Abstract
Systolic (SBP) and diastolic blood pressure (DBP) and mean arterial pressure (MAP) are risk factors for cardiovascular mortality (CVM). Pulse pressure (PP) is considered as an easily available marker of vascular stiffness and the double product (DP) as a marker of cardiac workload. Therefore, we have examined the predictive value of PP and DP in the Ludwigshafen Risk and Cardiovascular Health study, a monocentric cohort study of 3316 patients referred to coronary angiography. An increase of SBP or PP by 1mmHg increased the risk of CVM with hazard ratios of 1.009 (95% CI, 1.005‐1.012) and 1.016 (1.012‐1.020), respectively. Increasing DP by 100 mm Hg/min was associated with a 1.010 (1.007‐1.013) higher risk of CVM. In patient subgroups with coronary artery disease (CAD) and heart failure (HF), PP and DP predicted CVM better than SBP or MAP. In a multivariate analysis adjusted for sex, BMI, diabetes, eGFR, hazard ratios for CVM for z‐standardized PP, DP, SBP, and HR were 1.20, 1.16, 1.12, and 1.14. After adding age to the multivariate analysis, only DP and HR remained significant. We provide evidence that PP and DP are powerful predictors of CVM and all‐cause mortality in a CV medium‐ to high‐risk population, especially in patients with CAD and HF. While DP proved to be an independent predictor of cardiovascular and all‐cause mortality also in multivariate analysis, PP was no independent predictor in our cohort with widespread antihypertensive treatment (>85%). PP is associated with age, presence of diabetes, obesity, and impaired renal function.
Keywords: cardiac workload, coronary artery disease, double product, heart failure, hypertension, mortality, pulse pressure, rate pressure product, vascular stiffness
1. INTRODUCTION
Arterial hypertension is an independent risk factor for CV (cardiovascular) disease, CV events, and mortality. Pulse pressure (PP), the difference between systolic and diastolic blood pressure, is clinically easily available. It is influenced by the Windkessel model (central artery elasticity), left ventricular stroke volume, and the backwards directed wave reflection of the peripheral arterial tree.
In the Framingham Heart Study, an increase of 10 mm Hg in pulse pressure (PP), systolic blood pressure (SBP), and diastolic blood pressure (DBP) was independently associated with a 23%, 16%, and 14% higher risk for coronary artery disease (CAD). 1 In addition to that, PP was also associated with CV mortality in the MRFIT trial (Multiple Risk Factor Intervention Trial), particularly in men older than 45 years. 2 In the FAVORIT trial (Folic Acid for Vascular Outcome Reduction in Transplantation) of kidney transplant patients, an association between SBP and CV disease could be seen. 3 DBP higher than 70 mm Hg was not associated with any risk for CV disease and mortality, while lower values than 70 mm Hg were. 3
We have recently reported in a large retrospective analysis of 43 006 renal transplant recipients that higher 1‐year PP and SBP were significantly associated with inferior 10‐year patient and death‐censored graft survival. 4 SBP but not DBP was associated with 10‐year survival in patients older than 60 years, whereas DBP was associated with inferior 10‐year patient and death‐censored graft survival only in patients younger than 60 years.
In a large community‐based study with 169 613 participants, PP better predicted new‐onset CV disease than arterial stiffness index. 5 It was also shown that increased arterial stiffness leads to high PP even in cerebral microvessels causing structural, functional, hemodynamic, and metabolic changes in the brain that could contribute to neuronal dysfunction and cognitive decline. 6 Apart from this, a previous analysis of the LURIC data showed that patients in the highest heart rate (HR) quartile (HR at rest ≥ 84/min) have a shorter survival time of 1.2 and 1.4 years for all‐cause and CV mortality. 7
Ansari et al showed that a low double product (DP) reserve (difference between DP at rest and peak value under stress) in patients who underwent dipyridamole stress MPI (myocardial perfusion imaging) is significantly associated with CAD. 8 The increase of DP per 1 m walking distance during the 6‐minute walk was higher or highest in patients with HFpEF and HFrEF compared to the control group. 9 Furthermore, the mean increase of DP per 1‐m walk is negatively correlated with the ejection fraction. The rise of DP, also called rate pressure product, seems to be a good parameter to evaluate the left ventricular myocardial function, especially in heart failure patients. 9 In addition, a low DP reserve (<10 000 mm Hg/min) was the strongest predictor of cardiovascular death in a cohort of 1759 male veterans undergoing treadmill testing. 10 In the International Database on Ambulatory blood pressure in relation to Cardiovascular Outcomes (IDACO) study, 9937 patients were followed up for a median period of 11.0 years. However, it could not be proven that DP has an additive value in risk assessment compared to SBP and HR in a general population. 11
In the Acute Study of Clinical Effectiveness of Nesiritide in Subjects With Decompensated Heart Failure (ASCEND‐HF), an in‐hospital rise of DP from baseline to discharge but not DP at baseline alone was associated with higher 30‐day mortality/HF hospitalization in patients with HFpEF. 12
The aim of our study was to analyze the role of blood pressure components SBP, DBP, and PP as well as the double product on CV mortality and all‐cause mortality in a patient cohort with a high percentage of coronary artery disease (CAD) and heart failure (HF).
