High blood pressure is common, usually silent, and responds well to a handful of changes. Because it causes no symptoms, it is found by measuring. How you take the reading matters as much as any treatment. The wrong cuff or an unsupported arm can each add a large error, so getting the technique right comes first.
From there the levers are concrete. Cutting sodium, losing weight, and regular exercise each lower the top number by several mmHg, and the DASH eating pattern stacks on top. Medication comes in when the numbers stay high. Do not stop a prescription on your own.
Practice Ranking
Every practice we track for High Blood Pressure: measuring it right, and lowering it, ranked by how well the evidence supports it for this condition. Strength describes the evidence, not our endorsement.
8 practices · 2 to start with
| # | Practice | Evidence | Type | Cost | Effort | Results In | Add to plan |
|---|---|---|---|---|---|---|---|
| 1 | Sodium Reduction & DASH Diet: What Lowers Blood Pressure, and How to Do It DASH on its own lowered systolic pressure 11.4 mmHg, and about 11.5 mmHg more when sodium was also cut; swapping in a potassium salt substitute cut stroke 14% and death 12%. | Strong | Self-Directed | Free to $ | Moderate | Days to Weeks | |
| 2 | Weight Loss: The Single Strongest Lever for Metabolic Health, and What the Trials Actually Show About 1 mmHg lower systolic per kg lost. | Strong | Self-Directed | Free to $$$ | Moderate to Hard | Weeks to Months | |
| 3 | Walking: How Many Steps You Need, and Where the Curve Flattens Aerobic exercise like brisk walking lowers blood pressure, sitting just behind isometric training in the pooled ranking of exercise modes. | Strong | Self-Directed | Free | Easy | Days to Longer | |
| 4 | Isometric Handgrip Training for Blood Pressure Isometric handgrip work topped a head-to-head ranking of exercise for blood pressure at about 8.2/4.0 mmHg; office readings fell clearly, while 24-hour ambulatory pressure moved less. | Moderate | Self-Directed | Free to $ | Easy to Moderate | Weeks to Months | |
| 5 | Alcohol and Your Health: What the Evidence Shows Now Cutting back helps heavier drinkers most, about 5.5 mmHg in those who were drinking six or more a day. | Moderate | Self-Directed | Free | Moderate to Hard | Weeks to Months | |
| 6 | Tai Chi and Qi Gong: What They Do, the Falls Evidence, and How to Start A year of tai chi lowered systolic pressure about 7 mmHg, slightly ahead of aerobic exercise in a head-to-head trial. | Moderate | Self-Directed | Free to $$ | Easy to Moderate | Weeks to Months | |
| 7 | Magnesium: What It Does, Which Form to Take, and How Much Magnesium lowered systolic pressure about 2 mmHg at 368 mg a day, a small add-on rather than a mainstay. | Moderate | Supplement | $ | Easy | Days to Weeks | |
| 8 | Dietary Nitrate and Beetroot Juice: The Research, the Practice, and How to Use It Beetroot juice and nitrate did not lower pressure in people who already have hypertension; the roughly 2.42 mmHg drop showed up only in healthy people. | Moderate | Self-Directed | $ | Easy | Days | |
Default order puts the best-supported practices first, with self-directed care ahead of clinical options. Click any row to open the practice.
What It Is
Blood pressure is the push of blood against the artery walls, written as two numbers in millimeters of mercury (mmHg). The higher systolic number is the force while the heart beats; the lower diastolic number is the pressure between beats. Above normal, readings fall into named categories. Elevated is 120 to 129 systolic with a diastolic under 80. That early drift upward is the one to catch.
Stage 1 is 130 to 139 over 80 to 89, where lifestyle change is the foundation and whether you need medication depends on your overall cardiovascular risk. Stage 2 is 140/90 or above, usually treated with medication alongside the same changes, and the higher the reading the more likely a drug is needed.
High blood pressure almost never causes symptoms, so it is found by measuring. Most of it builds slowly over years from no single cause. In fewer than 1 in 10 people the cause is single, specific, and often treatable. These include kidney disease, a narrowed kidney artery, a hormone-producing adrenal problem, thyroid disease, and untreated sleep apnea. That is a reason to ask a doctor about the pattern.
Measuring It Right
Getting the reading right can shift the number more than most treatments do, so it is the first thing worth fixing. Upper-arm monitors read more reliably than wrist or finger devices; validated models are best. The largest error is the cuff size. In a crossover trial, a standard cuff on an arm that needed an extra-large one read 19.5 mmHg too high. A too-large cuff on a smaller arm read about 3.6 mmHg low. Measure the mid-upper-arm circumference, match the cuff to it, and set it on bare skin.
Position matters almost as much. Sit with your back supported and both feet flat on the floor, and keep the legs uncrossed. Rest the measured arm on a desk so the cuff sits at heart height. Holding the hand in your lap instead adds about 4 mmHg to the systolic reading; letting the arm hang unsupported at your side adds about 6.5 mmHg. Before the first reading, sit quietly for five minutes, empty a full bladder, and skip caffeine in the hour before. Stay silent through the measurement, since talking pushes the number up.
A single reading means little, because pressure swings minute to minute and day to day with a within-person variation of about 10%. With that much swing, one reading anywhere from 120 to 157 mmHg cannot reliably tell you whether your blood pressure is controlled. Averaging removes most of that noise, and most of the gain arrives by the fifth or sixth measurement.
Measured in a rush at the clinic, only 28 of 100 people looked controlled. Home readings put that at 47 of 100, and careful standardized measurement at 68 of 100. How the reading is taken, by itself, moves a large share of people from one side of the controlled/uncontrolled cutoff to the other.
A home monitor helps mainly through what people do with the numbers. Paired with advice and dose changes it dropped systolic pressure about 6.1 mmHg, while the device alone did little, about 1.0 mmHg. Acting on the readings is what lowers pressure.
