Pulsatile Versus Steady Component of Blood Pressure: A Cross-sectional Analysis and a Prospective Analysis on Cardiovascular Mortality

Transcription

1 392 Pulsatile Versus Steady Component of Blood Pressure: A Cross-sectional Analysis and a Prospective Analysis on Cardiovascular Mortality Bernadette Darne, Xavier Girerd, Michel Safar, Frauds Cambien, and Louis Guize Studies on the prognostic significance of blood pressure on cardiovascular disease have essentially investigated the levels of diastolic or systolic blood pressure. However, blood pressure may also be divided into two other components: steady (mean arterial pressure) and pulsatile (pulse arterial pressure). The relations of these two components with cardiovascular risk factors and cardiovascular mortality were investigated in 8,336 men and 9,35 women aged years, who were followed up for a mean period of 9.5 years. However, the interpretation of the relations is complicated by the strong correlation existing between these two components. A principal component analysis was performed to obtain two independent parameters: a steady and a pulsatile component index, strongly correlated with mean and pulse arterial pressure, respectively. In the cross-sectional analysis, relations were stronger with the steady component index than with the pulsatile component index; an association was found between left ventricular hypertrophy and the pulsatile component index in both sexes. The survival analysis was not performed in women under 55 as only cardiovascular deaths occurred in this group. The steady component index was a strong prognostic factor of all types of cardiovascular death in both sexes. In women, the pulsatile component index was positively correlated to death from coronary artery disease and inversely correlated to stroke. In conclusion, the steady component of blood pressure is a strong risk factor for cardiovascular death in both sexes; the pulsatile component could be a risk factor independent of the steady component in women older than 55 years. (Hypertension 989; 3: ) Many epidemiological studies have shown a close relation between the level of blood pressure and cardiovascular diseases. - 0 The respective roles of the systolic blood pressure (SBP) and diastolic blood pressure (DBP) have also been investigated, - 6 and it has been suggested that DBP is more strongly related to cardiovascular diseases before age 45; whereas SBP is more strongly related to cardiovascular diseases after age 45. Blood pressure is a periodic phenomenon that can be divided into two components: a steady component and a pulsatile component. 78 The steady component is determined exclusively by two hemo- From the INSERM U 258, Hdpital Broussais (B.D., F.C.), Investigations Prt-Cliniques (IPC) (B.D., L.G.), and Centre de Diagnostic, Hdpital Broussais (X.G., M.S.), Paris, France. Supported in part by research grants from the F&Je'ration Franchise de Cardiologie, the Laboratoires Merck-Clevenot, the INSERM (486007), and by the Caisse Nationale d'assurance Maladie. Address for reprints: Dr Bernadette Dame\ INSERM U 258, Hdpital Broussais-96, rue Didot 7504 Paris, France. Received July 6, 988; accepted December 3, 988. dynamic measurements: cardiac output and vascular resistance. The pulsatile component represents the variations of the pressure curve around the steady component and is influenced by other hemodynamic mechanisms: the changes in ventricular ejection and large artery compliance and timing of reflected waves. Usually the steady component is estimated by the mean arterial pressure [i.e., DBP+/3(SBP-DBP)], and the pulsatile component is estimated by the pulse pressure, which is the difference between the SBP and the DBP. It is well established that for the same level of mean arterial pressure, different patterns of pulse pressure may be observed according to the level of ventricular ejection and distensibility. Little has been done to investigate the respective roles of these two components of blood pressure in the risk of cardiovascular disease. The aim of this analysis is to investigate the association of the steady and the pulsatile components of blood pressure with parameters correlated to blood pressure and risk factors for cardiovascular disease and to investigate their prognos-

