MedicineLakex

 

Treatment of hypertension in patients 80 years of age or older

The new england journal of medicine established in 1812 Treatment of Hypertension in Patients 80 Years Nigel S. Beckett, M.B., Ch.B., Ruth Peters, Ph.D., Astrid E. Fletcher, Ph.D., Jan A. Staessen, M.D., Ph.D., Lisheng Liu, M.D., Dan Dumitrascu, M.D., Vassil Stoyanovsky, M.D., Riitta L. Antikainen, M.D., Ph.D., Yuri Nikitin, M.D., Craig Anderson, M.D., Ph.D., Alli Belhani, M.D., Françoise Forette, M.D., Chakravarthi Rajkumar, M.D., Ph.D., Lutgarde Thijs, M.Sc., Winston Banya, M.Sc., and Christopher J. Bulpitt, M.D., for the HYVET Study Group* Background
Whether the treatment of patients with hypertension who are 80 years of age or From Imperial College London (N.S.B.,
R.P., R.L.A., W.B., C.J.B.) and the London older is beneficial is unclear. It has been suggested that antihypertensive therapy School of Hygiene and Tropical Medicine may reduce the risk of stroke, despite possibly increasing the risk of death.
(A.E.F.) — both in London; the University of Leuven, Leuven, Belgium (J.A.S., L.T.); the Beijing Hypertension League Insti- tute, Beijing (L.L.); Spitalul Judetean Cluj, We randomly assigned 3845 patients from Europe, China, Australasia, and Tunisia Clinica Medicală 2, Cluj, Romania (D.D.); who were 80 years of age or older and had a sustained systolic blood pressure of the National Transport Multi-Profile Hospital, Sofia, Bulgaria (V.S.); the Uni- 160 mm Hg or more to receive either the diuretic indapamide (sustained release, 1.5 versity of Oulu, Oulu, Finland (R.L.A.); mg) or matching placebo. The angiotensin-converting–enzyme inhibitor perindopril the State Scientific Research Institute of (2 or 4 mg), or matching placebo, was added if necessary to achieve the target blood Internal Medicine, Novosibirsk, Russia (Y.N.); the George Institute for Internation- pressure of 150/80 mm Hg. The primary end point was fatal or nonfatal stroke.
al Health, Sydney (C.A.); L'Etablissement Public de Santé Charles Nicolle, Service de Cardiologie, Tunis, Tunisia (A.B.); Hôpital Broca, University Paris V, Paris The active-treatment group (1933 patients) and the placebo group (1912 patients) (F.F.); and the Brighton and Sussex Medi- were well matched (mean age, 83.6 years; mean blood pressure while sitting, cal School, Brighton, United Kingdom 173.0/90.8 mm Hg); 11.8% had a history of cardiovascular disease. Median follow-up (C.R.). Address reprint requests to Dr. Beckett at Care of the Elderly, Division of was 1.8 years. At 2 years, the mean blood pressure while sitting was 15.0/6.1 mm Hg Medicine, Imperial College London, Du lower in the active-treatment group than in the placebo group. In an intention-to- Cane Rd., London W12 ONN, United treat analysis, active treatment was associated with a 30% reduction in the rate of Kingdom. fatal or nonfatal stroke (95% confidence interval [CI], −1 to 51; P = 0.06), a 39% reduc- *The committee members and investiga- tion in the rate of death from stroke (95% CI, 1 to 62; P = 0.05), a 21% reduction in tors for the Hypertension in the Very the rate of death from any cause (95% CI, 4 to 35; P = 0.02), a 23% reduction in the Elderly Trial (HYVET) are listed in the rate of death from cardiovascular causes (95% CI, −1 to 40; P = 0.06), and a 64% re-duction in the rate of heart failure (95% CI, 42 to 78; P<0.001). Fewer serious adverse This article (10.1056/NEJMoa0801369) was events were reported in the active-treatment group (358, vs. 448 in the placebo group; published at www.nejm.org on March 31, N Engl J Med 2008;358:1887-98.
Copyright 2008 Massachusetts Medical Society. The results provide evidence that antihypertensive treatment with indapamide (sus-tained release), with or without perindopril, in persons 80 years of age or older is beneficial. (ClinicalTrials.gov number, NCT00122811.) n engl j med 358;18 www.nejm.org may 1, 2008 The New England Journal of Medicine Downloaded from nejm.org at NORTHWESTERN UNIVERSITY on June 14, 2013. For personal use only. No other uses without permission. Copyright 2008 Massachusetts Medical Society. All rights reserved. Blood-pressure reduction is effec- efits and risks of antihypertensive treatment in tive in preventing stroke and other vascu- patients 80 years of age or older.
lar events, including heart failure.1 In the past decade, treatment benefits have appeared to be broadly consistent among a range of subpopu-lations and among antihypertensive agents.2 How- The detailed protocol for HYVET has been pub-ever, as highlighted by earlier3 and more recent4 lished previously.19 HYVET was a randomized, guidelines, evidence that treating patients 80 years double-blind, placebo-controlled trial performed of age or older is beneficial is inconclusive. Al- in 195 centers in 13 countries in Western and East-though the risk of stroke increases continuously ern Europe, China, Australasia, and North Africa. with increasing blood pressures above approxi- Approval for the trial was obtained from the ap-mately 115/75 mm Hg, the association of blood propriate authorities and central or local ethics pressure and stroke attenuates with increasing committees, as required. All patients gave written age.1 Epidemiologic population studies have con- informed consent, except those who were illiter-sistently suggested that blood pressure and the ate, for whom an independent witness signed the risk of death are inversely related among people consent form.
