Principal Component Analysis-Derived Clusters of Postprandial Biomarker Responses Differed by Cardiometabolic Disease Risk but Not by Early Life Nutritional Exposure (P18-123-19)

Abstract

ObjectivesExposure to improved nutrition in early life has diverging impacts on cardiometabolic disease (CMD) risk factors. We used principal component analysis (PCA) to investigate the clustering of prandial challenge-induced biomarker responses, which may further characterize the differential biochemical pathways affected by early nutrition and by CMD risk factors. MethodsParticipants were members of a Guatemalan cohort, who had been randomized at the village level to receive either a protein-containing supplement (atole) or a low-calorie control (fresco) as children. In 2015–17, we collected fasting and 2-hr postprandial plasma samples (n = 1030, 59.5% women, age 44.1 ± 4.2y). We assayed lipids, glycemic markers and inflammation cytokines (including the adipokines leptin and adiponectin), and calculated prandial challenge-induced changes in their concentrations. We used PCA to characterize the variance-covariance matrix in postprandial biomarker responses. We compared PCA clusters across strata of early nutrition (exposed to atole from conception to 2y of age versus other), metabolic syndrome (MetS), obesity status, and sex. ResultsPostprandial responses in lipids, glycemic markers, and inflammation cytokines clustered separately. The primary cluster was dominated by variation in lipid responses (Figure 1); this was true across strata of early nutrition exposure, MetS, obesity status, and sex. The loadings of glycemic markers and cytokines on the subsequent clusters differed across strata of MetS (Figure 2), obesity status, and sex (Hotelling’s T-squared test P < 0.05), but did not differ between those exposed to atole in early life versus others. Compared with participants without MetS, those with MetS had greater postprandial changes in the concentrations of interleukin-6 (IL-6), IL-10, and monocyte chemoattractant protein-1 (MCP-1). Among those without MetS, we observed that adipokines contributed negatively to the primary cluster, opposing the loading of lipids. ConclusionsPCA-derived clusters of postprandial biomarker responses differed by CMD risk factor but not by early nutrition in this cohort. Inflammation and adipokine-mediated pathways may have been altered by MetS status. Funding SourcesThis work is supported by the National Institutes of Health, USA. Supporting Tables, Images and/or Graphs▪▪