- When patients with Afib are analyzed based on features that group them together, they cluster into 4 clinical subtypes that fall outside conventional Afib categories.
- These 4 types are clinically relevant, the study authors say, and differ in comorbidities and cardiovascular outcomes.
- These novel clusters suggest that conventional classifications involve considerable patient heterogeneity.
Why this matters
- Using individual-level data to define patient groupings based on comorbidities and other features might yield more targetable endpoints than using conventional parameters does.
- The analysis used here, called “cluster analysis,” has been used in other diseases, and the authors say that it enhances identification of subtypes.
- Cluster analysis of 9749 patients with Afib.
- Patients classified into low comorbidity (LC; n=4673), younger/behavioral disorder (YBD; n=963), device implantation (DI; n=1651), and atherosclerotic-comorbid (A-Co; n=2462) clusters.
- Funding: Agency of Healthcare Research and Quality.
- Risk for major adverse cardiovascular/neurological events (MACNE) was lowest in LC cluster (2.58 events/100 patient-years; P<.001).
- Compared with LC cluster, risk for MACNE was significantly higher in YBD (aHR, 1.49; 95% CI, 1.10-2.00); DI (aHR, 1.39; 95% CI, 1.15-1.68), and A-Co (aHR, 1.59; 95% CI, 1.31-1.92) clusters.
- Similar clusters were identified in external validation cohort.
- Limited generalizability.
Coauthored with Antara Ghosh, PhD