Low Diastolic Blood Pressure is Not Related to Risk of First Episode of Stroke in a High-Risk Population: A Secondary Analysis of SPRINT

Transcription

1 Low Diastolic Bloo Pressure is Not Relate to Risk of First Episoe of Stroke in a High-Risk Population: A Seconary Analysis of SPRINT Piotr Sobieraj, MD; Jacek Lewanowski, MD, PhD; Maciej Sinski, MD, PhD; Bartosz Symonies, MD, PhD; Zbigniew Gaciong, MD, PhD Backgroun- Hypertension is the most prevalent an leaing risk factor for stroke. SPRINT (The Systolic Bloo Pressure Intervention Trial) assesse the effects on cariovascular event risk of intensive compare with stanar systolic bloo pressure reuction. In this seconary analysis of SPRINT ata, we investigate how low on-treatment iastolic bloo pressure (DBP) influence risk for stroke events. Methos an Results- For this analysis, we use SPRINT_POP (Primary Outcome Paper) Research Materials from the National Heart, Lung an Bloo Institute (NHLBI) Biologic Specimen an Data Repository Information Coorinating Center. Data for 8944 SPRINT participants were analyze from the perio after target bloo pressure was achieve until the en of the trial. Overall, there were 110 stroke events, incluing 49 from the intensive-treatment arm an 61 in the stanar-treatment group. In participants with DBP <70, stroke risk was higher than with DBP 70 (hazar ratio, 1.467; 95% CI ; P=0.0445). Univariable Cox proportional hazar risk analysis showe that in the whole group, age an cariovascular an chronic renal iseases were stroke risk factors. These risk factors were relate to lower DBP an higher pulse pressure, however, not to stuy arm. Multivariable Cox proportional hazar analysis reveale that only age, history of cariovascular isease, current smoking status an on-treatment systolic bloo pressure were significantly relate to stroke risk. Conclusions- Low on-treatment DBP is not relate to the risk for the first stroke, in contrast to oler age, the history of cariovascular isease, current smoking status, an on-treatment systolic bloo pressure. Clinical Trial Registration- URL: Unique ientifier: NCT ( J Am Heart Assoc. 2019;8: e DOI: /JAHA ) Key Wors: bloo pressure iastolic bloo pressure hypertension J-shape curve SPRINT (Systolic Bloo Pressure Intervention Trial) stroke Hypertension is the most prevalent treatable risk factor for stroke an other vascular events. 1,2 Observational stuies have emonstrate that stroke risk ecreases continuously as systolic bloo pressure (SBP) ecreases own to The magnitue of safe bloo pressure (BP) reuction in stroke prevention has yet to be establishe in ranomize trials. Benefits from tight BP control in the HOT (Hypertension Optimal Treatment) trial (<120/70 ) were not obvious. 4 In the population with iabetes mellitus in the ACCORD (Action to Control Cariovascular Risk in Type 2 Diabetes) trial, 5 the group with tight BP control ha fewer stroke events, but no benefit was seen for other en points. In a meta-analysis of tight versus normal BP control, Lee et al conclue that achieving an SBP <130 compare with 130 to 139 provie aitional stroke protection only among people with risk factors. 6 The SPRINT (Systolic Bloo Pressure Intervention Trial) results showe that achieving SBP <120, compare with <140, resulte in lower rates of fatal an non-fatal major cariovascular events an eath from any cause; however, stroke events were not ecrease. 7 In 1979, Stewart reporte an association of lower iastolic BP (DBP) with increase myocarial infarction rates in patients with hypertension An increase in stroke risk in From the Department of Internal Meicine, Hypertension an Vascular Diseases, Meical University of Warsaw, Polan. Corresponence to: Piotr Sobieraj, MD, Department of Internal Meicine, Hypertension an Vascular Diseases, Meical University of Warsaw, Banacha 1a, Warsaw, Polan. piotr.sobieraj@wum.eu.pl Receive September 19, 2018; accepte January 4, ª 2019 The Authors. Publishe on behalf of the American Heart Association, Inc., by Wiley. This is an open access article uner the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use an istribution in any meium, provie the original work is properly cite, the use is noncommercial an no moifications or aaptations are mae. DOI: /JAHA Journal of the American Heart Association 1

