MedicineLakex

 

A randomized trial of intensive versus standard blood-pressure control

The new england journal of medicine established in 1812 November 26, 2015 A Randomized Trial of Intensive versus Standard Blood-Pressure Control The SPRINT Research Group* BACKGROUND
The most appropriate targets for systolic blood pressure to reduce cardiovascular The members of the writing committee
(Jackson T. Wright, Jr., M.D., Ph.D., Jeff morbidity and mortality among persons without diabetes remain uncertain.
D. Williamson, M.D., M.H.S., Paul K. Whelton, M.D., Joni K. Snyder, R.N., B.S.N., M.A., Kaycee M. Sink, M.D., We randomly assigned 9361 persons with a systolic blood pressure of 130 mm Hg M.A.S., Michael V. Rocco, M.D., M.S.C.E., David M. Reboussin, Ph.D., Mahboob or higher and an increased cardiovascular risk, but without diabetes, to a systolic Rahman, M.D., Suzanne Oparil, M.D., blood-pressure target of less than 120 mm Hg (intensive treatment) or a target of Cora E. Lewis, M.D., M.S.P.H., Paul L. less than 140 mm Hg (standard treatment). The primary composite outcome was Kimmel, M.D., Karen C. Johnson, M.D., M.P.H., David C. Goff, Jr., M.D., Ph.D., myocardial infarction, other acute coronary syndromes, stroke, heart failure, or Lawrence J. Fine, M.D., Dr.P.H., Jeffrey A. death from cardiovascular causes.
Cutler, M.D., M.P.H., William C. Cush­man, M.D., Alfred K. Cheung, M.D., and Walter T. Ambrosius, Ph.D.) assume re­ At 1 year, the mean systolic blood pressure was 121.4 mm Hg in the intensive- sponsibility for the overall content and integrity of the article. The affiliations of treatment group and 136.2 mm Hg in the standard-treatment group. The interven- the members of the writing group are tion was stopped early after a median follow-up of 3.26 years owing to a signifi- listed in the Appendix. Address reprint cantly lower rate of the primary composite outcome in the intensive-treatment requests to Dr. Wright at the Division of Nephrology and Hypertension, Univer­ group than in the standard-treatment group (1.65% per year vs. 2.19% per year; sity Hospitals Case Medical Center, Case hazard ratio with intensive treatment, 0.75; 95% confidence interval [CI], 0.64 to Western Reserve University, 1100 Euclid 0.89; P<0.001). All-cause mortality was also significantly lower in the intensive- Ave. Cleveland, OH 44106­6053, or at jackson . wright@ case . edu.
treatment group (hazard ratio, 0.73; 95% CI, 0.60 to 0.90; P = 0.003). Rates of seri- ous adverse events of hypotension, syncope, electrolyte abnormalities, and acute * A complete list of the members of the Systolic Blood Pressure Intervention kidney injury or failure, but not of injurious falls, were higher in the intensive- Trial (SPRINT) Research Group is pro­ treatment group than in the standard-treatment group.
vided in the Supplementary Appendix, available at NEJM.org.
This article was published on November 9, Among patients at high risk for cardiovascular events but without diabetes, target- 2015, at NEJM.org.
ing a systolic blood pressure of less than 120 mm Hg, as compared with less than N Engl J Med 2015;373:2103-16.
140 mm Hg, resulted in lower rates of fatal and nonfatal major cardiovascular DOI: 10.1056/NEJMoa1511939
events and death from any cause, although significantly higher rates of some adverse Copyright 2015 Massachusetts Medical Society. events were observed in the intensive-treatment group. (Funded by the National Institutes of Health; ClinicalTrials.gov number, NCT01206062.) n engl j med 373;22 nejm.org November 26, 2015 The New England Journal of Medicine Downloaded from nejm.org on April 7, 2016. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Hypertension is highly prevalent among patients without diabetes.24 The current in the adult population in the United article describes the primary results of the Sys- States, especially among persons older tolic Blood Pressure Intervention Trial (SPRINT), than 60 years of age, and affects approximately which compared the benefit of treatment of 1 billion adults worldwide.1,2 Among persons 50 systolic blood pressure to a target of less than years of age or older, isolated systolic hyperten- 120 mm Hg with treatment to a target of less A Quick Take sion is the most common form of hypertension,3,4 than 140 mm Hg.
is available at
NEJM.org and systolic blood pressure becomes more im-
portant than diastolic blood pressure as an inde- pendent risk predictor for coronary events, stroke, heart failure, and end-stage renal disease (ESRD).5-13 Study Design and Oversight
The Global Burden of Disease Study identified SPRINT was a randomized, controlled, open-la- elevated blood pressure as the leading risk fac- bel trial that was conducted at 102 clinical sites tor, among 67 studied, for death and disability- (organized into 5 clinical center networks) in the adjusted life-years lost during 2010.14 United States, including Puerto Rico (see the Clinical trials have shown that treatment of Supplementary Appendix, available with the full hypertension reduces the risk of cardiovascular text of this article at NEJM.org). A trial coordi- disease outcomes, including incident stroke (by nating center served as a data and biostatistical 35 to 40%), myocardial infarction (by 15 to 25%), core center and supervised the central laboratory, and heart failure (by up to 64%).5,15,16 However, the electrocardiography reading center, the mag- the target for systolic blood-pressure lowering is netic resonance imaging reading center, and the uncertain. Observational studies have shown a drug-distribution center. The rationale and pro- progressive increase in cardiovascular risk as tocol for the trial are publicly available,25,26 and systolic blood pressure rises above 115 mm Hg,10 the protocol is available at NEJM.org.
but the available evidence from randomized, SPRINT was sponsored by the NHLBI, with controlled trials in the general population of cosponsorship by the National Institute of Dia- patients with hypertension only documents the betes and Digestive and Kidney Diseases, the benefit of treatment to achieve a systolic blood- National Institute of Neurological Disorders and pressure target of less than 150 mm Hg, with Stroke, and the National Institute on Aging. An limited data concerning lower blood-pressure independent data and safety monitoring board targets.11,17-21 In a trial involving patients with monitored unblinded trial results and safety type 2 diabetes mellitus, the rate of major cardio- events. The study was approved by the institu- vascular events was similar with a systolic blood- tional review board at each participating study pressure target of less than 120 mm Hg and the site. The steering committee designed the study, commonly recommended target of less than gathered the data (in collaboration with investi- 140 mm Hg, though the rate of stroke was gators at the clinics and other study units), made lower with the target of less than 120 mm Hg.22 the decision to submit the manuscript for publi- A recent trial involving patients who had had a cation, and vouches for the fidelity of the study stroke compared treatment to lower systolic blood to the protocol. The writing committee wrote pressure to less than 130 mm Hg with treatment the manuscript and vouches for the complete- to lower it to less than 150 mm Hg and showed ness and accuracy of the data and analysis. The no significant benefit of the lower target with coordinating center was responsible for analyz- respect to the overall risk of another stroke but ing the data. Scientists at the National Institutes a significant benefit with respect to the risk of of Health participated in the design of the study hemorrhagic stroke.23 and as a group had one vote on the steering The hypothesis that a lower systolic blood- committee of the trial.
