Key takeaways
- Homocysteine is a functional readout of the methylation cycle — it accumulates only when a clearance pathway is short of what it needs.
- Common causes are B12, folate or B6 shortfall, MTHFR variants, impaired kidney function, hypothyroidism, long-term metformin and alcohol.
- MTHFR variants are carried by roughly a third of people; the practical consequence is to use methylated forms, not to treat the genotype.
- B vitamins lower the number reliably, but randomised trials did not show the cardiovascular benefit the association predicted — treat it as a prompt to investigate.
Homocysteine is one of the few cardiovascular markers that is cheap to measure, straightforward to interpret, usually easy to lower — and genuinely contested as to whether lowering it helps. All four of those things are true simultaneously, and holding them together is the whole skill of using this test well. It tells you something real about how a person's biochemistry is running. What it does not tell you is that fixing the number fixes the risk.
What homocysteine actually is
Homocysteine is a sulfur-containing amino acid, and unlike most amino acids it is not a building block. It is an intermediate — a transient product of methionine metabolism that the body is supposed to clear promptly through one of two routes:
- Re-methylation back to methionine, which requires methylated folate as the donor and B12 as the enzyme's cofactor
- Trans-sulfuration onward to cysteine, which requires B6
Both exits have to be open for clearance to keep pace. That is why homocysteine is a useful marker at all: it accumulates only when something in the machinery is short of what it needs, so its level is a functional readout of a pathway rather than a measurement of a substance the body wants.
Why the pathway matters beyond the arteries
The re-methylation route is not a disposal system that happens to exist. It is the regeneration step of the methylation cycle, which produces S-adenosylmethionine, the body's principal methyl donor. Methylation is used in DNA synthesis and repair, in gene expression, in the synthesis of neurotransmitters, in phospholipid production and in the processing of hormones and drugs for excretion.
So an elevated homocysteine is best read as a supply problem in a system with many customers, not as a poison accumulating in the blood. That reframing matters for what follows, because it explains why the associations run so wide — cardiovascular disease, stroke, cognitive decline, pregnancy complications — and also why lowering the number is not automatically the same as fixing what caused it.
What raises it
- Deficiency of B12, folate or B6 — the commonest cause by a distance
- Genetic variants that slow the enzymes involved, principally MTHFR
- Reduced kidney function, since the kidney is a major site of clearance
- Hypothyroidism
- Some medications, including methotrexate, long-term metformin (through B12 depletion) and certain anticonvulsants
- Heavy alcohol use, which impairs folate handling
- Smoking
- Age, and chronic inflammation
The ordering there is deliberate. A high homocysteine with an untested kidney function or an untested thyroid is a half-finished workup, and long-term metformin is the single most frequently missed contributor in an otherwise healthy metabolic patient.
Reading the number
| Level (µmol/L) | Interpretation | What it usually calls for |
|---|---|---|
| < 7 | Optimal | Nothing |
| 7-10 | Within the standard reference range | Context — retest if other risk is high |
| 10-13 | Borderline | Check B12, folate, thyroid and kidney function |
| > 13 | Elevated | Full cofactor workup and a look for a cause |
| > 20 | Marked | Investigate properly; rarely a simple dietary shortfall |
Note the gap between "normal" and "optimal" here, which is wider than for most markers and is a good illustration of why a laboratory reference range is not a target. A result of 12 will be reported without comment on most panels. Note too that homocysteine is not on most standard panels at all — it usually has to be requested specifically, which is why elevation so often goes years without being seen.
MTHFR, and how much weight it deserves
The MTHFR gene encodes the enzyme that produces the methylated form of folate. Two common variants reduce its activity: heterozygous C677T by roughly 30%, homozygous C677T by roughly 70% (Frosst et al., Nat Genet 1995), and roughly 30-40% of the population carries at least one.
That prevalence is the point. A variant carried by a third of everyone is a common polymorphism, not a diagnosis, and it has acquired a following out of proportion to what it does. Its practical implication is narrow and useful: if the enzyme producing methylfolate is slow, supply the methylated form and bypass the step. That is the entire clinical consequence for most carriers. Someone with a known variant and a homocysteine of 6 has nothing to act on, and treating the genotype rather than the phenotype is how people end up on supplement regimens that address a number they never had.
The part usually left out
B vitamins lower homocysteine reliably and dose-dependently (Homocysteine Lowering Trialists, Am J Clin Nutr 2005), and homocysteine is consistently associated with vascular disease across observational data (Boushey et al., JAMA 1995). The obvious conclusion is that lowering it should reduce events.
Large randomised trials tested exactly that, and the result was not what the association predicted: homocysteine came down and cardiovascular event rates largely did not follow, with a possible exception around stroke (homocysteine-lowering interventions for preventing cardiovascular events, PubMed). The honest interpretation is that homocysteine is at least partly a marker of a state rather than a cause of the damage — an indicator that something upstream, whether B-vitamin status, kidney function or general metabolic health, is not right.
The cognitive evidence is more interesting. In older adults with mild cognitive impairment and raised homocysteine, B-vitamin treatment slowed the rate of brain atrophy compared with placebo (Smith et al., PLoS One 2010) — a single trial in a selected population, not a general recommendation, but a signal that the strongest case for treating this marker may not be the cardiovascular one everyone quotes.
The clinical pearl: treat an elevated homocysteine as a prompt to investigate, not as a target to shoot at. It is cheap, it is easy to lower, and lowering it in isolation has not shown the cardiovascular payoff the association implied. What it reliably does is flag B12 or folate deficiency, unrecognised kidney or thyroid disease, or a medication effect — all of which are worth finding for their own sake.
What to do with an elevated result
The sequence matters more than the supplement.
- Confirm and contextualise. Repeat the test, and check serum B12 with methylmalonic acid, folate, thyroid function and kidney function. Correcting a B12 deficiency is a different task from mopping up a number.
- Review medications and alcohol before adding anything. Metformin-associated B12 depletion is common, silent and easily corrected — see B12 and methylation.
- Replace the cofactors that are actually short, in methylated forms where an MTHFR variant is known, with dosing and duration set by the physician who ordered the test.
- Retest at around three months. Most people who are going to respond have responded by then.
- Keep it in proportion. If cardiovascular risk is the real question, ApoB and Lp(a) carry considerably more decision-making weight than homocysteine does.
What to expect
The number itself usually moves, and moves fast — most people who respond to cofactor replacement show it by the three-month retest. What you should not expect is to feel different, unless the underlying finding was a genuine deficiency; a corrected B12 deficiency can produce a real change in energy and cognition, while normalising a homocysteine of 12 in someone who was not deficient typically produces no symptom change at all. Non-responders exist, and a homocysteine that will not come down on adequate cofactors is a reason to look harder at kidney function and thyroid rather than to escalate the supplements.
Bottom line
Homocysteine is a functional readout of the methylation cycle and an independent risk marker, and it is not on most standard panels — so it has to be asked for. Elevation usually means a B-vitamin shortfall, an MTHFR variant, or unrecognised kidney or thyroid disease, and it is generally straightforward to lower. Be honest about what that buys: randomised trials lowered the number without reliably lowering cardiovascular events, so its main value is as a prompt to find the upstream problem. Test it as part of a complete panel, act on what it points to, and keep ApoB and Lp(a) at the front of the cardiovascular conversation.
Educational content, not medical advice. Laboratory interpretation and any treatment decision are made by a licensed physician after individual evaluation. Individual results vary.