2 the Rotteram observational stuy began at DBP < Other authors have propose that the relationship between cariovascular risk an BP reuction may follow a J-shape curve. 12,13 However, the SHEP (Systolic Hypertension in the Elerly Program) trial foun fewer stroke events with DBP <70 achieve in the active treatment group. 14 Similarly, in the Hypertension in the Very Elerly Trial, which suggeste a cause-an-effect relationship between lower DBP an averse cerebrovascular outcomes, DBP <80 was associate with a significant reuction in fatal stroke. 15 A linear rather than a J-shape relationship between stroke events an BP has thus been postulate, but ata from large, ranomize controlle trials are neee to provie efinitive answers. 16,17 The SPRINT trial primary analyses i not present ata on DBP, although the reuction in SBP also influences DBP. For this reason, in this seconary analysis of SPRINT ata, we investigate how a low DBP influence the first stroke occurrence in the SPRINT population. Methos Clinical Perspective What Is New? Low iastolic bloo pressure (<70 ) is relate to the first stroke incience in hypertensive high cariovascular risk subjects. The impact of low iastolic bloo pressure on stroke risk is iminishe by patient s characteristic, especially age, prior meical history, an smoking habits. What Are the Clinical Implications? Among hypertensives at high cariovascular risk, low iastolic bloo pressure shoul not be an obstacle in achieving target systolic bloo pressure when stroke risk reuction is consiere. SPRINT was a ranomize, multicenter open-label trial evaluating how SBP reuctions to <120 versus <140 affecte cariovascular risk. In total, SPRINT enrolle 9361 participants. Through 3.26 years of follow-up, mean SBP values in the intensive an stanar treatment arms were an 134.6, respectively in the intensive an stanar treatment arm. A significant reuction in the primary outcome (a composite of myocarial infarction, other acute coronary synromes, stroke, heart failure, or eath from cariovascular causes) was foun with intensive compare with stanar treatment. The rationale, protocol, an results of SPRINT have been publishe an wiely iscusse elsewhere. 7,18 Stuy Population Accoring to the SPRINT protocol we ecie to analyze the interval from the 6-month visit until the en of the stuy because uring this perio, BP was relatively stable. 7,19 Figure 1 presents SBP an DBP over the uration of the trial with the analyze stuy perio highlighte. Our stuy inclue 8944 participants who were recognize as having high risk for cariovascular iseases an events, with the following risk factors: oler age, SBP 130 to 180, an a history of cariovascular or chronic kiney isease or Framingham Risk Score for 10-year cariovascular isease risk >15%. Inclusion an exclusion criteria are escribe elsewhere. 7 Participants were ranomize to intensive (target SBP <120 ) or stanar treatment (target SBP <140 ). A target DBP was not selecte, although after meeting the goal for SBP, participants were treate to achieve DBP <90. The SPRINT stuy protocol assume own-titration of antihypertensive agents if the SBP reache <130 at a single visit or <135 at 2 consecutive visits in the stanar treatment arm regarless of DBP. 19 BP an Clinical an Laboratory Measurements SBP an DBP were measure 3 times uring each visit using an automate office system (Moel 907, Omron Healthcare), an the mean of those measurements was calculate. For Figure 1. Systolic bloo pressure (SBP) an iastolic bloo pressure (DBP) in stanar an intensive treatment arms throughout the trial, with the current analysis perio highlighte in yellow. DOI: /JAHA Journal of the American Heart Association 2

3 each iniviual, we calculate the meian SBP an DBP measure at visits from the sixth month until the en of the stuy. For each stuy group, we calculate the mean of these meians. Anthropometric, laboratory, an other ata collecte uring the original stuy were use, an pulse pressure was compute. 7 The following events were efine as inicating clinical cariovascular isease: previous myocarial infarction or acute coronary synrome; percutaneous coronary intervention; coronary artery bypass grafting; caroti enarterectomy or stenting; peripheral artery isease with revascularization; electrocariographic changes on a grae exercise test; a positive imaging stuy; 50% iameter stenosis of a coronary, caroti, or lower extremity artery; an abominal aortic aneurysm 5 cm. 7 Data on the clinical primary en point an BP were collecte for ays. Outcome For this analysis, the first occurrence of stroke was consiere as the primary clinical en point (CE). In SPRINT, stroke was iagnose base on signs an symptoms an on compute tomography or magnetic resonance imaging of the brain an large vessels showing a new lesion. In the absence of a new lesion on imaging, clinical finings consistent with the occurrence of stroke that laste more than 24 hours were require. Data Source The ata for this manuscript were accesse from SPRINT_POP (Primary Outcome Paper) Research Materials obtaine from the National Heart, Lung an Bloo Institute (NHLBI) Biologic Specimen an Data Repository Information