Most of the error in a blood pressure reading is put there before anyone looks at the number. This is the technique half. What the numbers mean is in vital signs, and what counts as healthy against high is in the blood pressure chapter.
1. Blood pressure — the setup that decides the reading
A blood pressure taken badly is not a slightly worse blood pressure. It is a different number, reliably wrong in a predictable direction, and usually wrong by more than the width of the category it is being compared against.
| Do this | Because |
|---|---|
| Seated, back supported, feet flat on the floor, legs uncrossed | An unsupported back and crossed legs each add several mmHg |
| Arm supported on a table, cuff at heart height | The largest single positional error. An arm hanging at the side adds around 6.5 mmHg systolic, and about 9 in someone already hypertensive |
| Cuff on bare skin, not over a sleeve | And do not roll a sleeve up into a tight band above the cuff, which is worse than leaving it down |
| Silence for a minute before, and during | Talking raises it measurably. So does having just walked in, just had coffee, or needing the lavatory |
| Two or three readings a minute apart, average the last two | Blood pressure falls across successive readings in the same sitting. A single first reading is the highest one you will get |
Cuff size, which is where most of the remaining error lives
The inflatable bladder inside the cuff should be about 40% of the arm's circumference in width and 80% or more of it in length. Measure the arm at the midpoint between shoulder tip and elbow if there is any doubt.
- Too small reads high — by 10 to 15 mmHg commonly, and up to about 20. This is the error that matters, because it produces diagnoses of hypertension in people who do not have it, and it falls disproportionately on larger arms.
- Too large reads low, by a few mmHg, which hides hypertension.
- On a boundary between two sizes, go up.
- Children need paediatric cuffs, and cuff choice must be by measured arm circumference rather than by age or appearance. This is the commonest reason a child's blood pressure reading is not usable.
Children specifically
- A child's blood pressure has no single normal range. It is interpreted against age, sex and height centiles, which is why the bands on the vital signs page are approximate and why an isolated raised reading in a child is a reason to repeat rather than to act.
- It is not routinely measured in a well child and becomes important when they are acutely unwell — in which case do it, but do not delay treatment to get it.
- A falling blood pressure in a child is very late. They compensate hard and hold their pressure until they are close to collapse. Rising pulse, poor perfusion and altered behaviour arrive long before the pressure moves.
2. Why both arms, at least once
Almost everyone has some difference between arms, and it is usually a few mmHg and means nothing. A large or persistent difference is a genuine finding, and it is only ever found by someone who bothered to check.
What an inter-arm difference can mean
- It changes which arm you use from then on. NICE asks for both arms at diagnosis: if the difference exceeds 15 mmHg, repeat; if it persists, use the arm with the higher reading for all future measurements. Use the lower arm and you will under-treat indefinitely.
- Subclavian or peripheral arterial disease is the commonest explanation for a persistent difference — a narrowing on the low side. It is a marker of atherosclerosis elsewhere, which is why the difference itself is associated with higher cardiovascular risk in cohort data, independently of the pressure.
- Aortic dissection is the emergency version. A sudden, severe, tearing chest or back pain with a marked difference between arms — or an absent pulse on one side — is dissection until proven otherwise, and is a 999 call. This is the single reason the check earns its place in an emergency chapter rather than only a clinic one.
- Coarctation of the aorta in a child or young adult: high pressure in the arms with weak or delayed femoral pulses. Rare, congenital, and routinely missed for years.
- A technical difference — different cuff, different position, different moment — is the commonest explanation of all, which is why a single odd reading gets repeated rather than acted on.
How to do it properly: both arms, same cuff where possible, same position, ideally close together in time. If a device only allows sequential readings, do left then right then left again and compare like with like, because pressure drifts downward across a sitting and a simple left-then-right comparison will manufacture a difference that is not there.
3. The two observations no device does for you
- Respiratory rate. Count for a full minute in a child, and do it without announcing it — people change their breathing when they know it is being watched. Keep your fingers on their wrist as though taking a pulse. It is the most sensitive early sign of deterioration in almost every acute illness and the one most often estimated or left blank.
- Capillary refill. Press a fingertip or the sternum for five seconds and release; colour should return within two. It takes five seconds, needs nothing, and detects poor perfusion before blood pressure moves.