Taurine and ageing: the biomarker that collapsed in two years

I do not take taurine. This is not an experience article, it is an article about a story that played out in two years in the same journal, and about what it teaches anyone who builds a stack from headlines.

In June 2023, Science publishes a study that makes taurine a driver of ageing. The reasoning has three steps: its blood level falls with age in mice, monkeys and humans; giving it to mice extends their life; so its lack could be a cause. In June 2025, the same journal publishes repeated measurements in the same individuals. The level does not fall.

I have already told a neighbouring story with vitamin B3: a marker that rises without health following. Here it is more radical. The marker itself did not hold.

2023: a biomarker, some mice, a headline

The study by Singh and colleagues is solid on what it measures. Circulating taurine concentrations decline with age in mice, monkeys and humans. In mice, supplementation increases life span and health span; in monkeys, health span. Mechanistically, taurine reduces cellular senescence, protects against telomerase deficiency, limits mitochondrial dysfunction and DNA damage.

In humans, the study only has correlations: a lower level is associated with several age-related diseases, and the level rises after endurance exercise. The authors say it themselves: clinical trials in humans seem warranted to test whether taurine deficiency might drive ageing.

What the headline did not say is the nature of the human measurement. Comparing the taurine of people aged 30, 50 and 70 at one point in time is a cross-section. It mixes two things: what age does to a person, and what distinguishes generations that did not eat, move or live the same way.

2025: the same people, measured over time

Fernandez and colleagues made the measurement that was missing. Three human cohorts from different regions, monkeys and mice, measured both cross-sectionally and longitudinally, that is, several times in the same individuals over the years. Result: circulating taurine increases or remains stable with age.

And the link between taurine and health, motor function or energy metabolism, varies considerably from one individual to another. The authors' conclusion is measured and final: changes in circulating taurine are not a universal feature of ageing, and its effects probably depend on the temporal and physiological context of each person.

What is left standing: the metabolic trials

The fall of the biomarker does not cancel the controlled trials, which never depended on it. The most recent meta-analysis, 34 randomised trials, finds on taurine a fasting glucose about 6 mg/dL lower, glycated haemoglobin 0.21 points lower, triglycerides 14 mg/dL lower, total cholesterol 12 mg/dL lower, and blood pressure 4.4 mmHg lower systolic and 2.5 diastolic. The most effective range is 1.5 to 3 g per day.

A 2024 meta-analysis, 25 trials and 1,024 participants, at doses of 0.5 to 6 g per day for 5 to 365 days, finds the same orders of magnitude: 4 mmHg systolic, 1.5 diastolic, 6 mg/dL of glucose, 18 mg/dL of triglycerides, nothing on HDL, and no significant adverse effect compared with control groups. The 2020 one, smaller, sets the context: most trials involve people who already have diabetes, liver disease, obesity or metabolic syndrome.

These are real, modest effects, in people already affected, on intermediate markers. Nobody has measured whether those few mmHg and mg/dL translate into events avoided, let alone years of life.

What has never been measured in humans

Life span. No trial. The closest thing to an ageing question is a small Brazilian trial: 24 women aged 55 to 70, 1.5 g per day for 16 weeks. Plasma taurine rises and an antioxidant enzyme, superoxide dismutase, is maintained while it drops under placebo. That is one more marker, in 24 people.

Another trial, in 48 women aged 84 on average, compares exercise, taurine, both, or neither, for 14 weeks. Taurine alone lowers one ratio of inflammatory markers. The cognitive score only improves in the group combining exercise and taurine, and the authors conclude on exercise as the main tool.

The first trial to ask the ageing question with an outcome that comes close has just been registered: 80 healthcare workers, 3 g per day or placebo for six months, with glycated haemoglobin as the primary outcome and the PhenoAge biological age as a secondary one. It is a phase II trial, designed to decide whether a larger one should be launched. That is where things stand.

Why I do not take it

I am a young adult with no metabolic disorder. The measured effects of taurine address people with glucose or blood pressure to correct, and the argument that could have concerned me, ageing, has lost its biomarker. Adding a line to my stack for a benefit that does not target me is spending without any possible measurement. If the phase II trial finds something on PhenoAge, I will read again.

What this article does not say

• That taurine is useless. Thirty-four trials find real, modest metabolic effects in affected people. It is a useful molecule for some, not an anti-ageing one.
• That the 2023 study was wrong. Its results in mice and monkeys hold. It is the extrapolation to humans, from a cross-sectional measurement, that did not survive longitudinal measurement.
• That the ageing question is closed. A phase II trial is starting, with a biological age as an outcome. It will say whether a larger one is needed.
• That energy drinks are a source. Taurine comes in them with caffeine and sugar, at doses and in a context that have nothing to do with a controlled trial.
• That it is risk-free over years. No significant adverse effect in trials of one year at most. Nobody has looked further.

« A biomarker that falls with age in a cross-section does not necessarily fall in you. Only repeated measurement knows, and it is the only one that counts. »

What Helix does

This story is the one Helix replays with every blood test: a value compared with an age norm says nothing about what it does in a given person. You need the same value, in the same person, measured several times. That is the difference between the 2023 study and the 2025 one, and it is the difference between a blood test read once and a blood test followed.

In the engine, taurine is filed with what it has demonstrated: modest metabolic effects, a precise population, and a longevity promise marked as not measured in humans. When a biomarker falls, the card changes. When a headline falls, it does not move, because it never rested on it.