B12 Functional Deficiency

Healthcare · §608

A normal vitamin B12 result on your blood panel doesn't mean your cells are getting enough. The standard test measures mostly an inactive form of the vitamin, and the low-end cutoff was set on red-blood-cell adequacy — not on what your nerves and brain actually need. The cost of missing it is months of unexplained fatigue, tingling fingers, and brain fog, with some patients left holding damage that doesn't fully reverse. Two cheap follow-on tests close the gap.

Test · As-needed Evidence Moderate Chapter Healthcare

For someone who's actually deficient, the response to treatment is one of the cleanest in medicine — the fatigue, the foggy mornings, the pins-and-needles all lift in weeks. The catch is that the standard B₁₂ number on your panel routinely misses the diagnosis. The test costs about as much as a meal out; a year of pills costs less than that. The downside of catching it years late is permanent nerve damage. Vegans, anyone over 60, long-term metformin or acid-blocker users, and people with unexplained fatigue or neuropathy are the audience.

B₁₂ (cobalamin) does exactly two jobs in your body, and a shortage breaks both. The first job is keeping the methylation cycle running — the chemistry that recycles a compound called homocysteine into methionine, builds your DNA, and lays down the fatty insulation around your nerves. The second job helps your mitochondria break down certain fatty acids cleanly. When B₁₂ runs low, homocysteine and another compound called methylmalonic acid (MMA) build up in the blood — and the nerve insulation slowly degrades.

The reason most blood tests miss this: the standard "serum B₁₂" number adds up everything floating around, but only about a fifth of that — the part bound to a transport protein called transcobalamin — actually gets delivered to your cells. The rest is biologically idle. So your total B₁₂ can sit comfortably in the normal range while the active fraction is running short and the MMA is climbing. The deficiency is real; the lab number just isn't sensitive enough to catch it.

The diagnosis nobody catches

The seminal paper is older than most readers think. In 1988, Lindenbaum reported 141 consecutive patients with B₁₂-deficiency nerve and psychiatric symptoms — and 28% of them had no anemia at all. Forty had only neurological signs, no hematological clue whatsoever. All 141 responded to B₁₂ replacement. The conclusion that broke the older textbook story: the nerves can go first, the blood second, sometimes never Lindenbaum et al. 1988.

The two metabolite tests that close the diagnostic gap have been worked out since. Methylmalonic acid rises when B₁₂ falls short at the mitochondrial step; folate doesn't fix it, so it's specific for B₁₂. Homocysteine rises when either B₁₂ or folate is short; it's the broader marker, less specific. Both normalise within weeks of replacement. The combination — a borderline-low B₁₂ with an elevated MMA — is the classical fingerprint of functional deficiency in someone whose blood count looks fine Stabler 2013 Carmel 2008.

Population-scale numbers: roughly 3-6% of all adults and substantially more older adults run biochemical deficiency on metabolite testing, often with a "normal" serum B₁₂ Allen 2009. The cognitive end is where it gets striking.

A longitudinal Swedish cohort tracked older adults for seven years and found that lower active B₁₂ at baseline predicted a higher rate of conversion to Alzheimer's disease, independent of total B₁₂ — pointing the same direction as VITACOG Hooshmand et al. 2010.

What people get wrong

"My B₁₂ came back normal, so I'm fine." Not necessarily. The standard lower cutoff is set near 200 pg/mL — a number drawn from how low you can go before red blood cells start looking abnormal, not how low you can go before nerves start to suffer. People in the 200-450 pg/mL grey zone often turn out to have elevated MMA and real symptoms Aparicio-Ugarriza 2015.

"Anemia would show up first if it were B₁₂." Sometimes. Often not. Most B₁₂-deficient patients with neurological symptoms have a normal red-cell count and a normal cell size when they first walk into a clinic Lindenbaum et al. 1988.

"Folic acid will sort it." Folate (vitamin B₉) cleans up the anemia caused by B₁₂ deficiency without doing anything for the nerves — and by removing the blood-test clue, it can let the neurological damage progress quietly. This is why population folic-acid fortification of flour made metabolite testing more important, not less Reynolds 2006.

"Oral B₁₂ doesn't work if you have a stomach problem." The old textbook line. There's a small passive-diffusion pathway in the gut that doesn't need stomach acid or intrinsic factor. At a daily dose of 1000-2000 µg, that pathway delivers enough B₁₂ to correct even pernicious anemia — a Cochrane review puts oral on par with monthly injections for the standard endpoints Vidal-Alaball et al. 2005.