80 years of age or older,5-9 possibly reflecting in- HYVET was funded by grants from the British creased risks of therapy for blood-pressure reduc- Heart Foundation and the Institut de Recherches tion or reverse causation due to conditions that Internationales Servier. The trial was coordinated may be associated with blood-pressure reduction by staff of the Department of Care of the Elderly, (such as cancer, dementia, myocardial infarction, Imperial College London. Imperial College Lon-and heart failure).
don required all committee members and inves- A recent retrospective cohort analysis of pa- tigators to sign confidentiality agreements. All tients 80 years of age or older with hypertension, data management was performed by the authors of whom 84.5% were receiving antihypertensive who were at Imperial College London. The in-medication, reported a shorter survival for those terim analyses were performed by, and the final with systolic blood pressure levels below 140 analysis verified by, an academic author, indepen-mm Hg, even after adjustment for known predic- dently of Imperial College London. All the authors tors of death.10 Randomized controlled trials in- contributed to the writing of the manuscript and volving older adults either have excluded those the lead authors vouch for the completeness and 80 years of age or older11,12 or have recruited too accuracy of the results.
few to show an advantage of treatment.13-16 A Patients had to be 80 years of age or older meta-analysis of results regarding the treatment (confirmed by national documentation) with per-of hypertension specifically in this age group sistent hypertension (defined as a sustained sys-suggested that the benefit — a 36% reduction in tolic blood pressure of 160 mm Hg). Exclusion the risk of stroke — might be offset by possible criteria included a contraindication to use of the adverse effects, given a nearly significant in- trial medications, accelerated hypertension, sec-crease, by 14%, in the risk of death from any cause ondary hypertension, hemorrhagic stroke in the (P = 0.05).17 These positive results were not robust, previous 6 months, heart failure requiring treat-since addition of data from just one hypothetical, ment with antihypertensive medication, a serum properly designed trial that showed no treatment creatinine level greater than 150 μmol per liter effect would render the results not significant.
(1.7 mg per deciliter), a serum potassium level of The results of the pilot study for the Hyper- less than 3.5 mmol per liter or more than 5.5 tension in the Very Elderly Trial (HYVET)18 were mmol per liter, gout, a diagnosis of clinical de-consistent with those from the meta-analysis. mentia, and a requirement of nursing care.
Both results suggest that treatment for hyperten- Patients were instructed to stop all antihyper- sion was associated with a reduction in stroke but tensive treatment and to take a single placebo also a possible increase in death from any cause, tablet daily for at least 2 months and to undergo such that for each stroke prevented, there was one two blood-pressure measurements during each of death from a cause other than stroke. In the main two visits, 1 month apart, after having been seated HYVET study, we aimed to resolve persistent ar- for 5 minutes. On the third visit and thereafter, eas of clinical uncertainty about the relative ben- the standing blood pressure was taken twice, n engl j med 358;18 www.nejm.org may 1, 2008 The New England Journal of Medicine Downloaded from nejm.org at NORTHWESTERN UNIVERSITY on June 14, 2013. For personal use only. No other uses without permission. Copyright 2008 Massachusetts Medical Society. All rights reserved. Hypertension in the Very Elderly after the patient had been standing for 2 minutes. each visit, if desired. At the annual visits, infor-At the start of the trial, blood pressures were mation was collected on current diseases, medi-recorded with the use of either a mercury sphyg- cation, blood pressure, biochemical levels (sodium, momanometer or a validated automated device, potassium, urea, creatinine, glucose, uric acid), but by the end of the trial, a validated automated cholesterol levels (total and high-density lipopro-device was used in the majority of centers.
tein), and hematologic measures (hemoglobin, If the mean of the four systolic blood-pressure hematocrit), and electrocardiography and an as- measurements taken at the second and third vis- sessment of cognitive function with the use of
its (two at each visit) was between 160 and 199 the Mini–Mental State Examination were per-
mm Hg, patients underwent randomization, pro- formed. If the patient was enrolled in an optional
vided that all inclusion and exclusion criteria were add-on study, a quality-of-life questionnaire was
met. Randomization was stratified according to also completed. At the 3-month and 6-month vis-
age (80 to 89 years and 90 years or older) and its, only data on current diseases, medication, and
sex; permuted blocks of 4 and 6 of any 10 pa- blood pressure were collected.
tients were used to ensure roughly equal assign-
ment to each of the two groups within large End Points
centers. At the start of the trial in 2000, the The primary end point of the trial was any stroke
mean diastolic blood pressure while seated had (fatal or nonfatal). This end point did not include
to be 90 to 109 mm Hg, but in 2003 a protocol transient ischemic attacks. Secondary end points
amendment relaxed this criterion to be under included death from any cause, death from car-
110 mm Hg, allowing for the inclusion of pa- diovascular causes, death from cardiac causes,
tients with isolated systolic hypertension. The and death from stroke. All events that were pos-
criterion of systolic blood pressure while stand- sible end points were reviewed by an independent
ing remained the same throughout the trial, at committee, unaware of the group assignment,
140 mm Hg or more.
using predefined definitions from the protocol.19 After randomization, patients received either Events were classified as cardiovascular or non- indapamide (sustained release, 1.5 mg) or match- cardiovascular. Death from cardiac causes in-
ing placebo alone. At each visit (or at the discre- cluded fatal myocardial infarction, fatal heart
tion of the investigator), if needed to reach the failure, and sudden death. Heart failure was di-
target blood pressure, perindopril (2 mg or 4 mg) agnosed if the patient had at least one of four
or matching placebo could be added. The target symptoms (paroxysmal nocturnal dyspnea, dys-
systolic blood pressure was less than 150 mm Hg, pnea at rest, orthopnea, or symptoms consistent
and the target diastolic blood pressure was less with New York Heart Association class III heart
than 80 mm Hg. The use of additional antihyper- failure20) and at least two of seven signs (rales or
tensive agents for more than 3 months resulted crepitations, moderate ankle edema, tachycardia
in withdrawal of the patient from double-blind [120 or more beats per minute], a third heart
follow-up, with an option to enter open follow-up. sound, elevated jugular venous pressure, cardio-
Patients were also withdrawn from double-blind megaly, or radiologic signs characteristic of heart
treatment if they had received the maximum failure). If the two signs present were rales and
dose of the study drugs yet had a systolic blood ankle edema, a third sign was required.
pressure while sitting of 220 mm Hg or more or
if they had a diastolic blood pressure while sitting Data Monitoring
of 110 mm Hg or more on at least two consecu- An independent data monitoring committee met
tive visits that were 2 or more weeks apart.
