Lowering blood pressure naturally: what studies can measure

Maybe it was the machine at the pharmacy entrance, where you put your arm in more out of curiosity than anything else. And then there was a number that does not fit you. After all, you feel fine.

Or you have been taking a tablet for a long time and wonder what you can contribute yourself. Many people know both situations. Online you then find lists of home remedies, but hardly any of them tell you how large an effect is and how well it is supported.

Here I try a different approach. I sort the levers by what was measurable in mmHg in studies, and I mark visibly where the evidence gets thin.

Why high blood pressure is dangerous, precisely because you do not feel it

Imagine a garden hose that is under slightly too much pressure day and night. It does not burst, it does not drip. But the material is working, every minute, for years. Your arteries are that hose, and they have no warning light.

From the perspective of KPNI (Clinical Psychoneuroimmunology), blood pressure is the result of several control loops. The nervous system narrows or widens the vessels via the sympathetic nervous system. The hormonal system determines, via renin, angiotensin and aldosterone, how much salt and water the kidneys retain. Many levers in this article can act at exactly those points. That is mechanistically well founded, but in the end what gets measured is pressure in mmHg.

And then there is the invisibility. A worldwide analysis of 1,201 population-representative studies showed that in 2019 only 59 percent of women and 49 percent of men with hypertension had a diagnosis. Germany was among the countries with comparatively high treatment and control rates, but comparatively good does not mean complete (NCD-RisC, Lancet 2021, PMID: 34450083 [Pooled analysis of population-representative studies]).

In German routine data from 4.8 million people with statutory health insurance, 82.3 percent of those newly diagnosed in 2023 received medication within one year. A 24-hour measurement appeared in only 30.1 percent (Reitzle and colleagues, Dtsch Arztebl Int 2025, PMID: 41131981 [Routine data analysis]). This figure becomes important in the next chapter.

The thresholds: three guidelines, three types of measurement

This gets briefly confusing, because the guidelines do not use the same words. The European Society of Hypertension, ESH for short, published its own guideline in 2023, the European Society of Cardiology, ESC for short, in 2024. The last joint guideline of both dates from 2018. In Germany, the National Disease Management Guideline on Hypertension 2023, NVL for short, also applies.

So each type of measurement has its own threshold. A home average of 137 over 86 is already above the threshold; the same reading in the office is still below it.

The ESC has also dropped the words normal and optimal, because risk rises with every mmHg even in the lowest category. Someone who measures 125 over 82 in the office is normal according to the ESH and elevated according to the ESC. Both describe the same person.

The paths also diverge on the target under treatment. The ESC names a standard systolic target of 120 to 129 mmHg if tolerated, more relaxed from around age 85, with frailty or with dizziness on standing. The NVL considers a reading below 140/90 mmHg ideal, with an individual target corridor. Both positions rest on reasoned trade-offs. Which target applies to you is something you decide together with your doctor.

And now you know why a number without the type of measurement and without the guideline is only half the information.

Measuring correctly: the underestimated first lever

Many people know this pattern. At the doctor's office the reading is always too high, at home it is calmer. Or exactly the other way round, and nobody knows which number to believe. Before you think about salt, sport or tea, an honest look at the measuring itself is worthwhile.

Systolic mean values from two randomized crossover trials. They show the order of magnitude of individual error sources, not values for you personally.

Put these numbers next to the levers that follow: about one mmHg per kilogram of weight loss, a little over four mmHg with about 4.4 g less salt per day. These are figures from different studies and not a direct comparison. But they show that imprecise measuring can hide a real change or fake one.

Office, home or 24 hours: which measurement counts

Two patterns have their own names. With the white coat effect, the reading is elevated in the office but not in daily life. With masked hypertension, it is the other way round. According to the NVL, both are most reliably detected with a 24-hour measurement. The white coat effect is not imagination: the sympathetic nervous system can respond to the appointment situation by narrowing the vessels, which is mechanistically plausible.

The NVL therefore provides: if suspicion persists after office measurement, a 24-hour measurement should be recommended, or alternatively home measurement with a validated device. In the routine data mentioned above, 24-hour measurement appeared in just under a third of cases. You are allowed to ask for it.

