The introduction of the sphygmomanometer is attributed to Potain, who is probably best remembered by cardiologists as the discoverer of cardiac gallops and best remembered by the literate as the great Parisian diagnostician in Marcel Proust’s Remembrance of Things Past. Potain’s pupil, Riva Rocci, invented the mercury manometer, which led to the dissemination of indirect sphygmomanometry for the systolic pressure and, once Korotkoff had discovered the sounds known by his name, for the diastolic as well. The latter story is a fable of medicine, well worth telling.

Korotkoff was a surgeon in the Czar’s army, not an internist. He was doing experimental work on posttraumatic arteriovenous fistulas in the surgical dog. Pirogoff, his teacher (referred to in Dostoyevsky’s The Idiot), had taught him always to auscultate over any area before performing an incision. On one occasion, while auscultating over an artery just as he was releasing a tourniquet, he heard thumping sounds! Abandoning his original scientific problem, he attempted to quantitate the amount of pressure required to make these auscultated sounds appear and disappear. He noticed that the sounds correlated with systole and diastole, as could be determined by direct inspection of the flow of blood from the distally severed artery of the dog.

It is worth noting that when Korotkoff first found these sounds in humans and reported his research, some thought him to be quite mad. At the very least, his auditors found his suggestion that the sounds originated from pressure changes in the artery to be unacceptable because they “knew” that all such sounds had to emanate from the heart. Korotkoff’s response that the sounds could not come from the heart because they disappeared when the artery was completely occluded (Geddes et al., 1966) convinced no one, although apparently none of his critical professors had bothered to understand, let alone to replicate, his experiment. Returning to the panel of experts a second time, Korotkoff produced more evidence supporting the concept that the sounds came from the artery, not the heart, but the panel remained unconvinced (Segall, 1980). This story, a part of the oral tradition in which I was intellectually raised, was later confirmed by the written record (Multanovsky, 1970).

The final chapter of the story is unknown. Just as Lavoisier was guillotined during the French Revolution,¹ Korotkoff was said to have been arrested after the Russian Revolution. One version holds that he finally died during a Stalinist purge; another that he perished in 1920. As Stalin did not seize power until the late 1920s, this teaches us that history repeats itself and that historians repeat each other.

Shortly after the dissemination of the blood pressure cuff, workers realized that some cuffs were of insufficient length or width to transmit the bladder pressure efficiently through the intervening tissues to the brachial artery (Geddes et al., 1966). Pioneer after pioneer increased either the length or the width of the bladder in this cuff. (The length is the dimension of the bladder that is
wrapped around the arm.) Then, suddenly, the march of progress stopped short: short of a length or width sufficient to guarantee that arm circumferences greater than 27 cm would not generate spuriously hypertensive readings on occasion.

The above is the historic basis for the belief that one should use as long or as wide a blood pressure cuff bladder as possible. Dr Sapira recommended a thigh cuff; the scientific basis for this preference is reviewed later. This author finds a thigh cuff to be uncomfortable or painful for most patients and difficult to apply and inflate; she therefore prefers a large arm cuff for most patients.

According to the British Hypertension Society (BHS) guidelines, the bladder should encircle at least 80% of the arm’s circumference (Markandu et al., 2000). The width of the cuff should be approximately equal to two thirds of the distance between the axilla and the antecubital space; a 12-cm wide cuff is adequate for most adults. Erroneously low readings may occur with a bladder that is too wide (Kaplan, 2002).

Many authors, when discussing the proper size of the bladder, refer to the blood pressure “cuff,” when they really mean the bladder, which is inside the cuff. While the bladder, as a rule, has exactly the same width as the cloth covering (the cuff), the bladder is usually much shorter than the total cuff. Thus, before purchasing or making assumptions about a blood pressure cuff, one should inflate the bladder to see what its true dimensions are. (See the section entitled “The Fat Arm,” later in this chapter.)

As critical therapeutic decisions are based upon the blood pressure reading, it was shocking to discover that more than half of the mercury sphygmomanometers and their associated cuffs in a big London teaching hospital had serious problems that would have rendered them inaccurate (Markandu et al., 2000).

