Single Versus Combined Blood Pressure Components and Risk for Cardiovascular Disease
The Framingham Heart Study
Background— The utility of single versus combined blood pressure (BP) components in predicting cardiovascular disease (CVD) events is not established. We compared systolic BP (SBP) and diastolic BP (DBP) versus pulse pressure (PP) and mean arterial pressure (MAP) combined and each of these 4 BP components alone in predicting CVD events.
Methods and Results— In participants in the original (n=4760) and offspring (n=4897) Framingham Heart Study who were free of CVD events and BP-lowering therapy, 1439 CVD events occurred over serial 4-year intervals from 1952 to 2001. In pooled logistic regression with the use of BP categories, combining SBP with DBP and PP with MAP improved model fit compared with individual BP components (P<0.05 to P<0.0001). Significant interactions were noted between SBP and DBP (P=0.02) and between PP and MAP (P=0.01) in their respective multivariable models. Models with continuous variables for SBP+DBP and PP+MAP proved identical in predicting CVD events (Akaike Information Criteria=10 625 for both). Addition of a quadratic DBP2 term to DBP and SBP further improved fit (P=0.0016).
Conclusions— Combining PP with MAP and SBP with DBP produced models that were superior to single BP components for predicting CVD, and the extent of CVD risk varied with the level of each BP component. The combination of PP+MAP (unlike SBP+DBP) has a monotonic relation with risk and may provide greater insight into hemodynamics of altered arterial stiffness versus impaired peripheral resistance but is not superior to SBP+DBP in predicting CVD events.
Received June 9, 2008; accepted October 14, 2008.
Although hypertension is a well-established risk factor for cardiovascular disease (CVD) events, uncertainty exists regarding the relative importance of various blood pressure (BP) components in predicting risk. Historically, diastolic blood pressure (DBP) was considered a better predictor than systolic blood pressure (SBP) because it was thought to represent the resistance that the heart had to overcome to eject blood.1,2 It was not until the 1970s and 1980s that SBP was accepted as a clinically useful predictor of coronary heart disease, stroke, and heart failure; in older people, SBP was shown to be superior to DBP as a predictor of risk.3–5
Editorial p 210
Clinical Perspective p 250
More recently, in the 1990s, pulse pressure (PP), an indicator of arterial stiffness, was shown to be useful in predicting CVD events in the eldery.6–14 However, controversy persists regarding which BP component is superior as a predictor of CVD events.15–21 In an attempt to reconcile some of these differences, investigators at the Framingham Heart Study22 and others23–25 have shown age to play an important role in modifying the relation of BP components to CVD risk. With increasing age, there is a gradual shift from DBP to SBP and then to PP as predictors of risk. Importantly, a follow-up study of the Multiple Risk Factor Intervention Trial (MRFIT)26 participants concluded that CVD risk assessment was improved by considering both SBP and DBP jointly compared with SBP, DBP, or PP separately.
In summary, there still is uncertainty regarding the utility of single versus combined BP components in predicting CVD events. Furthermore, there is a suggestion in the literature that CVD risk attributable to hemodynamic factors may be assessed more clearly by a model that considers physiological rather than traditional (ie, SBP, DBP) components of BP, with PP serving as an indicator of large-artery stiffness (pulsatile load) and mean arterial pressure (MAP) serving as an indicator of peripheral resistance and cardiac output (steady flow load). We therefore asked the following questions: (1) Is it helpful for predicting CVD risk to use combined BP components instead of any 1 of the 4 single BP components? (2) Which combined model is more useful in predicting CVD risk: SBP and DBP or PP and MAP? (3) Are there nonlinear associations of 1 or more BP components in predicting CVD risk, and, if so, what is the clinical relevance?
The Framingham Heart Study began in 1948 with the enrollment of 5209 men and women, 28 to 62 years of age at entry, with participants undergoing repeated examinations biennially.27,28 In 1971, 5124 men and women who were children or the spouses of children of the original Framingham Heart Study were enrolled in the Framingham offspring cohort.29 The offspring cohort underwent repeat examinations approximately every 4 years. Each examination included an extensive CVD history and physical examination, 12-lead ECG, and various blood chemistries. Detailed descriptions of study design, the method for assessing BP, and the method of classifying CVD end points have been published elsewhere.30,31
The present study sample comprised 4760 participants from the original Framingham Heart Study cohort and 4897 participants from the Framingham offspring cohort who were free of CVD events at an index examination and not receiving antihypertensive therapy. We began with the third biennial examination in the original cohort (1952–1956) and the first examination of the offspring cohort (1971–1975) and extended our evaluation through November 1999, corresponding to the 25th examination (1997–1999) in the original and seventh examination (1998–2001) in the offspring cohort.
