Chrysin (5,7-dihydroxyflavone) is a plant flavone found in small amounts in honey and in higher amounts in propolis, the resin bees collect from plants. It is sold as a supplement and is often called a “natural aromatase inhibitor,” based on its ability to block the estrogen-making enzyme aromatase in lab assays.

The most important fact is that this effect has not been shown in people. In a 7-person absorption study, a 400 mg oral dose produced peak blood levels of free chrysin far below the concentration needed to inhibit aromatase. Three trials from one US lab gave men chrysin inside androgen-precursor mixtures, and estradiol still rose. We found no trial of chrysin alone that measured blood testosterone or estradiol.

This page summarizes what human absorption studies, small clinical trials and laboratory research report. It is for education only and is not medical advice.

Property	Detail
Class	Flavone (a subgroup of flavonoids)
Formula and weight	C₁₅H₁₀O₄, 254.24 g/mol (PubChem)
Main sources	Propolis, honey, passionflower and other herbal materials
Doses in cited human studies	400 mg once; 1,000 mg once and 1,000 mg/day for 30 days; 300–625 mg/day inside multi-ingredient androgen mixtures
Oral bioavailability	Estimated 0.003–0.02% in one study of 7 adults, using an assumed volume of distribution
Half-life	Apparent 4.6 h for unchanged chrysin after 400 mg; 9–11 h for standard capsules in a later study that used a different assay
Aromatase in cell assays	IC₅₀ (concentration giving 50% inhibition) about 1–7 µM
Effect on human hormones	No demonstrated reduction in estrogen or rise in testosterone
US status	Not an approved drug; FDA does not approve dietary supplements before sale
WADA status	Not named on the 2026 Prohibited List

What it is

Chrysin’s chemical name is 5,7-dihydroxy-2-phenylchromen-4-one (PubChem). It is closely related to apigenin, which has one more hydroxyl group. A 2021 review traces its first isolation and structure to 1893 and lists honey, propolis, some vegetables, fruits and mushrooms, and passion flowers as sources. In plants it occurs both as the free compound and bound to sugars (Gao et al., Biomed Pharmacother 2021).

How much is in food

Published figures disagree, and the honey numbers matter for how the one propolis-and-honey trial below is read:

Source	Reported chrysin content	Basis
Croatian mint honey	0.112–0.51 mg/100 g (1.12–5.1 mg/kg)	Measured by LC-MS/MS in five samples (Pavlešić et al., Food Technol Biotechnol 2022)
Honey, general	Up to 5.3 mg/kg	Review figure (Gao et al. 2021)
Honey, as quoted in an aromatase review	1 mg/g (1,000 mg/kg)	Review figure, much higher than the measurements above (Balam et al., Heliyon 2020)
Propolis, as quoted in the same review	54 mg/g	Review figure

The measured honey values fit the lower figures. At about 5 mg/kg, a 20 g serving of honey would hold roughly 0.1 mg of chrysin, which is thousands of times less than the supplement doses studied.

Pharmacokinetics

The first human study gave 7 healthy adults a single 400 mg oral dose. The capsules also contained lysophosphatidylcholine, and the participants ate a low-flavonoid diet beforehand (Walle et al., Br J Clin Pharmacol 2001). A 2025 crossover trial in 16 healthy adults compared three products (Ibi et al., Antioxidants 2025).

Study and product	Peak level	Total exposure (AUC)	Half-life	What was measured
Walle 2001, 400 mg	3–16 ng/mL (12–64 nM), at about 1 h	5–193 ng·h/mL	Apparent 4.6 h	Unchanged chrysin only
Walle 2001, chrysin sulfate	Not given numerically	450–4,220 ng·h/mL	Not reported	Main metabolite in plasma
Ibi 2025, unformulated	42.1 ± 28.4 ng/mL	456.9 ± 356.4 ng·h/mL	11.0 ± 3.5 h	Chrysin after glucuronidase digestion
Ibi 2025, non-micellar with quercetin and rutin	32.9 ± 23.5 ng/mL	345.7 ± 280.8 ng·h/mL	9.0 ± 3.3 h	Chrysin after glucuronidase digestion
Ibi 2025, micellar with quercetin and rutin	87.3 ± 59.4 ng/mL	914.8 ± 697.5 ng·h/mL	25.3 ± 59.1 h	Chrysin after glucuronidase digestion

The two studies should not be compared directly. Ibi’s samples were digested with β-glucuronidase before analysis, so its numbers likely include chrysin released from its glucuronide, while Walle’s chrysin values are the unchanged compound. Ibi’s paper attributes its higher values to the higher dose, but the assay difference may also contribute. Half-life differences between Ibi’s products were not statistically significant, and the micellar value has a very wide spread.

