Effect of Antihypertensive Therapy on Incident Stroke in Cohorts With Prehypertensive Blood Pressure Levels
A Meta-Analysis of Randomized Controlled Trials
Background and Purpose—Compared with normotensive individuals, there is a higher incidence of stroke in patients with hypertensive, as well as prehypertensive, blood pressure levels (ie, 120–139/80–89 mm Hg). Although several studies have shown that blood pressure reduction in hypertensive patients reduces the incidence of cardiovascular events, including stroke, it is still unknown whether treatment of prehypertensive blood pressure levels has a similar effect. We sought to determine whether reduction in blood pressure in the prehypertensive range reduces the incidence of stroke by performing a meta-analysis of randomized trials comparing an antihypertensive drug against placebo in cohorts with prehypertensive baseline blood pressure levels.
Methods—Randomized controlled trials performed with the 95 different antihypertensive agents available in the market were identified using MEDLINE, returning a total of 2852 results. Exclusion criteria included: average blood pressure of ≥140/90 mm Hg at baseline, crossover studies, and lack of a control group receiving placebo.
Results—A total of 16 trials involving 70 664 patients were included. Patients randomized to the active treatment arm had a statistically significant 22% reduction in the risk of stroke compared with placebo, with little heterogeneity among the trials (I2, 18.0%; RR, 0.78 [95% CI, 0.71–0.86]; P<0.000001). To prevent 1 stroke, 169 patients had to be treated with a blood-pressure-lowering medication for an average of 4.3 years.
Conclusions—The risk of stroke is significantly reduced with antihypertensive therapy in cohorts with prehypertensive blood pressure levels. These findings can have important clinical implications.
Hypertension is a modifiable risk factor for cardiovascular events. Treatment of hypertension results in a significant reduction in major cardiovascular events, including stroke, myocardial infarction, and heart failure.1–3 In fact, suboptimal blood pressure level is the number 1 attributable risk factor not only for stroke, but also for all-cause death throughout the world.4 Current guidelines recommend lowering blood pressure to ≤140/90 mm Hg (with lower targets in some patients, including those with diabetes and chronic kidney disease). The seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure identified a new category of blood pressure of 120 to 139 mm Hg systolic and 80 to 89 mm Hg diastolic blood pressure and designated it prehypertension.5 Prehypertension is highly prevalent, occurring in up to 40% of individuals depending on age, sex, and ethnicity of the studied population.6,7 Prehypertension usually progresses to hypertension over time,7 and even without such progression, it is associated with increased incidence of stroke and myocardial infarction.6 Importantly, professional societies do not currently recommend pharmacological treatment for prehypertension per se, based on the lack of prospective, randomized trials examining the effects of antihypertensive therapy on reducing cardiovascular events specifically in prehypertensives. However, a myriad of placebo-controlled trials have been performed in cohorts with prehypertensive average blood pressure levels for the treatment of other conditions (eg, atherosclerosis) with antihypertensives or for testing other hypotheses with these drugs (eg, whether some antihypertensives can also prevent diabetes).8–23 Given the strong association of stroke with blood pressure, our aim was to examine the effects of antihypertensive therapy on incident stroke using data from these clinical trials.
The aim of this meta-analysis was to include all published data about stroke incidence from randomized controlled trials comparing an antihypertensive drug against placebo in cohorts with average baseline blood pressure levels within the prehypertensive range. A systematic search of MEDLINE for randomized controlled human subject trials reported in the English language published before September 2010 was performed to identify morbidity and mortality trials for all antihypertensives. A separate search was performed for each antihypertensive drug available in the US market. The full search terms and a complete list of the 95 different antihypertensive agents examined are presented in the online-only Supplemental Appendix A (http://aha.strokejournal.org).
Predefined inclusion criteria were: average baseline systolic blood pressure between 120 and 140 mm Hg and baseline diastolic blood pressure <90 mm Hg in both active treatment and placebo groups, use of an antihypertensive drug and at least 1 comparison group treated with placebo, and reporting of stroke incidence (ie, fatal and nonfatal combined) in the antihypertensive arm(s) and in the placebo arm. Exclusion criteria were: lack of baseline blood pressure reporting, crossover studies, lack of a control group receiving placebo, and conditions in which certain antihypertensive drugs (such as angiotensin-converting enzyme [ACE] inhibitors, β-blockers, and angiotensin-receptor blockers) are known to change the natural history of disease and reduce cardiovascular events through mechanisms not necessarily related to blood pressure reduction (ie, systolic heart failure,24–26 acute myocardial-infarction,27–30 diabetic nephropathy, and other types of nephropathy31,32). There were no criteria for trial duration or sample size. Every search result retrieved with the search strategy presented in the online-only Supplemental Appendix A was examined carefully for inclusion and exclusion criteria. Of the 2852 potentially relevant articles, a total of 16 trials fulfilled the predefined criteria and were included in the meta-analysis.