Definition of BP Categories
We conducted 2 separate analyses: 6-by-6 cross-classification of combined SBP+DBP versus combined PP (defined as PP=SBP−DBP in mm Hg) and MAP (defined as MAP=1/3 SBP+2/3 DBP in mm Hg), using the Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC-6)32 BP categories for SBP+DBP and corresponding values for PP+MAP. The SBP categories were <120 (optimal BP), 120 to 129 (normal BP), 130 to 139 (high-normal BP), 140 to 159 (stage 1 hypertension), 160 to 179 (stage 2 hypertension), and ≥180 mm Hg (stage 3 hypertension). The DBP categories were <70, 70 to 79, 80 to 89, 90 to 99, 100 to 109, and ≥110 mm Hg. The PP categories were <40, 40 to 49, 50 to 59, 60 to 79, 80 to 99, and ≥100 mm Hg. The MAP categories were <90, 90 to 94, 95 to 99, 100 to 109, 110 to 119, and ≥120 mm Hg. Using these categories for PP+MAP (and SBP+DBP), we constructed 6-by-6 cross-classification tables and graphs.
Assessment of Risk Factors and Cardiovascular Events
Current cigarette smoking was defined as regularly smoking cigarettes at any time during the prior year. Body mass index was calculated as body weight (in kilograms) divided by the square of height (in meters). A fasting blood glucose level of ≥7.0 mmol/L (126 mg/dL) (in the offspring cohort), a nonfasting glucose of ≥11.0 mmol/L (≥200 mg/dL) (in the original cohort), or use of hypoglycemic medications (in both cohorts) defined diabetes mellitus. CVD events were defined as the occurrence of a first myocardial infarction, coronary insufficiency, sudden CVD death, stroke, or heart failure. The criteria for diagnosis of these specific CVD events are described elsewhere.31
The database was reconstructed to create 3- to 5-year intervals (averaging 4.24 years [minimum=3.00, maximum=4.99]), each of which was assessed for the occurrence of CVD events as described below. Pooled logistic regression was used in participants free of antihypertensive therapy at the index examination in each 4-year cycle, and multiple observations per subject were generated from the merged Framingham original and offspring cohorts to maximize the number of person-observations. After the initial occurrence of a qualifying CVD event, subsequent observations for that participant were truncated and did not contribute any further person-observations.
Pooled logistic regression was used to examine the relation of combined SBP+DBP (or PP+MAP) categories with CVD risk. Covariates in each probability model included male sex, nondiabetic, nonsmoking status, examinations 1 and 12 to estimate secular trend (in which examination 12 in the original cohort corresponded to examination 1 in offspring cohort), and mean values of age, total cholesterol, and body mass index. Odds ratios (ORs) and 95% CIs were determined for each individual cell in 6-by-6 cross-classification tables for models containing SBP+DBP or PP+MAP; each model included a reference cell and 35 indicator or dummy variables, constructing ORs for each, with lowest SBP/DBP or PP/MAP cells as the reference category. The cells containing <10 events were not collapsed but instead remained blank to indicate insufficient events per cell. Likelihood ratio χ2 statistics for model fit were compared in models containing SBP+DBP (or PP+MAP) with the addition of the interaction terms of SBP×DBP (or PP×MAP) versus only the main effects of SBP+DBP (or PP+MAP) to examine whether the interaction between these BP components significantly improved prediction of CVD events. In addition, ORs and 95% CIs for predicting CVD events were calculated from JNC-6 categories32 for SBP and DBP groups, including prehypertension33 (SBP 120 to 139 mm Hg and DBP 80 to 89 mm Hg), isolated systolic hypertension (ISH) (SBP ≥149 and DBP <90 mm Hg), and systolic-diastolic hypertension (SBP ≥140 and DBP ≥90 mm Hg).
Absolute rates (per 100 person-years) were also plotted in each cell for the combined SBP+DBP and combined PP+MAP cross-classification tables calculated from the predicted probability equation derived from the logistic regression program after setting covariates to mean levels: age 52.3 years, total cholesterol 219 mg/dL, body mass index 25.9, male, nonsmoker, nondiabetic, and secular trend set to visit 12.
