Sex hormones, pregnancy, and the life course of RLS

RLS is roughly twice as common in women as in men, and pregnancy is the single largest natural experiment in reversible RLS: onset or worsening in pregnancy, frequent (though not universal) postpartum remission, and elevated lifetime recurrence/chronic-RLS risk afterward. This synthesis covers the epidemiology of pregnancy-associated RLS, the mechanistic evidence for estrogen/progesterone/prolactin and iron-demand effects, sex differences outside pregnancy (menopause, hormone replacement therapy, oral contraceptives, parity), obstetric outcomes, and how management differs across the paediatric and older-adult ends of the life course. It also carries practical implications for this project's target population: women who may pass through pregnancy or menopause while on pramipexole or gabapentin. Every claim below cites a study note (filename) or PMID from this knowledge base.

What is established (grade A/B)

• Pregnancy prevalence and time-course are well established. A meta-analysis of 27 studies (51,717 pregnant subjects) found pooled RLS prevalence of 21% overall in pregnancy, rising by trimester (8% first, 16% second, 22% third) and falling to 4% postpartum (2018-chen-pregnancy-prevalence-meta-analysis.md, PMID 29169861, pre-existing). A large multicenter cohort (Life-ON, n=439) confirmed a 26.0% antenatal prevalence and 34.7% cumulative incidence to 12 months postpartum, with severity stable across trimesters rather than peaking sharply, and prior depression, family history, and prior RLS as significant risk factors (2026-hackethal-life-on-pregnancy-rls-cohort.md, PMID 40709818, pre-existing).
• Parity is a major, dose-dependent driver of the population sex difference. In a population-based German cohort of 4,310 adults, nulliparous women had RLS risk similar to men, but risk rose stepwise with each birth: OR 1.98 (95% CI 1.25-3.13) for one child, 3.04 (2.11-4.40) for two, 3.57 (2.30-5.55) for three or more (2004-berger-parity-sex-difference-rls.md, PMID 14744844). This is the primary quantitative support for the "parity explains the sex difference" hypothesis.
• A large classic cohort (n=642) established the transient nature of pregnancy RLS: 26% cumulative prevalence, concentrated in the third trimester, resolving around delivery, with affected women showing lower hemoglobin/MCV despite equivalent iron/folate supplementation to unaffected women (2004-manconi-pregnancy-puerperium-cohort.md, PMID 15452299).
• Abrupt loss of ovarian hormones (premenopausal bilateral oophorectomy) increases RLS risk: adjusted HR 1.44 (95% CI 1.08-1.92) versus non-oophorectomized age-matched women, concentrated in women without a benign surgical indication (HR 1.52, 1.03-2.25); post-surgical estrogen therapy did not reduce this excess risk (2021-huo-oophorectomy-rls-risk.md, PMID 33523190).
• The most consistent obstetric associations, per systematic review, are gestational hypertension, pre-eclampsia, and peripartum depression — not cesarean delivery, preterm birth, or low birth weight, where most primary studies found no association (2020-steinweg-pregnancy-rls-outcomes-review.md, PMID 32805557, grade A).
• A large US cohort (n=2,457) found no significant association between RLS and fetal growth trajectory by serial ultrasound, an important negative/reassuring finding for one specific obstetric-outcome domain (2024-na-pregnancy-rls-fetal-growth.md, PMID 38862351).
• Sex differences in RLS phenotype are measurable but severity itself is similar: in a 614-patient case-control study, women more often reported pain/awakening while men reported motor symptoms, and women had lower iron parameters and more frequent iron supplementation than men, but overall RLS severity did not differ by sex (2024-kim-sex-specific-rls-phenotype.md, PMID 37858283).

