Journal of Human Hypertension (2004) 18, 531533 & 2004 Nature Publishing Group All rights reserved 0950-9240/04

$30.00
www.nature.com/jhh

REVIEW ARTICLE

Epidemiological, pathophysiological and clinical significance of systolic, diastolic and pulse pressure
DG Beevers
University Department of Medicine, City Hospital, Birmingham, UK

Long-term population surveys and studies of untreated hypertensive patients both strongly suggest that over the age of about 45, the height of the systolic blood pressure is a better predictor of cardiovascular outcome than the diastolic pressure. Research into echocardiographic left ventricular size, endothelial function and thromborheology also back up this supposition. In clinical practice, the treatment of isolated systolic hypertension has been shown to be very worthwhile. There remains some uncertainty as to the relative Keywords: systolic; diastolic; pulse pressure

significance of diastolic blood pressure when it is frankly high, as in severe systo-diastolic hypertension. However some clinicians and epidemiologists have questioned whether diastolic pressure is still worth measuring, given the overwhelming importance of the height of the systolic blood pressure at predicting outcome. Journal of Human Hypertension (2004) 18, 531533. doi:10.1038/sj.jhh.1001702 Published online 11 March 2004

Introduction
Up until about 1990, almost all clinical decisions in hypertension were made on the basis of the height of the diastolic blood pressure (DBP). The importance of systolic blood pressure (SBP) was not understood and clinical trials of the treatment of isolated systolic hypertension (ISH) were not available. This concentration on DBP did not however seem to be particularly logical. Early studies of insurance policy holders had strongly suggested that the height of the SBP was a better predictor of cardiovascular risk than the DBP.1 In 1971, an important paper from the Framingham Heart Study demonstrated that, over the age of 45 years at least, the height of the SBP was a better predictor of cardiovascular risk than the DBP.2 Below this age there was some uncertainty partly because younger people are at lower cardiovascular risk whatever their blood pressure might be so that the number of cardiovascular events that had occurred was too small for confident conclusions to be drawn. The main difficulty in calculating the relative power of systolic vs diastolic pressures in order to predict cardiovascular risk is that these two measures of pressure are themselves closely related. Up
Correspondence: DG Beevers, University Department of Medicine, City Hospital, Birmingham B18 7QH, UK. E-mail: gareth.beevers@swbh.nhs.uk Published online 11 March 2004

until about 1985, it was not readily possible to investigate the relative risk in relation to these two measures of pressure when also taking into account other variables contributing to mortality and morbidity including age, sex, body mass index, cigarette and pre-existing cardiovascular disease. In 1985, the Whitehall study, based in London, was able to demonstrate that while SBP and DBP both predicted risk significantly, when the predictive power of DBP was adjusted to take into account the concurrent SBP, this ceased to be statistically significant.3 By contrast, when SBP was investigated, after adjusting for DBP, it remained a significant predictor of cardiovascular outcome. More recently, a major reanalysis of the Framingham Heart Study with longer follow-up data and more cardiovascular events showed that for all levels of SBP, including those within the so-called normal range, the height of the pressure accurately predicted coronary heart disease.4 By contrast, the relationship between DBP and cardiovascular risk, for a given level of SBP, was reversed, that is, the lower the diastolic pressure, the greater the mortality. Thus, a blood pressure of 170/ 70 carries a higher coronary heart disease hazard ratio than a blood pressure of 170/100. These data therefore strongly suggest that the pulse pressure (SBP minus DBP) is also a major independent predictor of coronary heart disease. In the original 1971 publication, pulse pressure was numerically but not significantly a more powerful predictor of risk than the SBP.

