B12 Deficiency Causes: The 9 Real Reasons Your Levels Drop (And How to Tell Which One Is Yours)

Why "Just Eat More Meat" Misses the Point of B12 Deficiency Causes

I’ve read a lot of these lists. Almost all of them stop at two things: vegan diets and pernicious anemia. Then they tell you to eat liver and see your doctor.

That’s not even close to the full picture. Most B12 deficiency causes in developed countries aren’t dietary at all. They’re absorption failures, and absorption of this particular vitamin is absurdly complicated compared to, say, vitamin C. NHANES-based analyses put low or borderline B12 status at roughly 1 in 20 adults under 60, and closer to 1 in 10 in adults over 60. Plenty of those people eat steak.

The two-bucket rule: intake vs. absorption

Every single cause in this article breaks one of two things: how much B12 goes in your mouth, or one of five specific handoffs required to get it from your stomach into your bloodstream. That’s the whole framework. Diet problems go in bucket one. Everything else (autoimmune attack, surgery, drugs, bugs, parasites, pancreatic failure) goes in bucket two.

Here’s what stopped me when I first went through the primary literature: serum B12, the test almost everyone gets, is a mediocre marker. It measures total cobalamin, and about 70 to 80% of that is bound to haptocorrin, a carrier protein your tissues can’t use. You can have a “normal” 350 pg/mL and still be functionally deficient at the cellular level. I’ll come back to this hard, because it’s the reason so many people get told they’re fine when they aren’t.

What I’m covering (and what the other articles skip)

Nine cause categories. A risk checklist. A testing strategy that goes beyond serum B12. Then a correction protocol with actual doses. In this first half I’ll cover the absorption machinery and causes one through six.

How B12 Absorption Actually Works (The 5-Step Chain Behind Most B12 Deficiency Causes)

Think of it as a relay race with five runners. Drop the baton at any handoff and the race is over, no matter how fast the first runner was.

Step 1: Stomach acid and pepsin free B12 from food protein

B12 in food is stuck to animal proteins. Hydrochloric acid and pepsin have to cleave it loose. No acid, no release. This step is why low stomach acid is one of the most underrated B12 deficiency causes on the list.

Step 2: R-binder (haptocorrin) hands it off

Freed B12 immediately binds haptocorrin (the old name is R-binder) from saliva and gastric juice, which protects it through the acidic stomach.

Step 3: Parietal cells make intrinsic factor

Gastric parietal cells secrete intrinsic factor, a glycoprotein that exists for one job: escorting B12 across the ileal wall. Lose your parietal cells and you lose Steps 1 and 3 simultaneously.

Step 4: Pancreatic enzymes complete the transfer

In the duodenum, pancreatic proteases (mainly trypsin) degrade haptocorrin so B12 can transfer to intrinsic factor. Untreated pancreatic insufficiency stalls the baton right here.

Step 5: Cubilin receptors in the ileum pull it in

The IF-B12 complex travels to the terminal ileum and binds cubilin receptors. This receptor-mediated route saturates at roughly 1.5 to 2 mcg per meal. Read that again, because it’s the single most practical fact in this article: you cannot force more than about 2 mcg of food-bound B12 across per sitting. Frequency beats dose size. Three modest servings across the day outperform one enormous one.

The passive absorption backdoor (1-2% of any oral dose)

Roughly 1 to 2% of any oral B12 dose crosses by passive diffusion, no intrinsic factor required. That tiny percentage is why high-dose oral tablets work even in people with zero intrinsic factor. Kuzminski and colleagues, publishing in Blood (1998), gave 2,000 mcg of oral cyanocobalamin daily to patients with cobalamin deficiency (including pernicious anemia) and matched the hematologic and neurologic response of intramuscular injections. Small trial, 33 patients, but it’s been replicated enough that oral megadose therapy is now standard practice in much of Europe.

One more number that explains everything about the timeline: your liver holds 2 to 5 mg of B12, roughly 1,000 days of reserve. That’s why the symptom lag after gastric surgery or starting a vegan diet runs one to four years. Deficiency arrives quietly, then all at once.

Absorption step	What breaks it
1. Acid + pepsin release B12 from food	Atrophic gastritis, H. pylori, PPIs, H2 blockers, gastrectomy, aging
2. Haptocorrin binding	Rarely broken in isolation (congenital haptocorrin deficiency)
3. Intrinsic factor secretion	Pernicious anemia, total/partial gastrectomy, gastric bypass
4. Pancreatic proteases free B12 for IF	Chronic pancreatitis, exocrine pancreatic insufficiency, cystic fibrosis
5. Cubilin receptors in terminal ileum	Crohn’s ileitis, ileal resection, metformin, Imerslund-Grasbeck syndrome, tropical sprue
Lumen competition (before any step)	SIBO, fish tapeworm, blind loop syndrome

Cause 1: You're Genuinely Not Eating Enough (Diet-Based Deficiency)

I’ll be straight: this is the easiest cause to fix and, in the average clinic, one of the least common. But when it happens, it can be severe.

Vegans, vegetarians, and the algae/seaweed myth

Pawlak et al. reviewed the pooled literature in Nutrition Reviews (2013) and the numbers are worse than most people expect: deficiency prevalence reached 62% in pregnant vegans, 25 to 86% across studies of vegan children, and 21 to 41% in vegan adolescents. Plants contain no meaningful B12. Full stop.

