Claims Review: Arugula, Garlic and Hibiscus for Blood Flow

This is the claim with the most research behind it and the weakest result, which is an uncomfortable combination to report.

Arugula is high in dietary nitrate. Mouth bacteria convert part of it to nitrite, and in a single-dose study an 800 mg nitrate drink made from arugula moved systolic pressure from 122 to 116 mmHg within 150 minutes in 18 healthy young adults (J Nutr, 2016). That is a real measurement. It is also a measurement of one concentrated drink over five hours, in people whose pressure was already normal, against their own baseline rather than against a placebo.

Then come the two trials built to answer the actual question. Thirty people with mildly raised pressure ate about 200 g a day of nitrate-rich vegetables for four weeks. Their blood nitrate and nitrite rose, exactly as the mechanism predicts, and their blood pressure did not fall — 24-hour systolic 127.4 against 128.6 and 126.2 mmHg in the comparison periods, with no statistical difference anywhere (Am J Clin Nutr, 2018). Two years later a larger trial randomised 243 adults aged 50 to 70 whose office systolic sat between 130 and 159 mmHg into leafy greens carrying 300 mg of nitrate a day, a 300 mg nitrate tablet, or neither, for five weeks. The three groups came out the same: −0.5, −1.2 and −0.6 mmHg (Am J Clin Nutr, 2020).

So the pathway is real, the single-dose effect is real, and the two trials designed to see whether eating the vegetables for weeks lowers pressure found that it did not. Our answer: not supported for sustained blood pressure, in the people who were tested.

That figure exists, and it is not about salad. A review pooling 13 trials in 325 people found systolic pressure 4.1 mmHg lower and diastolic 2.0 mmHg lower (J Hypertens, 2017). Two things about it are usually left out when the number travels.

First, the trials in it used inorganic nitrate or concentrated beetroot juice, not leaves on a plate. Second, the effect reached significance only in seated office readings. In the six trials that used 24-hour ambulatory monitoring and the three that used home monitoring — the measurements clinicians trust more — it was not there.

Our answer: the number is published, and it belongs to a supplement and to one way of measuring. Attaching it to a vegetable is where the claim stops being true.

Garlic has the opposite shape to arugula: the pooled numbers look strong, and what produced them is not the clove in your kitchen.

Pooling nine double-blind trials in 482 people with hypertension gave systolic pressure 9.1 mmHg lower and diastolic 3.8 mmHg lower, with the authors writing plainly that the evidence is not strong — the trials disagreed with each other a great deal, and when the analysis was restricted to the better ones the systolic effect shrank (Am J Hypertens, 2015). A second review of 17 trials found systolic 3.75 mmHg lower overall and 4.4 mmHg lower in people with hypertension, and no significant systolic effect in people whose pressure was normal (J Clin Hypertens, 2015). One 12-week trial of a standardised aged garlic extract — 1.2 g a day, 1.2 mg of S-allyl cysteine, in capsules — found systolic 5.0 mmHg lower than placebo in people whose hypertension was not controlled, and it was funded by the manufacturer (Integr Blood Press Control, 2016).

Against that, the best test of ordinary garlic: 192 adults ate the equivalent of one medium clove a day, six days a week, for six months, as raw garlic, garlic powder or aged extract, and their cholesterol did not move at all (Arch Intern Med, 2007). That trial measured lipids rather than pressure, which is its limit, but it is the only study here that tested a culinary amount for half a year.

Our answer: partly supported, for standardised extracts, in people who already have high pressure. Nobody has compared garlic with a prescribed medication head to head, so “works like a tablet” has no study behind it in either direction.

Hibiscus is the one with the most consistent numbers. In a double-blind placebo-controlled trial, 65 adults with pre-hypertension or mild hypertension drank three 240 mL servings a day for six weeks and came out with systolic pressure 7.2 mmHg lower against 1.3 in the placebo group; diastolic did not separate from placebo (J Nutr, 2010). Reviews pooling the trials land in the same region: 7.58 and 3.53 mmHg across five trials in 390 people (J Hypertens, 2015), and 6.67 and 4.35 mmHg across 13 trials in 1,205 people (J Cardiovasc Pharmacol, 2022).

And then the comparison the claim actually makes. In that same 2022 review, hibiscus measured against an active control — a blood pressure medication or another plant preparation — showed no difference at all, and the subgroup with metabolic syndrome showed no significant effect. In the one double-blind trial that put hibiscus against lisinopril directly, hibiscus lowered pressure and lowered it less than the drug (Planta Med, 2007). An earlier systematic review of the four trials then available concluded that they did not provide reliable evidence to recommend hibiscus for primary hypertension, almost all being short and of poor quality (J Hum Hypertens, 2010).

Our answer: the effect on pressure is supported. “As good as the prescription” is not — the one head-to-head trial found the opposite, and it is the claim most likely to lead somebody to stop taking something they need.

Half right, and the wrong half is the half that matters if you take medication. These are groceries and most healthy adults eat all three without a thought. The interactions that have been measured in people are specific and worth naming.

In six healthy volunteers given a single 40 mg dose of simvastatin, a hibiscus beverage raised the drug's clearance by 44.6% and cut its peak concentration by 18%, with the overall exposure ratio falling outside the accepted range for equivalence; the authors advised against taking them together (J Clin Pharm Ther, 2017). In a controlled study in healthy volunteers, 300 mL of a hibiscus beverage for three days reduced the amount of diclofenac recovered in urine, and those authors also advised caution (Phytother Res, 2007). In 78 people with newly diagnosed untreated hypertension, hibiscus increased urine volume more than lisinopril did and raised creatinine clearance, which lisinopril did not (J Physiol Sci, 2017) — a reason to raise it with a clinician if you already take a diuretic.

Two further cautions are honest only if we label them: the interactions reported with losartan, amlodipine and hydrochlorothiazide were measured in rats and rabbits, not in people, and the authors of those papers say so themselves. The same is true of the pregnancy data, which comes from rodent studies at doses far above a cup of tea. There is no human trial of hibiscus in pregnancy, and “no trial” is not the same as “safe”.

Our answer: not supported. Food interacts with medicine, it has been measured for one of these three, and the list above is short enough to read to a pharmacist in a minute.

There is research, which is exactly why this claim travels so well. What the research is, as a body, is worth stating in one place.

The trials above run from a single morning to six months. The largest has 243 participants and the smallest has six. Every endpoint measured is an intermediate one: millimetres of mercury, artery dilation, nitrite in plasma, cholesterol. Not one of these studies measured a heart attack, a stroke, or a death. Several of the reviews carry moderate to high heterogeneity, which is the statistical way of saying the trials did not agree with each other. Two of the strongest-designed vegetable trials came out null. One of the clearest positive results was paid for by the company selling the extract.

Our answer: supported, in the narrow sense, and routinely overstated. “There are trials” and “this will change your health” are different sentences, and the distance between them is this entire page.