Benefits
A source of vitamin C, vitamin K and fiber
Per 100 g, raw broccoli supplies about 89 mg of vitamin C, 102 micrograms of vitamin K and 2.6 g of fiber; boiled broccoli about 65 mg, 141 micrograms and 3.3 g; and frozen chopped broccoli about 56 mg, 81 micrograms and 3.0 g, according to USDA data. NIH notes that heat can reduce the vitamin C in food. These are nutrient contents, not tested health effects.
Gut bacteria composition on a controlled broccoli diet
In a randomized crossover feeding study of 18 healthy adults, adding 200 g a day of cooked broccoli (with 20 g of raw radish as a myrosinase source) to a controlled diet on 15 days of an 18-day period shifted the overall gut bacterial community versus the same diet without broccoli, with more Bacteroidetes and fewer Firmicutes. Diversity did not change, and health outcomes were not measured.
Oxidative DNA damage markers in young smokers
In randomized crossover studies in young Italian men comparing 10 days of broccoli (200 to 250 g a day) with a control diet, oxidized DNA bases in blood cells fell and cells resisted lab-induced DNA damage better after the broccoli period; the abstracts do not say whether this was tested against the control period. These are lab markers in 20 to 27 people per study; in a related analysis, TNF-alpha and IL-6 did not change.
Raw versus cooked broccoli and sulforaphane formation
In small human crossover studies, crushed raw broccoli gave about 37% sulforaphane bioavailability versus 3.4% for cooked, and fresh broccoli gave about three times more isothiocyanates in urine than steamed. Lightly microwaved broccoli gave about 3 times more sulforaphane than fully cooked. These measure exposure, not health outcomes.
LDL cholesterol with a high-glucoraphanin broccoli type
In two double-blind 12-week trials (130 UK adults at raised cardiovascular risk), 400 g a week of a specially bred high-glucoraphanin broccoli lowered LDL cholesterol by 7.1% and 5.1% from baseline, while standard broccoli gave no significant change. The two types differed significantly only when both studies were pooled. A seed company co-funded the work.
Blood pressure in a cruciferous vegetable soup trial
In a 2-week crossover trial in 18 older Australian adults with mildly raised blood pressure, soups of cruciferous vegetables (40% broccoli, with cauliflower, cabbage and kale) lowered 24-hour systolic pressure 2.5 mmHg more than root and squash vegetable soups. Cholesterol and glucose did not differ. Broccoli was not tested alone.
Broccoli sprouts eaten as food and stomach bacteria markers
In a randomized trial of 48 Japanese adults with H. pylori infection (the study population), 70 g a day of fresh broccoli sprouts for 8 weeks lowered breath test, stool antigen and stomach inflammation markers, while alfalfa sprouts did not (changes within each group, not a direct comparison). Values returned to baseline 2 months after stopping. This is not a treatment for infection, which needs medical care.
Mechanism of action
Glucoraphanin and myrosinase
In intact broccoli, glucoraphanin and the enzyme myrosinase are kept apart; cutting or chewing raw broccoli brings them together to form sulforaphane. Heat inactivates myrosinase, so after cooking the conversion relies on gut bacteria, which in human studies yielded much less sulforaphane, and later.
Cell-protective enzyme signaling (lab work)
Sulforaphane activates NRF2, a cell signal that switches on protective and detoxification enzymes, shown mainly in laboratory and animal studies (see our sulforaphane page). In the LDL trials the broccoli was blanched and frozen, which the authors note destroys myrosinase, so how much sulforaphane formed is uncertain.
Vitamin K and blood clotting
Broccoli is a source of vitamin K1 (phylloquinone), which the body needs for clotting proteins such as prothrombin. Warfarin works by opposing vitamin K, so changes in vitamin K intake can change how warfarin works.
Clinical trials
Two independent double-blind randomized parallel dietary trials comparing 400 g a week of high-glucoraphanin broccoli (bred with a Myb28 allele from the wild species Brassica villosa) with 400 g a week of standard broccoli for 12 weeks; the broccoli was blanched, frozen and steamed at home. Funded by UK research bodies with additional funding from Seminis Vegetable Seeds. (Armah et al. 2015, Mol Nutr Food Res)
130 UK adults aged 50 to 77 with a raised estimated 10-year cardiovascular risk (37 in study 1, 93 in study 2).
Within groups, high-glucoraphanin broccoli lowered LDL cholesterol by 7.1% (study 1) and 5.1% (study 2), while standard broccoli lowered it by 1.8% and 2.5% (not significant). Pooled across both studies, the difference between types was significant (P = 0.031). HDL and triglycerides did not change. The broccoli is patented, one author is a named inventor, and the rights are licensed to Seminis. There was no broccoli-free group.