We tried to test the following hypotheses: First, since DP reflects the cardiac workload and combines the information of two hemodynamic parameters, DP could offer a better mortality prediction than SBP, DBP, HR, and MAP. In addition, PP that reflects the vascular stiffness and the arteriosclerotic burden might also provide a better mortality prediction than standard blood pressure parameters.
2. METHODS
2.1. Study cohort
The Ludwigshafen Risk and Cardiovascular Health (LURIC) study is a monocentric hospital‐based cohort study that recruited 3316 patients of German ancestry living in the surroundings of the German cities Ludwigshafen, Mannheim, and Heidelberg, who underwent coronary angiography between July 1997 and January 2000. Ethical approval for this study has been given by the Landesärztekammer Rheinland‐Pfalz [#837.255.97(1394)]. Furthermore, the study was conducted according to the "Declaration of Helsinki” and all study participants have given written consent to participate in the study. Main indications for coronary angiography were acute chest pain or a positive non‐invasive cardiac stress test.
Exclusion criteria were other acute cardiac diseases, such as decompensated heart failure or decompensated valvular disease, or acute non‐cardiac diseases, such as infection, endocrine disease or any type of surgery within the previous three months, chronic polymorbid disease where non‐cardiac disease predominates (ie chronic renal failure and hemodialysis, severe rheumatic arthritis, persistent incapacitation after accident/trauma), history of malignant disease within the previous five years, and individuals incapable of understanding the purpose of the study. 13 After study inclusion, an exact physical examination including vital signs and a detailed blood test has been performed.
With the help of an automated oscillometric device (Omron MX4, Omron Healthcare GmbH, Hamburg, Germany), blood pressure was measured while being in a supine position for at least 10 minutes. At intervals of 30 seconds, at least three consecutive measurements of systolic and diastolic blood pressures were taken. If measurements varied > 10 mm Hg systolic,> 5 mm Hg diastolic, or heart rate > 5 beats per minute from each other (except for atrial fibrillation), they were regarded as invalid and have been repeated. Both valid and invalid measurements were recorded. The invalid ones were immediately identified as such. Only stable measurements that matched the reproducibility criteria were entered into the database. One part of the questionnaire investigated whether blood pressure had been measured according to protocol (supine, 10 minutes at rest). The study protocol and study procedures have been reported in some detail previously. 13
2.2. Laboratory measurements
Venous blood was drawn in study participants after an overnight fasting period under standardized conditions. Within 30 minutes after venipuncture, the remaining blood was centrifuged at 3000 g for 10 minutes, immediately aliquoted, and frozen at −80°C until further analysis. A summary of analytic methods has been reported previously. 13 Total cholesterol, high‐density lipoprotein (HDL) cholesterol, and triglyceride concentrations were determined enzymatically with a Synchron LX‐20 (Beckman Coulter, Munich, Germany). The LDL and very low‐density lipoprotein (VLDL) were separated by ultracentrifugation in a Beckman LM‐8 ultracentrifuge in 100 μL volumes with a VT‐51.2 rotor (Beckman Coulter). NT‐pro‐BNP was measured by electrochemiluminescence on an Elecsys 2010 (Roche Diagnostics). Creatinine was measured using the CREA assay (Roche, Germany) on a Hitachi 717 analyzer. The estimated glomerular filtration rate was calculated using the MDRD formula.
2.3. Clinical definitions
Coronary artery disease (CAD) was defined according to the American Heart Association by visible luminal narrowing of 20% stenosis or more in at least one of 15 coronary segments. 13 Coronary 1‐vessel disease, coronary 2‐vessel disease, and coronary 3‐vessel disease were defined as a stenosis of 50% or higher in one, two, or three of the major coronary arteries LAD (left anterior descending artery), RCX (ramus circumflexus), and RCA (right coronary artery). No or insignificant CAD was defined as a luminal narrowing of 0%‐49%. 13 The echocardiographic classification of heart failure has already been published. 13
Heart failure (HF) with preserved ejection fraction was defined as symptoms and signs of HF, a preserved left ventricular function with an ejection fraction >45% (echocardiographic or invasive) and the presence of diastolic HF according to the definition published by Paulus et al. 14 In particular, diastolic dysfunction was diagnosed in 388 patients (85%) based on hemodynamic criteria (mean pulmonary capillary wedge pressure > 12mmHg or a left ventricular end‐diastolic pressure > 16 mm Hg). In the remaining 71 patients (15%), diastolic dysfunction was identified by an elevated NT‐proBNP concentration (>220 pg/mL) and electrocardiographic evidence of atrial fibrillation. A summary of clinical definitions has been reported previously in more detail. 13
2.4. Follow‐up
Patients have been followed up for a median of 9.9 years. Information on vital status was obtained from local registries. Death certificates, medical records of local hospitals, and autopsy data were reviewed independently by two experienced clinicians. They were blinded to patient characteristics and classified the causes of death. In cases of disagreement or uncertainty concerning the coding of a specific cause of death, the decision was made by a principal investigator (WM).