Who should actually be testing

Five groups carry most of the population risk. If you're in one and you have any of the typical symptoms — persistent fatigue, tingling hands or feet, cognitive haze, mood that's slipped without an obvious reason — the standard B₁₂ on its own is the wrong test.

Strict vegans and long-term vegetarians. Animal foods are essentially the only natural source of B₁₂ in the human diet. Roughly half of unsupplemented adult vegans are biochemically deficient on testing; lacto-ovo vegetarians sit at intermediate risk Pawlak et al. 2013. A reliable daily supplement closes this gap entirely.
Adults over 60. The stomach lining changes with age; roughly a third of people over 70 produce less acid and pepsin — the chemistry that liberates B₁₂ from food protein in the first place. The result is "food-bound malabsorption": synthetic B₁₂ in a pill or fortified food still works fine, but the B₁₂ bound up in a steak doesn't get released Andrès et al. 2004.
Long-term metformin users. A four-year placebo-controlled trial in type-2 diabetes patients showed metformin dropped serum B₁₂ by an average of 19% and roughly doubled the rate of biochemical deficiency de Jager et al. 2010. Most metformin labels carry the warning; routine monitoring is uncommon in practice.
Long-term acid-blocker users. Daily proton-pump inhibitors (omeprazole, pantoprazole and the rest) or H₂-blockers for two years or more raise the risk of B₁₂ deficiency by about two-thirds in the largest population study to date Lam et al. 2013.
Pernicious anemia and gut surgery. Pernicious anemia is the autoimmune destruction of the stomach cells that make intrinsic factor (the carrier B₁₂ needs to cross the gut wall); it affects roughly 2% of adults over 60. Anyone with a gastric bypass, sleeve gastrectomy, Crohn's of the lower small intestine, or celiac disease falls in this group too Stabler 2013.

One more worth flagging: recreational nitrous oxide use (whippets, and in some patients repeated dental anesthesia) chemically inactivates the B₁₂ you already have. Heavy users can develop a full-blown spinal-cord injury on a perfectly normal serum B₁₂ number.

What happens if it gets missed

The cost of catching this late is paid in nerves. In a classical series of 153 patients with B₁₂-deficiency neurological disease, recovery after replacement scaled almost linearly with how long the symptoms had been allowed to run. Patients treated within weeks of onset recovered fully; patients treated past a year often kept a residual deficit that didn't go away Healton et al. 1991.

The way it actually looks from the inside is gradual and easy to dismiss. In the first months: afternoon fatigue that no amount of sleep clears, a heaviness in the legs at the end of a workday, occasional pins-and-needles in the fingers or toes that comes and goes. Most people put it down to stress, age, or a long week. Around the 18-month mark in a missed case, the symptoms become harder to ignore — your partner notices the slower pace on a familiar walk, you stop trusting your balance on stairs in the dark, you lose the thread of a conversation you'd have followed easily a year before. The blood work still says "normal-ish"; the B₁₂ result came back fine eighteen months ago and nobody re-checked it.

By the time the diagnosis lands, some of the damage is fixed in place. The part of the spinal cord that handles position sense and vibration — what tells you which way your foot is angled without looking — repairs slowly when it repairs at all. People treated late often keep a residual unsteadiness on uneven ground for the rest of their lives, even after the B₁₂ is replaced and the blood markers normalise.

The cognitive side runs in parallel. Lower active B₁₂ in your 60s and 70s carries an independent risk of progression to Alzheimer's disease over the following several years Hooshmand et al. 2010. Hearing belongs on the same list of slow costs — a long-missed deficiency is one of the quieter accelerators of age-related hearing loss and the ringing that comes with it. None of this is dramatic on a one-year horizon. It's a slow trade you'd rather not be making.

How to test and treat

The diagnostic pathway runs in three steps.

Start with total serum B₁₂. Below 200 pg/mL: treat, no further questions. Above 450 pg/mL with no symptoms: probably fine. The window in between is where the next step matters.
Add methylmalonic acid (preferred — specific for B₁₂) and / or homocysteine. MMA above 0.40 µmol/L in someone with normal kidneys confirms functional deficiency. Homocysteine above 15 µmol/L is supportive but less specific (it also rises in folate or B₆ shortage and in kidney disease) Stabler 2013 Devalia et al. 2014.
If the test is positive, look for cause: diet history, age, metformin or acid-blocker use, prior gut surgery, and — when pernicious anemia is plausible — an intrinsic-factor antibody test.

Total cost of the standing protocol is well under $50/year: a tablet a day costs less than a coffee a month, and the follow-up tests are usually insured when ordered with clinical reason.