pressure goal (e.g., <120 mm Hg) would reduce clinical events more than a standard goal was Study Population
designated by a National Heart, Lung, and Blood Participants were required to meet all the follow- Institute (NHLBI) expert panel in 2007 as the ing criteria: an age of at least 50 years, a systolic most important hypothesis to test regarding the blood pressure of 130 to 180 mm Hg (see the prevention of hypertension-related complications Supplementary Appendix), and an increased risk n engl j med 373;22 nejm.org November 26, 2015 The New England Journal of Medicine Downloaded from nejm.org on April 7, 2016. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Intensive vs. Standard Blood-Pressure Control of cardiovascular events. Increased cardiovascu- Takeda Pharmaceuticals International and Arbor lar risk was defined by one or more of the fol- Pharmaceuticals; neither company had any other lowing: clinical or subclinical cardiovascular dis- role in the study.
ease other than stroke; chronic kidney disease, Participants were seen monthly for the first excluding polycystic kidney disease, with an esti- 3 months and every 3 months thereafter. Medi- mated glomerular filtration rate (eGFR) of 20 to cations for participants in the intensive-treat- less than 60 ml per minute per 1.73 m2 of body- ment group were adjusted on a monthly basis surface area, calculated with the use of the four- to target a systolic blood pressure of less than variable Modification of Diet in Renal Disease 120 mm Hg. For participants in the standard- equation; a 10-year risk of cardiovascular disease treatment group, medications were adjusted of 15% or greater on the basis of the Framing- to target a systolic blood pressure of 135 to ham risk score; or an age of 75 years or older. 139 mm Hg, and the dose was reduced if sys- Patients with diabetes mellitus or prior stroke tolic blood pressure was less than 130 mm Hg were excluded. Detailed inclusion and exclusion on a single visit or less than 135 mm Hg on two criteria are listed in the Supplementary Appen- consecutive visits. Dose adjustment was based dix. All participants provided written informed on a mean of three blood-pressure measure- ments at an office visit while the patient was seated and after 5 minutes of quiet rest; the Randomization and Interventions
measurements were made with the use of an Eligible participants were assigned to a sys- automated measurement system (Model 907, tolic blood-pressure target of either less than Omron Healthcare). Lifestyle modification was 140 mm Hg (the standard-treatment group) or encouraged as part of the management strategy. less than 120 mm Hg (the intensive-treatment Retention in the study and adherence to treat- group). Randomization was stratified according ment were monitored prospectively and routinely to clinical site. Participants and study personnel throughout the trial.26 were aware of the study-group assignments, but outcome adjudicators were not.
After the participants underwent randomiza- Demographic data were collected at baseline. tion, their baseline antihypertensive regimens Clinical and laboratory data were obtained at were adjusted on the basis of the study-group baseline and every 3 months thereafter. A struc- assignment. The treatment algorithms were sim- tured interview was used in both groups every ilar to those used in the Action to Control Car- 3 months to obtain self-reported cardiovascular diovascular Risk in Diabetes (ACCORD) trial.22 disease outcomes. Although the interviewers These algorithms and our formulary are listed in were aware of the study-group assignments, they Figures S1 and S2 and Table S1 in the Supple- used the same format for interviews in the two mentary Appendix. All major classes of antihy- groups to minimize ascertainment bias. Medical pertensive agents were included in the formulary records and electrocardiograms were obtained and were provided at no cost to the participants. for documentation of events. Whenever clinical- SPRINT investigators could also prescribe other site staff became aware of a death, a standard antihypertensive medications (not provided by protocol was used to obtain information on the the study). The protocol encouraged, but did not event.
mandate, the use of drug classes with the stron- Serious adverse events were defined as events gest evidence for reduction in cardiovascular that were fatal or life-threatening, that resulted outcomes, including thiazide-type diuretics (en- in clinically significant or persistent disability, couraged as the first-line agent), loop diuretics that required or prolonged a hospitalization, or (for participants with advanced chronic kidney that were judged by the investigator to represent disease), and beta-adrenergic blockers (for those a clinically significant hazard or harm to the with coronary artery disease).5,27 Chlorthalidone participant that might require medical or surgi- was encouraged as the primary thiazide-type di- cal intervention to prevent one of the other uretic, and amlodipine as the preferred calcium- events listed above.30,31 A short list of monitored channel blocker.28,29 Azilsartan and azilsartan conditions were reported as adverse events if combined with chlorthalidone were donated by they were evaluated in an emergency department: n engl j med 373;22 nejm.org November 26, 2015 The New England Journal of Medicine Downloaded from nejm.org on April 7, 2016. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. hypotension, syncope, injurious falls, electrolyte Prespecified subgroups of interest for all out- abnormalities, and bradycardia. We also moni- comes were defined according to status with re- tored occurrences of acute kidney injury or acute spect to cardiovascular disease at baseline (yes vs. renal failure if they were noted on admission or no), status with respect to chronic kidney disease occurred during a hospitalization and were re- at baseline (yes vs. no), sex, race (black vs. non- ported in the hospital discharge summary as a black), age (<75 vs. ≥75 years), and baseline sys- primary or main secondary diagnosis. The Medi- tolic blood pressure in three levels (≤132 mm Hg, cal Dictionary for Regulatory Activities was used to >132 to <145 mm Hg, and ≥145 mm Hg). We classify the safety events. Coding was performed also planned a comparison of the effects of at the coordinating center, and up to three codes systolic blood-pressure targets on incident de- were assigned to each safety event. The relation- mentia, changes in cognitive function, and cere- ship of serious adverse events to the intervention bral small-vessel ischemic disease; these results was assessed by the trial safety officer and re- are not presented here.
viewed monthly by the safety committee.
