381‒Alzheimer’s disease in women: how hormonal transitions impact the brain, new therapies, & more
These notes were generated by AI and may contain inaccuracies.
Peter Attia introduces the Drive podcast episode with guest Lisa Mosconi, combining topics of women's health (pre-, peri-, and post-menopause) and brain health, specifically Alzheimer's disease in women.
Lisa Mosconi, born and raised in Florence, Italy, has a PhD in neuroscience and nuclear medicine (branch of radiology). She focuses on brain scans, memory, language, and cognition. Her family includes nuclear physicists and army members.
- Grew up in a multi-generational household in Florence.
- Grandmother (rock of the family, intelligent, survived WWII, raised family while grandfather was POW) developed cognitive decline living on the same landing.
- Grandmother diagnosed with Alzheimer's disease; lost memory, communication, cooking ability.
- One of four siblings: all three sisters developed Alzheimer's and died from it in late 80s (onset late 70s, ~10-year duration, subtle start with mild cognitive impairment she masked); brother spared despite same age.
"It's very typical unfortunately it is quite typical."
Research using biological markers (brain imaging, blood fluids) detects early risk. Cognitive tests are late-stage.
- Preclinical phase: lasts decades; proteins/lesions visible, no objective impairment, but subjective changes (patients feel off, not performing as before).
- Brain resilient, compensates long time.
- Gray area: patients anxious/depressed, tests normal.
- Awareness/distress until severe dementia (forget self, family, location).
Peter asks about awareness creating distress vs. when patients no longer suffer.
Dementia is umbrella term categorized by pathology/lesions, clinical symptoms, age of onset.
- Alzheimer's disease: ~70% of dementia cases; memory-focused.
- Frontotemporal dementia (FTD): earlier onset, aphasia/language issues (e.g., Bruce Willis); distinct brain scan patterns.
- Lewy body dementia: alpha-synuclein protein mutations.
- Vascular dementia: common, overlaps.
- Mixed dementia common (e.g., Alzheimer's + vascular/Lewy body).
Late clinical diagnosis improved by biological markers for earlier differentiation/therapeutics.
"We're trying to halt progress... catch people when they're still relatively healthy... delaying onset or preventing lesion accumulation."
Lisa worked at NYU School of Medicine with mentor Dr. Mony de Leon (pioneer in Alzheimer's prevention), focusing on under-65s (20 years ago, rare).
Prevalence overlap means totals >100%.
- Lewy body: ~10-20%.
- FTD: less.
- Vascular: intermixed, hard to quantify.
Other dementias not female-skewed:
- Vascular: 50/50.
- Parkinson's with dementia: more men.
- FTD: more men.
- Lewy body: 50/50.
Known since 1990s: after age, female sex strongest risk factor.
Not just longevity (women live 2-3 years longer, gap narrowing, e.g., 2 years in England; doesn't explain 2:1).
- Women don't have higher prevalence in other age-related diseases (cancer, CV) or dementias.
- Alzheimer's #1 cause of death for women in some low/middle SES countries, European countries, parts of US.
Debate persists: prevalence (cross-sectional) vs. incidence (new cases); if incidence equal, longevity explains accumulation.
- Incidence studies show higher in women, especially low/middle SES.
- Mortality: women die more from it.
Lisa's research (midlife, 45-65, high-risk: family history/APOE4):
- Women show more Alzheimer's red flags on brain imaging/biofluids vs. men same age.
- Faster lesion progression in women.
- Same symptom severity: women have more pathology.
- Women start earlier (midlife), compensate longer via higher cognitive reserve (verbal memory advantage), masking early diagnosis.
"Alzheimer's is not a disease of old age. It's a disease of midlife with symptoms that start in old age."
Analogy to osteoporosis: childhood disease manifesting later.
Beyond age: hormonal transitions (gradual male androgen loss vs. sudden female loss midlife).
Lisa's 2017 study (first): brain imaging pre-/peri-/post-menopause (no HRT, age-matched men).