What changes when you fix it

When the diagnosis is right, the response to treatment is one of the cleanest in medicine. Inside a week of starting replacement, fresh red blood cells flood out of the bone marrow — measurable on a blood test by day three to five if anyone bothers to check Carmel 2008.

The lived version: by the end of week two, the afternoon energy crash starts to soften — the version of you who needed a coffee and a sit-down at 3pm becomes the version that gets through the afternoon without thinking about it. By week four to six, the pins-and-needles retreat in the order they appeared — fingertips first, toes last. Glossitis (the sore, smooth, beefy-red tongue some patients develop) resolves over the same window. The mood symptoms — the low-grade irritability or unexplained depression that some patients didn't even connect to anything physical — lift in the same one-to-two-month window Lindenbaum et al. 1988.

Cognitive recovery is slower and patchier. Mild brain fog clears over a couple of months; the older the patient and the longer the deficit ran, the longer the tail. The structural payoff is what the brain-imaging data point at: the brain-atrophy curve flattens once the methylation cycle is back up to speed Smith et al. 2010.

At the year mark, the typical replaced patient barely remembers the pre-treatment baseline. The way people describe it, almost universally: "I didn't realise I'd been running on fumes."

Where this goes wrong in practice

Treating once and stopping. If the cause is permanent — pernicious anemia, gastric bypass, ongoing metformin — stopping replacement guarantees a relapse on a one-to-four-year timescale. The liver carries a meaningful B₁₂ reserve, but it isn't infinite.
The 25-µg multivitamin. A standard multi doses B₁₂ at a level that works for normal absorption and is useless for pernicious anemia or significant food-bound malabsorption. You need the 1000+ µg daily dose to drive the passive-diffusion pathway.
The "my urine turned yellow, must be working" trap. The bright yellow urine after a multivitamin is riboflavin (vitamin B₂) being excreted. It says nothing about whether your cells actually got any B₁₂.
Folate masking. A folic-acid-fortified diet or supplement can normalise the red-cell picture in a B₁₂-deficient patient while the nerve damage progresses underneath. This is why metabolite testing — not the blood count — is the right monitoring tool Reynolds 2006.
Checking serum B₁₂ to see if replacement is working. Serum B₁₂ always rises once you supplement; that number doesn't tell you whether enough is reaching your cells. Recheck the MMA and homocysteine — those are the actual response markers.
Kidney disease confounding. Both MMA and homocysteine rise in chronic kidney disease independently of B₁₂ status — which is why holotranscobalamin (the active-fraction test, widely available in Europe) is often the better choice in patients with reduced kidney function.
Pinning the nerves on B₁₂ when it's copper. Copper deficiency produces a near-identical nerve and spinal-cord picture, and turns up after gut surgery or heavy zinc use. If the metabolites come back clean but the neurology doesn't, check copper too before closing the case.

What's next door

A few adjacent topics worth looking up if this entry resonates: folate (vitamin B₉) deficiency, which interacts tightly with B₁₂ and is the other half of the megaloblastic-anemia story; pernicious anemia as a stand-alone autoimmune condition; the homocysteine-and-cardiovascular question (largely null in big trials, despite the brain-atrophy signal); and consumer-genetic MTHFR testing, which has a smaller real-world impact than its marketing suggests.

Related in the handbook

Worsens / aggravates

— Years on a PPI cut the stomach acid you need to pull B12 from food — a common, quiet cause of functional deficiency.

Diagnosis

— When the basic B12 looks fine but symptoms persist, these are the follow-up markers that reveal the true deficiency.

— One thing a missed B12 deficiency does over years: speed up hearing loss and ringing in the ears.
— Before pinning nerve symptoms on B12, copper deficiency can mimic it — check both.
— Folate and B12 run the same one-carbon pathway — a folate excess can mask B12 deficiency on a blood count, so check both together.
— Same trap, different nutrient: iron can be depleted while your blood count still looks fine.
— This is the textbook case of a lab 'normal' that isn't — the B12 cutoff was set for blood cells, not nerves.
— B12 deficiency swells red cells, so a high MCV is a clue — but it shows up late and a normal count won't clear you.
— Metformin is a classic driver of the functional B12 gap these extra tests are built to catch.
— A long-standing H. pylori infection is one cause of the B12 shortfall this covers.

