Dietary Silicon — Body Handbook

Supplements · §501

Dietary Silicon

There's a $30 bottle on the supplement shelf claiming to be your skin, hair, nail, and bone-density mineral, often labelled bamboo silica or orthosilicic acid. The trace mineral is real; the bottle mostly is not. A bowl of oats, a slice of wholegrain bread, and the occasional beer cover the daily intake associated with denser hips, slightly more elastic skin, and slower hair thinning. The effect is modest — silicon is a quiet co-factor, not a transformation. But the food version is free, and the more interesting question is whether the bottle is worth buying once your diet already covers it.

Do · Daily Evidence Mixed Chapter Supplements

A small but real lever, sitting in foods you probably already eat. The bone-density signal is consistent in men and women who still have their oestrogen — about a tenth more hip density at the high-intake end of the population. Skin and hair changes are slow and modest, not dramatic. The reason this entry exists is not that the effect is large; it's that the effect is honestly there, the food source is free, and the bottle is mostly unnecessary.

Collagen — the protein that scaffolds bone, skin, hair shafts, and nails — needs its proline residues hydroxylated before it can fold into the triple helix that gives it strength. That hydroxylation step depends on an enzyme called prolyl hydroxylase. In cell-culture work on human osteoblast-like cells, adding orthosilicic acid (the small, dissolved form of silicon) at micromolar concentrations bumped up type 1 collagen synthesis; block the enzyme, and the silicon effect disappeared Reffitt 2003. The same logic carries to dermal fibroblasts, which build the collagen that holds the skin taut. So the story isn't that silicon is the headline ingredient of collagen; it's that the enzyme that finishes collagen seems to work better with silicon around.

Most of the silicon you eat — bound up in plant tissues as silica — never makes it past the gut. The fraction that does cross is the dissolved monomeric form, orthosilicic acid, which is what shows up in your blood and ends up in urine within hours. That's why food source matters: beer and silicon-rich mineral water deliver silicon already dissolved, with about 50% absorption; the polymeric silica in horsetail tea is mostly inert Sripanyakorn 2004 Sripanyakorn 2009.

Does it actually work?

The cleanest population signal comes from Framingham. Across roughly 2,800 adults, the people in the top quarter for daily silicon intake — call it about 40 milligrams a day, hit by anyone eating wholegrains regularly — had hip bone density about 10% higher than the bottom quarter. The effect held for men and for women who hadn't gone through menopause. For postmenopausal women not on hormone replacement, it disappeared Jugdaohsingh 2004. The Aberdeen study, a few thousand women age 50 to 62, confirmed the pattern: the silicon–bone signal showed up only in women whose oestrogen was still active Macdonald 2012. Silicon seems to need oestrogen as a partner. Take the partner away and the lever stops moving.

For skin, hair, and nails, the human evidence is two double-blind trials from the same Belgian group, both at 10 mg of supplemental silicon a day. In the first, 50 women with sun-damaged facial skin took the supplement for 20 weeks; their skin's elasticity and surface texture improved measurably, and their hair and nail brittleness scores dropped, both significantly versus placebo Barel 2005. In the second, 48 women with thin hair took it for 9 months; the placebo group's hair lost about a tenth of its elasticity and break strength; the silicon group lost about a third of that. Their hair shafts also got measurably thicker Wickett 2007.

Two caveats on those trials are worth flagging. Both were run by the company that makes the supplement, both used the same proprietary formulation, and nobody else has reproduced them. They are not no evidence. They are not strong evidence either. The honest tier is "the effect is probably real and modest."

What the low-silicon decade costs you

Picture two versions of a 55-year-old who eats well in every other way. One has spent the decade on a no-grain pattern — keto-leaning, refined-carb-light, soft tap water, not a beer drinker. Daily silicon intake hovers around 10–15 milligrams. The other has kept oats most mornings, a wholegrain bread sometimes, a beer on a Friday. Daily intake sits around 35 milligrams.

At 35, neither of them noticed. At 45, the first one's nails started chipping a bit sooner, the second one's didn't, and neither connected it to anything. At 55, the first one's hip bone density is sitting roughly where the bottom-quartile-intake hips sit in the population data — about a tenth lower than their peer's at the same lifestyle Jugdaohsingh 2004. That gap is not the difference between independent and bedridden; the bigger drivers of that are training, protein, and not falling. But it is a real fraction of the gap, and it cost the second person nothing to claim.