Two mismatches show up only with out-of-office readings. White-coat hypertension reads high in the clinic and normal at home. Masked hypertension is the reverse and the more dangerous, normal in the clinic while high the rest of the day, so a single office check misses it. Both are caught with home readings, and where a fuller picture is needed, 24-hour ambulatory monitoring is the best out-of-office reference.
What Lowers It
The same stiffening and narrowing of arteries that raises pressure also drives stroke, heart attack, heart failure, and kidney disease. Lowering the number lowers those events. Every 10 mmHg drop in systolic pressure cuts major cardiovascular events about 20% and stroke about 27%. It cuts heart failure about 28% and death from any cause about 13%, however the drop is achieved.
The Research & Studies
Everything here is based on the research we have collected and checked, sorted into groups and ordered with the strongest evidence first. Click any claim to open the studies behind it.
Measurement And Diagnosis
A too-small cuff on a large arm reads about 19.5 mmHg too high
Blood pressure cuffs come in sizes, and using the wrong one skews the reading. On a large arm that needs an extra-large cuff, a regular cuff read 19.5 mmHg too high, and a too-large cuff on a smaller arm read about 3.6 mmHg low. Match the cuff to your arm before you trust the number.
Compared with a correctly sized cuff, using a regular cuff read 19.5 mmHg higher systolic (95% CI 16.1 to 22.9) in people whose arm required an extra-large cuff, 4.8 mmHg higher (3.0 to 6.6) in those requiring a large cuff, and 3.6 mmHg lower (-5.6 to -1.7) in those requiring a small cuff. Measured in: 195 community-dwelling adults in Baltimore with a range of mid-arm circumferences, mean age 54, 34% male, 68% Black, 51% with hypertension. Readings were taken with an automated oscillometric device in a research setting, so the figures describe the cuff error alone with every other technique variable controlled. In a real clinic the cuff error stacks with the others, not replacing them. The trial was single-site and heavily Black, and mid-arm circumference distribution drives who is affected.
Who this may not transfer to:67 of 195 participants (34%) were male, so the sample is women-weighted; the abstract does not report the cuff-size error separately by sex.
The study · 1
Ishigami et al., effects of cuff size on the accuracy of blood pressure readings: the Cuff(SZ) randomized crossover trial · JAMA Intern Med 2023;183(10):1061-1068
An unsupported arm reads about 6.5 mmHg too high, the lap about 4
Where your arm rests changes the reading. Against an arm supported on a desk at heart height, resting the hand in your lap added about 4 mmHg to the systolic number, and letting the arm hang unsupported at your side added about 6.5. Support the arm at heart height every time.
Against the arm supported on a desk at heart height, resting the hand on the lap overestimated systolic pressure by 3.9 mmHg (95% CI 2.5 to 5.2) and diastolic by 4.0 mmHg (3.1 to 5.0). An unsupported arm at the side overestimated systolic by 6.5 mmHg (5.1 to 7.9) and diastolic by 4.4 mmHg (3.4 to 5.4). Measured in: 133 adults aged 18 to 80 recruited in Baltimore, mean age 57, 53% female; 36% had systolic pressure of 130 mmHg or above and 41% had a BMI of 30 or above. Single-site trial with triplicate automated readings under research conditions, so it isolates arm position with everything else held constant. It does not tell you how often each position is actually used in clinics, which is the quantity that decides the population-level error.
Who this may not transfer to:70 of 133 participants (53%) were female. The trial reports results as consistent across demographic subgroups.
The study · 1
Liu et al., arm position and blood pressure readings: the ARMS crossover randomized clinical trial · JAMA Intern Med 2024;184(12):1436-1442
One reading cannot classify control; averaging 5 to 6 removes most of the noise
Blood pressure swings enough from reading to reading, about 10% within a person, that no single clinic reading between 120 and 157 mmHg can tell whether you are in control. Averaging several readings fixes most of that, with the benefit leveling off around five or six. The same people looked 28% controlled by clinic readings but 68% by careful measurement.
Across 111,181 systolic readings from 444 patients over 18 months, no single clinic reading between 120 and 157 mmHg could classify a patient as in or out of control with 80% certainty. Within-patient coefficient of variation averaged 10%. Averaging removed most of that noise, with most of the benefit accrued by 5 to 6 measurements. Baseline control rates were 28% by clinic measurement, 47% by home measurement and 68% by standardized research measurement in the same people. Measured in: 444 US veterans with hypertension followed in primary care clinics affiliated with the Durham VA Medical Center. What could explain it instead: Patients were enrolled on the basis of previously poor blood pressure control, so regression to the mean inflates the apparent short-term variability relative to a general hypertensive population, and the same selection makes the low clinic control rate at baseline expected, not surprising.. This is a secondary analysis of a management trial, not a study designed to measure variability, and control rates depend on the thresholds chosen (140 mmHg for clinic and research, 135 for home). The three settings used different devices as well as different conditions, so device differences are inside the gap between them.
Who this may not transfer to:The authors state the sample was mostly men, drawn from a Veterans Affairs clinic population with long-standing hypertension, so the variability estimates come from a male-predominant group and were not reported separately for women.
The study · 1
Powers et al., measuring blood pressure for decision making and quality reporting: where and how many measures? · Ann Intern Med 2011;154(12):781-8
Heart And Vascular
Home monitoring cut systolic pressure 6.1 mmHg with support, 1.0 alone
Measuring at home helps mostly because of what comes with it. Paired with advice and medication changes, home monitoring lowered systolic pressure about 6.1 mmHg at a year; the device on its own lowered it about 1.0 mmHg, which is indistinguishable from nothing. The monitor is useful only when you act on the numbers it gives.