Dose reporting in Ibi is also unclear. The methods state 1,000 mg per product, while the paper’s product table lists 412–558 mg of chrysin per capsule at one capsule per dose. Most authors are employees of ISURA, which ran the study, and one coauthor owns the Factors Group of Companies; the product is made by Natural Factors. The trial ran from December 2024 to March 2025 and was registered in July 2025.

Why so little reaches the blood

Conjugation. The gut wall and liver rapidly attach sulfate and glucuronic acid. In Walle’s study, plasma chrysin sulfate exceeded chrysin about 30-fold, and no oxidative metabolites were seen.
Excretion. Urine held only 1–7% of the dose, mostly as chrysin glucuronide. Stool collected from four people contained about 40, 160, 180 and 390 mg of chrysin, the highest being 98% of the dose (Walle et al. 2001).
Efflux and recycling. Transporters including BCRP and MRP2 pump chrysin conjugates back into the gut and bile. Gut enzymes can then free the chrysin again. This recycling keeps it concentrated in the lower gut rather than the rest of the body (Gao et al. 2021).
Protein binding. Walle estimated that over 99% of chrysin in plasma is bound to protein, leaving even less free to act on tissues.

Walle’s 0.003–0.02% bioavailability figure is an estimate. There is no intravenous human comparison; the authors borrowed the volume of distribution of a related flavonoid, quercetin.

Regulatory status

United States: chrysin is not an approved drug. Under the 1994 supplement law, FDA does not have the authority to approve dietary supplements before they are marketed, and firms generally need not give FDA their safety evidence (FDA).
Sport: chrysin is not named on the WADA 2026 Prohibited List. Class S4.1 (aromatase inhibitors) is open-ended (“including, but not limited to”) and names one synthetic flavone, α-naphthoflavone, as well as arimistane. We found no WADA statement that dietary flavones such as chrysin fall under S4.1. Androstenedione and DHEA, the steroids paired with chrysin in the trials below, are named under S1.1.

What the research shows

Aromatase inhibition in the lab

Aromatase converts testosterone to estradiol and androstenedione to estrone. The idea traces to a 1984 report that chrysin was a potent competitive inhibitor of the enzyme in human placental and ovarian microsomes (cell fractions) (Kellis & Vickery, Science 1984). Later cell work agrees that it inhibits the enzyme, though potency varies by model:

Study	Model	Chrysin result
Kellis & Vickery 1984	Human placental and ovarian microsomes	Competitive inhibitor; Ki (binding constant) of 2.6 µM, as cited by Walle et al. 2001
Saarinen et al. 2001	Human JEG-3 cancer cells; kidney cells carrying the human aromatase gene	IC₅₀ under 1.0 µM
van Meeuwen et al. 2007	Human mammary fibroblasts and MCF-7 breast cancer cells	IC₅₀ 1.5 µM
Sanderson et al. 2004	Human adrenal cancer cells (H295R)	IC₅₀ 7 µM; the steroidal drug control 4-hydroxyandrostenedione had an IC₅₀ of 20 nM

A 2020 systematic review found 20 relevant articles: one human study, two rat studies and the rest in vitro. All but one showed some aromatase inhibition, the exception being a study in human endometrial cells. The review concludes chrysin could be useful for hormone-dependent breast cancer, but its only human study was negative and it reports no trial in breast cancer (Balam et al. 2020).

Why the lab result does not carry over

Walle’s peak level of free chrysin was 12–64 nM, against a Ki of 2.6 µM (2,600 nM). That peak is about 40–200 times lower, before allowing for protein binding. The authors called any effect on androgens and estrogens in peripheral tissues “questionable” (Walle et al. 2001).

Better formulations do not clearly close the gap. The highest mean peak in the micellar trial, 87.3 ng/mL, equals about 0.34 µM, even with an assay that likely counts chrysin released from glucuronide. That trial did not measure sex hormones (Ibi et al. 2025).