Substance + claimed effects

B₁₂ functional deficiency is the state in which cells do not receive enough biologically active cobalamin to run the two B₁₂-dependent enzymatic steps — even though serum total B₁₂ sits inside the laboratory reference range (commonly 200–900 pg/mL, with most labs flagging only values below ~200 pg/mL). The deficit is detected biochemically: elevated methylmalonic acid (MMA, the substrate of the mitochondrial reaction) and elevated total plasma homocysteine (the substrate of the cytoplasmic reaction) are the two metabolite markers that turn abnormal when intracellular B₁₂ is insufficient, often months to years before serum B₁₂ drops below the conventional cutoff Stabler 2013 Carmel 2008. Holotranscobalamin (active B₁₂, the fraction bound to transcobalamin and delivered to cells) is the direct upstream measure and falls earlier than total B₁₂ Aparicio-Ugarriza 2015. Claimed consequences — covered as a single substance with multiple effects in this entry — span neurology (paresthesias, gait disturbance, subacute combined degeneration of the spinal cord, neuropsychiatric features), hematology (macrocytosis and megaloblastic anemia, often a late and unreliable sign), cognition (impaired memory, executive dysfunction, accelerated brain atrophy with elevated homocysteine), mood (depressive and irritability syndromes that resolve with replacement in deficient patients), and daily energy (fatigue from both anemia and impaired mitochondrial intermediary metabolism).

Evidence by addressing question

Mechanism

Cobalamin is the cofactor for exactly two human enzymes. The first, methionine synthase (cytoplasm), uses methylcobalamin to remethylate homocysteine to methionine, regenerating S-adenosylmethionine (SAM) — the universal methyl donor for DNA, RNA, neurotransmitters, and myelin basic protein. When methylcobalamin is low, homocysteine accumulates in plasma and SAM falls, methylation reactions slow, and 5-methyltetrahydrofolate is trapped (the "folate trap"), producing a secondary functional folate deficiency that drives megaloblastic erythropoiesis Reynolds 2006. The second, methylmalonyl-CoA mutase (mitochondria), uses adenosylcobalamin to convert methylmalonyl-CoA to succinyl-CoA in the propionate / odd-chain fatty acid pathway. When adenosylcobalamin is low, MMA accumulates and odd-chain (C15/C17) and branched-chain fatty acids are incorporated into the myelin sheath in place of the normal even-chain residues, destabilising the sheath and contributing to the demyelinating lesions of subacute combined degeneration of the dorsal and lateral columns Reynolds 2006 Healton 1991.

Holotranscobalamin (holo-TC) is the ~20% fraction of plasma B₁₂ bound to transcobalamin II, which carries the vitamin across the cell membrane via the TCblR/CD320 receptor. The remaining ~80% is bound to haptocorrin, has no known transport role for tissue delivery, and is cleared by the liver. Total serum B₁₂ therefore reports the haptocorrin pool dominantly, which is why a "normal" total B₁₂ can coexist with a low holo-TC, a rising MMA, and a clinically deficient patient Carmel 2008 Aparicio-Ugarriza 2015. The biochemical sequence as B₁₂ status declines is: holo-TC falls first, MMA and homocysteine rise next, total B₁₂ declines last, and overt anemia or macrocytosis appears last of all — sometimes never.

Evidence

The foundational clinical-epidemiology paper is Lindenbaum's 1988 NEJM series of 141 consecutive patients with cobalamin-deficiency neuropsychiatric disorders, of whom 28% had no anemia and no macrocytosis at presentation; in 40 patients, neurological signs were the only manifestation, and all 141 responded to cobalamin replacement Lindenbaum et al. 1988. This established the canonical claim that neurological disease can precede, and occur in isolation from, the hematological picture — the foundation of the functional-deficiency concept.

The Framingham Offspring Study and other large cohorts subsequently showed that plasma MMA and homocysteine are elevated in a meaningful fraction of community-dwelling older adults with "normal" serum B₁₂ in the 200–400 pg/mL range, and that supplementation normalises both metabolites — establishing MMA as the most specific functional marker (folate corrects homocysteine but not MMA) Allen 2009 Clarke et al. 2003. NHANES-class population data show population prevalence of low or borderline B₁₂ on the order of 3–6% overall and substantially higher in adults over 60 Allen 2009.

The VITACOG randomised trial in 271 adults with mild cognitive impairment showed that two years of high-dose B₁₂ + B₆ + folic acid reduced the rate of whole-brain atrophy by ~30% vs placebo, with the largest effect in the tertile with the highest baseline homocysteine — the most direct human evidence that the methylation-cycle disturbance of functional B₁₂ insufficiency has structural brain consequences and is partially reversible by replacement Smith et al. 2010. The longitudinal Kungsholmen cohort found that higher holotranscobalamin at baseline was associated with a lower seven-year risk of Alzheimer's disease (per pmol/L increase, hazard ratio 0.980), independently of total B₁₂ Hooshmand et al. 2010.