At 12 months, self-monitoring lowered clinic systolic pressure by 3.2 mmHg overall (95% CI 1.6 to 4.9). The effect depended almost entirely on what accompanied it: self-monitoring alone gave 1.0 mmHg (-1.2 to 3.3), which is not distinguishable from nothing, while self-monitoring plus intensive support gave 6.1 mmHg (3.2 to 9.0). In the four trials with ambulatory outcomes and little co-intervention, neither clinic nor ambulatory pressure differed from usual care. Measured in: Individual patient data from 25 randomized trials, with primary outcome data from 7,138 of 8,292 randomized adults with hypertension. Significant heterogeneity remained after pooling, driven by different inclusion criteria, monitoring regimes and blood pressure targets between trials. The benefit is attributable to the package, not the device, and the trials that isolated the device found no effect. The author reporting that self-monitoring alone changes little has ties to the blood-pressure monitor industry, so a finding that the device is not the active ingredient carries extra weight.
Who this may not transfer to:The individual patient data allowed a direct test, and the authors report no difference in efficacy by sex.
The study · 1
Tucker et al., self-monitoring of blood pressure in hypertension: a systematic review and individual patient data meta-analysis · PLoS Med 2017;14(9):e1002389
A large sodium cut lowered systolic pressure about 4.26 mmHg, more the more you cut
Eating less salt lowers blood pressure with a clear dose-response. A mean cut of 130 mmol in daily sodium, measured in urine, not food diaries, lowered systolic pressure 4.26 mmHg and diastolic 2.07. The effect was larger in older people, in those starting higher, and in longer trials.
A mean reduction of 130 mmol in 24-hour urinary sodium lowered systolic pressure by 4.26 mmHg (95% CI 3.62 to 4.89) and diastolic by 2.07 mmHg (1.67 to 2.48). Each 50 mmol reduction was worth 1.10 mmHg systolic (0.66 to 1.54). Trials shorter than 15 days found 1.05 mmHg per 50 mmol, less than half the 2.13 mmHg found in longer trials. Measured in: 133 randomized trials, 12,197 adults, with sodium intake measured by 24-hour urinary excretion, not food diaries. Effects were larger in older people, in non-white populations and in those starting with higher pressure, so a single pooled number understates the response in some groups and overstates it in others. These are blood pressure outcomes, not event outcomes, and most trials ran weeks, not years.
Who this may not transfer to:The pooled analysis reports subgroup effects by age, ethnicity and baseline pressure but not by sex, so any difference in sodium sensitivity between men and women is not visible in this review.
The study · 1
Huang et al., effect of dose and duration of reduction in dietary sodium on blood pressure levels: systematic review and meta-analysis of randomised trials · BMJ 2020;368:m315
DASH at low sodium ran 11.5 mmHg below a high-salt diet in people with hypertension
Combining the DASH eating pattern with low sodium stacks the two effects. Eating DASH at low salt sat 11.5 mmHg below the usual control diet at high salt in people with hypertension, and 7.1 mmHg below it in those without. This was a feeding study with all food provided, so it shows the biological ceiling, not what home cooking achieves.
On the control diet, moving from high to intermediate sodium lowered systolic pressure 2.1 mmHg and intermediate to low a further 4.6 mmHg. On the DASH diet the same steps gave 1.3 and 1.7 mmHg. DASH at low sodium sat 11.5 mmHg below the control diet at high sodium in participants with hypertension, and 7.1 mmHg below it in those without. Measured in: 412 adults randomized to a control or DASH diet, each eating high, intermediate and low sodium for 30 days in random order. This is a feeding study: all food was provided, which removes adherence from the result and makes it an estimate of the biological effect, not of what a person achieves cooking for themselves. Each sodium period ran 30 days, so nothing here speaks to whether the effect holds for years.
Who this may not transfer to:The trial reports the sodium effect held in both women and men, and in Black participants and those of other races, though the paper's abstract does not give the effect size separately by sex.
The study · 1
Sacks et al., effects on blood pressure of reduced dietary sodium and the Dietary Approaches to Stop Hypertension (DASH) diet · N Engl J Med 2001;344(1):3-10
The DASH diet alone lowered systolic pressure 11.4 mmHg in people with hypertension
The DASH pattern, built on fruit, vegetables and low-fat dairy with less saturated and total fat, lowers blood pressure even before you cut salt. With sodium and weight deliberately held constant, it lowered systolic pressure 5.5 mmHg overall and 11.4 mmHg in the participants who had hypertension. That effect adds to the sodium effect, not overlapping it.
The combination diet rich in fruit, vegetables and low-fat dairy with reduced saturated and total fat lowered systolic pressure 5.5 mmHg and diastolic 3.0 mmHg more than the control diet. Among the 133 participants with hypertension the reductions were 11.4 and 5.5 mmHg; among the 326 without, 3.5 and 2.1 mmHg. A fruit-and-vegetable diet without the dairy and fat changes gave 2.8 mmHg systolic. Measured in: 459 adults with systolic pressure under 160 mmHg and diastolic 80 to 95 mmHg, fed a control diet for 3 weeks then randomized for 8 weeks. Sodium intake and body weight were deliberately held constant, so this measures the eating pattern alone and the effect is additive to the sodium result, not overlapping it. All food was provided, and the trial ran 8 weeks.
Who this may not transfer to:The published abstract does not break the blood pressure effect down by sex, so the sex balance of the enrolled sample is not reflected in the reported effect sizes.
The study · 1
Appel et al., a clinical trial of the effects of dietary patterns on blood pressure (DASH) · N Engl J Med 1997;336(16):1117-24
A potassium salt substitute cut stroke 14% and death 12% over five years
Swapping ordinary salt for a potassium-enriched substitute is the rare lifestyle change tested against hard outcomes. Over about five years in a large rural trial, it cut strokes about 14%, major cardiovascular events about 13%, and death from any cause about 12%, with no rise in serious high-potassium events. It is not for people with impaired kidneys or on potassium-sparing drugs.