twice a year throughout the trial to monitor the quality of the data and also met at prespecified intervals (after every 70 reported stroke events) to Baseline characteristics were recorded during the perform interim analyses. The interim analyses placebo run-in period. After randomization, pa- of the primary end point were monitored accord-tients were seen at least every 3 months during ing to an O'Brien–Fleming guideline; death from the first year and at least every 6 months thereaf- any cause was monitored to detect a possible in-ter. Investigators were permitted to adjust the dose creased risk in the active-treatment group. On of the trial medication more frequently than at the basis of the committee's recommendations, n engl j med 358;18 www.nejm.org may 1, 2008 The New England Journal of Medicine Downloaded from nejm.org at NORTHWESTERN UNIVERSITY on June 14, 2013. For personal use only. No other uses without permission. Copyright 2008 Massachusetts Medical Society. All rights reserved. four centers were closed after the first year of the ble-blind or open). Patients who were withdrawn trial because of concerns that these centers failed from double-blind follow-up for reasons other to provide complete and accurate data. Specifi- than withdrawal of consent were followed in an cally, one center was closed because of a failure open-label fashion, with the same schedule of to identify all the biochemical data in the source visits. Events that occurred during the open fol-documents, one because of a failure to obtain low-up period were included in the intention-to-appropriate regulatory approval, one owing to treat analysis. A per-protocol analysis was also inappropriate drug delivery, and one owing to a performed; it excluded data from any patients who failure to use validated equipment for measuring underwent randomization in error, who were blood pressure.
stratified incorrectly for either age or sex, or who did not receive the correct, assigned study drugs. For this analysis, patient-years were calculated HYVET was designed to detect a 35% reduction from the date on which the study drugs were initi-in the rate of any stroke, with a statistical power ated through the time the patient became ineli-of 90% at the 1% level of significance and assum- gible to continue with double-blind follow-up, ac-ing an event rate of 40 per 1000 patient-years. We cording to the protocol19 (i.e., when the patient calculated that 10,500 patient-years of follow-up had an event that required withdrawal from dou-would be required, with equal numbers of pa- ble-blind treatment).
tients assigned to the active-treatment group and We compared the means of continuous vari- the placebo group. The first patient underwent ables by using the z-test, proportions by using the randomization in February 2001. After the first chi-square test, and incidence rates by using the interim analysis in 2005, the data monitoring log-rank test. In the log-rank and Cox analyses of committee recommended that the trial continue fatal or nonfatal strokes and death from specific with no change to the protocol. At the second causes, death from any cause and death from a interim analysis, in July 2007, after 140 strokes cause other than the specific cause, respectively, had been reported (through April 30, 2007; a to- were regarded as events for which data were tal of 7399 patient-years of follow-up), the active- censored. Cumulative-incidence curves were es-treatment group showed evidence of a reduction timated by means of the Kaplan–Meier method. in the rate of the primary end point of any stroke P values less than 0.05 were considered to indi-(relative risk, 0.59; 95% confidence interval [CI], cate statistical significance. For patients with more 0.40 to 0.88; P = 0.009) and, unexpectedly, for than one end point during the follow-up period, death from any cause (relative risk, 0.76; 95% CI, the time to the first relevant end point was used 0.62 to 0.93; P = 0.007). The trial was terminated in each analysis. Results for death from fatal or at that point, for ethical reasons. All final visits nonfatal stroke, death from any cause, and death were completed by October 12, 2007. Data and from cardiovascular causes were adjusted for sex, events reported before the final visit were includ- age, baseline systolic blood pressure measured ed in the final analysis.
while the patient was seated, and previous cardio- The primary analysis was performed accord- vascular disease. The proportional-hazard assump- ing to the intention-to-treat principle. Data from tion was verified by plotting the Schoenfeld re-patients were analyzed for the groups to which siduals. All reported P values are two-sided and the patients were assigned, regardless of which were not adjusted for multiple testing. Data man-study drugs (or which doses) the patients actu- agement and analyses were performed with the ally received and regardless of other protocol ir- use of Stata software (version 8.1) (StatCorp) and regularities. Patients from closed centers were SAS software (version 9.1.3) (SAS Institute).
included in the intention-to-treat population and contributed person-years and events up to the date of closure of the center, after which no further information was available.
A total of 4761 patients entered the placebo run-in In the intention-to-treat analysis, patient-years phase. Of these, 3845 were randomly assigned to were calculated from the date of randomization one of the two study groups. The reasons for non-through the date of death or the last available randomization are listed in Figure 1. Patients were visit, irrespective of the type of follow-up (dou- recruited from Western Europe (86 patients), East- n engl j med 358;18 www.nejm.org may 1, 2008 The New England Journal of Medicine Downloaded from nejm.org at NORTHWESTERN UNIVERSITY on June 14, 2013. For personal use only. No other uses without permission. Copyright 2008 Massachusetts Medical Society. All rights reserved. Hypertension in the Very Elderly 4761 Patients entered the placebo run-in phase 916 Did not undergo randomization 59 Were in run-in phase when trial was 12 Were eligible but did not undergo randomization before trial was stopped 280 Withdrew consent461 Did not meet protocol criteria 65 Were at a center that closed 9 Were at a center at which quota had 3845 Underwent randomization 1933 Were assigned 1912 Were assigned to active treatment Event for which data were censored 1933 Were included 1912 Were included Event for which data were censored in the intention-to- in the intention-to- 282 Declined to participate 266 Declined to participate 4 Were withdrawn by (4149 total patient-yr) (3942 total patient-yr) 5 Were withdrawn by 27 Had a protocol withdrawal 42 Had a protocol withdrawal event and no open follow-up event and no open follow-up 164 Were at centers closed by 166 Were at centers closed by data monitoring committee data monitoring committee 168 Had other administrative 171 Had other administrative 1086 Were alive at end of trial 1016 Were alive at end of trial 6 Were lost to follow-up 11 Were lost to follow-up 12 Were withdrawn for the per-protocol analysis 11 Were withdrawn for the 4 Underwent randomization per-protocol analysis 6 Underwent randomization 6 Did not receive correct 4 Did not receive correct 1 Was randomized to wrong 1 Was randomized to wrong 1 Had group-assignment-code envelope broken at the time of randomization Event for which data were censored 1922 Were included 1900 Were included Event for which data were censored 170 Died during double-blind in the per-protocol in the per-protocol 215 Died during double-blind 327 Declined to participate 325 Declined to participate 11 Were withdrawn by (4025 total patient-yr) (3672 total patient-yr) 10 Were withdrawn by 70 Had a protocol withdrawal 137 Had a protocol withdrawal event and no open follow-up event and no open follow-up 164 Were at centers closed by 166 Were at centers closed by data monitoring committee data monitoring committee 168 Had other administrative 171 Had other administrative 1006 Were alive at end of trial 866 Were alive at end of trial 6 Were lost to follow-up 10 Were lost to follow-up Figure 1. Entry, Randomization, and Follow-up of Patients in the Hypertension in the Very Elderly Trial.