The other thing the low-grain decade costs is harder to see directly: the bottle. The version of this reader who didn't read this entry signed up for a $30-a-month silica supplement when the bowl of oats they'd talked themselves out of would have done most of the job. The honest cost of getting this entry wrong is not collapse; it is paying for the thing your kitchen already covered.

How to actually get the dose

The target most of the bone research used as the "high intake" benchmark is somewhere between 25 and 40 milligrams a day. Three honest routes get you there.

Food, the default. Oats, wholegrain bread, brown rice, wholewheat pasta, and bran-bearing cereals are the heavy lifters. A bowl of oats and a couple of slices of wholegrain bread covers the bulk of a day's intake; vegetables and unrefined grains the rest. People who eat this way without trying are already at or above the high-intake benchmark Powell 2005.
Beer, the absorption champion. A 330 mL bottle of standard lager delivers around 6 to 10 milligrams of silicon, in the most absorbable form your gut sees from any food — about half of it reaches the bloodstream, comparable to a pure orthosilicic acid solution Sripanyakorn 2004. This is genuinely interesting biochemistry. It is also not a license to drink more beer; once you go past one or two a day the alcohol load on bone, sleep, and everything else swamps the silicon benefit.
Mineral water, the low-fuss top-up. Silicon-rich brands — Fiji at ~45 mg/L, Acilis/Spritzer at ~35 mg/L, Volvic at ~15–20 mg/L — deliver silicon already dissolved as orthosilicic acid. A litre covers a whole day's high-intake target. This is the cleanest top-up route for a low-grain, no-beer diet.
Supplement, if the diet really won't budge. Choline-stabilised orthosilicic acid (the form the clinical trials used) at 6 to 10 mg of silicon a day matches the trial doses. Look for the milligram count of elemental silicon on the label, not the total compound weight. Skip horsetail-extract products — most of their silicon is polymerised and your gut doesn't absorb it Sripanyakorn 2009.

One number worth keeping in mind: the European food safety body could not find a level at which silicon caused harm, even at intakes far above anything achievable from food EFSA 2004. The UK expert group landed on 700 mg per day from supplements as a safe ceiling. You will not get there by accident.

What the wellness market gets wrong

Three things popularly attributed to "silica" run ahead of what the evidence supports.

Horsetail tea is mostly inert. Horsetail is the original plant-based silica supplement and still common on shelves. Most of its silicon is locked into polymerised plant phytoliths — particles your gut doesn't break apart or absorb Sripanyakorn 2009. If the label says "silica from horsetail" with no mention of orthosilicic acid or a stabilised form, treat the silicon claim sceptically.

Silica does not regrow hair or reverse grey. The Belgian trials show preserved hair-shaft thickness and elasticity in women with already-fine hair over 9 months — that is a holding action, not regrowth Wickett 2007. Anyone selling silica for male-pattern baldness or for greying is selling a story without trial data.

Silicon-rich water is not Alzheimer's therapy. There is a real and interesting line of work — silicon binds dissolved aluminium in the gut and helps flush it through the kidneys — and a small pilot in 15 people with Alzheimer's showed reduced aluminium burden and modest cognitive improvements in 3 of them after 12 weeks of drinking high-silicon mineral water Davenward 2013. That is a hypothesis worth following, not a treatment. Anyone framing it as cure-grade is overreaching.

Where the bone signal does, and does not, show up

Both big population studies — Framingham in the United States, Aberdeen in Scotland — found the same uncomfortable pattern. The link between dietary silicon and stronger hips held in men, and in women who were still pre-menopausal or on hormone replacement. In post-menopausal women who weren't on hormone replacement, the link disappeared Jugdaohsingh 2004 Macdonald 2012.

That is a hard finding for the case for silicon, because post-menopausal women are the population most at risk for osteoporosis — exactly the people you'd want a bone nutrient to work for. The most likely explanation is that silicon helps build new bone, and once oestrogen drops, the rate of bone breakdown swamps anything new bone formation is doing. Silicon is probably a permissive co-factor, not an active brake on resorption.