Coorinating Center (BIOLINCC). This analysis an report o not necessarily reflect the opinions or views of SPRINT_POP or the NHLBI. SPRINT_POP Research Materials are available by BIOLINCC upon reasonable request. The analysis receive Meical University of Warsaw Ethics Committee approval. The SPRINT stuy was approve by the institutional review boar at each participating stuy site. Informe consent was obtaine from all participants. Statistical Analysis This is a retrospective analysis of SPRINT ata. Stuent t tests an analysis of variance were employe for the comparison of 2 groups, respectively. The chi square test was use to compare iscrete variables. Univariate an multivariate Cox proportional hazar risk moels were use to estimate stroke risk. A final multivariate Cox moel was establishe after removing non-significant variables (P=0.05 was selecte as the cut-off value). All Cox moels were teste for assumption of proportional hazars using scale Schoenfel resiuals. We use Kaplan Meier curves to present stroke-free survival with log-rank test comparison. The ifferences were consiere significant at P<0.05. Continuous variables are presente as means an stanar eviations. Discrete variables are expresse as percentages. All computations were performe using STATISTICA 13 (StatSoft, Tulsa, OK, USA) with coe programme in the R environment for statistical computations. 20 Stanar, survival, an survminer packages were use. 21,22 Results Our stuy inclue 8944 patients (95.5% SPRINT participants) 3156 (35.3%) men an 5788 (64.7%) women. The patients were ranomly allocate to stanar (4463; 49.9%) or intensive treatment (4481; 50.1%). In the stanar an intensive treatment arms, on-treatment SBP an DBP were, respectively, / an / (P<0.0001). Descriptive statistics for the whole stuy population an comparison between stroke an non-stroke groups are presente in Table 1. Stroke occurre in 110 (1.2%) participants uring the analyze perio: 61 in the stanar treatment group an 49 in the intensive treatment group (55.5% versus 44.5%; P=0.24). During the first 6 months of the original trial, 22 strokes occurre; we have not inclue these in our analysis. The mean (within participants) of meian (within iniviual participant visit) SBP values was ; for DBP, it was The prevalence of clinical cariovascular isease, chronic kiney isease, an age 75 years were higher in participants with stroke than in those without it uring the analyze perio (30% versus 16.5%, P<0.0002; 39.1% versus 28%, P=0.01; an 40% versus 27.7%, P=0.004, respectively). A etaile comparison between participants with DBP <70 an 70 is presente in Table 2. In the whole stuy population, the first-stroke rate was higher with DBP <70 (1.5% versus 1.0%, P=0.044). The stanar treatment group ha significantly higher stroke rates with DBP <70 (corresponing to the lowest DBP quartile) compare with higher DBP (2.0% versus 1.1%, P=0.03). In the intensive treatment group, however, the stroke rate i not iffer between participants with DBP <70 an those with DBP 70 (1.3% versus 0.9%, P=0.24). Overall, the stuy participants with DBP <70 ha a higher incience of clinical cariovascular iseases an chronic kiney isease, respectively (22.1% versus 12.1%, P<0.0001; an 34.7% versus 23.3%, P<0.0001). These participants were also oler ( versus years, P<0.0001). Despite lower SBP (123.6 versus 130.6, P<0.0001), DOI: /JAHA Journal of the American Heart Association 3

4 Table 1. Clinical Characteristics of the Group an Comparisons Between Those Who Ha an Di Not Have a Stroke Parameter Allocation to intensive treatment arm (n, %) All Participants (n=8944) Participants Who Met CE (n=110) Remaining Participants (n=8834) 4463 (49.9) 49 (44.5) 4432 (50.2) On-treatment SBP, On-treatment DBP, On-treatment PP, Baseline SBP, Baseline DBP, Baseline PP, Women (n, %) 3156 (35.3) 39 (35.5) 3117 (35.3) Age, y Smoking status (non/former/ current smokers) (%) 44.1/42.7/ /40/ /42.7/13.1 Black (n, %) 2793 (31.2) 28 (25.5) 2765 (31.3) BMI, kg/m Clinical cariovascular 1491 (16.1) 33 (30.0) 1458 (16.5) Chronic kiney Time of observation, Time to event /censoring, 2515 (28.1) 43 (39.1) 2472 (28) BMI inicates boy mass inex; CE, clinical en point; DBP, iastolic bloo pressure; PP, pulse pressure; SBP, systolic bloo pressure. participants with DBP <70 ha higher pulse pressure (61.3 versus 52.8, P<0.0001). DBP <70 was relate to higher stroke risk (hazar ratio [HR], 1.47; 95% CI, ; P=0.0445) (Figure 2). Univariate proportional hazar Cox risk moels reveale that age (HR, 1.05; 95% CI, ; P<0.001), history of clinical cariovascular isease (HR, 2.21; 95% CI, ; P=0.001), history of chronic kiney isease (HR, 1.66; 95% Table 2. Characteristics of Participants With DBP <70 an