And no, spirulina doesn’t count. Neither does chlorella, nori, tempeh, or unfortified nutritional yeast. Spirulina is loaded with pseudo-B12 analogues (mainly pseudocobalamin) that bind human transport proteins without doing any of the enzymatic work. They occupy the seat and refuse to drive. Some evidence suggests they may even compete with real cobalamin. I’ve seen supplement labels advertise spirulina as a B12 source for a decade and it’s still wrong.

The actual numbers

The RDA is 2.4 mcg/day for adults, 2.6 during pregnancy, 2.8 during lactation. Those figures assume normal absorption. For vegans the practical target is 25 to 100 mcg daily from a supplement or fortified foods, or 2,000 mcg once weekly, because the low absorption fraction of large doses means you need to overshoot dramatically.

Older adults eating less meat and dairy

Appetite drops, dentition gets worse, cooking for one gets tedious, meat gets expensive. Intake falls right as absorption is already declining. It’s a double hit, and I’ll cover the absorption half in Cause 3.

Low-income and food-insecure populations

Animal foods are the expensive ones. Food insecurity tracks with low B12 intake, and this group also has higher H. pylori prevalence. Two mechanisms, same person.

Infants of B12-deficient mothers

This is the danger window nobody talks about enough. An exclusively breastfed infant of a B12-deficient vegan or bariatric-surgery mother has almost no liver stores of their own, and breast milk B12 tracks maternal status closely. The case literature describes failure to thrive, hypotonia, developmental regression (losing skills already gained), and irritability at four to eight months. Most respond to treatment. Not all recover completely. If you’re vegan and pregnant or nursing, supplementing isn’t optional.

Lacto-ovo vegetarians aren’t off the hook either. Typical intakes hover around 1 to 1.5 mcg/day, which is below the RDA before you subtract anything for absorption.

Cause 2: Pernicious Anemia (The Autoimmune One)

What antibodies against intrinsic factor and parietal cells do

Pernicious anemia is autoimmune destruction of gastric parietal cells, usually with antibodies against intrinsic factor itself. It takes out Steps 1 and 3 at the same time: no acid to free food B12, no intrinsic factor to carry it. This is the most complete absorption failure on the whole list.

Who gets it and how common it really is

Roughly 0.1% of the general population, rising to 1.9 to 2% of adults over 60. Rates run higher in people of Northern European and African descent, and the African-descent cases tend to present earlier. Mean age at diagnosis sits around 60, but I’ve seen it in patients in their thirties.

The diagnostic pathway

Anti-intrinsic factor antibody is the workhorse: highly specific (a positive result is close to diagnostic) but only about 50 to 70% sensitive, so a negative doesn’t clear you. Anti-parietal cell antibodies are more sensitive, less specific. Then look at elevated serum gastrin and low pepsinogen I, the fingerprint of atrophic body gastritis. The Schilling test, which older articles still describe in loving detail, is basically extinct (radiolabeled B12 isn’t manufactured for it anymore).

The Hashimoto, type 1 diabetes, and vitiligo overlap

Autoimmune conditions travel in packs. Up to 40% of pernicious anemia patients have autoimmune thyroid disease. If you have Hashimoto’s, type 1 diabetes, vitiligo, or Addison’s, low B12 deserves a real workup rather than a shrug.

Why it raises gastric cancer risk

Chronic atrophic gastritis carries a 3 to 7-fold increased risk of gastric adenocarcinoma, plus a distinct risk of type 1 gastric carcinoid tumors driven by chronic hypergastrinemia. That’s the argument for a baseline endoscopy at diagnosis. I think it’s underappreciated in primary care.

One thing to be blunt about: this cause is permanent. Lifelong repletion, not a three-month course. I’ve watched people finish a course of injections, feel great, stop, and come back two years later with numbness in their feet.

Cause 3: Low Stomach Acid and Atrophic Gastritis (One of the Most Missed B12 Deficiency Causes)

Food-bound cobalamin malabsorption explained

Here’s the distinction almost every competitor article blurs. In food-bound cobalamin malabsorption you still have intrinsic factor. What you’ve lost is acid and pepsin, so B12 stays welded to food protein and never enters the relay. Crystalline B12 (the free form in supplements and fortified cereals) absorbs perfectly fine because it skips Step 1 entirely.

That single mechanistic difference changes treatment completely. Pernicious anemia needs high-dose or parenteral B12 forever. Food-bound malabsorption is often fully corrected by a modest daily supplement.

Why aging does this

Atrophic gastritis affects an estimated 20 to 50% of adults over 60 to some degree. Framingham Offspring Study data showed the pattern clearly: the worst B12 status showed up in older adults who had both low intake and impaired absorption, and plasma B12 correlated better with supplement and fortified cereal intake than with meat intake. Which is exactly why the Institute of Medicine recommends adults over 50 get their B12 from supplements or fortified foods rather than relying on meat.

H. pylori infection as an underrated driver

H. pylori causes the chronic gastritis that becomes atrophy. It’s the slow arsonist. Eradication therapy has been shown to raise B12 levels in deficient patients with active infection, sometimes enough to normalize them without any supplementation at all. If someone has unexplained low B12, dyspepsia, and low ferritin, I want an H. pylori test before I want anything else.

