Search for how to lower blood pressure and you get a list. Move more, eat less salt, drink less, sleep better, lose weight, breathe deeply. Every item is plausible and none of them come with a number.

The numbers exist. They come from meta-analyses covering tens of thousands of people, and they are more interesting than the list, because the levers are not the same size and several of them only work for some people.

Key takeaways

• Aerobic exercise lowered systolic pressure by 8.3 mmHg in people with hypertension and by a statistically insignificant 0.75 mmHg in people with normal pressure (Cornelissen and Smart 2013, JAHA).

• Cutting sodium works in proportion to how much you cut. Each 50 mmol/day reduction was worth about 1.1 mmHg, and roughly double that in trials lasting longer than two weeks (Huang et al. 2020, BMJ).

• Cutting alcohol showed no measurable effect below two drinks a day. Above that it scaled sharply (Roerecke et al. 2017, Lancet Public Health).

• The single largest effect in the exercise literature came from isometric training, at 10.9 mmHg, though from only five study groups.

Which habits actually have evidence behind them?

Four have been measured repeatedly at scale: exercise, sodium, dietary pattern, and alcohol. Here is roughly what each is worth on systolic pressure, the top number. If those numbers need a refresher, blood pressure numbers explained gently covers them.

Read that table by column, not just by row. The largest effect has the thinnest evidence, and the same intervention appears three times with wildly different results depending on who did it.

Why does exercise help some people and not others?

Because the effect depends almost entirely on where you start.

A meta-analysis of 93 randomised trials covering 5,223 people separated participants by baseline blood pressure, and the result is the most useful finding in this whole area (Cornelissen and Smart 2013):

If your blood pressure is already normal, aerobic exercise will do a great many good things for you. Measurably lowering your blood pressure is probably not one of them. That is not a reason to stop; it is a reason not to judge your training by this particular number.

The other surprise in that paper: isometric resistance training produced the largest systolic reduction of any modality, at 10.9 mmHg. Isometric means holding a contraction without movement, typically a handgrip squeeze held for a couple of minutes at a time. The catch is that this came from just five study groups, against 105 for endurance training. It is a promising signal on thin evidence, not an established equal.

Ordinary dynamic weight training came in at −1.8 mmHg systolic but −3.2 diastolic. As with cardio versus strength, the two do different jobs.

How much does cutting salt really do?

It works, it works in proportion to how much you cut, and the trials probably understate it.

Pooling 133 trials with 12,197 participants, an average reduction of 130 mmol of sodium a day (roughly 3,000 mg of sodium, or about 7.5 g of salt) lowered systolic pressure by 4.26 mmHg and diastolic by 2.07 (Huang et al. 2020, BMJ).

The dose-response matters more than the headline. Each 50 mmol/day reduction was worth about 1.10 mmHg systolic. But in trials running longer than 15 days, the same 50 mmol reduction was worth 2.13 mmHg, more than double. Short trials underestimate the effect, which means the real-world payoff of a sustained change is larger than a quick study suggests.

Three groups saw bigger reductions for the same cut: older people, non-white populations, and anyone starting from a higher blood pressure. So sodium is one of the levers where "it depends who you are" is a genuine answer rather than a hedge.

Worth knowing where sodium actually comes from: mostly packaged food, bread, sauces and restaurant meals rather than the salt shaker on your table.

Is the DASH diet worth the effort?

Yes, and less than its reputation suggests.

The DASH pattern (more vegetables, fruit, whole grains and low-fat dairy; less saturated fat and sugar) was tested across 30 randomised trials covering 5,545 people. Compared with a control diet it lowered systolic pressure by 3.2 mmHg and diastolic by 2.5 (Filippou et al. 2020, Advances in Nutrition).

You will see much larger figures quoted for DASH, often from the original 1990s trials. This more recent pooling, which accounts for methodological differences between studies, lands lower.

Two details are useful. Having hypertension did not change the size of the effect, which makes DASH unusual on this list. And the benefit was larger in trials where sodium intake was above 2,400 mg/day, meaning some of what DASH does overlaps with what cutting salt does.

Does cutting back on alcohol help?

Only if you drink more than two a day. This is the cleanest threshold finding in the area, and it is not what the general advice implies.

Across 36 trials with 2,865 participants, reducing alcohol in people who drank two or fewer drinks a day produced no significant blood pressure reduction at all. Above two a day the effect appeared and scaled with intake. The strongest result came from people drinking six or more a day who cut their intake by about half: systolic down 5.50 mmHg, diastolic down 3.97 (Roerecke et al. 2017).

So if you have a glass of wine most evenings and you are cutting back specifically to move your blood pressure, the evidence does not support expecting much. Alcohol does plenty of other things, including raising your resting heart rate overnight after a single drink. Blood pressure just is not where a small reduction shows up.

So what should you actually do?

The honest summary is that the answer depends on your starting point, which is exactly why a personal baseline beats a generic list.

• If your pressure is genuinely high, aerobic exercise is the strongest well-evidenced lever available to you at −8.3 mmHg, and it gets stronger the higher you start.

• If your pressure is normal, none of these will move it much. That is fine. Pursue them for the other reasons.

• Sodium is the most universal lever, working across nearly every subgroup, with the benefit growing the longer you sustain it.

• Alcohol only matters above two a day, at which point it matters a lot.

• Effects are not additive in any simple way. These trials tested one change at a time, and stacking them rarely produces the sum.

One measurement caution before you judge any of this. Single readings are noisy and clinic readings run high, so a change of a few mmHg is invisible in a one-off. Use a week of home readings against the previous week, and remember that readings at the doctor's office run higher for reasons that have nothing to do with you.

Frequently asked questions

How much does exercise lower blood pressure?

It depends almost entirely on your starting pressure. Endurance training lowered systolic pressure by 8.3 mmHg in people with hypertension, 2.1 mmHg in prehypertensive people, and 0.75 mmHg in people with normal pressure, which was not statistically significant (Cornelissen and Smart 2013).

How much salt do I need to cut to see a change?

Roughly proportionally. Each 50 mmol per day reduction in sodium (about 1,150 mg of sodium) was associated with about 1.1 mmHg lower systolic pressure, rising to 2.13 mmHg in trials lasting longer than 15 days (Huang et al. 2020). Bigger cuts sustained for longer do more.

Will cutting out my evening glass of wine lower my blood pressure?

Probably not measurably. In pooled trials, reducing alcohol among people drinking two or fewer drinks a day showed no significant blood pressure effect. The benefit appeared only above two a day and was largest in heavy drinkers halving their intake (Roerecke et al. 2017).

Does handgrip exercise really lower blood pressure?

The evidence is promising and thin. Isometric resistance training produced the largest systolic reduction of any exercise modality at 10.9 mmHg, but from only five study groups compared with 105 for endurance training (Cornelissen and Smart 2013). Treat it as a genuinely interesting signal rather than settled.

How long before I see a change?

Weeks to months, and longer is better. Sodium trials running beyond 15 days showed roughly double the effect of shorter ones. Compare a week of home readings against the previous week rather than day to day, since ordinary variation will swamp a few millimetres.

Knit shows patterns, not diagnoses. Talk to a clinician about readings that concern you.