Over a mean 4.74 years, stroke occurred at 29.14 versus 33.65 events per 1000 person-years (rate ratio 0.86, 95% CI 0.77 to 0.96), major cardiovascular events at 49.09 versus 56.29 (0.87, 0.80 to 0.94) and death from any cause at 39.28 versus 44.61 (0.88, 0.82 to 0.95). Serious adverse events attributed to hyperkalemia did not differ (rate ratio 1.04, 0.80 to 1.37). Measured in: 20,995 people in 600 villages in rural China, mean age 65.4, 49.5% female, 72.6% with a history of stroke and 88.4% with a history of hypertension. Open-label and randomized by village, so the effective sample is 600 clusters, not 20,995 individuals. The population was high-risk, rural Chinese, and cooked with added salt at home, which is where the intervention has room to work; the effect in a population eating mostly processed food is untested. Participants taking potassium-sparing diuretics or potassium supplements, and those needing a low-potassium diet, were excluded, so the hyperkalemia safety result does not extend to them.
Who this may not transfer to:49.5% of participants were female, so the trial is close to balanced, though the results are reported for the whole cohort, not by sex.
The study · 1
Neal et al., effect of salt substitution on cardiovascular events and death (SSaSS) · N Engl J Med 2021;385(12):1067-1077
Each kilogram lost lowered systolic pressure about 1 mmHg (11.2 lb (5.1 kg) gave 4.44)
Losing weight lowers blood pressure in proportion to the amount lost, about 1 mmHg of systolic pressure per kilogram. A net loss of 11.2 lb (5.1 kg) lowered systolic pressure 4.44 mmHg and diastolic 3.57. These trials predate modern weight-loss drugs, so they do not describe blood pressure change on a GLP-1 medication, where the loss is larger.
A net weight reduction of 11.2 lb (5.1 kg) lowered systolic pressure by 4.44 mmHg (95% CI 2.95 to 5.93) and diastolic by 3.57 mmHg (2.25 to 4.88), which is 1.05 mmHg systolic and 0.92 diastolic per kilogram lost. Populations losing more than 11 lb (5 kg) fell 6.63 mmHg systolic against 2.70 for those losing less. The diastolic effect was larger in populations taking antihypertensive drugs (5.31 mmHg) than in untreated ones (2.91). Measured in: 25 randomized controlled trials comprising 34 strata and 4,874 participants, published between 1966 and 2002, using energy restriction, increased physical activity or both. The trials predate current obesity pharmacotherapy, so nothing here describes what happens to blood pressure with GLP-1 receptor agonists, where the weight loss is larger and the drug has direct cardiovascular effects of its own. Trial durations were mostly months, and weight regain after a trial ends is not captured.
Who this may not transfer to:The pooled analysis reports subgroups by weight loss achieved and by antihypertensive use, not by sex, so a difference between men and women in blood pressure response per kilogram is not visible here.
The study · 1
Neter et al., influence of weight reduction on blood pressure: a meta-analysis of randomized controlled trials · Hypertension 2003;42(5):878-84
Isometric training led the exercise ranking at about 8.24/4.00 mmHg
Every kind of exercise lowers resting blood pressure, and across 270 trials the type mattered. Isometric training, holding a muscle under steady tension such as a wall squat, lowered it most at about 8.24/4.00 mmHg and ranked first, ahead of combined training (6.04), dynamic resistance (4.55) and aerobic exercise (4.49). The isometric arm rests on fewer and smaller trials, so its lead carries more uncertainty than the ranking alone suggests.
Against non-intervention controls: aerobic training lowered resting pressure 4.49/2.53 mmHg, dynamic resistance training 4.55/3.04, combined training 6.04/2.54, high-intensity interval training 4.08/2.50, and isometric training 8.24/4.00. In the network model the rank order for systolic pressure was isometric training (SUCRA 98.3%), combined training (75.7%), dynamic resistance (46.1%), aerobic (40.5%) and interval training (39.4%). Isometric wall squat ranked first among submodes for systolic pressure. Measured in: 270 randomized controlled trials published 1990 to February 2023, 15,827 participants, with interventions of at least 2 weeks. Every comparison is against a non-intervention control, so these are effects against doing nothing, not head-to-head trials, and the network ranking is inferred from that indirect structure. The isometric arm rests on far fewer and smaller trials than the aerobic arm, so its top ranking carries wider uncertainty than the SUCRA figure suggests. Baseline blood pressure varied widely between trials, and pooled samples include normotensive participants.
Who this may not transfer to:The network meta-analysis does not report the sex composition of its 270 pooled trials, and exercise trial populations are historically male-skewed, so the ranking may not describe women equally well.
The study · 1
Edwards et al., exercise training and resting blood pressure: a large-scale pairwise and network meta-analysis of randomised controlled trials · Br J Sports Med 2023;57(20):1317-1326
Every 10 mmHg drop cut major cardiovascular events about 20%
Lowering blood pressure prevents the events that matter, in proportion to how far it falls. Across 123 drug trials, every 10 mmHg reduction in systolic pressure cut major cardiovascular events about 20%, stroke about 27%, heart failure about 28% and death from any cause about 13% (the effect on kidney failure was not significant). The proportion held at higher and lower starting pressures.
Every 10 mmHg reduction in systolic pressure reduced major cardiovascular events (relative risk 0.80, 95% CI 0.77 to 0.83), coronary heart disease (0.83, 0.78 to 0.88), stroke (0.73, 0.68 to 0.77), heart failure (0.72, 0.67 to 0.78) and all-cause mortality (0.87, 0.84 to 0.91). The effect on renal failure was not significant (0.95, 0.84 to 1.07). Proportional reductions were similar at higher and lower baseline pressures. Measured in: 123 randomized trials of blood pressure lowering treatment, 613,815 participants, each trial with at least 1000 patient-years of follow-up per arm. These are drug trials, and the proportional relationship is derived by meta-regression across trials, not measured within individuals, so applying the same slope to a 10 mmHg fall achieved by diet or exercise is an extrapolation. Smaller reductions were seen in diabetes and chronic kidney disease, and beta blockers underperformed other classes for stroke and major events.
Who this may not transfer to:The tabular meta-analysis does not report the sex composition of its pooled trials or sex-stratified effect estimates, so proportional benefit in women is assumed from the whole, not demonstrated separately here.
