Abstract
Background: Biomarker science in heart failure (HF) is advancing quickly in our ability to diagnosis and treat patients with this complex syndrome. Researchers are urged to not use single-marker strategies, but instead evaluate biomarkers in patterns to better understand their relationship to one another, as well as disease progression. Latent class mixture modeling allows researchers to determine novel associations between biomarkers.
Objective: The objectives of this study were to identify and compare latent classes of cardiovascular biomarkers among patients with moderate to advanced HF.
Methods: This was a cross-sectional study of 96 participants with moderate to advanced HF. Latent class mixture modeling was used to identify unique classes of biomarkers and their associations to sociodemographic and clinical variables.
Results: The average age of the sample was 54 years, with most of the sample being men (77%) and having an average ejection fraction of 23%. Two unique classes of biomarkers were identified. Latent class 1 had higher levels of all biomarkers, whereas latent class 2 had lower levels. The higher biomarker class had, on average, more neurohormonal activation and fluid retention; however, the higher levels of biomarker class were not more likely to be diagnosed with advanced HF or have more comorbidities.
Conclusion: By identifying classes of biomarkers, providers may be better able to identify patients who are at risk of progressing into advanced HF quicker or those who are more likely to have more severe complications, such as fluid overload or renal disease.