Be sure that the sphygmomanometer reading is zero when the cuff is not inflated. If you are using a mercury sphygmomanometer, be sure that the column of mercury is vertical. The mercury column should rise smoothly during cuff inflation and stop immediately when inflation stops.

The mercury should be a clean silver color; it oxidizes with time, forming a black powder. If a large amount of black powder has accumulated, the mercury needs to be removed and the column and reservoir cleaned. A method for doing so has been described (Yeats, 1992), although such instruments are likely to be retired, in these times when mercury encased in glass is of greater concern than that in vaccines injected into babies or in amalgams used in dental restorations.

If the mercury column is inadequately damped, or “bouncy,” tighten the knurled nut at the top of the column (Reeves, 1995).

For your black bag, you will probably purchase a sphygmomanometer with an aneroid pressure gauge. Be sure that it is a kind that does not have a pin stop, that is, a device that keeps the gauge from reading below zero when the cuff is deflated. The pin stop keeps you from being able to see that the gauge is out of calibration.

The aneroid gauge should be checked periodically, about every 6 months, for accuracy, preferably against a mercury sphygmomanometer. You will need a Y connector and tubing so that the bladder in your cuff can be simultaneously connected to both the mercury manometer and the aneroid gauge. Alternately, you can roll up two blood pressure cuffs together, inflate them partially, and squeeze them, checking their gauges against each other (whether one is a mercury manometer or both are aneroids). One survey found that 30% of aneroid instruments were off by 10 mm Hg or more, usually reading too low (Reeves, 1995). Another survey found inaccuracy in more than 50% of the aneroid manometers, with errors greater than 13 mm Hg in 7% (Mion and Pierin, 1998).

The aneroid gauge is a mechanical pressure transducer. The increase in pressure in the cuff causes an expansion of corrugated metal bellows, which drives the indicator needle through a series of gears. Repeated expansion of the bellows leads to a loss of elasticity, causing greater inaccuracy at high readings. Any trauma to the instrument can disrupt the gear system, causing inaccuracy at all pressure levels. Testing at two or three points does not exclude inaccuracy at other pressure levels (Bailey et al., 1991).

The increased use of home blood pressure monitoring—as well as the sloppy techniques of clinicians that lead to inaccurate readings—has led to recommendations to use automatic devices with digital readouts. Such devices have been notorious for inaccuracy, and only a fraction of the hundreds of models available worldwide have been subjected to independent validation. There are two published standards for evaluating blood pressure devices: the American Association for the Advancement of Medical Instrumentation (AAMI) standard and that of the BHS. The devices currently available for self-measurement generally use an oscillometric technique, with an algorithmic method zealously guarded by manufacturers. Such techniques cannot measure blood pressure accurately in all situations, especially in the presence of arrhythmias. In some patients, they do not work for reasons that are not always apparent.

Of 30 tested devices, 28 did not meet the AAMI standard of a mean difference of less than 5 mm Hg from an intra-arterial measurement of systolic pressure, and 9 of 30 failed the standard for diastolic pressure. Additionally, 27 of 30 failed the AAMI requirement of a standard deviation of less than 8 mm Hg with regard to systolic pressure and 7 of 30 with regard to diastolic pressure (Van Egmond et al., 1993).

With monitors that passed AAMI and BHS validation criteria, more than half the patients tested may have average measurements that are in error by more than 5 mm Hg (Schwartz et al., 2003). Stiff arteries (Jones et al., 2003) and slow heart rates (Bendjelid, 2003) are additional sources of error.

Devices measuring blood pressure in the finger are not recommended because measurements may be distorted by peripheral vasoconstriction. Measurement at the wrist is also problematic. The upper arm devices are best. The usual advice concerning cuff size applies.