Signs your acid is the problem

• Bloating or heaviness within 30 minutes of a protein-heavy meal
• Low ferritin alongside low B12 (iron absorption is also acid-dependent, so they fail together)
• Early satiety, feeling full after a few bites
• Undigested food visible in stool
• Reflux symptoms that paradoxically got worse, not better, with age

Cause 4: Medications That Quietly Drain B12

This is the section I’d most want a reader to send to their parents.

Metformin: the dose- and duration-dependent effect

The Diabetes Prevention Program Outcomes Study, published in JCEM (2016), followed participants for over a decade. B12 deficiency at year 5: 4.3% on metformin vs 2.3% on placebo. By year 13: 7.4% vs 5.4%. Low-normal B12 was substantially more common still. Absolute risk isn’t enormous, but it compounds, and UK NICE guidance now recommends periodic B12 testing in metformin users with symptoms suggesting deficiency.

The mechanism is specific: metformin interferes with the calcium-dependent binding of the IF-B12 complex to cubilin receptors in the ileum. It breaks Step 5. Risk scales past roughly 1,000 mg/day and past three to four years of use. Here’s the clinical trap that annoys me most: metformin-induced B12 deficiency causes peripheral neuropathy, and so does diabetes. Guess which one gets blamed?

PPIs and H2 blockers

Lam and colleagues published a case-control analysis in JAMA (2013) covering 25,956 patients with diagnosed B12 deficiency against nearly 185,000 controls. Two or more years of PPI use carried an odds ratio of 1.65 for deficiency. H2 blockers came in at 1.25. Dose mattered too, with more than 1.5 PPI pills daily showing stronger association. Mechanism is Step 1 again: suppress acid, fail to liberate food-bound B12.

I’m not anti-PPI. They’re excellent drugs for the right indication. I am against the ten-year prescription nobody has revisited since 2016.

Nitrous oxide: the fastest-acting B12 destroyer

This one gets its own subsection because it’s genuinely dangerous and badly underreported. Nitrous oxide irreversibly oxidizes the cobalt core of cobalamin from Co(I) to Co(III), which inactivates methionine synthase directly. It doesn’t reduce your B12 level. It disables the B12 you already have.

Recreational “whippet” and canister use has produced subacute combined degeneration of the spinal cord in weeks. Not years. Weeks. Young patients presenting with ascending numbness, unsteady gait, and loss of proprioception, sometimes with a perfectly normal serum B12 (because the molecule is present, just dead). Homocysteine and methylmalonic acid are the tests that reveal it.

Surgical and dental nitrous oxide is safe for the general population, but prolonged anesthetic exposure in someone already borderline deficient has triggered neurologic decompensation. Worth knowing before a long procedure if you have any risk factors from this article.

The rest of the list

Colchicine (damages ileal mucosa), cholestyramine (binds B12 in the gut), chloramphenicol (marrow suppression blunting the response to B12), aminosalicylates, anticonvulsants like phenytoin and carbamazepine (mixed data, mostly folate but B12 too), and oral contraceptives, which lower serum B12 measurably though the functional significance is debated. I’d call the contraceptive effect a minor concern in isolation and a real one if you’re stacking risk factors.

Excess alcohol

Four mechanisms at once: worse diet, gastritis wrecking acid production, impaired hepatic storage and release, and disrupted transport. Chronic heavy drinkers often have deficiency across the whole B-vitamin family.

Drug	Step broken	Typical time to deficiency	Monitoring
Metformin	Step 5 (ileal receptor binding)	3-4+ years, dose-dependent	Serum B12 every 1-2 years; MMA if symptoms
PPIs (omeprazole etc.)	Step 1 (acid release)	2+ years	B12 at 2 years, then periodically
H2 blockers	Step 1 (partial)	2+ years	Lower priority, check if symptomatic
Nitrous oxide	Inactivates cobalamin itself	Days to weeks with heavy use	MMA + homocysteine, not serum B12
Colchicine (long-term)	Step 5 (ileal mucosa)	Months to years	B12 annually on chronic therapy
Cholestyramine	Binds B12 in lumen	Months	Separate dosing; annual B12
Chronic alcohol excess	Steps 1, 4 + storage	Years	B12, folate, thiamine together

Cause 5: Stomach and Intestinal Surgery

Gastric bypass and sleeve gastrectomy

After Roux-en-Y gastric bypass, reported B12 deficiency runs 12 to 33% in patients who aren’t supplemented properly. You’ve bypassed most of the acid-producing stomach and reduced intrinsic factor output. Sleeve gastrectomy is lower risk but not zero, because you’re still removing a large chunk of parietal cell mass.

Total and partial gastrectomy

Total gastrectomy guarantees deficiency. No stomach, no intrinsic factor, no receptor-mediated absorption at all. Repletion is lifelong, either parenteral or very high-dose oral relying on that 1 to 2% passive route.

Ileal resection and the 60 cm threshold

B12 absorption happens almost exclusively in the terminal 60 to 80 cm of ileum. Resections shorter than that are often tolerated. Go beyond it (common in Crohn’s surgery or after complicated adhesions) and deficiency is reliable rather than possible.

The ASMBS recommends 350 to 500 mcg oral daily, or 1,000 mcg intramuscular monthly, after bariatric surgery, with annual monitoring.

Now the part that matters most. Post-op patients feel fine for one to three years while those liver stores drain. Adherence to supplements and follow-up labs falls off a cliff around year two, which is precisely when it starts to count. I’ve seen more than one bariatric patient present with neuropathy at year four, having stopped their multivitamin at month eighteen because they felt great.

