Impact of protein-enriched soy milk on phenolic bioaccessibility and antioxidant capacity of coffee pulp infusions

Coffee pulp is a phenolic-rich by-product with antioxidant potential; however, its use in functional foods may be restricted by stability and bioaccessibility issues. This study aimed to evaluate how the incorporation of protein-enriched soy milk influences a coffee pulp infusion's digestibility and functional performance.

Eleven formulations were prepared by modifying both the proportion of soy milk added to the infusion (0–30%) and the protein content in the soy milk (3.9–11.7%), based on a central composite circumscribed design (CCC) within a response surface methodology (RSM) framework, to model and predict the effects of these two variables on the stability and bioaccessibility of phenolic compounds. The samples underwent standardized in vitro gastrointestinal digestion following the INFOGEST protocol. Total phenolic content (TPC), total flavonoid content (TFC), and antioxidant capacity (ABTS and FRAP) were determined before and after digestion.

The statistical models exhibited high validity and predictive performance (R² and Q² > 0.9 in most cases), capturing the complex influence of both the proportion of soy milk and its protein enrichment on the behavior of phenolic compounds during digestion. In particular, quadratic effects and protein–soy interactions accounted for a significant proportion of the explained variance, up to 65.5% in some cases, highlighting the nonlinear and synergistic nature of the formulation’s impact on antioxidant functionality. The results showed that a formulation containing 18% soy milk and 9.4% protein yielded the optimal functional profile, favoring the bioaccessibility of phenolics and their antioxidant capacity, likely due to interactions with soy proteins that enhance compound stability and release during digestion. Compared to the infusion without soy, the optimized formulation led to post-digestion bioaccessibility improvements of 33% for TPC, 46% for TFC, 51% for ABTS, and 42% for FRAP.

This study highlights the value of plant-based matrices and multivariate modeling for enhancing and optimizing polyphenol-rich functional beverages.