And now you know why this article starts with measuring: without reliable numbers, you cannot judge any other lever.

The big levers on the plate and in the glass: weight, DASH, salt, potassium, alcohol

Less salt. More vegetables. A few kilos less. You have heard these sentences often, and they sound like a wagging finger. I am not showing them to you here as an obligation, but as numbers from controlled trials, in some of which the food was even provided.

Weight: about one mmHg per kilogram

The NVL phrases it cautiously: with overweight, weight reduction should be recommended, and it rates the evidence for the blood pressure effect as low. And one sentence matters to me: weight is not purely a matter of willpower. Sleep, hormones, stress, medication and genes all play a part. You can find the bigger picture in Understanding weight holistically, and if the kilos stay despite all your effort, in Insulin resistance and weight loss.

DASH: an eating pattern, not a diet program

DASH stands for Dietary Approaches to Stop Hypertension. Behind it is not a product but an eating pattern: rich in vegetables and fruit, plus low-fat dairy products, less saturated fat and less fat overall.

In the original DASH trial, 459 adults ate provided meals for eight weeks, with a constant amount of salt and constant body weight. The combination diet lowered blood pressure by 5.5 mmHg systolic and 3.0 mmHg diastolic more than the control diet, and in the 133 participants with hypertension by 11.4 and 5.5 mmHg (Appel and colleagues, N Engl J Med 1997, PMID: 9099655 [RCT, feeding study]). So the composition of the food alone shifted blood pressure.

A later analysis of the same data in participants not taking blood pressure medication showed how much this depends on the starting point. The combination of low sodium and DASH, compared with high sodium and the control diet, was 5.3 mmHg lower at a baseline systolic reading below 130 mmHg, 7.5 at 130 to 139, 9.7 at 140 to 149 and 20.8 mmHg at 150 mmHg or above (Juraschek and colleagues, J Am Coll Cardiol 2017, PMID: 29141784 [RCT, secondary analysis]).

An honest limitation belongs here: the food was provided, and each phase lasted only a few weeks. In daily life with a canteen, family and time pressure, that is harder. The NVL therefore considers statements on the effectiveness of individual dietary interventions not possible, but sees Mediterranean and DASH diets as reference points for a change.

Salt: the lever that is already in the food

Salt is not only a question of the salt shaker. The NVL notes that hidden salts in foods are often not taken into account.

A dose-response analysis of 133 studies found a 1.10 mmHg lower systolic reading for every 50 mmol less sodium, more pronounced in older people and at higher baseline readings. Studies shorter than 15 days showed less than half the effect of longer studies (Huang and colleagues, BMJ 2020, PMID: 32094151 [Meta-analysis, RCTs, dose-response]). So patience pays off.

And do lower numbers mean fewer events? In the follow-up of the salt trials TOHP I and II, 10 to 15 years later, the group originally randomized to salt reduction had 25 percent fewer cardiovascular events (relative risk 0.75). The difference in mortality was not significant. It was an observational follow-up (Cook and colleagues, BMJ 2007, PMID: 17449506 [Cohort, long-term follow-up of RCTs]).

The guidelines agree on the direction: the NVL makes a strong recommendation for less than 6 g of table salt per day, the ESH for below 5 g of salt, the ESC for about 2 g of sodium. The NVL adds a caveat. With an increased risk of sodium deficiency, harm cannot be ruled out with certainty, for example in older age, with kidney disease, low fluid intake, a vegan diet or on medications such as SSRIs, carbamazepine or older antipsychotics. In those cases, the amount should be agreed with a doctor.

Salt substitutes: the large trial from China and its limits

Since 2025, the WHO has suggested in a conditional recommendation that adults who use table salt replace it with a potassium-containing salt substitute, with low overall certainty of evidence. The recommendation explicitly does not apply to people with impaired kidney function or impaired potassium excretion, to pregnant women or to children (WHO guideline 2025, no DOI [Guideline]). The ESC gives potassium salt a class IIa recommendation for people with hypertension without moderate to advanced kidney disease and with high sodium intake, because so far there is only one outcome trial, in a homogeneous Asian population.