The practical reading: if you are a man or a pre-menopausal woman, the bone case applies to you. If you are post-menopausal and not on hormone replacement, the silicon effect on bone is unproven for your situation — and the bigger bone levers (resistance training, adequate protein, vitamin D, and the bisphosphonate or hormone-therapy conversation with a clinician) deserve the attention.

The skin, hair, and nail findings come from trials run in women only, mostly between 40 and 65. Generalisation to men is plausible by mechanism — the same collagen biology is doing the work — but it hasn't been tested.

What else might do the same job

Silicon is a small lever next to bigger ones. If you're trying to hold onto bone density into your sixties and seventies, weight-bearing exercise, adequate protein, and the calcium-plus-vitamin-D base each move the needle further than silicon does. If you're trying to hold collagen integrity in skin, oral collagen peptides have more direct human evidence at the elasticity endpoint than silicon does, and topical retinoids have more direct evidence at the visible-photo-aging endpoint than either. For hair thinning that is actually male- or female-pattern loss, minoxidil and finasteride are working on a completely different mechanism and beat silicon at that endpoint by a wide margin.

The case for silicon is not that it competes with any of these. It's that it stacks underneath them at near-zero cost and effort. The wholegrain breakfast is already on the table for ten other reasons. Treat the silicon contribution as a small bonus you were getting anyway, not the reason you eat the oats.

What changes if you start

Honest answer: nothing this week. Silicon is silent biology — absorbed, used as a co-factor, cleared in urine, no felt signal day to day. The window where any of this becomes legible is months and years, not days.

At six months, if you'd been on the low-intake side and you've moved to the high-intake side, you've moved your urinary silicon (and presumably your tissue silicon) into a range that looks like the population's bone-favourable quartile. Nothing visible yet. The trial that measured a bone-formation marker in osteopenic women needed a full year on a supplement before that marker shifted significantly Spector 2008.

At one to two years, the skin-and-hair endpoints from the Belgian trials start to apply. Skin elasticity holds slightly better than it would have; hair shafts are slightly thicker than they would have been; nails are less brittle Barel 2005 Wickett 2007. None of this looks like a transformation — your hairdresser will not stop you on the street. The hairdresser of the version of you who didn't bother might notice that yours is holding up a bit better than expected.

At a decade or two, the bone story. The Framingham gap between the bottom and top intake quartiles was about a tenth of hip bone density Jugdaohsingh 2004. That kind of fraction is not what determines whether you fall at 75. It is a small piece of what determines whether the fall breaks anything. The payoff is a slightly straighter bone-density curve, on top of all the other things — strength training, protein, sunlight, sleep — that determine how that decade actually goes.

The non-biological payoff: you keep the $30 a month you would have spent on the silica supplement, and you stop being the target market for the next bamboo-extract bottle that lands in your feed.

Adjacent reading

If the bone-density thread is what brought you here, the heavier-lifting entries on resistance training, dietary protein, calcium and vitamin D, and (for post-menopausal women) hormone replacement and bone-screening cadence are where most of the actual force is. If the skin and hair thread is what brought you here, oral collagen peptides, topical retinoids, and (for hair) the minoxidil / finasteride conversation each carry more direct evidence than silicon does. The aluminium-clearance line of work that comes up around silicon-rich water is interesting and unresolved; if it matures, it belongs in its own entry alongside cognitive-decline prevention rather than here.

Substance + claimed effects

Dietary silicon is the form of the element silicon (Si) that humans actually ingest — almost always as orthosilicic acid (OSA, Si(OH)₄), the small, soluble, monomeric form that crosses the gut epithelium. It enters the diet via plant tissues (cereal grains, especially oats and wholegrains, and the bran/husk of rice and wheat), via beer (where the malting and brewing process leaches Si from hulls into the wort as OSA), via certain mineral waters, and via supplements: stabilised orthosilicic acid (ch-OSA, monomethylsilanetriol, OSA-vanillin complex), hydrolysed bamboo extract, and the older horsetail extracts Sripanyakorn 2004 Powell 2005 Sripanyakorn 2009. Claimed effects, which the article will cover holistically, cluster in four places: (1) bone density and bone-formation markers (the bulk of the evidence base); (2) skin elasticity and visible photo-ageing surrogates; (3) hair tensile strength and thickness; (4) nail brittleness. A secondary claim — silicon binds and accelerates the urinary clearance of aluminium, with possible relevance to cognition — has thin trial data and lives at the edge of scope.