2.5. Statistical Analysis
IBM SPSS® Statistics version 25.0 (SPSS Inc, IBM Corporation) and R v4.0.2 were used for all analyses. We used the Kaplan–Meier method to estimate the survival regarding cardiovascular and all‐cause mortality in association with SBP, DBP, mean arterial pressure (MAP), PP, heart rate (HR), and double product (DP). Hazard ratios were calculated by Cox proportional hazards regression. Variables were Z‐transformed before entering analyses, so the hazard ratios are not only calculated for an increase per one unit but also for a rise of one standard deviation of the respective marker. We performed univariate analyses and multivariate analysis with adjustment for age, sex, BMI, diabetes mellitus, and eGFR. All data were completed except for three study participants with missing data on eGFR. These were excluded from the adjusted analysis.
3. RESULTS
3.1. Whole LURIC cohort
69.7% of the patients were male, 39.9% were suffering from diabetes mellitus, and 23.4% were active smokers, while 41% were ex‐smokers and 23.2% were obese (Table 1). The mean age was 62.7 years (Table 2), while the mean blood pressure was 141/ 81 mm Hg (Table 2).
Table 1.
Patient subgroups given in absolute numbers and percentages
| Frequency | Percentage | |
|---|---|---|
| Number of patients in total | 3316 | 100% |
| Male | 2310 | 69.7% |
| Body mass index > 30 kg/m2 | 770 | 23.2% |
| Active Smoker (self‐reported + cotinine>15µg/L) | 777 | 23.4% |
| Ex‐Smoker | 1367 | 41.0% |
| Non‐Smoker | 1178 | 35.5% |
| Diabetes mellitus (HbA1c > 6.5%) | 1324 | 39.9% |
| CKD (GFR < 60 mL/min/1,73 m2 | 456 | 13,8% |
| No Heart failure | 2230 | 67.2% |
| HFrEF | 580 | 17.5% |
| HFpEF | 506 | 15.3% |
| No/insignificant CAD | 1035 | 31.7% |
| 1‐vessel CAD | 620 | 18.7% |
| 2‐vessel CAD | 623 | 18.8% |
| 3‐vessel CAD | 989 | 29.8% |
| Atrial fibrillation | 397 | 12.0% |
| LDL > 130 mg/dL | 1072 | 32.3% |
| HDL < 40 mg/dL | 1935 | 58.4% |
| Triglycerides > 150 mg/dL | 1533 | 46.2% |
| Use of statin therapy | 1555 | 46,9% |
| Use of antihypertensive medication | 2879 | 86,8% |
| Use of antiplatelet therapy | 2366 | 71.4% |
Abbreviations: CAD, coronary artery disease; CKD, chronic kidney disease; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction.
Table 2.
Baseline patient characteristics given as mean, standard deviation, and range
| Mean | Standard deviation | Minimum | Maximum | N | |
|---|---|---|---|---|---|
| Age (y) | 62.7 | 10.6 | 17.3 | 92.1 | 3316 |
| PP (mm Hg) | 60.2 | 18.2 | 17.0 | 148.3 | 3316 |
| DP (mm Hg/min) | 9713 | 2428 | 4009 | 26 910 | 3316 |
| MAP (mm Hg) | 101 | 14.1 | 58.0 | 151 | 3316 |
| SBP (mm Hg) | 141 | 23.6 | 73.7 | 230 | 3316 |
| DBP (mm Hg) | 81 | 11.5 | 46.7 | 121 | 3316 |
| HR (beats per minute) | 68.7 | 11.7 | 35.6 | 130 | 3316 |
| Body mass index (kg/m2) | 27.5 | 4.08 | 16.3 | 57.1 | 3316 |
| HDL cholesterol (mg/dL) | 38.7 | 10.8 | 2 | 104 | 3315 |
| LDL cholesterol (mg/dL) | 117 | 34.3 | 15 | 361 | 3315 |
| Triglycerides (mg/dL) | 171 | 121 | 33 | 2667 | 3315 |
| Homocysteine (µmol/L) | 13.5 | 5.96 | 2.00 | 98.0 | 3312 |
Abbreviations: DBP, diastolic blood pressure; DP, double product; HR, heart rate; MAP, mean arterial pressure; PP, pulse pressure; SBP, systolic blood pressure.
995 (30%) participants died during a median follow‐up of 9.9 years (range 0.1‐11.9 years). A total of 622 (18.8%) participants died due to cardiovascular causes. Cardiac mortality included the following categories: sudden cardiac death (n = 259, 7.8%), fatal myocardial infarction (n = 106, 3.2%), death due to congestive heart failure (n = 148, 4.5%), death after intervention to treat coronary artery disease (n = 26, 0.8%), and other causes of death due to CAD (n = 19, 0.6%). Sixty‐one participants died due to fatal stroke. Information for vital status is complete for all participants but the cause of death of 19 deceased patients was unknown, and these patients were included in calculations of all‐cause mortality but not in calculations considering different causes of death.