A direct animal test also failed. In immature rats, oral chrysin at 50 mg/kg, a dose the authors say considerably exceeds daily dietary amounts, did not reduce estrogen- or androgen-driven uterine growth. The title of that paper states the conclusion: “No evidence for the in vivo activity of aromatase-inhibiting flavonoids” (Saarinen et al. 2001).

“Natural aromatase inhibitor.” This label is accurate only for lab assays. No published human study shows that chrysin lowers estradiol. Prescription drugs such as anastrozole, letrozole and exemestane are in a different evidence category, because their estrogen suppression has been measured directly in patients.

Human hormone trials

Four trials measured testosterone or estrogen in people, and none isolates chrysin.

Study	Who	What was given (per day)	Result
Gambelunghe et al., J Med Food 2003	Men aged 25–30; 10 treated and 10 controls, per Balam et al. 2020	Propolis 1,280 mg plus honey 20 g for 21 days	Urinary testosterone unchanged at days 7, 14 and 21, compared with baseline and controls
Brown et al., Int J Sport Nutr Exerc Metab 2000	20 young men in 8 weeks of resistance training	Chrysin 625 mg, androstenedione 300 mg, DHEA 150 mg, Tribulus 750 mg, indole-3-carbinol 300 mg, saw palmetto 540 mg, or placebo	Free and total testosterone unchanged; estradiol up at weeks 2, 5 and 8; estrone up at weeks 5 and 8; strength gains similar to placebo
Brown et al., Int J Vitam Nutr Res 2001	55 men aged 30–59, 28 days	Same six-ingredient mix (n = 28) or placebo (n = 27)	Total testosterone and PSA unchanged; androstenedione +342%, free testosterone +38%, DHT +71%, estradiol +103%; HDL cholesterol down 5.0 mg/dL
Brown et al., J Am Coll Nutr 2001	55 men aged 30–58, 28 days	Chrysin 300 mg, androstenediol 300 mg, saw palmetto 480 mg, indole-3-carbinol 450 mg, gamma-linolenic acid 1,500 mg, Tribulus 1,350 mg (n = 28), or placebo (n = 27)	Total testosterone and PSA unchanged; androstenedione +174%, free testosterone +37%, DHT +57%, estradiol +86%; HDL cholesterol down 0.14 mmol/L (about 5.4 mg/dL)

The combination trials. The herbal extracts were added to raise testosterone and prevent conversion to estrogen. The authors concluded they did not prevent that conversion. Each trial compared the full mixture with placebo, with no androstenedione-only arm, so chrysin’s own contribution cannot be separated.

The same lab did run a separately randomized trial of androstenedione alone in a similar group. It is a useful comparison, though not a head-to-head test:

Change over 28 days	Androstenedione 300 mg/day alone, men aged 30–56 (Brown et al., JCEM 2000)	Androstenedione 300 mg/day plus chrysin 625 mg and other extracts, men aged 30–59 (Brown et al. 2001)
Estradiol	+68%	+103%
DHT	+83%	+71%
Free testosterone	+45%	+38%
Total testosterone	No change	No change

Adding chrysin and the other extracts did not visibly blunt the estradiol rise. Androstenedione alone also lowered HDL cholesterol by 10% in the 2000 trial, so the HDL drop cannot be attributed to chrysin. A 2021 review notes that apart from the disposition study, these clinical trials did not measure chrysin exposure, which limits how they can be read (Gao et al. 2021).

The propolis trial. This is the only human study built around the testosterone claim, and it was negative. It has several limits. Honey adds very little chrysin, as shown above. Its abstract reports urinary testosterone only, which is an indirect measure, and it does not report estrogen.

Androstenedione and androstenediol have been Schedule III anabolic steroids in the US since the 2004 steroid law took effect, so these mixtures are no longer legal supplements; see the androstenedione and 4-androstenediol pages.

Testosterone production in lab and animal models

Mouse Leydig cells (the testosterone-producing cells of the testis). Chrysin increased steroid output when cells were stimulated with cyclic AMP, apparently by raising a protein that moves cholesterol into mitochondria. It did nothing without that stimulus (Jana et al., J Endocrinol 2008).
Adult gerbils. Oral chrysin at 50 mg/kg/day for up to 21 days did not raise testosterone in males or females. Estradiol rose in females at 21 days. Prostates showed more androgen receptors, faster cell growth and glandular hyperplasia, though with less hyperplasia than testosterone cypionate (Campos et al., Int J Exp Pathol 2019).