Evidence by addressing question

Mechanism

Two converging mechanisms carry the literature. First, OSA stimulates collagen type 1 synthesis in human osteoblast-like cells in vitro at low micromolar concentrations; the effect is abolished when prolyl hydroxylase activity is blocked, implicating Si as a co-factor (or modulator) in collagen post-translational hydroxylation — the step that lets the triple helix stabilise Reffitt 2003. Second, OSA promotes osteoblast differentiation markers (alkaline phosphatase, osteocalcin) in the same in-vitro systems Reffitt 2003. The same prolyl/lysyl hydroxylase logic carries over to dermal fibroblasts and the connective tissue of skin, hair shaft and nail plate, which all rely on properly cross-linked collagen and (for hair/nail) keratin matrix integrity Jugdaohsingh 2007. Silicon is also a known ligand for aluminium: dissolved orthosilicic acid binds Al³⁺ as hydroxyaluminosilicate, which is poorly absorbed and renally cleared — the basis of the "aluminium chelator" claim Davenward 2013.

Evidence — bone

The strongest population-level signal is the Framingham Offspring analysis: in 1,251 men and 1,596 women (age 30–87, cross-sectional), dietary Si in the highest quartile (≥ ~40 mg/day) was associated with ~10% higher hip BMD vs the lowest quartile (≤14 mg/day) in men and premenopausal women, but the association vanished in postmenopausal women not on HRT Jugdaohsingh 2004. The Aberdeen Prospective Osteoporosis Screening Study (n=3,198 women age 50–62) confirmed the oestrogen-dependence: a significant Si–BMD association at the femoral neck only in oestrogen-replete women (premenopausal, or on HRT) Macdonald 2012. On the RCT side, Spector et al. randomised 184 osteopenic women to 1,000 mg Ca + 20 µg D3 plus ch-OSA at 3, 6, or 12 mg Si/day or placebo for 12 months; the type-I collagen formation marker PINP rose significantly at the 6 and 12 mg doses vs placebo, and a post-hoc subgroup analysis showed a femoral-neck BMD signal at 6 mg Si in women with baseline T-score < -1, but the dose-response was not clean and primary BMD endpoints did not reach significance Spector 2008. Effect-size summary: observational hip-BMD differences ~5–10% across the quartile range; trial-level bone-formation marker changes ~15–20%; no fracture data.

Evidence — skin, hair, nails

Two double-blind RCTs by the Antwerp/Brussels group anchor this evidence base, both using ch-OSA at 10 mg Si/day. Barel 2005 randomised 50 women with photodamaged facial skin for 20 weeks; the active arm showed significantly improved skin micro-relief and viscoelasticity (cutometer) and significantly lower hair- and nail-brittleness scores vs placebo Barel 2005. Wickett 2007 randomised 48 women with fine hair for 9 months; the placebo group's elastic gradient fell 11.9% and break-load fell 10.8%, while the ch-OSA group fell only 4.5% (elasticity, significant) and 2.2% (break-load, non-significant) — and hair-shaft cross-sectional area rose significantly in the ch-OSA arm only Wickett 2007. Both trials are small, single-formulation, sponsored by the ch-OSA manufacturer, and have not been independently replicated. No head-to-head data for monomethylsilanetriol, OSA-vanillin, bamboo extract, or horsetail at human BMD or visible-skin endpoints; bioavailability ranks ch-OSA > beer > OSA-vanillin > bamboo extract > horsetail (which is largely insoluble polymeric silica) Sripanyakorn 2009.

