3D body shape–based phenotyping for precision risk stratification of metabolic syndrome

3D body scanning serves as a noninvasive surrogate for dual-energy X-ray absorptiometry (DXA) in body composition evaluation, but the links between the derived body shape phenotypes and metabolic syndrome (MetS) remain unclear. This cross-sectional study (2024 China National Health and Nutrition Survey) applied unsupervised clustering to 144 3D body measurements and classified 5,174 adults into four body shape categories. Multivariate logistic regression was subsequently used to examine the associations between MetS and core metabolic conditions. A total of 5,174 participants were classified into four body shape clusters. Cluster 2 had the lowest MetS prevalence (9%), whereas Cluster 1 had the highest (55%), followed by Cluster 3 (44%) and Cluster 4 (28%). Compared with the reference group (Cluster 2: tall, slender torso, well-proportioned, symmetrical limbs), Cluster 3 (tall, broader torso, relatively anterior pelvic tilt, thoracic protrusion, longer upper limbs) showed the strongest association with MetS, followed by Cluster 4 (short, shorter torso, relatively posterior pelvic tilt, larger lower limbs) and Cluster 1 (short, stockiest torso, relatively posterior pelvic tilt, forward head posture, shorter, thicker limbs). In normal-weight individuals, Cluster 4 was significantly associated with MetS, whereas in overweight individuals, the associations were strengthened across all body shape clusters. All body shapes were linked to hypertension, with Clusters 1 and 4 also associated with hyperuricemia and Cluster 3 associated primarily with hyperlipidemia. The associations were stronger for women and individuals < 60 years. Body shape clusters showed differential associations with MetS and core conditions (hypertension, hyperuricemia, and hyperlipidemia), varying across body mass index (BMI) categories, sex, and age groups.