Cardiac MRI Reveals Impact of Aging on Heart Function

In a recent interview, Gareth Mathews, Clinical Lecturer in Cardiac Electrophysiology at the University of East Anglia, discussed his team’s innovative research exploring the impact of aging on heart structure and function using cardiac MRI. Motivated by the aging global population—and particularly the older adult demographics of Norfolk, UK—the study aimed to quantify “heart age” through AI-supported analysis of MRI data from healthy individuals across multiple international centers. The team identified left atrial volume and function as key markers most strongly correlated with chronological age, using them to develop a regression equation to estimate heart age. When applied to patients with common modifiable risk factors like hypertension, diabetes, obesity, and atrial fibrillation, the tool revealed an average heart age significantly older than the patients’ actual age, with diabetes and hypertension showing the most pronounced effects. Mathews emphasized the potential of heart age as a powerful communication tool to motivate early lifestyle or medical interventions and guide more personalized treatment decisions, while noting the need for further prospective studies to validate its use in routine clinical screening and patient management.

Transcript

Cardio Care Today: What inspired your team to investigate the concept of a heart’s “functional age” rather than relying solely on chronological age?

Gareth Mathews: Well, thank you to yourself and the Cardio Care Today audience for speaking to me about our paper today. Really, it was inspired by the fact that we’re all aware of the massive population, global public health crisis of aging. The WHO estimates that in the next 5 years, one and a half billion people will be over the age of 60. That’s one in six people. It’s not just a problem for rich and high-income countries. By 2050, it’s expected that 80% of older people will live in low- and middle-income countries. So, [it] really is a global health problem.

But our inspiration really came from our local population in Norfolk. Norfolk’s in the UK; it’s a popular retirement destination. It’s close to the coast and a big national park. What we find is that people really want to move there when they retire. We’ve got a very elderly population. In fact, some bits of Norfolk really are the oldest in the country of the UK.  We are really seeing that population crisis in our hospital when we treat patients.

When we talk to patients, everyone understands that aging is a natural part of life; they all understand that it’s going to happen. Their real priority is: “Can I age in a healthy way?” They want to maintain their independence, they want to have a good quality of life, and they want symptoms that are manageable as they age. So this is really what we set out to do. We wanted to understand the biology of aging in more detail. We already know that there are cellular and molecular processes that happen with aging and that lead to trouble. We wanted to see whether that also correlated with heart function and size changes using cardiac MRI. Our study was motivated by our own population, the global population, and the need to understand the biology of aging as it occurs.

Can you explain how your MRI-based technique measures heart aging and what specific metrics it analyzes?

What I would say about cardiac MRI is [that] it really is the gold standard in assessing the heart volumes, in assessing the heart function, in assessing tissue characterization, and increasingly in assessing hemodynamic physiology of the heart. It’s a great test for looking at all those different things and getting a sort of multi-parametric view of heart function. The other thing that’s great about cardiac MRI is [that] it doesn’t use any ionizing radiation, so it’s safe to apply. That was a particular consideration for us in our population because we’re using healthy volunteers and also patients who were younger.

We recruited a cohort of patients across several different countries—191 in total—who were healthy, who had no known cardiovascular disease or risk factors and across several different age groups. The centers involved were three from the UK—Norwich, Sheffield, [and] Leeds—but also [a] center in Singapore and the center in Spain. All these patients underwent a cardiac MRI. Then we used an AI [artificial intelligence]–supported algorithm to rapidly calculate the volumes and functions of all four chambers in the heart in these patients. They were all consistently analyzed across the different centers. Then we looked at how age of the participant affected all those different sizes and volumes and functions that we’d measured.

In general, what we tended to find was that the atria overall increased in size and reduced in function over the course of the aging process. The right ventricle tended to decrease in size but increase in function when you measured the ejection fraction with age. The left ventricle did tend to reduce in size; the function in terms of the ejection fraction stayed fairly static. There was a signal as you age that it might then start to increase in function in terms of ejection fraction measurements.

So we looked across all those different volumes, those different functions, and we were trying to ascertain which one changed most strongly with age, but which one was also most linearly predictive of age as well—both simplicity and strength of association. The two factors that we came out with were the size of the left atrium—so the left atrial and systolic volume—and also the function of the left atrium—so the left atrial ejection fraction. We used those two parameters in a regression equation to estimate the patient’s heart age.

It’s not really surprising to us that the left atrium comes out as this excellent predictor because it’s on the left side of the heart, so it’s exposed to the systemic circulation. It’s exposed to all the hypertension and vascular risk factors that we know about. But it’s also a thin-walled structure. So it’s much more likely to dilate when the pressures inside the heart increase. A lot of our work in the past has been about using volume metrics to estimate the pressure inside of the heart, and I suspect what we’re seeing is an age-related increase in pressure being reflected in the left atrial volume and therefore function.

In what ways do lifestyle factors like obesity, diabetes, and high blood pressure accelerate heart aging, according to your findings?

We derived the CMR [cardiac MRI] heart age and then we went on and applied it in a different cohort of patients who were aging; obviously, there was a spectrum of age, but [the patients] also had at least one risk factor for cardiovascular disease as they aged. The risk factors that we examined were hypertension, obesity, diabetes, and atrial fibrillation. We picked those risk factors because they’re common, and we also picked them because they’re modifiable as well. So there are things that you can do about [these risk factors] and hopefully improve outcomes. The total unhealthy cohort was 366 patients in size, and what we found overall was an average increase in the heart age of 4.6 years over and above what the patient’s actual true chronological age was.

