Until this chapter, this guide has been about one axis of cardiovascular risk: the particles. Blood pressure is the other, it is roughly as important, and it was the largest gap on this site.
It is also the one where the gap between what is measured and what is true is widest — wider even than the lipid problem in chapter 2. A blood pressure is a single instantaneous reading of a quantity that changes with every breath, every thought and every position of your arm. Nearly everything that goes wrong with it goes wrong before anyone interprets the number.
Why it matters as much as it does
Raised blood pressure is common, almost entirely symptomless until something breaks, and cheap to treat. It is a leading contributor to stroke, heart failure, kidney disease and dementia, and the relationship with risk is continuous — there is no threshold below which pressure stops mattering and above which it starts. The categories below are administrative conveniences drawn across a smooth curve, which is worth remembering when a reading lands one unit either side of one.
Two things make it different from cholesterol. It responds faster — weeks rather than months, and to things you can do yourself. And it can be measured at home, accurately, for the price of a takeaway, which is not true of anything else in this guide.
1. Getting the measurement right
The ARMS trial randomised 133 adults across three arm positions in a crossover design, so each person acted as their own control. Against the guideline standard — arm supported on a desk, mid-cuff at heart level — the results were:
| Arm position | Systolic error | Diastolic error |
|---|---|---|
| Supported on a desk at heart height | reference | reference |
| Hand resting in the lap | +3.9 mmHg | +4.0 mmHg |
| Arm unsupported at the side | +6.5 mmHg | +4.4 mmHg |
| Unsupported, in people already hypertensive | about +9 mmHg | — |
Read that against the thresholds in the next section. An unsupported arm is the difference between a systolic of 133 and one of 140. It is not a rounding error; it is a diagnosis.
The other avoidable errors, in rough order of size:
- A cuff that is too small reads high, sometimes substantially. This is the commonest error in larger arms and the easiest to fix — the bladder should wrap around about 80% of the arm.
- Talking during the measurement raises it, as does having just walked in, just drunk coffee, or needing the lavatory.
- Crossed legs and an unsupported back both raise it.
- Cuff over clothing, or a rolled-up sleeve tight enough to constrict.
- One reading only. Blood pressure falls over successive measurements in the same sitting. Take two or three a minute apart and use the average of the last two.
The same discipline as chapter 2, for the same reason. Blood pressure has large within-person variation — larger, in percentage terms, than LDL-C. A single reading near a threshold tells you very little. This is not an argument for ignoring it; it is an argument for measuring it more than once, properly, before anything changes.
2. What the numbers mean in the UK
What the two numbers are, since this is rarely said. The top number is the pressure while the heart is squeezing; the bottom number is the pressure while it relaxes between beats. Either one being high is enough to matter — you do not need both, and a reading of 150/78 is hypertension just as 132/95 is.
And “normal” covers a wide range that is not all equally good. The staging table below calls anything under 140/90 normal, because that is the treatment threshold — but risk rises continuously below it, and 105/65 and 138/88 are not the same situation:
| Clinic reading | Where it sits |
|---|---|
| Below 120 / 80 | Ideal. Risk from blood pressure is at its lowest here |
| 120–139 or 80–89 | Raised, but not hypertension. Not a diagnosis and not nothing. Whether it is worth treating depends on your overall cardiovascular risk rather than the number alone, and it is the band in which lifestyle change does most |
| 140 / 90 or above | Hypertension, once confirmed away from the clinic. Staged below |
| Below 90 systolic | Low, and only matters if it causes symptoms or is a drop from your own usual — see fainting |
Which guideline you are being quoted changes the label but not the threshold. The 2024 European guidelines simplified everything into three bands — non-elevated below 120/70, elevated at 120–139 or 70–89, and hypertension at 140/90 or above — and abandoned the stages entirely. That reclassifies a great many people: someone at 118/72 was “optimal” before and is “elevated” now, on the strength of a diastolic reading in the seventies. The European treatment target also came down to a systolic of 120–129 where tolerated. UK practice follows NICE, which keeps 140/90 and the staging below — so a European table and a UK table can describe the same person differently without either being wrong, and the diagnostic threshold is the same in both. Contested on where the lower boundary of “raised” should sit. If you are in the UK, NICE is the one that operates: it is what your GP is working to, and the table below is the one your diagnosis and treatment will follow.