2 Dame et al Steady and Pulsatile Pressure Components 393 tic significance on cardiovascular mortality in subjects who underwent a checkup between 972 and 977. Mean arterial pressure and pulse pressure are highly correlated; thus, their respective associations with other factors are difficult to interpret when both are introduced in the same model. Therefore, to investigate the roles of the steady and pulsatile components of blood pressure, two new variables were generated by performing a principal component analysis on SBP and DBP. By construction, these two new variables are uncorrelated but are independently highly correlated with mean arterial pressure and pulse pressure. Materials and Methods The population of this prospective study was composed of 8,336 men and 9,35 women aged years who were bora in France. Clerks and working executives represented 88% of men and 60% of women; 26% of women were housewives. The subjects agreed to undergo a health checkup between May 972 and May 977 in the Investigations Pr6-Cliniques (IPC) Check-up Center (Paris, France). Subjects who were receiving antihypertensive drug treatment were excluded from the analysis. Supine blood pressure was measured on the right arm with a manual sphygmomanometer by a nurse after a 0-minute rest period. The first and fifth KorotkofTs sounds were used to define systolic and diastolic pressures. Only one measurement was obtained. A standard 2-lead electrocardiogram was performed on each subject and interpreted by a cardiologist. Heart rate was coded from the electrocardiogram into four classes: <60, 60-80, 80-00, and >00 beats/min. Heart rate was not recorded for the 9,502 subjects who were examined before April 974. The diagnosis of left ventricular hypertrophy was defined as the presence, on electrocardiogram, of two or more of the following criteria: Sokolow (S vl +greatest of R^RyJ >35 mm, intrinsecoid deflection >0.05 seconds in V 6, QRS axis between 0 and -90, Lewis index: (R I -R in )+ (S U SL)>7 mm. 22 Weight and height were measured on each subject. Body mass index at entry was computed as weight on height squared. The examination included a self-administered questionnaire with dichotomic (yes or no) questions about tobacco consumption. Blood samples were collected 45 minutes after a 50 g oral glucose load was administered to subjects not reporting diabetes mellitus and without a glucose load for the others (,50 men and 489 women). Subjects were not requested to be in a fasting condition on the day of examination. Uricaemia, uremia, glycemia, and cholesterolemia were measured on a Technicon SMA2 (London, England). The study period ended on December 3, 984. The mean follow-up period was 9.5 years. At the end of the study period, an inquiry was made at the city hall of the birth town of each subject concerning their vital status (subjects lost to follow-up because of an incomplete address of the city hall represented less than 3% of the study population). Any information on deceased subjects relevant to the cause of death was collected, including all medical records obtained from hospital departments, family doctors, or relatives. This information was then reviewed by a medical committee;,5 deaths occurred amgng men and 29 among women. The cause of death remained unknown for 57 men and for 5 women. Cardiovascular disease was the cause of death in 305 men and 43 women and included coronary death, heart failure of any origin, stroke, and sudden death. Coronary death was assessed as myocardial infarction or sudden death in a subject who had an episode of chest pain in the 6 hours preceding death or in a person known to have atherosclerotic coronary disease. Stroke was assessed as either atherosclerotic or hemorrhagic cerebrovascular disease or sudden death with a suggestive syndrome. Statistical Analysis Statistical analysis was performed with SAS software. Two independent analyses were performed in men and women. Mean arterial pressure and pulse pressure were computed from SBP and DBP, mean arterial pressure was defined as DBP+/3 (SBP- DBP)i623 an(j pui se pressure as SBP-DBP. Pearson correlation coefficients between mean arterial pressure and pulse pressure were 0.42 (p<0~ 3 ) in men and 0.52 (p<0~ 3 ) in women. To obtain two independent factors, a principal component analysis was performed on SBP and DBP from their covariance matrix (PROC PRINCOMP). The principal component analysis is a multivariate analysis technique that uses a set of linear transformations of the variables studied to create new variables called the principal components. 