Machine learning identifies microbiome and clinical predictors of sustained weight loss following prolonged fasting

Prolonged fasting may benefit metabolic health, but data in healthy individuals remain limited. We conducted a randomized, waitlist-controlled study, in which 38 healthy participants completed a 5-day fasting intervention with a 12-week follow-up (LEANER study, ClinicalTrials.gov: NCT04452916). Fasting acutely reduced body mass index (BMI), primarily due to fat mass loss. These changes partially persisted at follow-up. Fasting altered the gut microbiome composition and induced metabolite shifts in plasma and feces. Long-term and post-fasting changes to gut microbiome alpha diversity after fasting correlated with baseline microbiome diversity. Long-term BMI response at follow-up could be predicted using baseline microbiome and clinical data, highlighting an unclassified Faecalibacterium sp., Oscillibacter sp. 50_27, LDL cholesterol, and systolic blood pressure as predictors. The model was successfully applied to three independent cohorts: first, patients with metabolic syndrome undergoing a 5-day fasting intervention followed by a dietary intervention; second, patients with multiple sclerosis undergoing two periods of prolonged fasting with intermittent fasting in between and afterwards; and third, healthy volunteers undergoing between 6 and 12 days of prolonged fasting. Our results show that prolonged fasting is a safe and effective metabolic intervention in healthy adults and demonstrate that baseline characteristics can predict individual metabolic responses to fasting across both healthy and diverse patient groups.