Taurine
Verdict: Suggestive — leaning toward downgrade
Last reviewed: 2026-04-24 Triangulated against
anchor: NMN (Suggestive)
TL;DR
The 2023 Yadav et al. Science paper showing taurine extends mouse
lifespan by ~10-12% generated enormous attention. The follow-up has been
notably negative: a 2025 Science analysis found taurine does
not reliably decline with age — a foundational premise of the
supplementation rationale — and large interindividual variation
undercuts the simple "replenish what's lost" framing. No human RCTs
exist. Verdict: Suggestive, with significant downgrade
risk.
What it is
A sulfonic acid (technically not an amino acid by strict definition;
commonly called a "conditionally essential amino acid"), abundant in
animal tissues and present in many energy drinks. Synthesized
endogenously from cysteine. Sold as supplement at 0.5-3g/day for
longevity-curious users.
Proposed mechanism
Hypothesized: taurine declines with age → restoring tissue taurine
improves mitochondrial function, reduces oxidative stress, restores
tissue homeostasis. Specific mechanistic claims include effects on
calcium signaling, bile acid metabolism, mitochondrial function,
neurotransmitter modulation.
Confidence: Plausible for individual mechanisms (taurine is
biologically active); Hypothetical for the specific aging claim
— the foundational premise (age-related decline) is now contested.
Evidence ladder
Invertebrate (T5)
Taurine extended healthspan in C. elegans in Yadav 2023. Modest
effects.
Mouse / rat (T4)
- Yadav et al. 2023, Science — taurine
supplementation extended median lifespan ~10-12% in mice (sex-dependent
magnitude); improved various healthspan markers. Single-lab; published
with high prominence.
- Replication status: limited. As of early 2026, no
major independent replication of the lifespan finding has been
published. ITP has not (publicly) tested taurine yet.
- Effect size: 10-12% median lifespan extension in
the original paper.
- Sex / strain / dose: Effects reported in both
sexes; magnitude varied. Strain dependence not extensively
characterized.
NHP (T4)
- Yadav 2023 included rhesus monkey biomarker improvements (~6-month
supplementation): improved metabolic, inflammatory markers. No lifespan
data; healthspan biomarker only.
Human (T0-T2)
- Observational — Yadav 2023 included epidemiological
correlation of higher serum taurine with better aging biomarkers.
Classic dietary-cohort confounding.
- Critical 2025 reanalysis (Science, Fang et al. / NIA
team) — analysis across multiple cohorts found taurine does
not systematically decline with age in healthy
individuals; large interindividual variation; concentrations often
increase during adult life in many cohorts. This directly
undermines a load-bearing premise of the original story.
- No completed human RCTs of taurine supplementation
for aging endpoints. Yadav himself has stated supplementation should
wait for clinical trials.
- Energy drink populations provide a "natural
experiment" with high taurine intake; no clear longevity signal from
those cohorts.
Confounds
- The central premise (age-related taurine decline) is
contested. The 2025 Science follow-up suggests the decline is
not robust.
- Single-lab original study without replication.
- Healthy-user / dietary confounding in the human
observational component.
- Marketing pressure — taurine is cheap and widely
available; supplement industry adopted it rapidly post-Yadav 2023, ahead
of replication.
Conflict of interest scan
- Yadav 2023 was independent academic work.
- The 2025 critical follow-up was also independent.
- Supplement industry interest exists but didn't generate the primary
data.
- No discount applied; the contestation here is scientific, not
COI-driven.
Human translation
Honest read: taurine got a high-profile paper, the science press
hyped it, and follow-up work has weakened the central premise. There is
no human RCT evidence. Anyone supplementing taurine for longevity right
now is doing so on the basis of one mouse study whose foundational
premise (taurine declines with age) has been challenged.
This does not mean taurine "doesn't work" — the original mouse data
is real and the lifespan extension number was meaningful. It means the
case is much weaker than the popular narrative suggests, and the 2-year
trajectory points downward.
Calibrated verdict
Suggestive — leaning downgrade. Per methodology,
Suggestive is "T3 or T4 evidence with replication, but human data absent
or null." Taurine fits T4 (single-lab mouse) without replication, no
human data. It's at the bottom of the Suggestive band, not the top.
Compared to NMN (Suggestive), taurine has
less replication (NR has been ITP-tested and failed; taurine
hasn't been ITP-tested at all), less human RCT volume, and a
contested foundational premise (NMN's NAD+-decline-with-age
premise is well-established; taurine's decline premise is now
disputed).
Compared to resveratrol (Mostly hype), taurine is
not yet at the "tested at higher tier and failed" stage. ITP testing
would be the decisive next step.
Confidence interval on
verdict
- Could move to Mostly hype if (a) ITP tests taurine
and pooled results are null, OR (b) the 2025 critique of the
age-related-decline premise gains broader acceptance, OR (c) human RCTs
report null on hard endpoints. Plausible within 2 years.
- Could move to Probable only with strong independent
mouse replication AND positive human RCT on a hard or near-hard
endpoint. Unlikely on the trial timeline.
- Most likely 2-year trajectory: stays at Suggestive;
moderate probability of downgrade to Mostly hype.
Open questions
- Q: Is taurine being or going to be ITP-tested? (Add to QUESTIONS.md
for resolver agent)
- Q: Has any major lab independently replicated Yadav 2023's mouse
lifespan finding? (Status as of 2026-Q2)
- Q: Are there subpopulations where taurine genuinely declines and
supplementation might be appropriate (e.g., sepsis recovery, aging cats
— taurine is famously essential in cats)?
- Q: Why did the original Yadav cohort show age-related decline when
the NIA 2025 reanalysis did not? Methodological vs
population-difference?
Sources
Produced under methodology locked 2026-04-24. Triangulated
against NMN anchor.