New Blood-Based Metabolic Vulnerability Index Predicts Mortality Risk in Older Adults

A simple blood test that reads the combined signatures of inflammation and metabolic dysfunction may offer clinicians one of the most powerful tools yet for predicting which older adults are at imminent risk of dying. In a new study published in PLOS Aging and Health, researchers report that a biomarker-derived measure called the Metabolic Vulnerability Index, or MVX, strongly predicted mortality over time horizons ranging from one year to a decade in a large longitudinal cohort of adults over the age of 70. The findings suggest that a single nuclear magnetic resonance scan of a blood sample could capture hidden physiological vulnerabilities that conventional clinical metrics miss entirely.

The research team, led by Margery A. Connelly of LabCorp and including investigators from Duke University and the University of Groningen, set out to address a persistent problem in geriatric medicine. Traditional clinical measures such as body mass index, cholesterol levels, and blood pressure fail to capture the complex interplay among metabolic health, chronic inflammation, and nutritional status that drives aging-related illness and death. Frailty questionnaires and self-reported health assessments attempt to fill this gap, but they are subjective, time-consuming, and often insensitive to the earliest biochemical shifts that precede clinical decline.

MVX was designed to close that gap by integrating four blood-borne signals into a single score. Two of the markers reflect chronic inflammation: the concentration of small high-density lipoprotein particles and GlycA, a nuclear magnetic resonance measure of systemic inflammation that reflects the abundance of acute-phase proteins such as C-reactive protein, haptoglobin, and alpha-1-acid glycoprotein circulating in the blood. The other two markers signal metabolic malnutrition: citrate, an intermediate of cellular energy metabolism, and branched-chain amino acids, which serve as indicators of protein intake and muscle metabolism. When these components fall into adverse patterns, the index rises, flagging what the investigators describe as metabolic vulnerability, a state in which inflammation and dysmetabolism converge.

To test whether this composite score carried real prognostic weight, the team applied a validated MVX algorithm to clinical nuclear magnetic resonance spectra from the Duke Established Populations for Epidemiologic Studies of the Elderly, a longitudinal cohort of 1,507 community-dwelling older adults with a mean age of 78.1 years, of whom 36 percent were men. Deaths from all causes were tracked over a median follow-up of 7.0 years, with follow-up extending as long as 18.9 years. This design allowed the researchers to ask not just whether MVX predicted death, but how far into the future its predictive signal remained useful.

The answer was striking. In statistical models adjusted for age, sex, and race, each one-standard-deviation increase in MVX score was associated with a threefold higher hazard of dying within one year, with a hazard ratio of 3.00 and a 95 percent confidence interval of 2.34 to 3.85, a result that reached a p-value below 0.0001. Even after the investigators further adjusted for body mass index, type 2 diabetes, smoking status, systolic blood pressure, total cholesterol, triglycerides, estimated glomerular filtration rate, and qualitative measures of health and physical function, the association persisted with only modest attenuation. The fully adjusted hazard ratio for one-year mortality was 2.35, with a confidence interval of 1.76 to 3.14.

The signal was not confined to the short term. Similar associations were observed for two-year, five-year, and ten-year mortality, although the strength of the association attenuated somewhat over longer follow-up periods, as would be expected for a biomarker snapshot taken at a single point in time. When participants were sorted into quartiles of MVX score, those in the highest quartile showed significantly greater mortality risk on Kaplan-Meier survival analysis, with a p-value below 0.0001. In other words, the small group of older adults whose blood profiles combined the strongest inflammatory and malnutrition signatures were the most likely to die, and they died sooner than their peers.

Perhaps the most clinically consequential finding concerns discrimination, the ability of a test to separate those who will experience an event from those who will not. The area under the receiver operating characteristic curve for one-year mortality was 0.77, with a confidence interval of 0.71 to 0.84. That figure outperformed questionnaire-based measures of health and physical function as well as each of the individual MVX components considered alone, and the differences were statistically significant at p below 0.0001. This pattern indicates that the value of MVX lies in the integration of its components: inflammation and metabolic malnutrition are each informative, but together they carry a prognostic signal that neither delivers on its own.

The biology behind the index offers clues about why this integration works. Small HDL particles are known to be remodeled during chronic inflammatory states, when enzymes and acute-phase proteins alter their composition and size. GlycA, meanwhile, aggregates the signal of multiple inflammatory glycoproteins into one quantitative readout, providing a stable measure of low-grade systemic inflammation that is common in aging. Low citrate and altered branched-chain amino acid levels, by contrast, point toward compromised energy metabolism and inadequate protein nutrition, conditions that predispose older adults to sarcopenia, frailty, and reduced physiological reserve. A single score that rises when both processes are active therefore captures a doubly vulnerable state that neither a cholesterol panel nor a routine physical examination would reveal.

The authors suggest that MVX may provide additional clinical information to support personalized risk stratification and guide targeted approaches for older adults. Because the test relies on clinical nuclear magnetic resonance spectra, an established laboratory platform, it could in principle be incorporated into routine blood work without new instrumentation. For clinicians, a high MVX score might prompt closer monitoring, nutritional assessment, or interventions aimed at reducing inflammation and preserving muscle mass, while a low score could offer reassurance. For researchers, the index provides a quantitative, mechanistically grounded endpoint for trials of interventions in aging populations.

As populations worldwide age, the demand for objective tools that identify older adults at greatest risk has never been greater. This study, drawing on a well-characterized cohort followed for nearly two decades, demonstrates that the convergent signatures of inflammation and dysmetabolism, read from a single blood draw, carry prognostic power that rivals and exceeds established clinical and functional assessments. Whether MVX can guide interventions that actually change outcomes remains to be tested, but as a window into the hidden metabolic vulnerabilities of aging, the index represents a meaningful step toward precision medicine for the oldest patients.

Subject of Research: A biomarker-derived Metabolic Vulnerability Index predicting all-cause mortality in older adults using nuclear magnetic resonance markers of inflammation and metabolic dysfunction

Article Title: Metabolic Vulnerability Index predicts mortality in older adults based on markers of inflammation and dysmetabolism

Article References: Connelly, M. A., Shalaurova, I., Ross, L. M., Dullaart, R. P. F., Kraus, W. E., & Byers Kraus, V. (2026). Metabolic Vulnerability Index predicts mortality in older adults based on markers of inflammation and dysmetabolism. PLOS Aging and Health, 1(4), e0000018. https://doi.org/10.1371/journal.page.0000018

Image Credits: AI Generated

DOI: 10.1371/journal.page.0000018

Keywords: Metabolic Vulnerability Index, MVX, mortality prediction, older adults, inflammation, GlycA, small HDL particles, branched-chain amino acids, citrate, nuclear magnetic resonance, biomarkers, aging