Data extraction from source documents was independently performed by 2 of the investigators (A.S. and S.M.D.) and was verified. The following information was systematically extracted: baseline systolic and diastolic blood pressure, final systolic and diastolic blood pressure, total number of patients in each arm, trial duration, stroke incidence, myocardial infarction incidence, all-cause death, age, sex, race, hypertension history, diabetes history, cardiovascular disease history, baseline low-density lipoprotein and total cholesterol, history of dyslipidemia, smoking status, baseline medical therapy (aspirin, statin, β-blockers, ACE- inhibitors, angiotensin-receptor blockers [ARBs], calcium channel blockers), adverse events, medication adherence, and drop-out rates.
To assess for the presence of publication bias, the Begg's rank correlation method was used. This was executed by calculating Kendall's tau. In addition, a funnel plot was generated.
Cochran's Q statistic was calculated to assess for statistical heterogeneity across trials. Heterogeneity was defined as an alpha value <0.10, and the degree of heterogeneity was represented by I2 values. Risk ratios (rather than hazard ratios) were calculated by using the number of incident strokes and number of patients randomized in each arm of all trials (ie, 2×2 tables), because data regarding the timing of each stroke event was unavailable. Meta-analytic risk ratios were calculated as the measure of association. In cases of 0 events in an arm of a trial, 0.5 was added to the number of events in each arm. Fixed effect models were used unless there was evidence of heterogeneity, in which case random effects models were used. A sensitivity analysis limited to studies with an average baseline systolic blood pressure <130 mm Hg and baseline diastolic blood pressure <85 mm Hg in both groups was also performed. Additional 1-study out analyses were performed to assess whether the results were driven by a single trial. The effect of potential moderator variables, including baseline stroke risk, baseline systolic blood pressure, difference in final systolic blood pressures between the 2 arms at study end, history of diabetes, age, and frequency of statin (or other antilipid agent) use were examined with meta-regression analyses. The baseline stroke risk was assessed by determining the yearly incidence of strokes in the control arm of each trial. In addition, trials were grouped according to drug class studied and risk ratios for stroke were also calculated for angiotensin-converting enzyme inhibitors, angiotensin-receptor blockers, and calcium channel blockers separately.
Probability values less than 0.05 were considered statistically significant. All reported probability values are 2-sided. Data were analyzed with Comprehensive Meta Analysis Version 2.2048 (Biostat Inc).
The MEDLINE search repeated a total of 95 times for each antihypertensive drug using the search terms in Supplemental Appendix A yielded 2852 results (Figure 1). Of these, 2677 trials were excluded because they did not meet systolic and/or diastolic blood pressure criteria. Another 54 trials were excluded because they involved systolic heart failure, acute myocardial infarction, diabetic nephropathy, or other types of kidney disease patients. Another 100 trials were excluded because they were substudies without additional information. Three crossover studies were also excluded. Two trials were excluded because they reported only fatal stroke. Ultimately, the following 16 trials enrolling a total of 70 664 patients met both the inclusion and exclusion criteria: Diabetes Reduction Assessment with Ramipril and Rosiglitazone Medication (DREAM),9 Nateglinide and Valsartan in Impaired Glucose Tolerance Outcomes Research (NAVIGATOR),15 Heart Outcomes Prevention Evaluation (HOPE),23 European Trial on Reduction of Cardiac Events with Perindopril in Patients with Stable Coronary Artery Disease (EUROPA),13 Comparison of Amlodipine versus Enalapril to Limit Occurrences of Thrombosis (CAMELOT),16 Gruppo Italiano per lo Studio della Sopravvivenza nell'Infarto Miocardico-Atrial Fibrillation (GISSI-AF),12 Irbesartan in Heart Failure with Preserved Ejection Fraction Study (I-PRESERVE),14 Perindopril in Elderly People with Chronic Heart Failure (PEP-CHF),11 Candesartan in Heart Failure: Assessment of Reduction in Mortality and Morbidity (CHARM-PRESERVED),22 A Coronary Disease Trial Investigating Outcome with Nifedipine GITS (ACTION),18 Simvastatin/Enalapril Coronary Atherosclerosis Trial (SCAT),21 Prevention of Renal and Vascular Endstage Disease Intervention Trial (PREVEND IT),8 Ischemia Management with Accupril Post-Bypass Graft via Inhibition of the Converting Enzyme (IMAGINE),19 Prevention of Events with Angiotensin Converting Enzyme Inhibition (PEACE),10 Prospective Randomized Evaluation of the Vascular Effects of Norvasc Trial (PREVENT),17 and Appropriate Blood Pressure Control in Diabetes (ABCD).20
