Vitamin B12: Drugs That Distort This Test

A serum vitamin B12 test measures total cobalamin bound to transport proteins (transcobalamin and haptocorrin) in a blood sample. It is distinct from methylmalonic acid (MMA), homocysteine, and holotranscobalamin (often called "active B12") testing, which measure different things and are used to confirm or refute what a serum B12 result suggests. Several common medications interact with this test in one of two very different ways: some genuinely lower the amount of B12 in the body over months to years, while others leave true B12 stores untouched but distort what the assay reports. Confusing the two leads to the wrong response.

The central distinction: true depletion versus assay distortion

The useful question when a patient on one of these medications has an abnormal B12 result is not "is the drug known to affect B12" but "does this drug lower real B12 stores, or does it only change what the test reports." Metformin, PPIs, and possibly oral contraceptives fall into the first category: over time they can genuinely reduce the amount of usable B12 in the body, and a low result reflects real risk. Nitrous oxide and high-dose biotin fall into the second and third categories in opposite directions: nitrous oxide destroys B12's biological activity while the serum assay still reads normal, and biotin distorts the immunoassay chemistry itself without touching true B12 status at all. Treating a biotin-distorted result as if it were true depletion, or trusting a normal serum value in a nitrous oxide user, both lead to missed or mismanaged deficiency.

Vitamin B12 (cobalamin) blood testing can be distorted by several common drug classes, but not in the same direction: metformin, proton pump inhibitors, and long-term oral contraceptive use are linked to genuinely lower B12 stores in observational and trial evidence, nitrous oxide inactivates B12 while serum levels can remain normal, and high-dose biotin (commonly above 5 mg/day) distorts the immunoassay itself. When a patient is on any of these and the serum B12 result is borderline (roughly 200 to 300 pg/mL) or clinically discordant with symptoms, methylmalonic acid testing is the standard way to determine whether functional deficiency is actually present (NIH Office of Dietary Supplements fact sheet: https://ods.od.nih.gov/factsheets/VitaminB12-HealthProfessional/).

Metformin

Metformin is thought to reduce B12 absorption through effects on calcium-dependent uptake in the terminal ileum. This is a case of true, gradual depletion rather than an assay artifact. Multiple observational studies and at least one randomized placebo-controlled trial have linked several years of metformin use to lower serum B12 and higher rates of biochemical deficiency, though reported effect sizes vary across studies and the exact figures from any single trial should be verified against the primary publication before being quoted to a patient.

The American Diabetes Association's Standards of Care recommends periodic B12 monitoring in patients on long-term metformin, particularly those who develop anemia or peripheral neuropathy (see the current ADA Standards of Care: https://diabetesjournals.org/care/issue/47/Supplement_1). A reasonable clinical approach, consistent with that guidance, is to check serum B12 periodically in patients who have been on metformin for several years, and to follow up a borderline value with MMA rather than assuming the low result is either benign or definitively pathological. Oral B12 supplementation is generally effective and does not usually require stopping metformin, given its established benefit for glycemic control.

Proton pump inhibitors and H2 blockers

Gastric acid is needed to release B12 from dietary protein so intrinsic factor can bind it in the small intestine. Long-term acid suppression from PPIs (omeprazole, esomeprazole, lansoprazole, pantoprazole, and others) or H2 blockers can impair this step. Observational studies have associated multi-year PPI use with increased odds of B12 deficiency, with H2 blockers showing a smaller association; exact odds ratios and duration thresholds reported in specific studies should be checked against the primary literature rather than treated as fixed cutoffs.

An important clinical nuance: PPIs impair absorption of B12 that is still bound to food protein, but they do not block absorption of crystalline (supplement-form) B12, because that form does not require acid-mediated release. This is why patients who develop low B12 on a PPI usually respond well to oral crystalline B12 supplementation without needing to stop the PPI, when the PPI is clinically indicated (for example, for erosive esophagitis).

Oral contraceptives and estrogen-containing therapies

Lower serum B12 in users of estrogen-containing oral contraceptives has been reported since older observational studies, but the mechanism is not settled. The leading explanation is redistribution of B12 away from circulating transport proteins into tissue, rather than true tissue-level depletion; a competing explanation involves altered bile acid metabolism affecting absorption.

Decision framework: what a B12 result on an interfering drug actually means

Because the mechanisms above pull in different directions, the same lab value should not be interpreted the same way across all these drugs. This framework organizes the decision by mechanism, not by drug name alone.

Two failure modes are worth naming explicitly. First, treating a biotin-distorted result as a true B12 level, in either direction, leads to unnecessary supplementation or missed real deficiency; the fix is pre-analytical (ask about supplement use and time the draw), not a confirmatory blood test. Second, trusting a normal serum B12 in a patient with known or suspected nitrous oxide exposure and new neurologic symptoms can delay diagnosis of a treatable but potentially disabling myeloneuropathy; in that setting the serum value should be disregarded in favor of MMA and homocysteine, and neurologic symptoms warrant prompt clinical evaluation rather than routine follow-up.