Cause 6: Gut Diseases and Competition for Your B12

Crohn’s disease and terminal ileum involvement

Location is everything. B12 deficiency shows up in roughly 20 to 40% of patients with ileal involvement or prior ileal resection, and at close to background rates in isolated colonic disease. If you have Crohn’s, the question isn’t “do I have Crohn’s,” it’s “where is my disease and how much ileum do I still have?”

Celiac disease (including seronegative and newly diagnosed)

Estimates put B12 deficiency at 8 to 41% at celiac diagnosis, which puzzled people for years because celiac damages the duodenum and jejunum, not the ileum. The likely explanations: bacterial overgrowth secondary to dysmotility, overlapping autoimmune gastritis (celiac patients get more of it), reduced intake, and in severe cases diffuse enteropathy reaching further distally. Most correct on a gluten-free diet alone, but not all.

SIBO and bacterial overgrowth stealing your B12

The mechanism here is beautifully simple and genuinely underdiagnosed. Anaerobic bacteria in the small intestine consume free cobalamin and can cleave the IF-B12 complex before it ever reaches the ileum. They eat your lunch, literally.

The giveaway lab pattern: low B12 with normal or high folate. Those same bacteria synthesize folate while destroying B12. If you see that combination alongside bloating and altered bowel habits, think overgrowth, blind loop, diverticula, or a stricture.

Chronic pancreatitis and exocrine pancreatic insufficiency

Step 4 failure. Without trypsin and other proteases, haptocorrin doesn’t get degraded and B12 never transfers to intrinsic factor. Enzyme replacement therapy fixes it, which is one of the more satisfying causes to treat.

Fish tapeworm (Diphyllobothrium latum)

Rare in most places, real in populations eating raw or undercooked freshwater fish. The worm absorbs B12 directly from the intestinal lumen with impressive efficiency, and heavy infestation can produce frank megaloblastic anemia. Sushi from marine fish isn’t the usual culprit; freshwater pike, perch, and salmon are.

Tropical sprue and Whipple disease

Both damage the small bowel diffusely, both cause B12 malabsorption, both are uncommon but worth remembering in a traveler or someone with unexplained malabsorption plus arthralgias and weight loss.

Pattern you see	Most likely gut cause
Low B12 + high folate + bloating	SIBO or blind loop
Low B12 + low iron + low vitamin D + diarrhea	Celiac disease
Low B12 + prior ileal surgery or ileal inflammation on imaging	Crohn’s
Low B12 + steatorrhea + fat-soluble vitamin deficiencies	Exocrine pancreatic insufficiency
Low B12 + eosinophilia + raw freshwater fish history	Fish tapeworm
Low B12 + high gastrin + positive anti-IF antibody	Pernicious anemia (not a gut cause, but it lives in this differential)

Six causes down. The remaining three are the ones that catch people who look low-risk on paper: genetics, kidney and liver disease, and the pregnancy-plus-thyroid group. That’s where the testing strategy starts to matter, because these are the patients whose serum B12 comes back “normal.”

Cause 7: Genetic and Inherited B12 Deficiency Causes

These are rare. I’m including them anyway, because the ones that get missed get missed for a decade, and the cost of that miss is measured in IQ points and permanent gait problems.

Imerslund-Gräsbeck syndrome (CUBN and AMN mutations)

The cubam receptor in the terminal ileum has two subunits: cubilin (CUBN) and amnionless (AMN). Break either one and the intrinsic factor to B12 complex arrives at the ileum and simply bounces off. Classic presentation is megaloblastic anemia in a child between one and five years old, plus benign proteinuria (cubilin also handles protein reabsorption in the proximal tubule). That protein in the urine is the giveaway. A kid with anemia and unexplained proteinuria who’s been worked up by nephrology for years is the textbook missed case.

Congenital intrinsic factor deficiency (GIF mutations)

Same clinical picture as pernicious anemia, minus the autoimmunity, minus the atrophic gastritis, and about fifty years early. Gastric acid production is normal. Antibodies are negative. Kids present at one to five years with pallor, failure to thrive, and a hemoglobin that makes the lab call twice.

Transcobalamin II deficiency

This is the one that will fool you. Transcobalamin II is the protein that actually delivers B12 to cells. When it’s defective, the cobalamin sits in the bloodstream bound to haptocorrin, so serum B12 reads normal or even high while the cells starve. Infants present in the first few months with severe pancytopenia, diarrhea, and immunodeficiency. MMA and homocysteine are the only markers that tell the truth here. If you take one thing from this section: a normal serum B12 does not rule out cellular B12 deficiency, and this disorder is the proof of concept.

Intracellular processing defects (cblA through cblG, MMACHC)

Once B12 gets inside a cell, it has to be stripped of its ligand, reduced, and converted into methylcobalamin (for the cytosol) or adenosylcobalamin (for the mitochondria). Roughly a dozen genes handle those steps. MMACHC mutations cause cblC disease, the most common of the group, and it produces isolated methylmalonic aciduria with homocystinuria, developmental delay, and retinopathy. Newborn screening catches many of these now. Not all.