- Premenopausal (regular cycles): barely any brain differences vs. men.
Distinction between imaging modalities (anatomical resolution, e.g., MRI with T1 vs. T2, FLAIR to highlight white matter vs. gray matter, vascular system) and functional modalities (e.g., FDG PET for metabolic activity of cells).
- CT scans are more anatomical.
- Key is what you're measuring: structural, functional, biochemical, or pathological information.
MRI for Anatomy, Volumetrics, and Alzheimer's Risk Factors
MRI (especially T1 MRI) used for:
- Brain volumetrics: Ensure dense neurons; atrophy as risk factor for dementia.
- Medial temporal lobe (primitive structures for memory and emotional regulation).
- Hippocampus as biomarker for Alzheimer's risk (larger volume better; reductions/thinning in hippocampus and parahippocampus indicate risk, not diagnosis).
- Check for tumors, vascular damage, stroke.
T2 and FLAIR scans for additional info:
- Gliosis (non-specific; sign of white matter integrity damage, punctuations with aging, inflammation, vascular insults).
- Monitor vascular system.
Advanced MRI Techniques
- DTI (diffusion tensor imaging): Structural connectivity of brain fibers between neurons.
- ASL (arterial spin labeling): Non-invasive, quick blood flow measurement.
- 31P MRS (phosphorus-31 magnetic resonance spectroscopy): ATP production (phosphocreatine to ATP ratio as biomarker for brain stress, energetic damage, neuronal metabolic stress). Only MRI technique (PET emerging).
- All done in <1 hour, multiple sequences, one table time (switch coil for spectroscopy from hydrogen to phosphorus; needs ≥36 coils; advantage in brain due to head immobilization, short distance).
PET Scans
- FDG PET: Metabolic activity.
- ¹¹C-PIB (Pittsburgh Compound B): Alzheimer's plaques (amyloid PET tracer; carbon-11 labeled, allows back-to-back with FDG, no wait; clearer signal-to-noise than fluorinated tracers, especially in younger for research).
- Requires on-site cyclotron, chemist; decays quickly.
- Fluorinated tracers commercially available alternative.
First attempt to measure estrogen/hormones in brain (hard due to reasons; motivated by menopause-Alzheimer's link in women, need proof beyond rats, tool for hormone therapy effects).
- Measures estrogen receptor (ER) density (not estrogen levels directly).
- Tracer: ¹⁸F-fluoroestradiol (estradiol labeled with fluorine-18; mimics estradiol, binds ERs like key in lock, "jams" receptor; emits gamma rays for imaging; kinetic modeling for density per brain region).
- 2024 proof-of-concept: Signal in pituitary gland (non-specific, not confounded by metabolites/blood-brain barrier).
- Expected high ER density in pituitary/hypothaus for FSH/LH feedback.
- Pituitary: Anterior not protected by BBB, posterior is; hypothalamus inside BBB.
Estrogen Transport Across Blood-Brain Barrier
- Transporters exist.
- Binds (to albumin?); needs kinetic modeling.
- Uptake: Minutes post-IV injection; time-activity curve vs. blood kinetics.
- Scanning: 90 min total, peak uptake 15-35 min, then saturation.
- Models disentangle blood flow (e.g., Logan plot).
- Reference region: Inferiormost gray matter layer of cerebellar cortex (posterior inferior; void of ER-alpha per postmortem studies; mostly beta receptors; invariant across pre/peri/postmenopausal).
- Collaborator: Dr. Roberta Diaz Brinton (estrogen receptor expert).
Tracer Specifics and ER Types
- Targets ER-alpha (not beta; 17-beta estradiol biologically active in brain but undeveloped tracer; alpha for tumors/breast).
- Repurposed oncology tracer (win-win).
Key Findings on ER Density
- Increases during perimenopause; higher postmenopause (up to age 65; against preclinical rodent models).
