IPM Take
Cardiology has traditionally met protein misfolding at the point where the heart becomes visibly infiltrated with amyloid.
That may be too late, and too narrow.
A new review in Nature Reviews Cardiology brings together evidence that misfolded proteins can damage cardiovascular tissues through multiple routes: amyloid deposition, soluble oligomer toxicity, inflammation, endothelial and microvascular dysfunction, thrombosis, autonomic disruption and impaired cellular protein quality control.
The clinical evidence is strongest for transthyretin and immunoglobulin light-chain amyloidosis, where cardiovascular involvement is established, diagnostic pathways exist and disease-modifying therapies can now alter outcomes.
But the review deliberately pushes further.
Amyloid-beta, better known for its role in Alzheimer’s disease, has been linked with vascular injury and adverse cardiovascular phenotypes. Alpha-synuclein disorders such as Parkinson’s disease frequently involve autonomic cardiovascular dysfunction. Medin, one of the most common human amyloid proteins, accumulates in ageing arteries and is being investigated as a contributor to vascular pathology.
These conditions should not be collapsed into one disease mechanism, and the evidence supporting cardiovascular intervention varies enormously between them.
The policy relevance lies elsewhere.
Once disease-modifying treatment exists, diagnostic delay becomes more expensive. And once routine ECGs, echocardiograms and electronic health records can be screened by algorithms for patterns suggestive of transthyretin amyloid cardiomyopathy, cardiology faces a familiar precision-medicine question: how aggressively should health systems search for disease before symptoms force the diagnosis?
The answer is not population-wide amyloidosis screening tomorrow.
It is recognition that protein misfolding is moving upstream in cardiovascular medicine, from pathology observed at advanced disease toward something that may increasingly be detected, phenotyped and treated earlier.
Executive Summary
A review published in Nature Reviews Cardiology on 5 November 2026 synthesises evidence on cardiovascular involvement across protein misfolding disorders and proposes a broader framework connecting proteotoxicity with myocardial, vascular and heart-brain disease.
The authors identify transthyretin amyloidosis and immunoglobulin light-chain amyloidosis as the prototypical cardiovascular protein misfolding disorders. Both can cause cardiomyopathy, microvascular dysfunction, arrhythmias, autonomic abnormalities and thrombotic complications.
The review argues that cardiovascular injury is not explained by amyloid deposits alone. Soluble misfolded proteins and oligomers may exert direct cellular toxicity before or alongside large extracellular deposits, while oxidative stress, inflammation and disturbances of protein quality-control systems can compound organ injury.
Protein misfolding may also affect vascular biology. Evidence reviewed includes endothelial dysfunction, microvascular injury, abnormal vascular stiffness and thrombosis in amyloidosis, alongside emerging research into amyloid-beta, medin and alpha-synuclein across vascular ageing, Alzheimer’s disease and Parkinson’s disease.
The strength of evidence differs substantially by condition. ATTR and AL amyloidosis are clinically established diseases with validated diagnostic pathways and disease-modifying therapies. Cardiovascular effects attributed to amyloid-beta, medin and other proteinopathies are scientifically important but remain much less mature as targets for routine cardiovascular screening or treatment.
The treatment landscape for ATTR cardiomyopathy has expanded rapidly. Current strategies include stabilising the transthyretin protein or reducing its production, while fibril-depleting approaches and in-vivo gene editing are being investigated.
Even with modern therapy, however, cardiovascular risk does not disappear. Established structural damage, arrhythmia, microvascular dysfunction and other mechanisms can continue to influence prognosis, creating what the review describes as meaningful residual cardiovascular risk.
The authors therefore propose earlier identification of disease, more systematic cardiovascular phenotyping and development of biomarker-anchored trials. They also highlight AI-based analysis of routine ECGs, echocardiograms and electronic health records as potential tools for finding patients who warrant definitive amyloidosis testing.
Why it matters
- HTA bodies: Cardiac amyloidosis increasingly illustrates why the value of diagnostics and therapies cannot be assessed independently. A treatment that works best before irreversible myocardial damage has accumulated changes the value proposition of earlier detection. Future assessment may therefore need to examine diagnostic-treatment pathways rather than evaluating an AI screening algorithm, scintigraphy test or amyloidosis therapy as isolated technologies.
- Payers: Disease-modifying therapies for ATTR cardiomyopathy create an incentive to identify disease earlier, but broad screening could generate substantial testing costs and false-positive referrals. The key reimbursement question is likely to be targeted case finding: which combinations of age, heart-failure phenotype, musculoskeletal clues, ECG abnormalities and imaging findings produce sufficiently high pre-test probability to justify further testing?
