In plain words

In two of five small trials, blood pressure fell more with grape seed extract than with a placebo. The other three found no difference from the placebo.

The short answer

Of five randomized trials against placebo, three found no difference: 70 adults on 300 mg/d for 8 weeks (Ras, 2013), 78 on 300 mg/day for 16 weeks (Schön, 2021) and 30 on 400 mg/day for 12 weeks (Odai, 2019). Two found a drop: systolic fell 5.6% in 36 adults on 300 mg/d for 6 weeks (Park, 2015), and both readings fell in people with metabolic syndrome on 150 or 300 mg/day for 4 weeks (Sivaprakasapillai, 2009; the ledger has no n). Three pooled analyses, which may include these trials, disagree about systolic pressure. Certainty is low because the trials are small and split.

Where the evidence sits

Lower than placebo

■ Randomized trial, Park et al., 2015 · 36 people · 6 weeks · diastolic only a trend
■ Randomized trial, Sivaprakasapillai et al., 2009 · ledger has no n · 4 weeks
◪ Pooled analysis, Feringa et al., 2011 · may overlap studies counted here · diastolic no different
◪ Pooled analysis, Zhang et al., 2016 · may overlap studies counted here

No difference from placebo

■ Randomized trial, Ras et al., 2013 · 70 people · 8 weeks
■ Randomized trial, Schön, 2021 · 78 people · 16 weeks · men lower than placebo
■ Randomized trial, Odai et al., 2019 · 30 people · 12 weeks · fell from baseline; non-smokers lower
□ Pooled analysis reported by NIH National Center for Complementary and Integrative Health, 2025 · may include trials listed here · no new data · diastolic lower

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

Evidence weight7676 / 100?

Who says what

SourceFindingDesignPeopleDose and lengthFunding and independence
Ras, 2013No differenceRandomized70300 mg/d, 8 weeksNot in ledger
Schön, 2021Lower in menRandomized78300 mg/day, 16 weeksNot in ledger
Odai, 2019No differenceRandomized30400 mg/day, 12 weeksNot in ledger
Park, 2015Systolic lowerRandomized36300 mg/d, 6 weeksNot in ledger
Sivaprakasapillai, 2009Both lowerRandomizedNot in ledger150/300 mg/day, 4 weeksNot in ledger
Feringa, 2011Systolic lowerMeta-analysis390Not in ledgerNot in ledger
Zhang, 2016Both lowerMeta-analysis810100 to 2000 mg/dNot in ledger
NCCIH, 2025Diastolic lowerPublisher describing review1,080Not in ledgerUS government
Ras, 2013

No difference

Design
Randomized
People
70
Dose and length
300 mg/d, 8 weeks
Funding
Not in ledger
Schön, 2021

Lower in men

Design
Randomized
People
78
Dose and length
300 mg/day, 16 weeks
Funding
Not in ledger
Odai, 2019

No difference

Design
Randomized
People
30
Dose and length
400 mg/day, 12 weeks
Funding
Not in ledger
Park, 2015

Systolic lower

Design
Randomized
People
36
Dose and length
300 mg/d, 6 weeks
Funding
Not in ledger
Sivaprakasapillai, 2009

Both lower

Design
Randomized
People
Not in ledger
Dose and length
150/300 mg/day, 4 weeks
Funding
Not in ledger
Feringa, 2011

Systolic lower

Design
Meta-analysis
People
390
Dose and length
Not in ledger
Funding
Not in ledger
Zhang, 2016

Both lower

Design
Meta-analysis
People
810
Dose and length
100 to 2000 mg/d
Funding
Not in ledger
NCCIH, 2025

Diastolic lower

Design
Publisher describing review
People
1,080
Dose and length
Not in ledger
Funding
US government

What did the trials find?

Ras (2013) gave 70 adults with pre- or stage I hypertension 300 mg/d or placebo for 8 weeks:

BP values were modestly, but not significantly, affected by the polyphenol-rich GSE treatment v. placebo with an effect of - 3.0 mmHg for systolic BP (95% CI - 6.5, 0.5)

Park (2015) gave 36 prehypertensive adults a 300 mg/d extract beverage or placebo for 6 weeks; systolic pressure fell 5.6% and differed from placebo, while diastolic fell 4.7%, only a trend. Four weeks after stopping, Park reports:

BP returned to baseline after the 4-week discontinuation period of GSE beverage.

Sivaprakasapillai (2009) reports both readings lower than placebo in people with metabolic syndrome after 4 weeks at 150 or 300 mg/day; the ledger has no n.

Did some groups respond more?

Schön (2021) gave 78 mildly hypertensive adults 300 mg/day or placebo for 16 weeks, with no significant difference overall. In the 45 men:

Significant differences between placebo and GSEe treatment were observed in men after 8 weeks and 16 weeks, leading to a statistically significant systolic and diastolic blood pressure reduction

Odai (2019) gave 30 prehypertensive adults 400 mg/day for 12 weeks; systolic fell 13 mmHg from baseline but did not differ from placebo.

What do the pooled analyses say?

These pooled analyses may include the trials above. Feringa (2011) pooled 9 trials with 390 people and found systolic pressure 1.54 mm Hg lower, diastolic unchanged; the ledger has no dose or duration. Zhang (2016) pooled 16 trials with 810 people at 100 to 2000 mg/d and found both readings lower; the ledger has no duration. NCCIH (2025) describes a 2022 review of 1,080 participants in 19 studies: diastolic lower, systolic not; the ledger has no dose or duration.

Why the sources disagree

Comparisons differ: Schön reports falls in both the extract and placebo groups, and Odai's drop from baseline did not differ from placebo. Schön's positive result is in men only. Populations range from healthy adults to people with metabolic syndrome. Zhang's pooled analysis, which may include the trials above, found a drop in parallel but not cross-over trials.

What would settle it

A pre-registered, parallel, placebo-controlled trial of grape seed extract at 300 mg/day for 16 weeks in adults with prehypertension, large enough to compare men and women and smokers and non-smokers separately, with 24-hour ambulatory readings against placebo and a check 4 weeks after stopping. A matching result from a second, independent group would raise the certainty.

How this page was made

The 8 sources are five randomized trials, two meta-analyses that may include those trials, and the NCCIH page, a government publisher describing a 2022 review. All came from a literature search; animal feed studies 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 the Longevity Circle evidence editor approved each one. "Contested" means independent trials point in both directions; certainty follows the GRADE levels used in medical reviews. Drafting was assisted by AI.

Changelog: 4 Oct 2026, first version, 8 sources.