Improving preclinical study design to increase predictive value in heart failure drug development
M UfnalAbstract
Background
Promising preclinical findings in heart failure (HF) frequently fail to translate into clinically meaningful benefit. A major contributor is a systematic disconnect between what is modelled and measured in animals and the clinical outcomes that determine success in pivotal trials and adoption in practice.
Purpose
To identify dominant, recurring causes of poor translation in HF drug development and provide pragmatic, implementable recommendations to improve the predictive value of preclinical packages.
Methods
Structured narrative review of commonly used HF animal models and preclinical endpoints. Endpoints, models and readouts were evaluated for alignment with outcomes that drive registration and adoption (mortality, HF hospitalisation, functional capacity and symptom status).
Results
Four recurring drivers of translational failure were identified: (1) preclinical–clinical endpoint mismatch, with overemphasis on short-term haemodynamic or biomarker shifts rather than durable measures linked to mortality/hospitalisation risk, congestion and volume status, filling pressures, physical capacity and arrhythmic risk; (2) reliance on young, simplified single-insult models that do not reproduce the aged cardiovascular substrate typical of HF; (3) underrepresentation of key comorbidities (hypertension, diabetes, obesity, chronic kidney disease) that shape HF phenotype and modify treatment response; and (4) evaluation of investigational therapies without background guideline-directed medical therapy, despite pharmacodynamic interactions that can materially influence disease course and the apparent effect of the investigational medicinal product.
Conclusions
HF therapies often fail clinically because preclinical programmes prioritise endpoints that weakly map to clinical success and are conducted in biologically "clean" conditions. Predictive value can be improved by incorporating aged and comorbidity-enriched models, adding background HF therapy in later-stage preclinical testing, and prioritising clinically relevant endpoints (mortality, congestion/volume status, filling pressures and functional capacity), supported by robust bias control and standardised reporting.