New Frontiers in Heart Failure Screening in People with Diabetes

The following is a verbatim transcript of Prof Lee-Ling Lim’s webinar. The transcript has been lightly edited for clarity.

Hi everyone, I’m Professor Dr. Lee-Ling Lim, who is an executive director of the Department of Clinical Research and Advancement from the University of Malaya. I’m also a senior consultant endocrinologist at the Department of Medicine, University of Malaya, Kuala Lumpur, Malaysia. I also hold honorary positions at the Department of Medicine and Therapeutics, The Chinese University of Hong Kong, and Baker Heart and Diabetes Institute. It’s really a pleasure for me to share with you insights on the new frontiers in heart failure screening in people with diabetes.

Throughout my presentation, we’ll be sharing a few of the key objectives. And one of them, of course, is really what would be the ways for us to improve the diagnosis, risk stratification, and management of early stage of heart failure among people with diabetes.

There are four stages of heart failure. What we have been managing right now is more towards the symptomatic heart failure that we term it as stage C and stage D of heart failure in collaboration with cardiologists. But we also have the earlier stages of heart failure, namely stage A, where many of these people do have type 2 diabetes, they even have type 1 diabetes, we have concomitant hypertension, obesity, we’re mainly treating the risk factors. But there’s a stage of the conditions that we call stage B heart failure. It’s the asymptomatic conditions. They haven’t presented to us with overt symptoms of heart failure, but in fact, they already have either a structural abnormality or biochemical abnormality in relation to the heart failure presentations.

So there are actually a few of the key surveys looking into what would be the key gaps in the entire ecosystem of heart failure management. One of them is actually done by the Asia Pacific Society of Cardiology that involved close to 260 practising clinicians, of whom 70% were cardiologists, then about 20% would be general practitioner or internists. And what’s striking enough from this survey is that 60% of them echoed that there’s a lack of screening programme for heart failure within the entire ecosystem. There’s also a lack of access to necessary tools, especially the essential blood test, including the biomarker to compensate and complement the use of cardiac imaging for the earlier diagnosis of heart failure.

And so how are we going to tackle this gap in the entire heart failure management? So let’s look back to the entire spectrum of the stages of heart failure again. We know that we do have the opportunity to prevent symptomatic heart failure, meaning the progression from an initial stage of stage A, to stage B, to the overt symptomatic of stage C and D. The use of natriuretic peptide here would be mainly for the risk stratification and monitoring for those with symptomatic heart failure. But what would be the role of natriuretic peptides in the earlier stages of heart failure, for example stage A and stage B?

So for stage A, for example, the diabetes people, natriuretic peptide, whether we’re talking about BNP or NT-proBNP, can be used as a screening tool and risk stratification. And most importantly, among those who already have structural abnormality or even biochemical abnormality, namely with the raised level of natriuretic peptide, in fact, this biomarker can be used for diagnostic tool for stage B heart failure needing further treatment and optimisation of guideline-directed medical therapy to prevent further disease progression.

So the first question that we need to ask ourselves or even to our other peer healthcare providers is how are we going to diagnose the early stage of heart failure, especially talking about these subsequent risk stratification and risk prediction?

This is a work that we have done over the past few years. It was published about 3 years ago in 2024, talking about the precision prognostication for cardiovascular disease among people with type 2 diabetes. This is the work alongside the American Diabetes Association and the European Association for the Study of Diabetes Precision Medicine in Diabetes initiative.

What we have done so far is that, we want to prove that there is any additional biomarker that can be used on top of the existing cardiovascular risk factors that could improve the prediction and prognostication of cardiovascular diseases among type 2 diabetes population. And what would be the established cardiovascular risk factors that we were looking at? What we have seen here as listed, it would be the usual traditional sociodemographic factors, we have concomitant risk factors looking into hypercholesterolaemia, whether there’s hypertension, diabetes, atrial fibrillation, obesity, and smoking. We also take into account the work that’s been done with regards to existing chronic conditions, for example, any of the microvascular diseases, the presence of chronic kidney disease, including the continuous variables with the use of estimated glomerular filtration rate or the albuminuria assessment.

And we actually utilise the stepwise evaluation of biomarkers framework as proposed by the American Heart Association looking into what if we use the additional biomarker for the cardiovascular disease prediction among people with type 2 diabetes. This is a Sankey plot that initially we have close to 200 studies, and as well as the identified non-established biomarkers, and through stringent evaluation criteria, we found that there were only three biomarkers that could improve the prognostication for cardiovascular disease among people with type 2 diabetes. And who were the three? Is actually the CCTA, which would be the CT coronary angiogram. Then you have the troponin T, as well as the NT-proBNP.