24 The principal components have the following properties: the principal component variables are uncorrelated; the first has the largest variance of any linear function of the original variables, and the second has the second largest variance, and so on. When a principal component analysis is performed on two variables, the newly generated variables are a weighed sum and a weighed difference of the original variables. Figure shows the geometric interpretation of these two components. Simple Pearson correlation coefficients were computed for quantitative parameters (PROC CORR). A test of a quadratic effect was effected with PROC RSREG. An analysis of variance was performed to assess the relation between smoking, heart rate, left ventricular hypertrophy, and the dependent variables (PROC GLM). After the exclusion of subjects who died of unknown causes, a survival analysis using a Cox model was performed (PROC PHGLM). The survival analysis was not performed in women under 55 since only cardiovascular deaths were observed in this age group. The analysis of cardiovascular

3 394 Hypertension Vol 3, No 4, April 989 6RP 22( ( 80 no GC ISO nn 3d 20 FIGURE. Geometric representation of steady component index (SCI) and pulsatile component index (PCI). To simplify, a transformation of SCI and PCI was performed. SBP, systolic blood pressure; DBP, diastolic blood pressure BO so ido 30 MO OBP mortality was performed with the principal components after adjustment on age. Results Principal Components Analysis The first factor, obtained by the principal component analysis, accounted for 9% of the total variation of SBP and DBP in men and 92% in women. This factor, conventionally called steady component index (SCI), was highly correlated to the "classical" mean arterial pressure, r=0.98, p<0~ 3 in both sexes (in men, SCI=0.85 SBP+0.52 DBP and in women, SCI=0.87 SBP+0.49 DBP). The second factor was conventionally called pulsatile component index (PCI): in men, PCI=0.52 SBP-0.85 TABLE. Simple Pearson Correlatkin Coefficients Between Blood Pressure Parameters Variable SBP DBP Mean arterial pressure* Pulse pressure! SCI PCI DBP and in women, PCI=0.49 SBP-0.87 DBP, and its correlation coefficient with pulse pressure was 0.80 (p< 0" 3 ) and 0.74 (p<0" 3 ) in men and women, respectively. Since the principal component analysis was performed independently in men and women, the mean values of SCI and PCI cannot be compared between sexes. Table gives the correlation coefficients of these two factors with SBP, DBP, mean arterial pressure, and pulse pressure in both sexes. The SCI was highly correlated with SBP and DBP, whereas PCI was correlated with these parameters to a lesser degree. The PCI was highly negatively correlated with DBP; the correlation of PCI with SBP was much smaller than that of pulse pressure. In both SBP DBP 0.% Mean arterial pressure Pulse pressure p<0" 3 for all coefficients. SCI, steady component index; PCI, pulsatile component index; SBP, systolic blood pressure; DBP, diastolic blood pressure. Mean arterial pressure is DBP+/3 (SBP-DBP). tpulse pressure is SBP-DBP.

4 Dame et al Steady and Pulsatile Pressure Components 395 PCI 3 FIGURE 2. Plot showing distribution of pulsatile component index (PCI) (mean±.96 SD) according to age for men. *. + 4 * ' ROE 60 6 sexes, SCI was linearly correlated with age (r=0.20 [p<0.0~ 3 ] in men, r=0.34 [p<0~ 3 ] in women). Figure shows the regression of PCI on age; the relation is not linear. An adjunction of an agesquared term in the regression model improved the fit (test of a quadratic effect F=50 [p<0" 3 ] in men and F=52 [p<\0~ 3 ] in women). In both sexes PCI appeared to be independent of age before 55 years and to increase linearly with age after 55 (Figures 2 and 3). Therefore, a further analysis was made independently in two age groups: less than 55 years, greater than or equal to 55 years, and the analysis PCI -2-3 was adjusted for age in subjects over age 55. Since SCI and PCI were computed independently of age, the independence of the two components cannot be assumed after subdivision on age; however, the correlation between SCI and PCI remains small. Correlation between SCI and PCI was 0. in men before age 55 and 0. after age 55 and 0.4 in women before age 55 and 0.08 after age 55. Cross-sectional Analysis Table 2 indicates the number of subjects, mean and standard deviation of blood pressure, body FIGURE 3. Plot showing distribution of pulsatile component index (PCI) (mean±.96 SD) according to age for women. 65 HOE