Emerging C

• Estrogen as a candidate trigger, but contested. A small longitudinal study (n=19) found significantly higher plasma estradiol in third-trimester RLS patients versus pregnancy-matched controls (34,211 vs 25,475 pg/mL, p<0.05), with no significant difference in prolactin, progesterone, testosterone, FSH, LH, iron, ferritin, or hemoglobin (2009-dzaja-estradiol-pregnancy-rls.md, PMID 19238803). A separate, similarly designed study (n=146) measuring the same hormones found no significant estradiol or prolactin difference between RLS and non-RLS pregnant women, and only a non-significant trend toward higher progesterone in RLS, while lower hemoglobin and less iron/vitamin supplementation were the significant discriminators (2007-tunc-pregnancy-rls-predisposing-factors.md, PMID 17285614). These two small studies point in different directions on estrogen specifically (see Contested, below).
• Obstetric-outcome associations replicate across populations. A Peruvian cohort (n=218) found pre-eclampsia in 17.5% of RLS-affected women versus 2.6% of matched controls (p=0.03) alongside less prophylactic iron supplementation in the RLS group (2013-ramirez-preeclampsia-rls-peru.md, PMID 23891236); a large Chinese survey (n=3,874) found RLS associated with higher rates of hypertension, cardiovascular disease, and pre-eclampsia (2015-liu-china-pregnancy-rls-complications.md, PMID 26847992); a US surrogate-symptom cohort (n=1,000) found a dose-response relationship between leg-movement symptom frequency and chronic hypertension (adjusted OR 3.74, 95% CI 1.31-10.72) and lower birth weight/earlier gestational age (2015-oyiengo-rls-pregnancy-neonatal-outcomes.md, PMID 26740290).
• Hormonal contraception and RLS show no significant association. In 145 combined-hormonal-contraceptive users versus 169 matched non-users, RLS prevalence and severity did not differ significantly by contraceptive status (p=0.53 and p=0.21); BMI, not contraceptive use, tracked with RLS presence and severity (2013-kamsheh-oral-contraceptives-rls.md, PMID 23432776).
• Menopause/HRT associations are inconsistent across studies (see Contested, below), and lifespan periodic-limb-movement patterns diverge by sex, rising earlier in females (before age 10) and again after age 55, versus a steadier male trajectory continuing to rise past 75 — the authors propose sex hormones as the underlying driver though this is not directly tested (2024-mogavero-sex-differences-plms-lifespan.md, PMID 39217969, pre-existing).
• Pediatric RLS responds substantially to iron correction. In a retrospective cohort of 97 children, 65% received iron and roughly 80% of those with follow-up improved or resolved, at a median 3.8 months, with median baseline ferritin only 22.7 ng/mL (2013-amos-pediatric-rls-iron-cohort.md, PMID 24198314).
• Regional pregnancy-prevalence studies broadly replicate the ~11-27% range seen in the pooled meta-analysis: Thailand 11.2% with a sharp third-trimester concentration and full postpartum resolution within a week (2019-panvatvanich-rls-pregnant-thai-women.md, PMID 30618223, pre-existing), Saudi Arabia 26.6% with calcium (not iron) as the significant micronutrient correlate in multivariate analysis (2023-alnaaim-saudi-pregnancy-rls-prevalence.md, PMID 37664361, pre-existing), and Turkey with 44% reporting two or more RLS complaints alongside broader sleep disturbance (2021-kadioglu-pregnancy-rls-sleep-turkey.md, PMID 34311492, pre-existing).
• A single case suggests pre-conception IV iron may prevent pregnancy RLS: complete symptom remission through pregnancy until relapse ~5 months postpartum (2012-picchietti-iv-iron-prepregnancy-case.md, PMID 23066374, grade E — hypothesis-generating only).

Contested or negative findings

• Does estrogen itself trigger pregnancy RLS? Dzaja 2009 found significantly elevated estradiol in RLS-affected pregnant women; Tunc 2007, measuring the same hormone panel in a larger sample, did not. Both are small, single-timepoint studies; this is genuinely unresolved, not settled science.
• Does HRT/estrogen help, harm, or do nothing to RLS? The Manconi 2011 gender-focused review states menopause does not clearly change RLS incidence and HRT does not clearly increase it, describing the relationship as "uncertain" (2011-manconi-gender-differences-rls-review.md, PMID 22075215). A national position statement similarly calls the MHT-RLS evidence "conflicting" even while endorsing MHT for vasomotor symptoms and insomnia (2019-silvestri-aims-menopause-sleep-guideline.md, PMID 31547910). A postmenopausal Iranian cohort (n=980) found HRT users were less likely to have RLS (2018-bagheri-postmenopausal-rls-prevalence-iran.md, PMID 29419360) — the opposite direction from what estrogen-as-trigger would predict — while Huo 2021 found estrogen therapy after oophorectomy did not reduce the elevated post-surgical RLS risk. These three results are not easily reconciled and should not be presented as a coherent HRT recommendation.
• Oral contraceptives and RLS: a clean negative result (Kamsheh 2013) argues against a meaningful effect of steady exogenous hormone dosing, in contrast to the abrupt hormonal swings of pregnancy onset/postpartum and surgical menopause that do appear to matter.

