The twisting may cause the pressure, not the other way around
Hidden in the delicate vessels at the back of the human eye, scientists have long glimpsed shadows of cardiovascular disease — but the direction of that relationship has now been reversed. Researchers at St George's, University of London, using artificial intelligence to examine retinal images from tens of thousands of people, have identified 119 regions of the genome that shape how blood vessels twist and coil, suggesting that arterial tortuosity may drive high blood pressure rather than follow from it. The finding reframes not only how we understand the origins of heart disease, but how an ordinary visit to the optometrist might one day become an act of prevention.
- A long-accepted medical assumption — that high blood pressure warps arteries — has been quietly overturned by genetic evidence hiding in the eye.
- AI analysis of retinal images from nearly 53,000 people uncovered 119 genomic regions influencing vessel shape, 89 of them tied specifically to how much arteries spiral and coil.
- The reversal of cause and effect creates urgency: if twisted arteries are a genetic precursor to cardiovascular disease, millions of people may be carrying risk that current screening never catches.
- Researchers validated their findings across two independent datasets, lending the associations unusual credibility and clearing a path toward clinical application.
- The technology needed for mass screening already exists in optometrists' offices — adding AI software could transform a routine eye exam into an early warning system for heart disease.
Scientists at St George's, University of London have used artificial intelligence to examine the blood vessels at the back of the eye — and in doing so, may have rewritten a foundational assumption of cardiovascular medicine. Their study, published in PLOS Genetics, drew on retinal images from nearly 53,000 participants in the UK Biobank, with results confirmed in a further 5,000 from the EPIC-Norfolk eye study. Using AI to distinguish arteries from veins and measure how much they twist, the team mapped 119 sections of DNA that influence the shape and size of retinal vessels — 89 of them linked specifically to arterial tortuosity.
The implications cut against decades of received wisdom. Medicine had assumed that high blood pressure caused arteries to become twisted — a downstream consequence of sustained vascular damage. This research suggests the opposite: genetic variants that predispose a person to more twisted arteries appear to be what elevates diastolic blood pressure and raises the risk of heart disease itself. Cause and effect, it turns out, may have been running in the wrong direction all along.
Professor Christopher Owen, who leads the chronic disease epidemiology group at St George's, pointed to the practical horizon this opens. Retinal imaging is already standard at high street optometrists. Layering AI analysis onto those same scans could allow a routine eye check to flag individuals at elevated cardiovascular risk — before symptoms appear, before damage accumulates. For those identified early, the knowledge could prompt lifestyle changes, medication, or closer monitoring. More broadly, it marks a step toward precision medicine: reading genetic and imaging data not to diagnose illness, but to anticipate it.
Scientists at St George's, University of London have used artificial intelligence to peer into the genetic architecture of a problem hiding in plain sight: the twisted blood vessels at the back of your eye that may be quietly predicting whether you'll develop high blood pressure or heart disease.
The research, published in PLOS Genetics, analyzed retinal images from nearly 53,000 people enrolled in the UK Biobank, with findings validated in another 5,000 participants from the EPIC-Norfolk eye study. Using AI to automatically distinguish between arteries and veins in these images, the team measured vessel width and the degree to which they spiral and coil. What emerged was a map of the genome itself: 119 distinct sections of DNA that influence how blood vessels in the retina take shape and size. Of those 119 regions, 89 were specifically linked to how much the arteries twist.
The discovery upends a long-held assumption about cause and effect. Doctors had believed that high blood pressure caused arteries to become twisted—a visible consequence of the condition's damage. The new work suggests the relationship runs the other way. Genetic variations that predispose a person toward more twisted arteries appear to be what drives elevated diastolic blood pressure, the force exerted between heartbeats, and increases the risk of heart disease itself. It is a reversal that reframes how we might think about prevention.
Professor Christopher Owen, who leads the chronic disease epidemiology group at St George's, framed the implications plainly: the genetic information now available could become a cornerstone of understanding cardiovascular risk and might open pathways to treatments not yet imagined. He noted that retinal imaging is already routine in high street optometrists' offices—a familiar part of an eye exam. The addition of AI analysis to those same images, he suggested, could transform them into a screening tool. During a standard eye check, the system could flag individuals at high risk of developing high blood pressure or heart disease, enabling doctors to intervene early, before symptoms emerge or damage accumulates.
The method relies on a technique called genome-wide association study, or GWAS, which searches for patterns in DNA across large populations to find correlations between genetic variants and observable traits. In this case, researchers looked for similarities in the genetic code of people whose retinal blood vessels shared particular characteristics. The consistency of findings across two independent datasets—the UK Biobank and EPIC-Norfolk—strengthens confidence that the associations are real and reproducible.
What makes this work significant is not just the number of genetic regions identified, which exceeds any previous study of retinal vessel characteristics, but what it reveals about the hidden architecture of cardiovascular disease. The shape and size of blood vessels in the retina have long been known to correlate with conditions like diabetes, obesity, and heart disease. But the genetic mechanisms driving those vessel characteristics remained largely mysterious. Now, with 119 genomic regions mapped, researchers have a much clearer picture of how inheritance shapes the vascular system itself—and how that inheritance might predispose someone to disease.
The practical next step is integration. Optometrists and general practitioners already have the imaging technology. Adding AI analysis would require software, training, and clinical validation, but the pathway is clear. A person coming in for a routine eye exam could leave with more than a new glasses prescription; they could leave with actionable information about their cardiovascular future. For those identified as high-risk, that knowledge could prompt lifestyle changes, medication, or closer monitoring—interventions that might prevent or delay disease. For the broader population, it represents a shift toward precision medicine: using genetic and imaging data to predict risk before it manifests as illness.
Notable Quotes
It had been thought that high blood pressure might cause twisted arteries, but our work unveils that it's actually the other way around.— Professor Christopher Owen, St George's, University of London
Our AI analysis of these images as part of routine eye checks could easily be done as part of a health check to identify those at high risk of developing high blood pressure or heart disease and in need of early intervention.— Professor Christopher Owen
The Hearth Conversation Another angle on the story
So the AI is reading the eye images and finding patterns that doctors missed before?
Not exactly. Doctors knew the patterns were there—twisted vessels in the retina have been visible for years. What the AI did was process thousands of images fast enough to let researchers connect those visible patterns to specific sections of DNA. It's the scale and speed that changed the game.
And the surprising part is that the twisting causes the high blood pressure, not the other way around?
Right. For a long time, people assumed high blood pressure damaged the arteries and made them twist as a side effect. This work suggests the genetics come first—some people inherit a tendency toward twisted arteries, and that twisting is what drives the blood pressure up.
How does a twisted artery cause higher blood pressure?
That's the honest answer: we don't fully know yet. But the correlation is clear in the data. The mechanism—how the geometry of the vessel affects the flow of blood and the pressure it exerts—that's part of what future research will need to untangle.
If someone gets flagged as high-risk during an eye exam, what happens next?
That's where it becomes practical medicine. They'd know they carry genetic risk factors. A doctor could start them on preventive treatment earlier, or recommend lifestyle changes before any damage shows up. It's the difference between waiting for symptoms and acting on prediction.
Does this mean we're close to preventing heart disease through eye exams?
Not quite. This is foundational work—mapping the genetics and proving the connection. The clinical tools still need to be built and tested. But yes, the direction is clear. In a few years, this could be routine.