A pooled analysis of 72 randomized trials reports that aerobic training measurably improves endothelial function, the earliest widely used marker of artery health. The same review finds that changes in body weight, blood pressure, fitness and insulin all travel alongside the endothelial function response.
Endothelial function, the ability of the thin cell layer lining an artery to widen on demand, improved by 2.35 percentage points after aerobic exercise in a systematic review and meta-analysis published on 8 October 2026 in Sports Medicine – Open. The review pooled 72 randomized controlled trials covering 3,357 participants, and then asked which other cardiovascular changes moved with it.
The analysis was assembled by Simon Iskra, Gasper Turnsek, Kristina Drole and Armin Paravlic with several co-authors. Paravlic works at the Faculty of Sport Studies at Masaryk University in Brno, Czech Republic. The protocol was registered on the Open Science Framework before the data were pooled.
Databases were searched from inception to 15 September 2023, with updated searches run on 15 February 2025 and again on 22 June 2026. Only randomized controlled trials were eligible. Participants averaged 54.2 years of age, with a standard deviation of 13.9 years and a span from 20.5 to 78 years.
Pooled with a random-effects model, aerobic exercise produced a mean difference in brachial artery flow-mediated dilation of 2.35% (95% CI 1.98 to 2.73; p < 0.0001). Flow-mediated dilation is the standard non-invasive index of endothelial function, so that figure is the central estimate of the paper.
At a Glance
- Published 8 October 2026 in Sports Medicine – Open, volume 12, article 150.
- 72 randomized controlled trials, 3,357 participants, drawn from both healthy and clinical groups.
- Mean participant age 54.2 years, standard deviation 13.9, range 20.5 to 78 years.
- Brachial flow-mediated dilation improved by 2.35% (95% CI 1.98 to 2.73; p < 0.0001).
- VO2max rose with a standardised mean difference of 0.81, systolic pressure fell at an SMD of minus 0.35 and diastolic pressure at minus 0.41.
- Heterogeneity was substantial at I2 of 90.40%, and 14% of trials were judged at high risk of bias.
- Egger testing flagged possible publication bias in the main endothelial function analysis.
What 72 Randomized Trials Were Asked to Show
Flow-mediated dilation is measured by inflating a cuff on the forearm, holding it for several minutes, then releasing it and tracking how far the brachial artery expands as blood rushes back. A healthy lining responds briskly. A stiffer, less responsive lining does not, and that muted response is what researchers describe as impaired endothelial function.
The reason the measure matters is timing. Endothelial function tends to deteriorate before plaque builds, before a diagnosis and before symptoms, which makes it a candidate early indicator rather than a late one. That is why a shift of 2.35 percentage points is treated as substantive rather than cosmetic.
The review went further than a single pooled number. Alongside endothelial function, the team gathered blood biomarkers, haemodynamic measures, cardiorespiratory fitness, body composition, glucose and lipids, then tested whether changes in those secondary outcomes tracked the change in artery function.
How Far Endothelial Function Actually Moved
The headline estimate was a mean difference of 2.35% in favour of aerobic exercise, with a 95% confidence interval of 1.98 to 2.73 and a p value below 0.0001. The interval is narrow and sits well clear of zero, so the direction of the endothelial function effect is not in question.
Consistency is a different matter. Heterogeneity in the primary model was high, at an I2 of 90.40% with p < 0.0001, meaning the size of the endothelial function response varied widely from trial to trial. A pooled average of that kind describes a field of research, not any one programme.
Methodological quality was mixed rather than poor. By the count of the review team, 47% of included trials had a low risk of bias, 39% raised some concerns and 14% were judged at high risk. That spread is ordinary for exercise research, where blinding participants to their own training is impossible.
Blood Pressure, Fitness and Body Fat Shifted in the Same Trials
Secondary outcomes moved in the expected directions. Cardiorespiratory fitness showed the largest change, with a standardised mean difference of 0.81 for VO2max at p < 0.001. Systolic pressure fell with an SMD of minus 0.35 and diastolic pressure fell with an SMD of minus 0.41, both at p < 0.001.
Body composition changes were smaller. Body weight dropped only trivially, at an SMD of minus 0.130 (p = 0.024), while body fat percentage fell with an SMD of minus 0.36 (p = 0.006). On those numbers the endothelial function gain did not depend on substantial weight loss.