Randomized crossover dietary study: 250 g a day of steamed broccoli or a control diet for 10 days each, with a washout; oxidized DNA lesions, resistance to lab-induced DNA damage and repair enzyme expression measured in blood cells. (Riso et al. 2010, Mutagenesis)
27 young healthy smokers in Italy.
After broccoli, oxidized DNA lesions were 41% lower (95% CI 10% to 72%) and resistance to lab-induced DNA strand breaks 23% higher (95% CI 13% to 34%). Repair enzyme expression and activity did not change. Lab markers in blood cells over 10 days; the abstract does not say whether the changes were tested against the control period or state funding, and no clinical outcomes were measured.
Randomized crossover design: a 200 g portion of broccoli a day or a controlled diet for 10 days each, separated by a 20-day washout; DNA damage by comet assay, IGF-I and HDAC activity. (Riso et al. 2009, Nutr Cancer)
20 young healthy men in Italy (10 smokers, 10 nonsmokers).
DNA strand breaks fell 22.2% after the broccoli diet in smokers and nonsmokers, and oxidized purines fell 51.0% in smokers only. IGF-I levels and HDAC activity did not change. Lab markers in a small group over 10 days; the abstract does not say whether the changes were tested against the control period.
Dietary intervention with a repeated measures crossover design: 250 g a day of broccoli for 10 days; carotenoids, folate, CRP, TNF-alpha, IL-6, IL-6 receptor and adiponectin measured. (Riso et al. 2014, Int J Food Sci Nutr)
Young male smokers in Italy (number not stated in the abstract).
Plasma folate rose 17% and lutein 39%. TNF-alpha, IL-6, IL-6 receptor and adiponectin did not change. CRP fell 48% only in a post-hoc analysis, so that result is exploratory.
Single-blind randomized crossover controlled feeding study: the same base diet with or without 200 g a day of broccoli (microwave-steamed from frozen) plus 20 g of raw daikon radish (added as a myrosinase source) on days 1 to 15 of each 18-day period, with a 24-day washout; on day 17 both periods included a broccoli test meal for metabolite testing. Funded by the USDA Agricultural Research Service and the National Cancer Institute. (Kaczmarek et al. 2019, J Nutr Biochem)
18 healthy adults in the US.
Broccoli changed overall gut community structure versus control (P = 0.03), lowered Firmicutes by 9% (P = 0.05) and raised Bacteroidetes by 10% (P = 0.03) and Bacteroides by 8% (P = 0.02). Alpha diversity did not differ. Effects were strongest in people with a BMI under 26. Microbial markers only; no symptoms or health outcomes were measured.
Randomized controlled crossover trial, participants blinded to which soup was active: about 300 g a day of cruciferous vegetables as soup (40% broccoli, 25% cauliflower, 25% cabbage, 10% kale) versus root and squash vegetables, 2 weeks each with a 2-week washout. Funded by Edith Cowan University and the Western Australia Department of Health. (Connolly et al. 2024, BMC Med)
18 adults (median age 68, 16 women) in Australia with systolic blood pressure of 120 to 160 mmHg; the planned sample size was 25.
24-hour systolic pressure fell 2.5 mmHg more than with control soup (P = 0.002) and triglycerides 0.2 mmol/L more (P = 0.047). Total, LDL and HDL cholesterol and glucose did not differ between soups, and nighttime aortic diastolic pressure was 2.9 mmHg higher with the crucifer soup (P = 0.014). No adverse events were reported. A blend, so the effect cannot be assigned to broccoli alone.
Randomized trial of 70 g a day of fresh broccoli sprouts (420 micromoles of glucoraphanin) versus the same weight of alfalfa sprouts for 8 weeks, published with mouse experiments. (Yanaka et al. 2009, Cancer Prev Res (Phila))
48 adults in Japan with H. pylori infection.
Urea breath test, stool antigen and serum pepsinogens I and II fell with broccoli sprouts but not with alfalfa sprouts, and returned to original levels 2 months after stopping. The abstract reports changes within each group rather than a direct between-group test. It did not test eradication and is not a treatment.
Randomized, free-living, open crossover trial of 200 g of crushed broccoli eaten raw or cooked with a warm meal; sulforaphane measured in blood and urine. (Vermeulen et al. 2008, J Agric Food Chem)
8 men.
Sulforaphane bioavailability was 37% from raw broccoli versus 3.4% from cooked (P = 0.002), and peak blood levels came at 1.6 hours versus 6 hours. Excretion half-lives were similar. An exposure study in 8 people, not a health outcome.