The MTHFR question, and what it doesn’t do

Let me be blunt, because the internet has made a mess of this. MTHFR polymorphisms (C677T, A1298C) affect folate cycling. They do not cause B12 malabsorption. The enzyme converts 5,10-methylenetetrahydrofolate to 5-methyltetrahydrofolate. It sits downstream of everything I’ve described in the absorption chain. A homozygous C677T genotype can nudge homocysteine up a bit, especially with low folate intake, but it does not stop your stomach from making intrinsic factor and it does not damage your ileum.

I’ve watched patients spend $400 on methylated supplement stacks because a direct-to-consumer genetic report flagged MTHFR, while nobody ever ordered an anti-intrinsic-factor antibody. That’s the actual harm: MTHFR becomes the explanation, and the real cause goes unfound for years.

FUT2 and TCN1 variants that shift the number without shifting your status

Here’s the underdiscussed part. Genome-wide association studies have consistently identified variants in FUT2 (the secretor gene, which also governs H. pylori susceptibility and gut mucosal glycosylation) and TCN1 (haptocorrin) as significant determinants of measured serum B12. The effect sizes aren’t trivial. Certain TCN1 variants lower circulating haptocorrin, and since haptocorrin carries 70 to 80% of the B12 in your blood, those people walk around with a serum B12 that looks borderline or frankly low while their functional status is completely fine. The reverse also happens.

Translation: some people get falsely reassured, some get falsely alarmed, and serum B12 alone can’t tell them apart. MMA can.

When genetic testing is actually worth ordering

Three situations, in my view. Unexplained deficiency in a child. Family clustering across siblings or generations without a dietary or autoimmune explanation. And persistent functional deficiency (high MMA) with documented normal absorption and negative antibodies in an adult. Outside those, you’re paying for a report you’ll misinterpret.

---

Cause 8: Higher Demand, Pregnancy, and Life-Stage Drains

Not every deficiency is a broken pipe. Sometimes the pipe is fine and the tap is just running harder than the supply can keep up with.

Pregnancy and lactation

Serum B12 drops physiologically across gestation, roughly 30 to 40% by the third trimester, driven by hemodilution, active transfer to the fetus, and shifts in binding protein concentrations. Some of that is a measurement artifact. Some of it isn’t.

The data that changed how I think about this: Molloy and colleagues, publishing in Pediatrics in 2009, examined maternal B12 status in an Irish cohort and found women in the lowest B12 quartile had roughly 2.5 to 3 times the risk of a neural tube defect-affected pregnancy compared with the highest quartile, independent of folate status. Folate fortification prevents most NTDs. B12 appears to be the second lever, and it’s the one nobody pulls.

Infancy and rapid growth

Breast milk B12 tracks maternal status almost directly. A vegan mother with unrecognized deficiency, or a mother with undiagnosed pernicious anemia, produces milk that looks fine and contains almost nothing. The infant runs on placental stores for a few months, then falls off a cliff.

The presentation is heartbreaking and specific: hypotonia, irritability, feeding refusal, and developmental regression at four to eight months, sometimes with a peculiar tremor once treatment starts. Babies who were rolling over stop rolling over. Early treatment reverses most of it. Late treatment doesn’t reverse all of it.

Chronic hemolysis and hyperproliferative marrow states

Sickle cell disease, thalassemia, autoimmune hemolytic anemia, polycythemia vera, myelofibrosis. Anywhere the marrow is churning out cells at three to five times normal rate, DNA synthesis demand climbs, and B12 turnover climbs with it. These patients also tend to have elevated haptocorrin from the expanded myeloid mass, which inflates serum B12 and hides the problem. Again: measure MMA.

Dialysis and chronic kidney disease

B12 is water-soluble and modestly protein-bound, so hemodialysis strips some of it out every session. Add altered transcobalamin binding, poor appetite, and dietary protein restriction, and standard supplement doses often undershoot. The complication: MMA is renally cleared, so it’s elevated in CKD regardless of B12 status. In a dialysis patient, holotranscobalamin is the more interpretable marker.

Practical point across all four: these aren’t absorption failures. The plumbing works. You just need a bigger dose, and once demand drops (post-partum, post-weaning, hemolysis controlled), requirements usually fall back to baseline.

---

Cause 9: Liver Disease, Kidney Disease, and Thyroid Dysfunction

I’m grouping these because they share one property: they distort the relationship between the number on your lab report and what’s happening inside your cells.

Liver disease. Hepatocytes store roughly half your body’s B12 reserve. Damage them (alcoholic hepatitis, cirrhosis, hepatocellular carcinoma, acute hepatitis) and stored cobalamin spills into the circulation while haptocorrin production rises. Result: a serum B12 of 900 pg/mL in a patient who is functionally depleted. Alcohol compounds it further by damaging the ileal mucosa and worsening intake.

Kidney disease. Covered above, but worth restating: elevated MMA in a patient with an eGFR under 45 means less than you think.

Thyroid. Autoimmune thyroid disease and pernicious anemia travel together. Both belong to the autoimmune polyglandular cluster, and reported prevalence of pernicious anemia in Hashimoto’s patients runs around 10 to 12%, far above background. If you have Hashimoto’s or Graves’ disease and persistent fatigue that levothyroxine didn’t fix, checking B12 status is one of the higher-yield tests available. Hyperthyroidism separately increases metabolic turnover.

---

Why B12 Deficiency Causes Can Hide for Years Before You Feel Sick

The 1,000-day liver reserve

Your liver holds 2 to 5 mg of B12. You lose about 1 to 2 micrograms a day through bile and shedding, and a chunk of that gets recycled through enterohepatic circulation. Do the arithmetic and a person who abruptly stops absorbing B12 today has somewhere between two and five years of buffer.

