IPM Take
Cardiovascular prevention usually waits for measurable risk to appear: obesity, hypertension, dyslipidaemia, insulin resistance or abnormal glucose.
This study asks whether the biology starts signalling trouble much earlier.
Researchers profiled circulating proteins in 273 Hispanic or Latino children and adolescents from a high-risk community on the US-Mexico border and linked them to 25 cardiovascular-kidney-metabolic disease, or CKMD, traits. More than 1,000 proteins were associated with at least one paediatric phenotype, spanning pathways related to insulin sensitivity, beta-cell function, liver biology, inflammation and lipid metabolism. (nature.com)
When the researchers carried childhood-derived protein signatures into adult datasets, those signatures were associated with corresponding cardiometabolic traits and with later clinical outcomes in adults, including type 2 diabetes and cardiovascular disease. (nature.com)
That is scientifically provocative. It does not mean a blood test can now tell an eight-year-old whether they will develop heart disease.
The stronger message is that precision prevention may eventually have to move upstream, identifying biological vulnerability before conventional thresholds declare someone “high risk.”
Executive Summary
The study, published in Nature Metabolism, analysed 273 Hispanic or Latino children and adolescents, with a mean age of 13.1 years, from the Border Health Research Cohort in Cameron County, Texas. More than one-third had obesity alongside features such as insulin resistance, elevated blood pressure or pro-atherogenic dyslipidaemia. (nature.com)
Researchers linked circulating proteins to 25 CKMD-associated phenotypes covering adiposity, liver health, vascular traits, kidney function and dysglycaemia. They identified 1,064 proteins associated with at least one paediatric CKMD phenotype and used machine-learning methods to construct multi-protein signatures representing six broader biological domains. (nature.com)
The childhood-derived signatures were then evaluated in adults. The authors found substantial concordance with cardiometabolic biology in adults from the same community and applied the signatures to 28,256 UK Biobank participants, where several were associated with incident type 2 diabetes, cardiovascular disease, fatty liver disease, sleep apnoea and all-cause mortality over long-term follow-up. (nature.com)
The investigators also compared paediatric CKMD-associated proteins with existing proteomic data from adults with obesity treated with semaglutide in the STEP 1 trial. Many proteins associated with more adverse childhood CKMD traits changed in a favourable direction during adult semaglutide treatment. However, the children in this study were not treated with semaglutide, and the analysis does not show that GLP-1 therapy reverses these molecular patterns or prevents future cardiovascular disease in children. (nature.com)
Why it matters
- HTA bodies: Proteomic risk tools could eventually create a new category of preventive diagnostic, but value would depend on whether testing improves prediction beyond simpler clinical measures and, crucially, whether acting on the result changes long-term outcomes.
- Payers: Earlier risk identification could help concentrate intensive prevention on children most likely to benefit, but population-scale proteomic screening could also create substantial testing and treatment costs. Evidence of clinical utility would be essential before reimbursement.
- Industry / innovation partners: The study provides a potential framework for biomarker-led paediatric prevention and for identifying biologically defined populations in obesity and cardiometabolic trials. The opportunity is substantial, but translating a molecular signature into a clinically useful test will require prospective validation across more diverse populations.
Cardiovascular disease rarely begins with the first heart attack.
Long before clinical events occur, metabolic, inflammatory and vascular changes can accumulate across decades. Childhood obesity, insulin resistance, hypertension and abnormal lipids are already known to track into adult cardiovascular risk.
What remains difficult is deciding which children with apparently modest abnormalities are on the most dangerous trajectory.
Researchers behind the new Nature Metabolism study approached that question through the circulating proteome, measuring thousands of proteins that reflect biological activity across tissues and organ systems.
Among 273 children and adolescents, they found 2,916 statistically significant associations between proteins and CKMD-related traits. Sixty-five proteins were associated with at least four disease domains, suggesting that some molecular signals may reflect interconnected processes spanning obesity, insulin resistance, liver disease and cardiovascular risk rather than a single abnormal laboratory value. (nature.com)
The researchers then created multi-protein scores representing broader biological patterns.
Some of the most concerning signatures were those related to pro-inflammatory adiposity, liver fat and fibrosis, and insulin resistance. When those childhood-derived scores were calculated in adults in UK Biobank, higher scores were associated with increased risks of type 2 diabetes and multiple cardiovascular and metabolic outcomes. (nature.com)
That does not mean the study followed these 273 children for decades and showed which of them developed disease.
It did not.
Instead, researchers identified protein patterns associated with childhood CKMD traits, then tested whether the same mathematical signatures carried meaningful information when applied to adult populations. That is an important distinction because it makes this a biomarker-discovery study, not a validated paediatric prediction tool.
The semaglutide finding is intriguing, but easy to overstate
One of the most attention-grabbing findings involves GLP-1 therapy.
The researchers compared 1,032 proteins associated with paediatric CKMD phenotypes with previously published proteomic data from adults with obesity who participated in the STEP 1 semaglutide trial.
Proteins associated with worse metabolic health in children tended to move in a more favourable direction among adults receiving semaglutide. Examples included leptin, FABP4, CES1 and ACY1, which were associated with adiposity, insulin resistance or liver abnormalities in the paediatric cohort and declined during treatment in adults. (nature.com)
That strengthens the biological plausibility that some of these proteins reflect modifiable metabolic states.
It does not demonstrate that semaglutide should be prescribed to children because of a proteomic result.
The semaglutide analysis used adults. Protein changes could also partly reflect weight loss or other systemic effects of therapy. Whether changing these proteins themselves matters for future cardiovascular outcomes remains uncertain.
This distinction becomes increasingly important as paediatric GLP-1 use grows. The authors note that prescriptions of GLP-1 receptor agonists among young people increased sharply between 2020 and 2023. Precision biomarkers could theoretically help distinguish children at particularly high long-term risk from those for whom the balance of pharmacological treatment, cost and uncertainty may be less favourable. (nature.com)
But that future requires considerably more evidence.
The paediatric cohort was relatively small and came from one predominantly Hispanic or Latino community with a high prevalence of metabolic risk. The signatures need prospective validation in larger and more diverse paediatric populations, ideally with long-term follow-up showing whether they outperform established risk factors and accurately predict who develops disease.
Most importantly, researchers will eventually need to show that knowing the proteomic result changes care in a way that improves health.
Without that step, proteomics risks producing more precise descriptions of risk without more effective prevention.
The study nevertheless challenges one of medicine’s most persistent assumptions: that cardiovascular prevention belongs mainly to adulthood.
The molecular biology of adult disease may already be visible decades earlier.
The policy question is whether health systems can learn to act earlier without turning biological risk into premature diagnosis and overtreatment.