NICE (NG136) does something unusual and sensible: a raised clinic reading is not a diagnosis. It is a trigger for measuring properly, away from the clinic, over 24 hours or a week. Ambulatory monitoring is the reference standard, with home monitoring as the alternative where ambulatory is not suitable.
| Stage | Clinic | Confirmed by ABPM or home average |
|---|---|---|
| Normal | below 140/90 | below 135/85 |
| Stage 1 | 140/90 to 159/99 | 135/85 to 149/94 |
| Stage 2 | 160/100 to under 180/120 | 150/95 or above |
| Stage 3 / severe | 180 systolic or 120 diastolic, or higher | — needs same-day assessment |
Notice that the home and ambulatory numbers are lower than the clinic numbers for the same stage. That is not a stricter standard; it is a correction for the fact that clinic readings run high. Comparing a home reading against a clinic threshold will reassure you wrongly.
White coat and its mirror image
White-coat hypertension — high in clinic, normal elsewhere — is common and is the reason NICE asks for confirmation before diagnosing. Masked hypertension is the reverse and the more dangerous: normal in clinic, high in ordinary life. It is invisible to exactly the process most people rely on, which is another argument for owning a monitor.
Treatment targets, once you are being treated
| Group | Clinic target | Home or ambulatory target |
|---|---|---|
| Under 80 | below 140/90 | below 135/85 |
| 80 and over | below 150/90 | below 145/85 |
| CKD with diabetes, or ACR 70 mg/mmol or more | below 130/80 | below 125/75 |
NICE adds an instruction that matters more than the numbers: use clinical judgement in people who are frail or have several conditions at once. A target that causes falls has not helped anyone.
3. The disagreement worth knowing about
American guidance treats 130/80 as the threshold for stage 1 hypertension; NICE keeps 140/90 for diagnosis and treats the decision to medicate at stage 1 as depending on cardiovascular risk. Both groups have read the same trials.
The trial at the centre of it is SPRINT, which found that pushing systolic toward 120 rather than 140 reduced cardiovascular events and deaths. NICE looked at the same evidence and declined to adopt the lower target, on the grounds that the benefit came with more harm elsewhere — syncope, hypotension, acute kidney injury and electrolyte abnormalities — and that SPRINT measured blood pressure in a way that reads several mmHg lower than routine practice — so a SPRINT target of 120 is not the same number as a 120 taken in a GP surgery.
This is not one side being behind the other. It is a genuine disagreement about how to weigh a real benefit against real harms in an older, frailer population than the trial recruited.
4. What actually lowers it
Roughly in order of how much they move the number, for someone who has room to move it. These are population averages; individual response varies a great deal, which is why measuring at home before and after a change is worth more than any table.
- Weight loss, where there is weight to lose. Among the largest effects available without a prescription, and it improves everything else in this guide at the same time.
- Salt reduction. Most UK salt intake is not from the salt cellar — it is in bread, processed meat, sauces and ready meals, which is why reading labels beats moving the shaker. Response varies: some people are markedly salt-sensitive and some barely at all, and there is no easy test for which you are except trying it.
- More potassium, from vegetables, fruit, beans and potatoes. Works partly by offsetting sodium. Not from supplements or potassium-based salt substitutes without medical advice if you have kidney disease or take an ACE inhibitor, ARB, spironolactone or another potassium-sparing diuretic, or trimethoprim — that combination can raise potassium dangerously.
- Regular aerobic exercise, and isometric work such as wall sits and static holds, which have a surprisingly good evidence base for blood pressure specifically. See the exercise chapter.
- Less alcohol. Dose-dependent and reversible, and one of the faster responders — a few weeks of reduction shows up on a home monitor.
- Sleep, and specifically sleep apnoea. Blood pressure that resists three drugs should prompt a question about snoring and daytime sleepiness. See the sleep chapter.
5. When three drugs are not enough — the question that gets skipped
Blood pressure that stays above target on three drugs at proper doses, one of which is a diuretic, is resistant hypertension. The standard next step is to add a fourth, and NICE sets out how: check potassium, and add spironolactone if it is 4.5 mmol/L or below.
That works, and it skips a question worth asking first.
Two things keep it hidden.
The first is a piece of teaching that turns out to be wrong. Generations were taught to suspect it when blood pressure is high and potassium is low. In practice only somewhere between 9% and 37% of confirmed cases have low potassium, so using hypokalaemia as the trigger misses most of them. A normal potassium is not reassurance.