We found statistically highly significant associations for almost all blood pressure parameters with CV mortality and all‐cause mortality except for DBP (Table 3, Figure 1). Specifically, increases of 1 mm Hg of SBP, MAP, and PP were associated with increases in CV mortality of 0.9%, 0.7%, and 1.6%, respectively. Similarly, increases of 1 mm Hg of SBP, MAP, and PP were associated with increases in all‐cause mortality of 0.9%, 0.6%, and 1.7%, respectively. DBP was neither associated with CV mortality nor all‐cause mortality. Another important finding was that an increase of heart rate by 1/min was associated with a higher risk of cardiovascular mortality of 1.5% and all‐cause mortality of 1.4%. Also, a rise of DP of 100 mm Hg/min was associated with an increased CV mortality of 1.0% and all‐cause mortality of 0.9%.
Table 3.
Association of baseline hemodynamic parameters and their z‐standardized data with CV mortality and all‐cause mortality given as hazard ratio and 95% confidence interval in an unadjusted model
|
Hazard ratio CV Mortality (95% CI) |
p‐Value |
Hazard ratio All‐cause Mortality (95% CI) |
p‐Value | ||
|---|---|---|---|---|---|
| MAP (mm Hg) |
1.007 (1.001‐1012) |
.020 |
1.006 (1.002‐1011) |
.005 | |
|
PP (mm Hg) |
1.016 (1.012‐1.020) |
<.001 |
1.017 (1.013‐1.020) |
<.001 | |
| DP (per 100 mm Hg/min) |
1.010 (1.007 ‐ 1.013) |
<.001 |
1.009 (1.007 ‐ 1.012) |
<.001 | |
|
SBP (mm Hg) |
1.009 (1.005‐1.012) |
<.001 |
1.009 (1.006‐1.012) |
<.001 | |
|
DBP (mm Hg) |
0.996 (0.989‐1.003) |
.308 |
0.995 (0.990‐1.001) |
.086 | |
|
HR (1/min) |
1.015 (1.009‐1.022) |
<.001 |
1.014 (1.008‐1.019) |
<.001 | |
|
Z‐MAP (mm Hg) |
1.097 (1.014‐1.187) |
.020 |
1.094 (1.028‐1.164) |
.005 | |
|
Z‐PP (mm Hg) |
1.330 (1.234‐1.433) |
<.001 |
1.353 (1.276‐1.436) |
<.001 | |
|
Z‐DP (mm Hg/min) |
1.261 (1.174‐1.355) |
<.001 |
1.257 (1.187‐1.330) |
<.001 | |
|
Z‐SBP (mm Hg) |
1.228 (1.136‐1.327) |
<.001 |
1.237 (1.164‐1.315) |
<.001 | |
|
Z‐DBP (mm Hg) |
0.959 (0.886‐1.039) |
.308 |
0.946 (0.888‐1.008) |
.086 | |
|
Z‐HR (1/min) |
1.191 (1.104‐1.285) |
<.001 |
1.171 (1.103‐1.244) |
<.001 | |
Hazard ratios were calculated by Cox proportional hazards regression.
Abbreviations: DBP, diastolic blood pressure; DP, double product; HR, heart rate; MAP, mean arterial pressure; PP, pulse pressure; SBP, systolic blood pressure.
Figure 1.
Kaplan–Meier curves for cardiovascular mortality according to 4 different PP: pulse pressure (A) and DP: double product (B) categories
In order to make the quantitative associations of these blood pressure parameters, of heart rate, and especially of DP (with a high numerical value) with CV mortality and all‐cause mortality better comparable, we performed a z‐standardization.
Increases of one standard deviation of SBP, MAP, PP, HR, and DP were associated with increases in CV mortality of 22.8%, 9.7%, 33.0%, 19.1%, and 26.1%, respectively. Similar findings were observed for all‐cause mortality (Table 3).
Besides the results in the whole patient cohort which demonstrate that the derived parameters PP and DP, in addition to the well‐known parameters SBP, MAP, and heart rate, are even stronger predictors of CV mortality and all‐cause mortality, we focused our further analysis on two clinically important subgroups, that is, patients with coronary artery disease and patients with heart failure.
3.2. Coronary artery disease subgroup
Similarly to the whole LURIC cohort, there was a significant association of the z‐standardized PP with the hazard ratio for CV mortality in all categories of CAD. The extent of CV mortality risk did not seem to be affected by the degree of pre‐existing CAD (3‐vessel CAD with a hazard ratio of 1.251, insignificant CAD with a hazard ratio of 1.336) (Figure 2). HR and DP showed similar associations. However, both SBP and MAP showed stronger association with CV mortality in the presence of more severe CAD. For SBP, only patients with a 2‐ or 3‐vessel CAD showed a significantly increased hazard ratio of 1.227 or 1.210 and for MAP only patients with a 3‐vessel disease (hazard ratio 1.134).
Figure 2.
Hazard ratio for CV mortality in patients with different degrees of CAD: coronary artery disease in association with PP: pulse pressure, DP: double product, HR: heart rate, SBP: systolic blood pressure, MAP: mean arterial pressure
In addition to the Cox regression model, we constructed Kaplan–Meier curves for patients with different degrees of CAD in different PP categories (Figure 3). Significant differences according to pulse pressure categories were confirmed with the exception of patients with coronary 1‐vessel disease.
