brand logo

Am Fam Physician. 2005;72(7):1391-1398

Diagnosis and treatment of hypertension depend on accurate measurement of auscultatory blood pressure. The lowering of target blood pressure for patients with diabetes or renal disease has made detection of small differences more important. However, blood pressure reading is one of the most inaccurately performed measurements in clinical medicine.

“True” blood pressure is defined as the average level over a prolonged duration. Thus, in-clinic blood pressure measurement, which generally makes no allowance for beat-to-beat variability, can be a poor estimation and may fail to catch high blood pressure that occurs only outside the clinic setting. In addition, faulty methods and the “white coat effect” (an increase in blood pressure when a physician is present) may lead to misdiagnosis of hypertension in normotensive patients.

To increase accuracy of clinic readings, and in recognition of major changes over the past 10 years (including the prohibition of mercury in many countries), the American Heart Association (AHA) has published a new set of recommendations for the measurement of blood pressure. The AHA scientific statement, written by Pickering and colleagues, was first published in the January 2005 issue of Hypertension and also appears in the February 8, 2005, issue ofCirculation. It can be accessed online at A summary of the AHA scientific statement follows.

Classification of Hypertension

Systolic and diastolic blood pressures are preferred for use in hypertension classification, rather than arterial or pulse pressure. A classification of hypertension and prehypertension from the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure is given in Table 1. Prehypertension has increasing health risks and can progress to hypertension.

ClassificationBlood pressure (mm Hg)
Normal119 or lower79 or lower
Prehypertension120 to 13980 to 89
Stage 1 hypertension140 to 15990 to 99
Stage 2 hypertension160 or higher100 or higher

Interpretation of blood pressure measurements in children must take into account the child’s age, sex, and height. Hypertension in children and adolescents is defined as systolic or diastolic pressure, or both, at or above the 95th percentile of distribution, using tables to determine normal and elevated levels.

In-Clinic Measurement

In the standard clinic procedure (mercury sphygmomanometer with the Korotkoff ’s sound technique), accurate measurement of blood pressure depends on the person doing the reading, or the “observer.” Proper training; use of an accurate, well-maintained device; correct selection and positioning of the cuff; appropriate positioning of the patient; and recognition of factors that may skew the measurement are critical. One of the most common observer errors is terminal digit bias (e.g., excessive recording of “zero” as the last digit, or fitting the measurement to a specific recognized threshold). Guidelines for in-clinic measurement are summarized in Table 2.

Patient should be seated comfortably, with back supported, legs uncrossed, and upper arm bared.Diastolic pressure is higher in the seated position, whereas systolic pressure is higher in the supine position.
An unsupported back may increase diastolic pressure; crossing the legs may increase systolic pressure.
Patient’s arm should be supported at heart level.If the upper arm is below the level of the right atrium, the readings will be too high; if the upper arm is above heart level, the readings will be too low.
If the arm is unsupported and held up by the patient, pressure will be higher.
Cuff bladder should encircle 80 percent or more of the patient’s arm circumference.An undersized cuff increases errors in measurement.
Mercury column should be deflated at 2 to3 mm per second.Deflation rates greater than 2 mm per second can cause the systolic pressure to appear lower and the diastolic pressure to appear higher.
The first and last audible sounds should be recorded as systolic and diastolic pressure, respectively. Measurements should be given to the nearest 2 mm Hg.
Neither the patient nor the person taking the measurement should talk during the procedure.Talking during the procedure may cause deviations in the measurement.

Observers should be assessed for physical and cognitive competency to perform the procedure, including vision, hearing, and eye/hand/ear coordination. Retraining of all health care professionals is strongly recommended by the AHA. Training methods using audiovisual tapes to test and retest accuracy are extremely effective. Online resources that may be useful include a guide to accurate blood pressure measurement at, and an instructional video from BMJ Books


According to the AHA, mercury sphygmomanometers still are the preferred device and should be used if available and properly maintained, although they are being removed from clinical practice for environmental reasons. The tubing between the device and the cuff should be 27.5 inches (70 cm) or more in the office setting. The system must be airtight, so the tubing and release valve should be inspected regularly.