Of the 461 patients who did not meet the protocol criteria, 229 did not meet the criteria on the basis of blood pressure, 18 on the basis
of age, 61 on the basis of serum potassium level, 20 on the basis of serum creatinine level, 26 because they were receiving other antihy-
pertensive treatment, and 107 for other reasons. The other administrative reasons for the censoring of data were death or retirement of a local investigator and change in national legislation as to where patients in clinical trials 3rd could be seen.
med 358;18 www.nejm.org may 1, 2008 AUTHOR, PLEASE NOTE:
The New England Journal of Medicine Figure has been redrawn and type has been reset.
Downloaded from nejm.org at NORTHWESTERN UNIVERSITY on June 14, 2013. For personal use only. No other uses without permission. Please check carefully.
Copyright 2008 Massachusetts Medical Society. All rights reserved. ern Europe (2144), China (1526), Australasia (19), group and 4159 in the active-treatment group. At and Tunisia (70). The two study groups were well the end of the trial, vital status was unknown in balanced at baseline (Table 1). The age range at 17 patients, 1882 patients were still undergoing entry was 80 to 105 years, with 73.0% of patients double-blind follow-up, and 220 patients were un-80 to 84 years of age, 22.4% of patients 85 to 89 dergoing open follow-up.
years of age, and 4.6% of patients 90 years of age According to the intention-to-treat analysis, at or older (interquartile range, 81.2 to 85.3). More 2 years, as compared with baseline, the systolic than 90% of patients were known to be hyper- and diastolic blood pressure values obtained while tensive, of whom approximately one third had the patient was seated had fallen by a mean (±SD) not been previously treated. A history of cardio- of 14.5±18.5 mm Hg and 6.8±10.5 mm Hg, re-vascular disease was reported in 11.8%, and dia- spectively, in the placebo group and by 29.5±15.4 betes was reported in 6.9%.
mm Hg and 12.9±9.5 mm Hg, respectively, in the The median duration of follow up was 1.8 years active-treatment group. At 2 years, the mean sys- (mean, 2.1; range, 0 to 6.5). The number of pa- tolic and diastolic blood pressure values obtained tient-years of follow-up was 3964 in the placebo while the patient was standing had decreased by Table 1. Baseline Characteristics of the Patients.*
Active Treatment (N = 1933)
Placebo (N = 1912)
Female sex — no. (%) Blood pressure — mm Hg Orthostatic hypotension — no. (%)† Isolated systolic hypertension — no. (%) Heart rate — beats/min Cardiovascular history
Cardiovascular disease — no. (%)
Hypertension — no. (%) Antihypertensive treatment — no. (%) Stroke — no. (%) Myocardial infarction — no. (%) Heart failure — no. (%) Cardiovascular risk factors
Current smoker — no. (%)
Diabetes — no. (%)‡ Total cholesterol — mmol/liter High-density lipoprotein cholesterol — mmol/liter Serum creatinine — μmol/liter Uric acid — μmol/liter Body-mass index§ * Plus–minus values are means ±SD. To convert values for cholesterol to milligrams per deciliter, divide by 0.02586; to con- vert values for uric acid to milligrams per deciliter, divide by 59.48; and to convert values for serum creatinine to milli- grams per deciliter, divide by 88.4.
† Orthostatic hypotension is defined as a drop in systolic blood pressure of more than 20 mm Hg or a reduction in dia- stolic blood pressure of more than 10 mm Hg while standing.
‡ Diabetes is defined as reported diabetes, the receipt of antidiabetes treatment, or a random blood glucose measurement of more than 11.1 mmol per liter (200 mg per deciliter).
§ The body-mass index is the weight in kilograms divided by the square of the height in meters.
n engl j med 358;18 www.nejm.org may 1, 2008 The New England Journal of Medicine Downloaded from nejm.org at NORTHWESTERN UNIVERSITY on June 14, 2013. For personal use only. No other uses without permission. Copyright 2008 Massachusetts Medical Society. All rights reserved. Hypertension in the Very Elderly 13.6±18.9 mm Hg and 7.0±10.9 mm Hg, respec- 0 to 21) being prevented because 1000 patients tively, in the placebo group and by 28.3±16.5 were treated for 2 years or 1 stroke being prevented mm Hg and 12.4±10.3 mm Hg, respectively, in because 94 patients were treated for 2 years. There the active-treatment group. There was a differ- were a total of 431 deaths during the trial, with ence of 15.0/6.1 mm Hg in blood pressure, mea- an overall rate of death of 53.1 per 1000 patient-sured while patients were seated, between the years. There was a 21% reduction (95% CI, 4 to 35; two groups at 2 years (Fig. 2). Also at 2 years, P = 0.02) in the rate of death from any cause in the target blood pressure was reached in 19.9% the active group. The rate of fatal stroke was re-of patients in the placebo group and in 48.0% in duced by 39% (95% CI, 1 to 62; P = 0.05). The rate the active-treatment group (P<0.001).
of death from cardiac causes was not significantly At 2 years, 25.8%, 23.9%, and 49.5% of patients reduced in the active-treatment group. The rate in the active-treatment group were receiving inda- of death from cardiovascular causes was reduced pamide alone, indapamide and perindopril (2 mg), by 23% (95% CI, −1 to 40; P = 0.06). The rate of and indapamide and perindopril (4 mg), respec- fatal or nonfatal heart failure was reduced by 64% tively; 14.2%, 13.4%, and 71.8% of patients in the (95% CI, 42 to 78; P<0.001), and the rate of any placebo group, respectively, were receiving the cardiovascular event (death from cardiovascular corresponding placebos. The final main intention- causes or stroke, myocardial infarction, or heart to-treat analysis included an extra 724 patient- failure) was reduced by 34% (95% CI, 18 to 47; years of follow-up, with an additional 13 strokes P<0.001).
and 54 deaths. This final analysis confirmed the The benefits of treatment began to be appar- treatment benefit for the risk of death from any ent within the first year (Fig. 3). When adjusted cause and yielded a nearly significant benefit for according to sex, age, baseline systolic blood the risk of fatal or nonfatal stroke.
pressure while seated, and previous cardiovascu- With regard to the primary end point (fatal or lar disease, the results did not materially change nonfatal stroke), 51 events occurred in the active- for the end points of fatal or nonfatal stroke, treatment group as compared with 69 events in death from any cause, or death from cardiovas-the placebo group, a reduction in the rate of stroke cular causes.
of 30% (95% CI, −1 to 51; P = 0.06) (Table 2 and According to per-protocol analyses, in the ac- Fig. 3). This is equivalent to 11 strokes (95% CI, tive-treatment group as compared with the pla- Active-treatment group Systolic blood pressure Blood Pressure (mm Hg)
Diastolic blood pressure No. at Risk
Placebo group
Active-treatment group Figure 2. Mean Blood Pressure, Measured while Patients Were Seated, in the Intention-to-Treat Population,
According to Study Group.
n engl j med 358;18 www.nejm.org may 1, 2008 The New England Journal of Medicine Downloaded from nejm.org at NORTHWESTERN UNIVERSITY on June 14, 2013. For personal use only. No other uses without permission. AUTHOR, PLEASE NOTE:
Figure has been redrawn and type has been reset.
