Endpoint: surrogate — postprandial/short-term plasma TMAO.
Randomized, controlled, dose-response feeding study: healthy adults consumed 0, 1, 2, or 4 hard-boiled eggs. Plasma TMAO rose in a dose-dependent manner in the hours following egg ingestion, alongside choline and betaine, directly demonstrating that eating eggs measurably raises TMAO.
relevance_note: the cleanest “yes, eating eggs does raise TMAO” primary in the pool, with a proper dose-response design — the harm-pathway anchor for thread (ii). Note the rise documented is acute/postprandial, which the chronic-feeding null studies in this pool (DiMarco 2017, Lemos 2018, Wilcox 2021) argue does not translate into a change in fasting/steady-state TMAO.
Correction (2a read): Miller 2014 is an n=6 within-subject dose-response (0/1/2/4/6 yolks), not a larger-n trial.
Methodology
Longitudinal, double-blind, randomized dietary intervention in 6 volunteers, each fed breakfast doses of 0, 1, 2, 4, or 6 egg yolks (crossover, every subject received all 5 doses, >2-week washout between doses) on a standardized low-choline background the day before and day of each dose. Plasma TMAO was sampled before and at intervals up to 24 h after each dose (peak ~6-8 h); 24-h urine TMAO, pre- and 24-h-post serum hsCRP, and plasma oxidized LDL were also measured. NIDDK-funded (Zeisel choline lab; grants DK56350, P30 DK034987). Full text via PMC (PMC4135488). PMID 24944063; AJCN 2014;100(3):778-786.
Results
O-103 - Egg ingestion raised plasma and urine TMAO dose-dependently, with ~11-15% of egg choline converted to TMAO (n=6)
Urinary TMAO ran 236, 433, 634, 944, 900 umol/24h across 0/1/2/4/6 yolks. Direct demonstration that egg choline feeds the TMAO pathway in humans.
Link to originalMethodology
Longitudinal, double-blind, randomized dietary intervention in 6 volunteers, each fed breakfast doses of 0, 1, 2, 4, or 6 egg yolks (crossover, every subject received all 5 doses, >2-week washout between doses) on a standardized low-choline background the day before and day of each dose. Plasma TMAO was sampled before and at intervals up to 24 h after each dose (peak ~6-8 h); 24-h urine TMAO, pre- and 24-h-post serum hsCRP, and plasma oxidized LDL were also measured. NIDDK-funded (Zeisel choline lab; grants DK56350, P30 DK034987). Full text via PMC (PMC4135488). PMID 24944063; AJCN 2014;100(3):778-786.
Link to original
O-104 - TMAO rose significantly only at -=2 yolks and urinary TMAO plateaued between 4 and 6 yolks
Threshold near 2 yolks and saturation beyond ~4 yolks: 236 → 433 → 634 → 944 → 900 umol/24h urinary TMAO.
Link to original
O-105 - The egg-induced TMAO rise was acute-postprandial (peak 6-8 h), a single-meal flux not a fasting steady-state change
Plasma TMAO was sampled before and at intervals up to 24 h after each dose. The design captures postprandial flux, not chronic steady-state TMAO under habitual eating.
Link to original
O-106 - Marked inter-individual variation in egg-to-TMAO conversion (-=4-fold), all subjects converted, men more than women
The 4-fold spread on n=6 flags a large host/microbiome effect on TMAO generation, though no discrete non-converter phenotype appeared.
Link to original
O-107 - No change in hsCRP or oxidized LDL after egg doses despite the TMAO rise
The two measured downstream harm surrogates (inflammation, LDL oxidation) did not move acutely with the TMAO rise; short 24-h window and n=6 limit power.
Link to original
Interpretation
H-42 - Eating eggs (-=2 yolks) raises TMAO production, feeding egg choline into the TMAO pathway
Well-supported for postprandial flux; its extrapolation to chronic exposure under normal (non-low-choline) diets and to hard outcomes is what remains uncertain.
Link to original
H-43 - The egg-induced TMAO rise may be clinically benign because it is postprandial and unaccompanied by inflammatory-oxidative changes
The paper’s more contestable interpretive claim, cutting against the TMAO-harm thread. Uncertain: n=6, only two surrogate harm markers over 24 h, and no hard-outcome data.
