Hibiscus

Hibiscus sabdariffa L. belongs to Malvaceae, the mallow family, and NCBI Taxonomy files it as taxon 183260. Linnaeus named it in Species Plantarum in 1753, giving its habitat as India, but the conserved type specimen is cultivated material from George Clifford’s garden at Hartekamp in Holland. In 2025 Barrett and colleagues (Australian Systematic Botany), noting that Hibiscus as traditionally defined is grossly polyphyletic, reinstated the genus Sabdariffa for the 117 species of Hibiscus section Furcaria. Because ‘Sabdariffa sabdariffa’ would break the naming rules, roselle had to take the next available epithet, from Philip Miller’s 1768 Hibiscus gossypiifolius, and is now Sabdariffa gossypiifolia; its relatives kenaf and false roselle moved with it. The authors acknowledge that the new name has very little recognition among users of the plant, and research and product labels still use Hibiscus sabdariffa. Roselle should not be confused with the ornamental Hibiscus rosa-sinensis or Hibiscus syriacus, the flowers most people picture as hibiscus (JSTOR Daily); research on roselle calyces does not apply to them.

Floras collected by World Flora Online describe roselle as an erect annual herb, usually 0.5–2 m tall, with sturdy, often purplish or reddish stems. The leaves are variable: those low on the stem are simple and ovate, while upper leaves are deeply divided into three to five, sometimes seven, narrow toothed lobes. Flowers grow singly in the leaf axils, pale yellow with a dark red or purplish-brown center, roughly 3–7 cm across. Beneath each sits an epicalyx of about 7 to 12 narrow bracts, often red, and a five-lobed calyx that keeps growing after flowering, becoming thick, fleshy, and deep red, up to about 3.5 cm long, around an egg-shaped, usually hairy capsule of kidney-shaped seeds. Britannica notes that roselle is technically perennial but grown as an annual from seed, is very sensitive to frost, and needs about 25 cm of rain a month in the growing season. Food crops are widely spaced, short, and branched, and their calyces are picked when plump and fleshy.

World Flora Online gives the native range as tropical Africa, from Ghana and Nigeria east to Chad and Sudan and south to Gabon and the Congo basin, and Barrett and colleagues infer a tropical African origin, with every other population spread by people. Britannica calls it probably native to West Africa and describes a fiber form, variety altissima, sown densely to grow 3–5 m tall; its retted stalks yield a lustrous fiber, used with jute for sacking and twine, and the Dutch East Indies began large plantings in the 1920s to make sugar sacks. The calyx chemistry explains the color and sourness. Da-Costa-Rocha and colleagues (Food Chemistry 2014) single out the anthocyanins delphinidin-3-sambubioside (PubChem CID 10196837, C26H29O16+) and cyanidin-3-sambubioside, organic acids including hydroxycitric acid and hibiscus acid, and phenolic acids such as protocatechuic acid. Hopkins and colleagues (Fitoterapia 2013) add citric and malic acids. Amounts in a cup are small: the tea in the McKay trial supplied about 7 mg of anthocyanins per serving.

Roselle’s story begins in tropical Africa, where it remains an everyday food and drink. JSTOR Daily lists some of its African names for the beverage, including bissap, wonjo, foléré, dabileni, tsobo, zobo, and sobolo, and Mahmoud and colleagues (Journal of Antimicrobial Chemotherapy 1994) describe karkadi as one of three common Sudanese drinks. Hopkins and colleagues found it consumed as a beverage in Nigeria and other West African countries, Egypt, Iran, India, Thailand, Taiwan, Mexico, and the United States, and recorded folk use against high blood pressure in Egypt, Jordan, and Trinidad and Tobago and against high cholesterol in Jordan, Greece, Brazil, and Trinidad. Intriguingly, ethnomedical studies in Jordan and Iraq recorded its use for low blood pressure, and in Iraq, Greece, and Brazil for obesity, a reminder that folk traditions do not always point the same way. World Flora Online notes that the leaves are edible too, and the fiber form has long been woven into sacking alongside jute.

Britannica says roselle was known in the West Indies early in the sixteenth century and was growing in Asia by the seventeenth. JSTOR Daily notes that it is widely assumed to have reached the Caribbean from Africa and that its migration is closely tied to the Atlantic slave trade, which brought about 10.7 million people to the Americas between 1526 and 1867. Drawing on Judith Carney and Richard Rosomoff’s In the Shadow of Slavery, it describes how enslaved people kept African knowledge alive in provision gardens, known as conucos, kunukus, or ground provisions, where hibiscus became a staple drink plant, called sorrel or agua de Jamaica in the Americas. Carney (African Ethnobotany in the Americas, 2013) notes the English name guinea sorrel, from the slave traders’ term for Africa, and the Brazilian dish arroz de cuxá, rice with sorrel, whose name comes from the Mandinka of West Africa. Barrett and colleagues, citing a 1911 account, credit Sloane’s 1707 natural history of Jamaica with the first record of roselle eaten as food there.

