Biomarker relationships with small bowel histopathology among malnourished children with environmental enteric dysfunction in a multicountry cohort study.

ABSTRACT

BACKGROUND: Validated biomarkers could catalyze environmental enteric dysfunction (EED) research. OBJECTIVES: Leveraging an EED histology scoring system, this multicountry analysis examined biomarker associations with duodenal histology features among children with EED. We also examined differences in 2-h compared with 1-h urine collections in the lactulose rhamnose (LR) dual sugar test. METHODS: Three cohorts of undernourished children unresponsive to nutrition intervention underwent esophagogastroduodenoscopy and duodenal biopsies. Histopathology scores were compared to fecal calprotectin (CAL), myeloperoxidase (MPO), neopterin (NEO), and urinary LR ratio and lactulose percentage recovery. Log-transformed biomarkers were used in linear regressions adjusted for age, center, and sample collection-biopsy time interval in multivariable models. RESULTS: Data on >1 biomarker were available for 120 Bangladeshi (CAL, MPO, NEO, and LR), 63 Pakistani (MPO, NEO, and LR), and 63 Zambian children (CAL). Median age at endoscopy was similar (19 mo) across centers. Median sample collection prior to endoscopy was consistent with each center's study design: 2 wk in Bangladesh (urine and stool) and Zambia (stool), and 6 (urine) and 11 (stool) mo in Pakistan. In multivariable models, intraepithelial lymphocytes were associated with CAL (exponentiated [exp.] coefficient 1.19; 95% confidence interval [CI] 1, 1.41), intramucosal Brunner's glands with MPO (exp. coefficient 1.33; 95% CI 1.05, 1.69) and NEO (exp. coefficient 1.37; 95% CI 1.1, 1.7), and chronic inflammation with NEO (exp. coefficient 1.61; 95% CI 1.17, 2.17). Intraepithelial lymphocytes were associated with lactulose % recovery (exp. coefficient 1.22; 95% CI 1.05, 1.41). LR recovery was substantially lower in 1-h collections than in 2-h collections. CONCLUSIONS: Four commonly used markers of enteric dysfunction were associated with specific histologic features. One-hour urine collection may be insufficient to reflect small bowel permeability in LR testing. While acknowledging the challenges with obtaining relevant tissue, these findings form the basis for further EED biomarker validation research.