Heart failure risk assessment using biomarkers in patients with atrial fibrillation: analysis of 32,041 patients from the COMBINE-AF study
Bibliographic record
Abstract
Abstract Introduction Heart failure (HF) is one of the most common comorbidities in patients with AF, resulting in a clinical need for robust risk assessment. Purpose To investigate the incremental value of adding NTproBNP, high-sensitivity cardiac troponin T (hs-cTnT), and growth-differentiation factor-15 (GDF-15) to clinical predictors for HF risk stratification in patients with AF, and the individual contribution of each biomarker accounting for inter-biomarker correlation. Methods We used pooled individual patient data from 3 large RCTs investigating DOACs versus VKAs (ARISTOTLE, ENGAGE AF-TIMI 48, RE-LY) from the COMBINE-AF cohort and included all patients with available biomarkers at baseline. The composite primary endpoint was hospitalization for HF (HHF) or CV death (CVD), with HHF as secondary endpoint. We employed analyses of absolute risk and Cox-regression adjusting for clinical factors to assess the association of the individual biomarkers with both endpoints, testing incremental discrimination with the likelihood ratio test from nested models. To address the inter-biomarker correlation, weighted quantile sum (WQS) regression analysis summarizing the adjusted risk of all biomarkers (per quartile) in one index was applied, thereby exploring the individual and additive contribution of each biomarker to risk assessment. Results Data were available in 32,041 patients (median age 71 [IQR 64-77] years, 36.7% female, 23.0% with paroxysmal AF, 41.6% with established HF). Higher biomarker values were associated with a graded increase in absolute risk for HHF/CVD and HHF (Figure 1). In a model adjusting for clinical variables and all biomarkers, hs-cTnT (HR per 1-SD 1.39 [95% CI 1.33-1.44]), NT-proBNP (HR 1.67 [95% CI 1.58-1.76]), and GDF-15 (HR 1.20 [95% CI 1.15-1.25]) were associated with HHF/CVD. The c-index increased with addition of biomarkers (0.70 [0.69, 0.70] to 0.77 [0.76, 0.78]; Likelihood ratio test p<0.001). Results were similar for HHF alone. NTproBNP was significantly correlated with hs-cTnT (r=0.40, p<0.0001) and GDF-15 (r=0.36, p<0.001). To address this correlation, WQS regression analysis was conducted (Figure 2). NTproBNP and hs-cTnT contributed nearly equally to the risk assessment for HHF/CVD and HHF, with GDF-15 providing statistically significant but less contribution to risk assessment (Figure 2). Conclusion Hs-cTnT, NT-proBNP and GDF-15 contribute individually and additively to the risk assessment for HHF/CVD and HHF in patients with AF. Our analysis suggests that hs-cTnT is as important as NTproBNP for HF risk assessment, with significant but less contribution of GDF-15. Our findings support the possible future routine use of biomarkers to distinguish patients with AF at low or high risk for HF and could guide the introduction of therapies to mitigate the risk of HF events in this growing population.Event incidence by biomarker quartilesWeighted quantil sum regression analysis
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.008 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.003 | 0.004 |
| Bibliometrics | 0.002 | 0.003 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".