We planned a 2-year recruitment period, with a Definitions of study outcomes are outlined in maximum follow-up of 6 years, and anticipated the Supplementary Appendix. A committee whose a loss to follow-up of 2% per year. With an en- members were unaware of the study-group as- rollment target of 9250 participants, we estimated signments adjudicated the clinical outcomes that the trial would have 88.7% power to detect specified in the protocol. The primary hypothe- a 20% effect with respect to the primary out- sis was that treatment to reach a systolic blood- come, assuming an event rate of 2.2% per year pressure target of less than 120 mm Hg, as in the standard-treatment group.
compared with a target of less than 140 mm Hg, Our primary analysis compared the time to would result in a lower rate of the composite the first occurrence of a primary outcome event outcome of myocardial infarction, acute coro- between the two study groups with the use of nary syndrome not resulting in myocardial in- the intention-to-treat approach for all randomly farction, stroke, acute decompensated heart assigned participants; for this analysis, we used failure, or death from cardiovascular causes. Cox proportional-hazards regression with two- Secondary outcomes included the individual sided tests at the 5% level of significance, with components of the primary composite outcome, stratification according to clinic. Follow-up time death from any cause, and the composite of the was censored on the date of last event ascertain- primary outcome or death from any cause.
ment. Interactions between treatment effect and We also assessed renal outcomes, using a dif- prespecified subgroups were assessed with a ferent definition for patients with chronic kidney likelihood-ratio test for the interaction with the disease (eGFR <60 ml per minute per 1.73 m2) at use of Hommel-adjusted P values.32 Interim baseline and those without it. The renal outcome analyses were performed for each meeting of the in participants with chronic kidney disease at data and safety monitoring board, with group- baseline was a composite of a decrease in the sequential stopping boundaries defined with the eGFR of 50% or more (confirmed by a subse- use of the Lan–DeMets method with an O'Brien– quent laboratory test) or the development of Fleming–type spending function.33 The Fine– ESRD requiring long-term dialysis or kidney Gray model for the competing risk of death was transplantation. In participants without chronic used as a sensitivity analysis.34 kidney disease at baseline, the renal outcome was defined by a decrease in the eGFR of 30% or more to a value of less than 60 ml per minute per 1.73 m2. Incident albuminuria, defined for Study Participants
all study participants by a doubling of the ratio A total of 9361 participants were enrolled be- of urinary albumin (in milligrams) to creatinine tween November 2010 and March 2013 (Fig. 1). (in grams) from less than 10 at baseline to Descriptive baseline statistics are presented in greater than 10 during follow-up, was also a Table 1. On August 20, 2015, the NHLBI director prespecified renal outcome.
accepted a recommendation from the data and n engl j med 373;22 nejm.org November 26, 2015 The New England Journal of Medicine Downloaded from nejm.org on April 7, 2016. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Intensive vs. Standard Blood-Pressure Control safety monitoring board of the trial to inform the investigators and participants of the cardio- 14,692 Patients were assessed vascular-outcome results after analyses of the primary outcome exceeded the monitoring bound- ary at two consecutive time points (Fig. S3 in the 5331 Were ineligible or declined Supplementary Appendix), thus initiating the 34 Were <50 yr of age process to end the blood-pressure intervention 352 Had low systolic blood pressure at 1 min after early. The median follow-up on August 20, 2015, was 3.26 years of the planned average of 5 years.
2284 Were taking too many medications or had systolicblood pressure that was out 718 Were not at increased The two treatment strategies resulted in a rapid cardiovascular risk and sustained between-group difference in sys- 703 Had miscellaneous reasons587 Did not give consent tolic blood pressure (Fig. 2). At 1 year, the mean 653 Did not complete screening systolic blood pressure was 121.4 mm Hg in the intensive-treatment group and 136.2 mm Hg in the standard-treatment group, for an average 9361 Underwent randomization difference of 14.8 mm Hg. The mean diastolic blood pressure at 1 year was 68.7 mm Hg in the intensive-treatment group and 76.3 mm Hg in the standard-treatment group (Fig. S4 in the Supple- 4678 Were assigned to intensive 4683 Were assigned to standard mentary Appendix). Throughout the 3.26 years of follow-up, the mean systolic blood pressure was 121.5 mm Hg in the intensive-treatment group and 134.6 mm Hg in the standard-treat- 224 Discontinued intervention 242 Discontinued intervention ment group, and the mean number of blood- 111 Were lost to follow-up 134 Were lost to follow-up 154 Withdrew consent 121 Withdrew consent pressure medications was 2.8 and 1.8, respec- tively. The relative distribution of antihypertensive medication classes used was similar in the two 4678 Were included in the analysis 4683 Were included in the analysis groups, though the use of each class was greater in the intensive-treatment group (Table S2 in the Figure 1. Eligibility, Randomization, and Follow-up.
Discontinued intervention refers to participants who discontinued the study
treatment but did not withdraw consent or become lost to follow­up.
A primary outcome event was confirmed in 562 participants — 243 (1.65% per year) in the inten- sive-treatment group and 319 (2.19% per year) in pendix. The relative risk of death from cardiovas- the standard-treatment group (hazard ratio with cular causes was 43% lower with the intensive intensive treatment, 0.75; 95% confidence inter- intervention than with the standard treatment val [CI], 0.64 to 0.89; P<0.001) (Table 2). Separa- (P = 0.005) (Table 2).
tion in the primary outcome between the groups The numbers needed to treat to prevent a was apparent at 1 year (Fig. 3A). The between- primary outcome event, death from any cause, group differences were consistent across the and death from cardiovascular causes during the components of the primary outcome and other median 3.26 years of the trial were 61, 90, and prespecified secondary outcomes (Table 2).
