Blood pressure management

Blood pressure damages the heart, brain, and kidneys continuously — well before it crosses the line most clinics call "high" — so the modern target is lower, around 130/80, and the reading that counts is your home average, not the clinic visit. The cheapest high-value moves are measuring at home, a two-minute isometric wall-squat hold a few times a week, and, if a drug is needed, starting with a single-pill combination built around an angiotensin receptor blocker.

Blood pressure sits among the most heavily evidenced and most underused longevity levers. It is the single most prevalent modifiable risk factor for cardiovascular disease, stroke, chronic kidney disease, and dementia worldwide. The major guidelines updated between 2024 and 2025 codified two changes that together reshape midlife preventive care: a more aggressive treatment target (closer to 130/80 mm Hg for most adults, with consideration of <120 in the highest-risk), and an explicit shift away from one-off clinic measurements toward structured out-of-office (home or ambulatory) monitoring. The pharmacology has also been clarified — single-pill combinations are now first-line, angiotensin receptor blockers (ARBs) are better tolerated than angiotensin-converting-enzyme (ACE) inhibitors though not proven to save more lives, and the decade-long claim that night-time dosing reduces cardiovascular events has been definitively settled in the negative. This article walks the modern framework.

Why blood pressure matters the way it does (Strong)

The relationship between blood pressure and downstream disease is log-linear and continuous. Across a million adults in 61 prospective studies, each 20 mm Hg higher usual systolic pressure was associated with roughly double the death rate from stroke and from ischaemic heart disease in middle age, with no threshold visible down to 115/75 mm Hg — the lowest pressure the data reach.[1] Damage begins well below the traditional hypertension cutoff. The reason is mechanical: every heartbeat propagates a pressure wave through arteries that were designed to be elastic and compliant. Sustained pressure stiffens them, fragments their elastin fibres, replaces them with rigid collagen, and bounces the pressure wave back to the heart earlier in the cycle. The downstream consequences are microvascular damage in the heart (driving heart failure with preserved ejection fraction), the brain (driving small-vessel disease and the cerebral white-matter lesions that precede vascular dementia), and the kidneys (driving chronic kidney disease).

Treating blood pressure isn't just about preventing strokes and heart attacks. It is one of the few dementia risk factors with randomised evidence behind it, and the 2025 US guideline now makes a systolic target below 130 mm Hg a first-class recommendation specifically for preventing mild cognitive impairment and dementia.[2] It isn't the only lever, or necessarily the biggest — the 2024 Lancet Commission on dementia counts fourteen modifiable risk factors, from hearing loss to LDL cholesterol — but it is among the best evidenced.[3] See Dementia prevention.

SPRINT and the case for aggressive targets (Strong)

The SPRINT (Systolic Blood Pressure Intervention Trial) trial randomised 9,361 adults aged 50 and older at high cardiovascular risk (excluding those with diabetes or prior stroke) to either intensive systolic blood pressure (SBP) control aiming below 120 mm Hg, or standard control aiming below 140 mm Hg.[4] Entry required a systolic pressure of 130 mm Hg or higher alongside at least one other cardiovascular risk factor, and diabetes, previous stroke and polycystic kidney disease were all exclusions; the trial was stopped early after a median 3.26 years of follow-up.[5] The intensive-treatment arm produced:

  • 25% lower major cardiovascular events — 1.65% of the intensive arm per year against 2.19% of the standard arm, among SPRINT's high-risk adults aged 50 and over
  • 27% lower all-cause mortality

The adverse-event signals were real and not negligible. Over the trial's 3.26 years, among these high-risk adults aged 50 and over, the intensive arm had roughly 1.8 more episodes of acute kidney injury, 1.4 more of serious hypotension and 1.1 more of syncope per 100 people treated. Set against that, SPRINT's own arithmetic was one primary cardiovascular event prevented for every 61 people treated and one death for every 90 — a favourable trade, but a trade. Orthostatic hypotension, counter-intuitively, was slightly less common in the intensive arm.[6] Injurious falls did not differ between arms even in adults aged 85 and older, dispelling the older worry about aggressive control in the elderly.[7]

One critical post-trial observation: when participants returned to usual care their blood pressure drifted back up — among the 2,944 participants whose routine clinic records were tracked, average outpatient systolic pressure in the intensive arm rose from about 133 mm Hg at five years to about 140 mm Hg at ten — and by a median 8.8 years of total follow-up there was no longer any evidence of benefit for cardiovascular or all-cause mortality.[8] Blood pressure control is not a finite intervention. It requires sustained lifelong management.

Those two exclusions — diabetes and previous stroke — were for a decade the strongest objection to generalising SPRINT. Two large trials have since closed the gap. ESPRIT randomised 11,255 adults at high cardiovascular risk — 4,359 of them with diabetes and 3,022 with a previous stroke — to the same two targets; over a median 3.4 years the composite of heart attack, revascularisation, heart-failure hospitalisation, stroke or cardiovascular death occurred in 9.7% of the intensive arm against 11.1% of the standard arm, with no difference in effect by diabetes status or stroke history.[9] BPROAD ran the same comparison in 12,821 adults with type 2 diabetes and found major cardiovascular events at 1.65 per 100 person-years on the intensive target against 2.09 on the standard one over a median 4.2 years.[10] Both benefits are smaller than SPRINT's — about an eighth and about a fifth fewer events, not a quarter — which is the more realistic expectation to carry, and both trials were run in China. The harms tracked SPRINT's: more syncope in ESPRIT, with no difference in hypotension, electrolyte abnormality, injurious falls or acute kidney injury; more symptomatic low blood pressure and high potassium in BPROAD, with overall serious adverse events similar between the arms.

