Blood pressure readings explained: how arm position, cuff size and other errors can shift the numbers

Measuring blood pressure (BP) is, by one account, the most commonly performed medical procedure in clinical practice. It is also one of the easiest to get wrong. A reading taken with the arm dangling can come out several points higher than one taken a few minutes later with the same arm resting on a desk.

How much higher was pinned down by a trial at Johns Hopkins University in Baltimore, published in JAMA Internal Medicine in 2024. The university’s account of it was republished at the end of September. An arm hanging at the side added 6.5 points to the top number of a reading, and a hand resting in the lap added about 4.

Arm position is one of more than two dozen things shown to distort a reading. That does not make any particular diagnosis wrong, and the trial did not look at diagnoses. It does explain why the guidance on how to take a reading is so specific.

What the arm position trial tested

A blood pressure reading has two numbers, both in millimeters of mercury (mm Hg), often just called points. The top number is the systolic pressure, meaning the force of blood when the heart pumps. The bottom number is the diastolic pressure, meaning the pressure in the arteries while the heart rests between beats.

Guidelines recommend arm support on a desk with the midcuff positioned at heart level, the midcuff being the middle of the inflatable band. The authors note that nonstandard positions are used in clinical practice all the same. The Johns Hopkins account describes the usual scene: a patient sitting on an exam table with little or no support for the arm, which is held by the clinician or left resting in the lap.

The team set out to measure what those positions do. Theirs was a crossover trial, meaning every participant was measured in every position and so served as their own comparison. It recruited adults between the ages of 18 and 80 years in Baltimore between August 2022 and June 2023.

The trial enrolled 133 participants, with an average age of 57. Seventy were women. Forty-eight had a systolic pressure of 130 or higher, and 55 had a body mass index of 30 or higher, the usual threshold for obesity.

The protocol was built to resemble a clinic visit done properly. Participants emptied their bladders and walked for two minutes, to imitate arriving at an appointment. They then rested while seated for five minutes, with their backs and feet supported. Each wore a cuff matched to the size of their upper arm, and an automated device took three readings in a row.

That sequence was run with the arm in three positions, in random order: supported on a desk, with the hand resting in the lap, and hanging unsupported at the side. Every participant then did a fourth round with the arm back on the desk.

The fourth round was the control. Blood pressure drifts from one set of readings to the next for reasons that have nothing to do with arm position, and the repeat desk measurement showed how much. The researchers subtracted that natural drift from the differences they saw with the lap and side positions.

How much arm position changed blood pressure readings

Both nonstandard positions pushed readings up.

Arm position	Top number (systolic)	Bottom number (diastolic)
Hand resting in lap	3.9 mm Hg higher	4.0 mm Hg higher
Arm hanging at side	6.5 mm Hg higher	4.4 mm Hg higher

The comparison in each case is with the arm supported on a desk. The patterns were generally consistent across subgroups, including people with and without high blood pressure, older and younger adults, and people with and without obesity.

Tammy Brady, the senior author and vice chair for clinical research in the Department of Pediatrics at the Johns Hopkins University School of Medicine, said arm position makes a large difference in obtaining an accurate blood pressure measurement.

The direction of the result was not a surprise. A systematic review from the University of Queensland, published in 2017, had already gathered the earlier evidence. Three studies addressed measurement in an unsupported arm compared with a supported arm, and all found higher readings. Ten more compared readings when the measured arm was at heart level with readings taken with the arm lower, and they too found increases, ranging from small to large. The studies gathered previously had varied in whether patients stood, sat or lay down, and seldom reported how far the arm was from heart level. What the Baltimore trial added was a controlled estimate, using a modern automated device, for two positions the authors describe as commonly used.

Why a few points on a blood pressure reading matter

Hypertension, meaning blood pressure that stays high over time, is defined by thresholds. A few points can put a person on one side of a line or the other.

Sherry Liu, one of the study’s authors, spelled out the arithmetic. A consistent overestimate of 6.5 mm Hg, Liu said, is “a potential difference between a systolic BP of 123 and 130, or 133 and 140”. Those are the levels at which US guidance separates normal from high, and one stage of hypertension from the next.

The stakes are spread across a great many people. Hypertension affects approximately 45% of the adult US population and is the most commonly diagnosed condition at outpatient office visits, according to the US Preventive Services Task Force, an independent panel that reviews the evidence for screening.

The authors’ own conclusion is that the lap and side positions may lead to misdiagnosis and overestimation of hypertension. The trial did not count misdiagnoses. It measured readings in a research setting, and the step from a biased reading to a wrong diagnosis is an inference.

Cuff size, the other blood pressure measurement error the team tested

The arm position trial was the second of a pair. A year earlier, the same group had run a near-identical experiment on cuff size, with a total of 195 adults in Baltimore.