Copyright 2008 Massachusetts Medical Society. All rights reserved. Please check carefully.
Table 2. Main Fatal and Nonfatal End Points in the Intention-to-Treat Population.
Rate per 1000 Patient-Yr
Unadjusted Hazard Ratio
End Point
(No. of Events)
Fatal or nonfatal 0.70 (0.49–1.01) Death from stroke 0.61 (0.38–0.99) 0.79 (0.65–0.95) From noncardiovascular or un- 0.81 (0.62–1.06) From cardiovascular cause 0.77 (0.60–1.01) From cardiac cause* 0.71 (0.42–1.19) From heart failure 0.48 (0.18–1.28) Fatal or nonfatal Any myocardial infarction 0.72 (0.30–1.70) Any heart failure 0.36 (0.22–0.58) Any cardiovascular event† 0.66 (0.53–0.82) * Death from cardiac causes was defined as fatal myocardial infarction, fatal heart failure, and sudden death.
† Any cardiovascular event was defined as death from cardiovascular causes or stroke, myocardial infarction, or heart failure.
cebo group, the rate of stroke was reduced by 34% (95% CI, 5 to 54; P = 0.03), the rate of any heart failure by 72% (95% CI, 52 to 83; P<0.001), and the The results of HYVET indicate that antihyperten-rate of death from cardiovascular causes by 27% sive treatment based on indapamide (sustained (95% CI, 3 to 45; P = 0.03). In addition, the rate of release, 1.5 mg), with or without 2 to 4 mg of per-death from any cause was decreased by 28% (95% indopril, significantly reduces the risks of death CI, 12 to 41; P = 0.001) and the rate of death from from stroke and death from any cause in very stroke by 45% (95% CI, 7 to 67; P = 0.02).
elderly patients. This finding for stroke is consis- Among the patients followed for at least 2 years, tent with that seen in the HYVET pilot study18 there were no significant differences between the and the Individual Data Analysis of Antihyper-two groups with regard to changes from baseline tensive Drug Intervention Trials (INDANA) group in the serum potassium level (−0.02 mmol and meta-analysis.17 The reduction in death from any 0.03 mmol per liter in the active-treatment group cause is a new and unexpected result.
and in the placebo group, respectively; P = 0.09), As compared with clinical trials involving uric acid (11.6 μmol and 3.5 μmol per liter [0.2 other age groups,11,14-16 the ratio of fatal to non-and 0.1 mg per deciliter], P = 0.07), glucose (0.16 fatal events was higher in HYVET. Although the mmol and 0.11 mmol per liter [2.9 and 2.0 mg number of strokes reported in HYVET (120) was per deciliter], P = 0.56), or creatinine (3.4 μmol and similar to that reported in the Systolic Hyperten-2.3 μmol per liter [0.04 and 0.03 mg per deciliter], sion in Europe (Syst-Eur)16 trial (128 strokes) and P = 0.30). The number of serious adverse events higher than the number reported in the Swedish reported was 448 in the placebo group and 358 in Trial in Older Patients with Hypertension (STOP–the active-treatment group (P = 0.001). Only five of Hypertension, 82 strokes),14 in HYVET the pro-these events (three in the placebo group and two portion of fatal strokes (57.5%) was higher than in the active-treatment group) were classified by that in Syst-Eur (28.9%) or STOP (18.3%). This the local investigator as possibly having been due difference probably reflects the older age of the to the trial medication.
patients in HYVET and the higher rates of stroke n engl j med 358;18 www.nejm.org may 1, 2008 The New England Journal of Medicine Downloaded from nejm.org at NORTHWESTERN UNIVERSITY on June 14, 2013. For personal use only. No other uses without permission. Copyright 2008 Massachusetts Medical Society. All rights reserved. Hypertension in the Very Elderly A Fatal or Nonfatal Stroke
B Death from Any Cause
No. of Events per 100 Patients
No. of Events per 100 Patients
No. at Risk
No. at Risk
Active-treatment group 1933 Active-treatment group 1933 C Death from Cardiovascular Causes
D Death from Stroke
No. of Events per 100 Patients
No. of Events per 100 Patients
No. at Risk
No. at Risk
Active-treatment group 1933 Active-treatment group 1933 E Heart Failure
No. of Events per 100 Patients
No. at Risk
Placebo group
Active-treatment group 1933 Figure 3. Kaplan–Meier Estimates of the Rate of End Points, According to Study Group.
For the active-treatment group as compare ICM d with the placebo group, the unadjusted hazard ratios (95% CIs) were as follows: for fatal or nonfatal stroke, 0.70 (0.49 to 1.01) (Panel A); REG death from any cause, 0.79 (0.65 to 0.95) (Panel B); for death from cardiovascular causes, 0.77 (0.60 to 1.01) (Panel C); for death from sCASE troke, 0.61 (0.38 to 0.99) (Panel D); and fRevised or heart failure, 0.36 (0.22 to 0.58) (Panel E).
n engl j med 358;18 www.nejm.or Combo AUTHOR, PLEASE NOTE:
Figure has been redrawn and type has been reset.
The New England Journal of Medicine Please check carefully.
