Non-ulcer dyspepsia and duodenal eosinophilia: an adult endoscopic population-based case-control study.

Functional abnormalities of the duodenum have been observed in non-ulcer dyspepsia. We aimed to identify whether eosinophils in the upper gastrointestinal tract are a biomarker for non-ulcer dyspepsia. A random sample of an adult Swedish population (n = 1001; mean age, 54 y; 51% female) underwent upper endoscopy. Non-ulcer dyspepsia cases (n = 51, Rome II) and randomly selected controls (n = 48) were identified. Two blinded independent observers assessed the gastroduodenal eosinophil counts. Eosinophils were quantified by counting the number per 5 high-power fields at each of 5 sites (cardia, body, antrum, D1 duodenal bulb, and D2 second portion of duodenum), and total counts were summed over the 5 fields at each site. The odds ratio for non-ulcer dyspepsia (vs asymptomatic controls) in subjects with high duodenal bulb eosinophil counts (median, >/=22, relative to <22) was 11.7 (95% confidence interval, 3.9-34.9), adjusting for age, sex, and H pylori; similar results were observed in D2 (odds ratio = 7.3; 95% confidence interval, 2.9-18.1). A significant association with the number of eosinophil clusters was detected in the duodenum, with higher values in non-ulcer dyspepsia (P < .01). By immunostaining with major basic protein antibody in a subset of duodenal biopsy specimens, eosinophil degranulation was observed in non-ulcer dyspepsia (7 of 15 vs 0 of 5 controls; P = .11). Gastric eosinophil counts were overall not significantly increased in non-ulcer dyspepsia vs controls. Early satiety was associated with eosinophilia in D1 (P = .01) and D2 (P = .02), adjusting for age, sex, and H pylori. Duodenal eosinophilia may characterize a subset of adults with non-ulcer dyspepsia.