SPRINT-MIND: cognition and dementia (Moderate)

The SPRINT-MIND sub-study followed the same participants for cognitive outcomes. The 2019 analysis is worth reporting honestly, because it is routinely oversold. Its primary outcome — adjudicated probable dementia — was not significantly reduced, and the investigators attribute that partly to the trial stopping early with fewer dementia cases than it was powered for. What was significant, on secondary outcomes, was a 19% lower rate of mild cognitive impairment in the intensive arm and a 15% lower rate of the combined outcome of mild cognitive impairment or probable dementia.[11] The brain-imaging substudies point the same way, with one caveat worth stating. Intensive control produced a smaller increase in white-matter-lesion volume — the imaging marker of cerebral small-vessel disease — over the substudy's follow-up, though both arms still worsened.[12] It also raised whole-brain blood flow rather than lowering it, the opposite of the perfusion harm clinicians have long feared from pushing systolic pressure below 120 mm Hg.[13] The caveat is that the same volumetric substudy found slightly more total brain-volume loss with intensive treatment — a small difference whose meaning is still unclear.[14] All of this rests on a few hundred scans out of the 9,361 randomised.

Extended follow-up to a median of seven years found the effect persisting but shrinking: mild cognitive impairment about 13% lower in the intensive arm, the combined outcome about 11% lower, and probable dementia alone still not significantly reduced.[15] Three and a quarter years of tight control in late midlife appears to leave a lasting mark on cognitive trajectory — but cognitive status was ascertained in only 4,232 participants (59% of those eligible), no imaging was repeated, and "lasting" is the strongest word the data will carry.

The strongest dementia evidence is no longer SPRINT's. The China Rural Hypertension Control Project cluster-randomised 326 villages — 33,995 adults aged 40 and over with uncontrolled hypertension — to a programme in which trained non-physician community health workers started and titrated medication towards a goal below 130/80 mm Hg, or to usual care. Over four years the programme achieved a systolic pressure 22 mm Hg lower than usual care and about 15% less all-cause dementia, with fewer serious adverse events than usual care, not more.[16] It is open-label, set in a single rural population, and achieved a bigger separation in blood pressure than most real-world programmes manage — but it is the first randomised trial to lower dementia itself, and the 2025 US guideline leans on it alongside SPRINT for its cognitive recommendation.

What the 2024/2025 guidelines actually say (Strong)

The 2025 American Heart Association / American College of Cardiology guideline and the 2024 European Society of Cardiology guideline converged on aggressive targets for most adults.[17][18]

CategoryAHA / ACC 2025ESC 2024
Normal / non-elevated<120/80 mm Hg<120/70 mm Hg ("non-elevated blood pressure")
Elevated120–129 / <80one wide band, 120–139 / 70–89 ("elevated blood pressure")
Stage 1 hypertension130–139 / 80–89no staging — hypertension begins at ≥140/90
Stage 2 hypertension≥140/90—
General treatment target<130/80 (lower if very high risk)120–129 / 70–79, if well tolerated

The two documents land in nearly the same place on where treated blood pressure should end up — the American guideline at below 130/80, systolic below 120 where that is feasible, the European at 120–129 / 70–79 — and disagree mainly about what to call the numbers on the way there. The 2024 European guideline retired the old "high normal" band and the grade 1/2/3 staging of its 2018 predecessor, declined to diagnose hypertension below 140/90, and instead treats people in the upper part of its wide elevated band — systolic 130–139 — once their SCORE2 10-year cardiovascular risk reaches 10%, or sooner in the presence of established cardiovascular disease, diabetes, kidney disease or organ damage.[19]

For its own risk calculation the 2025 US guideline recommends the PREVENT (Predicting Risk of Cardiovascular Disease Events) equations rather than the older Pooled Cohort Equations, which overpredicted risk roughly two-fold when tested against a contemporary sample of 3.3 million US adults.[20] PREVENT estimates 10-year risk of total cardiovascular disease — heart failure included, not just heart attack and stroke — from age, sex, cholesterol, systolic blood pressure, smoking, diabetes, blood-pressure and statin treatment, and kidney function (estimated glomerular filtration rate), with optional refinement from glycated haemoglobin (HbA1c), the urine albumin-to-creatinine ratio, and a Social Deprivation Index that captures neighbourhood-level disadvantage. Body mass index is collected, but it enters only the heart-failure-specific model, not the total-cardiovascular-disease equation the 7.5% threshold is read from, and there is no race term.[21] Drug treatment is recommended for every adult whose average blood pressure is 140/90 mm Hg or higher, whatever the calculated risk. Between 130/80 and 140/90 the guideline splits on risk: start medication immediately alongside lifestyle change if there is established cardiovascular disease, a previous stroke, diabetes, chronic kidney disease, or a 10-year PREVENT risk of 7.5% or more; below that risk level, start medication if the average is still 130/80 mm Hg or higher after a three-to-six-month trial of lifestyle change.[22] The 7.5% cut-off is not arbitrary — the guideline notes it is equivalent to the Framingham 10-year risk of 15% or more that was one of SPRINT's entry criteria.