The study characteristics of the randomized controlled trials included are presented in Table 1. Eight trials studied angiotensin-converting enzyme inhibitors (ramipril, perindopril, enalapril, fosinopril, quinapril, and trandolapril),8–11,13,19,21,23 four trials studied angiotensin-receptor blockers (valsartan, irbesartan, and candesartan),12,14,15,22 2 trials studied calcium channel blockers (extended release nifedipine and amlodipine),17,18 1 trial had both calcium channel blocker and angiotensin-converting enzyme inhibitor arms (amlodipine and enalapril),16 and 1 trial used an angiotensin-converting enzyme inhibitor and/or a calcium channel blocker.20 Both patients and investigators were blinded to treatment allocation in all trials, with the exception of I-PRESERVE and ABCD, which were single-blind. Baseline characteristics of patients enrolled in the trials are presented in Table 2, and concomitant medications are presented in the online-only Supplemental Appendix C. There was no major imbalance within the arms of trials with regards to baseline blood pressure, age, sex, smoking status, statin use, diabetes prevalence, hypertension history, cardiovascular myocardial infarction history, baseline aspirin, β-blocker, ACE-I/ARB, and calcium channel blocker use. There was also no major difference in baseline low-density lipoprotein, total cholesterol, and history of dyslipidemia between trial arms (data not shown). At the time of trial completion, adherence rates in the active drug arms were between 60% and 95% and were comparable in the placebo arm of each trial. Dropout rates, defined as those who discontinued the study drug because of an adverse event, withdrawal of consent, or were lost-to-follow-up, were overall very low and also comparable in the drug and placebo groups. Adherence and dropout rates for each trial are available in the online-only Supplemental Appendix B. As compared with placebo, active drug treatment was accompanied by a greater reduction in both average systolic and average diastolic blood pressure in all of the trials reporting on change in blood pressure over time. The average baseline and follow-up blood pressure levels in each trial are presented in Table 3. Follow-up blood pressures weighted by trial sample size for active drug and placebo arms are 130.5/76.5 and 134.2/78.4 mm Hg, respectively.
Quantitative Data Synthesis
Publication bias was examined by the Begg's rank correlation method. The test statistic for Begg's approach, Kendall's tau with continuity correction, was nonsignificant (P=0.33); this indicates no evidence of publication bias with this test. A funnel plot for publication bias is presented in Supplemental Appendix D.
One trial differentiated between fatal and nonfatal stroke, whereas all other studies reported only total stroke.16 HOPE, PEP-CHF, CHARM-Preserved, and ABCD trials had a stroke incidence of >1% per year in the placebo group. In all other studies, the incidence of stroke was <1% per year in the placebo group.
In 14 of 16 included trials, there were fewer strokes in the antihypertensive arm as compared with the placebo arm. On meta-analysis, randomization to active treatment led to a statistically significant reduction in the risk of stroke compared with placebo, with no significant heterogeneity among the trials (I2, 18.0%; RR, 0.78; [95% CI, 0.71–0.86]; P<0.000001; Figure 2). One-study-out analysis demonstrated that the risk reduction remained statistically significant with any 1 study being excluded at a time (P<0.01 in all occasions), confirming that a single study was not driving the results. Concomitant exclusion of the 3 major ACE-inhibitor trials in patients with coronary artery disease or high cardiovascular risk (HOPE, EUROPA, and PEACE) also did not alter stroke reduction (I2, 18.3%; RR, 0.78 [95% CI, 0.68–0.89]; P<0.001).
To assess whether the reduction in stroke with antihypertensive therapy extended to patients with even lower baseline blood pressure levels, an analysis limited to trials with an average baseline blood pressure less than 130/85 mm Hg was performed (ie, CAMELOT, SCAT, PREVEND IT, IMAGINE, and PREVENT trials). When analysis was limited to the above trials, there was still a statistically significant reduction in the risk of stroke compared with placebo patients (I2, 34.1%; RR, 0.65 [95% CI, 0.43–0.995]; P=0.048; Figure 3).
