ObjectivesTo identify the most suitable size of imaging-visible embolic agents with balanced safety and efficacy for bariatric arterial embolization (BAE) in a preclinical model. Materials and MethodsTwenty-seven pigs were divided into 3 cohorts. In Cohort I, 16 pigs were randomized to receive (n = 4 each) 40–100-μm microspheres in 1 or 2 fundal arteries, 70–340-μm radiopaque microspheres in 2 fundal arteries, or saline. In Cohort II, 3 pigs underwent renal arterial embolization with either custom-made 100–200-μm, 200–250-μm, 200–300-μm, or 300–400-μm radiopaque microspheres or Bead Block 300–500 μm with microsphere distribution assessed histologically. In Cohort III, 8 pigs underwent BAE in 2 fundal arteries with tailored 100–200-μm radiopaque microspheres (n = 5) or saline (n = 3). ResultsIn Cohort I, no significant differences in weight or ghrelin expression were observed between BAE and control animals. Moderate-to-severe gastric ulcerations were noted in all BAE animals. In Cohort II, renal embolization with 100–200-μm microspheres occluded vessels with a mean diameter of 139 μm ± 31, which is within the lower range of actual diameters of Bead Block 300–500 μm. In Cohort III, BAE with 100–200-μm microspheres resulted in significantly lower weight gain (42.3% ± 5.7% vs 51.6% ± 2.9% at 8 weeks; P = .04), fundal ghrelin cell density (16.1 ± 6.7 vs 23.6 ± 12.6; P = .045), and plasma ghrelin levels (1,709 pg/mL ± 172 vs 4,343 pg/mL ± 1,555; P < .01) compared with controls and superficial gastric ulcers (5/5). ConclusionsIn this preclinical model, tailored 100–200-μm microspheres were shown to be most suitable for BAE in terms of safety and efficacy.