In addition, pooled logistic regression probability models incorporated BP components as continuous variables, with the same adjustments for age, sex, and covariates that were used for categorical analysis. DBP2 and SBP2 terms were added to the combined SBP+DBP model, and MAP2 and PP2 terms were added to the combined PP+MAP model to test for linear or quadratic relationships. Finally, interaction terms of SBP×DBP and PP×MAP were added to the model containing the combined BP components and quadratic terms to determine whether the effects of SBP+DBP or PP+MAP depended on each other.
We tested for the interaction of age and sex in modeling for CVD risk. We used Akaike Information Criteria to test whether models of PP+MAP outperformed models of SBP+DBP with regard to predicting CVD risk. SAS statistical software (SAS Institute, Cary, NC) was used.34 Area under the curve of the receiver operating characteristic curve was used to quantify whether combined BP components improved prediction of CVD events over single BP components. Area under the curve comparison was computed with the use of the roccomp command in Stata version 10.0 (Stata Corporation, College Station, Tex).
The authors had full access to and take responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
Final Study Sample
There were a total of twelve 4-year intervals that generated 3345 CVD events and 89 060 person-observations. We excluded 790 CVD events and 35 502 person-observations because they did not occur during the structured 4-year intervals, 987 CVD events and 10 359 person-observations because of use of antihypertensive therapy, and 129 CVD events and 1675 person-observations because of missing covariates. After these exclusions, 41 524 person-observations and 1439 CVD events remained eligible for analysis. There was a mean of 1221 person-observations and 48 events (range, 11 to 106) per cell for the combined PP+MAP tables and a mean of 1339 person-observations and 54 events (range, 12 to 148) per cell for the combined SBP+DBP tables. Table 1 outlines the characteristics of study participants at the initial qualifying examination. Women constituted 54% of the sample. Participants at the first index examination had a mean age of 42 years.
Models Treating BP Components as Categories
The number of CVD events and person-observations per cell for the 6-by-6 cross-classification of mean SBP+DBP are shown in Table 2. (Mean values per cell for combined SBP+DBP are shown in the online-only Data Supplement, Table I.) The corresponding adjusted ORs and 95% CIs are presented in Table 3. The pooled regression model containing both SBP+DBP was superior to either SBP or DBP alone in predicting CVD events (Table 3, Table 4, Figure 1). Furthermore, the model of SBP+DBP containing the interaction term of DBP×SBP improved model fit over the main effects of SBP+DBP alone, indicating that the effect of 1 component on risk varied according to the levels of the second component. Similar findings were noted for absolute rates of CVD risk (online-only Data Supplement, Table II).
The number of CVD events and person-observations per cell for the 6-by-6 cross-classification of PP+MAP is shown in Table 5. (Mean values for combined PP+MAP per cell are shown in the Data Supplement, Table III.) The corresponding adjusted ORs and 95% CIs are presented in Table 6. The pooled regression model containing the combination of PP+MAP was superior to either PP or MAP alone in predicting CVD events (Table 6, Table 4, Figure 2). In addition, the model of PP+MAP, containing the interaction term of PP×MAP, significantly improved model fit over the main effects of PP+MAP alone, indicating that the effect of 1 component on risk varied according to the levels of the second component (Table 4). Similar findings were noted for absolute rates of CVD risk (Data Supplement, Table IV). In summary, these categorical analyses show that the combination of SBP+DBP or the combination of PP+MAP was superior to single BP components in predicting CVD risk.
Furthermore, models based on JNC-6 categories for SBP and various DBP groups, utilizing all 41 524 person-observations and 1439 CVD events, were constructed with optimal BP (SBP <120 and DBP <80 mm Hg) serving as the reference group. Persons with prehypertension (SBP 120 to 139 mm Hg) in whom DBP was <70 mm Hg (mean 127/65 mm Hg) had an OR=2.0 (95% CI, 1.5 to 2.6) for CVD events, which was comparable to those with stage 1 ISH (SBP 140 to 159 mm Hg) in whom DBP was 70 to 89 mm Hg (mean 147/81 mm Hg) (OR=2.0 [1.6 to 2.5]). Similarly, persons with stage 1 ISH (SBP 140 to 159 mm Hg) in whom DBP was <70 mm Hg (mean 147/64 mm Hg) had an OR=3.0 (2.1 to 4.3) for CVD events, comparable to those with stage 2 and 3 ISH (SBP ≥160 mm Hg) in whom DBP was <90 mm Hg (mean 171/81 mm Hg) (OR=3.1 [2.4 to 4.1]) or persons with stage 2 and 3 systolic-diastolic hypertension (SBP ≥160 mm Hg) in whom DBP was 90 to 99 mm Hg (mean 172/94 mm Hg) (OR=2.7 [2.0 to 3.6]) (all P<0.0001).