That buffer is why the dietary story is so misleading, and it’s why the person who had gastric bypass in 2019 shows up with numb feet in 2024 and nobody connects the two.

Stage 1 to Stage 4

The depletion model, in order:

1. Stage 1: negative balance. Holotranscobalamin falls. Total serum B12 is still normal. No symptoms.
2. Stage 2: depletion. Total serum B12 starts dropping into the gray zone. Still often asymptomatic.
3. Stage 3: functional deficiency. MMA and homocysteine rise. Cells are now short. Fatigue, paresthesia, and cognitive fog often begin here, with a completely normal blood count.
4. Stage 4: clinical deficiency. Macrocytosis, anemia, glossitis, and in the worst cases subacute combined degeneration of the spinal cord.

Most articles treat Stage 4 as “B12 deficiency” and everything before it as noise. That’s backwards. Stage 3 is where treatment is cheap and reversal is complete.

Neurological damage that outruns the anemia

This is the single most important clinical fact in the article, so I’ll state it as plainly as I can: in roughly 25 to 30% of patients with neurological manifestations of B12 deficiency, the MCV is normal and there is no anemia. Lindenbaum and colleagues documented this in the New England Journal of Medicine back in 1988, examining 141 consecutive patients with neuropsychiatric abnormalities from cobalamin deficiency, and 40 of them had no anemia or macrocytosis at all.

Waiting for a big MCV before you take B12 seriously is a mistake that costs nerves.

Symptoms, roughly in order of appearance

• Fatigue and exertional breathlessness
• Glossitis (smooth, sore, beefy red tongue), angular cheilitis, mouth ulcers
• Symmetrical paresthesia in hands and feet, usually feet first, often worse at night
• Memory lapses, word-finding trouble, “brain fog”
• Loss of vibration and position sense, then gait instability and a positive Romberg
• Irritability, depression, and in advanced cases psychosis (the old term was “megaloblastic madness”)
• Subacute combined degeneration: dorsal column and corticospinal tract demyelination

Suggested graphic: a horizontal timeline showing months 0 to 36, mapping symptom onset against declining holoTC, falling serum B12, rising MMA, and the late appearance of macrocytosis. ``

How reversible is it, honestly?

Depends almost entirely on duration. Symptoms present less than six months typically resolve fully with adequate repletion. Symptoms present more than twelve months frequently leave residue: persistent numbness, ongoing balance problems, incomplete cognitive recovery. Follow-up data from Healton et al. (Medicine, 1991) on 143 episodes of neuropsychiatric deficiency found severity at presentation was the strongest predictor of incomplete recovery.

That asymmetry (cheap to prevent, expensive to fix) is the entire argument for testing early.

---

Are You Actually at Risk? A 12-Point Self-Check

The checklist

Tick every box that applies to you:

1. ☐ Age 60 or older
2. ☐ Vegan or vegetarian for more than 3 years
3. ☐ Taking metformin (any dose, more than 12 months)
4. ☐ Taking a PPI or H2 blocker for more than 12 months
5. ☐ Any gastric or intestinal surgery (bypass, sleeve, gastrectomy, ileal resection)
6. ☐ An autoimmune condition (Hashimoto’s, type 1 diabetes, vitiligo, Addison’s)
7. ☐ Crohn’s disease, celiac disease, or chronic diarrhea
8. ☐ More than 14 alcoholic drinks a week
9. ☐ Pregnant, breastfeeding, or trying to conceive
10. ☐ Any recreational nitrous oxide use (whippets, balloons), even occasionally
11. ☐ Unexplained numbness, tingling, balance problems, or memory changes
12. ☐ A first-degree relative with pernicious anemia

How to score it

1 box ticked: get serum B12 tested at your next blood draw. Cheap, sensible.

2 or more boxes: don’t settle for serum B12 alone. Ask for MMA or holotranscobalamin alongside it, because serum B12 is exactly the test most likely to mislead you in this group.

Box 10 or 11 ticked, regardless of the others: test now, and treat on suspicion if symptoms are neurological. Nitrous oxide can produce myelopathy with a serum B12 in the normal range.

---

Testing: Why Serum B12 Fails and What to Order Instead

The gray zone (148 to 258 pmol/L, or 200 to 350 pg/mL)

Serum B12 measures total cobalamin, and 70 to 80% of that total is bound to haptocorrin, which delivers it precisely nowhere. Only the 20 to 30% carried on transcobalamin II reaches your cells. So the test is measuring mostly cargo that’s parked in the wrong truck.

Reported sensitivity of serum B12 for functional deficiency sits around 50 to 70% depending on the cutoff and population. In the gray zone specifically, roughly half of people with a “low-normal” result have metabolic evidence of deficiency and half don’t. A coin flip with a lab fee attached.

Methylmalonic acid: the functional gold standard

Without adenosylcobalamin, methylmalonyl-CoA can’t convert to succinyl-CoA, and MMA accumulates. It’s the most specific functional marker available. Cutoffs vary by lab, but elevation above roughly 0.27 to 0.37 µmol/L is the usual threshold. MMA rises before anemia, before macrocytosis, and often before serum B12 leaves the normal range.

Two confounders: renal impairment (MMA clears through the kidneys) and small intestinal bacterial overgrowth (gut bacteria produce propionate, which feeds the same pathway). Interpret MMA with an eGFR in front of you.

