Duration-dependent effects of water-only fasting on blood lipids: a systematic review, meta-analysis, and threshold meta-regression

Background/objectives: Water-only fasting is practiced for metabolic and therapeutic purposes, yet its specific effects on lipid fractions remain inconsistently reported. This systematic review and meta-analysis evaluated lipid responses to water-only fasting across varying durations and fasting protocols. Methods: PubMed, Scopus, and Web of Science (2000-2025) were searched for human studies reporting pre-post lipid measurements under water-only fasting. Thirty-two studies met eligibility criteria. Effect sizes were calculated as Hedges' g using random-effects models. Duration-dependent responses were evaluated through subgroup analyses (≤3 days, >3 days) and piecewise threshold meta-regression. Publication bias was assessed via funnel plots and Egger's tests. Results: Water-only fasting produced lipid-specific and duration-dependent adaptations. HDL decreased significantly overall (g = -0.233), with no change in ≤3-day fasts but clear reductions in >3-day fasts; threshold analysis identified an early decline within the first ~3 days. LDL increased significantly (g = 0.489) and across all duration subgroups, showing a biphasic trajectory with progressive elevation up to ~10 days followed by attenuation or partial reversal. Total cholesterol also increased (g = 0.343), with the largest effects in >3--day fasts and a nonlinear threshold at ~5 days marking stabilization or modest decline thereafter. Triglycerides showed no significant overall effect (g = -0.039), characterized by reductions in ≤3-day fasts, increases in >3-day fasts; a marked early-phase threshold was observed at ~2.5 days. VLDL exhibited small, non-significant changes (g = 0.203) with substantial heterogeneity and limited data. Evidence of publication bias was detected for LDL and total cholesterol but not for HDL, triglycerides, or VLDL. Conclusion: Water-only fasting induces distinct, duration-dependent lipid adaptations. LDL and total cholesterol demonstrate early increases followed by stabilization, HDL decreases mainly during multi-day fasts, while triglycerides and VLDL show no uniform pattern. These findings highlight the importance of considering fasting duration when evaluating cardiometabolic effects and underscore the need for rigorously controlled, longer-term clinical trials.