The J-curve and elderly patients (Moderate)

In some older cohorts, blood pressures below 120 mm Hg associate with higher mortality — the J-shaped curve. The 2024 European guideline is unusually direct about what that means: since 2018, it says, compelling evidence has shown that the J-curve seen in observational datasets is highly unlikely to reflect a causal process, and is better explained by residual confounding and reverse causation — frailty and the illness that accompanies it lower blood pressure while independently raising mortality. Low blood pressure in an older adult marks a high-risk state; it does not create one.[23] The direct trial evidence points the other way for older adults who are otherwise well: among SPRINT participants aged 75 and over, intensive control cut the composite cardiovascular outcome by about a third and all-cause mortality by about a third — 27 people needed treating for three years to prevent one cardiovascular event and 41 to prevent one death, roughly half the numbers needed in the trial as a whole — and the benefit held even in the frailest subgroup and among those with the slowest gait speed.[24][25] The practical implication: pursue the standard target in robust older adults, and relax it deliberately — by symptom, not by birthday — for people with symptomatic orthostatic hypotension, moderate-to-severe frailty, age over 85, or limited life expectancy. The European guideline is explicit that the very old and frail should not be denied the benefits of treatment down to 120–129 / 70–79 mm Hg simply for being old, while noting that the trial participants who support that position probably had no more than mild frailty.[26]

Measurement: the diagnosis runs on out-of-office monitoring (Strong)

A single clinic reading is a poor surrogate for the pressure organs actually live under. Both the 2024 ESC and the 2025 AHA/ACC guidelines build the diagnosis on out-of-office measurement — home blood-pressure monitoring (HBPM) or ambulatory blood-pressure monitoring (ABPM) — rather than on the clinic reading.[27][28] The two are not interchangeable. A 24-hour ambulatory recording is the reference standard, because it is the only method that sees blood pressure during sleep. Home monitoring is the practical instrument: more accessible, and what dose titration is actually built around. Use ambulatory monitoring to settle an ambiguous diagnosis; use home monitoring to run the treatment.

Two clinically critical phenotypes (Moderate)

The mismatch between clinic and home blood pressure produces two patterns that dramatically alter risk:

  • White-coat hypertension — clinic readings elevated (≥130/80), home or ambulatory readings normal, in someone not taking blood-pressure medication. Among adults whose office readings sit in the hypertensive range, 15–30% turn out to have it — a figure measured when the office threshold was 140/90, so it is a rough guide at today's lower cut-off.[29] Once dismissed as benign anxiety, but the risk turns on whether it's treated: untreated white-coat hypertension carries roughly a third higher risk of cardiovascular events and of death from any cause, and perhaps double the cardiovascular mortality, compared with people who are normotensive both in and out of the clinic — whereas a white-coat effect in someone already on treatment is largely benign.[30] The doubling is the least secure of those three figures: its confidence interval — the range of values the data leave plausible — runs from a 23% excess to a more than four-fold one, and a 2024 meta-analysis of eight studies found no significant excess cardiovascular mortality in white-coat hypertension at all.[31] White-coat hypertension is often a staging post rather than a stable state. In an unselected Finnish population sample followed for 11 years, 52% of people with white-coat hypertension went on to develop sustained hypertension, against 18% of those who were normotensive both in and out of the clinic — a reason to keep measuring rather than to discharge them.[32]
  • Masked hypertension — clinic readings normal, home or ambulatory readings elevated. The far more dangerous pattern: roughly 15–30% of adults whose office reading looks normal have it, and the figure goes higher still once night-time readings count towards the definition.[33] People look fine in the clinic and walk around unmedicated with cardiovascular risk roughly equivalent to sustained hypertension — about a doubling of cardiovascular mortality.[34]

Detecting either phenotype requires structured out-of-office measurement. The 2024 ANTI-MASK trial is the only randomised test of actually treating masked hypertension: 320 untreated adults, all with some existing organ damage, took allisartan (an angiotensin receptor blocker), with amlodipine added from month 4 if ambulatory pressure stayed up, or a matching placebo, for 48 weeks. Treatment lowered 24-hour blood pressure by about 9/5 mm Hg more than placebo and improved the organ-damage composite — arterial stiffness, urinary albumin, or left-ventricular thickening — in 52% of those treated against 29% on placebo, roughly one extra person improving for every four or five treated.[35] The endpoint was organ-damage markers over a single year rather than heart attacks or deaths, and the trial was funded by the maker of the drug it tested, which reports no role in its design or analysis. It is a reason to treat masked hypertension, not proof that treating it prevents events.

The home BP protocol that produces usable data

For diagnosis or treatment titration, the protocol the guidelines and the American Heart Association's measurement statement converge on:[36][37]

  • Device. A validated automated upper-arm (brachial) cuff. Wrist and finger monitors are unreliable. Check the exact model against an independent validation register before buying — the US Validated Device Listing run by the American Medical Association and the American Heart Association at [38], or [39] internationally.
  • Conditioning. No caffeine, tobacco, or vigorous exercise for 30 minutes beforehand. Empty bladder.
  • Posture. Seated, back supported, feet flat on the floor, legs uncrossed. Arm supported on a flat surface at heart level. Silent — across the studies pooled in the standard review of measurement error, talking during a reading raised systolic pressure by about 4 to 19 mm Hg and diastolic by about 5 to 14 mm Hg.[40]
  • Rest. Five full minutes of quiet sitting before the first reading.
  • Schedule. Twice daily at the same times — morning, before taking any blood-pressure medication, and evening, before going to bed.
  • Duration. Seven days is the standard; if you discard day 1, plan for eight. Three days — four if you discard day 1 — is the shortest run that yields a usable average, and the European guideline says to extend to the full week whenever the short average lands close to a treatment threshold.
  • Per session. Two readings, 1–2 minutes apart — which over a week of morning and evening sessions gives the 28 readings the protocol is built around.
  • Averaging. Discarding day 1, which often runs high while you get used to the cuff, is common practice rather than a settled rule — the European guideline averages every reading instead. If you do discard it, add a day at the end to make up the count.