On meta-regression analysis, the magnitude of risk reduction in stroke was not related to the average systolic blood pressure at baseline, mean difference in final systolic and diastolic blood pressure, history of hypertension, mean age, prevalence of diabetes mellitus, or statin therapy at baseline.
In additional sensitivity analyses, the meta-analysis was repeated for each particular antihypertensive drug class. Accordingly, patients randomized to angiotensin-converting enzyme inhibitors (n=41 798) had a 25% decrease in the risk of stroke compared with placebo (I2, 32.0%; RR, 0.75 [95% CI, 0.66–0.86]; P=0.00003). Patients randomized to calcium channel blockers (n=10 487) experienced a similar decrease in the risk of stroke compared with placebo (I2, 0%; RR, 0.75 [95% CI, 0.57–0.97]; P=0.03). Analysis limited to angiotensin-receptor blockers (n=17 899) revealed a trend toward reduction in incident strokes (I2, 0%; RR, 0.85 [95% CI, 0.72–1.01]; P=0.07).
Based on the observed differences in risks, values of number needed to treat were calculated. The absolute risk of stroke in the treatment arm was 2.01%. The absolute risk of stroke in the placebo arm was 2.61%. To prevent 1 stroke, 169 patients with prehypertensive blood pressure levels had to be treated with a blood-pressure-lowering medication for 4.3 years (weighted average duration of the included trials calculated by using mean or median follow-up weighed according to total number of patients in each trial).
The impact of antihypertensive therapy on other major cardiovascular end points (ie, myocardial infarction and cardiovascular death) was also examined. Randomization to active treatment did not lead to a statistically significant decrease in the risk of myocardial infarction compared with placebo (I2, 52.6%; random effects model used; RR, 0.91 [95% CI, 0.80–1.03]; P=0.14; Supplemental Appendix E). There was significant heterogeneity among the included trials, mostly originating from the ACE-inhibitor trials in patients with coronary artery disease or high cardiovascular risk (ie, HOPE and EUROPA trials), in which a significant reduction in the risk of myocardial infarction was observed in the treatment arm. The incidence of cardiovascular death was not significantly decreased in the active treatment arm compared with placebo (I2, 28.9%; fixed effects model used; RR, 0.94 [95% CI, 0.88–1.00]; P=0.07; Supplemental Appendix F). Again, there was a significant decrease in cardiovascular death in the HOPE trial and a trend for decrease in the EUROPA trial, leading to an overall trend for reduction in cardiovascular mortality on meta-analysis.
Adverse effects were reported in a nonuniform fashion, but hyperkalemia, renal failure, hypotension, and peripheral edema were usually more common with active treatment, where reported. Meta-analytic effect sizes could not be calculated given erratic reporting.
In this meta-analysis, we observed a statistically significant 22% reduction in incident strokes with antihypertensive treatment in cohorts with average blood pressure levels within the prehypertensive range, as compared with placebo (P<0.000001). This reduction was apparent even in trials where the average baseline blood pressure was <130/85 mm Hg. The reduction in stroke was evident among all the drug classes studied.