The second is that it is easier to treat the phenotype than to diagnose the cause. Adding spironolactone — an aldosterone antagonist — will often bring the pressure down whether or not anyone has established why it was high. The number improves and the question closes. What that misses is the subset in whom the excess comes from one adrenal gland, where surgery can be curative rather than merely controlling.
The screening test is the aldosterone-to-renin ratio, a blood test, with confirmation and imaging afterwards if it is positive. The order matters: the ratio should be measured before starting spironolactone, because a mineralocorticoid receptor antagonist suppresses the very ratio you would be screening with. Starting the drug first does not just delay the diagnosis; it obscures it. Guidance has moved: the 2024 ESC hypertension guideline suggests screening all adults with confirmed hypertension using the ratio (Class IIa, Level B), and the 2025 Endocrine Society guideline now makes the same suggestion as a conditional recommendation. NICE has not gone that far, and its step-4 potassium check is a treatment algorithm rather than a diagnostic one. Moderate
Why this sits on a prevention site. Almost nothing else here is potentially curable. Untreated aldosterone excess also does more cardiovascular and kidney damage than the same blood pressure from ordinary hypertension, so the diagnosis changes more than the drug choice. And the failure is not subtle: audits repeatedly find that most patients who meet existing screening criteria are never tested, and that a third or more of those who screen positive get no follow-up. If your pressure needs three drugs, asking whether this has been excluded is a reasonable question, not an exotic one.
The rest of the secondary causes are worth a thought at the same time, particularly in resistant or young-onset hypertension: obstructive sleep apnoea, which has its own section in the sleep chapter and is another classic cause of resistance; kidney disease; thyroid disease; and the everyday drug causes — NSAIDs, decongestants, some antidepressants, steroids, the combined oral contraceptive and liquorice.
6. When lifestyle is not enough
If lifestyle is not enough, the drugs are cheap, well understood and generally well tolerated, and NICE sets out a clear sequence based on age and ethnicity. The point of this chapter is not to argue against them. It is that the decision to start should rest on a properly measured number rather than a badly measured one.
What to do with this
Buy a validated upper-arm monitor
Upper arm, not wrist. Validated — the British and Irish Hypertension Society publishes a list, and unvalidated devices are common and can be badly wrong. Around £25–40.
Measure properly for a week, not once
Twice in the morning and twice in the evening, a minute apart, seated and supported, for seven days. Discard day one, average the rest. That number is worth more than any single reading taken anywhere.
Take the average to your GP, not the worst reading
A week of home readings is close to what NICE asks for diagnostically, and it removes both the white-coat effect and the argument about it.
There is a calculator for the measurement side of this. Measuring blood pressure properly covers position, cuff and timing, works out your inter-arm difference, and calculates an ankle-brachial index if you have ankle readings. And is your treatment working? takes a week of readings against your target and tells you what to ask.
7. The drugs, and the side effects that get blamed on getting older
Section 6 says the drugs are well tolerated, and they are. But the side effects that do occur have a particular character: they arrive slowly, and they look exactly like ageing. A bit more tired. A bit less steady standing up. A cough that will not clear. Swollen ankles. Each is easy to notice and quietly file under "my age" — which is how a fixable medication effect becomes a permanent feature of someone's life.
Which drug you are on is not arbitrary
NICE sets a sequence, and the starting point depends on age and family origin — which surprises people who assume the choice is a matter of preference.
| Step | What is offered |
|---|---|
| 1 | Under 55 and not of Black African or African-Caribbean family origin, or type 2 diabetes at any age: an ACE inhibitor or ARB — the "A" drugs. 55 or over, or of Black African or African-Caribbean family origin at any age: a calcium channel blocker — "C". If that is not tolerated, usually because of ankle swelling, a thiazide-like diuretic — "D" |
| 2 | A + C, A + D, or C + D |
| 3 | A + C + D |
| 4 | Resistant hypertension. Spironolactone if potassium is 4.5 or below, otherwise an alpha or beta blocker |
- Beta blockers are not routine first-line treatment for blood pressure in the UK, which often surprises people taking one. They appear at step 4, or earlier where there is a separate reason — angina, a previous heart attack, heart failure, rate control in atrial fibrillation. If you are on one, there is usually a reason beyond the blood pressure, and it is worth knowing what it is.
- An ARB is preferred to an ACE inhibitor in people of Black African or African-Caribbean family origin, partly because of a higher risk of angioedema.