- Rodents (ovariectomy): Initial ER overexpression, then crash (narrow window; Gaussian curve; <5 years post-FMP, density halves).
- Humans: No crash; high to 65 (study: 35-65yo, hormone-naive; mapped full window despite team concerns).
"We're going to do 35-65... we're going to try and map the whole window of opportunity."
- Premenopausal: Mostly never used or ≥3 years off OCPs.
- Postmenopausal: Never-users initially; now stratify past/current users.
- Current HT users (unpublished anecdote): Curve shifts; density preserved >65yo (question: good? Are receptors functional? Transcriptional pathways intact?).
ER Functionality and Mechanisms
Steroids (estrogen/testosterone) bind receptors → enter cell/nucleus → bind DNA → RNA → protein (transcription/translation).
- Assay measures binding only, not downstream.
- Stable feedback loop (brain-ovaries) implies function: ↑blood flow, energy, immunity, neuroplasticity, synaptic growth.
- Aging/disease: Malfunction, conformational changes → poor output.
Potential Evidence of ER Dysfunction
- Oxidative stress: ER in mitochondria (energy factory); electron transport chain produces ATP + free radicals (usually ATP-favored).
- Conformational change → more oxidative stress vs. ATP (still energy, but inefficient).
Suggested study: Hormone-naive women (35-65); inject estradiol → measure peripheral mRNA response (normalize to binding signal); expect declining mRNA with age.
"This question implies can we throw more estrogen at the problem?"
In periphery/brain: Studies show timing is really
Studies show timing is critical for brain: in healthy neuronal tissues, estrogen supports neurons; in diseased neurons with ischemic damage, amyloid pathology, or tangles, estrogen worsens it.
Evidence from Dr. Brinton's Studies and WHI
Dr. Brinton's studies showed estrogen impacts mitochondrial function. Clinical evidence from Women's Health Initiative (WHI) memory study confirms, unpacking age-related benefit-risk ratio for hormone therapy (HT) and brain health.
WHI women were too old to start therapy; used different formulations/higher doses than today. Confirmed timing hypothesis, especially sub-analyses showing MRI evidence of vascular lesions or white matter hyperintensities (WMH): women with pre-existing brain damage less responsive than those with healthier brains.
Distinctions: No Reversal vs. Exacerbation
- Estrogen unlikely to reverse set-in disease (one variation).
- Estrogen may exacerbate disease (separate).
WHI used oral conjugated equine estrogen (CEE), slightly increases coagulation—concern for exacerbating vasculopathy. No such evidence with topical estradiol: no increase in vasculopathy or ASCVD risk, safer than oral estradiol.
Progestin (e.g., MPA in WHI) likely culprit for breast cancer incidence uptick (not mortality) and vascular damage—reason not used today.
Couple with oral CEE "getting bad rap"; still serve purpose. Need research on specific HT types and brain effects—we have tools now.
Current Study Design
Normalize to same point; used bioidentical (injected). Collaborate with Weill Cornell Medicine menopause clinic/OB/GYN: women on transdermal estradiol +/- micronized progesterone (standard care). Recruit similar women.
Want to test compounded, oral estrogen, other formulations for differential signals. Most important: transdermal estradiol + oral micronized progesterone.
Brain compensates for estradiol changes by increasing estrogen receptor density. Brain built for stability; during menstrual cycle, predictable concentrations require little adjustment.
Receptors actively made by brain: high estradiol → fewer receptors; low estradiol → overexpress to capture circulating estradiol.
Mechanism eventually crashes: estradiol permanently low, receptor production metabolically expensive. Happens after 65 in studies.
Developing new ligands for better signal in hippocampus, amygdala, frontal cortex, cingulate cortex.
Assay Details
Fully quantitative; low radiation: 6 millicuries (< 1 mSv).
Proposed study: scan 35-year-old women day 1 of period, every 5 days for 30 days—natural experiment of estrogen/progesterone highs/lows. Compare to blood estrogen (poorly correlates with brain). Quantitative → sense compensation speed, highs/lows.