DBP 70 Parameter Allocation to intensive treatment arm (n, %) Participants With DBP <70 mm Hg (n=3792) Participants With DBP 70 mm Hg (n=5152) P Value 2604 (68.7) 1877 (36.4) < On-treatment < SBP, On-treatment < DBP, On-treatment < PP, Baseline SBP, Baseline DBP, < Baseline PP, < Women (n, %) 1377 (36.3) 1779 (34.5) 0.08 Age, y < Smoking status (non/former/ current smokers) (%) 43.3/47.7//9 44.6/39.2/16.2 < Clinical cariovascular Chronic kiney Time of observation, Time to event/ censoring, 839 (22.1) 652 (12.7) < (34.7) 1200 (23.3) < DBP inicates iastolic bloo pressure; PP, pulse pressure; SBP, systolic bloo pressure. CI, ; P=0.010), an on-treatment SBP (HR, 1.03; 95% CI, ; P<0.001) significantly affecte stroke risk. In accorance with other stuies, we evaluate multivariate proportional hazar risk moels beyon available parameters an on-treatment DBP to assess the impact on stroke risk of low on-treatment DBP Multivariate proportional hazar risk moels beyon age, clinical cariovascular isease, chronic kiney isease, current smoking status, ontreatment SBP, an on-treatment DBP (inclue as continuous variable or as ichotomous parameter using the cut-off of 70 ) are presente in Table 3. The application of these moels reveale that only age, history of clinical DOI: /JAHA Journal of the American Heart Association 4

5 Figure 2. Kaplan Meier plot of stroke-free survival in participants with iastolic bloo pressure (DBP) <70 an 70. P value compute for comparison between Kaplan Meier curves. cariovascular isease, current smoking status, an ontreatment SBP were significant. Baseline SBP an DBP were not significant factors after being entere into any of the moels. Figure 3 presents unajuste an ajuste HRs in both treatment arms accoring to DBP quartile. Table 4 provies a etaile comparison within quartiles of DBP in both treatment arms. Discussion The main fining of the present stuy is that in the stanar treatment group in participants with DBP <70, the rate of the first stroke was higher than for those with DBP 70 (2.0% versus 1.1%, P=0.03). Similar relationship in the stroke rate was foun in all investigate subjects with DBP <70 ie, in both intensive an stanar treatment groups (1.5% versus 1.0%, P=0.044). In contrast, the participants in the intensive-target group i not iffer in stroke rates with DBP <70 an DBP 70 (1.3% versus 0.9%, P=0.24, respectively). We suggest that the similar stroke rate between these 2 subgroups having intensive treatment traces to a reuce stroke risk from the lower SBP compare with the stanar arm ( versus , P<0.0001). The participants with stroke ha higher SBP an pulse pressure ( versus , P=0.001; an versus , P<0.0001; respectively). Table 3. Multivariate Cox Proportional Hazar Risk Moels for Stroke Parameter Hazar risk (95% CI) P Value Moel A Age, y 1.04 ( ) <0.001 History of clinical 1.92 ( ) cariovascular isease Current smoking status 1.78 ( ) 0.04 On-treatment SBP, 1.03 ( ) <0.001 Moel B Age, y 1.04 ( ) History of clinical 1.87 ( ) cariovascular isease Current smoking status 1.76 ( ) On-treatment DBP < ( ) On-treatment SBP, 1.03 ( ) <0.001 Moel C Age, y 1.05 ( ) History of clinical 1.93 ( ) cariovascular isease Current smoking status 1.78 ( ) 0.04 On-treatment DBP, 1.00 ( ) 0.96 On-treatment SBP, 1.03 ( ) Moel D Age, y 1.05 ( ) <0.001 History of clinical 1.97 ( ) cariovascular isease Current smoking status 1.84 ( ) 0.03 On-treatment DBP, 1.02 ( ) Allocation to intensive 0.89 ( ) treatment arm Moel E Age, y 1.05 ( ) History of clinical 1.92 ( ) cariovascular isease Current smoking status 1.79 ( ) 0.04 On-treatment DBP, 1.00 ( ) On-treatment SBP, 1.04 ( ) Female sex 0.98 ( ) 0.93 DBP inicates iastolic bloo pressure; SBP, systolic bloo pressure. The existence of a J-curve phenomenon reflecting an averse relationship between excessive BP reuction an cariovascular risk is wiely ebate. SPRINT results showe beneficial an safe effects of SBP reuction, an these ata have strongly influence the newest US guielines on hypertension management in aults. 7,26 SPRINT reveale that DOI: /JAHA Journal of the American Heart Association 5

