Background: Predictive disease risk biomarkers that can be linked to exposure have proved difficult to identify in case-control studies.
Methods: Parallel statistical analysis of the correlation between (1)H NMR profiles from plasma samples collected before disease onset (EPIC cohort), versus exposure to dietary compounds, and follow-up disease endpoints (colon and breast cancer) was performed.
Results: Metabonomic signatures associated with colon cancer and dietary fiber intake (a protective factor according to epidemiological studies) were identified.
Conclusion: This implementation of the novel "meet-in-the-middle" analytical strategy indicates how case-control studies nested in prospectively collected cohorts may reveal intermediate biomarkers linking exposure and disease.