Reviews/Polyphenols and Antioxidants/Cocoa flavanols lower blood pressure
CV · POLY-CV-03 · v2
Borderline: within 4 points of the Likely/Uncertain boundary, so a re-grade could land either side.
Several separate human trials found that cocoa flavanols lowered blood pressure by a small amount, with systolic drops of roughly 1 to 4 mmHg in healthy people [1][2] and bigger drops in people who already had high blood pressure or heart failure [5][8]. A very large population study pointed the same way [17]. But the picture is not uniform: some well-run trials found no effect at all [9][15], and the one big, long-running trial found no overall benefit, with improvement only in a slice of participants who started with the lowest blood pressure [3]. Because the typical effect is small and the results vary a lot from study to study, the panel treated the word 'modestly' as a good fit. The evidence supports a real but small effect rather than a strong, reliable one.
Independent randomised trials report directionally consistent systolic reductions from cocoa flavanols: −4.4 mmHg systolic and −3.9 mmHg diastolic in healthy middle-aged adults [1], −1.4 mmHg over 12 hours in normotensive volunteers [2], and larger effects in disease populations — −11.9 mmHg 24-hour systolic in grade-I hypertensives [5] and −6.7 mmHg diastolic in chronic heart failure [8]. An acute postprandial trial in older hypertensives showed −3.8 mmHg resting systolic [4]. A large cross-sectional cohort (N=25,618) associated higher flavan-3-ol biomarker with −1.9 to −2.5 mmHg lower systolic BP [17]. Against this, the only large long-term RCT (N=8905) found no overall effect on incident hypertension (HR 0.96, 95% CI 0.88–1.05), with reduction only in the baseline-SBP<120 subgroup [3], and trials [9] and [15] were null. The contested point: The evidence divides on whether a small average effect is real and sustained or an artefact of small, short, or subgroup-selected studies. The single large long-term RCT [3] is null overall and positive only in a post-hoc baseline-BP subgroup, while the largest trial effect [5] comes from the smallest, oldest hypertensive sample (N=20), and several controlled trials [9][15] show no effect.
All 17 papers behind this claim with their DOIs and what each does to it, the panel’s arbitration, the evidence gap map and the open research questions. Free, and sent to your inbox.
We’ll also send you other claims from Polyphenols and Antioxidants, one at a time, and tell you when a release or a pivotal trial changes a grade. Unsubscribe by replying.
One reviewer's only job is to attack the finding. The challenges below were weighed before the grade was set.
The only large, long-term RCT [3] (N=8905) found no overall effect on incident hypertension (HR 0.96, CI 0.88-1.05 crosses 1); the positive 24% figure is a baseline-SBP<120 subgroup with P-interaction=0.002 — a post-hoc subgroup, not the primary result.
Effect sizes are implausibly heterogeneous, spanning -1.4 to -11.9 mmHg SBP; the largest effect [5] comes from the smallest, oldest hypertensive trial (N=20), a pattern consistent with small-study bias.
Where it remains uncertain. The only large, long-term RCT found no overall effect on incident hypertension (HR 0.96, CI 0.88–1.05), with benefit confined to a baseline-SBP<120 subgroup [3], and several trials are null [9][15].
Aleatory
[2] documents substantial between- and within-subject variation in BP response (ICC 0.41), with effect sizes inversely correlated to baseline BP, so individual response is intrinsically variable.
Epistemic
[3] (N=8905) is the single large long-term test and was null overall for incident hypertension; the rest of the corpus is small, short trials, leaving long-term sustained effect under-tested.
Inferential
Several positives are acute postprandial measures over about 2 hours [4][11], which do not license an inference about sustained, chronic BP lowering implied by the unqualified claim.
The load-bearing papers from a corpus of 17, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
Parallel-group RCT in 100 healthy adults: modest SBP/DBP reduction.
“CF decreased systolic and diastolic blood pressure by 4·4 mmHg (95 % CI 7·9, 0·9 mmHg) and 3·9 mmHg (95 % CI 6·7, 0·9 mmHg)”
Panel auditInternal validity: high · scope of inference: Healthy, low-risk, middle-aged (35-60 y) men and women without CVD.Key limitations: BP a secondary endpoint · healthy low-risk population · 1-month duration
Rigorous n-of-1 crossover confirms a small but significant SBP reduction in free-living normotensives.
“The overall mixed model analysis showed that CF significantly decreased 12-h systolic blood pressure and PWV by -1.4 ± 0.3 mmHg and -0.11 ± 0.03 m/s, respectively”
Panel auditInternal validity: medium · scope of inference: Healthy, free-living young adults with normal blood pressure.Key limitations: n=11 · large response variability · normotensive free-living setting
Large long-term RCT null overall; benefit only in SBP<120 subgroup — caps the claim at modest/heterogeneous.
“cocoa extract supplementation had no significant effect on incident hypertension in an intention-to-treat analysis, with incidence rates of 7.1 and 7.4 per 100 person-years in cocoa and placebo groups, respectively (hazard ratio, 0.96 [95% CI, 0.88-1.05])”
Panel auditInternal validity: high · scope of inference: Older adults (women ≥65, men ≥60) free from baseline hypertension; outcome is incident hypertension over median 3.4 y.Key limitations: outcome is incident hypertension not BP change · self-reported component in outcome · subgroup effect only in baseline SBP <120
Stress-induced BP responses were similar between high- and low-flavanol conditions.
Panel auditInternal validity: medium · scope of inference: Young healthy men, acute single-dose BP response during a mental-stress task.Key limitations: BP only as stress response · acute single dose · healthy young men
4 of 17 papers shown. The full review lists all of them, with the panel’s arbitration →
Claims are graded at each quarterly release, and re-graded between releases only when a pivotal trial lands. Every change is a new version of the claim.
A real but small and heterogeneous BP-lowering effect is supported by consistent directionality across independent trials, while the null large RCT and several null studies keep it below the Established threshold.
Release 2026-q4 re-graded every claim from its corpus under protocol v0.19. An earlier release's grade is shown for the record; it came from an earlier protocol and is not a movement.
Assumptions that carry weight but are not themselves established.
The panel records the gap that would most move the claim, and the study that would close it.
Large long-term RCT with BP as primary endpoint. The only large long-term trial [3] measured incident hypertension and was null overall; a trial powered on continuous ambulatory BP as the primary outcome is missing. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: Multicentre, double-blind, placebo-controlled RCT with standardised cocoa flavanol dose and 24-hour ambulatory BP as the pre-registered primary endpoint over 12+ months.
Open question: Does cocoa flavanol supplementation produce sustained ambulatory BP lowering over months as a primary endpoint? Why it is still open: The only large long-term RCT [3] used incident hypertension and was null overall; smaller trials showing reductions [1][5] were short.
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 21 Oct 2026 · protocol v0.19