New Ways of Measuring Blood Pressure

Until relatively recently the measurement of blood pressure (BP) was very simple: a single technique (the auscultatory method) was used to provide a single pair of numbers (systolic and diastolic) to characterize an individual patient’s BP status. Although this method has served us very well for evaluating the risks of hypertension and the benefits of its treatment, it is becoming increasingly clear that this method gives an inadequate description of what is really going on. Blood pressure is a highly dynamic phenomenon, and varies in at least four different ways: first, the relationships between systolic and diastolic pressure are not constant; second, BP is continually changing over time; third, it varies according to the circumstances of measurement (in the medical setting or away from it); and fourth, the shape of the pressure wave changes as it passes through the arterial tree. There is a current flurry of research investigating these aspects of BP, all of which have the ultimate goal of improving the assessment of risk and effects of treatment of hypertension. New devices are available for evaluating the dynamic aspects of BP, with prices ranging from very cheap (about $50) to very expensive (about $20,000). Some of the major issues are reviewed below. Until relatively recently, diastolic pressure was considered to be more important than systolic, and many of the earlier intervention trials used it as the entry criterion for the study; but it is now clear that after the age of 50 years the predictive power of diastolic pressure is weakened.1 There is general consensus that, in older individuals, systolic pressure is a better predictor than diastolic, but there is an ongoing controversy as to whether pulse pressure may be better than systolic.2 The crux of this argument is whether a low diastolic pressure can be harmful. If it is, then pulse pressure must be a better predictor than systolic. Epidemiologic studies such as the Multiple Risk Factor Intervention Trial (MRFIT)3 have shown that in the presence of a very high systolic BP, the highest risk is seen in patients with the lowest diastolic pressures, and other studies have shown that a low diastolic pressure may be related to an increased risk of myocardial infarction but not stroke.2,4 The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7)5 makes no reference to pulse pressure as a risk factor, and in practical terms there is no obvious advantage for adopting guidelines based on pulse pressures. One of the problems here is that there is no accepted means of lowering systolic pressure while at the same time raising diastolic pressure, although this is theoretically possible with drugs that selectively reduce arterial stiffness. The inherent variability of BP has been acknowledged for many years to be one of the major limitations of the conventional clinic method of measuring BP. More than 40 years ago, Armitage and Rose6 showed that when sets of two readings were taken over a period of several months in healthy subjects, the average level could vary by 25 mm Hg or more, leading them to conclude that multiple readings needed to be averaged to obtain a valid measure of the true BP. Taking large numbers of readings at a single visit is impractical, although there are now automated devices that can take five or six readings in the clinic7; but this is easily done using home readings.8 A related issue is that there are systematic changes of BP throughout the day and night. These can be most easily assessed with 24-h ambulatory monitoring. Several prospective studies have shown that this technique gives a better prediction of individual risk than conventional clinic readings,9 but most of these have used the average daytime or 24-h average as the predictor, raising the issue that it may be the larger number of readings rather than the circumstances of measurement that leads to the improved prediction. Nevertheless, some studies have identified particular components of the diurnal rhythm of BP as superior predictors. These include the night-time pressure,10 dipping versus nondipping (usually expressed as the percentage difference between the daytime and night-time levels),11 the morning surge,12 and the short-term variability of BP.13 All of these are of potential interest but not yet sufficiently established to be adopted into routine practice. Two types of hypertension have been described in which the level of BP depends on the circumstances of measurement. In most patients the BP is consistently higher in the clinic than at other times, a phenomenon referred to as the white coat effect, which is conventionally defined as the difference between the clinic and the daytime pressures. White coat hypertension is defined as an elevated clinic pressure but normal daytime pressure, and is associated with a relatively benign prognosis14; whereas masked hypertension, defined as a normal clinic but elevated daytime pressure, is associated with a higher risk than predicted from the clinic pressure.15 Several methods have been used for recording the BP wave in the peripheral circulation and estimating either the central (aortic) BP, or measures of arterial stiffness, such as pulse wave velocity, or both.16 The rationale behind these procedures is that the increased stiffness of the conduit arteries that is related both to aging and hypertension results in increased pulse wave velocity, and hence acceleration of wave reflections from the peripheral to the central circulation. In younger subjects with compliant arteries the wave reflection is relatively slow, and the reflected wave impinges on