Published online Oct 14, 2005. doi: 10.3748/wjg.v11.i38.5988
Revised: January 23, 2005
Accepted: January 26, 2005
Published online: October 14, 2005
AIM: To examine whether the fasting levels of serum gastrin-17 (G-17) are lower in Barrett's esophagus (BE) patients than in non-Barrett controls.
METHODS: Nineteen patients with BE (presenting with a tubular segment ≥2 cm long in lower esophagus and intestinal metaplasia of incomplete type ("pecialized columnar epithelium") in endoscopic biopsies from the tubular segment below the squamocolumnar junction were collected prospectively from outpatients referred to diagnostic gastroscopy. The controls comprised 199 prospectively collected dyspeptic outpatients without BE or any endoscopically visible lesions in the upper GI tract. Fasting levels of serum G-17 (G-17fast) were assayed with an EIA test using a Mab highly specific to amidated G-17. None of the patients and controls received therapy with PPIs or other antisecretory agents.
RESULTS: The mean and median levels of G-17fast in serum were significantly lower (P = 0.001) in BE patients than in controls. The positive likelihood ratios (LR+) of low G-17fast to predict BE in the whole study population at G-17fast levels <0.5, <1, or <1.5 pmol/L were 3.5, 3.0, and 2.8, respectively. Among patients and controls with healthy stomach mucosa, the LR+ were 5.6, 3.8, and 2.6, respectively. In the whole study population, serum G-17 was below 2 pmol/L in 15 of 19 BE patients (79%). The corresponding prevalence was 66 of 199 (33%) in controls (P<0.001). The G-17fast was 5 pmol/L or more in only one of the 19 BE patients (5%). In controls, 76 of the 199 patients (38%) had such high serum G-17fast levels (P<0.01).
CONCLUSION: Serum levels of G-17fast tend to be lower in native patients with BE than in healthy controls.