Today roselle is a festive and commercial drink across the Atlantic world. NPR reports that Caribbean sorrel, steeped with ginger and spices, sweetened, and sometimes spiked with rum or wine, became a Christmas drink partly because the plant used to be available only at that time of year. Food historian Adrian Miller calls African American ‘red drink’ a nod to ancestral red drinks that crossed the Atlantic, and says Christmas sorrel has been gaining a foothold in the United States since at least the 1930s, with newspaper references to it being sold on New York streets to Caribbean immigrants. In Mexico, Ojeda and colleagues (Journal of Ethnopharmacology 2010) describe hibiscus beverages as widely used as a diuretic and for digestive complaints, liver disease, fever, high cholesterol, and hypertension, and researchers at the Mexican Social Security Institute ran the trials comparing hibiscus with captopril and lisinopril. Carney also notes that roselle is the plant that gives Red Zinger tea its color.

Several mechanisms have been proposed for the blood pressure effect, none fully proven in people. The first is a mild diuretic action. Herrera-Arellano and colleagues (Phytomedicine 2004) observed increased sodium excretion in people drinking hibiscus infusion, and in their 2007 trial serum chloride rose, sodium tended to fall, and potassium did not change. Hopkins and colleagues found that hibiscus extracts acted as diuretics in animal studies but in most cases did not significantly change electrolytes, and note that this profile has been compared with potassium-sparing diuretics such as spironolactone. A diuretic effect would also explain why interactions with the diuretic hydrochlorothiazide have been studied.

The second proposal is inhibition of angiotensin-converting enzyme (ACE), the target of drugs such as captopril and lisinopril. In the 2007 trial, plasma ACE activity fell from about 44 to 30 units in people taking a hibiscus extract. Ojeda and colleagues isolated delphinidin-3-sambubioside and cyanidin-3-sambubioside from hibiscus calyces and showed that they inhibit ACE in the test tube by competing with its substrate, with IC50 values of 84.5 and 68.4 µg/mL, similar to related plant flavonoids; test-tube activity does not show that the same happens in the body. Hopkins and colleagues also list relaxation of blood vessels through endothelial pathways and blocking of calcium entry into vessel muscle in laboratory studies. Whether enough anthocyanin reaches the bloodstream is unclear: McKay and colleagues detected none in the urine within an hour of drinking their tea (Hopkins).

For cholesterol and blood sugar the mechanisms are laboratory findings. Hopkins and colleagues describe inhibition of HMG-CoA reductase, the enzyme blocked by statins, and antioxidant protection of LDL cholesterol from oxidation, as well as reduced fat-cell development in cell studies, and they note that hibiscus acid has been proposed to inhibit triglyceride synthesis. Da-Costa-Rocha and colleagues add inhibition of the carbohydrate-digesting enzymes α-glucosidase and α-amylase. They caution that most evidence comes from extracts that are poorly characterized chemically, and call for trials using defined preparations. Some of its authors were affiliated with two German botanical companies. Laboratory activity is a reason to test a plant, not evidence that it works in people.

Blood pressure, placebo-controlled trials: the best-known trial is by McKay and colleagues (Journal of Nutrition 2010) at Tufts University. They randomized 65 adults aged 30 to 70 with prehypertension or mild hypertension, not taking blood pressure medicines, to three 240 ml servings a day of brewed hibiscus tea or a placebo drink for six weeks, in a double-blind design. Systolic pressure fell by 7.2 mmHg with hibiscus and 1.3 mmHg with placebo, a significant difference; diastolic pressure fell by 3.1 and 0.5 mmHg, which was not significantly different, and the change in mean arterial pressure was borderline. People with higher starting pressure responded more. Each serving was brewed from 1.25 g of dried hibiscus (Hopkins). It is a careful study, but it is small, short, and measured readings rather than strokes or heart attacks.

Head-to-head with medicines: Herrera-Arellano and colleagues (Phytomedicine 2004) gave 90 Mexican adults with hypertension either an infusion of 10 g of dried calyx in half a liter of water each morning or captopril 25 mg twice a day for four weeks; 75 completed. With hibiscus, pressure fell from about 139/91 to 124/80 mmHg, with no significant difference from captopril. In their 2007 double-blind trial (Planta Medica), 193 adults took a dried extract supplying 250 mg of anthocyanins or lisinopril 10 mg for four weeks; pressure fell from about 146/98 to 130/86 mmHg on hibiscus, but lisinopril worked significantly better. Wahabi and colleagues (Phytomedicine 2010) reviewed four early trials with 390 patients, including these two, judged all but one short and of poor quality, and found no reliable evidence for treating hypertension. A small four-week trial is not enough to show that a tea can replace a drug.

