Metabolic profiles to predict long-term cancer and mortality: the use of latent class analysis

Abstract

Background Metabolites are genetically and environmentally determined. Consequently, they can be used to characterize environmental exposures and reveal biochemical mechanisms that link exposure to disease. To explore disease susceptibility and improve population risk stratification, we aimed to identify metabolic profiles linked to carcinogenesis and mortality and their intrinsic associations by characterizing subgroups of individuals based on serum biomarker measurements. We included 13,615 participants from the Swedish Apolipoprotein MOrtality RISk Study who had measurements for 19 biomarkers representative of central metabolic pathways. Latent Class Analysis (LCA) was applied to characterise individuals based on their biomarker values (according to medical cut-offs), which were then examined as predictors of cancer and death using multivariable Cox proportional hazards models. Results LCA identified four metabolic profiles within the population: (1) normal values for all markers (63% of population); (2) abnormal values for lipids (22%); (3) abnormal values for liver functioning (9%); (4) abnormal values for iron and inflammation metabolism (6%). All metabolic profiles (classes 2–4) increased risk of cancer and mortality, compared to class 1 (e.g. HR for overall death was 1.26 (95% CI: 1.16–1.37), 1.67 (95% CI: 1.47–1.90), and 1.21 (95% CI: 1.05–1.41) for class 2, 3, and 4, respectively). Conclusion We present an innovative approach to risk stratify a well-defined population based on LCA metabolic-defined subgroups for cancer and mortality. Our results indicate that standard of care baseline serum markers, when assembled into meaningful metabolic profiles, could help assess long term risk of disease and provide insight in disease susceptibility and etiology. Electronic supplementary material The online version of this article (10.1186/s12860-019-0210-7) contains supplementary material, which is available to authorized users.

Metabolic subgroups and cardiometabolic multimorbidity in the UK Biobank

“…In a typical scenario, people with similar profiles are grouped together and the aggregate rates of disease outcomes are compared between the subgroups. Examples of biomarker-based analyses include five clusters of diabetes with divergent outcomes 14 , four metabolic profiles of cancer mortality 15 and six endotypes of heart failure 16 . We developed subgroups of diabetic complication burden in 2008 17 and validated them in 2018 with new previously unseen data on clinical outcomes 18 .…”