Figure 3.
Kaplan–Meier diagrams for CV mortality in patients with different CAD: coronary artery disease severity grades according to four different PP: pulse pressure categories
3.3. Heart failure subgroup
Similarly to the CAD subgroup analysis, we performed Cox regression and Kaplan–Meier tests for patients with heart failure, specifically HFrEF (heart failure with reduced ejection fraction) vs. HFpEF (heart failure with preserved ejection fraction).
Noteworthy, z‐PP and z‐DP showed statistically significant associations with CV mortality for all HF subgroups in contrast to SBP, MAP, and heart rate (Figure 4). Specifically, SBP showed a weak association with CV mortality in HFrEF patients which was not significantly different from unity (p = .069), while the hazard ratio for patients with HFpEF and no HF was highly significant and MAP was only associated with CV mortality in patients without HF (Figure 4). Heart rate did not show any significant associations (Figure 4). Furthermore, we report Kaplan–Meier curves for patients with different types of heart failure and no heart failure according to PP categories (Figure 5).
Figure 4.
Hazard ratio for cardiovascular mortality in patients with HFrEF: heart failure with reduced ejection fraction, HFpEF: heart failure with preserved ejection fraction and without HF: heart failure in association with PP: pulse pressure, DP: double product, HR: heart rate, SBP: systolic blood pressure, MAP: mean arterial pressure
Figure 5.
Kaplan–Meier diagrams for CV mortality in patients with HFrEF: heart failure with reduced ejection fraction, HFpEF: heart failure with preserved ejection fraction and without HF: heart failure according to four different categories of PP: pulse pressure
3.4. Adjusted models
After adjustment for sex, BMI, diabetes mellitus, and eGFR (model 1), significant results for all‐cause and CV mortality could be demonstrated for all examined parameters apart from MAP. After adding age to the adjustment model (model 2), significant hazard ratios for DP and HR remained (Table 4).
Table 4.
Association of baseline z‐standardized hemodynamic parameters with CV mortality and all‐cause mortality in two different adjusted models given as hazard ratio and 95% confidence interval
| All‐cause mortality | Cardiovascular mortality | |||
|---|---|---|---|---|
| HR (95% CI) | p‐Value | HR (95%CI) | p‐Value | |
| Adjusted for sex, BMI, diabetes, eGFR (model 1) | ||||
| Z‐PP | 1.25 (1.17‐1.32) | <.001 | 1.20 (1.11‐1.29) | <.001 |
| Z‐DP | 1.18 (1.11‐1.25) | <.001 | 1.16 (1.08‐1.25) | <.001 |
| Z‐SBP | 1.15 (1.08‐1.23) | <.001 | 1.12 (1.03‐1.21) | .006 |
| Z‐MAP | 1.04 (0.98‐1.11) | .176 | 1.03 (0.95‐1.11) | .480 |
| Z‐HR | 1.13 (1.07‐1.20) | <.001 | 1.14 (1.06‐1.23) | .001 |
| Adjusted for age, sex, BMI, diabetes, eGFR (model 2) | ||||
| Z‐PP | 1.04 (0.97‐1.11) | .269 | 1.01 (0.92‐1.10) | .880 |
| Z‐DP | 1.11 (1.04‐1.18) | .001 | 1.10 (1.01‐1.19) | .020 |
| Z‐SBP | 0.98 (0.91‐1.05) | .518 | 0.96 (0.88‐1.04) | .335 |
| Z‐MAP | 0.94 (0.88‐1.00) | .048 | 0.93 (0.86‐1.01) | .089 |
| Z‐HR | 1.18 (1.11‐1.26) | <.001 | 1.19 (1.10‐1.29) | <.001 |
Hazard ratios were calculated by Cox proportional hazards regression for z‐standardized data.
Abbreviations: DP, double product; MAP, mean arterial pressure; PP, pulse pressure; R, heart rate; SBP, systolic blood pressure.
4. DISCUSSION
The novel information of this study is that, especially in a CV high‐risk cohort, both PP and DP are stronger predictors for CV and all‐cause mortality than SBP, DBP, and MAP. A high PP resulting mainly from a low DBP at a normal or high SBP indicates stiffness of the large arteries, while a high PP resulting from an increased SBP and a DBP in the normal range points to a rise in peripheral vascular resistance. 15 Therefore, a mean SBP of 141 mm Hg and mean DBP of 81 mm Hg in our study point to a predominantly increased peripheral vascular stiffness.
In the Framingham heart study, Franklin et al could show that neither SBP nor DBP was superior to PP in predicting CAD risk in the general population. 1 Beyond that, our results show that PP is not only a powerful predictor of cardiovascular and all‐cause mortality in a cardiovascular risk population but also in patients with CAD, where both PP and DP were more powerful than SBP, MAP, DBP, and HR.
Our results therefore are in line with the results of the Framingham heart study, in which CV events were more strongly related to pulsatile stress during the ejection because of isolated systolic hypertension than the steady‐state stress due to increased resistance in combined systolic–diastolic hypertension. It seems that pulsatile stress is more harmful regarding CAD risk 1 and CV mortality, respectively, all‐cause mortality.