Protocol

Two clean food-based routes get a 25–40 mg/day intake. Cereal-heavy diets — oats, wholegrain bread, brown rice, wholewheat pasta — drive intake in most populations (Western diet median ~20–30 mg/day men, 18–25 mg/day women, with vegetarians and high-grain eaters at the upper end) Powell 2005 Jugdaohsingh 2007. Beer is the most absorption-efficient dietary source: a 330 mL bottle of standard lager delivers ~6–10 mg Si with ~50% bioavailability (urinary recovery), comparable to a pure OSA solution Sripanyakorn 2004. Silicon-rich mineral waters (Fiji ~45 mg/L, Spritzer/Acilis ~35–55 mg/L, Volvic ~15–20 mg/L) deliver OSA already monomeric and highly absorbable. Supplemental ch-OSA was given at 10 mg Si/day in the dermatology trials and 6 mg Si/day in the bone-marker trial. The UK Expert Group on Vitamins and Minerals set a safe upper level of 700 mg Si/day from supplements EFSA 2004.

Contraindications

None established at intake levels achievable from diet, water, or labelled supplements. Si is not retained in soft tissue; excess is renally cleared. The EFSA 2004 opinion concluded that no UL could be set from dose-response data because no convincing toxicity has been demonstrated at any tested intake EFSA 2004. Practical caveats are about form: silica gels and other polymeric particulates (E551 food additive at very high acute doses) are not the same molecule and not the same risk profile. People with kidney disease should be cautious about high-dose mineral-water protocols (volume, mineral load), not Si per se.

Misconceptions

Three things popularly attributed to "silica" overshoot the evidence. (1) Horsetail tea is widely sold as a "silica supplement"; most of its Si is polymerised, non-bioavailable phytoliths, and human absorption data are poor Sripanyakorn 2009. (2) "Silica reverses grey hair / regrows hair" — there is no controlled human evidence; the Wickett trial shows hair-shaft thickening and elasticity preservation, not regrowth. (3) "Drinking silicon water cures Alzheimer's" — the Davenward 2013 pilot (n=15) showed reduced urinary aluminium burden over 12 weeks and clinically-relevant cognitive improvement in 3/15; this is a hypothesis-generating signal, not therapy Davenward 2013.

Audience

The bone evidence skews male-and-premenopausal-female. Postmenopausal women without HRT do not show the Si–BMD association in either Framingham or Aberdeen, suggesting Si's bone effect needs oestrogen to act as a co-factor (or that bone turnover dominates over Si-mediated formation in the oestrogen-deficient state) Jugdaohsingh 2004 Macdonald 2012. The dermatology trials enrolled women only (photodamaged, age ~40–65 in Barel; fine-haired, age 25–55 in Wickett); generalisation to men is plausible by mechanism but not directly tested.

Alternatives

For bone density: weight-bearing exercise, adequate calcium + vitamin D, protein intake, and (when indicated) bisphosphonates dominate Si's effect size by an order of magnitude. For collagen-mediated skin/hair/nail goals: oral collagen hydrolysate (more direct human evidence at the skin elasticity endpoint), topical retinoids (skin), minoxidil/finasteride (hair, by a different mechanism), biotin (nails, only in true deficiency). Silicon's appeal is as a low-cost, low-effort add-on to diet, not as a primary lever.

Failure modes

The common screwup is buying polymeric silica supplements (horsetail extract, "silica gel" capsules with no spec on form) and assuming they deliver bioavailable Si Sripanyakorn 2009. The second is taking ch-OSA expecting BMD changes within 12 weeks; the bone-formation marker signal needed 12 months in Spector 2008 and the BMD signal was subgroup-only Spector 2008. The third is reaching for beer as a "bone tonic"; the dose response of alcohol on bone is U-shaped to negative beyond 1–2 drinks/day, and the alcohol harm dominates Si's benefit at any intake that would matter.

Practicalities

Achievable Si intake from food costs nothing — a daily bowl of oats, wholegrain bread, and brown rice puts most people at 25–35 mg/day Powell 2005. ch-OSA supplements (BioSil, JarroSil) cost ~$20–30 for a 60-day supply at 10 mg Si/day. Silicon-rich mineral water is the most expensive route (~$2–4/L), but is the cleanest way to top up a low-Si diet (low-grain, no beer). No prescription needed; no monitoring needed.