172, respectively. The effects of the intervention A total of 365 deaths occurred — 155 in the on the rate of the primary outcome and on the intensive-treatment group and 210 in the stan- rate of death from any cause were consistent dard-treatment group (hazard ratio, 0.73; 95% CI, across the prespecified subgroups (Fig. 4, and 0.60 to 0.90; P = 0.003). Separation in mortality Fig. S5 in the Supplementary Appendix). There between the groups became apparent at ap- were no significant interactions between treat- proximately 2 years (Fig. 3B). Causes of death are ment and subgroup with respect to the primary provided in Table S3 in the Supplementary Ap- outcome or death from any cause. When death n engl j med 373;22 nejm.org November 26, 2015 The New England Journal of Medicine Downloaded from nejm.org on April 7, 2016. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Table 1. Baseline Characteristics of the Study Participants.*
Intensive Treatment Standard Treatment
(N = 4678)
(N = 4683)
Criterion for increased cardiovascular risk — no. (%)† Chronic kidney disease‡ Cardiovascular disease Framingham 10­yr cardiovascular disease risk score ≥15% Female sex — no. (%) Among those ≥75 yr of age Race or ethnic group — no. (%)§ Non­Hispanic black Non­Hispanic white Baseline blood pressure — mm Hg Distribution of systolic blood pressure — no. (%) >132 mm Hg to <145 mm Hg Serum creatinine — mg/dl Estimated GFR — ml/min/1.73 m2 Among all participants Among those with estimated GFR ≥60 ml/min/1.73 m2 Among those with estimated GFR <60 ml/min/1.73 m2 Ratio of urinary albumin (mg) to creatinine (g) Fasting total cholesterol — mg/dl Fasting HDL cholesterol — mg/dl Fasting total triglycerides — mg/dl Fasting plasma glucose — mg/dl Statin use — no./total no. (%) Aspirin use — no./total no. (%) Smoking status — no. (%) Framingham 10­yr cardiovascular disease risk score — % n engl j med 373;22 nejm.org November 26, 2015 The New England Journal of Medicine Downloaded from nejm.org on April 7, 2016. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Intensive vs. Standard Blood-Pressure Control Table 1. (Continued.)
Intensive Treatment Standard Treatment
(N = 4678)
(N = 4683)
Body­mass index‖ Antihypertensive agents — no./patient Not using antihypertensive agents — no. (%) * Plus–minus values are means ±SD. There were no significant differences (P<0.05) between the two groups except for statin use (P = 0.04). To convert the values for creatinine to micromoles per liter, multiply by 88.4. To convert the values for cholesterol to millimoles per liter, multiply by 0.02586. To convert the values for triglycerides to millimoles per liter, multiply by 0.01129. To convert the values for glucose to millimoles per liter, multiply by 0.05551. GFR denotes glomer­ ular filtration rate, and HDL high­density lipoprotein.
† Increased cardiovascular risk was one of the inclusion criteria.
‡ Chronic kidney disease was defined as an estimated glomerular filtration rate of less than 60 ml per minute per 1.73 m2 of body­surface area.
§ Race and ethnic group were self­reported.
¶ Black race includes Hispanic black and black as part of a multiracial identification.
‖ The body­mass index is the weight in kilograms divided by the square of the height in meters.
was treated as a competing risk in a Fine–Gray participants in the intensive-treatment group model, the results with respect to the primary (4.7%) and 118 participants in the standard- outcome were virtually unchanged (hazard ratio, treatment group (2.5%) had serious adverse 0.76; 95% CI, 0.64 to 0.89).
events that were classified as possibly or defi- Among participants who had chronic kidney nitely related to the intervention (hazard ratio, disease at baseline, no significant between-group 1.88; P<0.001) (Table S5 in the Supplementary difference in the composite outcome of a de- Appendix). The magnitude and pattern of differ- crease in the eGFR of 50% or more or the devel- ences in adverse events according to treatment opment of ESRD was noted, though the number assignment among participants 75 years of age of events was small (Table 2). Among partici- or older were similar to those in the overall co- pants who did not have chronic kidney disease at hort (Table S6 in the Supplementary Appendix).
baseline, the incidence of the outcome defined by a decrease in the eGFR of 30% or more to a value of less than 60 ml per minute per 1.73 m2 was higher in the intensive-treatment group than SPRINT showed that among adults with hyperten- in the standard-treatment group (1.21% per year sion but without diabetes, lowering systolic blood vs. 0.35% per year; hazard ratio, 3.49; 95% CI, pressure to a target goal of less than 120 mm Hg, 2.44 to 5.10; P<0.001).
as compared with the standard goal of less than 140 mm Hg, resulted in significantly lower rates Serious Adverse Events
of fatal and nonfatal cardiovascular events and Serious adverse events occurred in 1793 partici- death from any cause. Trial participants as- pants in the intensive-treatment group (38.3%) signed to the lower systolic blood-pressure target and in 1736 participants in the standard-treat- (intensive-treatment group), as compared with ment group (37.1%) (hazard ratio with intensive those assigned to the higher target (standard- treatment, 1.04; P = 0.25) (Table 3, and Table S4 treatment group), had a 25% lower relative risk in the Supplementary Appendix). Serious adverse of the primary outcome; in addition, the inten- events of hypotension, syncope, electrolyte ab- sive-treatment group had lower rates of several normalities, and acute kidney injury or acute other important outcomes, including heart fail- renal failure, but not injurious falls or bradycar- ure (38% lower relative risk), death from cardio- dia, occurred more frequently in the intensive- vascular causes (43% lower relative risk), and treatment group than in the standard-treatment death from any cause (27% lower relative risk). group. Orthostatic hypotension as assessed dur- During the follow-up period of the trial (median, ing a clinic visit was significantly less common 3.26 years), the number needed to treat with a in the intensive-treatment group. A total of 220 strategy of intensive blood-pressure control to n engl j med 373;22 nejm.org November 26, 2015 The New England Journal of Medicine Downloaded from nejm.org on April 7, 2016. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Standard treatment Systolic Blood Pressure (mm Hg)
Intensive treatment No. with Data
Standard treatment
Intensive treatment Mean No. of Medications
Standard treatment
Intensive treatment Figure 2. Systolic Blood Pressure in the Two Treatment Groups over the Course of the Trial.
The systolic blood­pressure target in the intensive­treatment group was less than 120 mm Hg, and the target in the
standard­treatment group was less than 140 mm Hg. The mean number of medications is the number of blood­
pressure medications administered at the exit of each visit. I bars represent 95% confidence intervals.
prevent one primary outcome event was 61, and more frequently in the intensive-treatment group the number needed to treat to prevent one death than in the standard-treatment group (Table 3, from any cause was 90. These benefits with re- and Table S5 in the Supplementary Appendix). spect to both the primary outcome and death The differences in adverse renal outcomes may were consistent across all prespecified subgroups, be related to a reversible intrarenal hemody- including participants 75 years of age or older.
namic effect of the greater reduction in blood Owing in part to a lower-than-expected de- pressure and greater use of diuretics, angioten- cline in the eGFR and to the early termination of sin-converting–enzyme inhibitors, and angio- the trial, the number of renal events was small. tensin-receptor blockers in the intensive-treat- Among participants who had chronic kidney ment group.35,36 With the currently available data, disease at baseline, the number of participants there is no evidence of substantial permanent with a decrease in the eGFR of 50% or more or kidney injury associated with the lower systolic reaching ESRD over the course of the trial did blood-pressure goal; however, the possibility of not differ significantly between the two interven- a long-term adverse renal outcome cannot be ex- tion groups. Among participants who did not have cluded. These observations and hypotheses need chronic kidney disease at baseline, a decrease in to be explored further in analyses that incorpo- the eGFR of 30% or more to a value of less than rate more clinical outcomes and longer follow-up.
