Studies of biomarker responses to intervention with vitamin B-12: a systematic review of randomized controlled trials.

Systematic review of randomized controlled trials in humans, totaling 506 participants across 8 studies. Included populations were mainly adults and elderly participants with low or borderline baseline vitamin B-12 status, including several elderly cohorts with mild deficiency or risk of deficiency; one study involved children/adolescents.

Oral vitamin B-12 supplementation was evaluated across 8 randomized controlled trials, using cyanocobalamin or methylcobalamin delivered as capsules, added to milk, or food-based delivery. Doses ranged from very low physiologic amounts to pharmacologic regimens, including 1000 lg/d cyanocobalamin in one capsule arm and 1000 lg/d cyanocobalamin added to milk in another; most regimens were daily and lasted 4 to 24 weeks.

Oral vitamin B-12 supplementation reliably improved biomarkers of vitamin B-12 status. Serum/plasma total vitamin B-12 increased across 8 RCTs (WMD 185 pmol/L; 95% CI 107, 263; n = 506; P < 0.00001), while methylmalonic acid decreased (WMD 20.28 lmol/L; 95% CI 20.35, 20.22; n = 186; P < 0.00001) and total homocysteine decreased (WMD 23.3 lmol/L; 95% CI 24.5, 22.2; n = 282; P < 0.00001). Holotranscobalamin could not be evaluated reliably because it was measured in only one study. The authors concluded that serum/plasma total vitamin B-12, methylmalonic acid, and total homocysteine are responsive markers of intake after oral supplementation, although heterogeneity was substantial for total vitamin B-12.

The evidence base was small and heterogeneous, with varying doses, formulations, and study durations across the trials. Serum/plasma total vitamin B-12 showed substantial heterogeneity (I2 = 94.6%), and holotranscobalamin evidence was insufficient because only one study measured it. Some subgroup conclusions were limited by sparse data, and generalizability is constrained by the predominance of elderly participants with low or borderline baseline status.