chatGPT(6AstraMaxPaid):
This is a useful overview of possible vascular effects of supplements, but its originality is limited and several reporting problems weaken confidence in its evidence tables. Its main conclusion is appropriately cautious: improvements in vascular measurements do not establish that these supplements prevent heart attacks, strokes, or cardiovascular death.
The paper is Effects of Nutritional Supplements on Vascular Function: A Narrative Review, by Katsanaki and colleagues, published in July 2026. jcm-15-05693.pdf
Summary. The review examines supplements that might improve the endothelium, the cellular lining of blood vessels. Proposed mechanisms include increasing nitric oxide availability, reducing oxidative stress and inflammation, improving mitochondrial redox balance, and preserving the endothelial glycocalyx, the protective coating on the vessel surface.
Most studies measure intermediate outcomes:
- Flow-mediated dilation, or FMD: how much an artery widens in response to increased blood flow.
- Pulse wave velocity, or PWV: a measure of arterial stiffness.
- Blood pressure and biochemical markers: including inflammatory molecules and oxidised lipoproteins.
- Glycocalyx measurements: indirect assessments of the protective endothelial coating.
The principal findings reported in the review are:
| Intervention | Findings summarised | Main qualification |
|---|---|---|
| L-citrulline and watermelon | Some improvements in FMD, blood pressure, or arterial stiffness, consistent with increased arginine availability for nitric oxide synthesis | Results vary by population and endpoint; studies are generally small and short |
| Cocoa flavanols | Several trials report improved FMD and modest reductions in blood pressure or stiffness | The review incompletely covers evidence on actual cardiovascular events |
| Coenzyme Q10 | A meta-analysis of 12 trials involving 489 participants found improved FMD | No significant improvement in the adhesion molecules ICAM-1 or VCAM-1 |
| MitoQ | A trial of 20 older adults receiving 20 mg/day for six weeks reported a 42% relative improvement in FMD | Very small study; stiffness improved in participants with elevated baseline stiffness |
| Grape polyphenols | Some improvements in FMD and oxidative stress markers | Findings often concern particular formulations and need independent replication |
| Glycocalyx-supporting supplements | A psoriasis trial reported improved glycocalyx measurements and arterial stiffness | A separate trial in type 2 diabetes found no significant vascular benefit |
| Olive-derived compounds | Small crossover studies reported improvements in several vascular measurements | Human studies included only 15 or 30 participants and short treatment periods |
| Curcumin | Some favourable findings, particularly in studies of postmenopausal women | A 12-month trial in 88 people with advanced chronic kidney disease found no vascular improvement despite reduced IL-6 |
| S-allyl cysteine and black garlic | Cell experiments support effects on hydrogen sulfide signalling, nitric oxide, and oxidative stress | Those experiments do not establish clinical benefit |
The MitoQ figure is a relative change in FMD, not a 42-percentage-point increase or a demonstrated 42% reduction in cardiovascular risk.
Novelty. The contribution is primarily the organisation of existing evidence. It brings mitochondrial antioxidants and glycocalyx-directed interventions into the same discussion as more established nitric oxide and polyphenol approaches.
It provides no new experiments, clinical trial, pooled analysis, or independently validated mechanism. The most interesting element is its emphasis on several interacting components of vascular dysfunction, including endothelial surface structure. These concepts and the supporting studies predate this review.
Critique. The paper deserves credit for acknowledging small samples, formulation differences, uncertain long-term safety, and the distinction between vascular measurements and clinical outcomes. It also includes negative studies. However, there are substantive weaknesses beyond the usual limitations of supplement research.
-
The selection and grading of evidence are insufficiently transparent.
The authors explicitly describe an intentionally selective, non-systematic search. There is no reproducible screening process or formal assessment of bias across the included evidence.
Table 11 assigns categories such as moderate or moderate-high evidence quality without explaining a consistent grading procedure. These should be treated as author judgements, rather than formal certainty ratings.
The last search was on 30 June 2025, more than a year before publication, which also limits its claim to summarise current evidence.
-
It omits a major cardiovascular outcomes trial.
The cocoa section does not discuss COSMOS, published in 2022, despite repeatedly emphasising the shortage of trials measuring cardiovascular events.
COSMOS enrolled 21,442 adults and followed them for a median of 3.6 years. Cocoa extract did not significantly reduce the primary composite cardiovascular outcome: hazard ratio 0.90, with a 95% confidence interval of 0.78 to 1.02. Cardiovascular death was lower as a secondary outcome, but that finding requires caution because the primary outcome was not significant and multiple outcomes were assessed. Original COSMOS trial
This is a consequential omission: it excludes evidence directly addressing the gap the review identifies.
-
At least one human study is incorrectly classified as randomised.
Table 2 labels the Shatanawi study of L-citrulline in 25 people with diabetes as a randomised controlled trial.
The original paper describes a study without placebo control, analysing measurements before and after supplementation in the same participants. It does not report random allocation to parallel treatment and control groups. Calling it an RCT overstates its ability to establish causation. Original Shatanawi study
-
Some references do not match the evidence being described.
Table 1 attributes a watermelon meta-analysis to references 5 and 7. These are a 2017 review and a 2023 editorial, rather than the meta-analysis itself. The reported PWV and postprandial glucose findings match the Smeets meta-analysis, which is absent from the reference list. Smeets and colleagues
There are also biochemical editing errors: the text abbreviates glutathione peroxidase as GSH, while the figure caption calls glutathione an enzyme. GSH is reduced glutathione; glutathione peroxidase is a distinct enzyme.
These errors do not invalidate every conclusion, but they make checking the original studies necessary.
-
Mechanistic plausibility sometimes receives more weight than demonstrated mechanism.
Improved FMD alongside lower oxidative stress markers does not establish which molecular pathway caused the improvement. Similarly, correlations between glycocalyx measurements and arterial stiffness do not demonstrate that glycocalyx restoration mediated the stiffness change.
Several diagrams present a straightforward progression from supplementation to molecular changes to vascular protection. The evidence supporting each step differs substantially, particularly between cell experiments and human trials.
The curcumin trial illustrates this distinction well: reducing an inflammatory marker did not produce measurable vascular improvement.
-
The evidence does not support a reliable comparison between supplements.
Participants, formulations, doses, treatment durations, and measurement methods differ too much to rank interventions confidently. Numerous small studies also assess multiple endpoints, increasing the importance of prespecified outcomes and adjustment for multiple comparisons.
Results from one proprietary mixture cannot identify which ingredient was responsible or establish that another product will reproduce the effect. Nor does the review establish that combining apparently complementary supplements produces additive benefits.
For the mitochondrial biology, the paper mainly concerns oxidative stress and nitric oxide availability. It does not test mitochondrial citrate export, nuclear acetyl-CoA availability, histone acetylation, transcription, or splicing. Consequently, its vascular findings provide no direct evidence for changes in that metabolic and transcriptional pathway.