The unit is a height

Pressure is a push spread over an area. Engineers measure it in pounds per square inch and scientists in pascals. Medicine measures blood pressure by asking a different question: how tall a column of liquid could this push hold up?

That approach dates from a time before electronic sensors existed. If you connect an artery to a vertical tube, blood climbs the tube until the weight of the column balances the pressure pushing it. Read the height, and you have the pressure. Mercury became the liquid of choice because it is heavy. According to the conversion factors published by the National Institute of Standards and Technology, mercury is about 13.6 times as dense as water, so a mercury column stays short. A pressure of 120 mm Hg would lift water about 5.3 feet, which would call for a tube taller than most people. The same pressure lifts mercury less than 5 inches. A mercury gauge marked all the way to 300 mm Hg is about a foot tall.

To put the unit on a familiar scale, 1 mm Hg equals about 133 pascals, or about 0.019 pounds per square inch. A systolic reading of 120 mm Hg works out to roughly 2.3 pounds per square inch, and a diastolic reading of 80 works out to about 1.5. A typical car tire is inflated to around 32 pounds per square inch, roughly 14 times the pressure in your arteries at the peak of a heartbeat.

You may occasionally see other spellings or units. Monitors and journals print the unit as either mm Hg or mmHg, and they mean the same thing. If a source reports blood pressure in kilopascals (kPa), multiply by 7.5 to get mm Hg. A reading of 16.0 over 10.7 kPa is 120 over 80 in the units you know.

How blood pressure came to be measured this way

The first recorded blood pressure measurement used the height method literally. In 1733, an English clergyman and scientist named Stephen Hales connected a long glass tube to an artery in a horse and watched the blood rise. He recorded that it climbed 8 feet 3 inches above the level of the heart. Blood is close to water in density, so that column corresponds to roughly 185 mm Hg in today's unit, a plausible figure for a large, restrained animal.

Nobody wanted an 8 foot tube, and nobody wanted to open an artery to get a number. Over the next 150 years, inventors worked on ways to measure the pressure from outside the body. In 1881, Samuel von Basch built a device that pressed on an artery and recorded the systolic pressure as the height of a mercury column. In 1896, an Italian physician, Scipione Riva Rocci, introduced the design that still shapes every cuff you have ever worn: a rubber bag wrapped around the upper arm inside a sleeve of material that does not stretch, inflated with a hand bulb, and connected to a mercury manometer. The user pumped the cuff until the pulse at the wrist disappeared, then let air out slowly and noted the height of the mercury when the pulse returned. That gave the systolic number only.

The second number arrived in 1905. Nikolai Korotkoff, a Russian surgeon, presented his method to the Imperial Military Medical Academy in St. Petersburg in November of that year. He had placed a stethoscope over the artery just below a Riva Rocci cuff and listened as the pressure came down. The first tapping sound marked the systolic pressure. The point where the sounds faded away marked the diastolic pressure. Each was read as a height on the mercury column, which is why both numbers on your monitor share the same unit.

A separate article in this series explains what the two numbers describe about your heart and arteries. For this article, the point is that the unit came from the instrument. Doctors read blood pressure off a mercury column for most of the twentieth century, so blood pressure came to be spoken of in millimeters of mercury, the same way car engines are still rated in horsepower long after horses left the road.

Your monitor has no mercury in it

An automatic upper arm monitor works on a different principle called oscillometry. The cuff still squeezes the arm, but there is no column of anything to read. Instead, an electronic pressure sensor inside the device measures the air pressure in the cuff many times per second.

As the cuff deflates, each heartbeat sends a pulse of blood through the artery under the cuff, and each pulse produces a tiny wobble in cuff pressure. The wobbles are faint while the cuff is tight enough to block the artery, grow strongest when the cuff pressure is close to the average pressure in the artery, and fade again as the cuff loosens. The monitor's software reads this pattern of pulsations and estimates systolic and diastolic pressure from it. The sensor is calibrated at the factory so its output lines up with the traditional unit, and the screen shows mm Hg.

Two features of this method are worth understanding. First, the device does not hear anything. The tapping sounds Korotkoff described play no part. Second, the systolic and diastolic values are estimates produced by a calculation rather than direct observations of a column. A 2014 review written for health care professionals describes oscillometric devices as providing an estimation of blood pressure and notes that they became the clinical standard for practical reasons: listening with a stethoscope takes training, automatic devices cost less over time, and mercury has been phased out or banned in many states and countries. Each manufacturer writes its own software to turn the pulsation pattern into numbers, and those calculations are not identical from one brand to the next.

That is why validation matters. The 2025 American Heart Association and American College of Cardiology guideline says automated oscillometric devices are a reasonable replacement for the stethoscope method, but it recommends only devices that have passed a rigorous standardized testing protocol, in which a device's readings are compared with careful reference measurements in a group of volunteers. The guideline points to validatebp.org as the vetted list for the United States. It also notes that devices should be recalibrated according to the manufacturer's guidance, and that cuffless devices, which estimate pressure without inflating a cuff, are not recommended for diagnosing or managing high blood pressure.

Why the mercury gauge is disappearing

Mercury is toxic, and a broken gauge releases it. Under an international treaty called the Minamata Convention on Mercury, countries agreed to phase out or reduce mercury in a range of products, including measuring devices. According to the United States Environmental Protection Agency, the treaty text was finalized in January 2013, the United States signed it and became a party on November 6, 2013, and it entered into force in 2017. Many US hospitals and clinics had already moved to aneroid gauges (the round dial type) and automatic monitors before then. The 2014 review mentioned above noted that mercury had by that time been phased out or banned in many states and countries.