Cuffless wearables: not yet (Moderate)

Smartwatches and biosensor patches that estimate blood pressure optically (photoplethysmography plus pulse-transit-time algorithms) are appealing for continuous monitoring — they could capture nocturnal dips and morning surges that even a 7-day home protocol misses. The 2025 AHA/ACC guideline recommends against using them to diagnose or manage high blood pressure until they show greater precision and reliability, and the 2024 ESC guideline states that none of the cuffless modalities is currently recommended for routine clinical use. Validation standards are arriving but have not caught up: the 2022 international standard covers continuous devices built for operating theatres and intensive care, and the protocol that would cover the intermittent wearables people actually buy is still being drafted. Meanwhile most devices infer pressure from a signal that is not pressure and need repeated calibration — some against a cuff, some merely against age, sex and body-mass index — and none of the enormous volume of readings they generate has been linked to clinical outcomes. Regulatory clearance is not validation: formal accuracy testing against an established protocol is not required to obtain it.[41] They are a reasonable trend tool; they are not a diagnostic substitute.

Vascular age: pulse wave velocity and pulse pressure (Moderate)

Brachial blood pressure is a peripheral surrogate for what's actually happening at the central aorta and the microvasculature of the heart, brain, and kidneys. Two metrics provide complementary insight into the underlying biology.

Carotid-femoral pulse wave velocity (cfPWV) measures how fast the pressure wave travels down the aorta — a direct readout of aortic stiffness. Compliant young aortas absorb the systolic ejection (the Windkessel effect) and slow the wave. Stiff aged aortas transmit it fast, and the reflected wave returns to the heart during systole rather than diastole, amplifying central systolic pressure and reducing coronary perfusion. cfPWV predicts cardiovascular events — coronary heart disease and stroke — over and above the conventional risk factors. Pooling individual data on 17,635 people, a one-standard-deviation step up in log aortic pulse wave velocity carried roughly 30% higher risk of a cardiovascular event after adjustment for those risk factors, and adding the measurement improved reclassification of people at intermediate 10-year risk in some subgroups.[42] The signal is strongest in middle age and fades in the over-70s, where almost every aorta is stiff. It beats chronological age as a risk predictor and is the closest routinely obtainable stand-in for how old an artery behaves — but it is a soft measurement, not a lab value. The carotid-femoral route overestimates true aortic stiffness, and simply choosing a different way to estimate the distance between the two pulse points can move a reading by more than 3 m/s, which is why the same review stops short of calling it a clinical standard.[43]

Pulse pressure — the difference between systolic and diastolic (a reading of 120/80 has a pulse pressure of 40) — is a quick, clinical surrogate for the same biology. A widening pulse pressure in midlife and beyond (rising systolic, falling diastolic) is the visible signature of accumulating aortic stiffness. Pooling 14 prospective cohorts covering 510,456 people from the general population, every extra 10 mm Hg of pulse pressure came with about 13% higher cardiovascular mortality and about 9% higher death from any cause.[44] Among older people with hypertension enrolled in treatment trials, the same 10 mm Hg tracked with nearly 20% more cardiovascular deaths even after accounting for mean pressure — the pulsatile part of the reading, not the average, is what carried the risk.[45] Both are observational, so they show association rather than proving that narrowing a pulse pressure would help.

These aren't routine clinical metrics yet, and it is worth being clear why. Pulse wave velocity is a validated risk marker, but only one randomised trial has tested treating hypertension to a stiffness target rather than to a blood-pressure target: SPARTE assigned 536 higher-risk hypertensive patients to stiffness-guided or guideline-directed care and, after a median of four years, found no difference in cardiovascular events that could be distinguished from chance — on 41 events in total, having enrolled roughly a third of the patients it had planned. The result is compatible with a large benefit and with a large harm alike, which is another way of saying it settles nothing.[46] Measure it if you're curious; act on the risk factors either way. The intervention list for stiffening arteries overlaps almost entirely with the rest of this site — aerobic fitness (VO₂ max, zone 2 training), Mediterranean-pattern eating, and the pharmacological levers below.

Sodium: salt sensitivity is a continuum, not an exemption (Moderate)

The classic public-health message — eat less salt — is roughly right at the population level, and it applies to more people personally than the phrase "salt sensitivity" suggests. When 213 adults aged 50–75 spent a week on a high-sodium diet and a week on a low-sodium one with round-the-clock ambulatory monitoring, cutting sodium lowered mean arterial pressure (the average pressure across the whole cardiac cycle) in about three quarters of them, and the size of the fall did not differ meaningfully by whether they had hypertension, or by age, sex, race, diabetes or weight. The median within-person change in mean arterial pressure between the two weeks was 4 mm Hg; median systolic pressure was 126 mm Hg on the high-sodium week and 119 on the low-sodium one. The usual 5 mm Hg cut-off labelled 46% of the group "salt sensitive" — but that is a line drawn through a smooth distribution of responses, not a boundary between two kinds of person. The participants were 65% female and 64% Black, with a median age of 61.[47] Against that continuum, epidemiological estimates still put salt sensitivity of blood pressure at nearly half of people with hypertension and about a quarter of those with normal readings.[48] In salt-sensitive individuals, blood pressure climbs measurably when sodium intake rises, and they appear to progress to sustained hypertension faster than salt-resistant peers — though the evidence for that is thinner than it is usually made to sound. The most-quoted study re-examined just 36 participants of the Italian Olivetti Heart Study after 15 years: 14 of the 16 people in the higher salt-sensitivity group had become hypertensive, against 10 of the 20 in the lower.[49] That is a real signal in a very small sample, not a settled effect size. The phenotype also tracks with non-dipping nocturnal BP, insulin resistance, left-ventricular hypertrophy, and heart failure with preserved ejection fraction.