For testing and treatment, the British Society for Haematology guideline (Devalia 2014) and the Carmel 2008 Blood review concur that (a) total B₁₂ is insensitive in the 200–350 pg/mL range and metabolite testing is indicated when clinical suspicion is high, (b) high-dose oral cobalamin (1000–2000 µg daily) achieves passive absorption that bypasses the intrinsic-factor mechanism and corrects most deficiencies — including pernicious anemia, by Cochrane meta-analysis — and (c) parenteral hydroxocobalamin remains first-line for severe neurological disease, malabsorption, and patients who cannot reliably take a daily oral dose Devalia et al. 2014 Carmel 2008 Vidal-Alaball et al. 2005.

Misconceptions

Four widely-held but wrong beliefs:

"Normal serum B₁₂ means no deficiency." The conventional lower cutoff (~200 pg/mL) was set on hematological adequacy in unselected lab populations; it routinely misses biochemically and clinically deficient patients in the 200–450 pg/mL "grey zone" Aparicio-Ugarriza 2015 Carmel 2008.
"Anemia comes first." Lindenbaum's series and subsequent case-series data show that the majority of B₁₂-deficient patients with neuropsychiatric presentations have a normal MCV and a normal hemoglobin Lindenbaum et al. 1988 Reynolds 2006.
"Folic acid will fix it." Supplemental folate corrects megaloblastic anemia caused by B₁₂ deficiency but does not correct the underlying neurological lesion — and may worsen it by masking the hematological signal. This is why population folic-acid fortification raised the relative weight of metabolite testing for B₁₂ diagnosis Reynolds 2006 Stabler 2013.
"Oral B₁₂ doesn't work for pernicious anemia." A passive-diffusion ~1% absorption pathway exists independent of intrinsic factor; at 1000–2000 µg daily oral doses this is sufficient for most pernicious-anemia patients, as shown by RCTs and a Cochrane review Vidal-Alaball et al. 2005.

Audience

Population prevalence of functional B₁₂ insufficiency is concentrated in identifiable groups Allen 2009 Wolffenbuttel et al. 2019:

Vegans and long-term vegetarians. Pooled prevalence of biochemical B₁₂ deficiency in adult vegans approaches ~50% without supplementation; lacto-ovo-vegetarian rates run lower but still substantially exceed omnivore baseline Pawlak et al. 2013.
Adults over 60. Atrophic gastritis (progressive loss of parietal cells, ~30% prevalence by age 70) impairs the acid-pepsin step that releases B₁₂ from food protein; food-bound malabsorption is the dominant mechanism in this group, distinct from pernicious anemia (autoimmune intrinsic-factor antibody) Andrès et al. 2004 Wong 2015.
Long-term metformin users. The HOME RCT (390 type-2 diabetes patients, 4.3 years) showed metformin reduced serum B₁₂ by an absolute mean of 19% vs placebo and doubled the incidence of biochemical deficiency; absorption interference is calcium-dependent at the ileal receptor de Jager et al. 2010.
Long-term PPI / H₂-blocker users. A nested case-control study (Kaiser, n=25,956 cases) showed ≥2 years of PPI use was associated with a 1.65× increase in B₁₂ deficiency vs non-use Lam et al. 2013.
Pernicious anemia. Autoimmune destruction of gastric parietal cells; lifetime prevalence ~0.1% overall and ~2% in adults over 60. Lifetime replacement required Stabler 2013.
Post-gastric-surgery / ileal disease. Bariatric surgery (especially gastric bypass and sleeve), Crohn's disease involving the terminal ileum, celiac disease, and any ileal resection compromise the intrinsic-factor-mediated absorption pathway.
Nitrous-oxide exposure. N₂O irreversibly oxidises the cobalt centre of methylcobalamin; recreational ("whippet") use and repeated anesthesia precipitate functional deficiency, sometimes overt myelopathy, even on a normal-range serum B₁₂.

Protocol

Testing pathway with strongest yield in the suspected functional-deficiency range:

Total serum B₁₂ as the screen. Below 200 pg/mL: treat. Above 450 pg/mL with low clinical suspicion: usually adequate. In between, the test is uninformative on its own.
Add MMA (preferred — specific for B₁₂) and homocysteine if the clinical picture or risk profile justifies it. MMA > 0.40 µmol/L with no renal impairment supports functional deficiency; homocysteine > 15 µmol/L in the absence of folate/B₆ deficiency or renal disease is supportive but less specific Stabler 2013 Devalia et al. 2014.
Holotranscobalamin (where available) — direct measure of bioactive B₁₂. Below ~35 pmol/L is concerning Aparicio-Ugarriza 2015.
If deficient, look for cause: intrinsic-factor antibody (pernicious anemia), parietal-cell antibody, diet history, metformin / PPI history, prior gastric / ileal surgery.