These findings do not establish that chrysin raises testosterone in men, and gerbil prostate results may not apply to people.

Thyroid

We found no human study that measured TSH, T3 or T4 after chrysin. The 30-day safety panel in the micellar trial did not include thyroid hormones (Ibi et al. 2025). The lab and animal findings point in different directions:

Rat thyroid cells. In a 2026 screen of 424 natural compounds, chrysin raised iodide uptake 3.4-fold and activity of a reporter for the sodium/iodide symporter (the iodide pump) 2.3-fold, but it did not increase that pump’s mRNA or protein. Compounds were screened at micromolar concentrations, and the authors caution that in vitro potency needs interpreting carefully without human pharmacokinetic data (Psarias et al., Redox Biol 2026).
Thyroid peroxidase. This enzyme builds thyroid hormone. A 2021 paper lists chrysin among flavones with reported inhibitory potential, citing earlier work (Habza-Kowalska et al., Int J Mol Sci 2021). We did not find a chrysin-specific human result.
Thyroid cancer cells. Chrysin has been reported to suppress growth of anaplastic thyroid carcinoma cells through Notch1 signaling (Psarias et al. 2026).
Rats. In a 15-day study with 6 rats per group, a nanocrystal form at 5 or 10 mg/kg was tested with and without the pesticide chlorpyrifos. T3 and T4 were lower in all treated groups than in controls, including the groups given chrysin without the pesticide, and TSH did not change. Thyroid tissue changes were described in all groups, which makes the finding hard to interpret (Farkhondeh et al., Endocr Metab Immune Disord Drug Targets 2026).

Together these results raise a question rather than answer one. The clinical effect on thyroid function, if any, is unknown.

Side effects and risks

Human tolerability

Human safety data are small and short:

Micellar product, 30 days. In an open-label, placebo-free extension with 15 adults, there were no serious adverse events and no clinically meaningful liver or kidney changes. One person had bloating, heartburn and abdominal discomfort that improved with food. Another developed an itchy rash after week 2 and stopped; the link to the product was unclear. Fasting glucose fell in men (5.9 to 5.1 mmol/L) and women (5.2 to 4.2 mmol/L), without HbA1c changes. The authors could not say whether chrysin, quercetin or rutin was responsible (Ibi et al. 2025).
Cancer patients on irinotecan. In an uncontrolled pilot in 20 patients with advanced colorectal cancer, chrysin was given twice daily for a week before and after each infusion (250 mg/day total, per Gao et al. 2021). No toxicities were attributed to chrysin. Grade 3 diarrhea occurred in 10%, but the drug’s key metabolite ratio was similar to historical controls, and a randomized trial is needed to judge any benefit (Tobin et al., Cancer Chemother Pharmacol 2006).

Long-term safety in people is unknown.

Animal and cell toxicity

These results flag questions for study and do not translate directly into human risk:

Rats. In acute and sub-chronic oral testing, the paper reports an LD₅₀ of 4,350 mg/kg alongside 40% mortality at 5,000 mg/kg, figures that do not sit neatly together, so the acute value is imprecise. The no-observed-adverse-effect level was 500 mg/kg. At 1,000 mg/kg daily, body weight fell, male liver weight rose, and liver and kidney tissue changes appeared (Yao et al., Drug Chem Toxicol 2021).
Mutagenicity. Chrysin tested positive in a micronucleus assay in human HepG2 liver cancer cells at 1–15 µM and in Salmonella strain TA100 (Oliveira et al., J Toxicol Environ Health A 2012). We found no human data on this question.
Estrogen-like activity. In MCF-7 breast cancer cells, chrysin stimulated growth with an EC₅₀ (half-maximal effect) of 4 µM, in the same range as its aromatase inhibition. The authors urged caution with such compounds as supplements, and their co-culture model could not show which effect dominates (van Meeuwen et al. 2007).

Drug interactions

Apart from the irinotecan pilot above, we found no human study designed to test drug interactions. Lab and rat data suggest where to look.