Link to original
Arguments
A-39 - The -=2-yolk threshold and 4-6 yolk plateau imply saturating conversion capacity, bounding TMAO exposure from high egg intake
Validity verdict (step 6)
Link to original
approved,checked. Reconstruction: premise is the urinary TMAO dose series (236/433/634/944/900 umol/24h at 0/1/2/4/6 yolks) with a flat, statistically indistinguishable 4→6 interval; conclusion is a concave, saturating exposure-response that bounds/shapes TMAO from high egg intake. Traced the arithmetic: per-yolk increments run ~197 (0→1), ~201 (1→2), ~155/yolk (2→4), then ~-22/yolk (4→6). The fall in marginal increment to ≈0 is directly in the numbers — concavity does not need extra assumptions, so the “saturating dose-response / incremental TMAO falls at high intake” content ischeckedand holds. Considered the undercutting defeater that a urinary plateau could reflect saturation of renal excretion rather than of production, in which case plasma (harm-relevant) TMAO could keep rising while urine flattens, breaking “bounds exposure.” Judged it non-surviving under charity: in healthy young subjects at these modest TMAO levels renal TMAO clearance is not near saturation, so urinary output tracks exposure — a reasonable premise, not completion-by-force. The statement’s hedges (“indicate,” attributing to “conversion capacity” while the body allows gut-TMA and/or FMO3) keep it within what the data license; the mechanism attribution is the softest part but does not affect the exposure-bounding conclusion, which holds whichever step saturates. Scope note honored: bounds/shapes exposure, does not confirm or deny harm. The urinary TMAO series (236, 433, 634, 944, 900 umol/24h for 0/1/2/4/6 yolks) is not linear: the marginal increment falls from ~200 umol per yolk over the 0-4 range to essentially zero (944 vs 900) between 4 and 6 yolks, and the 4- and 6-yolk responses are statistically indistinguishable. A plateau at high substrate load is the signature of a rate-limited step reaching capacity — here the gut-microbial trimethylamine production and/or hepatic FMO3 oxidation. Two consequences follow. First, the exposure-response of TMAO to egg intake is concave: doubling eggs from 3 to 6 does not double TMAO, so the incremental TMAO risk (if any) of very high egg intake is smaller than a linear extrapolation from low doses would predict. Second, the ~2-yolk threshold means a single egg produces no detectable TMAO signal, so any TMAO-mediated risk would concentrate at higher per-meal doses. This bounds and shapes the exposure side of the TMAO-harm hypothesis rather than confirming or denying harm itself.
A-40 - A postprandial TMAO spike with unchanged inflammation-oxidation markers does not establish that egg-driven TMAO causes harm
Validity verdict (step 6)
Link to original
approved,checked. Reconstruction: premises are that the measured rise is a single-meal postprandial excursion (peak 6-8 h, toward baseline by 24 h) against a deliberately low-choline background, and that the two downstream harm surrogates (hsCRP, oxidized LDL) did not move; conclusion is the non-establishment claim that the study shows eggs raise TMAO flux but does not establish that egg-driven TMAO produces cardiovascular harm. Traced step: the harm hypothesis is choline→TMAO→atherosclerosis over years, i.e. a chronic steady-state exposure over a hard endpoint; an acute postprandial flux measured over 24 h is a different quantity and cannot on its own demonstrate the chronic causal chain — so “harm not established” follows straightforwardly. This is a claim about what the evidence fails to show, which is structurally hard to defeat: no undercutting alternative makes a 24-h surrogate-null over n=6 establish multi-year atherogenesis. The statement is correctly modest and, unlike the body’s parenthetical “mildly favouring the benign reading,” it does not assert benignity — it only denies establishment of harm, and the body itself flags the surrogate null as too weak (n=6, 24 h) to prove benignity. So no weaker form is needed. Author-blind: rests on the exposure/endpoint mismatch, not the source. The harm hypothesis for eggs runs choline → TMAO → atherosclerosis over years, which is driven by chronic (fasting/steady-state) TMAO exposure. What this design measures is different: a transient postprandial excursion following a single yolk challenge, with peak at 6-8 h and return toward baseline by 24 h, and it does so against a deliberately low-choline standardized background that maximizes the measurable conversion of the egg’s choline. Neither feature licenses a claim about chronic steady-state TMAO under an ordinary mixed diet, where competing choline sources and adaptation may blunt the egg-specific signal. On the outcome side, the study looked one step downstream and found nothing: hsCRP (inflammation) and oxidized LDL did not differ before vs 24 h after any dose. That absence is weak evidence (n=6, only two surrogates, a 24-h window far too short for atherogenesis), so it cannot affirmatively prove benignity either. Net, the logical content is that “eggs raise postprandial TMAO” and “egg-driven TMAO causes cardiovascular harm” are separate claims, and this study supports the first while leaving the second unestablished, mildly favouring the benign reading via the null on inflammation/oxidation.