The research, led by Constantin L. Palm of the University of Groningen together with Niels T. B. Scholte of Erasmus MC Rotterdam and a large team of Dutch collaborators, set out to track how plasma ketone bodies evolve in the months following a heart failure hospitalization, and whether those changing levels could predict who would die or be readmitted. Rather than relying on a single blood draw, the investigators measured ketones repeatedly, up to seven times over the course of a year in their primary cohort, using nuclear magnetic resonance spectroscopy, a technique that quantifies β-hydroxybutyrate, acetoacetate and acetone simultaneously with remarkable precision. The NMR results correlated almost perfectly with mass spectrometry reference methods, with R-squared values above 0.99 for each individual ketone species.

The primary dataset came from the TRIUMPH study, an observational prospective trial that enrolled patients hospitalized for acute heart failure across fourteen centers in the Netherlands. Of the 496 patients enrolled, 467 had detectable ketone body measurements and formed the analytical population. These were seriously ill individuals: the median age was 74 years, only 37 percent were women, and a striking 82 percent were in New York Heart Association class III or IV, meaning they experienced symptoms even during minimal physical activity. Blood was collected on admission, on days two and four of the hospital stay, on the day of discharge, and at follow-up visits spanning the subsequent year, yielding a median of four serial samples per patient.

The findings were unambiguous. Patients who went on to experience the combined endpoint of death from any cause or rehospitalization for heart failure, 185 patients or 40 percent of the cohort, carried continuously higher estimated ketone body levels than those who remained event-free. Using sophisticated joint models that combine longitudinal mixed-effects regression with Cox proportional hazards analysis, the researchers calculated that each doubling of total ketone body concentration at any point during follow-up was associated with a 2.10-fold increase in the hazard of the combined endpoint, after adjustment for age, sex, disease severity, kidney function, natriuretic peptide levels and a battery of other clinical confounders. Per doubling, the hazard ratios were 1.62 for acetone and 1.62 for β-hydroxybutyrate, both statistically significant.

Perhaps the most surprising discovery was temporal. The team had expected, based on earlier work showing that ketones fall as patients recover from acute decompensation, that ketone levels would spike in the days or weeks preceding a catastrophic event, much like rising troponin precedes a myocardial infarction. Instead, the estimated ketone trajectories remained persistently elevated in high-risk patients throughout the entire follow-up period, without a dramatic pre-event surge. This pattern suggests that ketone bodies are not simply an acute marker of hemodynamic stress, as natriuretic peptides are, but rather a window into a chronic, smoldering catabolic state that distinguishes patients destined for poor outcomes long before the event itself arrives.

To guard against the possibility that the finding was an artifact of one particular cohort, the researchers turned to Bio-SHiFT, an independent study of 397 patients with chronic heart failure followed at outpatient clinics in Rotterdam and Alkmaar. These patients were younger, with a median age of 64 years, and provided up to ten repeated blood samples over a median follow-up of 2.1 years. During that time, 124 patients, or 31.2 percent, reached the composite endpoint of cardiovascular death, heart failure hospitalization, left ventricular assist device implantation or heart transplantation. Once again, ketone bodies were consistently higher in those who reached the endpoint, with a hazard ratio of 5.36 per doubling of total ketone bodies and 4.24 per doubling of acetone in the adjusted models, an even stronger signal than in the acute cohort.

The biology behind these numbers is intricate. Ketone bodies are produced by the liver during fasting, prolonged exercise or metabolic stress, and in heart failure the failing myocardium increasingly relies on them as fuel, a shift that experimental studies suggest may actually be protective. Supplementation with ketone esters has improved cardiac function in animal models and in early clinical trials, which makes the association between high endogenous ketones and poor outcomes seem paradoxical. The authors are careful to address this: elevated ketones are unlikely to cause harm directly. Rather, they appear to be a compensatory response, a metabolic distress signal broadcast by a body under sustained siege from inflammation, sympathetic overactivation and catabolic wasting.

Several mechanistic threads support this interpretation. The researchers observed that higher ketone levels correlated with worse ejection fraction, higher NT-proBNP, and a reduced likelihood of tolerating beta-blocker therapy, all hallmarks of advanced disease. They propose that inflammation-driven malabsorption creates a fasting-like state that promotes ketogenesis, while heightened activity of the hypothalamic–pituitary axis floods the circulation with catecholamines and cortisol, driving muscle breakdown and insulin resistance that further fuel ketone production. Intriguingly, acetoacetate, the ketone species most subject to rapid interconversion and spontaneous decarboxylation into acetone, showed no association with outcomes in either cohort, whereas the more stable acetone and β-hydroxybutyrate did, suggesting that acetone may best reflect net ketogenesis over time.

The clinical implications are considerable. Dynamic discrimination analysis showed that serially measured ketone bodies performed comparably to NT-proBNP, the current gold-standard prognostic biomarker, with time-dependent area under the curve values ranging from 0.76 to 0.80 in TRIUMPH and 0.77 to 0.82 in Bio-SHiFT. Because ketones capture a dimension of heart failure, chronic metabolic stress, that established biomarkers largely ignore, repeated measurements could complement rather than replace existing tools, helping clinicians identify a subset of patients living in a state of continuous catabolic strain who might benefit from intensified monitoring, metabolic interventions or targeted nutritional strategies. The authors caution that their study was observational, confined to Dutch centers, and included few women, and that the era before SGLT2 inhibitor therapy, drugs known to raise ketone levels, may limit generalizability to contemporary patients. Still, the message is clear: in heart failure, what the liver tells the blood about the body’s metabolic state, told repeatedly over time, may be one of the most honest accounts of the heart’s declining fortunes.

Subject of Research: Serial measurement of plasma ketone bodies as a prognostic biomarker for adverse outcomes in heart failure patients

Article Title: Continuously elevated plasma ketone bodies are linked to adverse outcomes in patients with heart failure

Article References: Palm, C. L., Scholte, N. T. B., Lu, L., Kamar, S. A., Bakker, B. M., Voorrips, S. N., Mahmoud, B., Connelly, M. A., Suthahar, N., de Boer, R. A., Umans, V. A., Akkerhuis, K. M., Hillege, H., Boersma, E., Kardys, I., & Westenbrink, B. D. (2026). Continuously elevated plasma ketone bodies are linked to adverse outcomes in patients with heart failure. Clinical Research in Cardiology. https://doi.org/10.1007/s00392-026-02992-6

Image Credits: AI Generated

DOI: 10.1007/s00392-026-02992-6

Keywords: heart failure, ketone bodies, biomarkers, NT-proBNP, metabolism, prognosis, TRIUMPH study, Bio-SHiFT, nuclear magnetic resonance, cardiology, catabolism, risk stratification