What causes restless legs syndrome

The strongest working theory is that restless legs syndrome comes from a shortage of iron in specific brain regions that control movement — not the iron measured by a standard blood test. This local shortfall is thought to disrupt dopamine signalling, one of the two core mechanisms behind RLS symptoms.

Spinal-fluid and autopsy studies support this picture: people with RLS show a lower-iron protein signature in movement-control brain tissue even when their blood iron looks normal. But the evidence is not uniform — the largest brain-MRI study to date, plus a formal meta-analysis of the prior MRI literature, found no significant pooled evidence of lower brain iron in RLS. That is the single most important caveat to this hypothesis, stated honestly rather than glossed over.

This is also why a normal blood ferritin result does not rule out a brain-iron contribution to symptoms. See the full mechanism, iron to dopamine to hyperarousal, on How RLS works.

What the evidence says

• CSpinal-fluid ferritin is lower and transferrin is higher in RLS than in matched controls, despite normal blood ferritin — the original biochemical evidence for a brain-specific iron shortfall. PMID 10762522
• CAutopsy studies of the substantia nigra, a movement-control brain region, show a coordinated iron-handling protein signature consistent with local iron deficiency. PMID 15136682
• AThe largest brain-MRI study to date (72 RLS patients vs. 72 controls), plus a formal meta-analysis of the prior MRI literature, found no significant pooled evidence of lower brain iron in RLS. PMID 35500370
• BIn a randomized, placebo-controlled trial, IV iron produced significantly greater symptom improvement than placebo, alongside MRI-measurable increases in brain iron in the treated group. PMID 37437492

Restless legs syndrome is substantially genetic. The founding 2007 genome-wide studies identified MEIS1, BTBD9, and MAP2K5/SKOR1 as risk genes, each raising risk by more than 50%. A 2024 mega-analysis of over 116,000 cases expanded that list to 164 risk loci, including genes involved in glutamate signalling.

One of these genes gives a direct link back to the iron story: each BTBD9 risk variant a person carries lowers their average serum ferritin by about 13% — a plausible bridge between inherited risk and the brain-iron hypothesis above.

Genetics is not the whole story. Known common genetic variants explain less than 10% of the total heritability of RLS, despite MEIS1 being one of the strongest genetic signals found for any common disease — a substantial share of inherited risk remains unmapped (see What this doesn't explain yet, below).

What the evidence says

• BThe founding 2007 German genome-wide association study identified MEIS1, BTBD9, and MAP2K5/SKOR1 as genome-wide-significant RLS risk loci. PMID 17637780
• AA 2024 mega-meta-analysis of 116,647 cases expanded confirmed genetic risk to 164 loci, including glutamate-receptor genes. PMID 38839884
• BEach BTBD9 risk allele was associated with an average 13% reduction in serum ferritin. PMID 17634447
• DKnown common genetic variants explain less than 10% of RLS heritability — a persistent “missing heritability” problem. PMID 31551905

RLS is markedly more common in people with chronic kidney disease, especially those on dialysis, where prevalence runs 15-30% versus 5-10% in the general population.

Treating the kidney disease itself can resolve RLS entirely: in one case series, RLS symptoms disappeared completely within 1 to 21 days of kidney transplantation in every one of 11 patients studied, and symptoms returned quickly whenever the transplanted kidney later failed.

What the evidence says

• DA systematic review of the dialysis-population literature found chronic kidney disease-associated RLS prevalence of 15-30%, two to three times the general-population rate. PMID 37008995
• CIn an 11-patient case series, RLS resolved completely within 1-21 days of kidney transplantation and recurred rapidly whenever the graft later failed. PMID 12360562

Restless legs syndrome affects roughly 1 in 5 pregnancies, rising through the trimesters (about 8% in the first, 16% in the second, 22% in the third) and falling to around 4% shortly after delivery.

For the full picture on why pregnancy triggers or worsens RLS, what is safe to try, and what guidelines say about iron and medication in pregnancy, see Restless legs syndrome in pregnancy.

What the evidence says

• AA meta-analysis of 27 studies (51,717 subjects) found pooled RLS prevalence of 21% in pregnancy, rising by trimester and falling to 4% postpartum. PMID 29169861

Nerve damage in the legs (peripheral neuropathy) is another recognized trigger. In one specialist clinic cohort, about 30% of patients with polyneuropathy also had RLS — more often when the neuropathy affected small nerve fibers or had an acquired, immune-related cause, rather than the more common large-fiber, axonal type.

What the evidence says

• CRLS occurred in about 30% of a general polyneuropathy clinic cohort, more often with small-fiber or acquired/dysimmune neuropathy than large-fiber axonal disease. PMID 16685701

Iron-deficiency anemia unrelated to kidney disease or pregnancy also carries a strikingly high RLS rate. In one clinic-based study, 41% of women diagnosed with iron-deficiency anemia also had RLS — described by the study authors as an 8-to-10-fold increase over general-population estimates.

