In plain words

One trial found that taurine did not change insulin sensitivity in 20 men at raised risk of type 2 diabetes. In a second trial, taurine was associated with higher adiponectin, a different measure, than placebo in 16 obese women. A pooled analysis that may include these trials found lower fasting insulin.

The short answer

Two randomized trials, both lasting 8 weeks, point different ways: Brøns and colleagues (2004) gave 1.5 g a day to 20 men at raised diabetes risk and found no effect on insulin sensitivity, while in Rosa and colleagues (2013), 3 g a day with nutritional counseling in 16 obese women was associated with higher adiponectin than placebo. A 2020 review by Wu and colleagues reports a study that found better insulin sensitivity in people with high blood sugar. Tzang and colleagues (2024) found lower fasting insulin, a pooled estimate that may include the trials above. Certainty is very low, because both trials are small and measured different outcomes.

The 60-second version

Every study in the video is listed below, with its quote.

Watch on YouTube

Where the evidence sits

Insulin markers improved

■ Randomized trial, Rosa et al., 2013 · 16 people · 8 weeks
◪ Pooled analysis, Tzang et al., 2024 · may overlap studies counted here
□ Review, Wu et al., 2020 · summarises other studies

No effect found

■ Randomized trial, Brøns et al., 2004 · 20 people · 8 weeks · also reported by Ripps et al., 2012

■ measured by the source's own study · ◪ pooled analysis, may overlap studies counted here · □ reported, reviewed or argued, no new data

Evidence weight3434 / 100?

Who says what

SourceFindingDesignPeopleDose and lengthFunding and independence
Brøns et al., 2004No effect on insulin sensitivityRandomized trial20 men at raised diabetes risk1.5 g a day, 8 weeksNot in ledger
Rosa et al., 2013Higher adiponectin than placeboRandomized, placebo-controlled16 obese women3 g a day plus counseling, 8 weeksNot in ledger
Tzang et al., 2024Lower fasting insulin and HOMA indexMeta-analysis of randomized trialsNot in ledgerNot in ledgerNot in ledger
Wu et al., 2020Better insulin sensitivity with high blood sugarReview describing another group's studyNot in ledgerNot in ledgerNot in ledger
Ripps et al., 2012No effect on insulin sensitivityReview that appears to describe the Brøns trial201.5 g, 8 weeksNot in ledger
Brøns et al., 2004

No effect on insulin sensitivity

Design
Randomized trial
People
20 men at raised diabetes risk
Dose and length
1.5 g a day, 8 weeks
Funding
Not in ledger
Rosa et al., 2013

Higher adiponectin than placebo

Design
Randomized, placebo-controlled
People
16 obese women
Dose and length
3 g a day plus counseling, 8 weeks
Funding
Not in ledger
Tzang et al., 2024

Lower fasting insulin and HOMA index

Design
Meta-analysis of randomized trials
People
Not in ledger
Dose and length
Not in ledger
Funding
Not in ledger
Wu et al., 2020

Better insulin sensitivity with high blood sugar

Design
Review describing another group's study
People
Not in ledger
Dose and length
Not in ledger
Funding
Not in ledger
Ripps et al., 2012

No effect on insulin sensitivity

Design
Review that appears to describe the Brøns trial
People
20
Dose and length
1.5 g, 8 weeks
Funding
Not in ledger

Did insulin sensitivity change when it was measured directly?

Brøns and colleagues (2004) gave 20 men at increased risk of type 2 diabetes 1.5 g of taurine a day for 8 weeks in a randomized trial. They concluded:

Daily supplementation with 1.5 g taurine for 8 weeks had no effect on insulin secretion or sensitivity

Ripps and colleagues (2012), in a review, describe a study of 20 obese people with a genetic predisposition to type 2 diabetes, given 1.5 g for 8 weeks, with the same result. It appears to describe the Brøns trial, so it is one measurement, not two.

What did the trial in obese women measure?

Rosa and colleagues (2013) compared taurine with placebo in 16 obese women taking 3 g a day for 8 weeks alongside nutritional counseling. They report adiponectin rather than insulin sensitivity itself:

Eight weeks of taurine supplementation associated with nutritional counseling is able to increase adiponectin levels

The ledger has no insulin or HOMA result from this trial.

What do the pooled analysis and the reviews add?

Tzang and colleagues (2024) combined randomized trials and found lower fasting insulin and a lower HOMA index with taurine than with control. This is a pooled estimate that may include the trials above, and the ledger has no trial count, n, dose or duration for it:

fasting insulin levels (WMD: −1.521 mU/L, 95% CI: −2.591 to −0.451, p = 0.005)

Wu and colleagues (2020) report in a review that a study by Sarkar and colleagues found better whole-body insulin sensitivity in people with high blood sugar. The ledger has no n, dose or duration for that study.

Why the sources disagree

The studies measured different things: insulin secretion and sensitivity (Brøns), adiponectin (Rosa), and fasting insulin and the HOMA index (Tzang, a pooled estimate that may include both trials). The people differ: men at increased diabetes risk, obese women, and, in the study Wu describes, people with high blood sugar. Both trials lasted 8 weeks, with 20 people at 1.5 g a day and 16 at 3 g a day plus nutritional counseling. Two of the five sources are reviews of other groups' work, and the ledger has no funding details.

What would settle it

An independent, pre-registered trial in at least 100 adults, giving 3 g of taurine a day or placebo for at least 12 weeks with no other change in diet. It would measure insulin sensitivity directly, plus fasting insulin, the HOMA index and adiponectin, and report men, women and people with high blood sugar separately. A matching result from a second group would raise certainty.

How this page was made

The 5 sources come from a search of PubMed records on taurine; studies in animals, cells and disease groups such as heart failure were left out. Every claim was copied word for word from its source into the Longevity Circle ledger and checked against the original text, and a person approved each one. Certainty follows the GRADE levels used in medical reviews. Drafting was assisted by AI.

The ledger holds only 3 studies on this question.

Changelog: 23 Sep 2026, first version, 5 sources.