Potassium: the counterpart from vegetables and fruit

This explicitly does not mean potassium supplements. The ESH also prefers diet and excludes advanced kidney disease. And the points from the safety box above apply here just the same: with kidney disease or on ACE inhibitors, sartans, spironolactone or other potassium-sparing drugs, the potassium level in the blood can rise dangerously. In that case, even a deliberately potassium-rich diet should be discussed with a doctor beforehand, and you do not change anything about your prescribed medication yourself.

Alcohol: the amount makes the difference

The NVL recommends not exceeding the limits of low-risk drinking. What alcohol can set in motion elsewhere even in small amounts is described in Alcohol and the gut.

And now you know why I put the plate, the salt shaker and the glass ahead of any dietary supplement.

Exercise, and why an exercise without movement makes headlines

Some time ago a headline made the rounds in the media: an exercise in which you barely move is said to lower blood pressure more than jogging. The wall sit: back against the wall, knees bent, hold. Little time, no gym, big numbers. Let us look at what is behind it.

This section stays with blood pressure. What endurance fitness can reveal about life expectancy is covered in VO2max and life expectancy, calm endurance training as a foundation in Understanding Zone 2 training, and breaking up long periods of sitting in Sitting and lack of movement.

Sounds clear. But four points of context belong here before this turns into a ranking.

• A ranking is not an effect size. The SUCRA value describes how likely a training mode is to rank at the top, not by how many more mmHg it lowers blood pressure. Isometric training reached 98.3 percent, the wall sit as a single exercise 90.4 percent.
• The mmHg values come from separate comparisons, each against its own control groups. You cannot directly read off from this that it is twice as strong as jogging.
• The isometric studies are mostly small and usually measure in the office.
• Data for 24-hour measurement are missing.

An older meta-analysis of 93 studies adds that aerobic training with hypertension was associated with readings 8.3 and 5.2 mmHg lower, but hardly with any change at normal readings (Cornelissen and Smart, J Am Heart Assoc 2013, PMID: 23525435 [Meta-analysis, RCTs]). So those who already had elevated readings showed the larger changes.

And wall sit versus handgrip? In a crossover study with only 21 healthy people, both were associated with lower office blood pressure, systolic by 6.8 and 4.8 mmHg, with no significant difference in the systolic reading (Edwards and colleagues, Eur J Appl Physiol 2026, PMID: 42018000 [RCT, crossover, healthy participants]).

I deliberately do not describe exercise instructions. What this could look like for you belongs in a personal conversation.

The guidelines are broader than any headline. The ESC names 150 minutes of moderate or 75 minutes of vigorous aerobic activity per week, supplemented by dynamic or isometric resistance training two to three times a week. The NVL gives inactive people a strong recommendation for regular moderate activity and considers the form secondary. In 2013, the AHA rated aerobic training with the highest class and isometric handgrip with IIb (Brook and colleagues, PMID: 23608661 [Scientific Statement]). More on muscles as an organ in Strength training after 40 and Exercise as medicine.

And now you know why a headline about the wall sit can be accurate and still not tell the whole story.

The quieter levers: breathing, magnesium, beetroot, garlic, hibiscus, meditation

At the drugstore there is a whole shelf for this: magnesium with a heart on the pack, beetroot powder, garlic capsules. I take these questions seriously, because they come from a good wish to do something yourself. That is why every method gets the same grid: effect, study quality, type of measurement.

Above this section could stand: Mechanistically plausible, human studies small or contradictory. Many of these studies measure only in the office. If an effect disappears in 24-hour measurement, that is an important signal.

Slow breathing: yes when measured openly, no when blinded

What breathing techniques can set in motion in the nervous system is described in Breathing techniques for the nervous system, the physiology of stress in The vagus nerve and stress regulation.