There is also a mirror-image phenotype — inverse salt sensitivity — in which blood pressure falls on a high-sodium diet and rises when sodium is cut hard. In the one controlled-feeding study to count it, about 11% of 84 healthy, non-obese, normotensive adults averaging 37 years old qualified on the usual threshold of a 5 mm Hg change in 24-hour mean arterial pressure, and about 21% qualified under an alternative definition — so the honest answer is somewhere in the low tens of percent, measured in people younger and healthier than the typical reader here.[50] It is not a reason to salt more: inverse salt sensitivity predicts incident hypertension too.

Identifying salt sensitivity in routine practice is hard — the gold-standard diagnostic protocols (week-long controlled high-sodium and low-sodium phases, or rapid in-patient saline-and-furosemide challenges) are not widely available. The practical workaround the 2025 AHA/ACC guideline endorses:

  • Aim for under 2,300 mg of sodium a day — that is the 2025 AHA/ACC target for every adult, with or without hypertension, and it sits at the guideline's strongest tier of recommendation. Under 1,500 mg is better still, but the guideline frames it as the direction of travel rather than the pass mark.[51] In trials, moving from a high-sodium diet (4,500 mg a day or more) down to a low-sodium one (1,500–2,300 mg) lowered blood pressure by roughly 3/2 mm Hg in people with normal pressure and roughly 7/3 mm Hg in people with hypertension. Older adults and salt-sensitive people respond most.
  • Consider potassium-enriched salt substitutes — table salt with roughly a quarter to a little under a third of the sodium chloride replaced by potassium chloride. The 2025 AHA/ACC guideline rates them as reasonable rather than mandatory (Class 2a, its second-strongest tier) and puts the expected effect at about 5/1.5 mm Hg. They are explicitly not for people with chronic kidney disease, or for anyone taking a drug that reduces potassium excretion such as a potassium-sparing diuretic — there, serum potassium needs monitoring first.[52] The Salt Substitute and Stroke Study randomised 600 rural Chinese villages — 20,995 people, mean age 65, of whom 73% had survived a stroke and 88% were already known to be hypertensive — to a salt that was three-quarters sodium chloride and one-quarter potassium chloride, or to ordinary salt. Over a mean of 4.74 years, strokes ran at about 29 per 1,000 person-years on the substitute against about 34 on regular salt, roughly a 14% relative reduction, with similar reductions in major cardiovascular events and in death from any cause, and no increase in serious hyperkalaemia (dangerously high blood potassium).[53] One caveat sits underneath all of this: a 2024 review of 16 randomised trials found the hard-outcome evidence comes almost entirely from China and Taiwan, mostly in older or higher-risk people, and graded the survival benefit low certainty and the reduction in cardiovascular events very low.[54] Substitutes help most where the salt is added at home; where most sodium already arrives inside packaged and restaurant food, changing what is in the shaker moves a smaller share of the total.
  • The dietary-sodium harm is concentrated in packaged and restaurant foods, not the salt shaker — see Foods to limit.

Pharmacology: four classes, single-pill combinations, ARB-leaning (Strong)

When lifestyle alone can't sustainably keep blood pressure below the target, the modern approach starts with a single-pill combination of two first-line agents from different classes, rather than monotherapy titration. Two lower doses acting by different mechanisms blunt the compensatory responses that a single drug pushed to full dose provokes, and combinations are usually the better-tolerated option: pooling 716 double-blind trials in over 159,000 people — short trials, a mean of nine weeks — several combination regimens caused fewer people to stop treatment for side effects than single drugs did, and an ARB paired with a calcium channel blocker caused fewer discontinuations than placebo.[55] The four first-line classes:

  • Angiotensin receptor blockers (ARBs) — losartan, telmisartan, valsartan, candesartan. Block the angiotensin II receptor.
  • Angiotensin-converting-enzyme inhibitors (ACE inhibitors) — enalapril, lisinopril, ramipril, perindopril. Block conversion of angiotensin I to angiotensin II.
  • Dihydropyridine calcium channel blockers — amlodipine. Direct vasodilator.
  • Thiazide / thiazide-like diuretics — chlorthalidone, indapamide.

ARBs versus ACE inhibitors — the tolerability case is solid, the mortality case is not (Strong for tolerability; Weak for mortality)

For decades, ACE inhibitors and ARBs were treated as clinically interchangeable, with ACE inhibitors preferred on cost. A recent multi-database target trial emulation has complicated that: across UK Biobank and a Chinese hospital-based database (the China Renal Data System), starting on an ACE inhibitor was associated with roughly 12–13% higher all-cause mortality over five years than starting on an ARB. In absolute terms that was a gap of 0.4 percentage points in the community-based UK cohort (3.5% versus 3.0% dead at five years) and 1.5 points in the sicker hospital cohort (16.5% versus 15.0%). The excess in major cardiovascular events was 6% in the UK cohort and 12% in the Chinese one, and when the authors reanalysed the same data with a different confounding adjustment, the UK cardiovascular signal disappeared and the UK mortality signal shrank to the edge of no difference.[56] A separate analysis of the same UK Biobank cohort found ARB users with metabolic dysfunction-associated steatotic liver disease had about 6% lower all-cause mortality than ACE-inhibitor users — an effect whose confidence interval, the range of values the data are compatible with, stopped exactly at no difference — and about 8% fewer cardiovascular events.[57] Because it draws on the same cohort, it is not independent corroboration.

The randomised evidence does not support a mortality gap. ONTARGET randomised 25,620 people with established vascular disease or high-risk diabetes to telmisartan, ramipril, or both. Over a median of nearly five years the primary composite of cardiovascular death, myocardial infarction, stroke and heart-failure hospitalisation occurred in 16.7% on the ARB and 16.5% on the ACE inhibitor — equivalent within the play of chance — while the combination added low blood pressure, fainting and kidney injury without adding benefit.[58] A target trial emulation is an observational design with an RCT's vocabulary; when it disagrees with actual head-to-head randomised evidence, the randomised evidence wins. A Cochrane review of nine head-to-head trials in 11,007 people treated for primary hypertension found no difference between the two classes in deaths from any cause, cardiovascular deaths, or cardiovascular events.[59] Treat the 12–13% mortality figure as hypothesis-generating. The honest caveat runs the other way too: the Cochrane reviewers noted that, unlike ACE inhibitors, ARBs have not been tested against placebo for hypertension outcomes, so their efficacy record rests on head-to-head comparisons rather than on beating no treatment.