- Industry / innovation partners: Protein misfolding has become a crowded therapeutic field extending from stabilisers and gene silencers to gene editing and amyloid-depleting antibodies. Differentiation will increasingly depend on demonstrating not only reduction of the pathological protein, but meaningful improvement in cardiovascular structure, function, symptoms and hard clinical outcomes. Earlier diagnosis could also alter trial populations by shifting treatment toward less advanced disease.
Medicine has historically organised protein misfolding by the organ in which the damage becomes most visible.
Amyloid-beta belongs to Alzheimer’s disease. Alpha-synuclein belongs to Parkinson’s disease. Transthyretin and immunoglobulin light chains enter cardiology when enough abnormal protein accumulates in the heart to cause cardiomyopathy.
The biology is proving less respectful of those boundaries.
A major review published in Nature Reviews Cardiology argues that protein misfolding diseases share pathways capable of damaging both the heart and vasculature and that cardiovascular involvement deserves to be considered as part of a broader proteostasis problem rather than only as a complication of a few rare amyloidoses.
The strongest clinical example is cardiac amyloidosis.
In transthyretin amyloidosis, the normally soluble transthyretin protein becomes unstable, misfolds and aggregates into amyloid structures that can accumulate in the myocardium and other tissues. Disease can result from inherited TTR variants or occur without a pathogenic variant as wild-type ATTR, particularly in older adults.
In light-chain amyloidosis, the precursor is entirely different.
Abnormal plasma cells produce misfolded immunoglobulin light chains that can aggregate into amyloid deposits and exert direct proteotoxic effects. Cardiac involvement is particularly consequential because it strongly influences prognosis and requires rapid coordination between haematology and cardiology.
The diseases therefore share the phenomenon of protein misfolding without being interchangeable.
That distinction becomes important as researchers try to identify common mechanisms.
Historically, amyloid disease was often conceptualised largely as a problem of deposition. Enough insoluble material accumulates in an organ, tissue structure becomes distorted and function deteriorates.
The modern model is more complicated.
Soluble misfolded proteins and oligomeric intermediates can themselves be biologically toxic. In AL amyloidosis, for example, amyloidogenic light chains can directly impair cardiomyocyte function through oxidative and cellular stress pathways even before the burden of extracellular fibril deposition explains the entire phenotype.
The cardiovascular consequences can therefore emerge from both what accumulates and what remains soluble.
That matters because therapies targeting precursor production, protein stability or deposited fibrils are intervening at different points in the disease cascade.
The same review expands the problem beyond cardiomyocytes.
Amyloid diseases also affect the vasculature.
Microvascular dysfunction has been demonstrated in cardiac amyloidosis, and vascular deposition can occur in coronary and systemic vessels. Endothelial injury, abnormal vasoreactivity and autonomic dysfunction may contribute to symptoms that cannot be explained solely by myocardial wall thickness.
Thrombosis adds another layer.
Patients with cardiac amyloidosis can develop intracardiac thrombi and thromboembolic events, partly because amyloid infiltration can disrupt atrial mechanical function even when conventional rhythm-based risk assessment appears reassuring.
The result is a disease phenotype that does not fit neatly into the standard architecture of heart failure care.
The myocardium, microvasculature, conduction system, autonomic nervous system and coagulation pathways may all be involved simultaneously.
This increasingly explains why clinical guidance treats cardiac amyloidosis as more than a conventional hypertrophic cardiomyopathy.
Diagnosis has also changed dramatically.
ATTR cardiomyopathy once depended heavily on tissue biopsy.
Today, appropriately selected patients can be diagnosed non-invasively using bone-avid radionuclide scintigraphy when monoclonal protein testing has excluded a plasma-cell process.
That distinction is crucial.
A positive cardiac scan cannot simply be labelled ATTR without first excluding AL amyloidosis because the treatment pathways are fundamentally different and delay in AL can be particularly dangerous.
The 2026 European heart-failure guideline now incorporates this diagnostic pathway directly into heart-failure care, recommending serum and urine immunofixation, serum free light chains and appropriate bone scintigraphy when cardiac amyloidosis is suspected.
The therapeutic reason for finding ATTR has become increasingly compelling.
Tafamidis established that stabilising transthyretin could improve outcomes in ATTR cardiomyopathy. Acoramidis subsequently added another stabiliser, while RNA-targeting strategies such as vutrisiran demonstrated that reducing production of the precursor protein could also alter cardiovascular outcomes.
