Not Just Plumbing: Your Smallest Blood Vessels May Set the Pace of Aging in Every Organ

This narrative review argues that aging of the microvasculature (capillaries, their endothelial lining and supporting pericytes) is an active driver of organ decline, not a side effect of it. It describes a shared endothelial aging program that plays out differently in the brain, heart, kidney, lung, muscle, bone marrow and tumors. It separates rare senescent cells from a broader, population-wide drift in endothelial function, which has direct consequences for senolytic strategies. It then surveys candidate interventions, from exercise and metabolic drugs to VEGF signaling and endothelial reprogramming.

For more than three centuries, doctors have quoted Thomas Sydenham’s remark that a man is as old as his arteries. Two of Germany’s most prominent vascular biologists now argue that the saying points at the wrong vessels. In a review in Nature Cardiovascular Research, Stefanie Dimmeler and Hellmut Augustin make the case that the body’s smallest vessels, and the endothelial cells that line them, help set the pace of aging in every organ.

The central idea is that blood vessels are not plumbing. Endothelial cells send chemical instructions, called angiocrine signals, to the tissue around them. When those cells age, the instructions change. The review describes a shared aging program found in vessels throughout the body: less nitric oxide, more oxidative stress, DNA damage, a thinner sugar coating on the vessel lining, and stiffer surrounding tissue. The outcome depends on the organ. In the brain it appears as a leaky blood-brain barrier, in the heart as stiffening and heart failure with preserved ejection fraction, in the kidney as protein leaking into urine, and in muscle and bone as frailty and osteoporosis.

The most useful correction concerns senescent cells, the “zombie” cells that senolytic drugs are meant to remove. Single-cell studies show that truly senescent endothelial cells remain a small minority even in very old mice. Most of the endothelium instead drifts gradually into a less capable, mildly inflamed state. Killing a rare cell population cannot fix a problem that affects nearly every cell. It may also backfire: removing senescent endothelial cells worsened liver injury and pulmonary hypertension in mouse experiments.

Timing is another theme. Human protein data suggest blood vessels are among the first tissues to show aging, with an inflection near age 50. The authors argue that treatments helpful in midlife may be useless or harmful later. Statins illustrate the point. They partly mimic the vascular effects of exercise, yet their benefit shrinks in people over 75 without existing vascular disease, possibly because aged endothelial cells respond less to the drug’s signals.

On treatments, the review is candid about how little is proven. Exercise remains the best-supported way to improve endothelial function. Metformin, SGLT2 inhibitors and GLP-1 receptor agonists improve vascular measures, but nobody has shown whether they act on the vessel wall directly or simply improve whole-body metabolism. More exotic ideas, including boosting VEGF signaling, a mitochondrial drug called SUL-238 and reprogramming of endothelial cells, rest almost entirely on mouse work.

The authors end with a question they cannot answer: does rejuvenating vessels reverse organ aging, or only slow further decline? Readers should keep the paper’s nature in mind. It is a narrative review. Its value lies in the framework, and that framework is still a hypothesis.

Actionable Insights

The review contains no new trial data, so the numbers below come from the studies it cites.

  • Exercise is the only intervention the authors call established for vascular aging. The review gives no effect size for it.
  • Statins lower major vascular events by about 21 percent for each 1 mmol/L (39 mg/dL) drop in LDL cholesterol. If 20 of 100 people like you would have an event over ten years, that means about 16 instead, or 4 fewer. Over age 75 the reduction falls to about 13 percent, and in people over 75 without vascular disease it was not statistically distinguishable from zero.
  • SGLT2 inhibitors cut worsening heart failure or cardiovascular death from about 21 to 16 per 100 patients over 18 months in one landmark trial. That is 5 fewer per 100, in people who already had heart failure.
  • Senolytics are not ready. Fisetin did not extend lifespan in the most rigorous mouse testing program, and senescent cell removal caused harm in some models.
  • Vascular aging appears to accelerate around age 50, so blood pressure, lipids, urine albumin and fitness deserve attention in midlife rather than later.