The study · 1
Ettehad et al., blood pressure lowering for prevention of cardiovascular disease and death: a systematic review and meta-analysis · Lancet 2016;387(10022):957-967
Untreated white-coat hypertension carried about a third higher cardiovascular risk (HR 1.36)
White-coat hypertension, high in the clinic and normal at home, still carries risk. Left untreated it carried a 1.36 times higher rate of cardiovascular events and a roughly doubled rate of cardiovascular death versus pressure that was normal everywhere. When the high clinic reading was only a treated patient's white-coat effect, no added risk showed up.
Against sustained normotension, untreated white-coat hypertension carried a hazard ratio of 1.36 (95% CI 1.03 to 2.00) for cardiovascular events, 1.33 (1.07 to 1.67) for all-cause mortality and 2.09 (1.23 to 4.48) for cardiovascular mortality. In people already on treatment whose pressure was high only in the clinic, no significant association was found with cardiovascular events (1.12, 0.91 to 1.39) or mortality. Measured in: 27 observational studies, 25,786 people with untreated white-coat hypertension or a treated white-coat effect and 38,487 with normal blood pressure, followed a mean of 3 to 19 years. Every included study is observational, so the association cannot separate white-coat hypertension itself from what accompanies it, and definitions of white-coat hypertension varied between studies (home versus ambulatory thresholds). The confidence interval on cardiovascular events reaches 1.03, so the lower bound is close to no effect. The review notes few studies reported participant race or ethnicity.
Who this may not transfer to:The pooled analysis does not report a sex breakdown across the 27 included cohorts, so whether the risk of white-coat hypertension differs between men and women cannot be read off this review.
The study · 1
Cohen et al., cardiovascular events and mortality in white coat hypertension: a systematic review and meta-analysis · Ann Intern Med 2019;170(12):853-862
Potassium lowered pressure up to about 30 mmol a day, then raised it above 80
Potassium's effect on blood pressure is U-shaped. Blood pressure fell as potassium rose, the benefit faded above about 30 mmol a day of extra intake, and above roughly 80 mmol pressure rose again, mainly in people already on blood-pressure drugs. The lowering was strongest in those with hypertension and high salt intake.
The relationship is U-shaped, not linear. Blood pressure fell as the active-minus-control difference in 24-hour potassium excretion rose, the effect weakened above a difference of about 30 mmol/day, and blood pressure rose again above roughly 80 mmol/day. The increase at high intakes appeared in participants with drug-treated hypertension and not in their untreated counterparts. Lowering was strongest in people with hypertension and at higher sodium intakes. Measured in: 32 randomized trials of at least 4 weeks, mostly crossover designs in adults with hypertension, using supplementation doses from 30 to 140 mmol/day. The estimates at high potassium intake rest on few trials, which the authors flag directly, so the upward limb of the curve is the least certain part of the shape. The analysis models supplementation, not dietary potassium from food, and the two are not interchangeable in how quickly they are absorbed.
Who this may not transfer to:The dose-response model pools 32 trials without reporting a sex breakdown, so whether the U-shape sits at the same doses in women and men is not established by this analysis.
The study · 1
Filippini et al., potassium intake and blood pressure: a dose-response meta-analysis of randomized controlled trials · J Am Heart Assoc 2020;9(12):e015719
Cutting back helped heavier drinkers most: about 5.5 mmHg in those drinking six or more
Cutting alcohol lowers blood pressure only if you drink more than lightly. In people drinking two or fewer a day, reducing it did little. The clearest effect was in those drinking six or more a day who roughly halved their intake, whose systolic pressure fell about 5.5 mmHg and diastolic about 3.97. Most of this evidence is in men.
In people drinking two or fewer standard drinks a day, reducing intake did not produce a significant blood pressure reduction. In people drinking more than two, it did, and the effect was strongest in those drinking six or more a day who cut their intake by about half: systolic pressure fell 5.50 mmHg (95% CI 4.30 to 6.70) and diastolic 3.97 mmHg (3.25 to 4.70). Measured in: 36 trials, 2,865 participants, of whom 2,464 were men and 401 women. 86% of participants were men, and the threshold below which reduction did nothing is defined in standard drinks per day, which maps imperfectly onto how people actually drink in weekly patterns. Trials were short and relied on self-reported intake, which is generally under-reported.
Who this may not transfer to:2,464 men against 401 women. The review stratified by sex, and the threshold and dose-response are estimated far more precisely in men than in women.
The study · 1
Roerecke et al., the effect of a reduction in alcohol consumption on blood pressure: a systematic review and meta-analysis · Lancet Public Health 2017;2(2):e108-e120
Isometric training cut office pressure 7.47 mmHg but not 24-hour ambulatory pressure
Isometric training lowers the office reading more than it lowers round-the-clock pressure. It lowered office systolic pressure 7.47 mmHg, yet 24-hour and daytime ambulatory pressure did not fall significantly, and only night-time pressure dropped, about 4.28 mmHg. Ambulatory pressure predicts events better than an office reading, which tempers the office figure. That null rests on only 4 of the 12 trials, so it is weakly held.
Isometric resistance training lowered office systolic pressure 7.47 mmHg (95% CI 4.84 to 10.10), office diastolic 3.17 mmHg (1.04 to 5.29) and mean arterial pressure 7.19 mmHg (5.32 to 9.06). Night-time ambulatory pressure fell 4.28 mmHg systolic (0.67 to 7.88) and 2.22 diastolic. Neither 24-hour nor daytime ambulatory systolic or diastolic pressure fell significantly, and office pulse pressure and resting heart rate did not change. Measured in: 12 studies, 14 intervention groups, 415 adults with hypertension, randomized controlled and crossover trials against sedentary or sham controls. The 24-hour ambulatory null is the finding that limits the office result, because ambulatory pressure is the better predictor of cardiovascular events. Only a minority of the 12 trials collected ambulatory data at all, so the null rests on a smaller subset than the office figure, and the whole pooled sample is 415 people over interventions of a few weeks. That null itself rests on 4 of the 12 trials and 147 of the 415 people, with heterogeneity at I-squared 92%, so it is a weak null carrying wide uncertainty. It still belongs beside the office-pressure figure, because ambulatory pressure is the better predictor of events.