The unit survived the instrument. Guidelines, research trials, and the screens on every home monitor still speak in millimeters of mercury because more than a century of medical knowledge is written in that language. The blood pressure categories in the 2025 guideline, from normal below 120/80 to stage 2 hypertension at 140 or higher systolic or 90 or higher diastolic, are all expressed in mm Hg and rest on decades of studies that reported pressure in this unit.

A trained observer with a stethoscope and a calibrated gauge is still the reference in many validation studies. Your monitor is designed to match that reference on average, across a group of people. It is not guaranteed to match it exactly on your arm on a given morning.

What one millimeter can and cannot tell you

A home monitor displays whole millimeters, so a reading of 127 looks more precise than 130. The precision of the display and the accuracy of the device are two different things. Even a validated, properly calibrated monitor can read a few millimeters above or below the true pressure, and a monitor that has been dropped or used for years may drift further.

Research on device error puts numbers on this. A 2017 systematic review of 328 studies on sources of measurement error found that, depending on the study, between 1.4 and 69.7 percent of aneroid gauges and between 4.5 and 26 percent of automated devices were off by more than 3 mm Hg when checked against a reference. In another study, researchers tested 85 wrist and upper arm monitors that people had brought from home against two trained observers using stethoscopes. Of those devices, 69 percent differed from the reference by 5 mm Hg or more on the systolic or diastolic number. Twenty nine percent differed by 10 or more, and 7 percent by 15 or more. Larger arm circumference predicted higher readings, which points to cuff fit as one cause.

These figures do not mean your monitor is useless. They mean the number on the screen deserves the right amount of trust. A difference of 1 or 2 mm Hg between two readings, or between your monitor and the one at the clinic, is noise. A difference of 5 is within the range that devices commonly show and is not by itself a reason to distrust either machine. A consistent difference of 10 or more between your monitor and a clinic reading taken with good technique is worth investigating.

The number that matters for your health is not any single reading but the average of several. The 2025 guideline bases its categories on an average of two or more careful readings on two or more occasions, and it states that a single reading is inadequate for clinical decisions. The guideline also notes that across large populations, cardiovascular risk roughly doubles for each 20 mm Hg increase in systolic pressure and each 10 mm Hg increase in diastolic pressure. That is a statement about long term averages, not about the change between Tuesday and Wednesday.

Why arm height changes the number

The height idea behind the unit is not just history. It is at work in your arm every time you take a reading. Blood between your heart and the cuff behaves like a short column of fluid. If the cuff sits below the level of your heart, the weight of that column adds to the pressure the cuff feels. If the cuff sits above your heart, the column subtracts from it.

The size of the effect follows from the physics of fluid columns. Blood is close to water in density, and by the standards institute's conversion factors, a column of water 1 inch tall exerts about 1.9 mm Hg. So each inch the cuff sits below your heart adds roughly 2 mm Hg to the reading, and each inch above subtracts about the same. A cuff on an arm hanging at your side sits several inches below your heart.

A randomized crossover trial published in JAMA Internal Medicine in 2024 tested this in 133 adults. Compared with the arm resting on a desk at heart level, resting the arm on the lap raised the systolic reading by an average of 3.9 mm Hg and the diastolic by 4.0. Letting the arm hang at the side raised systolic by 6.5 and diastolic by 4.4. The systematic review mentioned earlier found studies in which an arm below heart level raised systolic readings by anywhere from 3.7 to 23 mm Hg.

Those are differences large enough to move a reading from one guideline category to another without any change in your actual blood pressure. The fix costs nothing. Sit at a table, rest your arm on it, and position the cuff so its middle is level with the middle of your chest. If you use a wrist monitor, the same rule applies and is harder to follow, because the wrist has to be held at heart level for the entire reading.

Reading the display with the unit in mind

Suppose your monitor shows 134/86 mm Hg this morning and 129/81 yesterday. Nothing about the unit lets you say your pressure went up 5. The difference is inside the range a validated device can show from one reading to the next, and it is smaller than the effect of resting your arm on your lap. What the unit does let you do is compare your numbers with the ranges in the guideline, once you have an average built from repeated readings taken the same way each time. A few habits follow from that.

1. 1 Treat single millimeters as noise and look at averages of several readings over several days. A separate article in this series covers what home, office, and ambulatory measurements each mean to your health care professional.
2. 2 Check whether your monitor appears on the validatebp.org list, and ask the manufacturer how often it should be recalibrated. A device that has been dropped may need checking sooner.
3. 3 Whenever your home readings and clinic readings seem to disagree, bring your monitor to an appointment and take a reading on both devices a minute or two apart, with the same arm, the same position, and the same rest beforehand.
4. 4 Keep the middle of the cuff level with your heart, with your arm supported on a table. Each inch off adds or subtracts about 2 mm Hg.
5. 5 When a reading surprises you in either direction, wait at least 1 minute and take it again before drawing any conclusion. The number is an estimate produced by software, not a column of mercury read by a nurse.

One exception to the wait and see approach is spelled out by the American Heart Association. If a reading is 180/120 mm Hg or higher, wait at least 1 minute and take it again. If it is still that high and you have chest pain, shortness of breath, back pain, numbness or weakness, a change in vision, or difficulty speaking, call 911. If it is still that high and you have no symptoms, contact your health care professional as soon as possible.

If your monitor and the clinic's device disagree by 10 mm Hg or more on repeated side by side checks, ask your care team which readings they want to rely on and whether your monitor should be replaced.