Baseline cardiovascular risk stratification in multiple myeloma: a simple clinical score outperforming established risk models
L Lorenzo Alves, T Branco, E Andrade, S Amorim, M Paiva, R RodriguesAbstract
Introduction
Cardiovascular events are increasingly recognised during treatment for multiple myeloma, yet prediction tools that are practical at treatment initiation are limited. Whether commonly used cardiovascular (HFA-ICOS) and myeloma-specific (R-MCI, ISS, R-ISS) scores discriminate risk of treatment-related cardiovascular events remains uncertain in real-world cohorts.
Purpose
To derive a simple baseline clinical risk score for cardiovascular events during multiple myeloma treatment and compare its discriminative performance with HFA-ICOS risk class, R-MCI class, ISS and R-ISS.
Methods
We performed a retrospective analysis of a real-world cohort of patients with multiple myeloma. The primary endpoint was a composite cardiovascular event during treatment (heart failure, atrial fibrillation, other arrhythmias, stroke, cardiovascular hospitalisation or cardiovascular death). Patients were considered evaluable if at least one component was recorded. Baseline predictors were screened using Fisher’s exact test. A parsimonious point-based score was constructed from the strongest baseline predictors. Discrimination was assessed by the area under the receiver operating characteristic curve (AUC) and compared with HFA-ICOS, R-MCI, ISS and R-ISS.
Results
Of 25 patients, 24 were evaluable for the composite endpoint; 9/24 (37.5%) experienced a cardiovascular event. Baseline dyslipidaemia was associated with events (OR 9.63, p=0.033), while prior atrial fibrillation and prior heart failure showed directionally higher risk but did not reach statistical significance due to small numbers (atrial fibrillation: separation with p=0.13; heart failure: OR 3.25, p=0.33). A simple score was derived assigning 2 points for dyslipidaemia, 2 points for prior atrial fibrillation and 1 point for prior heart failure (range 0–5). Event rates increased from 15.4% (score 0) to 60.0% (score 2), 50.0% (score 3) and 100% (scores 4–5). The derived score showed good discrimination (AUC 0.78, n=24). In comparison, discriminative performance was limited for established scores: HFA-ICOS AUC 0.52 (n=23), R-MCI AUC 0.38 (n=23), ISS AUC 0.46 (n=24) and R-ISS AUC 0.56 (n=24).
Conclusion
A simple baseline clinical score incorporating dyslipidaemia, prior atrial fibrillation and prior heart failure stratified risk of cardiovascular events during multiple myeloma treatment and outperformed HFA-ICOS, R-MCI, ISS and R-ISS in this cohort. This pragmatic approach may support targeted cardio-oncology surveillance at treatment initiation. External validation in larger cohorts is warranted.Cardiovascular event rate ROC curve comparison