How much estrone sulfate actually converts back to estradiol — and what that does and does not tell you about choosing one route over another. Built from the primary pharmacokinetic literature with strict confidence tiers and a public audit of the casual-conversation claims that prompted this page.
ρ = 0.014 · ~1.4 % of circulating E₁S flux appears as circulating E₂
Source: Ruder HJ, Loriaux DL, Lipsett MB. Estrone Sulfate: Production Rate and Metabolism in Man. J Clin Invest. 1972; 51:1020–1033. DOI 10.1172/JCI106862. Isotope-dilution method in humans. A separate E1S → E1 transfer factor ρ = 0.21 was measured in the same study; Longcope 1972 independently reported E1S → E1 at 0.15 (95% CI 0.12–0.18).
What this number does not say. The 1.4% is a plasma-level transfer factor measured by tracer dilution. It does not bound intracellular E2 formation in target tissues, because intracellular E1S that is taken up via OATP transporters and hydrolyzed by STS may not reappear in plasma at all. So the headline 1.4% is the right answer to "how much circulating E1S reaches the circulating E2 pool"; it is the wrong answer to "how much estrogenic effect does oral E2 exert through the E1S route." The mechanistic story is real, but it can't be reduced to a single percentage.
Especially do not translate this to "oral is really 15–25% bioavailable when you count E1S." That number was an arithmetic shortcut, not a published PK metric. The literature offers no validated single bioavailability figure that includes the reservoir, and using one papers over the tissue-specificity that actually drives the effect.
The trans-specific PK study with modern LC-MS/MS is Doll 2022 (n = 10, single-dose crossover). Older comparators (Price 1997, Kuhl 2005, Zumenon SmPC) extend the picture to multi-dose steady state and additional doses.
| Cohort & method | Route / dose | E₂ Cmax | Tmax | AUC / notes | Conf |
|---|---|---|---|---|---|
| Doll 2022, n=10 trans women, LC-MS/MS PMID 34781041 |
1 mg oral E₂ | 35 pg/mL | 8 h | AUC0–8 reference; E₂:E₁ ratio 0.7 ± 0.4 | C2 |
| Doll 2022, n=10 trans women, LC-MS/MS | 1 mg sublingual E₂ | 144 pg/mL | 1 h | AUC0–8 1.8× oral; E₂:E₁ ratio 1.1 ± 1.0 | C2 |
| Price 1997 (via Kuhl 2005) DOI 10.1016/S0029-7844(96)00513-3 |
1 mg sublingual | ~450 pg/mL | < 1 h | E₁ 165 pg/mL; E₂ fell to ~85 pg/mL by 3 h | C2 |
| Price 1997 (via Kuhl 2005) | 0.25 mg sublingual | ~300 pg/mL | < 1 h | Demonstrates supra-linear early peaks at low dose | C2 |
| Zumenon 1 mg SmPC postmenopausal women, steady state |
1 mg oral micronized E₂ | E₂ 48 ± 17 pg/mL E₁ 349 ± 129 pg/mL E₁S 10.5 ng/mL |
E₂ Tmax 3.9 h | AUC0–24: E₂ 751 pg·h/mL · E₁ 5487 pg·h/mL · E₁S 129 ng·h/mL. Cmin: E₂ 20.8 pg/mL · E₁ 146 pg/mL · E₁S 2.51 ng/mL | C2 |
| Kuhl 2005 review PMID 16112947 |
2 mg oral EV, day 21 | E₂ ~80 pg/mL | — | E₁ 4–6× E₂; E₁S ~200× E₂ | C2 |
| Kuhl 2005, transdermal | 50 µg/day patch | E₂ 30–60 pg/mL | steady | E₁:E₂ ratio ~1:1 (vs ~5:1 oral) | C2 |
| Kuhl 2005, injectable | 4 mg EV IM, day 2 → day 10 | E₂ ~400 → 150 pg/mL | depot | Not comparable to single-dose tablet Tmax | C2 |
Interpretation. Sublingual delivers about 4× the E₂ Cmax of swallowed oral at the same nominal dose, and roughly 1.8× the early-window AUC in the only trans-specific LC-MS/MS study. That advantage is narrower than the "sublingual feels stronger" Cmax-only narrative implies — once you let the curves integrate over 24 h, the gap further compresses because sublingual decays fast.
The E₁ / E₁S story is route-specific. Oral at 2 mg generates E₁ at 4–6× E₂ and E₁S at ~200× E₂ by mass (Kuhl). Sublingual still creates E₁/E₁S (some is swallowed; some is mucosal-metabolized) but at lower ratios. Transdermal patches generate roughly 1:1 E₁:E₂. Injectable depots run with low E₁/E₁S relative to oral.
