Heart failure & cardiomyopathiesCohort study

Serum Metabolomics Improves Risk Stratification for Incident Heart Failure

Eur J Heart FailPublished

  • Prospective cohort study (prediction model development and validation)
  • UK Biobank
  • elastic net models
  • per-metabolite Cox models

Authors' conclusion

Serum metabolomics improves incident HF risk prediction over PCP-HF. Scores based on age, sex and metabolomics exhibit similar predictive power to clinically-based models

Study context

Population
UK Biobank participants
Participants
68,311 participants · >0.8 million person-years
Exposure
PCP-HF + serum metabolomics vs PCP-HF alone
Outcome
Prediction of incident HF
Follow-up
Median ≈12.3 years

Main finding

As reported in the abstract

Incident HF prediction

PCP-HF + metabolomics

0.768

Harrel's C

PCP-HF

0.755

Harrel's C

Each group vs PCP-HFΔC (95% CI)
PCP-HF + metabolomics
0.013(0.004–0.022)
PCP-HF + metabolomics
continuous NRI 0.287 (95% CI 0.200–0.367)
PCP-HF + metabolomics
relative IDI 17.47% (95% CI 9.463–27.825)

Observational data: associations, not effects.

Further findings

OutcomePCP-HF + metabolomicsPCP-HFAge, sex and metabolomicsΔC (95% CI)
Age, sex and metabolomics model vs PCP-HF0.010 (−0.004 to 0.027)
Harrel's CNot reported in the abstract0.7550.745
OutcomeEstimate
Age, sex and metabolomics model vs PCP-HFcontinuous NRI 0.097 (95% CI −0.025 to 0.217)
Age, sex and metabolomics model vs PCP-HFrelative IDI 13.445% (95% CI −10.608 to 41.454)

Individual metabolite associations with incident heart failure (HF)

Metabolites independently associated, adjusted for age and sex
90/168
Metabolites independently associated, adjusted for PCP-HF
48/168

Original publication

Serum Metabolomics Improves Risk Stratification for Incident Heart Failure. Eur J Heart Fail. Published 16 April 2024. doi:10.1002/ejhf.3226

Summarised from the published abstract; every number is checked against it