The Gut–brain–adipose Axis in Ultra-processed Food and Obesity: A Mechanistic Synthesis and Its Implications for Food Classification

PURPOSE OF REVIEW: Consumption of ultra-processed food (UPF) tracks closely with obesity across populations, and the Nova classification has become the dominant tool for capturing that exposure. Why UPF promotes weight gain is a separate question, and the biological answer has accumulated in fragments. I draw those strands together and ask whether three mechanisms usually studied in isolation: gut microbial disruption, hypothalamic inflammation, and adipose tissue dysfunction, are better understood as one connected system, and what that would mean for how UPF is classified. RECENT FINDINGS: Three experimental literatures have converged on a shared pathway. Dietary emulsifiers and non-sugar sweeteners alter microbial composition and weaken the intestinal barrier, raising circulating lipopolysaccharide. In animal models this signal reaches the hypothalamus, where it activates inflammatory pathways, recruits glia, and blunts the leptin response that normally limits intake; though whether the same sequence operates in humans remains unestablished. Visceral fat that expands under this regime secretes its own inflammatory load, which feeds back onto both the gut and the brain. A 2025 UK Biobank analysis added a human imaging dimension, reporting structural differences in feeding-related brain regions that scaled with UPF intake and were only partly explained by adiposity. The evidence coheres best when the three arms are read as a single self-reinforcing loop in which each influences the others. Within that frame, the limitation of the Nova classification becomes specific and tractable: Group 4 mixes products that engage the loop strongly with products that barely touch it. This review sets out where the mechanistic evidence is firm, where it remains thin, and how an attribute-aware refinement of Group 4 might be tested.