Who this may not transfer to:The review does not report the sex composition of the 12 pooled trials. Isometric handgrip research has historically enrolled more men than women, so the effect size in women is less well characterized.
The study · 1
Baffour-Awuah et al., isometric resistance training to manage hypertension: systematic review and meta-analysis · Curr Hypertens Rep 2023;25(4):35-49
Acupuncture showed no lasting effect on blood pressure (only a 3.4 mmHg short-term flicker)
Controlled trials do not show acupuncture lowering blood pressure in a way that lasts. The one trial that looked beyond the session found nothing at three and six months. Sham-controlled trials showed a small 3.4 mmHg systolic difference lasting only 1 to 24 hours, at very low certainty. Trials reporting bigger effects compared acupuncture with drugs and were at high risk of bias.
No sustained blood-pressure-lowering effect was demonstrated. Only one trial looked beyond the immediate period and found nothing at three and six months. Four sham-controlled trials produced a short-term (1 to 24 hour) systolic difference of 3.4 mmHg (95% CI 0.9 to 6.0) and diastolic 1.9 mmHg (0.3 to 3.6) at very low certainty. Trials comparing acupuncture with ACE inhibitors and calcium antagonists reported acupuncture doing better, which the reviewers attributed to very high risk of bias. Measured in: 22 randomized trials, 1,744 adults with primary hypertension, searched without language restriction including the Chinese databases CNKI and VIP. Most included trials were at high risk of bias through lack of blinding, and the sham-controlled evidence was graded very low certainty, so the short-term figure is weakly held in both directions. Safety could not be assessed because only eight trials reported adverse events. The review's own recommendation is that future trials use sham controls and measure whether any effect lasts at least seven days.
Who this may not transfer to:The review does not report a pooled sex breakdown across its 22 trials.
The study · 1
Yang et al., acupuncture for hypertension (Cochrane review) · Cochrane Database Syst Rev 2018;11:CD008821
A year of tai chi lowered systolic pressure 7 mmHg, edging out aerobic exercise
In adults with prehypertension, a year of tai chi (four hour-long sessions a week) lowered office systolic pressure 7.0 mmHg, against 4.6 for aerobic exercise. Pooled trials in hypertensive people show larger figures, but English-language trials return about 10.4 mmHg and Chinese-language ones about 18.6, a publication-bias gap that is the reason to plan around the smaller number.
In adults with prehypertension, 12 months of four 60-minute sessions a week lowered office systolic pressure 7.0 mmHg against 4.6 mmHg for aerobic exercise, a difference of 2.4 mmHg (95% CI 0.4 to 4.4), with a similar edge on 24-hour ambulatory readings. In a bias-adjusted meta-analysis of the same intervention in hypertensive samples, English-language trials return about 10.4 mmHg and Chinese-language trials about 18.6 mmHg against non-exercising controls. Measured in: 342 adults with prehypertension at two tertiary hospitals in China, mean age 49.3, 176 women and 166 men; the pooled analysis covers 3,223 middle-aged adults, mean age 56.6. The 10.4 versus 18.6 mmHg split between English-language and Chinese-language trials for the same intervention is publication bias measured, not argued about, and it is the reason to plan around the smaller figure. The larger pooled numbers are against people doing no exercise, so they measure exercise as well as tai chi. The trial dose, four supervised hours a week for a year, is far above what most people sustain.
Who this may not transfer to:176 women and 166 men in the trial. The meta-analysis reports its pooled samples as roughly 69.5% women on average, with wide variation between trials.
The studies · 2
Li et al., effect of tai chi vs aerobic exercise on blood pressure in patients with prehypertension · JAMA Netw Open 2024;7(2):e2354937
Wu et al., Tai Ji Quan as antihypertensive lifestyle therapy: systematic review and meta-analysis · J Sport Health Sci 2021
Magnesium lowered systolic pressure about 2 mmHg at 368 mg a day
Magnesium supplements lower blood pressure a little. A median 368 mg a day for about three months lowered systolic pressure 2.00 mmHg and diastolic 1.78. That is roughly a quarter of what isometric training does and half of what cutting sodium does, so it is a minor addition to the main levers.
A median dose of 368 mg/day for a median 3 months lowered systolic pressure 2.00 mmHg (95% CI 0.43 to 3.58) and diastolic 1.78 mmHg (0.73 to 2.82), alongside a 0.05 mmol/L rise in serum magnesium. A dose of 300 mg/day or a duration of one month was sufficient to raise serum magnesium and move blood pressure. Measured in: 34 randomized double-blind placebo-controlled trials, 2,028 normotensive and hypertensive adults. The effect is around a quarter of what isometric training produced and half of what sodium reduction produced, so the size matters as much as the significance. Larger reductions clustered in higher-quality and lower-dropout trials, and residual heterogeneity remained after the authors accounted for those factors.
Who this may not transfer to:The pooled analysis does not report a sex breakdown across its 34 trials, so the dose threshold is not established separately in women and men.
The study · 1
Zhang et al., effects of magnesium supplementation on blood pressure: a meta-analysis of randomized double-blind placebo-controlled trials · Hypertension 2016;68(2):324-33
Beetroot juice and nitrate did not lower pressure in hypertension; 2.42 mmHg only in healthy people
Beetroot juice and other sources of dietary nitrate are marketed for blood pressure, but the benefit lands in the group with the least to gain. In healthy people, nitrate lowered systolic pressure a small 2.42 mmHg; in people who already have hypertension it did nothing measurable, a non-significant 0.82 mmHg. Trials were short and doses varied between them.