Treatment options, ranked by usual practice:

Oral cyanocobalamin or methylcobalamin 1000–2000 µg daily. First-line for most dietary or food-bound malabsorption deficiency, and effective for pernicious anemia via passive ~1% diffusion. Cochrane review found oral non-inferior to intramuscular for hematological and neurological endpoints at 90 days Vidal-Alaball et al. 2005.
Intramuscular hydroxocobalamin 1000 µg. Loading schedule (e.g., every other day for 1–2 weeks) then maintenance every 1–3 months. First-line for severe neurological disease, severe anemia, or when oral adherence is uncertain. Hydroxocobalamin has the longest half-life of the available forms Devalia et al. 2014.
Lifetime replacement when the cause is irreversible (pernicious anemia, ileal resection, ongoing metformin without alternative).

Form-of-B₁₂ debate: cyanocobalamin is the cheapest and most-studied; the body converts it to methyl- and adenosylcobalamin. Methylcobalamin and hydroxocobalamin are alternatives. No RCT shows a clinically meaningful difference for the typical deficiency patient; the form preference is largely commercial.

Stakes

The clinical hazard specific to B₁₂ functional deficiency is that the neurological lesion — subacute combined degeneration of the spinal cord plus peripheral and optic neuropathy — can become irreversible if treatment is delayed. In Healton's classical series of 153 patients, neurological recovery correlated tightly with duration of symptoms before treatment; patients with >12 months of symptoms had substantially less recovery than those treated within weeks Healton et al. 1991. Combined with Lindenbaum's finding that ~28% of B₁₂-deficient patients with neurological disease have no hematological clue, the population-level consequence is that an underused metabolite test plus an over-relied-on serum B₁₂ threshold lets some patients accumulate permanent demyelinating injury Lindenbaum et al. 1988.

Cognitive stakes: elevated homocysteine is an independent risk factor for dementia in multiple cohorts, with the VITACOG trial showing partial reversibility of the brain-atrophy trajectory when B₁₂ + folate + B₆ normalise homocysteine in mild cognitive impairment Smith et al. 2010 Hooshmand et al. 2010.

Payoff

Replacement reverses fatigue, paresthesias, glossitis, mood disturbance, and (when caught early) gait and cognitive symptoms. Hematological recovery is typically complete within 2–8 weeks; reticulocytosis appears within 3–5 days of treatment Carmel 2008. Neurological recovery is slower (weeks to months) and incomplete when treatment is delayed beyond ~6–12 months of symptom onset Healton et al. 1991. In the asymptomatic biochemical-deficiency case (raised MMA, no overt symptoms), the payoff is structural: normalised methylation cycle, lower long-term dementia risk per the homocysteine literature, and lower risk of progression to overt neuropathy.

Failure modes

Treating once and stopping. When the cause (pernicious anemia, ileal disease, ongoing metformin) is permanent, stopping replacement guarantees recurrence on a timescale of 1–4 years for healthy livers (large hepatic reserves), shorter for already-depleted patients.
Low-dose oral in malabsorption. The standard B₁₂ in a multivitamin (6–25 µg) is insufficient to drive passive diffusion in pernicious anemia or post-gastric-surgery patients; the dose has to be 1000+ µg to load the passive pathway.
Folate masking. A folic-acid-fortified diet plus a multivitamin can normalise the MCV in a B₁₂-deficient patient while the neurological lesion progresses, removing the hematological clinical signal Reynolds 2006.
Failure to retest. After oral replacement starts, serum B₁₂ rises predictably but tells you nothing about whether the dose is sufficient; MMA and homocysteine are the only validated response markers.
Renal-impairment confounding. Both MMA and homocysteine rise in chronic kidney disease, independent of B₁₂ status — limiting their specificity in patients with reduced GFR.

Practicalities

Cost of testing: total B₁₂ is typically included in basic metabolic panels at near-zero marginal cost. MMA and homocysteine each add roughly $30–80 out of pocket in the US; covered by most insurers when ordered with clinical suspicion. Holotranscobalamin is widely available in Europe, less so in the US. Cost of supplementation: 1000 µg oral cyanocobalamin tablets are widely available at <$15/year. Intramuscular hydroxocobalamin requires prescription and clinic visits or self-injection; cost is dominated by visits, not the vial.