Context/Source

  • Paywalled Paper: Vascular aging as a driver of organ dysfunction and systemic aging, Published 23 September 2026
  • Institutions: Institute of Cardiovascular Regeneration, Goethe University Frankfurt; German Center for Cardiovascular Research (DZHK); Cardio-Pulmonary Institute, Frankfurt; European Center for Angioscience, Heidelberg University, Mannheim; German Cancer Research Center, Heidelberg
  • Country: Germany
  • Journal: Nature Cardiovascular Research
  • Impact evaluation: The impact score of this journal is 12.6 (2025 Journal Impact Factor; 5-year 13.5), evaluated against a typical high-end range of 0 to 40 for top cardiovascular journals (0 to 60+ for top general science), therefore this is a High impact journal.

Related Reading:

Biomarker Data (Effect Size Extraction)

A standardized effect size such as Cohen’s d tells you how large a difference is relative to the normal spread between individuals. Calculating it requires group averages and standard deviations. The review reports none, so no Cohen’s d can be computed from it. The quantitative anchors below are what the review or its cited sources provide.

From the review itself:

  • Vascular aging becomes detectable in human proteomic data at roughly age 50. This is a descriptive inflection, not an effect size.
  • About 25 percent of cell types across 21 tissues show age-related population shifts in epigenomic single-cell data.
  • A ROCK2 inhibitor showed a trend toward reduced liver fibrosis in 5 of 6 participants (83 percent). With six people and no control group, the plausible true response rate runs from roughly 36 to nearly 100 percent. This is a signal, not a result. [Confidence: Low]

From Grunewald et al. 2021, checked against the original paper:

  • The intervention roughly doubled circulating VEGF from 8 months of age, about human age 40 by the authors’ estimate. This makes it a prevention study, not a reversal study.
  • Median and maximal lifespan increased in both sexes (log-rank P below 0.0001). The percentages appear only in a figure I could not read, so I am not quoting a number.
  • Treated 24-month-old females had 33 percent more tibial bone volume than littermate controls.
  • Kyphosis index was 38 percent better in treated mice at 24 months.
  • Mice given AAV-VEGF stayed on a rotating rod 49 percent longer than mock-treated littermates.
  • These are relative differences in means. No standard deviations are given in the text, so Cohen’s d is not calculable. Typical group sizes in such experiments are small, which tends to inflate apparent effects.

From clinical sources the review cites, with figures recalled from the original publications and not re-verified in this session:

  • Statins, Cholesterol Treatment Trialists 2019: rate ratio about 0.79 per 1 mmol/L LDL reduction overall (21 percent fewer major vascular events). Over age 75 it was about 0.87 (13 percent fewer). In primary prevention over 75 it was about 0.92, with a confidence interval spanning from a 27 percent reduction to a 16 percent increase, meaning no demonstrated benefit. [Confidence: Medium]
  • SGLT2 inhibitors, DAPA-HF as a representative trial: hazard ratio 0.74, meaning 26 percent fewer events per unit time. Absolute rates were 16.3 versus 21.2 percent over 18 months, a 4.9 point difference, or about 21 patients treated to prevent one event. [Confidence: Medium]

Hazard ratios and relative risks describe proportional change. The absolute benefit depends on your starting risk, which is why the same drug yields large gains in heart failure patients and small ones in healthy 45-year-olds.