5 3% Hypertension Vol 3, No 4, April 989 TABLE 2. Clinical and Biological Parameters Measured at Entry into Study in and, According to Age Variable years a55 years years Subjects (n) SBP (mm Hg) DPB (mm Hg) Mean arterial pressure (mm Hg) Pulse pressure (mm Hg) SCI (mm Hg) PCI (mm Hg) BMI (kg/m 2 ) Cholesterolemia (g/) Uremia (g/) Uricemia (mg/) Glycemia (g/) 2, (6)* 84(2) 00(3) 50(0) 57 (9) -.83(6.32) 25.2 (2.9) 2.2 (0.36) 0.34 (0.08) 6.3(0.9).67(0.4) 5,599 4 (20)* 87(3) 05 (4) 54(3) 65 (23) -0.23(7.0) 25.6 (3.2) 2.24 (0.35) 0.36 (0.08) 6.8(.7).74(0.44) 5,869 28(5) 80() 96() 48(9) 5 (7) (5.67) 23.2 (3.3) 2.2(0.36) 0.30 (0.07) 45.7 (0.39).66(0.40) 5:55 years 3, (20) 85 (2) 03 (4) 55 (3) 63 (22) (6.75) 24.2 (3.5) 2.4 (0.38) 0.35 (0.08) 5.0(0.4).79 (4.27) SBP, systolic blood pressure; DBP, diastolic blood pressure; SCI, steady component index; PCI, pulsatile component index; BMI, body mass index. * Values are mean±sd. mass index, and biological measurements in both sexes, in each age group at entry in the study. A SBP greater than or equal to 60 mm Hg or a DBP greater than or equal to 95 mm Hg was found in 4,029 (22%) men and,270 (4%) women. Simple Pearson correlation coefficients of SCI and PCI with body mass index and biological measurements are given in Table 3. SCI was correlated with body mass index and glycemia. In both sexes, correlation coefficients of PCI with the other parameters were quite small or not significant. As shown in Table 4, SCI was slightly associated with cigarette consumption in both sexes. In men, PCI expanded with cigarette consumption. This increase was statistically significant in men under 55 years and, after adjustment on age, in those who TABLE 3. Simple Pearson Correlation Coefficients Variable Age BMI* Glycemiat Cholesterolemia Uremia Uricemia Age (yr) a:55 & were 55 years or older. In women, PCI was larger among those smoking more than 20 cigarettes a day, but that increase was not statistically significant. As shown in Table 5, 230 (.7%) men and 5 (0.5%) women showed electrocardiographic evidence of left ventricular hypertrophy. The mean SCI was increased in all subjects with left ventricular hypertrophy. The mean PCI was also higher among subjects with left ventricular hypertrophy than among those without hypertrophy in each age category and in both sexes. This difference was not statistically significant among men under age 55 (p=0.08). SCI increased with heart rate (Table 6) in both sexes and each age category, whereas an increase of PCI with heart rate was observed only in younger men. SCI (NS) (NS) t SCI, steady component index; PCI, pulsatile component index; BMI, body mass index. *Body mass index is weight/height 2, tafter adjustment on glucose ingestion (NS) $ * (NS) (NS) -0.0 (NS) (NS) PCI (NS) $ $ (NS) 0.0 (NS) -0.04$ (NS) Other values prslo

6 Dame et al Steady and Pulsatile Pressure Components 397 TABLE 4. Variable Subjects (n) SCI (mm Hg) PCI (mm Hg) Comparison of Steady Component Index and Pulsatile Component Index According to Smoking Status Age (yr) =55 Never smoked 3,27,349 4,50 2,84 57 (9)* 64 (22) 5 (7) 64 (22) -.99(6.4) (6.8) -7.5(5.63) (6.83) Ex-smoker or smoking <9 c./day 4,353 2, (9) 66 (23) 49 (7) 60 (20) -.96(6.37) -0.7(7.3) (6.0) (6.74) Smoking 9-20 c./day 2,468, (9) 64 (22) 49 (9) 57 (9) -.84(6.29) -0.0 (7.07) (5.68) (6.28) c, cigarettes; SCI, steady component index; PSI, pulsatile component index. Values are mean±sd. tafter adjustment for age F=4.0; p= Smokinga20 c./day or > i pouch/day 2, (9) 65 (24) 49 (8) 57 (9) -.4(6.48) (6.99) (5.33) (5.60) Variance ratio F=5.83 F=2.30 F=6.65 F=.54 F=5.23 F=2.50 F=.94 F=0.55 Probability value p<0" 3 p=0.08 p<0" 3 p=0" 4 p=0.002 p=0.06t p=0.3 p=0.66 