Pooled evidence points to a modest effect with a large question mark. Serban and colleagues (Journal of Hypertension 2015) combined five trials with 390 participants and found systolic pressure 7.58 mmHg and diastolic 3.53 mmHg lower, using a fixed-effect model that can overstate certainty. Ellis and colleagues (Nutrition Reviews 2022) pooled 17 trials: systolic pressure was 7.10 mmHg lower than with placebo (95% confidence interval 1.20 to 13.00 mmHg lower), with extreme heterogeneity (I² of 95%). An umbrella review by Norouzzadeh and colleagues (Complementary Therapies in Medicine 2025) of 26 trials with 1,797 participants found dose-dependent reductions. But Cochrane’s 2021 update by Pattanittum and colleagues, which accepted only placebo or no-treatment comparisons in people with hypertension, found a single eligible trial of 60 people with type 2 diabetes, with no clear effect and very low certainty. Its 2010 version found none.

Cholesterol and blood sugar: Aziz and colleagues (Journal of Ethnopharmacology 2013) found six trials with 474 participants and no significant effect on any blood lipid compared with placebo, black tea, or diet, and NCCIH says hibiscus has been studied for cholesterol but not found effective. Later pooled analyses are more favorable: Ellis and colleagues found LDL cholesterol 6.76 mg/dL lower than with other teas or placebo, a borderline result, and Norouzzadeh and colleagues reported lower total and LDL cholesterol and fasting glucose. In a 12-week Mexican trial of 104 adults (Revista Médica del Instituto Mexicano del Seguro Social 2011), hibiscus extract lowered triglycerides, while pravastatin had the greatest effect on cholesterol. Bule and colleagues (Food Research International 2020) pooled a fall in fasting glucose of only about 4 mg/dL, too small to matter for people with diabetes.

Other uses: for weight, Dilokthornsakul and colleagues (Complementary Therapies in Medicine 2024) pooled six trials with 339 participants and found no meaningful effect on body weight, BMI, or waist size. A meta-analysis of three trials with 178 women with uncomplicated cystitis (Antibiotics 2021) found benefit from a medical device combining xyloglucan, hibiscus, and propolis, but the design cannot show what hibiscus itself contributed. Traditional uses for digestion, fever, and liver complaints have not been tested in good trials. A fair summary is that regular hibiscus tea may lower systolic pressure by a few points in people with raised readings, a modest effect that could be worth having alongside diet and exercise, but not one established well enough to treat hypertension, prevent heart disease, or substitute for medicine.

There is no standardized dose. In the McKay trial, each 240 ml serving was made by steeping 1.25 g of dried hibiscus in boiled water, taken three times a day, an amount the authors described as readily incorporated into the diet. The 2004 Mexican trial used a much stronger infusion, 10 g of dried calyx in half a liter of water once a day, supplying 9.6 mg of anthocyanins, while the 2007 trial used a dried extract providing 250 mg of anthocyanins per dose. Norouzzadeh and colleagues found that effects increased with dose. Trials lasted from two weeks to about three months, so longer-term effects of daily use are not known. Because calyces vary by cultivar, origin, and freshness, two cups of hibiscus tea may contain very different amounts of active compounds.

Hibiscus is sold as whole or cut dried calyces, in tea bags, often blended with other herbs, in bottled drinks, and as capsules and liquid extracts, some standardized to anthocyanins. McKay and colleagues noted that it is an ingredient in many herbal tea blends, so check labels if you are trying to avoid it. Traditional drinks can be strong and sweet. A Caribbean sorrel recipe published by NPR steeps 2 pounds of fresh sorrel with ginger, cinnamon, and allspice for two or three days and adds 1½ to 2 pounds of brown sugar, along with red wine and rum, for 8 cups of drink. The trials used unsweetened tea or capsules, so heavily sugared or alcoholic versions are a different proposition for anyone watching blood pressure or blood sugar.

If you have raised blood pressure and want to try hibiscus tea, agree a plan with your clinician, keep taking any prescribed medicine, and check home readings before and after a few weeks so that any effect, or any fall that is too large, is noticed. Do not take hibiscus capsules or extracts alongside blood pressure medicines without advice. Take chloroquine and other antimalarials with water, not hibiscus drinks. Hopkins and colleagues suggested taking acetaminophen three to four hours before drinking hibiscus, based on a small study showing faster elimination. Give any trial of hibiscus a few weeks: the effects in the McKay trial were measured at six weeks and those in the Mexican trials at four, and a single cup tells you little about your blood pressure.