All candidate interventions and the data behind them

Intervention Data the review provides or references Evidence level Caveats the authors state
Physical exercise Raises laminar shear stress, nitric oxide, KLF2 and KLF4. Protects aged mice from coronary endothelial senescence via FUNDC1 mitophagy. Improves cerebral blood flow and neurovascular coupling. Human physiology plus mouse Cannot separate endothelial from systemic effects. No effect sizes given.
Statins Reduce major vascular events across age groups. Partly mimic shear-stress programs independent of lipid lowering. Human meta-analysis of 28 trials Benefit attenuated over 75 without vascular disease. STAREE and PREVENTABLE trials ongoing.
SGLT2 inhibitors “Robust clinical benefits” in cardiovascular and renal disease, including non-diabetics. Attenuate endothelial senescence experimentally. Human meta-analysis plus cell and mouse Endothelial mediation unproven.
GLP-1 receptor agonists Reduce cardiovascular events. Endothelial GLP-1 receptor mediated liraglutide’s protection in hypertensive mice. Human trials plus mouse Direct versus indirect effect unresolved.
Metformin Activates AMPK in endothelial cells, preserves nitric oxide, improves barrier function, attenuates senescence phenotypes. Experimental; human trials ongoing Unclear if vascular effects are direct, or persist in advanced age.
ROCK2 inhibitor (TDI01) Phase 1 showed favorable pharmacokinetics and safety. Trend to reduced liver fibrosis in 5 of 6 participants. Human phase 1, N of 6 Uncontrolled, liver fibrosis only.
SUL compounds (SUL-138, SUL-238) Preserve complex IV activity and reduce superoxide. Protected against endothelial dysfunction and kidney damage in DNA damage-driven mice. Accelerated-aging mouse models Phase 2 trial is in early Parkinson’s disease, not vascular aging.
VEGF-A signaling restoration Low-dose systemic VEGF prevented capillary loss and extended healthspan and lifespan in mice. Naturally aged mice, one group Aged tissues already overexpress VEGFA mRNA; the defect is downstream.
VEGFB gene transfer Rescued cardiac denervation in aged mice, improved diastolic function, reduced fibrosis, raised heart rate variability. Aged mice Single study.
Liver exerkine GPLD1 Targets brain vasculature, restoring youthful transcriptional signatures and cognition. Mouse Single study.
Fisetin and senolytics Fisetin preferentially targets senescent endothelial cells and reversed endothelial dysfunction, partly via CXCL12. Mouse Removing p16-positive endothelial cells worsened liver injury; senescent cell clearance can worsen pulmonary hypertension. Senescent cells are a minority.
KDR overexpression (endothelial reprogramming) Reversed senescence-associated phenotypes and reduced liver fibrosis. Mouse Described as a “first approach”.
ZBTB16 overexpression Delayed cardiac fibrosis and preserved diastolic function. Mouse Dimmeler holds a patent application on this target.
NAD+ and hydrogen sulfide axis Described as a reversible cause of vascular aging and a key metabolic control point via sirtuins. Mouse No specific NAD+ precursor is named or endorsed.
Cytosolic NADPH enhancement Ameliorated vascular aging when raised in endothelial cells. Mouse Smooth muscle contribution cannot be excluded.
Endothelial Mfn2 deletion (mitohormesis) Improved antioxidant defense, lipid oxidation, systemic metabolism and healthspan. Mouse genetic A target concept, not a therapy.
Mitochondrial transfer to endothelial cells Tunneling nanotube transfer improved endothelial bioenergetics and engraftment. Bioengineering, mouse “Might in the future be applied”.
Endothelial transplantation Rejuvenated aged hematopoietic stem cell function. Mouse Mechanism uncertain.
Glycocalyx restoration Age-related loss of mucin-domain glycoproteins impairs the blood-brain barrier. Mouse No strategy exists that spares core glycoprotein synthesis.
Mechanotransduction mimics (KLF2/KLF4) Decades of flow biology. Concept “Not yet translated” into any clinical intervention.
Targeted vascular-immune modulation Proposed alternative to broad anti-inflammatory therapy. Concept Causality between endothelial dysfunction and inflammaging unresolved.
Gut microbiome (phenylacetic acid) This microbial metabolite induces endothelial senescence. Mouse Mentioned as a mechanism; no intervention proposed.

My ranking by strength of cited evidence

  • Tier 1, human outcome data: exercise, statins, SGLT2 inhibitors and GLP-1 receptor agonists. The outcome evidence is for cardiovascular and renal events in patients, not for slowing vascular aging in healthy people. [Confidence: High]
  • Tier 2, approved drug with mainly experimental vascular evidence: metformin. [Confidence: Medium]
  • Tier 3, early human signal: the ROCK2 inhibitor and SUL-238. The ROCK2 efficacy signal comes from six people, and SUL-238 is being tested for a different disease. [Confidence: Low]
  • Tier 4, mouse only: VEGF-A, VEGFB, GPLD1, the NAD+ axis, NADPH enhancement, KDR reprogramming, ZBTB16 and endothelial transplantation. VEGF-A has the broadest data, but from a single group. [Confidence: Low]
  • Tier 5, mouse data with documented harm signals: fisetin and other senolytics. The review’s own senescence versus drift argument undercuts them as a primary vascular strategy. [Confidence: Medium]
  • Tier 6, concept only: mechanotransduction mimics, glycocalyx restoration, mitochondrial transfer, vascular-immune modulation and Mfn2-based mitohormesis.