Mechanistic picture in 5 bullet points

1. Hormonal instability, not steady-state level, may be the more reliable correlate. Abrupt estrogen/progesterone/prolactin rise (pregnancy, especially third trimester) and abrupt estrogen/progesterone loss (bilateral oophorectomy) both raise RLS risk (2009-dzaja-estradiol-pregnancy-rls.md; 2021-huo-oophorectomy-rls-risk.md), while steady exogenous dosing (oral contraceptives) shows no effect (2013-kamsheh-oral-contraceptives-rls.md) and gradual natural menopause shows no clear directional effect (2011-manconi-gender-differences-rls-review.md) — a pattern consistent with hormonal change stressing dopaminergic/iron-handling systems more than any particular hormone level does.
2. Iron demand and hormonal change likely act on the same iron-dopamine-hyperarousal circuitry described in this project's brain-iron-deficiency topic file, not through a separate pathway. Pregnancy dramatically increases iron demand (expanded blood volume, fetal/placental transfer) precisely when RLS peaks (third trimester), and lower hemoglobin/iron-supplementation status tracks with RLS status in more studies than estrogen level does (2004-manconi-pregnancy-puerperium-cohort.md; 2007-tunc-pregnancy-rls-predisposing-factors.md; 2013-ramirez-preeclampsia-rls-peru.md).
3. Prolactin and dopamine receptor sensitivity are neuroendocrinologically linked even outside pregnancy. A neuroendocrine-challenge study found RLS patients have exaggerated nighttime (but not daytime) dopamine-receptor-mediated prolactin suppression, with prolactin levels correlating with periodic-limb-movement index — independent circumstantial support for a prolactin-dopamine circadian interaction that could plausibly be amplified by pregnancy's physiologic prolactin rise (2004-garcia-borreguero-neuroendocrine-ldopa-circadian.md, PMID 15283001, pre-existing).
4. Pregnancy-associated hormonal/iron stress appears to leave a durable trace on the system, not just a transient perturbation. Each additional pregnancy raises lifetime RLS odds in a dose-dependent way (2004-berger-parity-sex-difference-rls.md), and pregnancy-onset RLS predicts higher risk of both recurrence in future pregnancies and chronic primary RLS later in life (2015-prosperetti-pregnancy-rls-review.md, pre-existing) — suggesting the episode is not purely reversible at the population level even when an individual pregnancy's symptoms fully resolve.
5. Postpartum hormonal collapse coincides with, and plausibly drives, rapid symptom remission — RLS falls from a 22-26% antenatal prevalence to about 4% postpartum within weeks (2018-chen-pregnancy-prevalence-meta-analysis.md; 2004-manconi-pregnancy-puerperium-cohort.md) — but this reversal is probabilistic, not universal: cumulative incidence continued to rise to 34.7% by 12 months postpartum in the Life-ON cohort, meaning some women's RLS does not fully reset (2026-hackethal-life-on-pregnancy-rls-cohort.md).

Implications for a non-drug long-term plan for people currently on dopamine agonists or alpha-2-delta ligands