Other devices, such as aneroid and hybrid sphygmomanometers, may be used as a substitute or a supplement, but there is no widely accepted replacement. Mercury sphygmomanometers still are necessary for evaluating the accuracy of other devices.

Automated oscillometric devices may be useful for an increased number of readings and to avoid expensive training. Devices with linear deflation rates may be more accurate than those with stepwise deflation. Measurements with automated devices typically are lower than those taken by a physician, perhaps because of the white coat effect; correct patient position and cuff selection still are required. Some automated devices have been validated for use during pregnancy and may be useful as an alternative to mercury devices in the future. Automated devices are acceptable in newborn and young infants and in the intensive care setting, although their reliability is unclear.


The position of the patient can have a sizable impact on blood pressure measurements. For the most accurate measurement, the AHA recommends that the patient be relaxed and seated with legs uncrossed and back and arm supported. Children should have their feet on the floor rather than dangling above it. If possible, the patient should be seated five minutes before the reading. All clothing covering the cuff location should be removed (rolled-up sleeves, if tight, may create a tourniquet effect above the cuff).

The middle of the cuff on the upper arm should be level with the right atrium, at the midpoint of the sternum. If the upper arm is below the level of the right atrium, the readings will be too high; if the upper arm is above heart level, the readings will be too low In the supine position, the arm should be supported on a pillow to raise it above the level of the heart, which is situated about halfway between the bed and the sternum. In women who are pregnant, the left lateral recumbency position can be used, with measurement on the left arm.

PatientRecommended cuff size
Adults (by arm circumference)
22 to 26 cm12 × 22 cm (small adult)
27 to 34 cm16 × 30 cm (adult)
35 to 44 cm16 × 36 cm (large adult)
45 to 52 cm16 × 42 cm (adult thigh)
Children (by age)*
Newborns and premature infants4 × 8 cm
Infants6 × 12 cm
Older children9 × 18 cm

The patient should not talk during the procedure, because this may cause deviations in the measurement. Other factors that can affect the measurement include exercise, smoking, alcohol consumption, muscle tension, bladder distension, room temperature, and background noise.

In older patients, blood pressure should be measured routinely in the standing and seated positions to screen for postural hypotension


The most common error in blood pressure measurement is use of inappropriate cuff size. Considerable overestimation can occur if the cuff is too small. The bladder length recommended by the AHA is 80 percent of the patient’s arm circumference, and the ideal width is at least 40 percent. Error is minimized when the cuff width is 46 percent of the arm circumference, although for large adult and thigh cuffs this is not practical. In obese patients, longer, wider cuffs are needed to compress the brachial artery adequately. In children, cuff bladder width should be at least 40 percent of the arm circumference halfway between the olecranon and acromion; the cuff should then cover 80 percent or more of the arm circumference. Recommended cuff sizes are listed in Table 3.

For correct cuff placement, the midline of the cuff bladder should be positioned over the arterial pulsation in the patient’s upper arm following palpation of the brachial artery in the antecubital fossa. There should be a 2– to 3–cm space for the stethoscope between the lower end of the cuff and the antecubital fossa, unless this would require an undersized cuff. In patients with an arm circumference greater than 50 cm, the cuff should be wrapped around the forearm, supported at heart level, and the radial pulse felt at the wrist


The rate of deflation in indirect blood pressure measurement significantly impacts the reading. The AHA recommends that the cuff be inflated to at least 30 mm Hg above the point at which the radial pulse disappears. The cuff should then be deflated at a rate of 2 to 3 mm Hg per second (or per pulse when the heart rate is slow). Deflation rates greater than 2 mm Hg per second can cause the systolic pressure to appear lower and the diastolic pressure to appear higher. In pregnant women, the fifth Korotkoff ’s sound has been recommended as the diastolic measurement, although the fourth Korotkoff ’s sound should be used when sounds are audible with the cuff deflated.

For a child, overinflation of the cuff may cause discomfort. One technique to avoid this is to estimate the systolic pressure by inflating the cuff while palpating the pulse, and then inflate the cuff to 30 mm Hg above the estimated level when the pressure is auscultated.