Holotranscobalamin (active B12)

This measures the fraction that actually gets delivered. It’s the first marker to fall in Stage 1, before total B12 budges. Practical cutoff is around 35 pmol/L, with a gray band between roughly 25 and 45. Availability varies wildly by country; it’s routine in parts of Europe and hard to find in the US.

Combine holoTC with MMA and you’re close to a definitive answer. That’s the pairing I’d order if I could only order two things.

Homocysteine and its confounders

Homocysteine rises in B12 deficiency because methionine synthase stalls. It’s sensitive but not specific: folate deficiency, B6 deficiency, hypothyroidism, renal disease, and smoking all push it up. Useful as supporting evidence. Useless in isolation.

False normals and false lows

Falsely high or normal serum B12 despite real deficiency: - Liver disease (stored B12 leaking out, plus raised haptocorrin) - Myeloproliferative disorders and chronic myeloid leukemia - Recent supplementation, even a single multivitamin in the past week - Transcobalamin II deficiency - Certain TCN1 variants raising haptocorrin

Falsely low serum B12 without functional deficiency: - Pregnancy (that physiologic 30 to 40% drop) - Combined oral contraceptives - Folate deficiency - Multiple myeloma - TCN1 variants that lower haptocorrin

A decision tree you can actually use

1. Serum B12 below 148 pmol/L (200 pg/mL) with symptoms: treat, then investigate the cause.
2. Serum B12 in the gray zone: order MMA plus holoTC.
3. MMA elevated (with normal kidney function) or holoTC low: confirmed functional deficiency. Now find the cause.
4. Cause unclear and diet adequate: anti-intrinsic-factor antibody (highly specific, only ~50 to 70% sensitive) and anti-parietal-cell antibody (sensitive, less specific), plus fasting gastrin and pepsinogen I.
5. Antibodies or gastrin suggest autoimmune atrophic gastritis: upper endoscopy with biopsy, both to confirm and because these patients carry elevated gastric cancer and carcinoid risk that needs surveillance.
6. Absorption issue with normal stomach workup: think ileum, think drugs, think SIBO.

---

Fixing the Cause, Not Just the Number

Here’s my core argument, and it’s why this article is 4,000 words instead of 800. If you inject B12 without identifying why it dropped, three things can happen: you treat forever unnecessarily, you miss a drug that should have been changed, or you miss autoimmune atrophic gastritis and the gastric cancer surveillance that comes with it. Repletion is easy. Diagnosis is the part with value.

Match the treatment to the cause

Cause	Route and dose	Duration
Dietary (vegan/vegetarian)	50 to 250 mcg oral daily, or 2,000 mcg weekly	As long as the diet continues
Metformin-associated	1,000 mcg oral daily, recheck in 3 months	While on metformin
PPI-associated	1,000 mcg oral daily; reassess PPI necessity	While on the drug
Pernicious anemia	1,000 mcg IM loading, then 1,000 mcg IM monthly or 1,000 to 2,000 mcg oral daily	Lifelong
Post-gastrectomy / bypass	IM preferred, 1,000 mcg monthly	Lifelong
Ileal resection >60 cm	IM, 1,000 mcg monthly	Lifelong
SIBO / blind loop	Treat the overgrowth, then reassess	Often temporary
Nitrous oxide toxicity	Hydroxocobalamin IM plus stop exposure	Until recovery
Pregnancy demand	250 to 1,000 mcg oral daily	Through lactation

Oral high-dose vs. intramuscular: what the evidence says

I used to assume injections were categorically superior. The data made me revise that.

About 1 to 2% of an oral dose crosses the intestine by passive diffusion, completely bypassing intrinsic factor and the ileal receptor. That sounds negligible until you multiply it: 1% of 2,000 mcg is 20 mcg a day, roughly eight times the RDA, in someone with zero intrinsic factor.

Kuzminski and colleagues tested this directly in 1998 (Blood), randomizing patients with cobalamin deficiency (including pernicious anemia) to 2,000 mcg oral daily or an IM regimen. At four months, the oral group had higher serum B12 and lower MMA than the injection group. The Cochrane review by Wang et al. (2018) pooled the available randomized trials and concluded oral and intramuscular routes produced comparable normalization of serum B12 and similar hematologic responses.

So when would I still choose injections? Four situations:

• Acute or progressive neurological presentation (I don’t gamble with cord demyelination)
• Severe malabsorption or total gastrectomy
• Documented adherence problems, because a monthly injection someone actually receives beats a daily tablet they don’t take
• The first few weeks of severe deficiency, then switch to oral

Which form? Cyanocobalamin, hydroxocobalamin, methylcobalamin, adenosylcobalamin

Cyanocobalamin is cheapest, most stable, and best studied. It’s what nearly all the trial evidence used. The cyanide moiety is a rounding error (micrograms), irrelevant outside of tobacco amblyopia and Leber’s hereditary optic neuropathy, where it’s genuinely contraindicated.

Hydroxocobalamin has longer tissue retention and better protein binding, which is why UK practice uses it and why maintenance can stretch to every two or three months. It’s also the form used for cyanide poisoning and, in my view, the better choice for nitrous oxide toxicity.

Methylcobalamin is the marketing favorite. I’ll say this directly: the retention data supporting it as superior is weaker than the supplement aisle implies, and it’s less stable in light. It works fine. It isn’t magic, and paying triple for it isn’t justified by evidence.