Brain estradiol highly regulated, sheltered from circulation via active transporters—brain "calls for" hormones. Not passive diffusion; hard to get tracers in.
Blood estradiol fluctuations dramatic (near zero to 200 pg/mL); brain response partial, not as dramatic.
Daily scans over year (12 cycles): include flu, sleep deprivation, stress, diet variations. Use identical twins as controls; serves as own control.
Externalities impact peripheral/central physiology: predict lifestyle minor short-term driver vs. hormonal needs (brain tightly regulates to avoid cognitive impairment/confusion).
Poor lifestyle/sleep/stress over time: may impair receptor function/estrogen uptake, cause long-term glitches.
"We're only ankle deep" in understanding.
CARE (Cutting Alzheimer's Risk through Endocrinology): $50 million unrestricted by Wellcome Leap (Wellcome Trust subsidiary). Largest research program on women's brain health, menopause, Alzheimer's disease (AD).
3-year "sprint" (high-risk/high-reward, like Oppenheimer). Goal: halve AD risk for women by 2050; prevent 55 million new cases globally (330 million women).
Three Components
-
Neuroendocrine aging/hormones and AD risk in women: current models sex-aggregated, adjust sex as covariate (genderless). Re-analyze raw data.
- E.g., APOE4: women heterozygous → 4-fold dementia risk vs. non-carriers; two E4 copies → 12-15x (twice men's risk).
- APOE3/APOE4 (heterozygous) vs. APOE3/APOE3: sixfold increase in women (host note).
WHI Memory Study (WHIMS): only trial on HT and AD incidence.
- CEE + MPA: increased dementia risk.
- Post-hysterectomy estrogen-only: 50% higher risk (not significant).
Women long-postmenopausal. No midlife trials (20-30 years needed) measuring dementia.
Observational meta-analyses: timing and formulation matter.
- Hysterectomy (no uterus): estrogen-only, start <10 years final menstrual period (FMP): 32% reduced AD/dementia risk (consistent; one Northern Europe study no effect).
- Intact uterus, <10 years FMP: CEE + progestogen → 23% risk reduction (trend; some increased risk).
-
10 years FMP: no benefit estrogen-only; increased risk estrogen + progestogen (any type).
Insufficient data to separate progestins vs. micronized progesterone, specific HT types.
Better: start HT today, monitor estrogen receptors (brain imaging), AD biomarkers (delay progression vs. placebo)—CARE approach.
CARE Clinical Component
Prospective: follow women spontaneously starting MHT (perimenopause ideal) vs. not on AD biomarkers (3 years feasible, no full RCT).
Address biases:
- Healthy user bias: MHT women more health-conscious, overcome doctor resistance.
- Severe hot flashes (esp. night) associate with higher plasma amyloid beta, brain WMH.
Match carefully (e.g., breast cancer history).
Number one with education, colleagues are making sure women understand hormone therapy is on the table. Women can come to research sites with access to gynecology departments and menopause clinics where clinicians are open to hormone therapy if symptoms are present and patients are interested. No barrier to access if indicated, following professional guidelines.
For women in perimenopausal transition, caution with hormone therapy due to fluctuating hormone levels. Using birth control blocks ovulation for standard tested doses that can be monitored. With menopause hormone therapy, can't monitor daily; concern if estradiol levels spike and added hormones amplify them.
once women are symptomatic... benefits occur immediately with respect to vasomotor symptoms, bone health, cognitive performance in the short term, sexual health and sexual function.
Waiting until fully postmenopausal (FSH 40, estradiol unmeasurable) could take 5 years from symptom onset, with evidence of health steps backwards. Trade-off: loops with ovulations causing estradiol spikes, but clinically most women less bothered than reverse.
Requires nuance, not set it and forget it. Doctors titrate: e.g., 200mg micronized progesterone in full menopause, but 50mg now, 100mg next year. Precision medicine.