The gradient of peripheral to central arterial PP decreases during the lifespan because of aging processes and enlargement of central PP by early wave reflection. 16 , 17 Thereby, PP derived from a brachial artery cuff measurement can be regarded as more representative for central PP in the elderly than in the younger patients. 1 Therefore, in our study with an average age of 62.66 years, the peripherally measured PP can be considered as reasonably representative of the centrally measured PP.
In a recently published study of 43 006 first renal transplant recipients, we have shown that PP has a stronger predictive value than SBP or DBP separately for 10‐year death‐censored graft and patient survival. 4 Similarly, a PP higher than 70 mm Hg was significantly associated with progression to end‐stage renal disease in the RENAAL study. 18 Overall, it remains unclear whether PP is mediating CV or renal disease or is a mere marker of CV or renal disease.
In a recent study, preoperative pulse pressure greater than 62 mm Hg significantly predicted postoperative myocardial injury, notably independent from systolic blood pressure. 19 Like our current study, that study highlights the importance of PP in predicting CV events compared to systolic blood pressure.
A large metaanalysis of fourteen studies involving 510,456 subjects from the general population found a pooled risk ratio for CV and all‐cause mortality per 10 mm Hg pulse pressure rise of 1.13 and 1.09. 20 In addition, a study with 7336 hypertensive patients showed that pulse pressure greater than 60 mm Hg was significantly associated with the prevalence of carotid plaque and left ventricular hypertrophy. Furthermore, these patients had a 57% increased risk for major CV events. 21 These two large studies done in the general population or in hypertensive patients are in line with our findings in a high CV risk population undergoing coronary angiography.
Remarkably, in heart failure patients, be it HFpEF or HFrEF, CV mortality was predicted better by PP compared to all other blood pressure parameters. MAP showed no significant correlation at all, neither in HFpEF nor in HFrEF patients. SBP was associated with CV mortality in patients without heart failure and patients with HFpEF, but showed no significant association in HFrEF patients. The low predictive power for CV and all‐cause mortality of SBP and MAP may be due to the intensive treatment with antiheart failure medication, that also lowers blood pressure to similar levels in most HF patients.
Our findings emphasize the impact of vascular stiffness and pulsatile stress on CV mortality in HF patients. Interestingly, also DP predicted CV mortality in HFrEF patients in a stronger way, while SBP and HR alone showed no significant association. Furthermore, in 3‐vessel CAD patients DP was a stronger predictor for CV mortality than SBP or HR alone. Therefore, the cardiac workload given as DP should be considered in clinical practice, especially in patients suffering from severe CAD and HF.
A 10‐year lasting observational study using ambulatory blood pressure monitoring (ABPM) proved that 24‐hour PP and 24‐hour systolic BP at baseline predicted mortality with a HR of 1.25 and a HR of 1.07, respectively. 22 Furthermore, it could be shown that a well‐controlled daytime blood pressure at baseline was associated with a lower increase in 24‐hour pulse pressure after 10 years of follow‐up. 22 These results show not only that PP predicts mortality better than SBP, but also that poorly controlled blood pressure leads to a significantly higher increase in PP over time. Interestingly, these 24‐h ambulatory blood pressure measurements and our single office blood pressure measurements seem to provide similar results in terms of mortality prediction.
Of course, SBP and DBP are influenced to different degrees by any antihypertensive medication. However, if SBP and DBP are considered individually, the confounding effect with antihypertensive medication can be assumed to be higher than for PP. DP as a marker of myocardial workload allows to estimate the corresponding oxygen consumption. The combination of two markers, that is, HR and SBP, may be more robust and can therefore better represent the hemodynamic situation of a patient than the individual parameters. This may be particularly important for CAD patients.
In the adjustment model 1, all parameters remained significant except for MAP. Interestingly, the addition of age as a confounder in model 2 led to a loss of significance for PP and SBP. Therefore, it has to be assumed that treated PP is not an independent predictor of mortality, but strongly correlates with age. In contrast to the Framingham Heart Study, we did not examine a general population but a patient cohort in which 86.8% of the participants were treated with antihypertensive medication. PP may integrate the risk for cardiovascular and all‐cause mortality that is mainly due to age, presence of diabetes, obesity, and CKD in a high cardiovascular risk population under antihypertensive treatment and may as such serve as a surrogate for mortality risk. In contrast, PP per se confers increased cardiovascular and all‐cause mortality risk in untreated or less frequently/intensely treated hypertensive patient populations from the general population such as in the Framingham Heart Study, The Campania Salute Network and in UK Biobank participants, 1 , 5 , 21 as it also has been summarized in a metanalysis of prospective observational studies. 20
Regarding the limitations of our study, it has to be mentioned that we only took blood pressure measurements at the beginning of the study and did not perform additional measurements during follow‐up. Since only patients with German ancestry were included, the study results cannot be extrapolated to other populations. On the other hand, the strength of this study lies in the large number of patients and the relatively long follow‐up time of almost 10 years.