60 ml per minute per 1.73 m2 occurred more The results of SPRINT add substantially to frequently in the intensive-treatment group than the evidence of benefits of lowering systolic in the standard-treatment group (1.21% per year blood pressure, especially in older patients with vs. 0.35% per year). Among all participants, acute hypertension. Trials such as the Systolic Hyper- kidney injury or acute renal failure occurred tension in the Elderly Program trial,17 the Sys- n engl j med 373;22 nejm.org November 26, 2015 The New England Journal of Medicine Downloaded from nejm.org on April 7, 2016. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Intensive vs. Standard Blood-Pressure Control Table 2. Primary and Secondary Outcomes and Renal Outcomes.*
Hazard Ratio
no. of patients (%) % per year no. of patients (%) % per year (N = 4678)
(N = 4683)
Primary outcome† 0.75 (0.64–0.89) <0.001 Secondary outcomes Myocardial infarction 0.83 (0.64–1.09) Acute coronary syndrome 1.00 (0.64–1.55) 0.89 (0.63–1.25) 0.62 (0.45–0.84) Death from cardiovascular causes 0.57 (0.38–0.85) Death from any cause 0.73 (0.60–0.90) Primary outcome or death 0.78 (0.67–0.90) <0.001 Participants with CKD at baseline
(N = 1330)
(N = 1316)
Composite renal outcome‡ 0.89 (0.42–1.87) ≥50% reduction in estimated GFR§ 0.87 (0.36–2.07) Long­term dialysis 0.57 (0.19–1.54) Kidney transplantation Incident albuminuria¶ 0.72 (0.48–1.07) Participants without CKD at baseline‖
(N = 3332)
(N = 3345)
≥30% reduction in estimated GFR to <60 ml/ 3.49 (2.44–5.10) <0.001 Incident albuminuria¶ 0.81 (0.63–1.04) * CI denotes confidence interval, and CKD chronic kidney disease.
† The primary outcome was the first occurrence of myocardial infarction, acute coronary syndrome, stroke, heart failure, or death from cardio­ vascular causes.
‡ The composite renal outcome for participants with CKD at baseline was the first occurrence of a reduction in the estimated GFR of 50% or more, long­term dialysis, or kidney transplantation.
§ Reductions in the estimated GFR were confirmed by a second laboratory test at least 90 days later.
¶ Incident albuminuria was defined by a doubling of the ratio of urinary albumin (in milligrams) to creatinine (in grams) from less than 10 at baseline to greater than 10 during follow­up. The denominators for number of patients represent those without albuminuria at baseline.
‖ No long­term dialysis or kidney transplantation was reported among participants without CKD at baseline.
tolic Hypertension in Europe trial,11 and the blood-pressure target than that currently recom- Hypertension in the Very Elderly Trial18 showed mended in most patients with hypertension.
the benefits of lowering systolic blood pressure Comparisons between SPRINT and the below 150 mm Hg. However, trials evaluating ACCORD trial22 are inevitable, because the trials systolic blood-pressure levels lower than those examined identical systolic blood-pressure tar- studied in these trials have been either under- gets (<120 mm Hg vs. <140 mm Hg). In contrast powered19-21 or performed without specific sys- to the findings of SPRINT, the cardiovascular tolic blood-pressure targets.37 A major component and mortality benefits observed in the ACCORD of the controversy regarding the selection of the trial were not statistically significant and were systolic blood-pressure goal in this population of a lesser magnitude. Several important differ- has resulted from inadequate data on the risks ences between these trials should be noted. The versus benefits of systolic blood-pressure targets ACCORD trial enrolled participants with diabe- below 150 mm Hg.11,17-21,37 SPRINT now provides tes exclusively, whereas SPRINT excluded par- evidence of benefits for an even lower systolic ticipants with diabetes; in addition, the sample n engl j med 373;22 nejm.org November 26, 2015 The New England Journal of Medicine Downloaded from nejm.org on April 7, 2016. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. observed in SPRINT. The ACCORD trial also A Primary Outcome
used a factorial design that included compari- Hazard ratio with intensive treatment,0.75 (95% CI, 0.64–0.89) sons of standard and intensive glycemic and lipid treatment targets in the same trial. A sec- Standard treatment ondary analysis of the ACCORD results showed that, as compared with the combined standard Intensive treatment glycemia and blood-pressure treatments, inten- sive blood-pressure treatment alone reduced ma- jor cardiovascular outcomes by 26% without additional benefit from combining the two in- tensive treatments.38 Thus, the difference in re- sults between the trials could be due to differ- ences in study design, treatment interactions, or the play of chance. An inherent difference in the No. at Risk
cardiovascular benefits of systolic blood-pressure Standard treatment Intensive treatment lowering between the population with diabetes and the population without diabetes seems un- B Death from Any Cause
likely but cannot be ruled out.
Hazard ratio with intensive treatment, In the Secondary Prevention of Small Sub- 0.73 (95% CI, 0.60–0.90) cortical Strokes trial (intensive systolic blood- pressure goal <130 mm Hg)23 and in the Standard treatment ACCORD trial (intensive systolic blood-pressure goal <120 mm Hg), the lower blood-pressure Intensive treatment target was associated with a nonsignificant 19% lower incidence of stroke (P = 0.08) and a signifi- cant 41% lower incidence of stroke, respectively, than the incidence with higher targets. The intensive-treatment group in SPRINT had a non- significant 11% lower incidence of stroke, though SPRINT also excluded persons with prevalent No. at Risk
stroke or transient ischemic attack at baseline.