History

Silicon's biological role was first proposed by Edith Carlisle (UCLA) and Klaus Schwarz (NIH) in the early 1970s from chick and rat deficiency models — both groups showed Si-deficient diets produced skeletal abnormalities reversed by Si repletion. The role in humans remained essentially unstudied until the Framingham team's quartile analysis in 2004 created the first population-level signal, after which the Brussels group ran the cosmetic trials and Spector ran the osteopenia trial Jugdaohsingh 2007.

Stakes

The substance's absence in a low-grain, beer-free, soft-water diet (the modernised, refined-carb, "no-grains" pattern) drops Si intake to ~10–15 mg/day, which sits below the Framingham low quartile that scored worst for hip BMD Jugdaohsingh 2004. The effect over decades is small per year but cumulative — the population-level BMD gap between low- and high-Si quartiles (~5–10% at the hip) is on the same order as the gap good calcium + D supplementation closes. For skin and hair, the absence is invisible in any single week and roughly invisible at scale — Si is a permissive co-factor for collagen integrity, not a transformative dermatology lever.

Payoff

Honest payoff is modest and slow. Bone-formation marker changes need months; visible skin and hair changes in the Brussels trials needed 20 weeks (skin) and 9 months (hair). The realistic claim is preservation of trajectory — collagen integrity holds slightly better, hair shafts thin slightly slower, hip BMD declines slightly less — rather than reversal. Si is a small, cheap, low-effort additive lever that earns its place by stacking with the bigger ones (training, protein, calcium, vitamin D, sunlight), not by carrying any of them.

Credibility range

Optimist case. Si has a real mechanism (prolyl hydroxylase co-factor / collagen synthesis stimulator at micromolar OSA) demonstrated in human cell lines Reffitt 2003. Two large independent observational cohorts (Framingham, Aberdeen) show the predicted Si–BMD association in oestrogen-replete adults, with effect sizes that, if causal, would be clinically meaningful Jugdaohsingh 2004 Macdonald 2012. A 12-month RCT moved a bone-formation marker in the predicted direction Spector 2008. Two double-blind dermatology RCTs moved skin and hair endpoints in the predicted direction Barel 2005 Wickett 2007. The intervention is cheap, food-attainable, safe at any plausible intake, and lines up with a vegetarian/wholegrain dietary pattern that is independently good. The honest read on the totality: a small but real positive lever, especially valuable as a free add-on.

Skeptic case. Si is not classified as essential by EFSA or the IOM; no biochemical requirement has been demonstrated and no deficiency syndrome has been described in humans EFSA 2004. The Framingham and Aberdeen associations are cross-sectional, confound with overall diet quality (wholegrain eaters do many other things right), and absent in the population most at risk for osteoporosis (postmenopausal women without HRT). The single bone RCT missed its primary BMD endpoint; the bone-formation marker signal had no clean dose-response and the subgroup BMD signal is post-hoc Spector 2008. Both Brussels dermatology trials were sponsor-run, single-site, and use a proprietary formulation; no independent replication exists. Effect sizes on visible endpoints are modest. The aluminium-chelation Alzheimer's pilot is n=15 with no control group Davenward 2013. The honest read on the totality: a plausible mechanism with consistent-but-thin human signal, none of it strong enough to call settled.

Author's call. The mechanism is solid; the human signal is consistent but small; the intervention is essentially free and safe. Treat Si as a real, modest, low-effort lever — score the bone/skin/hair effects honestly at "small contribution" tier, the evidence at "preliminary but plausible mechanism" tier, and the controversy as low (no field fight: most clinicians are neutral or mildly positive). The article lands as "eat wholegrains and have the occasional beer, supplement only if your diet is genuinely Si-poor."

Stakeholder + incentive map

Commercial: Bio Minerals NV (Belgium) developed ch-OSA and sponsored the Barel, Wickett, and Spector trials; bamboo extract supplement manufacturers; the Brewers Association (incentivised to amplify "beer is good for bones").
Professional: minimal — Si is not in standard bone-health or dermatology guidelines (USPSTF, NOF, AAD, NICE). The literature is led by a small group at the MRC HNR (Powell, Jugdaohsingh) and the Brussels collaborators.
Cultural: low-key wellness/longevity adjacent — "silica beauty" content in women's magazines and bamboo-extract supplement marketing. Smaller than the collagen-peptide market it competes with.
Skeptic: nutrition science mainstream defaults to "not essential, no UL, no recommendation," which is technically correct and leaves the substance under-studied.