PP can be decreased by antihypertensive drugs, especially by thiazide diuretics, 23 , 24 while DP is also influenced by common antihypertensives and especially by beta‐blockers. Our results suggest that it should be considered to enable blood pressure monitors also to display pulse pressure and double product for more accurate risk stratification. In addition, further studies are needed to delineate thresholds for PP and DP that indicate increasing risk in specific patient subgroups.
In conclusion, we provide evidence that not only the standard blood pressure parameters SBP and MAP predict CV mortality. Rather PP and DP are even more powerful predictors of CV and all‐cause mortality in the LURIC cohort of patients with medium‐to‐high cardiovascular risk that underwent coronary angiography. PP and DP are superior predictors of a higher CV mortality in patients with heart failure and coronary artery disease. While DP proved to be an independent predictor of cardiovascular and all‐cause mortality also in multivariate analysis, PP was no independent predictor in a patient cohort with widespread antihypertensive treatment (>85%). In this patient, cohort‐treated PP is mainly associated with age and as such may serve as a surrogate risk marker.
CONFLICT OF INTEREST
Dr rer. nat. Marcus E. Kleber reports lecture fees from Bayer outside the submitted work. Prof. Dr med. W. März reports grants from Siemens Healthineers, grants and personal fees from Aegerion Pharmaceuticals, grants and personal fees from AMGEN, grants from Astrazeneca, grants and personal fees from Sanofi, grants and personal fees from Alexion Pharmaceuticals, grants and personal fees from BASF, grants and personal fees from Abbott Diagnostics, grants and personal fees from Numares AG, grants and personal fees from Berlin‐Chemie, grants and personal fees from Akzea Therapeutics, grants from Bayer Vital GmbH, grants from bestbion dx GmbH, grants from Boehringer Ingelheim Pharma GmbH Co KG, grants from Immundiagnostik GmbH, grants from Merck Chemicals GmbH, grants from MSD Sharp and Dohme GmbH, grants from Novartis Pharma GmbH, grants from Olink Proteomics, other from Synlab Holding Deutschland GmbH, all outside the submitted work. Prof. Dr med. B. K. Krämer reports lecture fees and/or advisory board memberships and/or study participation from Astellas, Bayer, Boehringer Ingelheim, Chiesi, Riepharm, Pfizer, Servier, and Vifor Pharma, all outside the submitted work. He is the past president of the German Hypertension Society DHL. For the remaining authors, none were declared.
AUTHOR CONTRIBUTIONS
Babak Yazdani and Bernhard K. Krämer designed research, wrote the first draft of the manuscript, did the literature research, and interpreted the results. Gökhan Yücel, Urs Benck, Marcus E. Kleber, Graciela E. Delgado, Winfried März, and Bernd Krüger corrected the manuscript and gave important advice and intellectual input on the evaluation and interpretation of the results. Marcus E. Kleber and Graciela E. Delgado made statistical analyses. Winfried März is principal investigator of the LURIC study and was instrumental in performing of the study and provided the data for evaluation.
Yazdani B, Kleber ME, Yücel G, et al. Association of double product and pulse pressure with cardiovascular and all‐cause mortality in the LURIC study. J. Clin. Hypertens. 2020;22:2332–2342. 10.1111/jch.14067
Funding information
This work has received funding from the European Union's Horizon 2020 research and innovation program under the ERA‐Net Cofound action N° 727565 (OCTOPUS project) and the German Ministry of Education and Research (grant number 01EA1801A).
REFERENCES
- 1. Franklin SS, Khan SA, Wong ND, Larson MG, Levy D. Is pulse pressure useful in predicting risk for coronary heart Disease? The Framingham heart study. Circulation. 1999;100(4):354‐360. [DOI] [PubMed] [Google Scholar]
- 2. Domanski M, Mitchell G, Pfeffer M, et al. Pulse pressure and cardiovascular disease‐related mortality: follow‐up study of the Multiple Risk Factor Intervention Trial (MRFIT). JAMA. 2002;287(20):2677‐2683. [DOI] [PubMed] [Google Scholar]
- 3. Carpenter MA, John A, Weir MR, et al. BP, cardiovascular disease, and death in the Folic Acid for Vascular Outcome Reduction in Transplantation trial. J Am Soc Nephrol. 2014;25(7):1554‐1562. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Krüger B, Döhler B, Opelz G, Krämer BK, Süsal C. Collaborative Transplant Study. Pulse Pressure and Outcome in Kidney Transplantation: Results From the Collaborative Transplant Study. Transplantation. 2019;103(4):772‐780. [DOI] [PubMed] [Google Scholar]