Arterial stiffness was harder to judge. Only 13 comparisons from 9 studies informed the carotid-femoral pulse wave velocity analysis, and 9 comparisons from 7 studies informed the augmentation index. Those are thin evidence bases next to the 72 trials behind the endothelial function estimate.
The pooled improvement in brachial artery flow-mediated dilation across 72 randomized trials and 3,357 participants.
What the Meta-Regression Does and Does Not Establish
The most novel part of the paper is its meta-regression. Changes in body weight, body fat, systolic pressure, diastolic pressure, VO2max and insulin each significantly moderated the endothelial function response. In plain terms, trials reporting bigger shifts in those measures tended to report bigger shifts in artery function as well.
Two coefficients give the flavour. Change in body weight carried a negative association with the training effect, at a beta of minus 4.021 (p = 0.008), and change in insulin carried a positive one, at a beta of 0.381 (p = 0.034). The pattern points toward several overlapping routes to better endothelial function rather than a single one.
The authors are careful about what that buys. Because the meta-regression rests on study-level associations, they state that the direction of causality cannot be established from it. Weight loss and improved insulin handling may drive endothelial function, follow it, or share an upstream cause with it.
Why the 24 Percent Risk Translation Needs Care
The paper also converts its result into risk language, using an established rule of thumb that a 1% improvement in flow-mediated dilation corresponds to roughly a 10% attenuation of cardiovascular risk. On that arithmetic the 2.35% gain works out to about a 24% lower risk of future cardiovascular events, a figure the text elsewhere renders as roughly 23.5%.
That conversion is a projection rather than an outcome anyone counted. No participant in these 72 trials was followed to a heart attack or a stroke as part of this analysis, and applying a population-level conversion factor to a pooled endothelial function estimate stacks one assumption on another.
Publication bias adds a second caution. Egger regression flagged possible bias in the overall endothelial function analysis and again in the analysis of VO2max, and the authors report that after adjustment the bias persisted at p = 0.0129. Small positive trials have always been easier to publish than small null ones.
How This Fits Alongside Earlier Vascular Research
Readers who have followed this beat will recognise the measurement. Fitness Living Magazine has reported that prolonged sitting impaired leg blood vessel function even in people who had completed 12 weeks of interval training, and that static stretching after lifting sped early vascular recovery in a small trial. Both sit alongside this review rather than inside it.
The haemodynamic findings echo earlier coverage of how often isometric exercise has to be repeated to hold a blood pressure drop, while the fitness result connects to work showing that estimated cardiorespiratory fitness tracked lower disease risk across three large cohorts. None of those studies measured endothelial function directly, which is part of why this pooled estimate is useful.
What Endothelial Function Research Changes for a Training Week
For most readers the practical message is unremarkable in the best sense. Regular aerobic exercise of the kind already recommended was linked to better artery lining function, and the pattern appeared in trials of healthy adults as well as adults living with disease.
What the review cannot supply is a prescription. It pooled trials with different modes, intensities, session lengths and programme durations, so it cannot say that one weekly structure produces one particular endothelial function gain. Anyone training around a diagnosed cardiovascular condition should work from clinical advice rather than from a pooled average.
- Aerobic work is the exposure examined here, not strength training, although the two are not in competition.
- Pressure, fitness and body fat all changed alongside endothelial function, so broad cardiovascular habits matter more than any single number.
- Flow-mediated dilation is a laboratory measurement taken with ultrasound, not something a wearable device reports.
The Fitness Living Takeaway
Across 72 randomized trials and 3,357 participants, aerobic exercise improved brachial flow-mediated dilation by 2.35%, the clearest pooled estimate yet of its effect on endothelial function.
The direction of that finding is firm, but the spread around it is wide, with heterogeneity at an I2 of 90.40% and publication bias that survived adjustment. The 24% risk reduction quoted in the paper is a projection from a conversion factor, not an event count, and the meta-regression linking weight, pressure, fitness and insulin to artery function describes associations between studies rather than proven mechanism.
Research & Sources
- Sports Medicine – Open: Impact of Aerobic Exercise on Endothelial Function and Other Cardiovascular Disease Markers in Health and Disease
- Study DOI
- Images: Gard Pro / Unsplash; Nathan Nuyda / Unsplash
This article summarizes peer-reviewed research for general information and is not individualized medical or exercise advice.