Downloaded from nejm.org at NORTHWESTERN UNIVERSITY on June 14, 2013. For personal use only. No other uses without permission. Copyright 2008 Massachusetts Medical Society. All rights reserved. that occur with increasing age.1 Increasing age stroke and death from any cause and the low is a major predictor of death from stroke,21-25 prevalence of previous cardiovascular disease at with rates as high as 52%21 in persons 80 years baseline. It is possible that the numbers of pa-of age or older, consistent with our results. A linear tients needed to be treated is an overestimate, association between blood pressure and stroke re- given the low rate of stroke, though a number duction has been clearly demonstrated, although needed to treat of 94 is acceptable in the context it is attenuated by increasing age.25 The reduction of prevention. A number needed to treat of 40 to in the risk of death from stroke with active treat- prevent one death during a 2-year period is more ment was significant in both the intention-to-treat impressive, supporting the effectiveness of the and the per-protocol analyses, as was the reduc- intervention used. However, it would be prema-tion in the risk of any stroke in the per-protocol ture to extrapolate the results from HYVET to pa-analysis. This reduction is consistent with previ- tients in this age group who are more frail.
ous results in younger hypertensive patients and The large reduction in the risk of heart failure in reviews of data from the very elderly.
in HYVET appears to be important. Heart failure An unexpected finding of our trial is the re- is common in people older than 70 years of age, duction in the risk of death from any cause with and hypertension is a major risk factor for heart active treatment, making HYVET one of the few failure.30 The combination of a diuretic and an individual studies of hypertension showing ben- ACE inhibitor is likely to confer a benefit. In the efits of blood-pressure reduction on mortality.14,26 Antihypertensive and Lipid-Lowering Treatment to The results for death from any cause show a sig- Prevent Heart Attack Trial (ALLHAT; ClinicalTrials.
nificant benefit of treatment, at odds with the gov number, NCT00000542), the lowest rates of results of the INDANA meta-analysis, the largest hospitalization and fatal heart failure were found analysis to date to focus on very elderly patients for the groups receiving a diuretic or ACE inhibi-with hypertension, which included 1670 patients tor, with a greater (albeit not significantly greater) 80 years of age or older recruited for interven- reduction in the diuretic group.31tion trials.17 The INDANA meta-analysis includ- A challenge and possible limitation in per- ed trials with treatment regimens based on high forming our study of the very elderly was the col-doses of diuretics other than indapamide or on lection of adequate information to validate the beta-blockers. Diuretics have not been shown to end-point data, especially for patients who died increase mortality among younger patients, but at home without receiving any direct medical the very elderly may be more prone to metabolic intervention. The end-point committee required disturbances from medications, particularly hy- evidence to support a diagnosis of a stroke, and pokalemia, which can predispose patients to ar- this evidence was not always available. Rapid and rhythmias and possibly sudden death. Beta-block- unexpected deaths were designated as deaths ers appear to be less effective than newer agents from cardiovascular causes by the committee. In in the treatment of hypertension.27 Given the the very elderly, it is difficult to establish the known changes in serum potassium level that exact cause of death, since patients are often not can occur with a thiazide-like antihypertensive monitored during the last hours of life and au-agent or angiotensin-converting–enzyme (ACE) in- topsies are rarely performed. Given these diffi-hibitor on its own, in combination, they are likely culties, the reduction of the risk of death from to have a neutral effect. This was evidenced by any cause in the active-treatment group clearly the similar serum potassium levels in the two indicates the overall advantage of active treat-groups, with 73.4% of patients in the active-treat- ment in HYVET.
ment group receiving both indapamide and per- In HYVET, we evaluated the benefit of treat- indopril at 2 years. Indapamide (sustained release) ing patients who had a sustained systolic blood has also been shown to have a neutral effect on pressure of 160 mm Hg or higher. The fall in blood glucose and lipids,28 and the combination blood pressure in the active-treatment group was of indapamide and perindopril has also been consistent with results for indapamide-based shown to confer a benefit with regard to stroke.29 strategies in other studies.32,33 The results sup- As in most trials, the patients in HYVET were port a target blood pressure of 150/80 mm Hg in generally healthier than those in the general popu- patients receiving treatment, since that target lation, as evidenced by the low overall rates of was reached in nearly 50% of such patients in n engl j med 358;18 www.nejm.org may 1, 2008 The New England Journal of Medicine Downloaded from nejm.org at NORTHWESTERN UNIVERSITY on June 14, 2013. For personal use only. No other uses without permission. Copyright 2008 Massachusetts Medical Society. All rights reserved. Hypertension in the Very Elderly HYVET after 2 years. Whether further reduction Supported by grants from the British Heart Foundation and is beneficial still needs to be established.
the Institut de Recherches Internationales Servier.
Drs. Beckett and Peters and Mr. Banya report receiving grant Elevated blood pressure is common in persons support from the Institut de Recherches Internationales Servier; 80 years of age or older,34 a group constituting Dr. Staessen, consulting fees from Pfizer, Tanabe, Daiichi-San-the fastest-growing segment of the general popu- kyo, and Sigma-Tau and speakers' fees from Pfizer, Tanabe, and Bayer; Dr. Anderson, consulting fees from Boehringer Ingelheim lation.35 HYVET provides unique evidence that and Servier and speakers' fees from Boehringer Ingelheim, Ser-hypertension treatment based on indapamide (sus- vier, AstraZeneca, and Sanofi-Aventis; Dr. Forette, consulting fees tained release), with or without perindopril, in from Wyeth Elan, Sanofi-Aventis and Bristol-Myers Squibb and speakers' fees from Servier, AstraZeneca, and Sanofi-Aventis; Dr. the very elderly, aimed to achieve a target blood Rajkumar, speakers' fees from Schering-Plough, Merck Sharp & pressure of 150/80 mm Hg, is beneficial and is Dohme, and Menarini; and Dr. Bulpitt, consulting fees from Im-associated with reduced risks of death from perial College Consulting, a consultancy funded by a grant from the Institut de Recherches Internationales Servier. No other po- stroke, death from any cause, and heart failure. tential conflict of interest relevant to this article was reported.
The committee members and investigators for HYVET were as follows: Coordinating Center: C.J. Bulpitt (lead investigator), A.E.
Fletcher (coinvestigator), N.S. Beckett (trial coordinator), R. Peters (deputy trial coordinator), HYVET coordinating team at Imperial
College London (1999 to 2008); Steering Committee: T. McCormack, J. Potter, B.G. Extremera, P. Sever, F. Forette, D. Dumitrascu, C.
Swift, J. Tuomilehto, J. Coope (retired in 2001), C. Nachev (deceased); Data Monitoring Committee: J. Staessen, L. Thijs, R. Clarke, K.
Narkiewicz; End Points Committee: C. Davidson (retired in 2003), J. Duggan, G. Leonetti, N. Gainsborough, M.C. De Vernejoul, J.