So from here, you may realise that the American Diabetes Association has incorporated the use of this biomarker into the early diagnosis of heart failure algorithm. In fact, this is a call to action to all of us who are managing people with diabetes, that apart from looking after people who present to us with symptoms of heart failure would be the stage C and stage D again, in which where we tend to refer them to cardiologists for co-management, we have to look after those with stage A heart failure of whom that we’ve been managing so far. But what about the middle stage of heart failure, stage B, of which they’re already having structural abnormalities, including increased biomarkers, signifying increase of the filling pressure? So for this group of people, the incorporation of biomarkers, as what has been proven by the ADA/EASD Precision Medicine in Diabetes work, incorporation of NT-proBNP or BNP or even with the troponin T, would actually help to identify this group of people, specifically for NT-proBNP in which then the American Diabetes Association has suggested a universal threshold of 125 pg/mL.

So if you’re seeing a person with type 2 diabetes as a screening tool, you have ordered for NT-proBNP and the level has exceeded 125 pg/mL, then you should refer them for additional echocardiogram. So when there’s a presence of abnormal structural abnormalities on the echocardiogram, you could actually label them as having heart failure, depending on this spectrum of disease.

The European Society of Cardiology has also have a more defined algorithm for the early diagnosis of heart failure. This is for outpatient people, when you have suspected de novo heart failure, apart from the history, physical examination, and ECG, the incorporation of NT-proBNP into the routine practice as a screening and diagnostic tool will be essential. So one of them, the rule-out level will be lower than 125 pg/mL, then heart failure is actually very unlikely. There’s also an age-adjusted rule-in as proposed by the ESC: when it’s beyond 75 years old, it’s 500 pg/mL; when it’s younger than 50 years old, it’s more than 125 pg/mL; and heart failure will be likely, you need to arrange for echocardiography within the 6 weeks of the testing itself.

There are actually other questions that have been posted to us that how we can integrate the early diagnosis of heart failure into our routine healthcare system. Are we talking about clinical effectiveness? Showing you here would be data from the STOP-HF randomised trial, it was conducted way back about 10 years ago, but it is a key landmark trial telling us how we can incorporate BNP leverage, which is the risk-based heart failure prevention trial.

So what does STOP-HF involve? It’s a 1-to-1 randomisation, and the study population is really reflective of what we’re seeing nowadays. They are actually the primary prevention group. They haven’t developed any known history of heart failure. So actually age is more than 40 years old plus one of the following risk factors as what have been listed here, for example, presence of hypertension, high cholesterol, diabetes, obesity, and all these cardiovascular risk factors were very common within our community. And close to 1,400 of them have been randomised into the intervention group. The intervention group would have received the BNP screening, and there’s a designated protocol for referral when the BNP was more than 50 pg/mL for the specialist cardiovascular centre for collaborative care. And what do they receive under this intervention arm? They receive risk education, there’s an optimisation of the risk factor management, there’s also ways to improve the treatment adherence, and of course, with the focus on the healthy lifestyle education. Then what about those who have been randomised to the usual care? They will actually receive the usual care per the routine management during that time.

Showing you here would be the baseline characteristics of the STOP-HF randomised trial. And it’s really surprising to see that 1 in 3 people who belonged, who were eligible for STOP-HF actually had a BNP of more than 50 pg/mL, which is raised. They were asymptomatic of heart failure. And looking into the presence of the comorbidities, close to 60 to 70% of them have concomitant hypertension, diabetes at about 20%, for obesity can be up to 30%, and those with concomitant hypercholesterolaemia will be as high as 50 to 60%.

One key thing that we have found from the STOP-HF trial would be showing you here in the middle panel, those with a BNP of more than 50 pg/mL, 1 in 3 of the study population. And on your right panel would be those with a BNP of lower than 50 pg/mL. What we have seen that amongst those with a raised BNP, the end point events, irrespective whether you are looking into asymptomatic of left ventricular systolic dysfunction, you are looking into those with left ventricular diastolic dysfunction remain asymptomatic, whether it’s a control or intervention, people with a raised BNP level, they do have a higher number of the end point events compared to those with a lower level of BNP. Then of course with those who have received intervention with an optimal risk factor management, with the healthy lifestyle and risk education, in fact, they would actually fare better compared to the usual care arm. And this Kaplan-Meier curve would even show things in a better manner when compared to the control group, which is the usual care treatment group, those who have received more intensive treatment, where with a raised BNP level, they have experienced up to 35% relative risk reduction of the major adverse cardiovascular event. And what about specifically those with a BNP of more than 50 pg/mL? You may realise that the odds ratio were actually better, they actually have a higher level of the effect sizes of 35% compared to 30% among all participants. And who are the participants with a BNP of more than 50 pg/mL? If you look back at the diagnostic criteria, they are those with a stage B heart failure. No symptoms of heart failure yet, but they are the high risk group, where they only have the raised level of BNP being done as a routine practice.