What is solid is tolerability. In trials of enalapril, the most-studied ACE inhibitor, about one in nine people developed the bradykinin-mediated dry cough and roughly one in forty stopped the drug because of it, with similar rates across the class.[60] In ONTARGET, which had already screened out during a run-in the people who could not tolerate either drug, cough affected 4.2% on the ACE inhibitor versus 1.1% on the ARB, and the rare-but-serious angioedema 0.3% versus 0.1%.[61] Across the randomised hypertension trials, switching class avoids one withdrawal for side effects for every 55 people treated over about four years — an absolute difference of under two percentage points.[62] That is a good reason to prefer an ARB when starting from scratch, and a good reason not to switch someone off an ACE inhibitor who is doing well. One caveat cuts against reflex switching: in placebo-controlled trials 13.5% of people on an ACE inhibitor reported a cough — but so did 8.5% of people on placebo, so a new cough is worth a look for another cause before the prescription is blamed, even though in hypertension specifically most of it does turn out to be the drug.[63] It is not evidence that the drug class saves lives the other doesn't.

The geroprotective angle (Weak / preliminary)

Angiotensin II, acting at its main vascular (AT1) receptor, doesn't only raise blood pressure; chronic activation drives mTOR (the master anabolic kinase whose chronic activation accelerates aging), upregulates inflammation through the master inflammatory switch NF-κB, and suppresses the catabolic AMP-activated protein kinase (AMPK) — see Hallmarks of aging and Protein. Blocking the system pharmacologically pushes each of those signals the other way, and the resulting molecular picture has been argued to resemble that of caloric restriction. That argument comes from a narrative review by the group that has been developing this hypothesis, which cites its own earlier papers for the caloric-restriction resemblance and is honest about its limits: there is no direct evidence tying renin-angiotensin blockade to these pathways in the way proposed, and the authors call the chain a conjecture.[64] Genetic or pharmacological manipulation of the renin-angiotensin system extends lifespan in nematodes, fruit flies and rodents, according to the same review, which sources the claim to an earlier review rather than to primary data; whether any of it transfers to humans is unproven, but it adds a non-haemodynamic argument for ARBs as a preferred class in the longevity frame.

Chronotherapy: settled in the negative (Strong)

For roughly a decade, the Spanish MAPEC and Hygia Chronotherapy trials drove an international wave of "take your BP medication at bedtime" guidance. Hygia, the larger of the two, reported that bedtime dosing cut major cardiovascular events by about 45% and cardiovascular deaths by more than half. The effect sizes were biologically implausible — far larger than the actual difference in 24-hour blood pressure between the two groups — and the trial has since drawn two formal expressions of concern from the journal that published it, both still standing; it has not been retracted.[65][66] What settled the question was not the controversy but the replication trials it prompted.

Both replication trials — the UK TIME trial in over 21,000 participants and the Canadian BedMed trial in over 3,300 — published between 2022 and 2025, found no difference in major cardiovascular events, heart attacks, strokes, or all-cause mortality between morning and bedtime dosing.[67][68] A 2025 systematic review pooling all five trials and 46,606 patients found the picture only resolves once trial quality is taken into account: across every trial the results disagreed so violently that the pooled estimate was uninterpretable, but among the three trials judged at low risk of bias the answer was a clean null — no reduction in cardiovascular events from evening dosing, with no disagreement between them.[69] The optimal time to take blood-pressure medication is whatever time of day produces the highest adherence — for most people, that's whenever they reliably remember. BedMed also checked the harms that had been alleged against night-time dosing — falls and fractures, new glaucoma, cognitive decline at 18 months — and found none of them, so anyone already dosing at bedtime has no reason to switch.[70]

Lifestyle

Isometric exercise: the underrated finding (Weak / preliminary)

The standard lifestyle interventions for blood pressure are well-known — a DASH-pattern diet (Dietary Approaches to Stop Hypertension), weight loss, alcohol reduction, regular aerobic activity. A 2023 British Journal of Sports Medicine network meta-analysis of 270 randomised trials in over 15,000 participants reshuffled the priority list.[71] Across exercise modalities (aerobic training, dynamic resistance, combined training, high-intensity intervals, isometric training), the largest single effect on resting blood pressure came from isometric exercise — sustained static muscle contraction without movement.

The headline result: isometric exercise reduced systolic blood pressure by 8.2 mm Hg and diastolic by 4.0 mm Hg on average — roughly double the effect of aerobic training, and about what one blood-pressure drug at a standard dose achieves: 6.8 mm Hg systolic for an ACE inhibitor, 7.3 for a thiazide, 8.4 for a calcium channel blocker, 9.3 for a beta-blocker, each measured against placebo.[72] That comparison is not like for like: the drug figures are placebo-subtracted in people with hypertension, while the exercise figure is office pressure against a control group that knew it was not exercising. Read the headline with its foundations in view. Of the 358 separate effects pooled in the analysis, 24 came from isometric trials and only four from wall squats; no trial in the network compared isometric training head-to-head with another mode; and every mode worked substantially better in people who already had high blood pressure.[73] A dedicated pooled analysis of isometric training found the same office-pressure benefit but could not show a change in 24-hour ambulatory pressure — though that rests on only four studies and 147 people, so it is an absence of evidence rather than evidence of absence.[74]

Within isometric protocols, the isometric wall squat scored highest. The validated protocol:

  • Stand with your back against a wall and slide down to a knee angle you can hold for two minutes at an effort of about 4/10. The trials set that angle two ways, as alternatives rather than as one rule: an incremental test stepping from 135° of knee flexion down to 95° to find the position that produces 95% of peak heart rate, or self-selection to a target effort.[75] Pick one angle and keep it for all four sets: the effort is meant to climb as your legs fatigue — roughly 4, 5.5, 7 and 8.5 out of 10 across the four bouts — not to start at 8.[76]
  • Hold the position for 2 minutes.
  • Rest for 1–4 minutes.
  • Repeat for 4 sets per session.
  • 3 sessions per week.

