Vitamin B12 and one-carbon biomarker assay interference: a systematic review, with a consensus-derived diagnostic algorithm proposed for Gulf primary care

Abstract

Vitamin B12, holotranscobalamin, and folate are measured almost entirely by immunoassay, a technique susceptible to analytical and pre-analytical interference, while methylmalonic acid and homocysteine are measured by heterogeneous methods, including mass spectrometry. In the Gulf Cooperation Council, where measured deficiency is high, concentrates in metformin-treated type 2 diabetes, and is investigated on immunoassay-dominated platforms with limited access to confirmatory functional markers, that susceptibility carries particular clinical weight. This systematic review, reported in accordance with PRISMA 2020 and the Synthesis Without Meta-analysis guideline and registered retrospectively, searched four bibliographic databases last run on 24 March 2026 and identified 24 eligible studies yielding 73 discrete interference observations: nine case reports or series, six observational studies, five method comparisons, and four in-vitro interference studies. The largest effects were based on single or non-independent studies. Findings were grouped into four categories that differ in the corrective action they imply: direct immunoassay interference, platform and method bias, pre-analytical effects, and biological confounding. Macrocobalamin was the most frequently reported mechanism, falsely raising both total B12 and holotranscobalamin, so the marker used to adjudicate an implausible result is displaced by the same interferent; the two polyethylene glycol criteria disagreed, and neither cross-platform measurement nor serial dilution reliably excluded it. Biotin produced analyte-specific errors on a streptavidin-biotin platform, falsely raising total B12 at low concentrations while falsely lowering folate by up to 25.9%, and susceptibility tracked capture chemistry rather than assay format. Antibody interference moved results in both directions, and a transcobalamin gene variant lowered holotranscobalamin approximately tenfold, the largest error identified. Haemolysis, icterus, and lipaemia lowered homocysteine at interferent concentrations below those at which no interference is claimed, so the functional-marker safety net failed in the same direction as the error it exists to detect. Evidence of direct interference was strong for total B12, moderate to strong for holotranscobalamin, moderate for folate and homocysteine, and absent for methylmalonic acid. The clinically consequential error is the falsely reassuring result that conceals a treatable deficiency. A pre-analytically-aware diagnostic pathway is proposed for Gulf primary care; it is an unvalidated synthesis-and-consensus construction and requires prospective evaluation.