• A woman on gabapentin who becomes pregnant: existing IRLSSG consensus guidance (2014-picchietti-irlssg-pregnancy-lactation-guideline.md, pre-existing) and later reviews (2020-jahani-kondori-rls-pregnancy-lactation-management.md) recommend non-pharmacologic measures first, iron-status testing and repletion as a foundational step, and continuing or adjusting medication only under physician supervision with attention to pregnancy-specific safety data. Pooled first-trimester gabapentin-exposure registries (294 exposures) found a major-malformation rate (1.7%) comparable to the general-population background (1.6-2.2%), a cautiously reassuring but still limited (non-randomized, modest-sample) safety signal (2014-guttuso-gabapentin-pregnancy-safety.md, PMID 25195202). Per this project's non-negotiable rules, any change to a dopamine agonist must never be abrupt, given augmentation and DAWS risk — pregnancy is not an exception.
• A woman on pramipexole approaching menopause: the honest answer, per the evidence above, is that neither menopause itself nor starting HRT can be confidently predicted to improve or worsen her RLS — the guideline-level literature explicitly calls this relationship conflicting (2019-silvestri-aims-menopause-sleep-guideline.md) — so the standard iron-first, gabapentinoid-preferred, augmentation-aware protocol from this project's core guideline notes (2026-winkelman-rls-jama-clinical-review.md; 2026-silber-updated-rls-management-algorithm.md, both pre-existing) should remain the primary lever rather than hormone therapy tailored to RLS specifically.
• Iron repletion is the one hormone-adjacent, non-drug intervention with a plausible preventive role across the whole life course — from pediatric iron-responsive RLS (80% improvement/resolution with iron in Amos 2013) through pregnancy (Tunc 2007, Manconi 2004) to the IRLSSG's general iron-treatment guideline (2018-allen-irlssg-iron-treatment-guidelines.md, pre-existing, PMID 29425576) — though pre-conception iron repletion specifically preventing pregnancy RLS rests on one case report (Picchietti 2012) and needs a controlled trial.
• Obstetric vigilance, not RLS-specific drug escalation, is the appropriate response to the pre-eclampsia/hypertension/depression associations. Because this evidence is observational (Steinweg 2020's systematic review, Liu 2015, Ramirez 2013), the responsible inference is closer obstetric and mood monitoring for RLS-affected pregnant patients, not a claim that treating RLS prevents these outcomes — that causal question remains untested.
• Paediatric management differs mainly in sequencing, not philosophy: non-drug measures and iron correction are used even more strictly as the first and often sufficient step before considering any dopaminergic or gabapentinoid medication, given a much thinner pediatric pharmacologic safety/efficacy evidence base than in adults (2018-delrosso-pediatric-rls-treatment-review.md; 2013-amos-pediatric-rls-iron-cohort.md).
• Older-adult management differs mainly in caution around comorbidity and polypharmacy, not in the core iron-first/gabapentinoid-preferred algorithm, per this project's existing life-course guideline notes (2026-winkelman-rls-jama-clinical-review.md; 2026-silber-updated-rls-management-algorithm.md), which explicitly extend their recommendations to older patients alongside pregnancy and pediatric populations.

What pregnancy remission tells us about reversibility of RLS

This is inference from this project's synthesis, not a proven causal claim from any single study (per Master Plan honesty rule 4), and it is not a treatment or a cure (per Master Plan honesty rule 1). The fact that RLS prevalence falls from roughly 22-26% in the third trimester to about 4% within weeks postpartum (2018-chen-pregnancy-prevalence-meta-analysis.md; 2004-manconi-pregnancy-puerperium-cohort.md) — coinciding with the rapid postpartum collapse of estrogen, progesterone, and prolactin, and with restoration of normal blood volume and iron dynamics — is the strongest existing human evidence that RLS symptom expression can be substantially and rapidly reversible when its physiological drivers (hormonal and iron-related) are removed. If that inference is correct, it would suggest that in non-pregnant patients, a sufficiently thorough and sustained correction of the analogous drivers (iron/ferritin status, removal of aggravating medications, treatment of comorbidities) might likewise permit substantial, though not guaranteed, symptom reduction — consistent with this project's mission framing of "long-term management" and "reduce need for medication," never "cure." The important counter-evidence that must accompany this inference: pregnancy-onset RLS also raises lifetime chronic-RLS risk (2015-prosperetti-pregnancy-rls-review.md) and cumulative incidence continued rising to 34.7% at 12 months postpartum in the Life-ON cohort (2026-hackethal-life-on-pregnancy-rls-cohort.md) — so remission in pregnancy is real and informative about reversibility in principle, but it is partial, probabilistic, and not evidence that any given individual's RLS will fully or permanently resolve outside the unique physiology of pregnancy.