Wang, V. Stoyanovsky; Dementia Validation Committee: J. Tuomilehto, R. Clarke, A. Waldman, I. Walton, C. Ritchie; Ethics Committee:
R. Fagard, J. Grimley Evans, B. Williams; Investigators: Australia — R. Warne and I. Puddey (national coordinators), M. Woodward, R.
Penhall, C. Inderjeeth, S. Roger, R. Scholes, C. Johnson; Belgium — H. Celis (national coordinator), G. Adriaens, W. Onsea, K. Cor-
nelli, D. Vantroyen, P. Cleen, P. de Voogt; Bulgaria — C. Nachev (deceased) (national coordinator from 1998 to 2005), V. Stoyanovsky
(national coordinator after 2005), P. Solakov, R. Prokopova, E. Mantova, D. Smilkova, S. Mantov, K. Yankulova, R. Kermova, D. Popov,
V. Sirakova, V. Gergova, D. Kamenova, F. Grigorov, T. Vassileva, R. Alahverdian, M. Tzekova; China — L. Liu (national coordinator), H.
Ge, S. Wang, J. Wang, W. Zhang, S. Jin, L. Ge, Y.F. Lu, S. Ma, L. Shen, J. Guo, Z. Lv (deceased), R. Huang, X. Li, B. Guo, T. Zhang, L.
Zhang, J. Feng, Z. He, J. Wang, L. Deng, L. Liu, Q. Yuan, F. Zhang, H. Li, D. Wang, K. Yang, M. Sun, H. Liu, X. Yan, F. Ren, J. Tang;
Finland — R. Antikainen (national coordinator), T. Strandberg, T. Konttila, A. Hynninen, M. Jääskivi, J. Airas, T. Jääskeläinen, J. Tu-
omilehto, H. Litmanen; France — F. Forette (national coordinator), J. Doucet, J. Belmin, A. Benetos, G. Berrut, T. Boge, M. Bonnefoy,
A. Carre, N. Charasz, J. Covillard, T. Dantoine, M. Escande, Y. Frances, R. Joire, C. Jeandel, S. Legrain, A. Lion, M. Maillet-Vioud, J.P.
Escaillas, S. Meaume, P. Pfitzenmeyer, F. Puisieux, E. Quercy, O. Rodat, J. Soubeyrand, B. de Wazieres, H. Hindennach, L. Lugassy, J.
Rossi, M. Martel, J.-M. Paladel, C. Ravier, A. Visconti, J.P. Gallet, D. Zygouritsas, D. Charles, F. Flamand, G. Grandmottet, M. Grand-
mottetegermann, C. Gevrey, P.L. Mesnier, G. Robert, C. Besset-Prat, A. Brousse, P. Lafont, J. Morelli, P. Vernede, A. Volkmann, X.
Bodin, B. Destrube, R. Eoche, A. Boye, F. Seropian, P. Gernigon, D. Meker, J. Thomere, Y. Thual, F. Volny, E. Grassart, M. Herent, D.
Lejay, J.-P. Lopez, B. Mannessier, G. Pruvost, J.-C. Urbina; Ireland — J. Duggan (national coordinator); New Zealand — C. Anderson
(national coordinator), S. Lillis, J. Gommans; Poland — T. Grodzicki (national coordinator), Z. Chodorowski, Z. Gaciong; Romania — D.
Dumitrascu (national coordinator), M. Comsa, V. Sandru, G. Prada, M. Dunca-Moisin, D. Jianu, D. Jinga-Lazar, V. Enachescu, C. Zaha-
ria; Russia — Y. Nikitin (national coordinator), A. Kirichenko, L. Olbinskaya, A. Martynov, V. Zadionchenko, V. Moiseev, G. Storohza-
kov, S. Nedogoda, R.S. Karpov, O. Barbarash, G. Efremushkin, V. Kostenko, M. Boyarkin, S. Churina, T. Tyurina, M. Ballyuzek, L.
Ermoshkina, A. Timofeev, S. Yakusheva, N. Shilkina, V. Barbarich; Tunisia — A. Belhani (national coordinator), E. Boughzela, S.
Soraya, B. Youssef-Zouari, A.B. Khalfallah, M.H. Houman, A.K. Abida; United Kingdom — C. Rajkumar (national coordinator), M.
Wilkins, N.D. Pandita-Gunawardena, J. Potter, E. Ekpo, M. Price, N. de Kare-Silver, A. Starczewski, S. Chandran, N. Nasar, M. Datta-
Chaudhuri, T. McCormack, N. Majmudar, A. Gordon, L. Brawn, T. Solanki.
References
1. Lawes CM, Bennett DA, Feigin VL, pertension of the European Society of Hy-
8. van Bemmel T, Woittiez K, Blauw GJ,
Rodgers A. Blood pressure and stroke: an pertension (ESH) and of the European et al. Prospective study of the effect of overview of published reviews. Stroke Society of Cardiology (ESC). J Hypertens blood pressure on renal function in old 2004;35:1024.
2007;25:1105-87. [Erratum, J Hypertens age: the Leiden 85-Plus Study. J Am Soc 2. Neal B, MacMahon S, Chapman N. Ef-
fects of ACE inhibitors, calcium antago- 5. Mattila K, Haavisto M, Rajala S, 9. Satish S, Freeman DH Jr, Ray L, Good-
nists, and other blood-pressure-lowering Heikinheimo R. Blood pressure and five win JS. The relationship between blood drugs: results of prospectively designed year survival in the very old. Br Med J (Clin pressure and mortality in the oldest old. overviews of randomised trials. Lancet Res Ed) 1988;296:887-9.
J Am Geriatr Soc 2001;49:367-74.
6. Langer RD, Criqui MH, Barrett-Con-
10. Oates DJ, Berlowitz DR, Glickman
3. World Health Organization–Interna-
nor EL, Klauber MR, Ganiats TG. Blood ME, Silliman RA, Borzecki AM. Blood tional Society of Hypertension guidelines pressure change and survival after age 75. pressure and survival in the oldest old. for the management of hypertension. Hypertension 1993;22:551-9.
J Am Geriatr Soc 2007;55:383-8.
J Hypertens 1999;17:151-83.
7. Rastas S, Pirttilä T, Viramo P, et al. 11. Coope J, Warrender TS. Randomised
4. Mancia G, De Backer G, Dominiczak Association between blood pressure and trial of treatment of hypertension in el-
A, et al. 2007 Guidelines for the manage-
survival over 9 years in a general popula- derly patients in primary care. Br Med J ment of arterial hypertension: The Task tion aged 85 and older. J Am Geriatr Soc (Clin Res Ed) 1986;293:1145-51.