6 Figure 3. Plots of hazar ratios (HRs) with 95% CIs in quartiles of iastolic bloo pressure (DBP) in stanar an intensive treatment arms. HRs were compute in relation to the thir quartile of DBP in each treatment arm. Ajustment of HR was performe as efine in a multivariate proportional hazar risk moel. A an B, unajuste an ajuste HRs in the stanar treatment arm, respectively; C an D, unajuste an ajuste HRs in the intensive treatment arm, respectively. targeting SBP <120 resulte in lower rates of its composite primary outcome; nevertheless, the analysis of its iniviual components showe that SBP ecrease i not reuce the number of myocarial infarctions (HR, 0.83; 95% CI, , P=0.19) or strokes (HR, 0.89; 95% CI, ; P=0.50). 7 Several stuies have aresse the influence of low DBP on cariovascular events, but none has irectly assesse the relationship between low DBP an stroke risk. A few post hoc analyses have been publishe base on SPRINT ata, evaluating the effects of DBP lowering on cariovascular risk. Behu et al showe an association of the lowest quintile of baseline DBP with an increase risk for the composite cariovascular outcome in both intensive an stanar therapy groups ( an , respectively). 27 Stroke, however, was assesse as a part of the composite primary outcome an not as a separate en point. These authors inicate that a higher rate of cariovascular episoes observe DOI: /JAHA Journal of the American Heart Association 6

7 Table 4. Comparison Within s of DBP in Both Treatment Arms Intensive Treatment Arm Stanar Treatment Arm First (n=1211) Secon (n=1159) Thir (n=1056) Fourth (n=1055) P Value First (N=1188) Secon (N=1073) Thir (N=1151) Fourth (N=1051) P Value On-treatment SBP, < <0.001 On-treatment DBP, < <0.001 On-treatment PP, < <0.001 Baseline SBP, Baseline DBP, < <0.001 Baseline PP, < <0.001 Women (n, %) 429 (35.4) 391 (33.7) 387 (36.6) 394 (37.3) (39.1) 386 (36.0) 391 (34.0) 314 (29.9) <0.001 Age, y < <0.001 Smoking status (non/former/ current smokers) (%) 43.4/49.6/ /44.1/ /40.2/ /34.8/22.3 < /50.2/ /46.2/ /40.5/ /34.1/22.5 <0.001 Clinical cariovascular 295 (24.4) 201 (17.3) 137 (13.0) 115 (10.9) < (25.9) 183 (17.1) 152 (13.2) 100 (9.5) <0.001 Chronic kiney 461 (38.1) 337 (29.1) 246 (23.3) 234 (22.2) < (39.1) 306 (28.5) 257 (22.3) 210 (20.0) <0.001 Time of observation, < <0.001 Time to event/ censoring, < <0.001 DBP inicates iastolic bloo pressure; PP, pulse pressure; SBP, systolic bloo pressure. DOI: /JAHA Journal of the American Heart Association 7

8 at lower levels of DBP was not a irect response to DBP reuction but rather the result of patients clinical characteristics associate with low DBP. That inference is in line with our finings on stroke. In a multivariate hazar risk moel, we foun that age an a history of clinical cariovascular isease significantly affecte stroke risk. Stensru et al foun that after ajustment for treatment, low DBP (<60 ) was associate with a poor primary outcome, incluing stroke (HR, 1.9; 95% CI, ). 28 After ajustment for emographic an clinical covariates, however, HR values significantly improve (HR, 1.04; 95% CI, ]). A further analysis confine to the participants over 75 years foun no harmful effects of DBP <60 on the primary outcome. The results of both stuies showe that the effect of low DBP on the risk of the composite en point is remove following ajustment for confouning variables. Neither Behu et al nor Stensru et al inclue a separate analysis for stroke risk. 27,28 A limite number of trials have aresse the possibility of a J-shape relationship between low DBP an stroke risk. Voko et al foun that in elerly participants treate to DBP <65, stroke risk was clearly higher than with DBP 65 to 74. This fining supports the iea of the J- curve phenomenon. 11 In patients age >60 years who participate in SHEP, those who were actively treate to DBP 68 ha a significantly lower relative risk of stroke (0.64, P=0.0003) compare with the placebo group with DBP In the ACCORD trial, which inclue patients with iabetes mellitus, DBP 64.4 versus 70.5 resulte in a significant reuction in all strokes an non-fatal strokes. 5 In another analysis, the participants in ACCORD who met the eligibility criteria for SPRINT intensive BP control ha a significantly reuce risk of the main composite outcome. This outcome inclue stroke, with the tren to the reuction in non-fatal stroke risk. 29 A subsequent analysis evaluate intensive an stanar BP control in ACCORD participants 4 years after its termination. 30 After >9 years of follow-up, similar to the previous stuy, intensive BP control reuce the risk for the composite primary outcome by 25% (HR, 0.75; 95% CI, ; P=0.02); however, the risk for non-fatal stroke remaine unchange. The effects were observe espite similar BP achieve uring observational follow-up. None of these stuies irectly reporte DBP, but their results support the favorable effects of intensive BP reuction observe in the ACCORD stuy. In contrast to the ACCORD stuy, the HOT trial showe no irect effect on stroke risk of DBP reuctions to 90, 85, 80, or 70 in the overall stuy population or in the subgroup of participants with iabetes mellitus. 