The AHA recommends that at least two readings be taken, with a one-minute interval between them, and the average of the measurements recorded. The first reading in a series is usually the highest. Additional readings should be taken if the difference between the first two is greater than 5 mm Hg

At the first visit, blood pressure should be measured in both arms, which may be useful for identifying coarctation of the aorta and upper-extremity arterial obstruction. If there is a consistent difference in measurement between the arms, the highest pressure should be recorded. In children, the right arm is always preferable for consistency and comparison with reference tables.


Children should not be diagnosed with hypertension without confirmation from repeated visits unless they are symptomatic or have profoundly elevated levels. The most precise measurement is the average of multiple readings taken over weeks or months, because this allows for reduction of anxiety. In children, a difference of several millimeters of mercury often is found between the fourth and fifth Korotkoff’s sounds. Children with repeatedly elevated measurements should have leg blood pressure measured to screen for coarctation of the aorta. This can be done by auscultation over the popliteal fossa, with use of a thigh cuff or oversized arm cuff. A systolic thigh blood pressure that is more than 10 mm Hg lower than the systolic arm pressure is cause for additional coarctation testing.

Ambulatory Monitoring

Ambulatory blood pressure monitoring is a noninvasive, automated process that records blood pressure over an extended period. Typically, readings are taken every 15 to 30 minutes for 24 hours, with around 50 to 100 readings in total. Data from the device are downloaded into software and can be translated into a report.

Indications for ambulatory blood pressure monitoring are listed in Table 4. Ambulatory monitoring may predict risk for morbidity more accurately than in-clinic blood pressure readings. It usually is used for diagnosis in patients with suspected “white coat” hypertension—who are thought to be at lower risk of blood pressure–related complications than those with sustained hypertension—and recently has been approved by the Centers for Medicare and Medicaid Services for this purpose. Another use is the identification of “nondipping” blood pressure (i.e., less than 10 percent reduction from daytime to nighttime pressure), which is thought to increase a patient’s risk for complications. Increasing evidence suggests that patients with hypertension whose blood pressure is lower at night have less risk for cardiovascular morbidity than those whose pressure remains high. Ambulatory measurements in older patients can identify episodic hypotension. In addition, ambulatory monitoring may be useful in patients with refractory hypertension but little organ damage, those with suspected autonomic neuropathy, and those who have wide differences between readings from home and clinic, as well as for monitoring treatment.


Blood pressure measurements taken on home monitors (typically oscillometric devices that record brachial artery pressure) may, according to prospective studies from Japan and France, be more accurate predictors of morbidity than clinical readings. Increasing evidence also suggests better target organ damage prediction with home readings, and self-measurement in older patients may aid physicians in antihypertensive medication dosing decisions. Electronic home monitoring is easy to use, cost-effective, and may improve therapeutic compliance. However, only five devices so far have passed proper validation tests (a list is available online at

Episodic hypotension
Monitoring treatment
“Nondipping” blood pressure that does not drop overnight Suspected autonomic neuropathy
Suspected overtreatment with resultant iatrogenic hypotension
Suspected “white coat” hypertension and discrepant readings between home and clinic

Patients interested in self-monitoring should be instructed to sit in a comfortable chair for three to five minutes with the upper arm at heart level before taking a measurement, and should be reminded not to exercise or eat directly before the reading. Three readings should be taken at intervals of one minute or longer, and the average of the three recorded. Early morning and evening readings are especially helpful. Normal blood pressure at home is lower than that in the clinic. The upper limit suggested by the American Society of Hypertension is 135/85 mm Hg. Because of the potential for incorrect reporting to the physician, the AHA recommends that devices with memory or printouts be used. A lower home blood pressure goal is recommended for pregnant women, patients with diabetes, and those with renal failure, among others.

Continue Reading

More in AFP

Copyright © 2005 by the American Academy of Family Physicians.

This content is owned by the AAFP. A person viewing it online may make one printout of the material and may use that printout only for his or her personal, non-commercial reference. This material may not otherwise be downloaded, copied, printed, stored, transmitted or reproduced in any medium, whether now known or later invented, except as authorized in writing by the AAFP.  See for copyright questions and/or permission requests.