Chrysin and its main metabolites inhibited several drug-handling proteins in vitro, and the authors advise care when combining chrysin supplements with medications (Mohos et al., Drug Metab Dispos 2020). IC₅₀ values below come from that study, as tabulated by Gao et al. 2021:

Target	Chrysin (µM)	Chrysin sulfate (µM)	Chrysin glucuronide (µM)
CYP2C9	3.2	2.7	Over 100
CYP2C19	4.6	Over 100	Over 100
BCRP	0.4	0.6	19.8
OATP1B1	Over 100	0.8	4.4
OATP2B1	4.8	0.5	0.3

The metabolites matter because they circulate at far higher levels than chrysin itself. In rats, oral chrysin at 200 mg/kg raised the AUC of the antibiotic nitrofurantoin 1.76-fold and its peak 1.72-fold, an effect the authors linked to BCRP (Wang & Morris, Drug Metab Dispos 2007). Other animal studies found BCRP inhibition by chrysin played only a minor role, which the Gao review partly attributes to low solubility.

Chrysin also induces the detoxifying enzyme UGT1A1 in intestinal and liver cancer cell lines. The Gao review notes that its rapid metabolism may limit this in the body.

Who has no safety data

The micellar trial excluded pregnant and breastfeeding women. We found no human safety data for chrysin supplements in pregnancy, breastfeeding or children, and none in people with hormone-sensitive conditions.

For other products marketed for estrogen control, see DIM and arimistane. Its close relative apigenin has a similar pattern of lab aromatase activity without human hormone data.

Key takeaways

A lab aromatase inhibitor, not a proven one in people. Chrysin blocks aromatase in cell assays at about 1–7 µM, but no human study shows lower estradiol.
Very low absorption. After 400 mg, peak free chrysin was 12–64 nM in 7 adults, and one stool collection held 98% of the dose.
It did not stop estrogen conversion in combination trials. In two 28-day trials, men taking androgen precursors with 300–625 mg/day of chrysin had estradiol rises of 86–103%.
The testosterone claim is unsupported. A 21-day propolis and honey trial found no change in urinary testosterone, and we found no trial of chrysin alone that measured serum testosterone.
Better formulations are untested for hormones. A micellar product about doubled exposure compared with unformulated chrysin, but that trial did not measure sex hormones.
Safety data are thin. Short human studies report few adverse effects, while cell data on drug-handling proteins, rat toxicity at high doses and thyroid findings remain unresolved.