Models Treating BP Components as Continuous Variables
In a model adjusted for age, sex, and other covariates, the odds of CVD events increased with increasing SBP (OR per SD=1.39; P=0.0001) or increasing DBP (OR per SD=1.25; P=0.0001). Model fit was not improved when DBP was added to SBP; however, adding a DBP2 term (but not SBP2) to DBP+SBP improved model fit (Table 4), indicating a nonlinear relation between CVD risk and DBP. Adding SBP2 to DBP+SBP did not improve the fit. Therefore, for any given SBP ≥120 mm Hg, odds of CVD events increased at both high and low extremes of DBP; in contrast, for any given DBP, odds of CVD events increased monotonically with increasing SBP. Furthermore, model fit improved with addition of an SBP×DBP interaction term, indicating that the effect of SBP or DBP on CVD risk is modified by the level of the alternate BP component (Table 4), consistent with the categorical model.
Similarly, in separate models, the odds of CVD increased with increasing PP (OR per SD=1.34; P<0.0001) or increasing MAP (OR per SD=1.35; P<0.0001). When MAP was added to PP, the fit improved significantly (Table 4). In contrast, adding MAP2 or PP2 to MAP+PP did not improve the fit, consistent with a linear relation. Therefore, for any given PP, the odds of CVD events increased linearly with increasing MAP; similarly, for any given MAP, the odds of CVD events increased linearly with increasing PP. Model fit was improved after addition of a PP×MAP interaction term, indicating that the effects of PP and MAP on CVD risk are modified by the level of the alternate BP component, consistent with the categorical models.
There was no significant age or sex interaction in modeling for CVD risk. Akaike Information Criteria was identical (10 625.04) for models of SBP+DBP versus MAP+PP when BP components were treated as continuous variables. Receiver operating characteristic curves for each model comparison (Table 4) show better fit for combined BP components than for single ones when fully adjusted for covariates and clinically relevant improvement in fit without adjustments.
The present study showed that the combined BP components of SBP+DBP or those of PP+MAP were superior to single BP components in predicting CVD risk. Indeed, the findings that combined SBP+DBP was superior to SBP alone in predicting CVD risk confirmed the MRFIT study26 and extended the findings to older people and to women. Moreover, combined SBP+DBP and combined PP+MAP were equally predictive of CVD risk, with neither model outperforming the other by Akaike Information Criteria. Finally, of the 4 BP components, only DBP showed a nonlinear, quadratic relation with CVD risk. The J-shaped relation to CVD risk that is associated with DBP presumably reflects, in large measure, increased arterial stiffness as manifested by a low DBP (and,by definition, a high PP).
Utility of Paired PP and MAP Versus Paired SBP and DBP
Because PP+MAP is derived exclusively from SBP+DBP, it cannot be superior to the latter pairing in predicting CVD risk, and this was confirmed by Akaike Information Criteria. Nevertheless, combined PP+MAP can be considered superior to combined SBP+DBP in interpreting the relative role of both stiffness and resistance in contributing to CVD risk. Individual components can assess stiffness or resistance but cannot assess both of these factors. These findings were not unexpected, in that DBP rises with increases in vascular resistance and falls with increased large-artery stiffness.9 On the other hand, the traditional combined SBP+DBP remains the clinical model of choice because these components are measured by the physician, and clinical trials of BP-lowering therapy were predicated on SBP and DBP but not PP or MAP. Moreover, despite the quadratic relation of DBP to CVD risk, increased vascular resistance can be deduced indirectly from concordantly high DBP and SBP, whereas increased arterial stiffness can be deduced indirectly from discordantly high SBP and low DBP, resulting in a high PP.9,26
The importance of PP as a risk predictor was not emphasized in the 2003 JNC-733 recommendations. In contrast, the 2007 European Guidelines for the Management of Arterial Hypertension35 recommended that patients with systolic hypertension in association with low DBP of <60 to 70 mm Hg (ie, ISH grades 1, 2, or 3) be regarded as having an additional risk as a result of “advanced organ damage.” Our findings support the high-risk designation of the 2007 European Guidelines and further extend an increased risk designation to those individuals with prehypertension and DBP <70 mm Hg (PP ≥60 mm Hg). Prehypertension and ISH with DBP <70 mm Hg involve a substantial number of individuals,36 and the recognition of their CVD risk by virtue of cross-classification of BP components may represent an important insight to guide preventive and therapeutic care.