Nitrous oxide: a normal test with an inactivated vitamin

Nitrous oxide oxidizes the cobalt atom at the center of the B12 molecule, which inactivates its enzymatic function without removing it from circulation. The standard serum B12 immunoassay measures total cobalamin and cannot distinguish active from inactivated molecules, so it can report a normal or even high value in someone with significant functional deficiency. This has been described both in surgical patients receiving prolonged nitrous oxide anesthesia and in recreational users of nitrous oxide ("whippets"), some of whom have developed subacute combined degeneration of the spinal cord while their serum B12 remained within the reference range.

Because serum B12 is unreliable in this setting, MMA and homocysteine are the tests that matter. Anyone with known or suspected significant nitrous oxide exposure who develops numbness, tingling, gait or balance problems, or weakness should be evaluated promptly rather than waiting on a repeat serum B12, since these symptoms can progress and are not fully reversible if treatment is delayed.

Anticonvulsants and antimetabolites

Some anticonvulsants (phenytoin, phenobarbital, carbamazepine, valproic acid) have been associated with lower serum B12 in observational studies, though reported effect sizes are inconsistent and the mechanism appears to involve both altered folate-dependent B12 metabolism and, for some agents, impaired absorption. Methotrexate works differently: it blocks the folate-dependent methylation cycle that uses B12 as a cofactor, so functional B12 activity can drop even when the serum level looks normal, and MMA results are variable depending on which part of the pathway is most affected. Patients on methotrexate for rheumatoid arthritis, psoriasis, or oncologic indications are typically monitored for folate and B12 status as part of routine care, often alongside concurrent folic acid supplementation to reduce hematologic toxicity, per standard rheumatology and oncology practice.

Colchicine, at high doses or with prolonged use, has been reported to reduce B12 absorption by affecting ileal mucosal cells; the effect described in older literature was modest and reversible with dose reduction.

These associations are less robustly established than the metformin and PPI evidence, and clinicians should treat exact effect-size claims from any single older study as needing verification rather than as settled figures.

Biotin and other supplement interference

High-dose biotin does not touch true B12 status. Instead, it interferes with the streptavidin-biotin chemistry that many laboratory immunoassays use, including some B12 platforms. Biotin doses above roughly 5 mg per day, commonly found in hair, skin, and nail supplements and used at much higher doses (100 to 300 mg/day) in some multiple sclerosis regimens, can distort results. Depending on the specific assay design, the error can make B12 look falsely normal, falsely low, or falsely elevated.

The FDA has issued a safety communication warning that biotin in a patient's blood or urine sample can cause significantly incorrect results on lab tests that use biotin-based technology, and recommended that patients tell their doctor and lab about biotin supplement use. The practical fix is to stop biotin supplements for at least 72 hours before a B12 blood draw whenever possible, and to flag recent biotin use to the ordering clinician and lab if the test cannot be delayed. This is a pre-analytical problem; no confirmatory test performed on the same sample can correct for it after the blood is drawn.

Confirming true B12 status: MMA, homocysteine, and holotranscobalamin

When any of the drugs above is present and the serum B12 result is borderline (roughly 200 to 300 pg/mL) or inconsistent with the clinical picture, serum B12 alone is not enough.

Methylmalonic acid (MMA) rises when B12-dependent methylmalonyl-CoA mutase activity is impaired and is considered more specific to B12 status than homocysteine, which also rises with folate deficiency, hypothyroidism, and renal impairment. Elevated MMA together with elevated homocysteine in a patient on a B12-affecting drug makes true functional deficiency substantially more likely than either marker alone. Holotranscobalamin ("active B12"), which measures the fraction of B12 bound to transcobalamin II, responds earlier to depletion than total serum B12 and appears less affected by estrogen-related redistribution, though it is not yet universally available or used as a first-line test.

What is established, what is plausible, and what is not established

Established: PPIs and metformin are associated with reduced B12 absorption or availability over long-term use, this is reflected in professional guidance to monitor B12 in these populations, and MMA is a more specific confirmatory marker of functional B12 deficiency than serum B12 alone. Biotin's interference with streptavidin-biotin immunoassays is an established analytical problem recognized by the FDA.

Plausible but not fully settled: the exact magnitude of B12 reduction from oral contraceptives, anticonvulsants, and colchicine, and whether oral contraceptive-associated low B12 usually represents true tissue depletion versus redistribution. Reported percentages and odds ratios from individual older studies in this space vary and specific figures should be checked against the primary paper before being used in patient-facing communication.

Not established from the material reviewed here: a validated, universally adopted screening interval for every drug class discussed (only metformin and PPI use have reasonably clear guideline-level monitoring recommendations), and a single confirmatory test that performs equally well across all these mechanisms, MMA is the best available option but is not perfectly specific or sensitive in every scenario described above.

When to seek care rather than wait for a repeat test

New numbness, tingling, weakness, gait or balance problems, or cognitive changes in someone on a B12-affecting medication, or with any history of nitrous oxide exposure, warrant prompt medical evaluation rather than a routine follow-up lab draw, because B12-related neurologic damage can become permanent if treatment is delayed. A single low or borderline B12 result without symptoms is not an emergency and can typically be worked up with MMA and a discussion with the prescribing clinician about whether the causative medication needs to be continued, adjusted, or supplemented around.