The credibility range

Optimist case. The functional-deficiency framing is one of the cleanest examples in clinical biochemistry of a "normal lab, real disease" gap. The mechanism is fully worked out at the enzyme level; MMA is a specific substrate marker; Lindenbaum demonstrated isolated neurological disease as standard rather than rare; the VITACOG trial showed structural reversibility in MCI; and oral replacement is cheap, safe, and effective. Reference ranges set decades ago on hematological adequacy are demonstrably too low to protect the nervous system, and the cost of a higher threshold is essentially zero — B₁₂ toxicity is not a clinically meaningful problem at oral or even parenteral doses. The conservative answer is to lower the conventional cutoff and use metabolite testing more freely.

Skeptic case. Population-screening trials for B₁₂ + homocysteine have not consistently shown morbidity-mortality benefits outside the MCI subgroup, and several large homocysteine-lowering RCTs (HOPE-2, VISP, NORVIT) failed to reduce cardiovascular endpoints despite normalising homocysteine — suggesting that homocysteine elevation may be a marker rather than a causal mediator outside the brain. Population frequency of truly symptomatic functional deficiency is uncertain because both MMA and holo-TC have grey-zone cutoffs and renal interference. Aggressive treatment of borderline numbers in asymptomatic patients has no high-quality trial backing. And the "B₁₂ fixes fatigue" claim, broadly marketed, is true in deficient patients and not clearly true in replete ones — the supplement industry blurs that distinction.

Author's call. The functional-deficiency concept is settled clinical biochemistry. The bias on the testing side runs one way — under-diagnosis from over-reliance on a single insensitive assay — and the cost of correcting that bias (a 1000-µg daily tablet) is trivial. The article should land confidently on: in any patient with unexplained neuropathy, cognitive decline, fatigue, or in any of the named risk groups, do not rely on total serum B₁₂ alone; order MMA (and/or homocysteine, and/or holo-TC) when clinical suspicion outweighs a borderline-normal result; treat decisively when the picture supports deficiency; and prefer oral 1000–2000 µg daily as a low-risk standing protocol for high-risk groups (vegans, long-term metformin users, adults over 60 with relevant symptoms). Controversy score is low on the diagnosis-side framing; moderate on the cardiovascular-prevention claims.

Stakeholder + incentive map

Hematology / neurology specialists — long-standing advocates of metabolite testing; the Lindenbaum / Carmel / Stabler tradition. Push for lowered serum B₁₂ thresholds and more liberal MMA ordering.
Primary-care guidelines bodies (USPSTF, NICE) — historically conservative on routine screening; require RCT-level evidence of net benefit, which exists for treatment of biochemical deficiency but not for asymptomatic-population screening.
Reference-lab industry — sets cutoffs based on populations that are not deficiency-free, biasing thresholds downward (the 200 pg/mL cutoff was set when much of the "reference" population had subclinical disease).
Pharma — metformin and PPI manufacturers — have historically under-emphasised the B₁₂ interaction; label warnings exist but routine monitoring is uncommon in practice.
Supplement industry — over-markets B₁₂ as an energy product to the general population, blurring the distinction between deficient and replete consumers and generating skepticism on the medical side.
Vegan / plant-based community — strongly aware and largely well-informed; routine high-dose supplementation is the standard recommendation from credible vegan dietary bodies.

Population variability

Several axes matter:

Age. Atrophic gastritis prevalence rises steeply with age — under 30: rare; 60+: ~30%; 80+: ~40%. Older adults have less reserve, slower onset of overt symptoms, and disproportionately neurological presentations.
Diet pattern. Strict vegans without supplementation will become deficient on a 2–5 year timescale; lacto-ovo-vegetarians at intermediate risk; omnivores eating meat or eggs daily are usually replete on diet alone (animal foods are essentially the only natural dietary source).
Genetics. Common polymorphisms in MTHFR (C677T), MTR, MTRR, TCN2, and FUT2 modulate the methylation cycle and B₁₂ absorption / transport, but the clinical impact in well-nourished individuals is modest; MTHFR-driven recommendations are oversold in consumer-genetics marketing relative to the trial evidence.
Renal function. MMA and homocysteine both accumulate in CKD, reducing the specificity of metabolite testing — holo-TC is less affected and preferred where available.
Liver reserve. Hepatic B₁₂ stores are large (1–5 mg) relative to daily losses (~1–3 µg), so the onset of deficiency after intake stops is slow (months to years) — and replenishment is slow once depleted.
Pregnancy. Fetal demand depletes maternal stores; deficiency in a vegan mother causes severe and sometimes irreversible developmental neurological injury in the infant — a separate failure mode worth flagging.

