Plant- and soil-derived polyphenols shape soil microbiomes and crop outcomes: a systematic review and meta-analysis

Background Polyphenols from root exudates and polyphenol-rich organic inputs may shape soil microbiomes, enzyme activities, and crop nutrient outcomes, with relevance to regenerative agriculture. Objectives To quantify the effects of plant- and soil-derived polyphenols/polyphenol-rich soil inputs on i) soil microbial alpha-diversity and enzyme activity and ii) crop yield and nutrient quality and to summarise mechanistic pathways and evidence certainty. Methods Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement, we searched PubMed, Web of Science, Scopus, Embase, and CAB Abstracts (1990–March 2025) plus grey literature. Fifteen controlled studies (163 observations) met the Population, Intervention, Comparator, Outcomes, and Study design (PICOS) eligibility. Random-effects meta-analysis pooled standardised mean differences (SMDs; Hedges’ g ) for continuous outcomes. To ensure reproducibility and statistical independence, we prespecified decision rules for selecting endpoints/time points and aggregated multiple within-study outcomes into study-level composites within each domain. We explored moderators [polyphenol source class, purified vs . complex inputs, environment (field vs . controlled), soil pH, soil organic carbon (SOC), duration, and dose, where reported] via subgroup analyses and mixed-effects meta-regression. We assessed certainty using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach. Results Polyphenol inputs increased soil microbial alpha-diversity (pooled across Shannon/Simpson indices) [SMD = +0.58, 95% confidence interval (CI) 0.32–0.83; high certainty] and soil enzyme activity (SMD = +0.49, 95% CI 0.24–0.73; moderate certainty). Crop yield increased modestly (+6.8%, 95% CI 2.7–10.9; moderate certainty), while micronutrients iron (Fe) and zinc (Zn) improved modestly (low–moderate certainty) due to limited reporting and imprecision. Yield and nutrient-quality outcomes were secondary endpoints in most included studies and were available in a smaller, uneven subset, limiting generalisability. Heterogeneity was moderate (I 2 ≈ 50%) and was partly consistent with differences between purified compounds vs . complex organic amendments and field vs . controlled environments. Conclusion and future directions Plant- and soil-derived polyphenol inputs meaningfully enhance soil microbial diversity and enzyme activity and may deliver smaller, context-dependent gains in yield and crop micronutrient quality; future multi-season field trials with standardised polyphenol chemistry/dose reporting and functional multi-omics are needed to clarify persistence, dose–response relationships, and causal mechanisms.