Systolic, diastolic and pulse pressure DG Beevers 532

In the Framingham Heart Study, taking all hypertensive patients, with SBPs greater than 160 mmHg and/or DBPs below 95 mmHg, ISH (4160/o95 mmHg) was present in 57.4% of men and 65.1% of women over the age of 65 years.5 More conventional systodiastolic hypertension (SDH) (4160/495 mmHg) was present in 30.3% of men and 27.7% of women. Isolated diastolic hypertension (IDH) (o160/495 mmHg) was present in 12.4% of men and 7.1% of women in this age band. Thus, ISH is the commonest form of hypertension in older people. The measurement of blood pressure in clinical and epidemiological practice is open to problems of the white coat effect, where some participants may have blood pressures that are falsely raised, related to the stress and unfamiliarity with clinical examination. Some of these problems can be reduced by 24 h ambulatory blood pressure monitoring (ABPM). A recent study of 110 hypertensive patients who had blood pressures measured in clinical practice and also by 24 h ABPM showed that the Gensini score of coronary artery atheroma and the mean internal carotid artery intima-media thickness was more closely related to the 24 h pulse pressure than to any other parameter of blood pressure on ABPM or in clinical practice.6 Several clinical trials have been conducted in which the value of treating ISH was investigated. In addition, several other trials included some patients with ISH as well as those with conventional SDH. A recent analysis of the placebo or control groups in these trials among hypertensive patients who received no treatment clearly demonstrates that for a given level of SBP the risk was greater when the DBP was low.7 Thus, a blood pressure of 200/68 carried a greater risk of death than a blood pressure of 200/92. Several studies have shown that the degree of left ventricular hypertrophy (LVH) on ECG or echocardiogram correlates more closely with the 24 h blood pressure than with using clinical blood pressure reading and that, numerically speaking, the 24 h SBP and pulse pressure was a greater predictor of risk than the 24 h DBP.8 In the LIFE study, echocardiographic LV mass in relation to body surface area was the same in patients with ISH and patients with SDH, strongly suggesting that the DBP did not matter.9 Similarly, in a study from Birmingham, patients with ISH (average blood pressure 193/82) were compared with patients with SDH (average blood pressure 199/112).10 The proportion of patients who had echocardiographic LVH was 77% in the ISH group and 65% in the SDH group. This study also showed that patients with ISH and those with SDH broadly share the same abnormalities of endothelial function (von Willebrand factor), haemorheological abnormalities (plasma fibrinogen, plasma viscosity and haematicrit) and markers of thrombogenesis (plasminogen activator inhibitor PAI1, fibrin D-dimer and soluble P-selectin).10 Hypertension is associated with a
Journal of Human Hypertension

prothrombotic tendency and this is found equally in patients with ISH and those with SDH. Turning to the value of antihypertensive treatment in patients whose diastolic pressures were not raised but who had ISH, the SHEP study and the SYST-EUR trial both showed highly significant reduction in stroke rates and reductions of approximately 20% in coronary heart disease rates.11,12 The reduction in coronary heart disease was not statistically significant in the SYST-EUR trial mainly due to small numbers.12 Pooling the results of all the patients with ISH included in the SHEP and SYSTEUR studies as well as the other studies of the treatment of hypertension where some patients with ISH were included, demonstrates a significant reduction in total mortality, cardiovascular mortality, cardiovascular vents, fatal or nonfatal stroke and fatal or nonfatal coronary heart disease.7 This overview analysis was the first to show convincingly that lowering blood pressure also reduces all-cause mortality. An early attempt in the Glasgow Blood Pressure Clinic to investigate the relative power of systolic and diastolic pressure showed that these two parameters at follow-up were equally predictive of mortality.13 The problem with these early studies is that in those days patients with ISH were not included as there was then no trial evidence of benefit from drug treatment. Therefore, the study was not in a strong position accurately to investigate the relative significance of SBP and DBP in treated patients. The epidemiological, clinical and pathological significance of ISH is so great that this has led some investigators to consider that maybe we should even discontinue measuring DBP and simply concentrate all clinical decisions on the height of the SBP.14

Isolated diastolic hypertension

There remains the unresolved question of the significance of IDH where the SBP is not raised, but the DBP is greater than 90 or 95 mmHg. The problem with such studies is that they are difficult to perform. Patients with IDH are usually young and are therefore at low cardiovascular risk. In some cases, the measurement of DBP may have been inaccurate due to observer problems and the subjectivity of diastolic endpoint sounds. Furthermore, IDH is relatively uncommon and it is difficult therefore to be sure of its significance in population studies. However, a recent study from Finland reported a follow-up over 32 years of 3267 healthy men aged 3045 years.15 Subjects whose blood pressures were less than 160 and less than 90 mmHg were given an arbitrary relative risk of 1.0. As expected, patients with SDH and ISH both had approximately double the relative risk. Taking the patients with IDH, those whose SBP was below 140 had a relative risk of 1.12 for all-cause mortality and