RCT, n=36Partial finding, cell culture
A newer approach is training the inspiratory muscles against resistance. In a double-blind, sham-controlled trial with 36 people aged 50 to 79, the systolic office reading fell from 135 to 126 mmHg after six weeks, while sham training had no effect. Part of the mechanistic data comes from cell cultures exposed to the participants' serum, that is, from the lab (Craighead and colleagues, J Am Heart Assoc 2021, PMID: 34184544 [RCT, sham-controlled]). Interesting, but tested in very few people.

Magnesium: well studied, small effect

Magnesium does not appear in the lifestyle recommendations of the ESH and ESC, as summarized by the reviews analyzed. A general medical precaution that does not come from this study: with impaired kidney function, magnesium is not taken without medical advice. You can find the forms compared in Which magnesium is best.

Beetroot and nitrate: one finding, one cross-check

The cross-check: a meta-analysis of 13 studies with 325 participants lasting one to six weeks found readings that were 4.1 and 2.0 mmHg lower overall, but significant only in office measurement, not in 24-hour and home measurement (Ashor and colleagues, J Hypertens 2017, PMID: 28319596 [Meta-analysis, RCTs]). Mechanistically, nitrate is plausible, because the body can use it to form nitric oxide, which can widen blood vessels. Whether this holds up in daily life over the long term is open.

Garlic: the stricter the study, the smaller the effect

An updated meta-analysis by Ried found systolic readings 8.7 mmHg lower with hypertension. It comes from a single author who searched only one database (J Nutr 2016, PMID: 26764326 [Meta-analysis, RCTs]). Rohner and colleagues pooled 9 double-blind trials with 482 people with hypertension and found 9.1 and 3.8 mmHg, with high heterogeneity. In the methodologically stricter trials, the systolic effect was smaller, and the authors themselves do not call the evidence strong (Am J Hypertens 2015, PMID: 25239480 [Meta-analysis, double-blind RCTs]).

Whether garlic protects against heart attack or stroke is open. A Cochrane review found only two suitable studies up to its 2011 search, and none reported on such events (Stabler and colleagues, Cochrane Database Syst Rev 2012, PMID: 22895963 [Meta-analysis, Cochrane review]).

Hibiscus tea: promising, on a narrow basis

Meditation: possible, with a question mark

In 12 randomized trials with 996 participants, transcendental meditation was associated with readings 4.26 and 2.33 mmHg lower, with a possible risk of bias (Bai and colleagues, J Hum Hypertens 2015, PMID: 25673114 [Meta-analysis, RCTs]). In 2013, the AHA rated it IIb, and other forms of meditation and yoga class III, meaning not recommended for lowering blood pressure. For stress, sleep and quality of life, they can still mean a great deal.

Clinical tradition without a strong study basis: hawthorn or mistletoe are often mentioned in guidebooks. For this article I did not review any studies on them and therefore make no statement about their effect.

And now you know why, with these methods, I always ask first about the type of measurement.

Sleep and sleep apnea: the blood pressure that develops at night

You lie in bed for eight hours and wake up as if you had slept four. Someone tells you about loud snoring and pauses in breathing. Many people know this pattern without connecting it to blood pressure.

How sleep apnea is recognized and treated is described in Recognizing sleep apnea. Here it is only about the connection to blood pressure. In obstructive sleep apnea, the upper airways collapse again and again during sleep, oxygen drops, and the sympathetic nervous system kicks in, often many times per hour. That blood pressure can then remain higher during the day as well is mechanistically well founded.

Sleep counts even without apnea. A meta-analysis of prospective cohorts found a relative risk of 1.21 for new hypertension with short sleep and of 1.20 with difficulty staying asleep (Meng and colleagues, Hypertens Res 2013, PMID: 24005775 [Meta-analysis, prospective cohorts]). What can be influenced about sleep is described in Putting sleep hygiene into practice.

An analysis of individual data from 36 randomized trials with 9,434 people with sleep apnea found that CPAP was associated with a reduction in blood pressure only when the office reading had previously been uncontrolled: systolic by 2.6 mmHg versus 0 mmHg with controlled readings (Pengo and colleagues, Eur Respir J 2025, PMID: 39401854 [Meta-analysis of individual data, RCTs]).

And now you know why the question about your sleep belongs in every blood pressure assessment.