Longitudinal Analysis The annual death rate per thousand was six in men and three in women. Table 7 gives the number of cardiovascular deaths observed in each age and in each sex category and presents the results of the analyses of the associations between SCI, PCI, and mortality from cardiovascular deaths, myocardial infarction, and stroke after adjustment for age. The number of cardiovascular deaths was higher among men than women. SCI was a strong prognostic factor of cardiovascular death, myocardial infarction, and stroke in both sexes. In women, PCI was positively correlated to myocardial infarction (after adjustment for age 0=0.09 [SD=0.03] p<0" 3 ) and inversely correlated to stroke (after adjustment for age /3=-0.07 [SD=0.03] p=0.05). PCI was not a prognostic factor of total cardiovascular deaths. Discussion This analysis was based on the premise that the pulsatile component of blood pressure could be a risk factor independent of the steady component. Studies on biomaterials have shown that, whatever the mean pressure, the higher the pulsatility the faster the damage of the biomaterial. 725 Furthermore, a study on patients with arteriosclerosis obliterans of the lower limbs has shown that, after adjustment for mean arterial pressure, an elevated pulse pressure was associated with the severity of the intermittent claudication. 26 In populations, a TABLE 5. Steady Component Index and Pulsatile Component Index and Electric Left Ventricular Hypertrophy Variable Age (yr) non-lvh LVH p value Subjects (n) SCI (mm Hg) PCI (mm Hg) 2,627 5,473 5,855 3, (9)* 64 (22) 5(7) 63 (22) -.84(6.3) (7.00) (5.66) -5.5 (6.67) (30) 86 (36) 96 (26) 9 (3) (7.38) 2.95 (0.22) (7.96) -0.8(0.3) LVH, left ventricular hypertrophy; SCI, steady component index; PCI, pulsatile component index. 'Values are mean±sd. P=IO- 4 p=0" 4 P<IO- 3 P=IO- 4 p=0.08 p<0" 3 p<0' 3 p=0.002

7 398 Hypertension Vol 3, No 4, April 989 TABLE 6. Steady Component Index and Pulsatile Component Index and Heart Rate < Variable Age (yr) (beats/min) (beats/min) (beats/min) 2=00 (beats/min) Variance ratio Probability value Subjects (n) SCI (mm Hg) PCI (mm Hg) &55 s=55, (6)* 59 (20) 43 (5) 58 (20) (5.9) -0.0 (6.55) (5.36) (6.22) 5,424 2,296 2,96, (8) 65 (2) 50(6) 62 (20) -2.03(6.4) (6.90) (5.4) (6.5), (20) 72 (23) 56(7) 69 (24) -.46(6.53) 0.4(6.48) (5.98) (6.63) (23) 76 (23) 66(2) 78 (24) (7.44) -0. (8.02) (6.70) (7.37) F=58 F=35.8 F=06 F=3.3 F=6.5 F=4.68 F=0.42 F=0.79 p=0-4 p=0-4 P=IO- 4 P=IO-" p= p=0.003 p=0.74 p=0.50 SCI, steady component index Values are mean±sd. PCI, pulsatile component index. wide range of pulse pressure is observed for a given mean arterial pressure. However, mean arterial pressure and pulse pressure are strongly correlated: the higher the mean arterial pressure, the higher the extent of the pressure oscillations around the mean Thus, the present report did not examine the prognostic significance of mean arterial pressure and pulse pressure, per se. Instead, a principal component analysis was performed that led to two uncorrelated components: a weighed sum and a weighed difference of SBP and DBP. Interestingly, the weights for both factors are almost the same in TABLE 7. Variable Death (n) SCI (mm Hg) PCI (mm Hg) men and women and similar values were also obtained when using the data of the Paris Prospective Study (unpublished results). Thus, if a relation is shown between a parameter and PCI, this relation is independent of SCI. The purpose of this analysis was to determine whether variables, which have been shown to be related to the level of blood pressure and hypertension, are related to SCI only or to PCI only or to both. The results demonstrate that SCI and PCI are not correlated with the same factors and that the relation between PCI and the studied factors varies according to age and sex. Cardiovascnlar Mortality B Coefficient of Cox Model Adjusted on Age and Standard Deviation Age (yr) Total cardiovascular mortality * (0.008)$ 0.06 (0.003) 0.02 (0.005) 0.03 (0.05) 0.04 (0.00) 0.09 (0.02) SCI, steady component index; PCI, pulsatile component index. *fi (standard deviation). Myocardial infarction death (0.005) 0.04 (0.004) (0.009) (0.020) (0.05) (0.027) Stroke death (0.008)$ 0.09 (0.005) 0.02 (0.008)t (0.037) (0.07) (0.034)t tp^o.ol p=s0.00.