The 2026 European heart-failure guideline now gives a Class I recommendation to disease-modifying treatment with a proven TTR stabiliser or silencer in appropriate patients with hereditary or wild-type ATTR cardiomyopathy and symptomatic heart failure.
That changes the economics of missed diagnosis.
A decade ago, identifying ATTR earlier might have improved explanation, prognosis and supportive management while leaving relatively limited disease-specific treatment options.
Now diagnostic delay can mean allowing a treatable proteinopathy to progress while therapies capable of slowing the disease remain unused.
But therapy has not made the disease simple.
Even when production of the pathogenic precursor is reduced or the transthyretin tetramer is stabilised, established cardiovascular abnormalities may remain.
Amyloid already deposited in the myocardium does not necessarily vanish immediately. Fibrosis, microvascular dysfunction, autonomic impairment, arrhythmia and altered atrial mechanics can continue to influence outcomes.
This is the residual-risk problem highlighted by the review.
The therapeutic objective may eventually have to extend from slowing additional amyloid formation toward removing existing deposits, restoring organ function and treating the downstream cardiovascular consequences that remain after the protein pathway is modified.
Fibril-depleting antibodies are one potential approach.
Gene editing represents another more radical strategy upstream.
In-vivo CRISPR editing has already demonstrated that transthyretin production can be reduced by modifying the TTR gene in the liver, although the long-term balance of efficacy, safety and reversibility remains under investigation.
These technologies illustrate how dramatically amyloidosis has moved from diagnostic obscurity toward therapeutic innovation.
They also sharpen the screening problem.
More effective treatment creates pressure to find disease earlier.
The traditional diagnostic pathway often begins when a patient has already developed heart failure, unexplained left-ventricular thickening, characteristic imaging abnormalities or a combination of extracardiac red flags such as bilateral carpal tunnel syndrome and lumbar spinal stenosis.
Some of those clues can precede overt cardiomyopathy by years.
The question is whether health systems can identify the pattern before a clinician manually connects every clue.
This is where AI is entering the field.
Deep-learning systems have been developed to detect patterns associated with ATTR cardiomyopathy from ordinary electrocardiograms, echocardiograms and other routinely collected data.
A 2026 JAMA diagnostic study evaluated an image-based deep-learning model applied to routine 12-lead ECGs. Performance varied across external cohorts, but the study demonstrated that a widely available cardiovascular test could potentially be used to prioritise which patients should undergo definitive amyloidosis evaluation.
That is different from allowing AI to diagnose amyloidosis autonomously.
The clinically defensible role is triage.
An algorithm can identify a patient whose ECG contains a pattern associated with ATTR and prompt additional assessment. Definitive diagnosis still requires the appropriate clinical and laboratory pathway.
That distinction matters for implementation because the prevalence of ATTR in unselected populations is much lower than in specialised amyloidosis cohorts.
Even a good diagnostic model can generate large numbers of false-positive alerts when deployed indiscriminately across millions of low-risk patients.
This is why the review’s proposal for systematic screening needs careful interpretation.
Screening high-risk groups is not the same as population-wide testing.
Older adults with unexplained increased ventricular wall thickness, heart failure with preserved ejection fraction, characteristic musculoskeletal history, conduction disease or other red flags represent a fundamentally different pre-test population from healthy adults with no cardiac abnormalities.
Precision screening means choosing the population as carefully as choosing the test.
That may become even more important as the concept of protein misfolding expands beyond ATTR and AL.
The review examines amyloid-beta as one component of a heart-vascular-brain axis linking neurodegeneration and cardiovascular disease.
Amyloid-beta has been detected in cardiovascular tissues, linked with endothelial dysfunction and atherosclerotic processes, and circulating amyloid-beta measures have been associated with cardiovascular outcomes in some cohorts.
These observations are biologically provocative.
They are not equivalent to the established clinical role of transthyretin in ATTR cardiomyopathy.
The same caution applies to medin.
Medin is generated from the protein lactadherin and is extraordinarily common in ageing arteries. Experimental work suggests it can contribute to endothelial dysfunction, vascular inflammation and interactions with other amyloid proteins, including amyloid-beta.
It may therefore be relevant to vascular ageing and cerebral amyloid pathology.
But there is currently no clinical medin-screening pathway comparable with ATTR testing, no established medin-directed cardiovascular therapy and no basis for routine clinical measurement.
Alpha-synuclein provides another example of the heart-brain relationship.
Parkinson’s disease and other synucleinopathies commonly involve autonomic dysfunction, which can manifest as orthostatic hypotension, altered blood-pressure regulation and other cardiovascular abnormalities.