Another recent publication by Professor Rodica Pop-Busui would be just published in December in Diabetes Care, trying to incorporate and provide strong evidence to support the potential use of natriuretic peptide level within the type 1 and type 2 diabetes population without any known heart failure episode, meaning they are the asymptomatic group. And they are actually using the Optum Market Clarity data from 2016 up until 2023. And looking back at the characteristics of this study population, the mean age was about 64 years old, and about 55% were actually female. The mean A1C was 7%, and the mean BMI close to 35 kg/m². And 80% of them had a good kidney function of more than 60 mL/min. And this Optum Market Clarity data population, 70% of them have received a background RAS inhibitor, and just about 10% of them were receiving a background of the SGLT2 inhibitors. Looking at the NT-proBNP level, it’s not so markedly elevated. The median level is 110 pg/mL, and the third quartile of the NT-proBNP can be up to 325 pg/mL.

And what they have found, just want to draw attention when this incorporation of the NT-proBNP into routine practice, so this with NT-proBNP with a level of 125 to 300 pg/mL compared to those with a normal level lower than 125, and over this column will be NT-proBNP of more than 300 pg/mL, and looking into the heart failure-free survival, and on your right panel will be looking into the overall survival clinical endpoint. And we realise that irrespective whether we utilise NT-proBNP early in the disease course, incorporate into the routine practice for the screening, diagnosis, and risk stratification, irrespective whether he or she has type 1 diabetes or type 2 diabetes, those with a raised level of NT-proBNP would have been associated with the much increased risk of the heart failure mortality among this diabetes population and really support that when we were to incorporate the NT-proBNP into the diabetes population without known heart failure, this will actually be used as a useful prognostication tool for heart failure and mortality in the diabetes population.

And the next question that people will tend to ask is, yes, you have shown to us that there’s strong evidence to support the integration of NT-proBNP into our routine practice, especially among people with type 2 diabetes. But we have been using a lot of the guideline-directed medical therapy, for example, SGLT2 inhibitors have shown robust evidence in the prevention of the new-onset or heart failure hospitalisation. So how are we going to prove that NT-proBNP would be a useful marker for the early diagnosis of heart failure?

One data that we wish to share with all of you here is from the Swiss Diabetes Registry and how they incorporate NT-proBNP into early diagnosis of heart failure protocol. So close to 550 study populations, of whom most of them have type 2 diabetes, pretty similar to the US insurance database, the median age about 61 years old with a median diabetes duration of 14, and a median HbA1c about 7%, and 40% of them have a BMI of beyond 30 kg/m². Similarly, they have a good eGFR with a median level of close to 90 mL/min. And we should draw attention for this Swiss cohort, the use of SGLT2 was higher than the American cohort, close to 25%.

And what we have shown from here, that would be focus screening for heart failure at the Swiss diabetes annual control. So they actually do a one-time measurement of NT-proBNP annually, and they also check whether he or she would have any symptoms of heart failure. If they’re symptomatic, then of course they won’t, they will actually record when the echocardiogram was done. But one important thing is when the patient reports there’s no symptoms of heart failure, then do we need to assess the probability of heart failure? And what are the questions to be asked? Whether there’s a presence of hypertension, whether they have been treated with any diuretics, and the following will be the symptoms of heart failure or the signs of heart failure. Then you may say that when there’s at least one being positive, meaning even not having any of the signs or symptoms of heart failure as been listed here, when there’s a diabetes population plus one hypertension, this would have been justifiable for the utilisation and integration of NT-proBNP, using the cut-off as recommended by ADA of 125 ng/L or pg/mL, then it’s justified for referral echocardiography because this group would constitute the stage B heart failure group.

And what’s been shown here is they realise that when they incorporated this annual screening protocol, close to 11% of people were being diagnosed as having heart failure, and most of them were actually HFpEF, heart failure with preserved ejection fraction. About one-third of them have heart failure with reduced ejection fraction, and 32% of this group of people when being newly diagnosed, they actually have a prior use of SGLT2 inhibitors. And further evaluation had realised that amongst those who have been newly diagnosed of heart failure using the biomarker, if this Swiss Diabetes Registry didn’t incorporate NT-proBNP into their routine protocol, 60% of these people would have gone undiagnosed. So this will tell us that it’s really important for us to move beyond the traditional symptoms classification. There’s need for us to have an early incorporation of certain biomarkers, for example NT-proBNP with a proven efficacy and effectiveness in the early diagnosis of stage B heart failure among the high-risk population including people with type 1 and type 2 diabetes.