Standard treatment In SPRINT, significant between-group differ- Intensive treatment ences were noted in some adverse effects that Figure 3. Primary Outcome and Death from Any Cause.
were attributed to the intervention (Table S5 in Shown are the cumulative hazards for the primary outcome (a composite the Supplementary Appendix). Orthostatic hypo- of myocardial infarction, acute coronary syndrome, stroke, heart failure, or tension as assessed during a clinic visit (Table 3) death from cardiovascular causes) (Panel A) and for death from any cause was observed less frequently in the intensive- (Panel B). The inset in each panel shows the same data on an enlarged y axis. treatment group than in the standard-treatment CI denotes confidence interval.
group (P = 0.01), but syncope was more common among participants in the intensive-treatment size of the ACCORD trial was only half that of group than among those in the standard-treat- SPRINT (4733 vs. 9361). SPRINT enrolled an ment group (3.5% vs. 2.4%, P = 0.003), as was older cohort (mean age, 68 years, vs. 62 years hypotension (3.4% vs. 2.0%, P<0.001). There was in the ACCORD trial), with 28% of participants no between-group difference in injurious falls 75 years of age or older, and also included partici- (hazard ratio, 1.00; P = 0.97). There was a higher pants with chronic kidney disease. The ACCORD rate of acute kidney injury or acute renal failure trial showed a (nonsignificant) 12% lower risk in the intensive-treatment group, as noted above. of its primary composite cardiovascular out- These adverse events need to be weighed against come, with a 95% confidence interval that in- the benefits with respect to cardiovascular events cluded the possibility of a 27% lower risk, which and death that are associated with intensive con- is consistent with the cardiovascular benefit trol of systolic blood pressure.
n engl j med 373;22 nejm.org November 26, 2015 The New England Journal of Medicine Downloaded from nejm.org on April 7, 2016. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Intensive vs. Standard Blood-Pressure Control P Value for
Hazard Ratio (95% CI)
no. of patients with primary outcome/total no. (%) 0.75 (0.64–0.89) 0.70 (0.56–0.87) 0.82 (0.63–1.07) 0.80 (0.64–1.00) 0.67 (0.51–0.86) 0.84 (0.62–1.14) 0.72 (0.59–0.88) 0.77 (0.55–1.06) 0.74 (0.61–0.90) Previous cardiovascular disease 0.71 (0.57–0.88) 0.83 (0.62–1.09) Systolic blood pressure 0.70 (0.51–0.95) >132 to <145 mm Hg 0.77 (0.57–1.03) 0.83 (0.63–1.09) Intensive Treatment Better
Standard Treatment Better
Figure 4. Forest Plot of Primary Outcome According to Subgroups.
The dashed vertical line represents the hazard ratio for the overall study population. The box sizes are proportional to the precision of
the estimates (with larger boxes indicating a greater degree of precision). The subgroup of no previous chronic kidney disease (CKD)
includes some participants with unknown CKD status at baseline. Black race includes Hispanic black and black as part of a multiracial
identification.
The strengths of SPRINT include a large United States, which suggests that control to even sample size, the diversity of the population that level is challenging.39 We excluded patients (including a large proportion of patients 75 years with more severe hypertension, and control of of age or older), and its success in achieving the systolic blood pressure to less than 120 mm Hg intended separation in systolic blood pressure required, on average, one additional antihyper- between the two intervention groups throughout tensive drug. In addition, the median systolic the trial. The lack of generalizability to popula- blood pressure in the intensive-treatment group tions not included in the study — such as per- was just above 120 mm Hg, which indicates that sons with diabetes, those with prior stroke, and more than half the participants had a systolic those younger than 50 years of age — is a limi- blood pressure above the 120 mm Hg target. tation. It is also worth noting that we did not These observations suggest that achieving a sys- enroll older adults residing in nursing homes or tolic blood-pressure goal of less than 120 mm assisted-living facilities. In addition, the effects Hg in the overall population of patients with of the lower blood pressure on the central ner- hypertension would be more demanding and vous system and kidney cannot be reasonably time-consuming for both providers and patients interpreted until analysis of these end points has than achieving a goal of 140 mm Hg, and would been completed.
necessitate increased medication costs and clin- The SPRINT results raise important practical ic visits.
issues. Hypertension control to a blood pressure In conclusion, targeting a systolic blood pres- of less than 140/90 mm Hg is achieved in only sure of less than 120 mm Hg, as compared with about 50% of the general population in the less than 140 mm Hg, in patients at high risk for n engl j med 373;22 nejm.org November 26, 2015 The New England Journal of Medicine Downloaded from nejm.org on April 7, 2016. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Table 3. Serious Adverse Events, Conditions of Interest, and Monitored Clinical Events.
Intensive Treatment
Standard Treatment
(N = 4678)
(N = 4683)
Hazard Ratio
no. of patients (%) Serious adverse event* Conditions of interest Serious adverse event only Electrolyte abnormality Injurious fall† Acute kidney injury or acute renal failure‡ Emergency department visit or serious adverse Electrolyte abnormality Injurious fall† Acute kidney injury or acute renal failure‡ Monitored clinical events Adverse laboratory measure§ Serum sodium <130 mmol/liter Serum sodium >150 mmol/liter Serum potassium <3.0 mmol/liter Serum potassium >5.5 mmol/liter * A serious adverse event was defined as an event that was fatal or life­threatening, that resulted in clinically significant or persistent disability, that required or prolonged a hospitalization, or that was judged by the investigator to represent a clinically significant hazard or harm to the participant that might require medical or surgical intervention to prevent one of the other events listed above.
† An injurious fall was defined as a fall that resulted in evaluation in an emergency department or that resulted in hospitalization.
‡ Acute kidney injury or acute renal failure were coded if the diagnosis was listed in the hospital discharge summary and was believed by the safety officer to be one of the top three reasons for admission or continued hospitalization. A few cases of acute kidney injury were noted in an emergency department if the participant presented for one of the other conditions of interest.
§ Adverse laboratory measures were detected on routine or unscheduled tests; routine laboratory tests were performed at 1 month, then quar­ terly during the first year, then every 6 months.
¶ Orthostatic hypertension was defined as a drop in systolic blood pressure of at least 20 mm Hg or in diastolic blood pressure of at least 10 mm Hg at 1 minute after the participant stood up, as compared with the value obtained when the participant was seated. Standing blood pressures were measured at screening, baseline, 1 month, 6 months, 12 months, and yearly thereafter. Participants were asked if they felt dizzy at the time the orthostatic measure was taken.
cardiovascular events but without diabetes re- The content is solely the responsibility of the authors and sulted in lower rates of fatal and nonfatal major does not necessarily represent the official views of the National cardiovascular events and death from any cause. Institutes of Health (NIH), the Department of Veterans Affairs, or the U.S. Government.