That did vary depending on which condition you were looking at and which risk factor. Hypertension overall is [responsible for] up to a decade increase in the heart age across many age groups. Diabetes led to several decades worth of increased age, particularly in lower age groups. Atrial fibrillation, again, a consistent increase in age across all age groups. Obesity was a little bit more complicated; it did depend on the degree of obesity. It was not significant, although visually, it was slightly higher on the lower BMIs. But as you get up to BMIs of about 35 or 40, then you’re seeing several decades of increased functional age of your heart because of that risk factor.

What we’re measuring there is the downstream effects of these risk factors and all the cellular processes that are associated with them. Left ventricle hypertrophy, inflammation, oxidative stress, telomere shortening—all the kinds of mechanistic factors that we associate with aging—are being picked up as subtle volumetric and cardiac dysfunction that you can see with cardiac MRI.

How accurate is your heart age formula when compared to the actual clinical outcomes or known risk factors in patients?

The purpose of this study was to introduce the concept of the heart age to derive the equation and to show its utility in some common risk factors. We didn’t directly look at any hard outcomes associated with the heart age, so we didn’t go on and follow up for death, heart attack, stroke, heart failure—all those things that would be useful to look at when you’re thinking about risk factors. That would require further study, which we hope to conduct prospectively looking at outcomes.

We did specifically choose the risk factors of interest because we know that they associate with outcomes. There’s a good chance that heart age will associate with patient outcomes in the future, and we also chose them because we know that they’re modifiable. There’s already a good evidence basis that we should treat blood pressure, that we should treat diabetes. It stands to reason that heart age might support those treatments.

What implications could this technology have for routine cardiovascular screening and early intervention strategies?

Early intervention is a definite implication. The main use for the heart age will be in patient communication. It’s a patient communication tool. Everybody understands what aging is. Everybody goes through it. Everybody sees people age. It’s a universal experience. A patient knows what a 20-year-old looks like compared to what an 80-year-old looks like. If they have a cardiac MRI scan and you’re in clinic and you’re discussing lifestyle factors, you’re discussing starting medication for treating blood pressure and the patient’s 50, you could potentially tell them that “Actually, although you are 50, your heart age is actually performing the same as [the heart of] a 70-year-old.” And taking the communication from there, you can ask them how they feel about it and what factors they’d like to think about to try and improve their heart age and their outcomes. That’s the big immediate application of it.

It [heart age] could help clinicians because it’s a single simple number that boils down lots of different processes going on. It could be quite difficult to integrate all the different heart volumes. We’ve really picked out the left atrium as a very good one to look at in the course of this study. Giving a single number would allow someone who’s reporting the MRI scans to say, “Actually there is a big discrepancy between this patient’s heart age and their chronological age; perhaps they need screening for these modifiable risk factors.”  They could put that on the report to trigger further downstream testing.

Also, it has the potential to personalize treatment strategies more. You might be quite aggressive in your risk factor modification for a 20-year-old who’s got the heart age of a 60-year-old. You might pursue different treatment avenues compared to say an 80-year-old who has an 84-year-old heart; the difference is much lower, and probably, you wouldn’t want to accept potential risks or adverse outcomes for that patient. Those are the three immediate areas where it might benefit.

Your question also related to routine screening. That’s a much more difficult question. For routine screening, you would want a validated prospective study, particularly with outcomes showing that you could improve outcomes by using the test. Also, you need to have an acceptable rate of false positives, false negatives going forward, and [the] test has to be available and acceptable to everyone. That’s a much bigger study involving health economics as well. I don’t think it’s immediately going to be applicable for routine screening. It might be applicable for pragmatic screening. So [if] you’ve got a patient who’s having a cardiac MRI, you can report the heart age and then use it when you follow them up in clinic.

Any closing takeaways?

The test will enter mainstream practice, but it does require further validation. Those prospective studies that we talked about, [which] you would want if you were communicating to patients to be able to say, “Your heart age is 70 now, but actually, if you do this, perhaps your heart age will reduce to that of a 60-year-old.” That’s the kind of information we need going forward from this study—and in terms of also increasing its applicability, integrating it more into the general work stream, [making it] simple to calculate for doctors. The test has a good potential for that because everything in it is routinely measured on every CMR that’s done in the world. They’re easy-to-measure volumes, and they’re also very easy to automate in terms of AI. What I would see is more and more integration of the CMR heart age and any subsequent refinements to it being automatically calculated by the scanner and therefore given to clinicians as an almost free piece of information that they can then go on to use with patients.

There are challenges to its implementation, so it needs to be accepted into guidelines, which requires that evidence basis. CMR is obviously a specialist test; it’s not available everywhere, so there’s an access issue and a cost issue. It will get better as more and more centers get CMR, and the fact that it’s operator independent and amenable to AI actually will help that process as well.

The final thing to say is [that] it’s also going to require a period of clinician experience and familiarization and also education about how you communicate the information to patients and how you use it as a motivational strategy to improve health outcomes.