A lifetime prediction model to estimate survival and treatment benefit in HFrEF

Lifetime prediction of survival and treatment benefit in patients with HFrEF – the LIFE-HF model

Presented at the ESC congress 2022 by: Stefan Koudstaal, MD, PhD – Gouda, The Netherlands

Introduction and methods

The 2021 HF guidelines recommend four classes of drug treatment as first-line therapy for patients with HFrEF (LVEF<40%). Implementation of these drugs in clinical practice is limited though and underuse of all classes, particularly ARNI and MRAs, is seen in real-world studies. For patients, clinical trial results (HR, absolute risk reduction, NNT, survival) may be difficult to comprehend, and life expectancy and gain in life expectancy by drugs may help to improve intrinsic motivation of patients.

The aim of this study was to develop a lifetime prediction model to inform patients about their individual lifetime survival free of HF hospitalization and/or CV mortality. And to inform patients how medical therapy for HFrEF will influence their individual lifetime risk for HF outcomes.

To develop the model, age was used as a time scale with left truncation and right censoring. The PARADIGM and ATMOSPHERE trials were used as datasets (n=15,415). External validation of the model was done using the Swede-HF, ASIAN-HF registries and the DAPA-HF trial (n=51,286). Outcome measures were CV mortality or first HF hospitalization; CV mortality; non-CV mortality. (Pooled) HRs for outcomes with drug treatments were used of different large trials or meta-analyses.

Main results

• The top 10 predictors for the lifetime model were: sex, NYHA class III/IV, prior hospitalization for HF, diabetes, extracardiac vascular disease, SBP, LVEF, NT-proBNP, total bilirubin, and uric acid.
• C-statistic was 0.68 (95%CI: 0.67-0.69) for the outcome of CV death or first hospitalization for HF in the LIFE-HF model. Performance was similar across geographical regions and gender.
• External validation resulted in similar C-statistics (0.66, 95%CI:0.66-0.67 in SWEDE-HF; 0.66, 95%CI: 0.65-0.68 in ASIAN-HF; and 0.70, 95%CI: 0.68-0.72 in DAPA-HF).

Conclusion

The LIFE-HF model provides individualized estimates of years free of HF hospitalizations using life-tables. And therefore provides the translation from HRs to individual treatment benefit in life-years gained. The visual aspect of the LIFE-HF calculator may help to motivate patients for better adherence to guideline-recommended therapies and facilitate shared decision-making.

Stefan Koudstaal ended his presentation by mentioning that when the study is published, the interactive LIFE-HF calculator will become available for free.

• Our reporting is based on the information provided at the ESC Congress -