Health status trajectories using Markov modeling in patients with heart failure with preserved ejection fraction.

In 30 seconds

This post hoc multistate Markov analysis of the PARAGON-HF randomized trial examined serial Kansas City Cardiomyopathy Questionnaire scores and clinical outcomes in 4707 patients with heart failure with preserved ejection fraction. Patients moved in both directions among four health-status categories, with death as a fifth state. Annual mortality rose as health status worsened, from 3.1% in the excellent state to 15.6% in the poor state. Overall time spent in health states did not differ between sacubitril/valsartan and valsartan, although the poor-health subgroup showed more favorable transitions with sacubitril/valsartan.

Key findings

• Annual mortality increased across worsening health states: 3.1% (excellent), 4.3% (good), 9.4% (fair), and 15.6% (poor).
• The excellent state had the longest sojourn time at 1.57 years (95%CI, 1.51-1.63); transitions between health states occurred in both directions.
• Among patients in the poor state, sacubitril/valsartan was associated with greater likelihood of improvement to fair than valsartan (93.1% vs 86.9%) and lower risk of death (6.9% vs 13.1%).
• Overall sojourn times did not differ between sacubitril/valsartan and valsartan.

Why it matters

The findings show that health status in HFpEF can change over time and that serial patient-reported assessments may help characterize prognosis beyond a single baseline score. The graded association between poorer KCCQ state and higher mortality underscores the clinical relevance of tracking symptoms and functional impact, while the treatment subgroup result warrants interpretation in the context of a post hoc analysis.

What to keep in mind

This was a post hoc analysis using a multistate model; the reported treatment-associated findings in the poor-health subgroup do not establish a causal treatment effect.