Need more research in parallel to guide clinicians basing on experience and patient relationships. No natural owner for studies as drugs are cheap, not protectable.
SERMs interesting: different estrogen receptors (ERα, ERβ, GPER). ERα more in reproductive tissues (pituitary, hypothalamus); ERβ more in cognitive brain areas.
Compounds selectively target ERβ. Dr. Roberta Brinton developed first neuroSERM from plants/phytoestrogens, high affinity to ERβ.
PhytoSERM Details
Tested in animals; phase 1 trial done, now phase 2b in women. Leaves reproductive organs alone, targets brain: stimulates cognition (memory, executive function), mitochondrial activity, neurogenesis (preclinical). Improves brain energy, memory, reduces Alzheimer's plaques risk. Early postmenopausal women. FDA-approved supplement (plant-derived), trials for claims on cognition, brain energy, Alzheimer's risk. Separate trial for hot flashes, VMS.
symptoms... start in the brain. Wonderful to have molecule never associated with reproductive cancer risk, targets brain.
Misses sexual organs, bones benefits (though ERβ in bones, not with this SERM). Future SERMs possible. But estradiol perfect.
No evidence estrogen causes breast cancer, fallacy. Women's Health Initiative (WHI): no additional breast cancer deaths with conjugated equine estrogen (CEE). Increase due to MPA (medroxyprogesterone acetate), few women take today; use bioidentical micronized progesterone.
"Cause" misleading: not zero risk to cancer from estrogen.
Testosterone (endogenous/exogenous) does not drive prostate cancer. Androgen deprivation therapy for treatment if not surgical candidate, but resume testosterone post-surgery, no risk increase.
Urologists advantaged by PSA biomarker: follow PSA, MRI, biopsy. E.g., Gleason 3+3 prostate cancer: watch/wait, no castration even if low testosterone/symptoms—give testosterone. No PSA equivalent for breast tissue creates fear, but doesn't change pathophysiology.
estrogen is very important for their brains.
Stuck with WHI for decades despite newer data. Similar to dietary cholesterol phobia (no yolks, shrimp).
WHI: >$1B 25 years ago, not replicated. Fatal flaws: garbage estrogen/progestin formulation (CEE+MPA, what doctors prescribed then—forgivable). Fault with PIs/media promoting false agenda; rare NIH press conference.
19 years post-WHI: no additional breast cancer deaths in CEE+MPA vs placebo.
Northern Europe retrospective studies: hormone therapy (pre-WHI) associated with dementia risk—all over news. Protective studies (e.g., 500k US women) ignored.
Studies correlational, biases: post-WHI dropouts predictably biased.
No more observational epi. Need biomarkers (C2N, p-tau, brain metabolomics) for rigorous prospective RCTs. Surrogate outcomes as hard outcomes too costly/long.
Clinically: C2N, commercial p-tau, Simoa ready for tracking interventions? Yes, using in CARE: assays same as commercial, surrogate Alzheimer's risk outcomes. Blood biomarkers cheaper, easier than scans. Need long-term predictive value data; used diagnostically/research.
Prospective cohorts, large biobanks (UK Biobank). CARE: female-specific data from 6 continents, >20M women longitudinal data. Establish neuroendocrine aging, reproductive history (puberty, pregnancy, postpartum, menopause) as stress test for cognitive decline/Alzheimer's or resilience.
E.g., preeclampsia previews chronic hypertension (worse Alzheimer's risk in women?). Midlife depression stronger risk in women. Hormonal history as vital sign.
45% Alzheimer's cases from 14 modifiable sex-aggregated risk factors. No female/male-specific. Section on menopause/hormone therapy: may increase risk, especially hysterectomy—shocking.
I'm not dismissive. I'm really like, what can we do to change this? Because
Speakers debate studies claiming hormone therapy after oophorectomy increases Alzheimer's risk. Oophorectomy before menopause is associated with increased Alzheimer's risk, but hormone therapy is beneficial, especially when started immediately after surgery and continued until natural menopause age. Preclinical research supports this biological plausibility.