E₁S is the sulfated, albumin-bound, hydrophilic conjugate of estrone. It is weakly or not directly estrogenic until hydrolyzed by steroid sulfatase (STS) to E₁, after which reductive 17β-HSD enzymes can convert E₁ to E₂. Oral E₂ creates a large E₁S pool because intestinal and hepatic metabolism conjugates much of the absorbed dose before it ever sees a receptor.
| Quantity | Value | Confidence | Source |
|---|---|---|---|
| E1S plasma half-life | 10–12 h | C2 | Kuhl 2005; Lobo & Cassidenti 1992 |
| E2 / E1 unconjugated plasma half-life | 20–30 min | C2 | Kuhl 2005 |
| E1S Cmax, 1 mg oral E2 steady state | 10.5 ng/mL | C2 | Zumenon SmPC |
| E1S Cavg, 1 mg oral E2 steady state | 5.28 ng/mL | C2 | Zumenon SmPC |
| E1S, 1 mg oral E2, 15 mo follow-up (RIA) | 38.8 ng/mL | C3 | Slater 2001 — PMID 11355042 |
| E1S : E2 mass ratio, 1 mg oral steady state | ~166× | C3 | Derived from Zumenon SmPC |
| E1S : E2 mass ratio, 2 mg oral EV | ~200× | C2 | Kuhl 2005 |
| E1S metabolic clearance rate | 157 L/day (range 70–292) | C1 | Ruder 1972 |
Assay caveat that contaminates the whole literature. The Zumenon SmPC says 5.28 ng/mL E₁S Cavg on 1 mg oral; Slater 2001 says 38.8 ng/mL on 1 mg oral after 15 months. That's a 7× difference for nominally the same dose. The dominant explanation is method (LC-MS/MS vs RIA) and study design, not biology. Older RIA-derived steroid numbers — especially for conjugates like E1S — can run several-fold high vs modern LC-MS/MS, and you cannot blend the two casually in a single argument.
E₁S doesn't activate the estrogen receptor itself. To produce a biological effect via the reservoir, this whole chain has to run inside a target tissue:
So whether a high-E₁S serum number produces a real local E₂ signal depends on three rate-limiting things in that specific tissue: transporter expression, STS activity, and the HSD17B1/2 balance. A tissue can have plenty of circulating E₁S available and still produce minimal local E₂ if any of those gates is closed.
| Tissue | STS / uptake | 17β-HSD direction | Evidence |
|---|---|---|---|
| Breast (postmenopausal cancer) | STS active; OATP expression documented | HSD17B1 supports E₁ → E₂; HSD17B2 opposes | C2 |
| Breast (normal adult) | STS present; quantitative tissue:plasma E₂ less skewed than tumor data | Balance less characterized | C3 |
| Endometrium | STS + HSD17B important in endometriosis/cancer biology | HSD17B1/7 activate; HSD17B2 inactivates | C2 |
| Bone | E₁S "major source of local estrogen formation" in ex vivo human bone (PMID 15356081) | Osteoblasts can form E₂ from E₁S | C3 |
| Brain | STS expressed; neurosteroid sulfate biology real | Outcome relevance for HRT unquantified | C4 |
| Adipose / skin | STS broadly expressed | Less quantitative HRT data | C3 |
| Liver / intestine | Major sulfation / conjugation site | HSD17B2 + SULT1E1 push oral E₂ → E₁ / E₁S | C2 |
Breast (the thing most trans HRT cares about). Postmenopausal breast cancer tissue can run E₂ at 10–20× plasma and E₁ at 2–10× plasma in some studies — this is where the "intracrine dominance" framing comes from. But it is a malignant-tissue finding. In the closest normal breast measurement (Lønning 2015, LC-MS/MS), median serum E₂ ≈ 29.7 pg/mL and median breast tissue E₂ ≈ 30.6 pg/g in cycling women — essentially the same. The strongest transfemme cohort signal goes the other direction: ENIGI Amsterdam saw oral E₂ users with ~7× higher estrone than transdermal users, but no association between estrone and 12-month breast development, and oral vs transdermal breast development did not differ clearly (de Blok 2022).
Bone. The one tissue with the most direct support for the "E₁S as major local source" story (Muir 2004, PMID 15356081). Plausible relevance to bone-density preservation on HRT, but not quantified head-to-head across routes in trans cohorts.
Brain. STS is expressed in multiple regions and neurosteroid-sulfate biology is real, but no route-specific HRT outcome data ties this to mood, cognition, or anything else worth dosing on.
Liver. The one tissue where route matters in the opposite direction: oral E₂'s high first-pass hepatic exposure drives SHBG up, drives coagulation factors up (VTE risk asymmetry), and pushes hepatic protein synthesis in directions transdermal does not. This is where the strongest non-breast reason to choose route lives.
Route choice.
Monitoring. Standard clinical monitoring still centers on serum E₂ and testosterone. Adding E₁ is sometimes useful for investigating unusual oral-metabolism patterns or assay interference. E₁S is not a validated routine target for trans HRT and chasing it as a dosing signal isn't supported by current cohort data.