8 Dame et al Steady and Pulsatile Pressure Components 399 Possible Methodological Bias The limitation of the results is related to the possible bias introduced by the blood pressure measurement, as previously reported. 3 The fact that a single blood pressure measure was performed on each subject may explain the relatively high percentage of subjects appearing to be hypertensive in this population. The main consequence of performing a single blood pressure reading results in a loss of power in the analysis. However, even if a series of measurements does improve the predictability, it has been shown that a single blood pressure measurement in a group of subjects predicts which individuals are more likely to develop cardiovascular diseases Cross-sectional Analysis The present study has demonstrated a linear relation between SCI and age, and a nonlinear one between PCI and age. The increase in PCI with age, observed in this population only after the age of 55, may be explained partly by the greater increase in SBP than DBP with age after 55 years. 303 The mechanisms involved to explain this phenomenon are an increase in peripheral resistance (probably from shrinkage of the vascular bed) and an increase in arterial stiffness Despite an increase in vascular resistance, DBP may remain constant or even fall if arterial stiffness increases to a relatively greater degree than resistance The increase in arterial stiffness observed with advancing age contributes to the explanation of the higher PCI observed in the older subjects in this study. Otherwise, a clinical investigation in hypertensive subjects has shown that factors correlated to pulse pressure vary with age; increased pulse pressure for a given mean arterial pressure is related to increased ventricular ejection in younger subjects, to increased arterial stiffness in older subjects, and to a combination of these factors in middle-aged subjects In the present report, for all factors studied, the correlation was stronger with SCI than with PCI. Although the steady component of blood pressure is a main factor of left ventricular hypertrophy, this study also shows that the pulsatile component might also be a risk factor of left ventricular hypertrophy, independent of the steady component. The results of clinical and experimental studies suggest that the pulsatile component of blood pressure is related to the severity of cardiac hypertrophy. 34 Furthermore, a study on middle-aged subjects with essential hypertension has shown that, for a given mean arterial pressure, pulse pressure is correlated to the cardiac mass measured on echocardiography. 35 Other physiological studies have shown that the heart load does not depend solely on the SBP or DBP levels but also on the other components of the impedance spectrum, especially arterial stiffness and distensibility A decrease in arterial distensibility is associated with an increase in pulse pressure The fact that, in the present investigation, left ventricular hypertrophy was more strongly related to PCI in the group above age 55 seems to confirm that an increase in arterial stiffness is one of the main factors associated with the increased pulsatility observed in subjects with left ventricular hypertrophy. Longitudinal Analysis In the present study, the steady component of blood pressure was a strong cardiovascular risk factor in both sexes, whereas the pulsatile component appeared to be a risk factor only in women. This last result must be considered with caution because of the small number of cardiovascular deaths observed in women. In the Chicago studies, 3 a parameter computerized as rescaled SBP+DBP, appeared to be a better prognostic factor of cardiovascular disease in both sexes than SBP or DBP alone. Relative risks were similar for mean arterial pressure, SBP, and DBP for stroke and coronary artery disease in the Multiple Risk Factor Intervention Trial conducted in men years old 39 ; this study also showed that, for all levels of DBP, an increase in SBP and therefore an increased pulse pressure is a risk factor of cardiovascular mortality. A study in Framingham 4 has shown that mean pressure is a better risk factor predicting coronary artery disease in men; whereas, in women, pulse