However, some adverse events occurred signifi- Supported by contracts (HHSN268200900040C, cantly more frequently with the lower target.
HHSN268200900046C, HHSN268200900047C, n engl j med 373;22 nejm.org November 26, 2015 The New England Journal of Medicine Downloaded from nejm.org on April 7, 2016. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Intensive vs. Standard Blood-Pressure Control HHSN268200900048C, and HHSN268200900049C) and an inter- of Illinois: UL1TR000050; University of Pittsburgh: UL1TR000005; agency agreement (A-HL-13-002-001) from the NIH, including University of Texas Southwestern: 9U54TR000017-06; University the National Heart, Lung, and Blood Institute (NHLBI), the Na- of Utah: UL1TR000105-05; Vanderbilt University: UL1TR000445; tional Institute of Diabetes and Digestive and Kidney Diseases, George Washington University: UL1TR000075; University of Cali- the National Institute on Aging, and the National Institute of fornia, Davis: UL1TR000002; University of Florida: UL1TR000064; Neurological Disorders and Stroke. Several study sites were University of Michigan: UL1TR000433; and Tulane University: supported by Clinical and Translational Science Awards fund- P30GM103337 COBRE Award NIGMS). The trial was also sup- ed by the National Center for Advancing Translational Sciences ported in part with respect to resources and the use of facilities of the NIH (Case Western Reserve University: UL1TR000439; by the Department of Veterans Affairs.
Ohio State University: UL1RR025755; University of Pennsyl- Disclosure forms provided by the authors are available with vania: UL1RR024134 and UL1TR000003; Boston University: the full text of this article at NEJM.org.
UL1RR025771; Stanford University: UL1TR000093; Tufts Univer- We thank the study participants, without whom this trial sity: UL1RR025752, UL1TR000073, and UL1TR001064; University would not have been possible.
The affiliations of the members of the writing group are as follows: the Division of Nephrology and Hypertension, University Hospitals Case Medical Center, Case Western Reserve University (J.T.W., M.R.), and Division of Nephrology and Hypertension, Louis Stokes Cleveland Veterans Affairs (VA) Medical Center (M.R.), Cleveland; Sticht Center on Aging (J.D.W., K.M.S.), Section on Nephrology (M.V.R.), and Department of Biostatistical Sciences (D.M.R., W.T.A.), Wake Forest School of Medicine, Winston-Salem, NC; Depart- ment of Epidemiology, Tulane University School of Public Health and Tropical Medicine, New Orleans (P.K.W.); Clinical Applications and Prevention Branch, National Heart, Lung, and Blood Institute (J.K.S., L.J.F., J.A.C.), and Division of Kidney, Urologic, and Hema- tologic Diseases, National Institute of Diabetes and Digestive and Kidney Diseases (P.L.K.), Bethesda, MD; Divisions of Cardiovascular Diseases (S.O.) and Preventive Medicine (C.E.L.), University of Alabama at Birmingham, Birmingham; Department of Preventive Medi- cine, University of Tennessee Health Science Center (K.C.J.), and the Preventive Medicine Section, VA Medical Center (W.C.C.), Mem- phis; School of Public Health, University of Colorado, Aurora (D.C.G.); and Division of Nephrology and Hypertension, University of Utah, Salt Lake City (A.K.C.).
References
1. Kearney PM, Whelton M, Reynolds K, 8. Levy D, Larson MG, Vasan RS, Kannel fects of ACE inhibitors, calcium antago-
Muntner P, Whelton PK, He J. Global bur- WB, Ho KK. The progression from hyper- nists, and other blood-pressure-lowering den of hypertension: analysis of world- tension to congestive heart failure. JAMA drugs: results of prospectively designed wide data. Lancet 2005; 365: 217-23.
1996; 275: 1557-62.
overviews of randomised trials. Lancet 2. Go AS, Mozaffarian D, Roger VL, et al. 9. MacMahon S, Peto R, Cutler J, et al. 2000; 356: 1955-64.
Heart disease and stroke statistics — Blood pressure, stroke, and coronary 16. Psaty BM, Smith NL, Siscovick DS, et
2014 update: a report from the American heart disease. Part 1: prolonged differ- al. Health outcomes associated with anti- Heart Association. Circulation 2014; 129(3): ences in blood pressure: prospective ob- hypertensive therapies used as first-line servational studies corrected for the re- agents: asystematic review and meta- 3. Franklin SS. Cardiovascular risks re-
gression dilution bias. Lancet 1990; 335: analysis. JAMA 1997; 277: 739-45.
lated to increased diastolic, systolic and 765-74.
17. SHEP Cooperative Research Group.
pulse pressure: an epidemiologist's point 10. Lewington S, Clarke R, Qizilbash N, Prevention of stroke by antihypertensive
of view. Pathol Biol (Paris) 1999; 47: 594- Peto R, Collins R. Age-specific relevance drug treatment in older persons with iso- of usual blood pressure to vascular mor- lated systolic hypertension: final results 4. Franklin SS, Jacobs MJ, Wong ND, tality: a meta-analysis of individual data of the Systolic Hypertension in the Elderly
L'Italien GJ, Lapuerta P. Predominance for one million adults in 61 prospective Program (SHEP). JAMA 1991; of isolated systolic hypertension among studies. Lancet 2002; 360: 1903-13.
middle-aged and elderly US hyperten- 11. Staessen JA, Fagard R, Thijs L, et al. 18. Beckett NS, Peters R, Fletcher AE, et
sives: analysis based on National Health Randomised double-blind comparison of al. Treatment of hypertension in patients and Nutrition Examination Survey placebo and active treatment for older pa- 80 years of age or older. N Engl J Med (NHANES) III. Hypertension 2001; 37: 869- tients with isolated systolic hypertension. 2008; 358: 1887-98.
Lancet 1997; 350: 757-64.
19. JATOS Study Group. Principal results
5. Chobanian AV, Bakris GL, Black HR, 12. Vasan RS, Larson MG, Leip EP, et al. of the Japanese trial to assess optimal sys-
et al. The seventh report of the Joint Na- Impact of high-normal blood pressure on tolic blood pressure in elderly hyperten- tional Committee on Prevention, Detec- the risk of cardiovascular disease. N Engl sive patients (JATOS). Hypertens Res 2008; tion, Evaluation, and Treatment of High J Med 2001; 345: 1291-7.