the diastolic downslope of the incident wave and is thus hemodynamically neutral. In older subjects or patients with very stiff arteries (eg, patients with end-stage renal disease), the reflected wave returns earlier and augments the systolic peak of the incident wave, thus increasing the pulse pressure and the afterload on the heart.17 This phenomenon has two important consequences. First, the relationship between the BP measured at the brachial artery and the aortic pressure is not constant: in older patients or those with extensive arterial disease the aortic systolic pressure is similar to the brachial pressure, whereas in younger patients it is lower. The heart is one of the principal target organs affected by hypertension and is exposed to the central aortic pressure. Second, the relationship between central and peripheral pressure is affected differently by different types of antihypertensive drugs. A good example of this is the Losartan Intervention For Endpoint reduction in hypertension (LIFE) study,18 which compared the consequences of treating hypertensive patients with left ventricular hypertrophy using a regimen based on either atenolol or losartan. Although the changes in the brachial artery pressure during the trial were identical, losartan caused a greater regression of left ventricular hypertrophy than did atenolol, which may be explained by the greater reduction in central aortic pressure. More recently, the Anglo-Scandinavian Cardiac Outcomes Trial Conduct Artery Function Evaluation (ASCOT CAFE) study showed that the central aortic pressure was lowered to a greater extent by a treatment regimen based on amlodipine than by one using atenolol, whereas the effects on brachial artery pressure were not very different.19 In addition, the central pressure was a stronger predictor of clinical events than was brachial pressure. Exactly how these newer measures of BP will be incorporated into clinical practice remains unclear at the present time, but it is becoming increasingly clear that the traditional method of clinic measurement of BP gives an inadequate assessment not only of a patient’s risk of cardiovascular events but also of the effects of treatment. Some generalizations, however, can be made. First, the debate concerning the relative importance of pulse pressure versus systolic pressure is one issue that does not require any new technology for BP measurement, and is at present of little consequence, because there is no good way of selectively lowering pulse pressure with the currently available medications. Second, “out-of-office” BP measurement is going to play an increasingly important role. Ambulatory BP monitoring is gradually finding its way into clinical practice for the initial diagnosis of hypertension (particularly white coat hypertension), but it is poorly suited for the routine follow-up of BP changes. Home BP monitoring is inexpensive and easy to use, and offers the prospect of becoming the standard method for evaluating the response to antihypertensive treatment. Although its use has been endorsed by JNC 7 and other guidelines, it is not yet a part of the routine measurement of BP and is not reimbursed. This situation is likely to change in the near future. Third, the increased use of home and ambulatory BP monitoring will diminish the exclusive reliance on clinic measurements, but will not replace them in the foreseeable future. It is likely that in the near future physicians will be reimbursed according to the extent to which their hypertensive patients’ BP are below the target levels (ie, pay-for-performance), and these readings will almost certainly be based on clinic measurement. The better documentation of clinic BP will provide an incentive to use more objective measures such as automated devices that can take multiple readings (which have the additional advantage of also tending to be lower than clinicians’ readings). Fourth, the role of techniques such as the measurement of central aortic pressure is harder to define at present. It is possible that in the future diagnostic assessments and treatment decisions might be based on central rather than on brachial artery pressure, but this will require more evidence that such decisions lead to improved patient outcomes. In addition, the currently available devices are expensive and relatively difficult to use. One of the major public health issues related to hypertension is the poor degree of BP control despite the availability of many potent drugs. The gradual introduction of these newer methods of measuring BP will provide a much more precise estimate of how high a patient’s BP really is, which is the critical first step in improving the control of hypertension. Two aspects stand out as being of major importance, both related to the loss of faith in the traditional methods of measuring clinic pressure. The first is a need for more reliable and objective measures of clinic BP; mercury is on its way out and is going to be replaced by automated devices, which at any rate for the immediate future will use the oscillometric method. When insurance companies start to reimburse on the basis of pay for performance, it is likely that they will require objective documentation of the BP that is not susceptible to observer bias. The second major change will be the widespread adoption of home monitoring, not only for diagnosing hypertension but also for evaluating the response to treatment. A campaign, supported by several major organizations, is now underway to get home BP monitoring recognized as an essential part of the management of hypertension, as well as for reimbursement of the cost of the devices.