Multi-variable biomarker approach in identifying incident heart failure in chronic kidney disease: results from the Chronic Renal Insufficiency Cohort study

Scott E. Janus, Jamal Hajjari, Tarek Chami, Haytham Mously, Anshul K. Badhwar, Mohamad Karnib, Herman Carneiro, Mahboob Rahman, Sadeer G. Al-Kindi

Research output: Contribution to journalArticlepeer-review

7 Scopus citations


Aims: Heart failure (HF) is one of the leading causes of cardiovascular morbidity and mortality in the ever-growing population of patients with chronic kidney disease (CKD). There is a need to enhance early prediction to initiate treatment in CKD. We sought to study the feasibility of a multi-variable biomarker approach to predict incident HF risk in CKD. Methods and results: We examined 3182 adults enrolled in the Chronic Renal Insufficiency Cohort (CRIC) without prevalent HF who underwent serum/plasma assays for 11 blood biomarkers at baseline visit (B-type natriuretic peptide [BNP], CXC motif chemokine ligand 12, fibrinogen, fractalkine, high-sensitivity C-reactive protein, myeloperoxidase, high-sensitivity troponin T (hsTnT), fibroblast growth factor 23 [FGF23], neutrophil gelatinase-associated lipocalin, fetuin A, aldosterone). The population was randomly divided into derivation (n = 1629) and validation (n = 1553) cohorts. Biomarkers that were associated with HF after adjustment for established HF risk factors were combined into an overall biomarker score (number of biomarkers above the Youden's index cut-off value). Cox regression was used to explore the predictive role of a biomarker panel to predict incident HF. A total of 411 patients developed incident HF at a median follow-up of 7 years. In the derivation cohort, four biomarkers were associated with HF (BNP, FGF23, fibrinogen, hsTnT). In a model combining all four biomarkers, BNP (hazard ratio [HR] 2.96 [95% confidence interval 2.14–4.09]), FGF23 (HR 1.74 [1.30–2.32]), fibrinogen (HR 2.40 [1.74–3.30]), and hsTnT (HR 2.89 [2.06–4.04]) were associated with incident HF. The incidence of HF increased with the biomarker score, to a similar degree in both derivation and validation cohorts: from 2.0% in score of 0% to 46.6% in score of 4 in the derivation cohort to 2.4% in score of 0% to 43.5% in score of 4 in the validation cohort. A model incorporating biomarkers in addition to clinical factors reclassified risk in 601 (19%) participants (352 [11%] participants to higher risk and 249 [8%] to lower risk) compared with clinical risk model alone (net reclassification improvement of 0.16). Conclusion: A basic panel of four blood biomarkers (BNP, FGF23, fibrinogen, and hsTnT) can be used as a standalone score to predict incident HF in patients with CKD allowing early identification of patients at high-risk for HF. Addition of biomarker score to clinical risk model modestly reclassifies HF risk and slightly improves discrimination.

Original languageEnglish (US)
Pages (from-to)988-995
Number of pages8
JournalEuropean Journal of Heart Failure
Issue number6
StatePublished - Jun 2022


  • Biomarker
  • Chronic kidney disease
  • Heart failure
  • High-sensitivity troponin

ASJC Scopus subject areas

  • Cardiology and Cardiovascular Medicine


Dive into the research topics of 'Multi-variable biomarker approach in identifying incident heart failure in chronic kidney disease: results from the Chronic Renal Insufficiency Cohort study'. Together they form a unique fingerprint.

Cite this