The historic definition of hypertension as a blood pressure of 140/90 mm Hg and above is arbitrary. Although there is a linear relationship between blood pressure over 140/90 mm Hg and overall cardiovascular events, the relationship between blood pressure and cardiovascular events extends to a systolic blood pressure of at least 115 mm Hg.1 Among cardiovascular events, stroke has the strongest relationship with blood pressure.33 The population attributable risk to hypertension for incident strokes is reported to be approximately 40%.33 The more recently introduced category of prehypertension is also associated with adverse cardiovascular events, especially stroke. A study of over 60 000 women who were followed over a period of 7 years showed that prehypertensives had an adjusted hazard ratio of 1.93 for stroke, 1.76 for myocardial infarction, and 1.58 for cardiovascular death.6 Regarding stroke, the Perindopril pROtection aGainst REcurrent Stroke Study (PROGRESS) trial found a significant risk reduction with antihypertensive therapy in patients with established cerebrovascular disease who were hypertensive, as well as in those who were not hypertensive (ie, average baseline blood pressure, 136/79 mm Hg).34 A subsequent post hoc analysis of this trial demonstrated that the lowest risk of stroke recurrence was among patients with the lowest follow-up blood pressure levels (median, 112/72 mm Hg), and the risk increased progressively with higher follow-up blood pressure levels.35 More recently, The Action to Control Cardiovascular Risk in Diabetes - Blood Pressure (ACCORD-BP) trial demonstrated a significant reduction in strokes in patients with intensive blood pressure management compared with standard therapy (average systolic blood pressure after 1 year of follow-up, approximately 119.3 versus 133.5 mm Hg).36 Given these findings and the strong and continuous relationship between blood pressure and stroke, we undertook the current meta-analysis of all randomized controlled trials of antihypertensive therapy in cohorts with prehypertensive average blood pressure levels and observed a highly statistically significant 22% reduction in incident stroke, which is consistent with the findings of the ACCORD-BP trial. There was little heterogeneity despite the different disease states and patient populations studied. This effect was observed across all the drug classes studied, including ACE-inhibitors, calcium channel blockers, and possibly angiotensin-receptor blockers. The meta-regression analysis showed that the magnitude of risk reduction with antihypertensive therapy in this group of patients was not related to average baseline blood pressure levels, baseline stroke risk, history of hypertension, average age, history of diabetes, or statin therapy. Importantly, meta-regression analyses also did not show a relationship between the magnitude of average blood-pressure-lowering and the magnitude of risk reduction. However, it should be pointed out that in these analyses, average blood pressure levels of the whole trial arms (and not of the individual patients) were used, thereby limiting the interpretation of the results of these meta-regression analyses. Moreover, lack of homogeneity of the trials in other regards (such as other baseline characteristics, concomitant therapies, and drug class studied) further limits the reliability of the results of the meta-regression analyses.
In this meta-analysis, we specifically excluded the trials studying conditions where antihypertensives have been shown to change the natural history of the disease and reduce cardiovascular events; this occurs through mechanisms unrelated to blood pressure reduction, most notably in systolic heart failure and myocardial infarction. With regards to stroke, because renin-angiotensin aldosterone system antagonists and β-blockers improve left ventricular structure and function in these conditions, reduction in strokes would be at least partially driven by reduction in cardioembolic strokes; this would lead to overestimation of the benefits of blood pressure reduction in these conditions. Thus, such trials were excluded from this meta-analysis. Conversely, the ACE inhibitor trials in patients with established atherosclerotic disease or high vascular risk (ie, HOPE, EUROPA, and PEACE trials) were included in this analysis given lack of consensus on whether the benefit observed in these trials is blood-pressure-dependent.37 However, in a sensitivity analysis where all 3 of these trials were concomitantly excluded, the results for risk reduction in stroke did not change (RR, 0.78; P<0.001). This suggests that the relationship between blood pressure and stroke is very strong, and blood pressure reduction within the prehypertensive range may reduce stroke risk regardless of the baseline cardiovascular risk.
Although in this meta-analysis there was no statistically significant reduction in myocardial infarction and cardiovascular death with antihypertensive therapy, there were signs of risk reduction for these end points. However, these trends were likely driven by the ACE-inhibitor trials in patients with established atherosclerotic disease or very high cardiovascular risk (ie, HOPE and EUROPA trials). Exclusion of the above trials caused the trend toward risk reduction to disappear. Many of the trials performed in patients with other conditions, such as impaired glucose tolerance (eg, NAVIGATOR), did not show any reduction in cardiovascular mortality or myocardial infarction with antihypertensive therapy. Therefore, trends of reduction in these conditions may not be applicable to other prehypertensive patient populations.
