[Pathophysiology-based phenotyping and clinical characterization of obesity]

Objective: To evaluate the applicability of a pathophysiology-based obesity phenotyping approach in Chinese patients with obesity and characterize the clinical features of different obesity phenotypes. Methods: In this cross-sectional study, patients with obesity (body mass index ≥28 kg/m²) were recruited from the obesity and weight management clinic of Peking University People's Hospital between December 2023 and November 2024. Data were collected regarding demographics, body composition, hunger perception [Visual Analogue Scale (VAS)], depressive status (Hospital Anxiety and Depression Scale), basal metabolic rate, and metabolic parameters. Using pre-established cut-off values derived from previous studies, participants were categorized into brain hunger, gastrointestinal hunger, emotional hunger, hypometabolic, or untyped phenotypes. The Kruskal-Wallis H test and other methods were used for comparisons among multiple groups, and the Bonferroni correction applied for multiple comparisons. Results: Among the 900 patients [mean age (34.9±8.3) years; 61.3% female, 552/900], the brain hunger phenotype (22.1%, 199/900) exhibited the highest 30-min VAS score [M (Q₁, Q₃), 33 (21, 43)], significantly exceeding that of the gastrointestinal hunger phenotype [19 (0, 30), all P<0.05]. The gastrointestinal hunger phenotype (43.9%, 395/900) demonstrated a significantly higher 2-hour VAS score [41 (30, 60)] than either the emotional hunger [20 (0, 41)] or hypometabolic [20 (0, 40)] phenotype (both P<0.05). The emotional hunger phenotype (34.1%, 307/900) had the highest depression score [8 (7, 9)], significantly exceeding those of the brain hunger [4 (2, 7)] and gastrointestinal hunger [4 (2, 7)] groups; its fasting C-peptide level [1.21 (0.95, 1.52) nmol/L] was significantly higher than those of the brain hunger [1.11 (0.84, 1.38) nmol/L] and gastrointestinal hunger [1.11 (0.85, 1.48) nmol/L] groups (all P<0.05). The hypometabolic phenotype (54.2%, 488/900) exhibited the lowest ratio of measured basal metabolic rate to predicted basal energy expenditure [82% (80%, 84%)], which was significantly lower than that of the brain hunger phenotype [86% (82%, 89%); all P<0.05]. This phenotype had the highest body mass index [34.22 (31.50, 37.38) kg/m²], which was significantly higher than that of the brain hunger phenotype [32.21 (30.20, 35.37) kg/m²; all P<0.05]. Furthermore, uric acid levels in the hypometabolic group [431 (370, 496) μmol/L] were significantly higher than those in the gastrointestinal hunger [405 (343, 473) μmol/L] and emotional hunger [413 (353, 476) μmol/L] phenotypes (both P<0.05); this group also had the highest body fat percentage [44.5% (39.3%, 48.5%)] and visceral fat level [20 (18, 20)] among all phenotypes (all P<0.05). Additionally, 16.3% (147/900) of patients remained untyped, while 49.0% (441/900) presented with multiple phenotypes. Conclusions: The pathophysiology-based phenotyping approach can identify subtypes with different clinical characteristics among Chinese patients with obesity. Each phenotype exhibits distinct characteristics in hunger regulation, emotional status, energy metabolism, and body composition.