For decades, medicine has relied on a deceptively simple assumption: that many measurements in the blood remain relatively stable within each person. A cholesterol value, immune marker or liver enzyme can fluctuate from day to day, yet clinicians often interpret it against an individual’s broader biological baseline. A new 20-year study now suggests that this principle applies unevenly across the human blood proteome. Some proteins appear to preserve a remarkably consistent personal signature, while others shift substantially with age, disease and changing physiology. The findings offer a detailed view of how the body maintains systemic homeostasis—and how that stability may begin to fail long before serious illness becomes visible.
Published in Nature Health, the study by Wu, Yang, Cheng and colleagues followed 1,298 people from middle and later adulthood into older age. Researchers examined blood samples collected at four time points over two decades and measured 10,776 circulating protein markers. This scale allowed the team to move beyond the conventional study of a small number of clinical biomarkers and instead examine the temporal behavior of thousands of proteins simultaneously. The result is described as the first comprehensive map of long-term proteomic stability, revealing that the bloodstream is not governed by one universal pattern of ageing-related change.
Proteins circulating in the blood perform an enormous range of functions. They transport hormones and nutrients, regulate immune reactions, influence clotting, communicate between organs and help repair or remodel tissues. Their concentrations can be affected by genetic variation, diet, infection, medication, inflammation and organ function. Because of this complexity, two people of the same chronological age may have very different molecular profiles. The new study focused on a central question: when a protein level differs between individuals, does that difference represent a stable biological trait, or is it simply a temporary fluctuation?
The answer varied sharply across the proteome. Certain proteins maintained highly individual-specific levels across the 20-year observation period, effectively acting as molecular fingerprints. Other proteins were far more volatile, changing in response to life stage or health status. This heterogeneity distinguishes the broader proteome from routine clinical markers, whose within-person stability is generally more uniform. In technical terms, the researchers examined how much of the total variation in protein abundance was attributable to persistent differences between individuals and how much reflected changes within the same person over time. The balance between these two forms of variation differed considerably from one protein to another.
The researchers also found that long-term protein stability was largely preserved across sex and across the life stages represented in the cohort. That result suggests that many protein-specific patterns are not simply temporary features of adulthood or older age. At the same time, selected proteins displayed changes that may signal transitions between stages of life. These shifts could reflect evolving immune regulation, metabolic remodeling or changes in the function of tissues and organs. Rather than portraying ageing as a uniform molecular decline, the findings support a more nuanced model in which some biological systems remain tightly regulated while others become increasingly dynamic.
Genetics appeared to be a major influence on the most stable proteins. Proteins with persistent, person-specific levels were strongly shaped by inherited factors, indicating that an individual’s long-term circulating profile may be partly encoded in their genome. These stable proteins were also enriched for functions mediated through the circulation, particularly immune regulation. This does not mean that their concentrations are immune to environmental influences or disease. Instead, it suggests that genetic architecture helps establish a personal operating range, or homeostatic set point, against which later deviations can be measured.
That concept formed the basis for the study’s clinical applications. The investigators used stable proteins to construct biological fingerprints and then tested whether these molecular patterns could improve health prediction. In validation analyses using data from the UK Biobank, they defined a sparse and reliable protein panel for estimating biological age. Unlike chronological age, which counts years since birth, biological-age measures attempt to capture the accumulated condition of physiological systems. The researchers reported that the protein-based panel improved biological-age prediction and strengthened its association with ageing-related outcomes, suggesting that a carefully selected subset of stable markers may provide more useful information than simply measuring thousands of proteins indiscriminately.
The study also addressed a critical problem in risk assessment: whether a biomarker represents a person’s enduring susceptibility or merely a momentary response. The researchers distinguished lifelong-stable risk biomarkers from volatile markers for cardiovascular disease and dementia. This distinction could eventually help clinicians interpret blood tests more intelligently. A persistently elevated protein may indicate a long-term biological predisposition, whereas a rapidly changing protein may be more informative as an early warning signal of an active disease process. Both types can be clinically valuable, but they answer different questions and should not necessarily be treated in the same way.
Perhaps the most striking application involved 21 proteins used to define personalized homeostatic baselines. Instead of comparing every patient with a population average, this approach establishes what is normal for that individual and then searches for meaningful departures. The researchers reported that deviations from these personal baselines independently predicted mortality. The signals spanned metabolic, immune and tissue-remodeling pathways, implying that disruption of homeostasis can emerge across several physiological systems at once. The finding raises the possibility that future blood tests could detect a person’s loss of biological balance before conventional diagnostic thresholds are crossed.
The study does not mean that a single protein profile can determine an individual’s future, nor that routine medical testing should immediately be replaced. Protein concentrations are influenced by factors ranging from acute infection to medication use, and translating population-level associations into clinical decisions requires careful replication and prospective testing. Nevertheless, the work provides a framework for separating stable biological identity from short-term molecular noise. By combining long-term observation, broad proteomic measurement and external validation, it points toward a form of precision medicine built not only on disease thresholds, but also on each person’s own molecular history. The bloodstream, these results suggest, is both a record of who we are and a sensitive readout of when our internal systems begin to drift.
Subject of Research: Long-term stability and temporal dynamics of the human blood proteome during ageing and disease.
Article Title: Temporal dynamics of the human blood proteome in ageing and disease
Article References: Wu, W., Yang, Z., Cheng, L. et al. “Temporal dynamics of the human blood proteome in ageing and disease.” Nature Health (2026). https://doi.org/10.1038/s44360-026-00183-1
Image Credits: AI Generated
DOI: https://doi.org/10.1038/s44360-026-00183-1
Keywords: blood proteome, protein stability, ageing, systemic homeostasis, precision medicine, biological age, cardiovascular disease, dementia, mortality, personalized biomarkers, UK Biobank, immune regulation