4 Only the group with ischemic heart isease ha a 43% stroke reuction with DBP 80 compare with a 90 target. Vial-Petiot et al analyze patients with coronary artery isease in ifferent ranges of systolic an iastolic BP 31 an foun no effect on stroke risk of DBP <60 or 60 to 69 compare with 70 to 79. These results seem to argue against a J-curve phenomenon. The stuy, however, was base on the population from routine meical practice with no preefine BP interventions, an the number of stroke episoes was smaller than the numbers for other investigate en points. McEvoy et al foun similar outcomes in a cohort from the Atherosclerosis Risk in Communities stuy, investigating the association between DBP <60 an cariovascular en points incluing stroke. 32 Over a meian follow-up of 21 years, DBP reuction to <60 was associate with a higher risk of myocarial amage an mortality compare with DBP 80 to 89 ; however, the risk of stroke was not increase with the lower DBP (HR, 1.13; 95% CI, ). These authors emphasize that the sensitivity analysis evaluating a SPRINT-eligible subpopulation was unerpowere because of a small sample size. 32 The iverse effects of DBP reuctions on stroke risk an myocarial infarction are possibly relate to physiological ifferences in regulation of cerebral an coronary bloo flow. Coronary bloo flow occurs preominantly uring the iastolic phase of the cariac cycle an epens mostly on heart rate an DBP ifferences between aorta an mean pressure in the right atrium. Consequently, anything that significantly ecreases DBP or increases heart rate will ecrease coronary bloo flow an lea to heart muscle amage. 31 Cerebral bloo flow in normotensive aults is preserve between about 60 an 160. In hypertension, however, these values can shift upwar. 33 All BP fluctuations beyon the range of autoregulation mean that cerebral flow shows a linear epenence on mean BP. Our analysis has got some limitations. Firstly, as it is necessary with a post hoc ata analysis, the conclusions shoul be interprete an applie with caution. Secon, SPRINT was stoppe prematurely, after a meian follow-up of 3.26 years; a longer follow-up coul substantially have increase the number of strokes an altere these final results. Also it is worth noting that SPRINT i not inclue the participants with previous stroke, iabetes mellitus, or lower cariovascular risk, so these results o not reflect potential effects in these subpopulations. Some concerns have been raise about the specific BP measurement metho in SPRINT. During the stuy, 3 unattene BP recorings were mae after a 5-minute perio of rest an then average. It was note that the automate metho of BP measurement might not correspon to the routine clinic BP practices because of the elimination of patient- an clinician-relate factors. Agarwal foun that 3 average measurements calle research-grae SBP/DBP were lower by 12.7/12 than routine clinic DOI: /JAHA Journal of the American Heart Association 8

9 measurements. 34 Filipovsky et al inicate ifferences between automate an office SBP/DBP recorings ( /8.07.3, respectively), with even lower automate than home BP recorings. 35 These observations suggest that BP values achieve in SPRINT were consierably higher than the intene goals of the therapy. One may assume that, the cut point in our stuy for the lowest range of DBP shoul be increase to 70 to 80. However, the recent survey results showe that both BP an risk reuction were similar in the intensive SPRINT group regarless of the metho. 36 The choice of meian as a parameter to escribe BP in our stuy requires explanation. Similar publishe reanalyses have use means covering the entire trial perio. 37,38 In our opinion, meian is a better central tenency parameter to characterize DBP an SBP because of the strategy use to achieve BP goals an the lower susceptibility of meians to outlier measurements. We performe aitional analyses an conclue that the choice of meian or mean i not affect our conclusions (ata not shown). Conclusion In summary, our results inicate that in a high-risk SPRINT population with low DBP (<70 ), stroke risk was increase only if the participants were oler, ha a history of clinical cariovascular isease, were smokers, or ha a high SBP. We emphasize that the current stuy searches for causal relationship between low DBP an the risk of stroke. What we foun was that apart from DBP, other clinical variables play significant roles. Therefore, DBP oes not appear to be useful for preiction purposes. Our finings strongly support the iea that BP goals shoul be tailore to a patient s iniviual characteristics. In each case, the patient s clinical profile shoul be carefully evaluate. Further stuies are neee to resolve the association of low DBP an stroke risk in some subpopulations, incluing those with the history of stroke. Ongoing stuies (such as the Optimal Bloo Pressure an Cholesterol Targets for Preventing Recurrent Stroke in Hypertensives trial) shoul yiel new ata an help to generate appropriate recommenations. 