A physician who is going to rely on self-measurement by patients should have the patient bring the equipment to the office and demonstrate his technique. A mercury sphygmomanometer should be used to check the calibration. Be sure that the patient is placing
the cuff at the level of the heart on the arm with the higher blood pressure reading (vide infra). The diagnostic threshold may be different for home measurements, but data from longitudinal studies are lacking (O’Brien et al., 2001).

There are two steps here: First, measuring the palpable systolic blood pressure, and second, the more customary auscultatory determination, also called the indirect determination of the systolic and diastolic pressures.

Sometimes, in high output states with vasodilatation, the sounds do not disappear until the cuff is fully deflated. In this case, you would record the blood pressure as K1/K4/0 (where K1 or the first Korotkoff sound is the systolic pressure, K4 is the point of muffling if heard) or K1/0 (Grim and Grim, 2008).

The American Heart Association recommends taking a second reading after the patient has rested for at least 30 seconds and averaging the two (American Heart Association, 1980).

If the Korotkoff sounds are very faint, you may have the patient, wearing the cuff, raise his arm and open and close his fist several times. Then, inflate the cuff; lower the arm with further inflation, if needed; and listen again (Reeves, 1995).

It is possible to measure both systolic and diastolic pressures by palpation. The procedure was independently discovered by Ehret in 1909 and by Segall in 1940 (Enselberg, 1961). In Segall’s original study (Segall, 1940), using light palpation over the vessel to sense the Korotkoff vibrations, it was possible to get values within 10 mm of the auscultatory measurements of both systolic and diastolic pressures in all 100 subjects. In more than half the cases, there was no difference between the pressures as determined by palpation versus auscultation.

If you place your thumb lightly over the brachial artery, you should be able to feel the “sharp” (phase 4) Korotkoff sounds as sharp knocks a little before the diastolic pressure reading. After a slight increase in sharpness, they suddenly disappear, and then the normal brachial pulse can be felt. The disappearance of knocks was found to have an excellent correlation with the diastolic pressure determined by auscultation (correlation coefficient 0.99) in 50 adult inpatients (Vaidya and Vaidya, 1996).

There is really no such thing as the blood pressure. A person’s blood pressure varies throughout the day. Try the following as a self-study:

Check the effect of various substances on the blood pressure. Smoking tobacco is said to increase systolic blood pressure by 20 mm Hg within 4 minutes (Kaplan, 2001). What is the effect of alcohol or caffeine? A number of widely used drugs may increase blood pressure (Kauffman, 2006), including COX-2 inhibitors, antimigraine drugs, anti-incontinence drugs, nonsteroidal anti-inflammatories (Gurwitz et al., 1994), oral contraceptives, cold remedies, cyclosporin, and tricyclic antidepressants (Joint National Committee, 1988).

In the 1970s, the upper limit for resting blood pressure was 160/95 mm Hg. Like “acceptable” cholesterol levels, blood pressure goals have been migrating downward. According to one major cardiology textbook (Black et al., 2001), the upper limit of acceptable in 2001 was 140/90, which corresponds to a home measurement of about 133/84. Still lower levels are, however, considered desirable. The “optimal” level of 120/80 mm Hg or less is actually observed in only 25% of adult men for systolic pressure and in 36% for diastolic pressure (Kaplan, 2001).

Using actual death rates from the Framingham study (Port et al., 2000a), rather than values derived by a computer curve-fitting algorithm, one can deduce that serious risk elevation begins at 165 mm systolic pressure if one is conservative and at 185 mm if one is not (Kauffman, 2006). Blood pressure rises with age, and hypertension is less damaging in women. Port et al. give treatment thresholds by age that are much higher than those recommended by official sources in 2004, such as the Health Services Advisory Group (HSAG), while warning that risk rises more rapidly with pressure than previously thought for persons in the upper 20% of pressures for their age and sex, indicating the need for more aggressive treatment (Port et al., 2000b). HSAG, by contrast, set a target of 75% of patients having their most recent blood pressure below 130/80, regardless of age. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure calls blood pressures between 120 and 139 systolic or 80 to 89 mm Hg diastolic “prehypertensive” (Chobanian et al., 2003).