The proposed mechanism: the brief local compression of the vessels during the held contraction is followed by a rebound surge of flow, and the shear stress that surge puts on the vessel wall is thought to drive nitric-oxide-mediated improvement in the endothelium’s capacity to dilate.[77] That account is plausible and widely repeated, but what has been measured is largely local, mostly in conduit vessels such as the brachial artery, and the chain from held contraction to lower resting pressure has not been demonstrated end to end in humans. The 2024 European guideline does now mention isometric training, but as a complement to aerobic exercise rather than a peer of it: it names aerobic exercise first-line, calls the isometric evidence inconsistent, and prescribes three sets of 1–2 minutes rather than the four-by-two-minute protocol above.[78] Whether adding it to aerobic training is additive has not been tested directly.

The foundations (Strong)

Other foundational interventions remain critical:

  • Mediterranean / DASH-pattern eating (see Dietary patterns). In the original trial the DASH combination diet lowered systolic pressure by 11.4 mm Hg among the 133 participants who were already hypertensive, and by 3.5 mm Hg among the 326 who were not.[79] Note the conditions: every meal was provided, and sodium intake and body weight were held constant, so this is the effect of the food pattern by itself, eaten properly, not the average result of trying to eat more like it. Entry required a systolic pressure under 160 mm Hg, so the hypertensive group was a mild one.
  • Aerobic and resistance exercise (VO₂ max, Zone 2, Resistance training).
  • Dietary fibre — worth eating for other reasons, but a much smaller blood-pressure lever than the rest of this list. Pooled across 24 placebo-controlled trials of fibre supplements, adding about 11.5 g/day moved systolic pressure by roughly 1 mm Hg, and the range of values consistent with the data still included no effect at all; the diastolic fall of about 1.3 mm Hg was clearer, and the effect tended to be larger in older and hypertensive people.[80] The short-chain-fatty-acid story — gut microbes fermenting fibre into acetate and butyrate — has one small randomised human test, in 20 untreated hypertensive adults given a purpose-built acetylated and butyrylated starch for three weeks, which did lower 24-hour systolic pressure.[81] Promising lead, not a reason to expect a blood-pressure return on eating more bran — though the pooled trials' own conclusion is that raising fibre intake in Western diets may still help prevent hypertension. For intake targets see Fibre.
  • Weight reduction in overweight adults — about 1 mm Hg of systolic reduction per kilogram lost, pooled across 25 randomised trials in 4,874 participants.[82] Those trials averaged only about 5 kg of net loss, and the relationship is not linear beyond it: in the trials whose participants lost more than 5 kg the systolic fall averaged 6.6 mm Hg in total, well under a millimetre per kilogram. Trust the rule of thumb over a few kilograms, and do not scale it up to the much larger losses now achievable on incretin drugs.
  • Alcohol restriction. See Alcohol — minimum-risk dose is near zero, and the blood-pressure relationship is linear with no safe floor: across seven cohorts and about 19,500 people, one standard drink a day tracked with systolic pressure roughly 1.3 mm Hg higher than not drinking, four drinks a day with about 4.9 mm Hg higher; the systolic pattern was linear in both men and women, though the female diastolic curve was not.[83] These are observational cohorts rather than trials, so the numbers describe what drinkers’ pressures did over about five years, not what quitting would do to yours.
  • Sleep quality, particularly screening and treating obstructive sleep apnea, which accompanies 70–85% of resistant hypertension.[84] Treating it helps without resolving it: across 12 randomised trials, continuous positive airway pressure lowered 24-hour systolic pressure by about 6 mm Hg in this group, with the effect concentrated at night.[85]
  • Stress management (Stress).

