.2 as a predictor of the presence of Barrett's oesophagus, the sensitivity was 79% with a specificity of 88%.
Conclusion: The majority of patients with GERD do not appear to develop Barrett's oesophagus when it is not present on the index endoscopy.