Choline and betaine concentrations in plasma discriminate levels of dietary choline intake in healthy adults: analysis of a double-blind randomized crossover controlled feeding study.

Abstract

BACKGROUND: Choline is an essential nutrient, and insufficient intake negatively affects the liver, brain, and muscles. In the United States, habitual choline intake remains below the adequate intake (AI). To date, no circulating metabolites have been validated to distinguish between low and adequate choline intake. OBJECTIVES: We tested whether plasma concentrations of choline and its metabolites could discriminate adequate compared with low dietary choline intake and whether liver elastography (FibroScan) could detect diet-induced changes in liver fat. METHODS: In a double-blind, randomized, crossover feeding study, participants followed three 15-d dietary arms providing ∼100%, 50%, and 25% of the choline AI in the form of choline chloride. On day 12 of each dietary arm, participants consumed a single bolus of 2.2 mmol trimethyl-d9-choline. Targeted assays quantified plasma choline, betaine, phosphatidylcholine (PtdCho), and total homocysteine (tHcy) concentrations. Liver fat content was measured using FibroScan. RESULTS: Plasma concentrations of d9-choline, betaine, and their isotopic enrichment ratio (IER) varied with dietary intake (q < 0.0001), and PtdCho IER also differed significantly (q = 0.001). In targeted analysis, choline and betaine concentrations were highly responsive to dietary choline intake, whereas PtdCho and tHcy were not. Compared with the 100% AI arm, plasma choline was lower in the 25% AI arm [β= -2.20, 95% confidence interval (CI): -2.72, -1.68]. Receiver operator characteristic analysis showed strong discrimination for plasma choline [area under the curve (AUC) = 0.81, 95% CI: 0.74, 0.88], and betaine (AUC = 0.80, 95% CI: 0.73, 0.88), with improved discrimination when combined (AUC = 0.85, 95% CI: 0.79, 0.91). Fibroscan identified a subset of participants with increased liver fat in response to the 25% AI compared with 100% AI choline diet, though patterns varied among individuals. CONCLUSIONS: Plasma choline and betaine concentrations discriminate low compared with AI under controlled feeding. These findings support targeted metabolite profiling to improve choline intake assessment and reveal individual differences in liver response to low choline intake. This study was registered at Choline Nutritional Status: Development of a Biomarker Panel as NCT03726671 (www. CLINICALTRIALS: gov) registered 31 October, 2018.