What about the cost? Everything that we add to our routine practice, we need to come with some cost-effectiveness analysis. We need to remember there’s always been 10 principles when we were to screen for any types of health conditions. This condition will need to be an important health problem. Heart failure has actually proven to be one of them. And there is a need for the entire spectrum of accepted treatment for patients with the recognized condition. We do have guideline-directed medical therapy for the heart failure population, whether you are talking about reduced ejection fraction or even those with heart failure with preserved ejection fraction. Do we have available facilities for diagnosis and treatment? And the answer is yes. Is there any early asymptomatic stage for heart failure? Yes, I’ve shown you just now there are four stages, stage A would be what we have been managing, those with metabolic risk factors, stage B heart failure is a well-recognized early stage of the heart failure spectrum. Whether there were any suitable tests or examinations? Yes, I’ve shown you, apart from cardiac imaging like echocardiography, now we do have blood-based biomarkers, either BNP or NT-proBNP. And the examination or the additional test is acceptable, and the answer again is yes, because it’s just minimally invasive from the blood taken for the additional BNP or NT-proBNP measurement. Do we know about the natural history of heart failure, the entire spectrum? I’ve shown you at the beginning of the presentation. Yes, we do know. Is there any agreed policy on whom to treat? Yes, that already having established clinical professional guidelines on who to treat and how to treat the stage A to stage D of heart failure. The cost of case finding, I’ll show you afterwards, and this case finding should be a continuing process, it’s not a once and for all, and this has been incorporated as recommended by ADA that we should incorporate as an annual screening protocol for people with diabetes. So those highlighted in green, that I’ve given you the justification that heart failure, the early stages fulfill the 10 principles for the screening for health conditions. What I’m going to show you now is pertaining to point 9, what will be the cost-effectiveness analysis. This is the latest analysis that we had about the use of NT-proBNP early for heart failure identification and diagnosis among people with type 2 without any known history of heart failure. This is using the Brazilian private healthcare system and just counted for direct cost, meaning hospitalisation, have not incorporated the indirect cost, which is the productivity loss. And the standard cost-effectiveness analysis in which the costing has been discounted at an annual rate of 5%.

And how was the analysis done? They actually compare the conventional heart failure risk assessment pathway, which is what we’re doing right now, or the NT-proBNP heart failure risk assessment. If you are following the ADA criteria, you are doing the annual NT-proBNP screening, they are able to identify the high-risk patients for further optimisation or initiation of cardioprotective treatment, and of course the low-risk patients, they will actually do not need further optimisation of the guideline-directed medical therapy. And using the Brazilian private healthcare setting, we realise that incorporation of NT-proBNP as one annual screening tool among people with type 2 diabetes has been leading to the improvement of quality, and it has been cost-dominant when we calculated for ICER, the incremental cost-effectiveness ratio, and proven to be cost-saving with a reduction of the heart failure hospitalisation burden.

So one good example would be using patients, this Madam T. Madam T is a teacher, 48 years old, when she came to me a few years ago, during that time, she presented with catabolic symptoms for a newly diagnosed type 2 diabetes. And when she came in 2022, her blood pressure was well-controlled, her BMI was at the overweight, with a close to 25 kg/m², but clinically, there weren’t any signs for heart failure.
And this shows the list of the investigation that she had done on the point of assessment. A1C was high, close to 12% with a fasting glucose of 16 mmol/L or equivalent of close to 280 mg/dL. Her kidney function was well, even though her urine albumin-to-creatinine ratio was not checked, but the eGFR was close to 100 mL/min. She has hypercholesterolaemia, the fundus camera done by the ophthalmologist was normal, but the NT-proBNP was already raised at that time. She did not complain of any symptoms of heart failure and the clinical examinations were normal for any signs of heart failure, but she has raised NT-proBNP.

So if we were to integrate NT-proBNP into early diagnosis of heart failure into our routine practice, just imagine if Madam T did not receive any of the annual screening using NT-proBNP, we as the managing physician will only diagnose her having type 2 diabetes and dyslipidaemia. But with the integration of NT-proBNP into routine practice, Madam T will have been diagnosed to have type 2 diabetes, dyslipidaemia, and stage B heart failure, and this entirely changed her disease prognostication. It also changes her disease trajectory and how we will be able to manage her and optimise her guideline-directed medical therapy to prevent further disease progression.

So as a conclusion, we’ve shown that there’s certain markers that can help to enhance cardiovascular risk stratification and prediction among people with type 2 diabetes beyond the established risk factors that we have known so far. And the strongest evidence that we’ve proven and we’ve published, it would be NT-proBNP, the troponin T, and the CT coronary angiogram. The American Diabetes Association and the European Society for Cardiology algorithms have already recommended early diagnosis of heart failure, and this rely heavily on biomarker analysis such as the NT-proBNP. As I’ve shown you a few examples, the best use of NT-proBNP would be among the high-risk populations, those with cardiovascular kidney metabolic conditions as proven to be cost-effective or even cost-saving in the certain healthcare setting, and one example shown here would be using the Brazilian healthcare system. With that, thank you for your kind attention, and I look forward to the further implementations for the enhancement of the clinical outcomes of our people living with type 2 diabetes. Thank you.