Nutraceuticals and phytoceuticals for the treatment of major depressive disorder: a systematic review and network meta-analysis

Background “Nutraceuticals” and “phytoceuticals” may have antidepressant activity, prompting an in-depth network meta-analysis of randomized clinical trials (RCTs). Methods Systematic review and network meta-analysis (PubMed/MEDLINE/EMBASE/Scopus/PsycINFO/CENTRAL/ClinicalTrials.gov (until 01/07/2025) of RCTs vs. placebo or treatment as usual (TAU = antidepressants) testing nutraceuticals/phytoceuticals in major depressive disorder (MDD). Change in depressive symptomatology (standardized-mean-difference = SMD), treatment response, and all-cause discontinuation (acceptability) (risk ratio = RR) were co-primary outcomes. Secondary outcomes were anxiety symptom changes, adverse event-related discontinuation (“tolerability”), and symptom remission. We conducted subgroup, transitivity, and sensitivity analyses to explore/reduce heterogeneity, and assessed global/local inconsistency, publication bias, risk of bias (RoB), and confidence in the evidence (CINeMA/AMSTAR-Plus). Results Across 163 studies (n = 15,757, distinct compounds’ combinations n = 137), several molecules/classes outperformed placebo with very large effect sizes. Restricting analyses to compounds with ≥2RCTs/non-TAU antidepressants/low RoB/non-sponsored trials/non-outliers, higher-than-expected/unrealistic effect sizes emerged vs. placebo for eicosapentaenoic acid (EPA) (k = 20, n = 8483; SMD = −1.67;95%C.I. = −2.38;−0.97), vitamin D3 (k = 3, n = 344; SMD = −1.59;95%C.I. = −2.71;−0.46), and St. John’s-wort-extract-ZE117 (k = 3,n = 207;SMD = −1.02;95%C.I = −1.90;−0.15) regarding depressive symptomatology. Substantial heterogeneity (I 2 = 85.3%;95%C.I. = 81.9%;88.0%) and global inconsistency (Q-between-designs = 186.91, p < 0.0001) emerged, explained by subgroup and sensitivity analyses, without publication bias. Additionally, a medium effect size emerged vs. placebo for curcumin (k = 2,n = 89;SMD = −0.58;95%C.I. = −0.99;−0.18) regarding anxiety reduction in low RoB analyses. Shugan granules, EPA, EPA+Docosahexaenoic Acid, ZE117, and fluoxetine+EPA response and St. John’s wort extract WS5570/ZE117 remission significantly outperformed placebo. Citicoline was associated with significantly more dropouts than placebo. A meta-regression of efficacy effect sizes against adapted AMSTAR-Plus scores showed that larger SMDs were associated with lower AMSTAR scores, indicating lower study quality. Confidence in the evidence was low to very low. Conclusions Upon filtering low-quality evidence, only a few nutraceutical/phytoceuticals showed potential for depressive symptoms, warranting more rigorous trials.