- An ACE inhibitor and an ARB are not combined. They act on the same pathway, and the combination adds harm without adding benefit.
- ACE inhibitors and ARBs must be stopped in pregnancy and avoided by anyone planning one. This is an MHRA safety position, not a precaution.
The six worth recognising
| What you notice | What is going on |
|---|---|
| Dizzy or unsteady on standing any BP drug |
Standing normally triggers a reflex — the heart speeds up, vessels tighten — within a second or two. Age blunts it, and blood pressure drugs work by the very mechanisms it needs. The result is orthostatic hypotension, affecting roughly one in five independently living adults over 60 and around half of care home residents. Most never mention it; they just learn to pause before walking. Confirmed with a lying-and-standing blood pressure, which takes two minutes. The answer is often a timing change or a smaller dose, not a different drug — the dose that suited someone at 62 may not at 72 |
| A dry, tickly cough ACE inhibitors only |
Not productive, not a cold, lasting weeks or months. People get chest X-rays and inhalers while the cough continues. ACE also breaks down bradykinin; blocking it lets bradykinin accumulate in the airway lining. Estimates range from about 1 in 20 to 1 in 3, most often quoted around 1 in 10, more common in women, starting anywhere from days to months in. ARBs do not raise bradykinin and very rarely cause it, so the switch is straightforward — and NICE names cough explicitly as the reason to make it |
| Swollen lips, eyelids or face ACE inhibitors |
The same bradykinin mechanism, presenting as angioedema instead of cough. It looks allergic, so people take antihistamines and hunt for a food or a washing powder — and antihistamines do not help much, because this is bradykinin rather than histamine. It can appear after months or years of uneventful treatment, which is why it is so often missed — and in June 2026 the MHRA required the product information for all ACE inhibitors to be updated to strengthen the warning about delayed-onset angioedema, precisely because of that. A Yellow Card review to 10 June 2026 found that roughly half of cases with a dated onset began 30 days or more after starting the drug — against the 20–30% that manufacturer product information had been describing. If it is suspected, the drug is stopped and not restarted. The mechanism matters clinically: bradykinin-mediated angioedema does not respond to standard anaphylaxis treatment, so if adrenaline and antihistamines are not working, that is itself a clue rather than a reason to give more. More common in people of Black African or African-Caribbean family origin, and the reason NICE prefers an ARB in that group. Swelling of the tongue or throat, or any difficulty breathing or swallowing, is a 999 call — this is the one airway-capable condition on this page, and it can progress. See anaphylaxis and airway swelling. Lip or facial swelling alone is not usually an emergency but still needs reporting promptly so the drug can be reviewed |
| Swollen ankles and feet amlodipine and relatives |
Amlodipine widens arteries more than veins, so capillary pressure in the legs rises and fluid is pushed into the tissue — a tap opened wider than the drain. It is local and mechanical, not fluid retention, which is the distinction that matters. Dose-related: roughly 1 in 50 at 2.5 mg rising to about 1 in 10 at 10 mg, about three times placebo, more common in women, worse by evening and better after lying flat overnight. A class effect shared with nifedipine and felodipine |
| Everything in second gear beta blockers |
They cap how hard and fast the heart responds to demand, and the heart cannot distinguish stairs from television — so it responds less to everything. Some people also report vivid dreams or unrefreshing sleep. The evidence is genuinely mixed: several studies find no meaningful loss of exercise capacity, others report fatigue in up to 3 in 10. Read that upper figure carefully — it is the proportion reporting fatigue, not the proportion caused by the drug, and the placebo-controlled estimate of excess fatigue is far smaller, in the region of 18 additional cases per 1,000 patient-years. What is consistent is that response varies between individuals and between drugs in the same class, so intolerance of one does not predict intolerance of another |
| Things nobody volunteers thiazides, beta blockers |
Erectile dysfunction is associated with both classes, is very commonly put down to age, and is very rarely raised in the appointment. It is worth raising, because it often has a fixable cause. Thiazide-like diuretics also cause gout, low sodium and low potassium; spironolactone at step 4 causes high potassium and breast tenderness or enlargement in men |
Two things the blood tests are actually for — and the combination that causes most harm
- ACE inhibitors and ARBs change how the kidney handles pressure, sodium and potassium. That is often protective, particularly with diabetes or kidney disease. It is also why kidney function and potassium are checked within one to two weeks of starting or increasing the dose, and at least yearly after. The blood test is not administrative — it is looking for a rising creatinine and a rising potassium, both of which are silent.