pressure is a better predictor (but in both sexes SBP was the best predictor). In another study, 40 atherothrombotic brain infarction was closely linked to the mean arterial pressure and the SBP, while DBP or pulse pressure did not improve the estimate of risk after adjustment for mean arterial pressure or SBP. It might be important to consider the relation between PCI and death from ischemic heart disease observed in the present report in women. As previously mentioned, in subjects over 55 years of age, an increased arterial stiffness leads to an increase in the pulsatility because of an increase in the SBP and a decrease in the DBP. 32 On the one hand, an increased SBP is a determinant of cardiac hypertrophy and, on the other hand, a decreased DBP may alter the coronary perfusion. The coronary perfusion depends more on DBP than on SBP, unlike the cerebral perfusion. l7r7-33 This hemodynamic pattern could explain why downward fluctuations in DBP appear to be more dangerous for the heart than for the brain, 4 and why antihypertensive treatments prevent stroke but not myocardial infarction. The inverse relation between PCI and stroke, observed in women, could explain why PCI was not found to be a prognostic factor of total cardiovascular deaths. In the Framingham study, a negative relation between stroke and pulse pressure, measured as SBP-DBP, has been noted, and this only in women. 40 In conclusion, the present report shows that the steady component of blood pressure is a strong risk factor for death from cardiovascular disease in both sexes. This study also suggests that the

9 400 Hypertension Vol 3, No 4, April 989 pulsatile component is a cardiovascular risk factor independent of the steady component in women older than age 55. Further studies are needed to validate these results. References. Kannel WB, Dawber TR, McGee DL: Perspectives on systolic hypertension: The Framingham Study. Circulation 980;6:l Ducimetiere P, Richard JL, Claude JR, Wamet JM: Les cardiopathies ische'miques. Incidence et facteurs de risque. Paris, INSERM Ed 98, pp Cambien F, Chretien JM, Ducimetiere P, Guize L, Richard JL: Is the relationship between blood pressure and cardiovascular risk dependent on body mass index? Am J Epidemiol 985; 22: Curb JD, Borhani NO, Entwisle G, Tung B, Kass E, Schnaper H, Williams W, Berman R: Isolated systolic hypertension in 4 communities. Am J Epidemiol 985; 2: Doyle AE: Vascular complications of hypertension, in Robertson JIS (ed): Handbook of Hypertension. Clinical Aspects of Hypertension. New York/Amsterdam, Elsevier Science Publishing Co, Inc, 985, vol, pp Kannel WB, Stokes J III: Hypertension as a cardiovascular risk factor, in Robertson JIS (ed): Handbook of Hypertension. Epidemiology of Hypertension. New York/Amsterdam, Elsevier Science Publishing Co, Inc, 985, vol 6, pp Conti S, Szklo M, otti A, Pasquini P: Risk factors for definite hypertension: Cross-sectional and prospective analysis of two Italian rural cohorts. Prev Med 986;5: Johnson JL, Heineman EF, Heiss G, Hames CG, Tyroler HA: Cardiovascular disease risk factors and mortality among black women and white women aged years in Evans County, Georgia. Am J Epidemiol 986; 23: Hagman M, Wilhelmsen L, Wedel H, Pennert K: Risk factors for angina pectoris in a population study of Swedish men. J Chron Dis 987;40: Tverdal A: Systolic and diastolic blood pressure as predictor of coronary heart disease in middle-aged Norwegian men. Br MedJ 987;294: Dyer AR, Stamler J, Shekelle RB, Shoenberger JA, Stamler R, Shekelle S, Berkson DM, Paul O, Lepper MH, Lindberg HA: Pulse pressure. I. Level and associated factors in four Chicago epidemiologic studies. J Chron Dis 982;35: Dyer AR, Stamler J, Shekelle RB, Shoenberger JA, Stamler R, Shekelle S, Berkson DM, Paul O, Lepper MH, Lindberg HA: Pulse pressure. II. Factors associated with follow-up values in three Chicago epidemiologic studies. J Chron Dis 982^5: Dyer AR, Stamler J, Shekelle RB, Shoenberger JA, Stamler R, Shekelle S, Berkson DM, Paul O, Lepper MH, Lindberg HA: Pulse pressure. III. Prognostic significance in four Chicago epidemiologic studies. J Chron Dis 982^5: Kannel WB, Gordon T, Schwartz MJ: Systolic versus diastolic blood pressure and risk of coronary heart