Scientists found a link between Mediterranean diet blood markers and longer survival

By Vijay Kumar MalesuReviewed by Susha Cheriyedath, M.Sc.Oct 20 2026

Could a blood test reveal what food questionnaires miss? Researchers examined 67 blood metabolites to investigate how dietary habits and the body's metabolic responses relate to long-term survival.

Study: A metabolomic signature of the Mediterranean diet predicts mortality in 4 independent cohorts. Image Credit: Spalnic / Shutterstock

In a recent study published in The American Journal of Clinical Nutrition, researchers evaluated whether a 67-metabolite plasma signature of Mediterranean diet (MedDiet) adherence is associated with long-term all-cause and cause-specific mortality and whether the association can be replicated across independent cohorts.

Background

Can a blood sample provide a more objective picture of dietary habits and their relationship with long-term survival? The MedDiet includes olive oil, fruit, vegetables, legumes, nuts, and fish. This dietary pattern has been linked to lower risks of cardiovascular diseases, type 2 diabetes, and other chronic diseases. Yet dietary adherence has traditionally been assessed using self-reported food frequency questionnaires (FFQs), which may misclassify actual intake. A previously developed plasma metabolomic signature comprising 67 metabolites was associated with lower future cardiovascular risk. Further research is needed to establish whether this biological signature predicts long-term mortality across diverse populations.

About the study

The researchers analyzed data from the Spanish multicenter PREvención con DIeta MEDiterránea (PREDIMED) randomized nutritional intervention trial, conducted from 2003 to 2011. The trial enrolled adults at high cardiovascular risk who did not have cardiovascular disease at baseline. Participants were randomly assigned to three groups: one received the MedDiet supplemented with extra-virgin olive oil, another followed the MedDiet supplemented with mixed nuts, and a third followed a low-fat diet. Of the 7,447 volunteers, 1,880 had baseline plasma metabolomic data, and 1,648 also had measurements after 1 year, with approximately 13.2 years of follow-up.

The researchers applied a previously developed and validated signature of 67 metabolites, selected from 302 analyzed compounds, to assess dietary adherence. Plasma samples were analyzed using liquid chromatography-tandem mass spectrometry, including hydrophilic interaction liquid chromatography (HILIC) methods. Mortality was assessed using medical records and death registries. The main outcome was all-cause mortality, whereas cardiovascular and cancer mortality were the secondary outcomes.

The findings were independently replicated in 12,704 participants from the Nurses’ Health Studies I and II (NHS I and NHS II) and the Health Professionals Follow-up Study (HPFS), with an average follow-up of 25 years. The authors used Cox regression models to estimate mortality associations after adjusting for relevant sociodemographic and clinical factors. Sensitivity analyses examined whether the findings remained consistent under different assumptions.

Study results

In the PREDIMED analysis, participants had a mean age of 67.1 years, and 57.5% were women. Those with higher baseline metabolomic signature scores generally demonstrated better adherence to the MedDiet and were less likely to smoke or have type 2 diabetes.

Participants in the highest third of signature scores had a 23% lower relative hazard of death than those in the lowest third. Each standard deviation (SD) increase in the signature was associated with a 19% lower relative hazard of death. The finding remained similar after adjustment for reported dietary adherence and other relevant factors. Sensitivity analyses excluding deaths during the first 2 years, limiting follow-up to 10 years, stratifying participants by age, sex, and diabetes status, and accounting for medication use produced broadly consistent findings.

The signature measured after 1 year also corresponded to lower all-cause and cardiovascular mortality. Each SD increase in the 1-year signature was associated with a 26% lower relative hazard of cardiovascular death. The association with cancer mortality was not statistically significant. The average signature across baseline and year 1 showed similar relationships with all-cause mortality and cardiovascular mortality. Changes in the signature from baseline to year 1 were not significantly associated with mortality.

In analyses combining the baseline signature with randomized dietary assignment, participants in the control group with the lowest signature scores had the highest all-cause mortality. The interaction between signature level and intervention assignment was not statistically significant, so the findings did not establish that the dietary intervention modified the association.

The external replication involved 12,704 participants, with a mean age of 54.7 years; 83.2% were women. Throughout the follow-up period, a total of 4,111 deaths were recorded, comprising 845 deaths from cardiovascular illnesses and 1,231 deaths from cancer. Those in the highest fifth of the metabolomic signature scores had a 23% lower relative hazard of death than those in the lowest fifth. Each SD increase was associated with a lower relative hazard of all-cause and cardiovascular death. After full adjustment, the association with cancer mortality was not statistically significant. Inverse probability weighting analyses yielded similar findings, supporting the consistency of the results.

Conclusions

A previously established plasma metabolomic signature of MedDiet adherence was inversely associated with long-term all-cause mortality in the PREDIMED trial and three independent United States (US) cohorts. Higher signature scores were consistently associated with lower mortality, supporting the value of objective biological markers alongside self-reported dietary assessments. Associations with cardiovascular mortality were also observed, particularly for the 1-year signature in PREDIMED, while cancer mortality findings varied across analyses.

Because the signature-mortality analyses were observational, these associations do not establish that increasing the score or changing diet directly reduces mortality. The study also had limitations, including a smaller PREDIMED sample for assessing individual causes of death, metabolomic changes measured over only 1 year, and the profiling platform's inability to measure certain phenolic compounds and antioxidants. The selection of participants for the US metabolomic analyses may also limit how broadly the results apply.

Further research is needed to assess the signature's usefulness in clinical care. The signature may capture longer-term dietary and metabolic status beyond short-term changes and reflect individual metabolic responses that dietary questionnaires do not fully measure.