If your 35-year-old sister needed oophorectomy due to car accident or ovarian cyst rupture, she's chemically menopausal with hot flashes; recommend hormones.
Both agree hormones do not increase Alzheimer's risk and are standard care per professional guidelines, especially for early/surgical menopause. Criticism of observational studies suggesting otherwise as flawed, cherry-picked data ignoring biological plausibility (e.g., Women's Health Initiative misapplied). Commissions fail to include neuroendocrine factors in Alzheimer's prevention recommendations or position them incorrectly.
Independent of weight loss/insulin sensitivity, GLP-1 agonists may protect the brain. Biological plausibility required for studies: preclinical work first, then powered observational studies. Concern over studies ignoring preclinical data or doing fishing expeditions.
Someone developing ligands/tracers for GLP-1s in the brain.
Mechanisms beyond metabolic: improve insulin resistance, reduce obesity-related inflammation.
Hypothetical: For healthy high-risk woman (e.g., like guest), micro-dose tirzepatide (e.g., 2.5 mg) without weight loss/sarcopenia risk—does it protect brain? Needs prospective study with brain biomarkers (blood/CSF markers of neuroinflammation/protein aggregation), not 20-year wait.
Funding challenge: expensive, needs NIH sponsorship as national bipartisan issue. Interest high, shown by ligand/tracer development.
Unpublished pilot data (n=20): 2.5 mg tirzepatide reduces blood/CSF markers of neuroinflammation/protein aggregation.
For women prioritizing brain health (sleep, nutrition, insulin sensitivity managed): consider hormone replacement therapy (HRT) seriously if perimenopausal/menopausal/post-menopausal, discuss long-term beyond symptoms given fast-moving research.
Alzheimer's Prevention Clinic at Weill Cornell Medicine (launched ~2013 by Richard Isaacson): midlife women often fear early dementia; symptoms often menopause-related. Prevention: behavioral/lifestyle (ABCs)—manage hypertension, dyslipidemia, insulin resistance, diabetes, obesity via diet, exercise, stress reduction, sleep hygiene, pharmaceuticals if needed.
Few women check all boxes perfectly. Consistency key: stick to specific patterns (avoid switching keto/vegan/etc.).
Exercise for Midlife Women
Association between intensity and health gains (not muscle mass) is inverted U: moderate intensity (e.g., zone 2) frequently yields greatest benefits. Intensity/duration matter; less time → prioritize intensity; more time → lower intensity optimal. Sweet spot: modest intensity, possibly north of zone 2.
Maximize lifestyle levers to buy time. More optimistic now vs. 5 years ago (previously despaired over protein folding).
Promising: klotho injections (host co-founder of company developing); GLP-1 agonists; hormones; exercise mimetics; irisin/BDNF.
Delay onset 10 years via lifestyle = normal vs. shortened cognitive life. No reversal for today's 75-year-olds with dementia, but for 60-year-olds, consistency positions them for future rescues.
Brain built for stability (unlike body for change): quick body results (lose inches, gain muscle) vs. brain needs frequent/long-term hits for epigenetic/structural changes.
Good news: takes time for damage; bad news: takes time for resilience—but achievable.
Lifestyle creates cognitive resilience/brain reserve: movement → BDNF, irisin for synaptic/dendritic health; reduce inflammation/oxidative stress.
Daily choices matter: invest now for cognitive lifespan matching lifespan (shows in 10 years).
Guest launched high-speed research (Clarity?) via grant, aims for answers in 3 years (before daughter's puberty/menopause).
Manhattan Project ~$34B today's dollars—paltry for impact. $3-4B trivial for US gov't; $50M accomplishes much in 3 years—imagine moonshot scale. Private sector (e.g., Wellcome Trust) filling gaps. Scientists worldwide ready.
Keep Peter Attia MD in your library
Save the videos and channels worth coming back to, and find them again in one place.