Adenosylcobalamin is sold as the “mitochondrial” form. Your cells interconvert these anyway. Skip the combination products charging a premium.

Sublingual and nasal: worth it?

Sublingual tablets absorb essentially the same as swallowed tablets at equivalent doses, because both routes are dominated by passive diffusion. Comparative studies have repeatedly failed to show an advantage. If you prefer the taste or you have swallowing difficulty, fine. Don’t pay extra expecting better absorption. Nasal spray (cyanocobalamin gel) is a legitimate option for maintenance in people who hate needles and won’t take tablets, but it costs substantially more per month.

Loading and maintenance

• Severe deficiency with neurological signs: 1,000 mcg IM every other day for one to two weeks, then weekly for a month, then monthly.
• Deficiency without neurological signs: 1,000 mcg IM three times weekly for two weeks, or 1,000 to 2,000 mcg oral daily from the start.
• Maintenance for permanent causes: 1,000 mcg IM monthly, or 1,000 to 2,000 mcg oral daily.
• Dietary causes: 50 to 250 mcg daily is plenty. Fortified foods count.

The potassium and iron drop nobody warns you about

Two things happen during rapid hematologic recovery, and both catch people off guard.

Hypokalemia. New red cells pull potassium into themselves at speed. In severe megaloblastic anemia, this can drop serum potassium into arrhythmia territory within the first 48 to 72 hours of treatment. Severe cases warrant a potassium check early in repletion.

Functional iron deficiency. The marrow suddenly starts building hemoglobin at full tilt and burns through iron stores. Ferritin can crater in two weeks. If someone’s fatigue improves and then stalls at week three, check iron.

The reticulocyte response peaks at day 5 to 7. If it doesn’t happen, either the diagnosis is wrong or something else (iron, folate, thyroid, marrow disease) is also going on.

What to retest, and when

Retest MMA and homocysteine at 4 to 8 weeks. Not serum B12. Serum B12 after supplementation will look spectacular and tell you exactly nothing about whether your cells are supplied. This is the most common follow-up error I see.

---

Five Mistakes I See Constantly

1. Taking folic acid that masks the anemia while nerves degenerate. High-dose folate lets thymidine synthesis proceed through an alternate route, so the megaloblastic anemia corrects and the MCV normalizes. The methylmalonyl pathway damage continues untouched. You’ve deleted the warning light while the engine seizes. Fix: never treat a macrocytic anemia with folate before checking B12.

2. Stopping treatment once the number normalizes. Pernicious anemia, gastrectomy, ileal resection, and congenital defects are permanent. The absorption problem doesn’t heal because your lab value improved. Stopping is the single most common reason people relapse two years later, usually with worse symptoms than the first time. Fix: if the cause is permanent, the treatment is permanent.

3. Trusting a “normal” serum B12 in someone with textbook symptoms. I’ll say it one more time, as loudly as I can: a normal serum B12 does not exclude B12 deficiency, and a normal MCV does not exclude B12-related nerve damage. Roughly a quarter to a third of neurological cases have neither anemia nor macrocytosis. Fix: order MMA.

4. Blaming MTHFR instead of finding the real cause. A genetic report flagged a folate-cycle variant and now nobody’s looking for antibodies, gastrin, or a drug culprit. MTHFR does not cause malabsorption. Fix: work the absorption chain, not the SNP panel.

5. Ignoring nitrous oxide exposure. Young adults with subacute myelopathy and a “normal” B12 get worked up for MS for months. Nobody asks about balloons at parties or the boxes of chargers in the recycling. Nitrous irreversibly oxidizes the cobalt core of cobalamin and inactivates methionine synthase, so the vitamin is present but useless. Fix: ask the question, and treat with hydroxocobalamin without waiting for the level.

Bonus mistake: taking a B-complex with 6 mcg of B12 when your absorption is broken and you need 1,000. Dose matters more than the label’s promises.

---

The Short Version

• Nine cause categories: dietary intake, autoimmune pernicious anemia, age-related atrophic gastritis and H. pylori, drugs (metformin, PPIs, nitrous oxide), surgical anatomy changes, gut disease (celiac, Crohn’s, SIBO, pancreatic insufficiency), genetic defects, increased demand states, and organ disease that distorts the numbers.
• Absorption beats intake. Outside of vegans and breastfed infants, low B12 is almost always a broken absorption chain, not a food problem.
• Serum B12 alone is not enough. It measures mostly haptocorrin-bound cobalamin your cells can’t use, and its sensitivity in the gray zone is roughly a coin flip.
• MMA and holotranscobalamin settle the question. Order them together when the picture is unclear, and retest MMA (not serum B12) at 4 to 8 weeks.
• Treatment length depends entirely on the cause. Pernicious anemia and surgical causes are lifelong. Drug- and diet-related causes resolve when the input changes.
• Don’t wait for anemia. Nerve damage frequently arrives first, and after a year it may not fully leave.

One closing opinion, and I’ll be direct about it. If you’re over 60, taking metformin, on a PPI, or eating plant-based, get tested once. Serum B12 plus MMA runs well under $100 in most places, often far less. The downside of skipping it isn’t a bad week of fatigue. It’s a spinal cord that doesn’t fully recover. That’s a lopsided bet, and I’d take the test every time.

Frequently Asked Questions