- The "triple whammy" is a well-recognised and largely preventable cause of drug-induced kidney injury in general practice: an ACE inhibitor or ARB, plus a diuretic, plus an NSAID — ibuprofen or naproxen, very often bought over the counter for a bad back and never mentioned. Each is tolerable alone. Together they remove the kidney's three main ways of protecting its own filtration pressure at once. If you take the first two, ask before taking the third, and say so if you already are.
- Sick day guidance. ACE inhibitors, ARBs and diuretics protect the kidney at normal fluid volume but can injure it during dehydration from vomiting, diarrhoea or fever — so prescribers sometimes advise pausing them during such an illness and restarting once eating and drinking normally again. Whether that applies to you is an individual decision, not a rule: in the UK this guidance is given case by case, and for some people — above all anyone taking a diuretic for heart failure — stopping can cause serious harm of its own. The useful action is to ask while you are well: ask your prescriber or pharmacist whether sick-day guidance applies to your medicines, and what your plan should be. And if you cannot keep fluids down for more than about 24 hours, that is a reason to seek medical advice in itself.
- And the pairing to question: a diuretic started for ankle swelling caused by amlodipine. The swelling is not fluid overload, so the diuretic does not address it — it just adds a second drug, and in older adults that combination is linked to more falls. Ask whether the amlodipine could be the cause before accepting a water tablet for it; a dose reduction, a switch, or adding an ACE inhibitor or ARB — which offsets this specific effect — is usually the cleaner answer.
What exists elsewhere — the first new mechanism for resistant hypertension in decades
What it showed, and the three things to hold alongside it
- BaxHTN, 796 people, published in the New England Journal of Medicine. On top of standard care — two drugs including a diuretic for uncontrolled hypertension, three or more for resistant — the placebo-adjusted seated systolic fall at 12 weeks was 9.8 mmHg at 2 mg and 8.7 mmHg at 1 mg. In resistant hypertension specifically, 24-hour ambulatory monitoring showed around 14 mmHg. Those are large numbers for an add-on in people already on three drugs. The chief investigator, Bryan Williams, is at University College London.
- It is approved on blood pressure, which is a surrogate. No cardiovascular outcome trial has reported. Blood pressure is a better-validated surrogate than most — this site says so elsewhere — but the same discipline applies as to the LDL agents: lowering the number is the beginning of the case, not the whole of it.
- Hyperkalaemia is the risk that matters, reported in up to about one in ten at the higher dose, along with low sodium. That is the same problem spironolactone has, from a drug acting one step earlier in the same pathway.
- It has not been compared with spironolactone. There is no head-to-head trial — and BaxHTN specifically excluded people already taking a mineralocorticoid receptor antagonist or potassium-sparing diuretic, so it cannot tell you how baxdrostat compares with, or adds to, spironolactone. The honest position is that spironolactone remains the UK step-4 answer
And the practical consequence of that hyperkalaemia figure: potassium and sodium are checked before starting and periodically during treatment. That is on the label, and it is the reason this is a monitored drug rather than a prescription you collect and forget. Hyponatraemia is a labelled warning too.
Where it stands in the UK, as of August 2026. Baxdrostat has no MHRA licence; the last public statement on that was a parliamentary answer in October 2025, and NICE has an appraisal in development. So this is a drug to know about rather than one to ask for. The 24-hour ambulatory figure quoted above comes from Bax24, which reported a placebo-adjusted reduction of about 14 mmHg in confirmed resistant hypertension — a large number, still on a surrogate, still with no outcome trial.
where potassium allows, and nothing about this approval changes that. Lorundrostat, a second drug in the same class, is under FDA review.
Why it is worth knowing about here. Resistant hypertension is the situation in this chapter with the fewest options, and this is the first genuinely new mechanism aimed at it in a long time. If you are on three drugs and still above target, the useful actions remain the ones in the panel above — confirm the readings, confirm adherence, look for a secondary cause, and consider spironolactone. But it is reasonable to know that the pipeline behind those options is no longer empty.
The habit worth building. When something new appears — the cough, the tiredness, the unsteadiness, the swelling — ask when before asking what. If it began within weeks of a new tablet or a dose change, say so out loud at the appointment, because the person prescribing it may not have the timeline in front of them. Sometimes it lines up and sometimes it does not, but it is a question worth asking rather than a conclusion worth assuming.