When blood pressure has a reason: aldosterone, kidneys, thyroid, licorice, painkillers

You are doing a lot right. Little salt, regular exercise, hardly any alcohol. And blood pressure stays high. Many people know this frustration, and some quietly blame themselves.

Sometimes blood pressure has a reason that can be found. Medicine speaks of secondary hypertension, and when it is suspected, the NVL makes a strong recommendation for further diagnostic workup.

Aldosterone: more common than long assumed

Aldosterone is a hormone of the adrenal gland. It tells the kidneys to retain salt and water. If the adrenal gland produces more of it than the body needs, this is called primary aldosteronism. Potassium in the blood can fall, but according to the ESC, it remains unremarkable in most of those affected.

Here the guidelines differ markedly. The ESC recommends considering screening for aldosterone and renin in all adults with confirmed hypertension (class IIa), based on an estimated prevalence of about 5 to 20 percent. The ESH gives no formal recommendation on this. The NVL provides for further diagnostics when there are clinical indications, including grade 2 hypertension before the age of 60, grade 3 or readings uncontrolled despite at least three active substances, low potassium occurring spontaneously or on diuretics, a tumor of the adrenal gland and a corresponding family history.

I consider both positions well founded. But aldosterone and renin are not values to interpret yourself. The workup belongs in experienced hands, and you do not stop anything on your own for it.

Kidneys, thyroid and rarer causes

• Kidneys: high blood pressure that begins before the age of 30, a missing nighttime dip in 24-hour measurement, blood or protein in the urine.
• Thyroid: an overactive thyroid, possibly also an underactive one. Which values count is covered in Thyroid values that matter, and more on overactivity in Graves' disease and hyperthyroidism.
• Rare hormonal causes: Cushing's syndrome, pheochromocytoma or acromegaly, investigated when matching signs are present.

Source: NVL Hypertension 2023, Table 6 [Guideline], in everyday language.

Licorice: sweet, black and not harmless

Licorice root contains glycyrrhizic acid. It can inhibit an enzyme in the kidney that inactivates cortisol. Cortisol can then dock onto the aldosterone receptor, and the kidneys can retain more salt. A meta-analysis of 18 studies with at least 100 mg of glycyrrhizic acid daily found an increase of 5.45 mmHg systolic and 3.19 mmHg diastolic, and blood potassium fell by 0.33 mmol/L (Penninkilampi and colleagues, J Hum Hypertens 2017, PMID: 28660884 [Meta-analysis]).

The content varies greatly between types of licorice, which is why I do not give amounts in grams. If your blood pressure is high and you eat licorice regularly, bring it up.

Painkillers and other medications

When taking the medical history, the NVL asks about medications that can raise blood pressure, for example painkillers, the contraceptive pill, immunosuppressants, antidepressants and systemic cortisone, as well as about salt, licorice and caffeine intake and about fitness and diet products. What coffee can set in motion in the body is described in Coffee, cortisol and adenosine.

And now you know why, with persistent high blood pressure, the question of why counts just as much as the question of how much.

Where lifestyle is enough and where it is not

Maybe you are hoping to manage without tablets after this article. Maybe you already take some and experience that as a defeat. Both feelings are understandable. Let us look at what the three guidelines say.

The direction is similar in all three. With low risk and slightly elevated readings, lifestyle comes first; with higher readings or high risk, medication is added alongside it, and lifestyle remains. For every 10 mmHg lower systolic reading, the relative risk of stroke in the medication trials was 0.73. A medication that can give you this protection is not a defeat but a tool.

The NVL describes, based on clinical experience, that weight reduction may under certain circumstances also allow drug therapy to be reduced. But this decision is never made by the reading and never by you on your own, but by your doctor together with you.

Functional medicine additionally looks at the question of what drives the pressure: sleep, sources of salt, lack of exercise, alcohol, medication, hormones. This does not stand against the guidelines but alongside them. What this thinking in overall risks rather than single values looks like for blood lipids is described in Statins from 35.

And now you know why the question is not lifestyle or tablet, but how both fit together for you.

Frequently asked questions about blood pressure and lifestyle

Read on

These articles lead further from here.