While treatment standards may vary, one constant remains: the need to measure blood pressure frequently and accurately.

There has long been a controversy as to whether the muffling or the disappearance of the Korotkoff sounds should be used as the diastolic pressure. The first committee of the American Heart Association recommended that the fourth phase (muffling) of the Korotkoff sounds be adopted as the correct locus for diastolic pressure. In 1959, a second committee recommended that it be the fifth phase (disappearance). In 1967, a third committee switched back to recommending the muffling (fourth phase), but the published recommendation included an appendix stating why disappearance (fifth phase) was superior! In fact, one of the members of that committee used disappearance for determining the diastolic blood pressure in the famous Veterans Administration Cooperative Study. The next committee recommended using disappearance in adults and muffling in children. However, a Finnish study showed that the fourth phase sounds were absent in between 3% and 6% of children aged 6 to 18 years (Uhari et al., 1991). A Dutch study showed that the fourth sound could not be detected at all in 23% of pregnant women and, if present, could not be precisely measured (κ = 0.42 for the fourth sound and κ = 0.96 for the fifth sound) (Franx et al., 1998). It should be clear even to the casual observer that expert committees are not necessarily correct.

A thorough review of the literature clearly indicates to me that in most persons, the disappearance of the sounds should be used for the diastolic pressure (vide infra). The various drug studies proving the value of antihypertensive therapy have all used diastolic pressures determined by disappearance.

Admittedly, there are situations of high stroke volume in which the disappearance gives an artifactually low pressure, sometimes even an impossible 0 mm Hg. (Apparently, in such cases, the artery continues to return toward baseline throughout diastole.) Examples include severe aortic insufficiency; patent ductus arteriosus; and high-output cardiac failure, as in pregnancy, fever, anemia, thyrotoxic heart disease, and beriberi. If you suspect one of these entities, you should follow the American Heart Association’s current recommendation of recording all three pressures (i.e., systolic, muffling, and disappearance, e.g., 140/80/0). But which one should be “used”?

In aortic insufficiency, at least, we know that muffling best approximates the directly measured intra-arterial pressure when the indirect diastolic blood pressure is “zero” (Goldstein and Killip, 1962). As muffling was only 2 mm different from the directly measured pressure, we may assume that in other situations with an impossible “zero” diastolic pressure, muffling should also be used as the diastolic pressure.

The muffle point may also be used for the diastolic pressure in severe bradycardia in which the disappearance point may be nearly zero, although the high stroke volume is not accompanied by a high cardiac output (Goldstein and Killip, 1962).

I had originally intended to write an entire chapter on the confusion about the indirectly determined diastolic blood pressure, not only to clarify and convince, but also to show how the history of the subject is an example of how fuzzy thinking and sloppy scholarship can confound such a common activity. However, intra-arterial lines are now so ubiquitous that the student is advised to measure indirect blood pressures on such monitored patients and convince himself of the truth. Accordingly, I will simply cite a few examples of welldone experiments that demonstrate disappearance to be closer to the true diastolic pressure than muffling, except in situations of high stroke volume (Karvonen et al., 1964; London and London, 1967; Raftery and Ward, 1968).

The standard adult blood pressure cuff has a bladder of the correct size for arms whose brachial circumference is up to 27 cm. If the circumference is greater than 27 cm, the larger the arm, the greater the overestimation of both systolic and diastolic pressures if the indirect blood pressure is performed with a standard cuff (King, 1967). The pharmacologic treatment of normotensive patients who have “cuff hypertension” may actually increase their mortality (Kaplan, 1983).

Some authors have attempted to apply epidemiologic formulas to correct the systolic and diastolic pressures in patients with large
arms (Pickering et al., 1954). The epidemiologic corrections were derived from the experimental data (Ragan and Bordley, 1941), which were faulted in several ways, according to the original report. In any event, the correction factors are invalid in individual patients because of variability in pulse wave contour, radius of the brachial artery, brachial artery to biceps ratio, and so forth, all of which affect the exact degree of blood pressure overestimation from an undersized cuff.