Assay choice. Use LC-MS/MS for low E₂ values when possible, especially when oral E₂ creates a high estrone background. Common direct-E₂ immunoassays can read falsely low on oral therapy due to cross-reactivity with metabolites and conjugates (PMC8859944).
| Claim | Verdict | Why |
|---|---|---|
| "Sublingual is just better." | refined | Higher Cmax and higher short-window AUC, but no direct evidence it produces better feminization than oral / transdermal / injectable when overall treatment is adequate. The Cmax advantage is brief. |
| "AUC doesn't matter, only Cmax." | refuted | Cmax explains acute sensations and peak labs; AUC and trough explain sustained receptor exposure and testosterone suppression. Both matter. |
| "E₁S is inactive, so oral is useless." | refuted | E₁S is inactive at ER but is hydrolyzed by STS in target tissues. Ruder 1972 directly measured E₁S → E₂ transfer. The reservoir is real. |
| "Oral is secretly 15–25% bioavailable after E₁S." | refuted | Not a published PK metric. The plasma transfer factor is 1.4%, and tissue activation cannot be compressed into a single "effective bioavailability" number. |
| "High E₁S predicts breast growth." | unsupported | No direct trans cohort measuring E₁S and breast development located. ENIGI measured estrone and found no association with 12-month breast development. |
| "Aromatase inhibitors prove STS bypass saves oral estrogen." | overstated | STS bypasses aromatase mechanistically, but AIs do reduce breast/tumor estrogens in primary studies. No trans oral-E₂ + AI outcome study supports the bypass claim clinically. |
| "SULT1E1 / STS polymorphisms make some people 2× converters." | weak | Some in vitro SULT1E1 variants shift activity several-fold; no validated trans HRT pharmacogenomic dosing rule exists for SULT1E1 / STS / HSD17B1. |
This page exists because a chat-style answer was made with 13 specific quantitative claims and no citations. Each one was checked against primary literature. Five survived, four needed refining, three were refuted or unverifiable, and one was weakly plausible. The full claim-by-claim log is at research/oral-vs-sublingual-e2/critique-of-prior.md.
| # | Original claim | Verdict | Correction |
|---|---|---|---|
| P1 | E₁S half-life 10–12 h, some to 30 h | confirmed | 10–12 h confirmed (Kuhl). The "to 30 h" anchor wasn't located. |
| P2 | E₂ half-life 1–3 h | refined | Free E₂ shorter still (~20–30 min). Oral terminal profile is prolonged by metabolite recycling, not by E₂ itself. |
| P3 | E₁S trough 3,000–8,000 pg/mL on oral 2 mg | refined | Method-dependent. SmPC 1 mg gives Cmin 2.51 ng/mL; Slater 2001 long-term 1 mg gives 38.8 ng/mL (RIA). 3–8 ng/mL is plausible but not a stable anchor. |
| P4 | E₂ trough 20–40 pg/mL on oral 2 mg | refined | 1 mg SmPC Cmin 20.8 pg/mL; chronic 2 mg ranges 50–120 pg/mL depending on cohort and timing. |
| P5 | E₁S:E₂ ratio 50–300× | refined | 1 mg oral SmPC ratio ~166× by mass; Kuhl says ~200× on 2 mg EV. |
| P6 | Plasma E₁S → E₂ ≈ 1.5% | confirmed | Ruder 1972 gives 0.014. Attribution was correct. |
| P7 | Tissue E₁S → E₂ 5–15%, dominant local source | mixed | STS pathway is important but no universal tissue fraction was verifiable. Dominance varies by tissue and disease state. |
| P8 | Tissue E₂ 5–50× serum in STS-rich tissues | refined | Postmenopausal breast cancer: ~10–20× E₂. Normal breast: tissue ≈ serum (Lønning 2015 LC-MS/MS). |
| P9 | Oral "true" exposure 15–25% with reservoir | refuted | Not a published PK metric. Do not use this number. |
| P10 | E₁S predicts breast development | unsupported | No direct cohort located. ENIGI estrone-vs-breast was null. |
| P11 | AIs don't kill oral E₂ because STS bypasses aromatase | refined | Mechanistically plausible but no trans oral-E₂ + AI study supports the clinical claim. |
| P12 | SL E₁S:E₂ peak 1:1 or 2:1 | refined | Kuhl reports SL E₁:E₂ ratio about 1:3. The specific "1:1 to 2:1" anchor wasn't verified. |
| P13 | SULT1E1 / STS polymorphisms shift pathway 2× | weak | In vitro SULT1E1 variants shift activity several-fold but no validated trans HRT dosing rule. |
Most consequential single error. P9 — the "oral is really 15–25% bioavailable when you count E₁S" line. It was arithmetic dressed up as a published PK metric. It also smuggles in the wrong frame: bioavailability is a plasma metric; tissue activation via the reservoir is a different question that doesn't compress into one percentage. The honest summary is the one in section 02: plasma E₁S → E₂ is 1.4%; tissue effects are real but depend on which tissue and cannot be reduced to a single number.
For a chronological history-of-science walk through 100 years of primary literature — Doisy 1929 → Schachter & Marrian 1938 → Ruder 1972 → STS gene cloning 1987 → HSD17B isoforms → Tebbens 2022 — with verbatim quotes from each paper's abstract, contestation noted, and a candid audit of community lore: see the E₁S history deep dive. Includes a complete table of every human study that administered E₁S and measured serum estrogens (1972–2022).
Full evidence map at darkarts.wiki/research/oral-vs-sublingual-e2/ · 16 claim files, 8 section files, plus the prior-claims critique log.