- 5. Said MA, Eppinga RN, Lipsic E, Verweij N, van der Harst P. Relationship of Arterial Stiffness Index and Pulse Pressure with Cardiovascular Disease and Mortality. J Am Heart Assoc. 2018;7(2):e007621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Thorin‐Trescases N, de Montgolfier O, Pinçon A, et al. Impact of pulse pressure on cerebrovascular events leading to age‐related cognitive decline. Am J Physiol Heart Circ Physiol. 2018;314(6):H1214‐H1224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Hartaigh B, Bosch JA, Pilz S, et al. Influence of resting heart rate on mortality in patients undergoing coronary angiography (from the Ludwigshafen Risk and Cardiovascular Health [LURIC] study). Am J Cardiol. 2012;110(4):515‐520. [DOI] [PubMed] [Google Scholar]
- 8. Ansari M, Javadi H, Pourbehi M, et al. The Association of Rate Pressure Product (RPP) and Myocardial Perfusion Imaging (MPI) Findings: A Preliminary Study. Perfusion. 2012;27(3):207‐213. [DOI] [PubMed] [Google Scholar]
- 9. Domka‐Jopek E, Jopek A, Bejer A, Lenart‐Domka E, Walawski G. The Importance of the Double Product in the Six‐Minute Walk Test to Predict Myocardial Function. Biomed Res Int. 2018;4(2018):3082690. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Sadrzadeh Rafie AH, Sungar GW, Dewey FE, Hadley D, Myers J, Froelicher VF. Prognostic value of double product reserve. Eur J Cardiovasc Prevention Rehab. 2008;15(5):541‐547. [DOI] [PubMed] [Google Scholar]
- 11. Schutte R, Thijs L, Asayama K, et al. International Database on Ambulatory blood pressure in relation to Cardiovascular Outcomes (IDACO) Investigators. Double product reflects the predictive power of systolic pressure in the general population: evidence from 9,937 participants. Am J Hypertens. 2013;26(5):665‐672. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Verma AK, Sun JL, Hernandez A, et al. Rate pressure product and the components of heart rate and systolic blood pressure in hospitalized heart failure patients with preserved ejection fraction: Insights from ASCEND‐HF. Clin Cardiol. 2018;41(7):945‐952. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Winkelmann BR, März W, Boehm BO, et al. LURIC Study Group (LUdwigshafen RIsk and Cardiovascular Health). Rationale and design of the LURIC study–a resource for functional genomics, pharmacogenomics and long‐term prognosis of cardiovascular disease. Pharmacogenomics. 2001;2(1 suppl 1):S1‐73. [DOI] [PubMed] [Google Scholar]
- 14. Paulus WJ, Tschöpe C, Sanderson JE, et al. How to diagnose diastolic heart failure: a consensus statement on the diagnosis of heart failure with normal left ventricular ejection fraction by the Heart Failure and Echocardiography Associations of the European Society of Cardiology. Eur Heart J. 2007;28(20):2539‐2550. [DOI] [PubMed] [Google Scholar]
- 15. Berne RM, Levy MN. Cardiovascular Physiology. St Louis, MO: CV Mosby; 1992:113‐144. [Google Scholar]
- 16. Benetos A, Laurent S, Hoeks AP, Boutouyrie PH, Safar ME. Arterial alterations with aging and high blood pressure. A noninvasive study of carotid and femoral arteries. Arterioscler Thromb. 1993;13(1):90‐97. [DOI] [PubMed] [Google Scholar]
- 17. Kawasaki T, Sasayama S, Yagi S, Asakawa T, Hirai T. Non‐invasive assessment of the age related changes in stiffness of major branches of the human arteries. Cardiovasc Res. 1987. 21(9):678‐687. [DOI] [PubMed] [Google Scholar]
- 18. Bakris GL, Weir MR, Shanifar S, et al. RENAAL Study Group. Effects of blood pressure level on progression of diabetic nephropathy: results from the RENAAL study. Arch Intern Med. 2003;163(13):1555‐1565. [DOI] [PubMed] [Google Scholar]
- 19. Abbott TEF, Pearse RM, Archbold RA, et al. Association between preoperative pulse pressure and perioperative myocardial injury: an international observational cohort study of patients undergoing non‐cardiac surgery. Br J Anaesth. 2017;119(1):78‐86. [DOI] [PubMed] [Google Scholar]
- 20. Zhao L, Song Y, Dong P, Li Z, Yang X, Wang S. Brachial pulse pressure and cardiovascular or all‐cause mortality in the general population: a meta‐analysis of prospective observational studies. J Clin Hypertens (Greenwich). 2014;16(9):678‐685. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Mancusi C, Losi MA, Izzo R, et al. Higher pulse pressure and risk for cardiovascular events in patients with essential hypertension: The Campania Salute Network. Eur J Prev Cardiol. 2018;25(3):235‐243. [DOI] [PubMed] [Google Scholar]
- 22. Balietti P, Spannella F, Giulietti F, et al. Ten‐year changes in ambulatory blood pressure: The prognostic value of ambulatory pulse pressure. J Clin Hypertens (Greenwich). 2018;20(9):1230‐1237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Cushman WC, Materson BJ, Williams DW, Reda DJ. Pulse pressure changes with six classes of antihypertensive agents in a randomized, controlled trial. Hypertension. 2001;38(4):953‐957. [DOI] [PubMed] [Google Scholar]
- 24. Chang JJ, Luchsinger JA, Shea S. Antihypertensive medication class and pulse pressure in the elderly: analysis based on the third National Health and Nutrition Examination Survey. Am J Med. 2003;115(7):536‐542. [DOI] [PubMed] [Google Scholar]