Some workers (Devetski, 1963) have suggested wrapping the arm cuff around the forearm and taking the blood pressure at the radial artery, with the arm in a supinated position. (The mean arterial blood pressure at the radial artery is from 3 to 5 mm less than at the brachial.) My own preference is to obtain a very large cuff. One can use either a very wide (20 cm) cuff such as a thigh cuff or a very long (42 cm) bladdered cuff (King, 1967). Despite statements to the contrary, I know of no good evidence that the use of a large cuff in an adult can produce a spuriously low blood pressure. In fact, there is evidence to the contrary (Karvonen et al., 1964; King, 1967; Linfors et al., 1984; Montfrans et al., 1987; Nielsen and Janniche, 1974).

For the Attending. As a teaching exercise, pick a medical student with large biceps and take his blood pressure. Then, take some foam rubber padding, about 1 in. thick, and wrap it around the upper extremity to imitate the mechanical effect of fat. (Alternately, use a few Turkish towels, a lab coat, or any pliable material that resembles normal fat by being compressible.) Note that you have not actually altered the subject’s cardiovascular system. Place your blood pressure cuff around the foam rubber “fat” and take the blood pressure again. Remove the “fat” and take the blood pressure yet again. What is the effect of increasing the upper arm circumference on the indirectly determined blood pressure?

Admittedly, obesity tends to increase the blood pressure even beyond the effect of circumference. However, here we are concerned only with the effect of arm circumference on the accuracy of the indirectly determined blood pressure.

Many clinicians have very sharp cutoff points above which they treat the blood pressure and below which they do not. You can estimate from this experiment that a large number of people with fat arms are being treated for a disease, essential hypertension, which they in fact do not have.

I regretfully use the word “essential” to modify the noun “hypertension” for convenience in communicating. Along with Sir George Pickering, I do not believe that hypertension is a disease sui generis. However, if it is, I do not see how it can be “essential.”

Back at the turn of the century, arterial occlusive disease could only be diagnosed at autopsy. As some of these patients were found to have elevated blood pressures, as determined by the newly popular sphygmomanometer, some prestigious authorities hypothesized that the hypertension was secondary to the diffuse arterial occlusive disease and that it was an adaptive response, “essential” in that it provided a high head of pressure for driving the blood through the stenotic vessels. The ghost of this disproven hypothesis lingers among those who fear that lowering the blood pressure in malignant hypertension will deprive vital organs of blood flow.

As various specific etiologies of hypertensive states were discovered, the adjective “essential” came to refer to those forms of hypertension of unknown etiology, or more than 90% of cases. It should be distinguished from hypertension of unexamined etiology.

When this author was in medical school, considerable emphasis was placed on ruling out treatable causes of hypertension before consigning a patient to the “essential” category. These days, a minimal workup is recommended before starting therapy, but “more extensive testing for identifiable causes is not indicated generally unless BP control is not achieved” (Chobanian et al., 2003).

Oddly, the new adjectival meaning of “essential” has metastasized to a variety of other conditions. For instance, one reads of an entity called “essential migraine,” although one is hard pressed to explain to the migrainous patient exactly what is essential about his suffering. (Migraine itself is a useful word only if you know that it is derived from “hemicrania.” It then becomes a diagnosis that suggests itself when evaluating a patient with unilateral cephalalgia.)

Have you noticed how the text is changing? Words and concepts that you have not previously encountered are starting to spring up. The material is no longer oriented to the layperson. If you are a sophomore reading this in sequence, you are confronted with a problem that will continue for the next few years: what to do when you are no longer sure that you understand all the words. My advice is to stop and look up the strange words in a medical dictionary or textbook. Remember this is not a textbook of medicine. This is similar to a book on how to sail your boat; you still have to buy some charts in order to set your course.

On the wards, you will also hear residents and staff using words and phrases that you do not know. You can raise your apparent IQ by 10 points simply by carrying with you at all times a small notebook for recording all the strange words of the workday. Each night look up all the words you collected during the day.