Open questions

• Testosterone and androgen data in men are essentially absent. A dedicated PubMed search for "restless legs syndrome AND testosterone AND men" returned zero results at the time of this sweep — a genuine, notable gap given how much attention estrogen/progesterone/prolactin have received in women; not reported.
• Does estradiol elevation actually drive pregnancy-onset RLS (Dzaja 2009) or is iron/hemoglobin status the dominant factor with hormones as a secondary correlate (Tunc 2007, Manconi 2004)? No study has yet measured both with adequate power in the same cohort.
• Why do HRT/estrogen-exposure studies point in different directions (protective in Bagheri 2018, no effect in Manconi 2011's review, ineffective at reversing oophorectomy-related risk in Huo 2021)? Confounding by indication, dose/formulation differences, and age-at-exposure differences are all plausible but untested explanations.
• Does treating RLS during pregnancy (rather than only observing it) reduce the incidence of pre-eclampsia, gestational hypertension, or peripartum depression? Steinweg 2020 explicitly flags this as untested.
• Does pre-conception iron repletion reliably prevent or reduce pregnancy-onset RLS? Only one case report exists (Picchietti 2012); a controlled trial is needed.
• What happens beyond 12 months postpartum? The largest prospective cohort (Life-ON) only follows to 12 months; longer-term recurrence/chronic-RLS trajectories rely on retrospective recall in older studies.
• Has the parity-explains-sex-difference hypothesis (Berger 2004) been retested in a modern cohort with objective iron and hormone measurement and stronger confounder control? Not found in this sweep.

Notes in this topic

New notes added by this sweep:

• 2011-manconi-gender-differences-rls-review.md — Workshop review: women 2x male RLS risk, parity raises risk, menopause/HRT effects uncertain (PMID 22075215).
• 2004-manconi-pregnancy-risk-factor-review.md — Early review naming iron/folate and hormonal (prolactin/progesterone/estrogen) hypotheses for pregnancy RLS (PMID 15165540).
• 2007-tunc-pregnancy-rls-predisposing-factors.md — n=146: hemoglobin/iron supplementation, not estradiol/prolactin, distinguished RLS from non-RLS pregnant women (PMID 17285614).
• 2009-dzaja-estradiol-pregnancy-rls.md — n=19: estradiol significantly elevated in third-trimester RLS versus controls; other hormones and iron markers not different (PMID 19238803).
• 2015-prosperetti-pregnancy-rls-review.md — Review: pregnancy-onset RLS predicts recurrence in future pregnancies and higher later chronic-RLS risk (PMID 26329442).
• 2004-manconi-pregnancy-puerperium-cohort.md — n=642 classic cohort: 26% prevalence, third-trimester peak, resolution near delivery (PMID 15452299).
• 2024-na-pregnancy-rls-fetal-growth.md — n=2,457: no association between RLS and fetal growth trajectory (PMID 38862351).
• 2004-berger-parity-sex-difference-rls.md — n=4,310: parity dose-dependently raises RLS risk, nulliparous women resemble men (PMID 14744844).
• 2024-kim-sex-specific-rls-phenotype.md — n=614+465 controls: sex differences in symptom type and iron status, not severity (PMID 37858283).
• 2021-huo-oophorectomy-rls-risk.md — n=3,306: premenopausal bilateral oophorectomy raises RLS risk; estrogen therapy does not reverse it (PMID 33523190).
• 2018-bagheri-postmenopausal-rls-prevalence-iran.md — n=980: 16% postmenopausal RLS prevalence; HRT users less likely to have RLS (PMID 29419360).
• 2019-silvestri-aims-menopause-sleep-guideline.md — National guideline: MHT helps vasomotor/insomnia symptoms but RLS evidence is "conflicting" (PMID 31547910).
• 2020-jahani-kondori-rls-pregnancy-lactation-management.md — Review + case: non-drug first, iron core intervention, medication data sparse in pregnancy/lactation (PMID 32054396).
• 2014-guttuso-gabapentin-pregnancy-safety.md — Pooled registries (n=294): gabapentin first-trimester malformation rate comparable to background (PMID 25195202).
• 2013-kamsheh-oral-contraceptives-rls.md — n=314: no significant association between combined hormonal contraceptive use and RLS (PMID 23432776).
• 2012-picchietti-iv-iron-prepregnancy-case.md — Case report: pre-conception IV iron associated with RLS remission through pregnancy (PMID 23066374, grade E).
• 2020-steinweg-pregnancy-rls-outcomes-review.md — Systematic review (16 studies): gestational hypertension, pre-eclampsia, peripartum depression most consistently associated (PMID 32805557).
• 2015-liu-china-pregnancy-rls-complications.md — n=3,874: RLS associated with hypertension, cardiovascular disease, pre-eclampsia (PMID 26847992).
• 2013-ramirez-preeclampsia-rls-peru.md — n=218: pre-eclampsia more frequent with RLS (17.5% vs 2.6%, p=0.03) (PMID 23891236).
• 2015-oyiengo-rls-pregnancy-neonatal-outcomes.md — n=1,000: RLS symptom frequency associated with chronic hypertension and lower birth weight (PMID 26740290).
• 2018-delrosso-pediatric-rls-treatment-review.md — Review: iron-first, non-drug-first approach in children given thin pediatric pharmacologic evidence (PMID 31229174).
• 2013-amos-pediatric-rls-iron-cohort.md — n=97 children: ~80% improved/resolved with iron treatment, median 3.8 months (PMID 24198314).