Force for the Management of Arterial Hy- 12. Medical Research Council trial of
n engl j med 358;18 www.nejm.org may 1, 2008 The New England Journal of Medicine Downloaded from nejm.org at NORTHWESTERN UNIVERSITY on June 14, 2013. For personal use only. No other uses without permission. Copyright 2008 Massachusetts Medical Society. All rights reserved. Hypertension in the Very Elderly treatment of hypertension in older adults: 21. Hollander M, Koudstaal PJ, Bots ML, 28. Ambrosioni E, Safar M, Degaute JP, et
principal results. BMJ 1992;304:405-12.
Grobbee DE, Hofman A, Breteler MM. In- al. Low-dose antihypertensive therapy with 13. Amery A, Birkenhäger W, Brixko P, et cidence, risk, and case fatality of first ever 1.5 mg sustained-release indapamide: re-
al. Mortality and morbidity results from stroke in the elderly population: the Rot-
sults of randomised double-blind con- the European Working Party on High terdam Study. J Neurol Neurosurg Psy- trolled studies. J Hypertens 1998;16:1677- Blood Pressure in the Elderly trial. Lancet chiatry 2003;74:317-21.
22. de Jong G, van RL, Kessels F, Lodder 29. PROGRESS Collaborative Group.
14. Dahlöf B, Lindholm LH, Hansson L, J. Stroke subtype and mortality: a follow-
Randomised trial of a perindopril-based Scherstén B, Ekbom T, Wester PO. Mor- up study in 998 patients with a first cere- blood-pressure-lowering regimen among bidity and mortality in the Swedish Trial bral infarct. J Clin Epidemiol 2003;56:262- 6,105 individuals with previous stroke or in Old Patients with Hypertension (STOP- transient ischaemic attack. Lancet 2001; Hypertension). Lancet 1991;338:1281-5.
23. Kammersgaard LP, Jørgensen HS, 358:1033-41. [Errata, Lancet 2001;358:
15. Prevention of stroke by antihyperten-
Reith J, Nakayama H, Pedersen PM, Olsen 1556, 2002;359:2120.] sive drug treatment in older persons with TS. Short- and long-term prognosis for 30. Stewart S, MacIntyre K, Capewell S,
isolated systolic hypertension: final re-
very old stroke patients: the Copenhagen McMurray JJ. Heart failure and the aging sults of the Systolic Hypertension in the Stroke Study. Age Ageing 2004;33:149-54.
population: an increasing burden in the Elderly Program (SHEP). JAMA 1991;265: 24. Zuliani G, Cherubini A, Ranzini M, 21st century? Heart 2003;89:49-53.
3255-64.
Ruggiero C, Atti AR, Fellin R. Risk factors 31. Davis BR, Piller LB, Cutler JA, et al.
16. Staessen JA, Fagard R, Thijs L, et al. for short-term mortality in older subjects Role of diuretics in the prevention of
Randomised double-blind comparison of with acute ischemic stroke. Gerontology heart failure: the Antihypertensive and
placebo and active treatment for older pa-
Lipid-Lowering Treatment to Prevent tients with isolated systolic hypertension: 25. Lewington S, Clarke R, Qizilbash N, Heart Attack Trial. Circulation 2006;113:
the Systolic Hypertension in Europe (Syst-
Peto R, Collins R. Age-specific relevance 2201-10.
Eur) Trial Investigators. Lancet 1997;350: of usual blood pressure to vascular mor- 32. Puig JG, Marre M, Kokot F, et al. Ef-
tality: a meta-analysis of individual data ficacy of indapamide SR compared with 17. Gueyffier F, Bulpitt C, Boissel JP, et al. for one million adults in 61 prospective enalapril in elderly hypertensive patients
Antihypertensive drugs in very old people: studies. Lancet 2002;360:1903-13.
with type 2 diabetes. Am J Hypertens a subgroup meta-analysis of randomised 26. Five-year findings of the Hyperten-
controlled trials. Lancet 1999;353:793-6.
sion Detection and Follow-up Program. I. 33. Emeriau JP, Knauf H, Pujadas JO, et al.
18. Bulpitt CJ, Beckett NS, Cooke J, et al. Reduction in mortality of persons with A comparison of indapamide SR 1.5 mg
Results of the pilot study for the Hyper-
high blood pressure, including mild hy- with both amlodipine 5 mg and hydro- tension in the Very Elderly Trial. J Hyper- pertension. JAMA 1997;277:157-66.
chlorothiazide 25 mg in elderly hyperten- 27. Dahlöf B, Sever PS, Poulter NR, et al. sive patients: a randomized double-blind
19. Bulpitt C, Fletcher A, Beckett N, et al. Prevention of cardiovascular events with controlled study. J Hypertens 2001;19:
Hypertension in the Very Elderly Trial an antihypertensive regimen of amlodi-
(HYVET): protocol for the main trial. pine adding perindopril as required ver- 34. Callow AD. Cardiovascular disease
Drugs Aging 2001;18:151-64.
sus atenolol adding bendroflumethiazide 2005 — the global picture. Vascul Phar- 20. The Criteria Committee of the New as required, in the Anglo-Scandinavian macol 2006;45:302-7.
York Heart Association. Nomenclature Cardiac Outcomes Trial-Blood Pressure 35. U.S. Census Bureau. International Data
and criteria for diagnosis of diseases of Lowering Arm (ASCOT-BPLA): a multi-
Base (IDB). (Accessed April 7, 2008, at the heart and great vessels. 9th ed. Bos- centre randomised controlled trial. Lan- ton: Little, Brown, 1994.
Copyright 2008 Massachusetts Medical Society. receive immediate notification when
a journal article is released early
To be notified when an article is released early on the Web and to receive the table of contents of the Journal by e-mail every Wednesday evening, sign up through our Web site at n engl j med 358;18 www.nejm.org may 1, 2008 The New England Journal of Medicine Downloaded from nejm.org at NORTHWESTERN UNIVERSITY on June 14, 2013. For personal use only. No other uses without permission. Copyright 2008 Massachusetts Medical Society. All rights reserved.

Source: https://canvas.northwestern.edu/files/1783295/download?download_frd=1&verifier=bUtAdJqerNHtYboB1hwKekQGBEZdcWFxpOAon7LZ