A provocative new study is suggesting a blood test tracking 14 different biomarkers can predict a person’s risk of dying within the next 10 years better than any conventional models. The research is still a long way off being broadly deployed in clinical environments and the test may be of most use in human drug trials as a surrogate endpoint for mortality.

The prospective blood test tracks 14 biomarkers the researchers identified as associated with all-cause mortality. The biomarkers identified were not related to specific diseases but instead represent general health, including markers for fatty acid metabolism, inflammation and glycolysis.

The researchers validated the test across a sample set of over 44,000 subjects spanning all age groups. The study suggests the predictive mortality score generated by the test is accurate across men and women of all ages.

One limitation in the research, raised by dementia researcher Amanda Heslegrave who did not work on this particular study, is the cohort is made up of people primarily of European descent, meaning results would need to be validated in other ethnicities before any clinical applications are engaged.

“The large numbers in the study are good and also the fact that they have a large number for outcome – in this case mortality – makes the data more viable,” says Heslegrave. “However, it is limited by the fact that being only European data it may not apply to other ethnic groups without further studies.”

One of the interesting outcomes for this kind of blood test is it could be used as a surrogate endpoint in clinical studies to represent overall mortality. The researchers note in the study that many clinical trials are obviously unable to run for long enough durations to encompass total mortality endpoints, so this test could be used to generate overall mortality scores evaluating a drug’s potential efficacy.

“… it may be used as a surrogate endpoint for clinical trials in older individuals, since showing (a reduction in) the total mortality endpoint is mostly not feasible due to the limited duration and number of cases in a regular clinical trial,” the researchers write in their recently published paper.

Eline Slagboom, one of the researchers on the paper from Leiden University, also suggests an outcome from this test could be to help clinicians determine more effective treatment strategies in elderly patients.

“The calendar age just doesn’t say very much about the general state of health of elderly people: one 70-year-old is healthy, while another may already be suffering from three diseases,” says Slagboom. “We now have a set of biomarkers which may help to identify vulnerable elderly people, who could subsequently be treated.”

This isn’t the first blood test designed to predict a person’s risk of dying. A Yale University study from last year, for example, revealed a similar test, based around nine key blood-based biomarkers. Perhaps even more striking are the number of AI systems being developed that purport to be able to study a patient’s medical scans and effectively predict how much longer they have to live.

The new research was published in the journal Nature Communications.