31: 2115-27.
Blood Pressure: the JNC 7 Report. JAMA 13. Sundström J, Arima H, Jackson R, et 20. Ogihara T, Saruta T, Rakugi H, et al.
2003; 289: 2560-72.
al. Effects of blood pressure reduction in Target blood pressure for treatment of 6. Collins R, Peto R, MacMahon S, et al. mild hypertension: a systematic review isolated systolic hypertension in the elder-
Blood pressure, stroke, and coronary and meta-analysis. Ann Intern Med 2015; ly: Valsartan in Elderly Isolated Systolic heart disease. 2. Short-term reductions in 162: 184-91.
Hypertension study. Hypertension 2010; blood pressure: overview of randomised 14. Lim SS, Vos T, Flaxman AD, et al. 56: 196-202.
drug trials in their epidemiological con- A comparative risk assessment of burden 21. Verdecchia P, Staessen JA, Angeli F, et
text. Lancet 1990; 335: 827-38.
of disease and injury attributable to 67 al. Usual versus tight control of systolic 7. Hsu CY, McCulloch CE, Darbinian J, risk factors and risk factor clusters in 21 blood pressure in non-diabetic patients
Go AS, Iribarren C. Elevated blood pres- regions, 1990-2010: a systematic analysis with hypertension (Cardio-Sis): an open- sure and risk of end-stage renal disease in for the Global Burden of Disease Study label randomised trial. Lancet 2009; 374: subjects without baseline kidney disease. 2010. Lancet 2012; 380: 2224-60.
Arch Intern Med 2005; 165: 923-8.
15. Neal B, MacMahon S, Chapman N. Ef-
22. Cushman WC, Evans GW, Byington
n engl j med 373;22 nejm.org November 26, 2015 The New England Journal of Medicine Downloaded from nejm.org on April 7, 2016. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved. Intensive vs. Standard Blood-Pressure Control RP, et al. Effects of intensive blood-pres- Group. Major outcomes in high-risk hyper- ards model for the subdistribution of a sure control in type 2 diabetes mellitus. tensive patients randomized to angioten- competing risk. J Am Stat Assoc 1999; 94: N Engl J Med 2010; 362: 1575-85.
sin-converting enzyme inhibitor or calcium 496-509.
23. Benavente OR, Coffey CS, Conwit R, channel blocker vs diuretic: the Antihy-
35. Bakris GL, Weir MR. Angiotensin-
et al. Blood-pressure targets in patients pertensive and Lipid-Lowering Treatment converting enzyme inhibitor-associated with recent lacunar stroke: the SPS3 ran- to Prevent Heart Attack Trial (ALLHAT). elevations in serum creatinine: is this a domised trial. Lancet 2013; 382: 507-15.
JAMA 2002; 288: 2981-97.
cause for concern? Arch Intern Med 2000; 24. Working group report: Expert Panel 29. Ernst ME, Carter BL, Goerdt CJ, et al. 160: 685-93.
on a Hypertension Treatment Trial Initia- Comparative antihypertensive effects of 36. Apperloo AJ, de Zeeuw D, de Jong PE.
tive meeting summary, 2007. Bethesda, hydrochlorothiazide and chlorthalidone A short-term antihypertensive treatment- MD, National Heart Lung and Blood Insti- on ambulatory and office blood pressure. induced fall in glomerular filtration rate tute (http://www .nhlbi .nih .gov/ sites/ www Hypertension 2006; 47: 352-8.
predicts long-term stability of renal func- .nhlbi .nih .gov/ files/ hypertsnsion-full .pdf).
30. Office for Human Research Protec-
tion. Kidney Int 1997; 51: 793-7.
25. Systolic Blood Pressure Intervention tions. OHRP guidance on unanticipated 37. Liu L, Zhang Y, Liu G, Li W, Zhang X,
Trial (SPRINT) protocol. November 1, problems and adverse events. 2007 (http:// Zanchetti A. The Felodipine Event Reduc- 2012 (https:/ / www .sprinttrial .org/ public/ www .hhs .gov/ ohrp/ policy/ advevntguid tion (FEVER) Study: a randomized long- term placebo-controlled trial in Chinese 26. Ambrosius WT, Sink KM, Foy CG, et 31. Food and Drug Administration. Code hypertensive patients. J Hypertens 2005;
al. The design and rationale of a multi- of Federal Regulations. Title 21CFR312.32a, 23: 2157-72.
center clinical trial comparing two strate- 2013 (http://www .accessdata .fda .gov/ 38. Margolis KL, O'Connor PJ, Morgan
gies for control of systolic blood pressure: scripts/ cdrh/ cfdocs/ cfcfr/ cfrsearch .cfm?fr TM, et al. Outcomes of combined cardio- the Systolic Blood Pressure Intervention =312 .32).
vascular risk factor management strate- Trial (SPRINT). Clin Trials 2014; 11: 532- 32. Hommel G. A stagewise rejective gies in type 2 diabetes: the ACCORD ran-
multiple test procedure based on a modi- domized trial. Diabetes Care 2014; 27. Lindholm LH, Carlberg B, Samuels-
fied Bonferroni test. Biometrika 1988; 75: 1721-8.
son O. Should beta blockers remain first 383-86.
39. Nwankwo T, Yoon SS, Burt V, Gu Q.
choice in the treatment of primary hyper- 33. Proschan MA, Lan KKG, Wittes JT. Hypertension among adults in the United
tension? A meta-analysis. Lancet 2005; Statistical monitoring of clinical trials: States: National Health and Nutrition Ex- 366: 1545-53.
a unified approach. New York: Springer, amination Survey, 2011–2012. NCHS Data 28. ALLHAT Officers and Coordinators 2006.
Brief 2013; 133: 1-8. for the ALLHAT Collaborative Research 34. Fine JP, Gray RJ. A proportional haz- Copyright 2015 Massachusetts Medical Society.
receive immediate notification when an article
is published online first
To be notified by e-mail when Journal articles are published Online First, sign up at NEJM.org.
n engl j med 373;22 nejm.org November 26, 2015 The New England Journal of Medicine Downloaded from nejm.org on April 7, 2016. For personal use only. No other uses without permission. Copyright 2015 Massachusetts Medical Society. All rights reserved.

Source: http://ajmed.fr/wp-content/uploads/2016/04/Etude-Sprint.pdf