With the exception of Serenoa repens (oral and topically) and Salvia rosmarinus (topically), none of these herbs has been shown to have this activity in humans. Extrapolation from in vitro or animal studies to humans is highly unlikely to lead to clinical efficacy.
* Serenoa repens (saw palmetto) fruit extract standardized to 70–95% fatty acids and sterols, 68% inhibition in one human clinical trial (PMID 9759701), 32% in another (PMID 11337315)
Note: As far as could be determined, this is the only human study in which the effect of an herb on 5α-reductase was determined. And no herb has been shown to decrease prostate size more than 1–2% (and only saw palmetto has really ever been shown to have even this tiny degree of shrinking effect), unlike the 30–50% reduction we see with drug 5α-reductase inhibitors. Thus it is likely 5α-reductase inhibition by herbs is working together with other properties of herbs to result in a net clinical effect that is manifestly different than drugs.
* Prunus africanum (pygeum, African plum) bark extract (which has a chemistry very similar to saw palmetto), in vitro ED50 0.78 mg/ml (PMID 23194959); inactive in another in vitro assay (PMID 8381228)
* Salvia rosmarinus = Rosmarinus officinalis (rosemary), in mice inhibited 5α-reductase 82% (200 µg/mL) and 95% (500 µg/mL) (PMID 22517595). This fits with a human clinical trial showing rosemary volatile oil topically improves hair growth (PMID 25842469).
* Phyllanthus niruri (stonebreaker), crude extract in vitro inhibited 5α-reductase by 24.3% (at 50 µg/mL) and 64.6% (at 200 µg/mL); its constituent stigmasterol glucoside IC40 = 27.2 µM; extract grew hair on mice at 5 mg po dose/day (PMID 30068884 )
* Ageratum conyzoides (billygoat weed, Brazilian chickweed), ex vivo suppressed mRNA expression of 5α-reductase type 1 and 2 genes (PMID 29048765)
* Lespedeza cuneata (Chinese bushclover), in rat model of BPH inhibited serum DHT by 51–54% (25–100 μg/ml) and expression of 5α-reductase by 54% (25 μg/ml), 61% (50 μg/ml), and 74% (100 μg/ml) (PMID 30724641 )
* Acorus gramineus (Japanese sweetflag), in mice 62–100% 5α-reductase inhibition (PMID 28830494)
* Curcuma longa (turmeric) crude extract IC50 = 9 μg/ml (27789379)
* Impatients balsamina (rose balsam) crude extract IC50 = 5.4 μg/ml (27789379)
* Carthamnus tinctorius (safflower) crude extract in vitro had a finasteride equivalent inhibition of 5α-reductase of 24 mg finasteride/1 g extract; it was the most potent of several herbal extracts at growing hair applied topically to mice (PMID 22178180 )
* Ganoderma lucidum (reishi) in vitro (PMID 23710567)
* Benincasa hispida (wax gourd) in vitro (PMID 23710567)
* Sphaeranthus indicus (East Indian globe thistle) in vitro (PMID 23710567)
* Cuscuta reflexa (giant dodder) in vitro (PMID 23710567)
* Torilis japonica (Japanese hedge parsley) a crude methanol extract was active in vitro, its compound torilin IC50 = 31.7 μM (PMID 12802730 ). Note that finasteride's IC50 was 0.38 μM in this assay.
* Brassica rapa (bok choy, napa cabbage) pollen, one compound (pollinastanol) had IC50 = 30 μM (PMID 19336936 )
* Cynomorium songaricum (suo yang) in mouse model of BPH, inhibited upregulation of 5α-reductase type 1 and 2 genes (PMID 30708032)
* Swertia bimaculata (double-spotted swertia), various xanthones inhibited 5α-reductase in vitro by 39–49% (PMID 28030961 )
* Avicennia marina (grey or white mangrove), in vitro crude extract (10 μg/ml) inhibited 5α-reductase 52%; avidequinone C inhibited 5α-reductase 1 with IC50 = 38.8 μM (PMID 24858268)
Notable Negatives
* Urtica dioica (stinging nettle) root in vitro (PMID 23194959; PMID 8381228)