Pre-existing notes (from other sweeps) referenced here:

• 2018-chen-pregnancy-prevalence-meta-analysis.md — Meta-analysis (27 studies, n=51,717): pooled 21% pregnancy prevalence, trimester gradient, postpartum drop to 4% (PMID 29169861).
• 2014-picchietti-irlssg-pregnancy-lactation-guideline.md — IRLSSG consensus guideline for RLS diagnosis/treatment in pregnancy and lactation (PMID 25553600).
• 2014-wesstrom-sweden-pregnancy-rls-depression.md — n=1,428: pre-pregnancy-onset RLS linked to perinatal depression; de novo pregnancy RLS not (PMID 24812538).
• 2026-hackethal-life-on-pregnancy-rls-cohort.md — n=439: 26% antenatal prevalence, 34.7% cumulative incidence to 12 months postpartum (PMID 40709818).
• 2019-panvatvanich-rls-pregnant-thai-women.md — n=214 Thai cohort: 11.2% prevalence, sharp third-trimester rise, full postpartum resolution (PMID 30618223).
• 2020-jafarimanesh-thermotherapy-cryotherapy-pregnancy-rls.md / 2020-jafarimanesh-thermotherapy-pregnancy-rct.md — RCT (n=80): cold-water leg immersion reduces pregnancy RLS severity more than warm water (PMID 32444058).
• 2021-kadioglu-pregnancy-rls-sleep-turkey.md — n=189 Turkish cohort: 44% reported 2+ RLS complaints; linked to broader sleep/QoL impairment (PMID 34311492).
• 2022-hansen-iv-vs-oral-iron-pregnancy-rct.md — RCT: IV ferric derisomaltose outperforms oral iron for persistent iron deficiency in pregnancy (PMID 36107229).
• 2023-akbas-progressive-muscle-relaxation-pregnancy-rct.md — RCT: progressive muscle relaxation improves RLS symptoms and sleep quality in pregnancy (PMID 37146530).
• 2023-alnaaim-saudi-pregnancy-rls-prevalence.md — n=459: 26.6% prevalence; calcium (not iron) deficiency significantly associated (PMID 37664361).
• 2024-esteves-pregnancy-exercise-lifeon.md — n=316 (Life-ON): physical activity associated with better objective sleep/fewer PLMs in pregnant RLS patients (PMID 39477367).
• 2026-rhode-island-pens-pregnancy-trial.md — Recruiting trial: peroneal nerve stimulation device feasibility/safety in pregnancy (trial registry).
• 2024-mogavero-sex-differences-plms-lifespan.md — n=184 lifespan cohort: sex-divergent periodic-limb-movement trajectories, proposed hormonal driver (PMID 39217969).
• 2004-garcia-borreguero-neuroendocrine-ldopa-circadian.md — Neuroendocrine challenge: exaggerated nighttime dopamine-prolactin sensitivity in RLS (PMID 15283001).
• 2026-winkelman-rls-jama-clinical-review.md — 2026 JAMA review: sex ratio, life-course prevalence (including older adults), current treatment algorithm (PMID 41563785).
• 2026-silber-updated-rls-management-algorithm.md — 2026 RLS Foundation algorithm covering children, pregnancy/lactation, and other special populations (PMID 42203073).
• 2018-allen-irlssg-iron-treatment-guidelines.md — IRLSSG iron-treatment task force, adults and children (PMID 29425576).