Systolic, diastolic and pulse pressure DG Beevers 533

this was not significantly greater than the normotensive group. Examinees with IDH with SBPs between 140 and 159 had a relative risk of 1.50, which was significantly greater than the normotensive group. Thus, one can conclude that mild IDH probably does not carry a great risk, but IDH with SBPs greater than 140 is a hazard to health. The revolution in our thinking must make one speculate as to why clinicians have been so reliant on the DBP alone until the last 12 years. Were we taught wrongly? The answer would appear that we were indeed taught wrongly but there is a reason for this. Up until about 1970, there were insufficient long-term follow-up studies of well-examined unselected populations. The Framingham project was started in 1950 and therefore reliable data could not become available until well into the 1970 s. More importantly, the statistical techniques (including Coxs proportional hazard model) were not readily available until the mid-1970s.16 In addition however, the computer hardware and software to investigate the relative power of SBP and DBP were not available until more recently. Clinicians observed that SBP is highly variable, with many people showing quite dramatic reductions in blood pressure on rechecking. DBP, by being less variable, appeared therefore to be a more reliable measure of the blood pressure load. There remain some areas of uncertainty. It is still not known whether SBP over-rides DBP in patients with severe hypertension. Thus, we do not yet know whether a blood pressure of 220/140 carries a different prognosis to a blood pressure of 220/100. It is possible that DBP starts to become an independent risk factor when it is high, for example, over 110 mmHg.17 One could speculate that because in population surveys and in hypertensive patients lower DBPs have a poorer prognosis, one should try to raise DBP while bringing the SBP down. There are as yet no pharmacological techniques to do this, and the overriding priority must be to reduce the SBP and thus prevent heart attacks and strokes. The change in emphasis from DBP to SBP that has occurred over the last 10 years has been evidence based.18,19 It is, however, probable that many doctors and nurses who qualified more than 10 years ago will be unaware of this information and will tend therefore to ignore raised SBPs when the DBPs are not raised. In view of the impressive benefits of treating such patients, a major educational exercise is necessary. It is interesting to note that in about 1990, the great epidemiologist Professor Geoffrey Rose once mused quietly that one sometimes wishes that Nicolai Korotkoff had not described the diastolic sounds (GA Rose, 1990, personal communication).

References
1 May O, Hyperpiesia IV. Mortality in relation to hyperpiesia. Br Med J 1925; ii: 11661167. 2 Kannel WB, Gordon T, Schwartz MJ. Systolic versus diastolic blood pressure and risk of coronary heart disease. Am J Cardiol 1971; 27: 335345. 3 Lichtenstein MJ, Shipley MJ, Rose GA. Systolic and diastolic blood pressure as predictors of coronary heart mortality in the Whitehall Study. Br Med J 1985; 291: 243245. 4 Franklin SS et al. Is pulse pressure useful in predicting risk for coronary heart disease? The Framingham Heart Study. Circulation 1999; 100: 354360. 5 Wilking SVB et al. Determinants of isolated systolic hypertension. JAMA 1988; 260: 34513455. 6 Zakopoulos NA et al. Pulse pressure in normotensives, a marker for cardiovascular disease. Am J Hypertens 2001; 14: 195199. 7 Staessen JA et al. Risk of untreated and treated isolated systolic hypertension in the elderly: meta-analysis of outcome trials. Lancet 2000; 355: 865872. 8 White WB, Dey MH, Shulman P. Assessment of the daily blood pressure load as a determinant of cardiac function in patients with mild to moderate hypertension. Am Heart J 1989; 118: 782795. 9 Papademetriou V et al. Similar effects of isolated systolic hypertension and combined hypertension on left ventricular geometry and function. The LIFE Study. Am J Hypertens 2001; 14: 768774. 10 Lip GYH, Blann AD, Beevers DG. Prothrombotic function, endothelial function and left ventricular hypertrophy in isolated systolic hypertension compared with systo-diastolic hypertension. J Hypertens 1999; 17: 12031207. 11 SHEP cooperative Research Group. Prevention of stroke by antihypertensive drug treatment in older persons with isolated systolic hypertension. JAMA 1991; 265: 32553264. 12 Staessen JA et al. Randomised double-blind comparison of placebo and active treatment for older patients with isolated systolic hypertension. Lancet 1997; 350: 757764. 13 Beevers DG, Davies P, Johnston JH, Larkin H. The relative importance of systolic, diastolic and mean arterial pressure treated hypertensives. Br J Clin Pract, Symp Suppl 1982; 15: 15. 14 Sever PS. Abandoning diastole. Br Med J 1998; 318: 1773. 15 Strandberg TE et al. Isolated diastolic hypertension: pulse pressure and mean arterial pressure as predictors of mortality during a follow-up of up to 32 years. J Hypertens 2002; 20: 399404. 16 Cox DR. Regression models and life tables. J R Stat Soc B 1972; 34: 187220. 17 He J, Whelton PK. Elevated systolic blood pressure or a risk factor for cardiovascular and renal disease. J Hypertens 1999; 17 (Suppl 2): S7S13. 18 Black HR. The paradigm has shifted to systolic blood pressure. Hypertension 1999; 34: 386387. 19 Swales JD. Systolic versus diastolic pressure: paradigm shift or cycle? J Human Hypertens 2000; 14: 477479. 20 Rose GA. Personal Communication, 1990.

Journal of Human Hypertension