The patient diagnosed with hypertension will be expected to have far more interaction with the medical system than he probably would like and might be expected to take expensive medication with troublesome side effects. Results are likely to be much better if the physician strives for concordance (see Chapter 2) rather than compliance.

The higher the patient’s blood pressure, the larger the benefits that the treatment offers him. For more moderate hypertension, the benefits of treatment are long-term and possibly detectable only over a large population. In selecting treatment goals with the patient, the physician should keep the effect of age and sex differences in mind, as well as the evidence in the literature concerning all-cause mortality, not just surrogate endpoints such as population-based blood pressure targets.

Although it is clear that hypertension is a risk factor for (i.e., is correlated with) cardiovascular disease, remember that it is a sign, not a disease (vide supra). While secondary hypertension may be uncommon, keep the possibilities in mind while performing
your history and physical examination, or you will miss them. The possibilities include endocrine disorders (such as pheochromocytoma or adrenocortical hyperfunction) or renal disease (including renovascular stenosis, see Chapter 18). Severe hypertension has even been described in a 9-year-old girl in association with signs of scurvy (see Chapter 7); all signs, including the hypertension, resolved with treatment with vitamin C (Weinstein et al., 2001). Be aware of possibly contributory factors such as obstructive sleep apnea (see Chapters 13 and 14) and hyperinsulinemia secondary to diet and obesity. If a patient becomes nonresponsive to a previously effective regimen, look for a supervening cause such as renovascular disease.

Even for “essential” hypertension, the mechanisms in the individual patient should be considered in prescribing therapy. Blood pressure is a function of cardiac output and peripheral resistance. If cardiac output rises to compensate for anemia, fever, beriberi, aortic valve defects, hyperthyroidism, or stiff arteries, the systolic pressure rises. Peripheral resistance is inversely proportional to the fourth power of the internal radius of the blood vessels (see Chapter 18), so tiny decreases in the lumen by atherosclerotic plaques will have a large effect on blood pressure.

Laragh distinguishes V-type (low renin) and R-type (high renin) hypertension. In the former, sodium retention and expanded plasma volume support increased cardiac output. In the latter, total peripheral resistance, set by the renin-angiotensin system, is high. V-type hypertension, predominantly found in black patients, responds well to sodium restriction and diuretics. R-type hypertension responds to beta-blockers and angiotensin converting enzyme inhibitors (Laragh, 2001). The history of previous response to various drugs is extremely important for choosing therapy that is effective and safe, especially if the patient presents with hypertensive crisis (vide infra) (Blumenfeld and Laragh, 2001).

In the current fad for uniform (“one size fits all”) treatment protocols, dietary sodium restriction is recommended for everyone—despite the fact that blood pressure is salt sensitive in only a subset of hypertensives and that the average decrease in systolic blood pressure 13 to 60 months after the initiation of salt restriction in 11 long-term randomized controlled trials was only 1.1 mm Hg (Hooper, 2002). The poor results might be explained by poor compliance and the natural history of hypertension, which tends to increase with time (T. Fagan, personal communication, 2009). Your patient may be in the 20% to 30% with low-renin hypertension who respond well to salt restriction; in the minority whose blood pressures may actually increase; or in the majority for whom this onerous intervention makes little difference.

Supervised home blood pressure monitoring, which can closely follow the effects of various habits, drugs, or interventions, helps to make the patient your ally. Always remember that your patient is Mrs Jones, not her blood pressure chart.

Postural hypotension refers to hypotension in the erect position relative to the recumbent position. The two main causes are volume depletion (due to anything from gastrointestinal hemorrhage to adrenal cortical insufficiency to diuretics) and neurogenic factors (e.g., due to certain antihypertensive medications; the various forms of autonomic vasomotor dysfunction; or even prolonged bed rest or weightlessness, as with the original astronauts). Less common causes include heart failure (the heart being unable to increase output when the patient stands) and pheochromocytoma (a rare disease in which hypovolemia compounds the problem caused by down-regulation of the noradrenergic receptors).