Knowledge gaps

Open questions:

The "true" lower bound of adequate B₁₂ for the nervous system in healthy adults is not settled — proposed cutoffs range from 200 pg/mL to 400 pg/mL depending on which biomarker (MMA, holo-TC, clinical outcome) the lower bound is anchored to.
Whether asymptomatic biochemical deficiency in younger adults (MMA elevated, no symptoms) carries long-term neurological or cognitive risk is plausible but not directly trialled.
The size of the cognitive-aging effect — VITACOG was a 271-patient MCI trial, has not been independently replicated at scale, and several follow-up trials in lower-baseline-homocysteine populations were null. The "B₁₂ + folate slow Alzheimer's" claim is more cautious than it sometimes appears in the trade press.
Cardiovascular endpoints from B-vitamin-driven homocysteine lowering: largely null in HOPE-2, VISP, NORVIT, suggesting homocysteine outside the brain is more marker than mediator.
Optimal monitoring interval after starting replacement — established practice is empirical, not trial-based.
Comparative efficacy of cyano- vs methyl- vs hydroxocobalamin in head-to-head clinical-outcome trials is weak; current preferences are largely tradition, regional regulation, and commercial.

Scoping. All four consequences named in the brief — neurological function, hematology, fatigue, cognition — are covered end to end across mechanism, evidence, stakes, payoff. Mood is added as a fifth named consequence since the literature (Lindenbaum 1988) explicitly includes psychiatric presentations and the dimension scores reflect it. No narrowing relative to the brief.

Holistic-scoring caveat. Per entry/entry.md §1a, dimension scores reflect what the substance does for the population for whom it applies — i.e., genuinely deficient patients. The 4 on energy and the 3s on focus and mood are "for the affected subgroup" calls; for a replete reader these are not benefits B₁₂ supplementation will deliver. The article makes this conditional explicit ("for someone who's actually deficient"); the pitches inherit that framing.

Rating difficulties.

Evidence at 4, not 5. Mechanism is settled enzyme biochemistry, Lindenbaum and Carmel codify the clinical picture, and Vidal-Alaball gives Cochrane-grade evidence for oral non-inferiority. But population-screening trials have not consistently shown net benefit outside the mild-cognitive-impairment subgroup, and the cardiovascular homocysteine-lowering RCTs (HOPE-2, VISP, NORVIT) were null. Held at 4 to honour the genuine asymptomatic-treatment uncertainty.
Longevity at 3. The dominant longevity benefit is preventing irreversible neurological injury in the at-risk subgroup, not population mortality reduction. Adding the homocysteine-Alzheimer's signal (Hooshmand 2010, Smith 2010) gets it to a meaningful-but-not-dominant 3.
Controversy at 2. Core diagnosis and treatment are not contested; what is contested is the optimal lower cutoff (200 vs ~400 pg/mL) and the broader homocysteine-cardiovascular story. Held at 2 rather than 1 to acknowledge that real expert debate exists.

Excluded.

Detailed cyano- vs methyl- vs hydroxocobalamin comparison. Mentioned briefly in protocol; the literature does not support a clinically meaningful difference for the typical patient, and a deeper treatment would dwarf the article's diagnostic-gap thesis.
MTHFR consumer-genetic testing critique. Flagged in out-of-scope for a future entry; cleaner as its own piece than tucked here.
The full homocysteine-and-cardiovascular argument. Flagged in out-of-scope; warrants its own entry given the null trials despite the brain-atrophy signal.
Pediatric / pregnancy presentations of maternal B₁₂ deficiency causing infant neurological injury. Real and serious but distinct enough in audience and intervention that it would dilute this entry; flagged for a future entry on B₁₂ in pregnancy.

Separate-entry candidates. Folate / B₉ deficiency. Pernicious anemia as a standalone autoimmune condition. Homocysteine as a cardiovascular risk marker. MTHFR genetic testing critique. Maternal B₁₂ in pregnancy / breastfeeding.

Future links. When the entries above exist, this article should cross-link from out-of-scope (currently signposted in prose) and from related in meta.

Citation hygiene. Healton 1991 (Medicine journal) was added to the library to support the "neurological recovery scales with treatment delay" claim; the DOI string captured for that entry may need an editor's verification — the Medicine (Baltimore) DOI scheme for 1991 papers is patchy in the wild. The article relies on the claim, not the URL render, so this is a low-priority hygiene flag.