Population variability

Sex/hormonal status: Si–BMD effect appears confined to men and oestrogen-replete women in both Framingham and Aberdeen Jugdaohsingh 2004 Macdonald 2012. Postmenopausal women without HRT are the population most at risk for osteoporosis and the one for which Si shows no benefit signal.
Diet pattern: vegetarians, wholegrain eaters, and beer drinkers run at the high end of Si intake (35–60 mg/day); refined-carb, low-grain, low-beer diets (paleo, keto, no-grain Mediterranean variants) run low (10–20 mg/day) Powell 2005.
Age: intake declines with age (less beer, less wholegrain), and absorption may decline with reduced gastric acid in the elderly.
Geography: water supply matters more than usually assumed — populations served by silicon-rich aquifers (volcanic / sedimentary rock geologies) get a notable additional baseline.

Knowledge gaps

The decisive trial would be a Si-only (no calcium/D3 co-administration) 24-month RCT with BMD and fracture endpoints in pre- and postmenopausal women, powered to clarify the oestrogen interaction — this does not exist. Independent replication of the Brussels dermatology trials (different sponsor, different formulation arm) does not exist. Head-to-head bioavailability and clinical-endpoint data comparing ch-OSA, monomethylsilanetriol, OSA-vanillin, bamboo extract, and horsetail are sparse. The aluminium-clearance / cognition signal warrants a properly powered RCT and currently has only a 15-person pilot. And the dietary-pattern confounder (wholegrains correlate with many other healthy behaviours) cannot be ruled out without a clean Si-only intervention.

Brief vs scope. The brief named bone density, skin elasticity, hair, and nails. All four are covered in the body, mostly through evidence and payoff. The bamboo-extract framing in the brief is addressed in misconceptions (largely a marketing form, mostly inert if not stabilised to orthosilicic acid) and not given its own positive treatment because the bioavailability evidence does not support it.

Hard scoping calls.

The silicon–aluminium chelation / Alzheimer's thread (Davenward 2013 pilot, Exley group's wider work) is parked in misconceptions as a sceptic-frame correction rather than developed positively. The n=15, uncontrolled pilot does not earn a positive-lever framing in this entry, but it is real and interesting enough that suppressing it entirely would leave readers who encountered the claim elsewhere with no honest reading. If a properly powered RCT lands, this warrants its own entry under cognitive-decline prevention.
The post-menopausal-without-HRT null is given a full audience section because pretending otherwise would be dishonest to a population that is the natural target for any "bone nutrient." This costs the entry some of its applicability impact and is the right call.
The bone-RCT (Spector 2008) missed its primary BMD endpoint; only a bone-formation marker (PINP) moved cleanly. The article reflects this honestly rather than leaning on the post-hoc subgroup BMD signal.

Rating difficulties.

The beauty and longevity dimensions were the closest calls. The bone signal in oestrogen-replete adults is consistent across two large cohorts, which could justify a 3 on longevity; the absence in postmenopausal women without HRT and the missed RCT primary endpoint pulled it to 2. Similar logic on beauty_direct and beauty_cumulative: real, replicated-within-one-sponsor, modest effect sizes, unreplicated independently — honest tier is 2 each.
evidence: 2 rather than 3 because the human trial base is small, single-sponsor on the dermatology side, and the bone RCT missed its primary endpoint. The mechanism is solid; the trial yield is preliminary.
controversy: 1 because the field is not actively fighting — silicon is under-studied rather than disputed.
health_short_term: 1 because the effect is silent in any felt-experience sense within weeks.

Future-link candidates. collagen-peptide-supplementation, calcium-vitamin-d-bone, resistance-training-for-bone-density, topical-retinoids, aluminium-cognitive-decline (if the chelation literature matures), hrt-bone-density-postmenopause.

Separate-entry candidates. The silicon-rich-mineral-water / aluminium-clearance line, if any future RCT confirms the cognitive signal. The beer-bone protective effect more broadly (silicon being only one mechanism among several proposed) could warrant its own entry that nets out the alcohol downside more rigorously than this entry's protocol section needed to.