Sources

National Center for Biotechnology Information. PubChem Compound Summary for CID 5281607, Chrysin. PubChem.
Gao S, Siddiqui N, Etim I, et al. Developing nutritional component chrysin as a therapeutic agent: bioavailability and pharmacokinetics consideration, and ADME mechanisms. Biomed Pharmacother. 2021;142:112080. PubMed; PMC.
Pavlešić T, Poljak S, Mišetić Ostojić D, et al. Mint (Mentha spp.) honey: analysis of the phenolic profile and antioxidant activity. Food Technol Biotechnol. 2022;60(4):509–519. PubMed; PMC.
Balam FH, Ahmadi ZS, Ghorbani A. Inhibitory effect of chrysin on estrogen biosynthesis by suppression of enzyme aromatase (CYP19): a systematic review. Heliyon. 2020;6(3):e03557. PubMed; PMC.
Walle T, Otake Y, Brubaker JA, et al. Disposition and metabolism of the flavonoid chrysin in normal volunteers. Br J Clin Pharmacol. 2001;51(2):143–146. PubMed; PMC.
Ibi A, Chang C, Kuo YC, et al. Comparative pharmacokinetics and safety of a micellar chrysin-quercetin-rutin formulation: a randomized crossover trial. Antioxidants (Basel). 2025;14(11):1313. PubMed; PMC.
US Food and Drug Administration. Questions and answers on dietary supplements. FDA.
World Anti-Doping Agency. World Anti-Doping Code International Standard: Prohibited List 2026. WADA 2026 Prohibited List.
Kellis JT Jr, Vickery LE. Inhibition of human estrogen synthetase (aromatase) by flavones. Science. 1984;225(4666):1032–1034. PubMed.
Saarinen N, Joshi SC, Ahotupa M, et al. No evidence for the in vivo activity of aromatase-inhibiting flavonoids. J Steroid Biochem Mol Biol. 2001;78(3):231–239. PubMed.
van Meeuwen JA, Korthagen N, de Jong PC, et al. (Anti)estrogenic effects of phytochemicals on human primary mammary fibroblasts, MCF-7 cells and their co-culture. Toxicol Appl Pharmacol. 2007;221(3):372–383. PubMed.
Sanderson JT, Hordijk J, Denison MS, et al. Induction and inhibition of aromatase (CYP19) activity by natural and synthetic flavonoid compounds in H295R human adrenocortical carcinoma cells. Toxicol Sci. 2004;82(1):70–79. PubMed.
Gambelunghe C, Rossi R, Sommavilla M, et al. Effects of chrysin on urinary testosterone levels in human males. J Med Food. 2003;6(4):387–390. PubMed.
Brown GA, Vukovich MD, Reifenrath TA, et al. Effects of anabolic precursors on serum testosterone concentrations and adaptations to resistance training in young men. Int J Sport Nutr Exerc Metab. 2000;10(3):340–359. PubMed.
Brown GA, Vukovich MD, Martini ER, et al. Effects of androstenedione-herbal supplementation on serum sex hormone concentrations in 30- to 59-year-old men. Int J Vitam Nutr Res. 2001;71(5):293–301. PubMed.
Brown GA, Vukovich MD, Martini ER, et al. Endocrine and lipid responses to chronic androstenediol-herbal supplementation in 30 to 58 year old men. J Am Coll Nutr. 2001;20(5):520–528. PubMed.
Brown GA, Vukovich MD, Martini ER, et al. Endocrine responses to chronic androstenedione intake in 30- to 56-year-old men. J Clin Endocrinol Metab. 2000;85(11):4074–4080. PubMed.
Jana K, Yin X, Schiffer RB, et al. Chrysin, a natural flavonoid enhances steroidogenesis and steroidogenic acute regulatory protein gene expression in mouse Leydig cells. J Endocrinol. 2008;197(2):315–323. PubMed.
Campos MS, Silva JPA, Lima DS, et al. Short-term exposure to chrysin promotes proliferative responses in the ventral male prostate and female prostate of adult gerbils. Int J Exp Pathol. 2019;100(3):192–201. PubMed; PMC.
Psarias G, Ziros PG, Mageiropoulou A, et al. Limited Nrf2 activation and heterogeneous thyroidal effects in a 424-compound multi-assay screen call for rigorous testing of purported antioxidant and health-promoting supplements. Redox Biol. 2026;94:104222. PubMed; PMC.
Habza-Kowalska E, Kaczor AA, Bartuzi D, et al. Some dietary phenolic compounds can activate thyroid peroxidase and inhibit lipoxygenase-preliminary study in the model systems. Int J Mol Sci. 2021;22(10):5108. PubMed; PMC.
Farkhondeh T, Ahrari F, Rajabi S, et al. The effect of chrysin nanocrystal on the thyroid gland of rats exposed to chlorpyrifos. Endocr Metab Immune Disord Drug Targets. 2026;26:e18715303329277. PubMed.
Tobin PJ, Beale P, Noney L, et al. A pilot study on the safety of combining chrysin, a non-absorbable inducer of UGT1A1, and irinotecan (CPT-11) to treat metastatic colorectal cancer. Cancer Chemother Pharmacol. 2006;57(3):309–316. PubMed.
Yao W, Cheng J, Kandhare AD, et al. Toxicological evaluation of a flavonoid, chrysin: morphological, behavioral, biochemical and histopathological assessments in rats. Drug Chem Toxicol. 2021;44(6):601–612. PubMed.
Oliveira GA, Ferraz ER, Souza AO, et al. Evaluation of the mutagenic activity of chrysin, a flavonoid inhibitor of the aromatization process. J Toxicol Environ Health A. 2012;75(16-17):1000–1011. PubMed.
Mohos V, Fliszár-Nyúl E, Ungvári O, et al. Effects of chrysin and its major conjugated metabolites chrysin-7-sulfate and chrysin-7-glucuronide on cytochrome P450 enzymes and on OATP, P-gp, BCRP, and MRP2 transporters. Drug Metab Dispos. 2020;48(10):1064–1073. PubMed.
Wang X, Morris ME. Effects of the flavonoid chrysin on nitrofurantoin pharmacokinetics in rats: potential involvement of ABCG2. Drug Metab Dispos. 2007;35(2):268–274. PubMed.

Chrysin