In hypertensive participants, inorganic nitrate did not significantly lower systolic pressure (-0.82 mmHg, 95% CI -2.53 to 0.90), diastolic pressure (-0.03, -1.35 to 1.30), 24-hour ambulatory systolic (-0.22) or 24-hour ambulatory diastolic (-0.33). In healthy participants it lowered systolic pressure 2.42 mmHg (-4.28 to -0.57) with no effect on diastolic or mean arterial pressure. Measured in: 19 randomized controlled trials of inorganic nitrate, split between healthy and hypertensive populations. The positive result sits in the population with the least to gain and the null in the population the supplement is marketed to. Trials were short, doses and nitrate sources varied between them, and the pooled hypertensive subset is smaller than the healthy one, so the null carries wider uncertainty than a flat zero implies.
Who this may not transfer to:The review does not report the sex composition of the 19 pooled trials.
The study · 1
Zhang et al., regulatory effect of dietary nitrate on blood pressure: a meta-analysis of randomized controlled trials · Food Funct 2023;14(4):1839-1850
Daily licorice raised systolic pressure about 5.45 mmHg
Regular licorice raises blood pressure and lowers potassium. A daily intake of about 100 mg or more of glycyrrhizic acid, the active compound, raised systolic pressure 5.45 mmHg and lowered plasma potassium, with the rise tracking the dose. This includes Chinese formulas containing Gan Cao (licorice root), whose content is rarely on the label, so it is a question for the herbalist.
Chronic ingestion of a product containing at least 100 mg of glycyrrhizic acid daily raised systolic pressure 5.45 mmHg (95% CI 3.51 to 7.39) and diastolic 3.19 mmHg (0.10 to 6.29), and lowered plasma potassium by 0.33 mmol/L, plasma renin activity by 0.82 ng/mL/hour and plasma aldosterone by 173 pmol/L. Daily glycyrrhizic acid dose correlated with the systolic rise (r squared 0.55) and the diastolic rise (0.65). Measured in: 18 studies, 337 participants, in trials where the treatment group ingested at least 100 mg of glycyrrhizic acid daily. Individual studies are small and mostly short, and the glycyrrhizin content of a given herbal formula or confectionery is rarely stated, so translating this into a dose from a prescribed decoction requires the herbalist, not the label. Susceptibility varies substantially between people, and the pooled mean hides individuals who responded much more strongly.
Who this may not transfer to:The pooled analysis does not report a sex breakdown, though separate clinical series of licorice-induced pseudoaldosteronism are heavily female, which may reflect prescribing patterns, not susceptibility.
The study · 1
Penninkilampi et al., the association between consistent licorice ingestion, hypertension and hypokalaemia: a systematic review and meta-analysis · J Hum Hypertens 2017;31(11):699-707
Garlic supplements lowered systolic pressure about 8.3 mmHg in pooled trials
Pooled trials put garlic supplements at about 8.3 mmHg systolic and 5.5 diastolic in people with hypertension, which sounds large. It sits at the preliminary tier because a third of those trials come from the reviewing author's own group, which also disclosed sponsorship from the maker of the garlic preparation, so the figure is held loosely until independent trials confirm it.
Pooled across 12 trials, garlic supplements lowered systolic pressure by an average of 8.3 ± 1.9 mmHg and, across 8 trials with 374 participants, diastolic by 5.5 ± 1.9 mmHg. Measured in: 12 trials, 553 hypertensive participants. Four of the twelve pooled trials come from the reviewing author's own group. The review declares no competing interests while also declaring symposium and travel sponsorship from the manufacturer of the garlic preparation involved, which is the more useful disclosure of the two.
Who this may not transfer to:The review does not report a sex breakdown across the pooled trials.
The study · 1
Ried, garlic lowers blood pressure in hypertensive subjects, improves arterial stiffness and gut microbiota: a review and meta-analysis · Exp Ther Med 2020;19(2):1472-1478
Sodium has the most evidence behind it, with a clear dose-response: the more you cut, the more the pressure falls. About 70% of the sodium in the American diet is built into packaged and restaurant food before it reaches the table (CDC). Manufacturers put it there; the eater rarely adds it. The DASH pattern of fruit, vegetables, and low-fat dairy lowers pressure and stacks with a low-salt diet for one of the largest single effects available.
Exercise of every kind lowers pressure. Static holds such as a wall squat came out on top, at about 8 mmHg systolic, so a short set is a high-return few minutes. Tai chi lowers pressure about as much as aerobic exercise in people with prehypertension. Weight loss lowers pressure in proportion to the amount lost, and cutting back on alcohol helps heavy drinkers while doing little for light ones.
Magnesium lowers pressure about 2 mmHg at a typical dose. Garlic shows a bigger average effect across studies, but much of that evidence comes from one research group with ties to a manufacturer. Potassium follows a U-shape: it lowers pressure up to a point, then above that point can raise it in people already on blood-pressure drugs. Beetroot juice and dietary nitrate help healthy people slightly and did nothing measurable in those with hypertension. Acupuncture showed no lasting blood-pressure effect in controlled trials. None of these replaces cutting salt, losing weight, and moving.
When the numbers stay high, medication does the rest, and four everyday classes are effective and generally well tolerated: ACE inhibitors, angiotensin-receptor blockers, calcium-channel blockers, and thiazide-type diuretics. Medication is added on top of the changes above, working alongside them.
Do not stop a blood-pressure medication, or cut the dose, on your own. The good numbers are in part the drug at work, and if it stops the pressure climbs back within days to weeks, usually with nothing you can feel. Coming off is sometimes the right call, made with a prescriber and with monitoring in place.
What To Do First
None of this needs a prescription to begin, and it works alongside whatever your prescriber has you on. Start with measuring well, since it makes everything after it trustworthy, then add the levers you can hold steadily.
Get a validated upper-arm monitor, use the right cuff size for your arm, sit quietly for five minutes with your back supported and arm resting at heart height, and take morning and evening readings for a week. The average is what counts, and it is the reading worth acting on.
Most sodium comes from bread, processed meat, and restaurant food, not the shaker. Cutting it lowers pressure with a clear dose-response, and a potassium-based salt substitute goes further, though not if your kidneys are impaired or you take a potassium-sparing drug.