In this meta-analysis, the number needed to treat to prevent 1 stroke was 169 (average duration of treatment, 4.3 years). Although this number is somewhat larger than the number needed to treat for a hypertensive patient (eg, 118 for 5 years for a diastolic blood pressure between 90–110 mm Hg), number needed to treat is likely to vary substantially depending on the patient population studied.38 Also, the number needed to treat obtained from this meta-analysis for (169 patients for 4.3 years) should be viewed in the context of the number needed to treat of 642 patients for 5 years to prevent 1 stroke with statin therapy for primary prevention.39
Our meta-analysis is limited by lack of access to individual patient level data. For example, in the HOPE and NAVIGATOR trials, the average systolic blood pressure at baseline was reported to be 139 mm Hg; this means that close to half of the patients studied had a baseline blood pressure in the hypertensive range. This could have overestimated the benefit of antihypertensives in reducing the incidence of stroke. However, analysis limited to trials with baseline blood pressure <130/85 mm Hg still demonstrates a significant reduction in the incidence of stroke. Also, although the average blood pressure of the cohorts was within the prehypertensive range, many patients actually had a diagnosis of hypertension and with antihypertensive treatment antedating their enrollment to the clinical trials, their blood pressure was reduced to prehypertensive levels. Therefore, the risks reduction in stroke observed in the current analysis comes not only from treatment of truly prehypertensive individuals with no history of hypertension, but also from hypertensive individuals with prehypertensive blood pressure levels on antihypertensives. Because we did not have access to patient level data, the meta-analysis could not be limited to those without a history of hypertension. However, analyses of nonhypertensive individuals from the PEACE and Post-stroke Antihypertensive Treatment Study (PATS) trials also show a statistically significant reduction in strokes with antihypertensive therapy, supporting our findings.24,40
The included trials did not report drug related adverse events uniformly. However, in the reported studies, hyperkalemia, renal failure, and hypotension were more common with active antihypertensive therapy, signaling the possible harm associated with tighter blood pressure control. Hypotension is a particularly concerning adverse event in the elderly population. Thus, although the risk reduction in incident strokes in patients with prehypertensive blood pressure levels is well elucidated in our meta-analysis, the clinical implications are uncertain. With an estimated 53 million adults having prehypertension in the United States alone, the cost of treating these patients and the associated adverse events can be substantial. These costs and risks will have to be weighed against the benefit that is derived from treatment to explore the clinical implications fully.
The included trials did not uniformly report the number of fatal strokes, level of functional impairment after nonfatal stroke, or differentiate between ischemic and hemorrhagic strokes. Therefore, the type and severity of strokes prevented by antihypertensive therapy in those with prehypertensive blood pressure levels is not known.
This meta-analysis of over 70 000 patients shows that the risk of incident strokes is significantly reduced with antihypertensive therapy in cohorts with baseline blood pressure within the prehypertensive range. Among the included trials, there was little heterogeneity for risk reduction in incident strokes.
In the last 2 years, I.S. received only lecture honoraria from Ranbaxy; D.I.S. serves on advisory boards for Daiichi-Sankyo, Portola, and Schering-Plough, and has received lecture honoraria from Daiichi-Sankyo, Eli Lilly, Sanofi-Aventis, and Schering-Plough.
I.S. had full access to all the data used in the study and takes responsibility for the integrity of the data and accuracy of the analysis. Study concept and design: I.S.; acquisition of data: I.S., A.S., S.M.D.; analysis and interpretation of data: I.S., A.S., S.M.D.; drafting of the manuscript: I.S., A.S., V.N.; revision of the manuscript for content: I.S., A.S., V.N., S.M.D., D.I.S., J.C.F.; statistical analysis: I.S., S.M.D.
The online-only Data Supplement is available at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.111.636829/-/DC1.
- Received August 23, 2011.
- Accepted October 7, 2011.
- © 2012 American Heart Association, Inc.
- Wang JG,
- Staessen JA,
- Franklin SS,
- Fagard R,
- Gueyffier F
- Hsia J,
- Margolis KL,
- Eaton CB,
- Wenger NK,
- Allison M,
- Wu L,
- et al
- Asselbergs FW,
- Diercks GF,
- Hillege HL,
- van Boven AJ,
- Janssen WM,
- Voors AA,
- et al
- Cleland JG,
- Tendera M,
- Adamus J,
- Freemantle N,
- Polonski L,
- Taylor J
- Pitt B,
- Byington RP,
- Furberg CD,
- Hunninghake DB,
- Mancini GB,
- Miller ME,
- et al
- Rouleau JL,
- Warnica WJ,
- Baillot R,
- Block PJ,
- Chocron S,
- Johnstone D,
- et al
- Teo KK,
- Burton JR,
- Buller CE,
- Plante S,
- Catellier D,
- Tymchak W,
- et al
- Køber L,
- Torp-Pedersen C,
- Carlsen JE,
- Bagger H,
- Eliasen P,
- Lyngborg K,
- et al
- Arima H,
- Chalmers J,
- Woodward M,
- Anderson C,
- Rodgers A,
- Davis S,
- et al
National Heart, Lung, and Blood Institute; Biologic Specimen and Data Repository Information Coordinating Center. PEACE study formal data request 383. National Heart, Lung, and Blood Institute Web site. https://biolincc.nhlbi.nih.gov/login/?next=/requests/data-formal-request/383/. Retrieved March 20, 2011.