Strengths and Limitations
The strengths of our investigation included a wide age range, inclusion of both men and women, and the standardized measurements available from the well-characterized Framingham Heart Study cohorts. The pooled logistic regression utilized repeated 4-year intervals, which had the advantage of maximizing the person-observations available for the study and minimizing interaction by age and sex. On the other hand, the 4-year follow-up intervals are not representative of longer-term risks. Indeed, we have shown important effect modifications by age and sex with longer-term risk in a previous report in this study sample.22
We examined single and combined BP components defined categorically on the basis of JNC-6 categories, as well as continuously for assessing CVD risk. The analysis using categories is more suitable to clinical interpretation, whereas when the BP components are defined continuously, it allows for a more direct comparison of their relative effects as well as assessment of linear and nonlinear relations of BP components in predicting CVD risk.’
Our sample consisted primarily of white individuals and therefore may not be representative of other ethnic or racial groups. Finally, our data do not provide insight into how more contemporary cohorts of treated persons may fare with respect to CVD risk.
Previous studies that championed a single BP component as the “the best” predictor of CVD risk examined a limited spectrum of the overall hypertensive population. Our results confirm the importance of combining BP components, such as SBP and DBP or PP and MAP, to improve stratification of CVD risk. Indeed, when PP, a measurement of stiffness, was combined with MAP, a measurement of resistance, we were able to relate the 2 major physiological components of hydraulic load to clinical outcome; single BP components cannot do this. Our results have a bearing on the current US guidelines33 that use both SBP and DBP, whichever is higher, for determining BP stage. Although these guidelines take into account the importance of increased vascular resistance, they undervalue the importance of increased arterial stiffness (ie, increase PP and low DBP), which is common in older persons, especially those with prehypertension and ISH with DBP <70 mm Hg.
Sources of Funding
The Framingham Heart Study is supported by National Institutes of Health/National Heart, Lung, and Blood Institute contract N01-HC-25195. Dr Vasan was supported in part by National Institutes of Health/National Heart, Lung, and Blood Institute contract 2K24HL04334.
Pachon V. A new approach to sphygmomanometry. Presse Med. 1913; 12: 229–331.
MacWilliams JA. Systolic and diastolic blood pressure estimation. BMJ. 1914; 2: 693–697.
Darne B, Girerd X, Safar M, Cambien F, Guize L. Pulsatile versus steady component of blood pressure: a cross-sectional analysis and a prospective analysis on cardiovascular mortality. Hypertension. 1989; 13: 392–400.
Franklin SS, Gustin W, Wong ND, Larson MG, Weber MA, Kannel WB, Levy D. Hemodynamic patterns of age-related changes in blood pressure: the Framingham Heart Study. Circulation. 1997; 96: 308–315.
Benetos A, Safar M, Rudnichi A, Smulyan H, Richard JL, Ducimetiere P, Guize L. Pulse pressure: a predictor of long-term cardiovascular mortality in a French male population. Hypertension. 1997; 30: 1410–1415.
Franklin SS, Khan SA, Wong ND, Larson MG, Levy D. Is pulse pressure useful in predicting risk for coronary heart disease? The Framingham Study. Circulation. 1999; 100: 354–360.
Gasowski J, Fagard RH, Staessen JA, Grodzicki T, Pocock S, Boutitie F, Gueyffier F, Boissel JP. Pulsatile blood pressure component as predictor of mortality in hypertension: a meta-analysis of clinical trial control groups. J Hypertens. 2002; 20: 141–151.
Psaty BM, Furberg CD, Kuller LH, Cushman M, Savage PJ, Levine D, O'Leary DH, Bryan RN, Anderson M, Lumley T. Association between blood pressure level and the risk of myocardial infarction, stroke, and total mortality: the Cardiovascular Health Study. Arch Intern Med. 2001; 161: 1183–1192.