39 Disclosures None. References 1. Go AS, Mozaffarian D, Roger VL, Benjamin EJ, Berry JD, Blaha MJ, Dai S, For ES, Fox CS, Franco S, Fullerton HJ, Gillespie C, Hailpern SM, Heit JA, Howar VJ, Huffman MD, Ju SE, Kissela BM, Kittner SJ, Lacklan DT, Lichtman JH, Lisabeth LD, Mackey RH, Magi DJ, Marcus GM, Marelli A, Matchar DB, McGuire DK, Mohler ER III, Moy CS, Mussolino ME, Neumar RW, Nichol G, Paney DK, Paynter NP, Reeves MJ, Sorlie PD, Stein J, Towfighi A, Turan TN, Virani SS, Wong ND, Woo D, Turner MB. Heart isease an stroke statistics upate: a report from the American Heart Association. Circulation. 2014;129:e28 e MacMahon S, Peto R, Cutler J, Collins R, Sorlie P, Neaton J, Abbott R, Gowin J, Dyer A, Stamler J. Bloo pressure, stroke, an coronary heart isease. Part 1, Prolonge ifferences in bloo pressure: prospective observational stuies correcte for the regression ilution bias. Lancet. 1990;335: Lewington S, Clarke R, Qizilbash N, Peto R, Collins R. Age-specific relevance of usual bloo pressure to vascular mortality: a meta-analysis of iniviual ata for one million aults in 61 prospective stuies. Lancet. 2002;360: Hansson L, Zanchetti A, Carruthers SG, Dahlof B, Elmfelt D, Julius S, Menar J, Rahn KH, Weel H, Westerling S. Effects of intensive bloo-pressure lowering an low-ose aspirin in patients with hypertension: principal results of the Hypertension Optimal Treatment ranomise trial. Lancet. 1998;351: Cushman WC, Evans GW, Byington RP, Goff DC Jr, Grimm RH Jr, Cutler JA, Simons-Morton DG, Basile JN, Corson MA, Probstfiel JL, Katz L, Peterson KA, Frieewal WT, Buse JB, Bigger JT, Gerstein HC, Ismail-Beigi F. Effects of intensive bloo-pressure control in type 2 iabetes mellitus. N Engl J Me. 2010;362: Lee M, Saver JL, Hong KS, Hao Q, Ovbiagele B. Does achieving an intensive versus usual bloo pressure level prevent stroke? Ann Neurol. 2012;71: Wright JT Jr, Williamson JD, Whelton PK, Snyer JK, Sink KM, Rocco MV, Reboussin DM, Rahman M, Oparil S, Lewis CE, Kimmel PL, Johnson KC, Goff DC Jr, Fine LJ, Cutler J, Cushman WC, Cheung AK, Ambrosius WT. A ranomize trial of intensive versus stanar bloo-pressure control. N Engl J Me. 2015;373: Stewart IM. Relation of reuction in pressure to first myocarial infarction in patients receiving treatment for severe hypertension. Lancet. 1979;1: Stewart IM. Lowering bloo-pressure. Lancet. 1979;2: Stewart IM. Optimise bloo-pressure. Lancet. 1979;2: Voko Z, Bots ML, Hofman A, Koustaal PJ, Witteman JC, Breteler MM. J-shape relation between bloo pressure an stroke in treate hypertensives. Hypertension. 1999;34: Cruickshank JM, Thorp JM, Zacharias FJ. Benefits an potential harm of lowering high bloo pressure. Lancet. 1987;1: Banach M, Aronow WS. Bloo pressure J-curve: current concepts. Curr Hypertens Rep. 2012;14: SHEP Cooperative Research Group. Prevention of stroke by antihypertensive rug treatment in oler persons with isolate systolic hypertension. Final results of the Systolic Hypertension in the Elerly Program (SHEP). JAMA. 1991;265: Beckett NS, Peters R, Fletcher AE, Staessen JA, Liu L, Dumitrascu D, Stoyanovsky V, Antikainen RL, Nikitin Y, Anerson C, Belhani A, Forette F, Rajkumar C, Thijs L, Banya W, Bulpitt CJ. Treatment of hypertension in patients 80 years of age or oler. N Engl J Me. 2008;358: Malyszko J, Muntner P, Rysz J, Banach M. Bloo pressure levels an stroke: J- curve phenomenon? Curr Hypertens Rep. 2013;15: Lacklan DT, Carey RM, Conforto AB, Rosenorff C, Whelton PK, Gorelick PB. Implications of recent clinical trials an hypertension guielines on stroke an future cerebrovascular research. Stroke. 2018;49: Ambrosius WT, Sink KM, Foy CG, Berlowitz DR, Cheung AK, Cushman WC, Fine LJ, Goff DC Jr, Johnson KC, Killeen AA, Lewis CE, Oparil S, Reboussin DM, Rocco MV, Snyer JK, Williamson JD, Wright JT Jr, Whelton PK. The esign an rationale of a multicenter clinical trial comparing two strategies for control of systolic bloo pressure: the Systolic Bloo Pressure Intervention Trial (SPRINT). Clin Trials. 2014;11: Reboussin D. Systolic bloo pressure intervention trial (SPRINT) protocol, Version 4.0. November 1, 2012, Available at: public/protocol_current.pf. Accesse December 29, R Core Team. R: A Language an Environment for Statistical Computing. R Founation for Statistical Computing. Available at: org/. Accesse December 29, Therneau T. A Package for Survival Analysis in S. Version Available at: Accesse December 29, Kassambara A. Drawing survival curves using ggplot2. Version Available at: Accesse December 29, Ayyagari R, Vekeman F, Lefebvre P, Ong SH, Faust E, Trahey A, Machnicki G, Duh MS. Pulse pressure an stroke risk: evelopment an valiation of a new stroke risk moel. Curr Me Res Opin. 2014;30: DOI: /JAHA Journal of the American Heart Association 9

