Background: Identification of healthy people at high risk for severe COVID-19 is a global health priority. We investigated whether blood biomarkers measured by high-throughput metabolomics could be predictive of severe pneumonia and COVID-19 hospitalisation years after the blood sampling.
Methods: Nuclear magnetic resonance metabolomics was used to quantify a comprehensive biomarker profile in 105,146 plasma samples collected in the UK Biobank during 2007-2010 (age range 39-70). The biomarkers were tested for association with severe pneumonia (2507 cases, defined as diagnosis in hospital or death record occurring during a median of 8.1-year follow-up) and with severe COVID-19 (195 cases, defined as diagnosis in hospital between mid-March to mid-June 2020). A multi-biomarker score was derived for prediction of severe pneumonia based on half of the study population and validated in the other half. We explored how this biomarker score relates to the risk of severe COVID-19.
Findings: The biomarker associations with risk of severe COVID-19 followed an overall pattern similar to associations with risk of severe pneumonia (correlation 0.83). The multi-biomarker score, comprised of 25 blood biomarkers including inflammatory proteins, fatty acids, amino acids and advanced lipid measures, was strongly associated with risk of severe pneumonia (odds ratio 1.67 per standard deviation [95% confidence interval 1.59-1.76]; 3.8-fold risk increase for individuals in upper vs lower quintile). The multi-biomarker score was also associated with risk of severe COVID-19 (odds ratio 1.33 [1.17-1.53]; 2.5-fold risk for upper vs lower quintile) and remained significant when adjusting for body mass index, smoking, and existing respiratory and cardiometabolic diseases. Mimicking the decade lag from blood sampling to COVID-19, severe pneumonia events occurring after 7-11 years associated with the multi-biomarker score to a similar magnitude (odds ratio 1.43 [1.29-1.59]; 2.6-fold risk for upper vs lower quintile) as for severe COVID-19. However, the short-term risk of severe pneumonia events associated to the multi-biomarker score at even 3 times higher magnitude (odds ratio 2.21 [1.95-2.50]; 8.0-fold risk for upper vs lower quintile in analysis of the first 2 years after blood sampling).
Interpretation: In decade-old blood samples from the UK Biobank, a biomarker score measured by high-throughput metabolomics is indicative of the risk for severe COVID-19. The molecular signature of biomarker changes reflective of risk for severe COVID-19 is similar to that for severe pneumonia, in particular when accounting for the time lag to the COVID-19 pandemic. The even stronger association of the biomarker score with 2-year risk for severe pneumonia lends support to promising screening possibilities for identifying people at high risk for severe COVID-19.
Blood biomarker score identifies individuals at high risk for severe COVID-19 a decade prior to diagnosis: metabolic profiling of 105,000 adults in the UK Biobank