Microvasculature is a subject matter that does interest me. First, let’s note, the relative failure of senolytic agents - although I’m a long time skeptic of the utility of the field (I think the effect size of senolytic treatments is minor), perhaps it can be argued that it’s down to poor agents. But that’s not the point - the point rather is that the population of truly senescent cells is relatively small, and the elimination of those does not measurably impact the pathological environment, and can even be counterproductive as this snippet illuminates:

“The most useful correction concerns senescent cells, the “zombie” cells that senolytic drugs are meant to remove. Single-cell studies show that truly senescent endothelial cells remain a small minority even in very old mice. Most of the endothelium instead drifts gradually into a less capable, mildly inflamed state. Killing a rare cell population cannot fix a problem that affects nearly every cell. It may also backfire: removing senescent endothelial cells worsened liver injury and pulmonary hypertension in mouse experiments.”

Among the agents that might impact microvasculature, they mention SGLT2i and statins. But they don’t mention all possible drugs.

In the course of my research reading back in the day, I’ve also come across papers suggesting that another class of drugs might have some impact too: ARBs - notably telmisartan, another in the plethora of reasons why I elected to take it. Some literature:

!!Chinese paper.

Telmisartan ameliorates vascular endothelial dysfunction in coronary slow flow phenomenon (CSFP)

https://pubmed.ncbi.nlm.nih.gov/29314204/

“Coronary slow flow phenomenon (CSFP) is a coronary microvascular disorder with an increasing morbidity, and currently, available therapies are of limited clinical value for its cure. Hence, it is urgent to find a novel approach to CSFP treatment. Several studies show that endothelial dysfunction plays a critical role in the aetiology of CSFP. Telmisartan (TMST) is a clinically available anti-hypertensive medicine and has shown its potential properties for improving vascular endothelial function. Thus, we aimed to investigate the effect of TMST on endothelial dysfunction in CSFP, Endothelial-dependent flow-mediated vasodilation, serum levels of nitric oxide, adiponectin, and endothelin-1 were surveyed before and after 3 months of TMST treatment. And the percentages of vasodilator response to acetylcholine (Ach) were detected after 12 weeks of TMST treatment. Compare with pretreatment, flow-mediated vasodilation, nitric oxide, and adiponectin were substantially improved after TMST treatment; meanwhile, endothelin-1 was decreased in the TMST group (all P < .01). Compared with the model group, the vasodilator response to Ach was enormously increased after TMST intervention. Additionally, administration of SU11274 or GW9662 would partially reverse the protective effects of TMST on accumulative concentration-vasodilator responses to Ach (P < .01). We demonstrated that administration of TMST could remarkably increase the mRNA and/or protein levels of hepatocyte growth factor, mesenchymal-epithelial transition factor, peroxisome proliferation-activated receptor γ, whereas dramatically diminish mRNA and/or protein levels of p-JNK1/2, mitogen-activated protein kinase, and nuclear factor kappa B (P < .05). Our results thus implicate that TMST ameliorates endothelial dysfunction in CSFP. It is suggested that TSMF may play an important role in the medication of CSFP.”

Telmisartan protects against microvascular dysfunction during myocardial ischemia/reperfusion injury by activation of peroxisome proliferator-activated receptor gamma

Telmisartan exerts pleiotropic effects in endothelial cells and promotes endothelial cell quiescence and survival

Telmisartan enhances mitochondrial activity and alters cellular functions in human coronary artery endothelial cells via AMP-activated protein kinase pathway

https://pubmed.ncbi.nlm.nih.gov/25682036/

Telmisartan induces proliferation of human endothelial progenitor cells via PPARÎł-dependent PI3K/Akt pathway

https://pubmed.ncbi.nlm.nih.gov/19193378/

In selecting medications for my stack, one reason why I do so much exploration of off target effects, is that I’m trying to minimize my stack by selecting drugs that have multiple pleiotropic effects. That way, instead of taking drugs which have narrow applicability specific to a particular indication, I can try to have drug cover more ground - and instead of taking, say, 3 drugs I can take one which does the job of three. One of the reasons I picked telmisartan is that it seems to have many effects, perhaps this way it can work synergistically with other drugs in my stack. YMMV.

1 Like