This connects directly to the brain-iron hypothesis above: correcting iron status is one of the best-supported, most mechanistically targeted non-drug steps available, and it is Step 1 of the long-term plan on this site.

What the evidence says

• C41.1% of women with iron-deficiency anemia unrelated to kidney disease or pregnancy also met criteria for RLS, an 8-to-10-fold increase over general-population estimates. PMID 29638100

Some medications prescribed for entirely unrelated reasons can trigger new RLS symptoms or make existing symptoms worse. Reviewing the full medicine cabinet — not just the RLS treatment itself — is often the highest-return, lowest-risk step available, and it is not always asked about.

The clearest signal is for two drugs with antihistamine-like activity: mirtazapine (an antidepressant) and quetiapine (an antipsychotic). Together they account for the large majority of drug-induced RLS cases identified in a large pharmacovigilance program, and mirtazapine alone provoked or worsened RLS in over a quarter of new starts in one prospective study — compared with none on reboxetine, a chemically different antidepressant.

Not every antidepressant carries the same risk. In a very large pharmacovigilance analysis, SSRIs specifically were the one antidepressant class not significantly linked to RLS — a genuine nuance worth knowing before assuming every antidepressant is a problem.

Metoclopramide, an anti-nausea drug that blocks dopamine, and acid-suppressing drugs (proton-pump inhibitors and H2-blockers) are both independently associated with RLS.

Sedating first-generation antihistamines (such as diphenhydramine) are widely assumed to aggravate RLS by a similar mechanism, but no dedicated controlled or cohort study in this evidence base confirms that. It is a plausible question worth raising with a prescriber, not an established fact.

The full table of aggravating drugs — including melatonin, and exactly what to ask a prescriber about each one — lives on Why the medication fails, since the same drug classes are also tied to dopamine-agonist augmentation.

• Highest-risk: mirtazapine and quetiapine.
• More nuanced: SSRIs specifically were not significantly linked to RLS in one large analysis.
• Also implicated: metoclopramide, proton-pump inhibitors, and H2-blockers.
• Suspected but unconfirmed: sedating first-generation antihistamines.

What the evidence says

• BMirtazapine provoked or worsened RLS in 28% of new starts, versus 0% on reboxetine, in a prospective multi-site cohort. PMID 18468624
• BIn a 340,099-patient pharmacovigilance program, mirtazapine and quetiapine together accounted for over 80% of all drug-induced RLS cases. PMID 42251748
• CIn a pharmacovigilance analysis of over 14 million reports, SSRIs specifically were the one antidepressant class not significantly linked to RLS. PMID 32546134
• CRLS was significantly more common among systemic sclerosis patients taking metoclopramide than those not taking it. PMID 23456369
• BProton-pump inhibitors and H2-receptor antagonists were associated with RLS across two national blood-donor cohorts, independent of measured iron stores. PMID 33119070

RLS and Parkinson's disease are related but different conditions — having RLS does not mean someone will develop Parkinson's. Most people with RLS never develop Parkinson's disease.

That said, the association is real and measurable at a population level. In a very large German case-control study of over 138,000 Parkinson's cases, a prior RLS diagnosis was associated with roughly four times higher odds of a later Parkinson's diagnosis — and the association remained strong even when RLS was diagnosed 5 to 10 years earlier, suggesting RLS can be an early signal rather than only something that appears after Parkinson's develops.

An odds ratio like this describes a population-level pattern, not an individual forecast. It does not mean a person with RLS has a 4-in-1 chance of developing Parkinson's — the baseline risk of Parkinson's is low, so even a four-fold increase leaves the great majority of people with RLS unaffected.

What the evidence says

• BIn a case-control study of 138,345 Parkinson's disease cases and 276,690 matched controls, RLS was associated with a subsequent Parkinson's diagnosis (adjusted OR 4.19), including when RLS was diagnosed 5-10 years earlier (OR 3.73). PMID 36342675

This page states the leading hypotheses as they stand today, not a settled, complete picture. A few open questions are worth naming honestly:

None of this uncertainty changes the practical starting point: checking iron status and reviewing the medicine cabinet are still the best-supported, lowest-risk first steps, laid out step by step in the long-term plan.

• Why in-vivo brain-MRI evidence for iron deficiency is so inconsistent — including a large, well-powered study finding no pooled deficit — when spinal-fluid and autopsy evidence point more consistently toward it.
• What explains the “missing heritability”: known genetic variants account for less than 10% of the inherited risk for RLS.
• Whether treating a specific secondary cause (kidney transplant, correcting anemia, removing an aggravating drug) reliably and durably resolves RLS, versus only reducing its severity — most of this evidence is observational, not from controlled trials.