Here again, the cardiovascular phenotype is clinically real while the mechanistic route from alpha-synuclein aggregation to every observed cardiovascular outcome remains less defined than the amyloid cardiomyopathy pathway.
The review’s value is therefore not that it turns all protein misfolding disorders into one cardiovascular diagnosis.
It does the opposite.
It creates a framework for asking which cardiovascular mechanisms are shared and where disease-specific biology still matters.
Proteotoxic stress, inflammation, endothelial dysfunction and impaired protein quality control may provide common language.
The precursor protein, tissue tropism, organ phenotype and therapeutic target remain specific.
That distinction should shape drug development.
Therapies that reduce transthyretin production cannot simply be extrapolated to AL amyloidosis, where a pathological plasma-cell clone generates the precursor protein. An anti-amyloid strategy developed for Alzheimer’s disease cannot be assumed to remove cardiac transthyretin deposits safely or effectively.
Even within ATTR, inherited variants can produce different combinations of cardiac and neurological disease.
Precision medicine is therefore not optional in protein misfolding.
The disease is defined partly by which protein is misfolding, why it is misfolding and where the resulting structures accumulate.
The review pushes that logic even further by discussing conformational “strains”, in which aggregates formed from the same precursor protein may adopt different structures with different biological behaviours.
If that field matures clinically, future treatment could potentially depend not only on the gene and protein involved but also on the structural form of the pathological aggregate.
That remains a research frontier.
The more immediate policy issue is earlier recognition of diseases for which treatment already exists.
Cardiac amyloidosis provides an unusually clear example of what happens when diagnostic technology, targeted therapy and machine learning converge.
The heart-failure clinic identifies a phenotype. Laboratory testing distinguishes ATTR from a monoclonal process. Nuclear imaging can establish ATTR non-invasively in appropriate patients. Genetic testing can separate hereditary from wild-type disease. Disease-modifying therapies can then target the transthyretin pathway.
AI may increasingly sit at the beginning of that sequence, helping determine who deserves a closer look.
That could be valuable if it shortens diagnostic delay.
It could also overwhelm clinical services if algorithms generate more suspected cases than specialist pathways can evaluate.
Implementation therefore needs capacity planning, referral criteria and evidence that earlier algorithmic detection actually changes outcomes.
An AUROC is not a healthcare pathway.
The same applies to biomarker development.
The review calls for biomarkers capable of capturing vascular stress, proteotoxic burden and dynamic changes in proteostasis.
Such markers could transform trials by showing whether a therapy is affecting the underlying biological process before conventional clinical endpoints change.
But a biomarker becomes clinically useful only when its measurement changes a decision.
Protein misfolding research is full of measurable biological signals. Health systems need to know which of those signals justify imaging, referral, treatment or surveillance.
This is where HTA and reimbursement policy will increasingly enter a field once considered too rare for mainstream cardiovascular planning.
ATTR cardiomyopathy is more commonly recognised than it was a decade ago, treatment options are multiplying and diagnostic pathways are moving into general heart-failure guidance.
That expansion creates predictable questions.
Should all older patients with heart failure and increased wall thickness undergo amyloidosis testing?
Which extracardiac red flags justify screening before overt heart failure appears?
Should AI-ECG tools be deployed automatically in high-risk populations?
What is the cost of finding more early disease when disease-modifying treatment itself may be expensive?
And how should health systems compare the cost of earlier intervention with the cost of allowing irreversible myocardial damage to accumulate?
Those questions will not be answered by molecular biology alone.
They are implementation questions created by scientific progress.
The broader protein-misfolding framework makes the horizon even wider.
If research eventually establishes actionable cardiovascular phenotypes in Alzheimer’s disease, Parkinson’s disease, medin-associated vascular disease or other proteinopathies, cardiovascular medicine could become part of disease management far earlier than it is today.
But the evidence threshold should remain high.
A mechanistic connection is not automatically a screening indication.
An association is not a treatment target.
And a biomarker is not useful merely because it can be measured.
ATTR and AL amyloidosis show what mature translation looks like: defined pathological proteins, clinically recognisable phenotypes, diagnostic pathways and therapies aimed at disease biology.
The rest of the protein-misfolding field is at different points along that trajectory.
The new review matters because it places them on the same cardiovascular map without pretending they have all arrived at the same destination.
Cardiology is learning that protein misfolding is not simply something pathologists find after organs fail. In the diseases where treatment now exists, the strategic question is increasingly whether health systems can recognise the proteotoxic process early enough to change what happens next.