10 24. Morillas P, Pallares V, Facila L, Llisterri JL, Sebastian ME, Gomez M, Castilla E, Camarasa R, Sanin M, Garcia-Honrubia A. The CHADS2 score to preict stroke risk in the absence of atrial fibrillation in hypertensive patients age 65 years or oler. Rev Esp Cariol (Engl E). 2015;68: Willey JZ, Moon YP, Kahn E, Roriguez CJ, Runek T, Cheung K, Sacco RL, Elkin MS. Population attributable risks of hypertension an iabetes for cariovascular isease an stroke in the northern Manhattan stuy. J Am Heart Assoc. 2014;3:e DOI: /JAHA Whelton PK, Carey RM, Aronow WS, Casey DE Jr, Collins KJ, Dennison Himmelfarb C, DePalma SM, Giing S, Jamerson KA, Jones DW, MacLaughlin EJ, Muntner P, Ovbiagele B, Smith SC Jr, Spencer CC, Staffor RS, Taler SJ, Thomas RJ, Williams KA Sr, Williamson JD, Wright JT Jr ACC/AHA/ AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA guieline for the prevention, etection, evaluation, an management of high bloo pressure in aults: executive summary: a report of the American College of Cariology/ American Heart Association Task Force on Clinical Practice Guielines. Hypertension. 2018;71: Behu S, Chertow GM, Cheung AK, Cushman WC, Rahman M, Greene T, Wei G, Campbell RC, Conroy M, Freeman BI, Haley W, Horwitz E, Kitzman D, Lash J, Papaemetriou V, Pisoni R, Riessen E, Rosenorff C, Watnick SG, Whittle J, Whelton PK. Influence of baseline iastolic bloo pressure on effects of intensive compare with stanar bloo pressure control. Circulation. 2018;137: Stensru MJ, Strohmaier S. Diastolic hypotension ue to intensive bloo pressure therapy: is it harmful? Atherosclerosis. 2017;265: Buckley LF, Dixon DL, Wohlfor GF IV, Wijesinghe DS, Baker WL, Van Tassell BW. Intensive versus stanar bloo pressure control in SPRINT-Eligible participants of ACCORD-BP. Diabetes Care. 2017;40: Buckley LF, Dixon DL, Wohlfor GFT, Wijesinghe DS, Baker WL, Van Tassell BW. Effect of intensive bloo pressure control in patients with type 2 iabetes mellitus over 9 years of follow-up: a subgroup analysis of high-risk ACCORDION trial participants. Diabetes Obes Metab. 2018;20: Vial-Petiot E, For I, Greenlaw N, Ferrari R, Fox KM, Tarif JC, Tenera M, Tavazzi L, Bhatt DL, Steg PG. Cariovascular event rates an mortality accoring to achieve systolic an iastolic bloo pressure in patients with stable coronary artery isease: an international cohort stuy. Lancet. 2016;388: McEvoy JW, Chen Y, Rawlings A, Hoogeveen RC, Ballantyne CM, Blumenthal RS, Coresh J, Selvin E. Diastolic bloo pressure, subclinical myocarial amage, an cariac events: implications for bloo pressure control. J Am Coll Cariol. 2016;68: Phillips SJ, Whisnant JP. Hypertension an the brain. The National High Bloo Pressure Eucation Program. Arch Intern Me. 1992;152: Agarwal R. Implications of bloo pressure measurement technique for implementation of systolic bloo pressure intervention trial (SPRINT). JAm Heart Assoc. 2017;6:e DOI: /JAHA Filipovsky J, Seilerova J, Kratochvil Z, Karnosova P, Hronova M, Mayer O Jr. Automate compare to manual office bloo pressure an to home bloo pressure in hypertensive patients. Bloo Press. 2016;25: Bauer F, Seibert FS, Rohn B, Bauer KAR, Rolshoven E, Babel N, Westhoff TH. Attene versus unattene bloo pressure measurement in a real life setting. Hypertension. 2018;71: Messerli FH, Mancia G, Conti CR, Hewkin AC, Kupfer S, Champion A, Kolloch R, Benetos A, Pepine CJ. Dogma ispute: can aggressively lowering bloo pressure in hypertensive patients with coronary artery isease be angerous? Ann Intern Me. 2006;144: Kjelsen SE, Berge E, Bangalore S, Messerli FH, Mancia G, Holzhauer B, Hua TA, Zappe D, Zanchetti A, Weber MA, Julius S. No evience for a J-shape curve in treate hypertensive patients with increase cariovascular risk: the VALUE trial. Bloo Press. 2016;25: Zanchetti A, Liu L, Mancia G, Parati G, Grassi G, Stramba-Baiale M, Silani V, Bilo G, Corrao G, Zambon A, Scotti L, Zhang X, Guan TR, Zhang Y, Zhang X, Berge E, Reon J, Narkiewicz K, Dominiczak A, Nilsson P, Viigimaa M, Laurent S, Agabiti-Rosei E, Wu Z, Zhu D, Roicio JL, Ruilope LM, Martell-Claros N, Pinto F, Schmieer RE, Burnier M, Banach M, Cifkova R, Farsang C, Konrai A, Lazareva I, Sirenko Y, Dorobantu M, Postazhiyan A, Accetto R, Jelakovic B, Lovic D, Manolis AJ, Stylianou P, Dicker D, Wei G, Xu C, Xie H, Coca A, O Brien J, For G. Continuation of the ESH-CHL-SHOT trial after publication of the SPRINT: rationale for further stuy on bloo pressure targets of antihypertensive treatment after stroke. J Hypertens. 2016;34: DOI: /JAHA Journal of the American Heart Association 10