Miura. K, Dyer AR, Greenland P, Daviglus ML, Hill MA, Liu K, Garside DB, Stamler J. Pulse pressure compared with other blood pressure indexes in the prediction of 25-year cardiovascular and all-cause mortality rates: the Chicago Heart Association Detection Project in Industry Study. Hypertension. 2001; 38: 232–237.
Mosley WJ, Greenland P, Garside DB, Lloyd-Jones DM. Predictive utility of pulse pressure and other blood pressure measures for cardiovascular outcomes. Hypertension. 2007; 49: 1256–1264.
Glynn RJ, L'Ialien GJ, Sesso HD, Kackson EA, Buring JE. Development of predictive models for long-term cardiovascular risk associated with systolic and diastolic blood pressure. Hypertension. 2002; 39: 105–110.
Asia Pacific Cohort Studies Collaboration. Blood pressure indices and cardiovascular disease in the Asia Pacific Region: a pooled analysis. Hypertension. 2003; 42: 69–75.
Franklin SS, Larson MG, Khan SA, Wong ND, Leip EP, Kannel WB, Levy D. Does the relation of blood pressure to coronary heart disease risk change with aging? The Framingham Heart Study. Circulation. 2001; 103: 1245–1249.
Khattar RS, Swales JD, Dore C, Senior R, Lahiri A. Effect of aging on the prognostic significance of ambulatory systolic, diastolic, and pulse pressure in essential hypertension. Circulation. 2001; 104: 783–789.
Assman G, Cullen P, Evers T, Petzinna D, Schulte H. Importance of arterial pulse pressure as a predictor of coronary heart disease risk in PROCAM. Eur Heart J. 2005; 26: 2120–2126.
Kannel WB, Wolf P, Garrison R. The Framingham Study: Section 35: Survival Following Initial Cardiovascular Events. Bethesda, Md: National Institutes of Health; 1998: 5.
Cupple LA, D'Agostino RB Sr. Some risk factors related to the annual incidence of cardiovascular disease and death using pooled repeated biennial measurements: Framingham Study, 30-year follow-up. In: Kannel WB, Wolf PA, Garrison RJ, eds. The Framingham Heart Study: An Epidemiologic Investigation of cardiovascular Disease. Washington, DC: National Institutes of Health; 1987: 87–203.
Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL, Jones DW, Materson BJ, Oparil S, Wright JT, Roccella EJ, and National High Blood Pressure Education Program Coordinating Committee. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. JAMA. 2003; 289: 2560–2572.
SAS Institute Inc. SAS/STAT Software; Changes and Enhancements Through Release 6.11. Cary, NC: SAS Institute Inc; 1996: 807–884.
The Task Force for the Management of Arterial Hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). 2007 Guidelines for the management of arterial hypertension. Eur Heart J. 2007; 28: 1462–1536.
Franklin SS, Milagros JJ, Wong ND, L'Italien GJ, Lapuerta P. Predominance of isolated systolic hypertension among middle-aged and elderly US hypertensives: analysis based on National Health and Nutrition Examination Survey (NHANES III). Hypertension. 2001; 37: 869–874.
Although hypertension is a well-established risk factor for cardiovascular disease, uncertainty exists regarding the relative importance of various blood pressure (BP) components in predicting risk. Our results, with the use of data from participants in the Framingham Heart Study, confirmed the importance of combining BP components, such as systolic blood pressure (SBP) and diastolic blood pressure (DBP) or pulse pressure and mean arterial pressure, to improve cardiovascular risk prediction beyond that of any single BP component. Previous studies that championed a single component examined a limited spectrum of the overall hypertensive population. Although both of these 2-component models were equally predictive of cardiovascular risk, only the model of pulse pressure and mean arterial pressure has a linear relation with risk that allows a comparison of the relative role of stiffness versus resistance in contributing to this risk; in contrast, risk increases at both high and low extremes of DBP when combined with SBP. Therefore, pulse pressure and mean arterial pressure as surrogates for stiffness and resistance, respectively, are not only valuable in quantifying risk but also provide greater insight into the underlying hemodynamic perturbation. Although current national guidelines consider both SBP and/or DBP, whichever is higher, in determining staging of BP, they undervalue the significance of increased arterial stiffness, as manifested by high SBP and low DBP. This latter pattern occurs frequently in older persons with prehypertension or isolated systolic hypertension when accompanied by DBP <70 mm Hg and is associated with greater cardiovascular risk than calculated by traditional guidelines.
Guest Editor for this article was Paolo Verdecchia, MD.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.108.797936/DC1.