disease: The Framingham study. Am J Cardiol 97l;27: Kannel WB, Castelli WP, McNamara PM, McKee PA, Feinleib M: Role of blood pressure in the development of congestive heart failure. The Framingham study. N Engl J Med 972;287: Ferguson JJ, Randall OS: Systolic, diastolic, and combined hypertension. Differences between groups. Arch Intern Med 986; 46: O'RourkeMF: Vascular impedance: The relationship between pressure and flow, in Arterial Function in Health and Disease. Edinburgh, Churchill Livingston Publisher, 982, pp 94-32, McDonald DA: Blood Flow in Arteries. London, Edward Arnold Publisher, 960, pp 7-54, 8-45, , Kannel WB: Prevalence and natural history of electrocardiographic left ventricular hypertrophy. Am J Med 983 (suppl);75:3-ll 20. Kromhout D, Bosschiter EB, De Lezenne Coulander C: Potassium, calcium, alcohol intake and blood pressure. The Zutphen Study'- 3. Am J Clin Nutr 985;4: Kannel WB: Hypertension, relationship with other risk factors. Drugs 986;3(suppl ):- 22. Simonson E: Differentiation Between Normal and Abnormal in Electrocardiography. St Louis, CV Mosby Co, 96, pp Safar ME, Laurent S, Safavian A, Pannier B, London GM: Pulse pressure in sustained essential hypertension: A haemodynamic study. J Hypertens 987^: Morisson DF: Multivariate Statistical Methods, ed 2. New York, McGraw-Hill Book Co, 978, pp O'Rourke MF: Hemodynamic basis for the concept of resistance and impedance in hypertension, in Safar ME (ed): Arterial and Venous Systems in Essential Hypertension. Dordrecht, Martinus Nijhoff Publishers, 987, pp Safar ME, Toto-Moukouo JJ, Asmar RE, Laurent S: Increased pulse pressure in patients with arteriosclerosis obliterans of the lower limbs. Arteriosclerosis 987;7: Milnor WR: Hemodynamics. Baltimore/London, Williams & Wilkins, 982, pp Gordon T, Sorlie P, Kannel WB: Problems in the assessment of blood pressure. The Framingham study. Int J Epidemiol l976;5: Harris T: Blood pressure experience and risk of cardiovascular disease in the elderly. Hypertension 985,7: Roberts J (ed): Blood Pressure Levels of Persons 6-74 Years, United States US Dept of Health, Education, and Welfare publication No. HRA US Government National Center for Health Statistics, Kannel WB, Wolf P, McGee DL, Dawber TR, McNamara P, Castelli WP: Systolic blood pressure, arterial rigidity, and risk of stroke. The Framingham study. JAMA 98; 245: Randall O, Van Den Bos G, Westerhof N: Systemic compliance. Does it play a role in the genesis of essential hypertension? Cardiovasc Res 984;l8: Safar ME, Simon AC: Hemodynamics in systolic hypertension, in Zanchetti A, Taruzi RC (eds): Handbook of Hypertension. Pathophysiology of Hypertension, Cardiovascular Aspects. New York/Amsterdam, Elsevier Science Publishing Co, Inc, 986, pp Safar ME, Toto-Moukouo JJ, Bouthier JA, Asmar RE, Levenson JA, Simon AC, Levenson JA, London GM: Arterial dynamics, cardiac hypertrophy and antihypertensive treatment. Circulation 987;75(suppl I):I-56-I Pannier B, Brunei P, Laurent S, Asmar R, Safar M: Pulse pressure and echocardiographic parameters in essential hypertension. J Hypertens (in press) 36. O'Rourke MF: Pulsatile arterial hemodynamics in hypertension. Aust NZ J Med 976;6(suppl 2): Bouthier JD, De Luca N, Safar ME, Levenson JA, Simon AC: Cardiac hypertrophy and arterial distensibility in essential hypertension. Am Heart J 985;09: Safar ME, London GM: Arterial and venous compliance in sustained essential hypertension. Hypertension 987; 0: Rutan GH, Kuller LH, Neaton JD, Wentworth DN, McDonald RH, McFate Smith W: Mortality associated with diastolic hypertension and isolated systolic hypertension among men screened for the Multiple Risk Factor Intervention Trial. Circulation 988;77:5O Kannel WB, Dawber TR, Sorlie P, Wolf PA: Components of blood pressure and risk of atherothrombotic brain infarction: The Framingham study. Stroke I976;7: Strandgaard S, Haunso S: Why does antihypertensive treatment prevent stroke but not myocardial infarction? Lancet 9872: KEY WORDS mean arterial pressure pulse pressure blood pressure cardiovascular risk factors

10 American Heart Association Scientific Sessions New Orleans, Louisiana November 3-6 Dallas, Texas November 2-5 Anaheim, California November 8-2 New Orleans, Louisiana November 9-2