Any regular activity lowers pressure. Isometric holds such as a wall squat, a few times a week, topped the exercise ranking, so a short wall-squat set is a high-return addition to walking or whatever you already do.
About 1 mmHg of systolic pressure comes off for every kilogram lost, and the eating changes above do most of that work. Steady loss held over months matters more than the exact method.
If you drink six or more a day, halving that is worth about 5.5 mmHg systolic. If you already drink little, this lever is not for you.
A week of readings is the basis for starting, adjusting, or reviewing treatment. That is a shared decision, and any change to medication belongs there rather than with you alone.
Go Deeper
- Sodium, potassium and the DASH diet: how far cutting salt and a potassium-based substitute move the number, and how to build the DASH pattern.
- Isometric handgrip training: the held-contraction work studied for lowering pressure, and how to run a set at home.
- Resistance training: steady strength work that lowers pressure and protects the heart alongside it.
- Walking and aerobic activity: the everyday movement that lowers pressure and adds up across a week.
- Heat and sauna bathing: regular heat that relaxes blood vessels and tracks with lower pressure.
- Tai chi and qi gong: slow, low-impact movement with a blood-pressure benefit in early hypertension.
- Sleep apnea: untreated apnea is a common secondary driver, and treating it can bring pressure down.
- Pranayama: slow breathing practices studied for a modest drop in blood pressure.
The Chinese Medicine View
The Chinese Medicine View
Chinese medicine has no classical correspondence to high blood pressure, because it is a measurement, not a symptom, and the tradition predates the cuff. What it treats are the presentations that often accompany a high reading, read as a pattern, not by the number alone.
Headache at the temples or crown, dizziness, tinnitus, a red face and irritability, with a wiry pulse. Usually Liver and Kidney Yin failing to anchor the Yang, so deficiency underneath and excess on top. The cooling, downward-draining herbs used here are contraindicated in someone cold, pale, loose in the bowels and depleted.
The same picture intensified, with numbness or tremor of the limbs, a stiff tongue and severe vertigo. Classically the territory of wind-stroke.
A heavy, muzzy head, chest oppression, nausea, and a thick greasy tongue coat, often alongside a richer diet and a heavier body.
Dizziness of a different quality, worse on standing and better on rest, with pallor and fatigue. Here the number may be normal or low.
Cautions
Extra restraintEverything to be aware of is here, in one place. This practice suits most healthy people; a few situations call for real care.
Licorice and Gan Cao raise blood pressure
Regular licorice, and Chinese formulas containing Gan Cao (licorice root), can raise blood pressure and lower potassium. A daily glycyrrhizic acid intake of about 100 mg or more raised systolic pressure around 5.45 mmHg in pooled studies, and the content of a given sweet or decoction is rarely on the label, so this is a question for the herbalist. Susceptibility varies, and some people respond far more strongly than the average.
Common painkillers and decongestants can raise it
Regular NSAID painkillers such as ibuprofen and naproxen, and decongestants such as pseudoephedrine, can raise blood pressure and blunt some medications. Occasional use is minor; a daily habit is worth swapping or flagging to your prescriber.
Do not stop or change medication on your own
Everything here can lower blood pressure, and stacking a large change onto a prescription can drop it too far. Any change to medication is a conversation with the person who prescribed it.
Start slow, be smart, read the research, and consult a professional if you have any concerns. This is here to inform your choice, not make it for you.
When to See Someone
A high number with no symptoms is not an emergency. It is a reason to measure properly over several days and to see a doctor about the pattern. A few combinations do need care the same day or sooner:
- A very high reading (around 180/120 or above) together with a severe or sudden headache, chest pain, breathlessness at rest, a sudden change in vision, one-sided weakness or numbness, trouble speaking, confusion, or a seizure. That combination is a hypertensive emergency: call an ambulance, not a clinic(seek urgent care)
- In pregnancy, a reading of 140/90 or above, or any new headache, visual disturbance, upper abdominal pain, or sudden swelling, which can signal pre-eclampsia(seek urgent care)
- A home average of 135/85 or higher over a week, which is worth taking to a doctor to start or review treatment
- Pressure that is high when you are under 40, rose suddenly, or stays high on three medications, which can point to a secondary cause worth investigating
- Repeated dizziness or faintness on standing, which can mean treatment is lowering pressure too far and is a reason to review it
If a week of good readings makes medication look unnecessary, take that to your doctor as a case for reviewing the dose. Deprescribing is sometimes appropriate, and it is made with monitoring in place.
Common Questions
What is a normal blood pressure?
Below 120/80 mmHg is the range where the heart and arteries work with the least strain. A home average under 120/80 is the target most people are aiming for.
How should I measure at home?
Take two or three readings a minute apart, and use the average of the set instead of any single reading.
How much can I lower blood pressure without medication?
A lot, if the changes stack. The DASH eating pattern at low sodium reached about 11.5 mmHg systolic in people with existing hypertension. Salt reduction, weight loss, and regular activity each add several mmHg on their own.
What is the single most effective change?
A potassium-based salt substitute stands out. Unlike most changes, it was measured against strokes and deaths themselves, and it cut strokes about 14% and deaths about 12% over five years. It is not for you if you have impaired kidneys or take a potassium-sparing diuretic.
Is white-coat hypertension harmless?
Not entirely. Even untreated, it carried about a third more cardiovascular events than pressure that reads normal everywhere, so the pattern is worth tracking over time.
Explore Related
Other pages this one connects to, by the evidence they share, the outcomes they touch, and the ground they cover.
All 22 sources on this page independently checked and cross-referenced.
Thomas Dehli, Founder & Editor, Sacred Lotus
Sacred Lotus has published Chinese medicine reference material since 2001. Integrative pages are held to the same standard as the herb and formula library: cite the source, grade the claim at its real strength, and say where the research has not looked. This page is educational and it is not medical advice. Last reviewed and updated August 9, 2026.
Evidence strength
How confidently the research supports a claim. Strength describes the evidence, not our endorsement.