Practical guidance

  1. Know your number — at home, not just at the clinic. Get a validated upper-arm cuff. Run the 7-day morning-and-evening protocol and average the readings (if you drop day 1, add a day at the end). The home average is what you act on.
  2. Treat target organs, not just numbers. For most healthy midlife adults, aim for a home-average <130/80 mm Hg. Consider <120 systolic if you're at high cardiovascular risk and tolerate it without orthostatic symptoms. Relax them deliberately — by symptom, not by age — for symptomatic orthostatic hypotension, moderate-to-severe frailty, age over 85, or limited life expectancy.
  3. Screen for masked hypertension. If you have any risk factor (family history, metabolic syndrome, sleep apnea, kidney disease) and your clinic BP runs at the upper end of normal, run a 7-day home protocol. The diagnostic yield is high enough to justify the effort.
  4. If you need medication, ask about a single-pill combination first. ARB + amlodipine or ARB + thiazide are common starting combinations and produce faster control with better adherence than titrating monotherapy.
  5. Take it whenever you reliably remember. The chronotherapy claim is dead; adherence is what matters.
  6. Add the wall squat protocol. Four 2-minute holds, three times a week, zero equipment. Pooled across isometric trials the average systolic fall was about 8 mm Hg — but that pools 19 trials, the wall-squat ranking rests on four, and the effect was larger in people who already had high blood pressure than in those who did not.[86] Treat it as a free addition to aerobic exercise, not a replacement for anything on this list.
  7. Consider potassium-enriched salt substitutes if you have hypertension or are at risk — a Class 2a ("reasonable") 2025 recommendation, and not for anyone with chronic kidney disease or on a potassium-sparing drug without potassium monitoring. Most of your sodium comes from processed and restaurant food anyway; cut packaged food first.
  8. Screen for obstructive sleep apnea if your hypertension is resistant to two or three drugs, you snore, you have atrial fibrillation, or you’re overweight. See Sleep-disordered breathing.
  9. If your pressure is resistant, or stage 2, ask to be screened for primary aldosteronism. An adrenal gland producing too much aldosterone causes roughly 5–10% of all hypertension and up to 20% of resistant hypertension, and it is treatable — with a mineralocorticoid receptor antagonist such as spironolactone, or with surgery when one gland is responsible. The 2025 AHA/ACC guideline widened the recommendation to everyone with resistant or stage 2 hypertension whether or not their potassium is low, because low potassium is absent in most cases and waiting for it is why the diagnosis gets missed: fewer than 2% of eligible patients are currently screened. If you are already taking spironolactone or eplerenone, they have to be stopped for at least four weeks before the test.[87]
  10. Treat aggressively in midlife specifically. The cognitive-protection signal from SPRINT-MIND persists, though it shrinks with time, and it depends on control starting well before symptoms appear.

What's overrated

  • Cuffless smartwatch BP readings as diagnostic. Useful for trends; not a substitute for a validated cuff. The 2025 American guideline says reliance on them should be avoided until the devices demonstrate greater precision and reliability, and the 2024 European one says none of these modalities is currently recommended for routine clinical use.[88][89] The reason is one the validation protocols expose: a cuffless device has to be calibrated against a real cuff on the individual wearer and then keep tracking that person’s changes as the calibration drifts, so the procedures written for cuffs do not test what actually matters.[90]
  • Bedtime dosing of antihypertensives as a survival strategy. Settled by TIME and BedMed. Whatever time gets you to take the pill.
  • One-off clinic readings as the diagnostic basis. Misses both white-coat and masked hypertension. People with masked hypertension sit only about 6 mm Hg above normal in the clinic but nearly 19 mm Hg above it outside it, which is why a single office reading cannot find them, and their risk of a cardiovascular event runs about 70% higher across fifteen studies, with cardiovascular mortality roughly doubled on a smaller body of three.[91]
  • The idea that sodium restriction only matters if you happen to be “salt-sensitive.” The crossover trial described above lowered mean arterial pressure in about three quarters of participants, and the fall was much the same whether or not they had hypertension, and across age, sex, race, diabetes and weight.[92] The conventional cut-off labelled only 46% “salt sensitive,” and earning that label took a two-week diet protocol with round-the-clock ambulatory monitoring, so it is not something you can establish about yourself in a clinic. A target under 2,300 mg of sodium a day plus a potassium-enriched salt substitute where suitable remains the right practical goal. What is overrated is the reassurance that you are personally exempt.
  • Switching a well-controlled patient from an ARB to an ACE inhibitor on cost alone. Not because ARBs are proven to save more lives — ONTARGET says they don't — but because the cough and angioedema risk buys you a real chance of a discontinuation you didn't need.

Further reading

  • Wright JT Jr et al. (SPRINT Research Group). A randomized trial of intensive versus standard blood-pressure control. NEJM 2015.[93] Trial overview: [94] Long-term follow-up: Jaeger BC et al., JAMA Cardiol 2022.[95]
  • Williamson JD et al. Effect of intensive vs standard blood pressure control on probable dementia — SPRINT-MIND. JAMA 2019.[96] 7-year follow-up: Reboussin DM et al., Neurology 2025.[97]
  • Cohen JB et al. Cardiovascular events and mortality in white coat hypertension — a systematic review and meta-analysis. Ann Intern Med 2019.[98] Masked hypertension: Tewari J et al., Egypt Heart J 2024.[99]
  • 2025 AHA/ACC guideline for prevention, detection, evaluation and management of high blood pressure in adults.[100]
  • McEvoy JW et al. 2024 ESC Guidelines for the management of elevated blood pressure and hypertension. Eur Heart J 2024.[101]
  • Bachlitzanaki M et al. Salt sensitivity of blood pressure — a cardiorenal phenotype. Life (Basel) 2026.[102]
  • Comparative effectiveness of ACE inhibitors versus ARBs — multidatabase target-trial emulation. 2025.[103]
  • Yusuf S et al. (ONTARGET Investigators). Telmisartan, ramipril, or both in patients at high risk for vascular events. NEJM 2008.[104]
  • Li EC et al. Angiotensin converting enzyme (ACE) inhibitors versus angiotensin receptor blockers for primary hypertension. Cochrane Database Syst Rev 2014.[105]
  • de Cavanagh EMV et al., Physiol Rep 2024. Renin-angiotensin system inhibitors positively impact on multiple aging regulatory pathways: could they be used to protect against human aging?[106]
  • Mackenzie IS et al. Cardiovascular outcomes in adults with hypertension with evening versus morning dosing of usual antihypertensives in the UK (TIME study). Lancet 2022.[107]
  • Garrison SR et al. Antihypertensive Medication Timing and Cardiovascular Events and Death: The BedMed Randomized Clinical Trial. JAMA 2025.[108]
  • Li Y et al. Controversies in hypertension therapy: bedtime dosing or daytime dosing? J Hypertens 2025.[109]
  • Edwards JJ et al. Exercise training and resting blood pressure — pairwise and network meta-analysis of 270 RCTs. Br J Sports Med 2023.[110]

— § —