Some settings keep one regular-size cuff and use it on everyone. The trial compared that habit with a cuff chosen to fit. For people with larger arms, the regular cuff read high. Among those who needed a large cuff, it added 4.8 mm Hg to the systolic reading. Among those who needed an extra-large cuff, it added 19.5.

The error ran the other way for people with slender arms. A regular cuff on someone who needed a small one read 3.6 mm Hg too low.

The authors called the problem particularly concerning for settings where 1 regular BP cuff size is routinely used in all individuals, regardless of arm size. An error of nearly 20 points is far larger than anything arm position produced.

The other things that shift a blood pressure reading

The Queensland review remains the fullest catalog. A total of 328 empirical studies were included. They investigated 29 potential sources of inaccuracy, sorted into four groups: the patient, the device, the procedure and the person taking the reading. A measurable effect turned up for 27 of them. Single sources moved the systolic number by anything from 23.6 mm Hg too low to 33 mm Hg too high.

Some of the sources are fleeting states of the patient. The review lists a recent meal, caffeine, nicotine, a full bladder and being cold among them. Others are matters of procedure:

• Rest. Patients had higher readings if they did not rest for a sufficient period of time prior to measurement.
• Talking. Six studies on the effect of the patient talking during a reading all found increases in both numbers.
• Posture. Crossed legs and an unsupported arm each raised readings in the studies reviewed, and a lone study of back support found the bottom number rose without it.
• The setting itself. Readings taken by a clinician tended to run higher than a person’s usual daytime level, the pattern known as the white-coat effect.

The review drew a distinction that matters for the arm trial. Passing influences such as caffeine fade, so a later reading can be closer to the truth. Posture is different. For crossed legs or an unsupported arm, the reviewers wrote, repeated measurements cannot eliminate their influence. A second and third reading taken in the same slouch will be wrong in the same direction.

How often blood pressure measurement goes wrong

The evidence here is one test of trainees, and they did poorly. In 2015 the American Medical Association tested students at a national meeting on a mock patient. One hundred fifty-nine students from medical schools in 37 states took part, and each was scored on 11 skills, among them resting the patient, uncrossing the legs, supporting the arm and choosing the right cuff.

Only one student demonstrated proficiency on all 11 skills. The mean number of elements performed properly was 4.1, out of the 11.

The patient actor in that test began on a raised stool with no support for the arm, back or feet, next to an empty chair and a table. The students were being checked on whether they would move the patient. All but one left something out.

The American Heart Association’s scientific statement on measurement, published in 2019, puts the weight on training. It calls for initial and ongoing training of technicians and healthcare providers along with devices that have been validated and calibrated.

How blood pressure readings at home fit in

A single clinic reading is not supposed to settle a diagnosis. The US Preventive Services Task Force recommends obtaining blood pressure measurements outside of the clinical setting for diagnostic confirmation before starting treatment.

One reason is the gap between settings. Some people have what is called white-coat hypertension, meaning readings that are high in the office but not outside the office. Home monitoring is one way to find them. Several national and international hypertension guidelines endorse it, according to a joint statement from the American Heart Association and the American Medical Association.

Home readings are open to the same errors of posture. Brady made that point in the university’s account, saying patients “must advocate for themselves in the clinical setting and when measuring their BP at home”. The university’s account lists the standard steps, which are the same wherever the cuff goes on: the right cuff size, the back supported, the feet flat on the floor with the legs uncrossed, and the arm resting on a desk or table with the cuff at heart level.

None of this is a reason to doubt a diagnosis already made or the treatment that followed it. What a set of readings means for one person is assessed by their doctor. Do not stop or change a prescribed medicine without talking to your doctor.

What the arm position trial cannot show

Readings from other devices. Every measurement was taken by one kind of automated machine. The investigators say the results may apply only to blood pressure measurements taken with automated devices, and may not carry over to other kinds of equipment.

Diagnoses or outcomes. The trial compared readings minutes apart in 133 volunteers at a single visit. It did not follow anyone to see whether a diagnosis or a prescription changed.

Real clinic conditions. Participants rested for five minutes in a quiet, private room and were asked not to talk or use their phones. A hurried clinic may add other errors on top of arm position, in either direction, and the trial did not measure how they combine.

Every patient. The volunteers came from one city. The published results give averages, and an average does not say how much one person’s reading would move.

The work was funded by Resolve to Save Lives, a health initiative backed by Bloomberg Philanthropies and the Bill and Melinda Gates Foundation. The cuff size trial that preceded it came from the same research group, so the two share investigators and methods.

Letting the arm hang at the side raised automated blood pressure readings by 6.5 points on average in a 133-person trial, a result that shows how much technique matters without showing how often it changes a diagnosis.