Achieved Blood Pressure and Outcomes in the Secondary Prevention of Small Subcortical Strokes TrialNovelty and Significance
Abstract—Studies suggest a J-shaped association between blood pressure and cardiovascular events in the setting of intensive systolic blood pressure control; whether there is a similar association with stroke remains less well established. The Secondary Prevention of Small Subcortical Strokes was a randomized trial to evaluate higher (130–149 mm Hg) versus lower (<130 mm Hg) systolic blood pressure targets in participants with recent lacunar infarcts. We evaluated the association of mean achieved blood pressure, 6 months after randomization, and recurrent stroke, major vascular events, and all-cause mortality. After a mean follow up of 3.7 years, there was a J-shaped association between achieved blood pressure and outcomes; the lowest risk was at ≈124 and 67 mm Hg systolic and diastolic blood pressure, respectively. For example, above a systolic blood pressure of 124 mm Hg, 1 standard deviation higher (11.1 mm Hg) was associated with increased mortality (adjusted hazard ratio: 1.9; 95% confidence interval: 1.4, 2.7), whereas below this level, this relationship was inverted (0.29; 0.10, 0.79), P<0.001 for interaction. Above a diastolic blood pressure of 67 mm Hg, a 1 standard deviation higher (8.2 mm Hg) was associated with an increased risk of stroke (2.2; 1.4, 3.6), whereas below this level, the association was in the opposite direction (0.34; 0.13, 0.89), P=0.02 for interaction. The lowest risk of all events occurred at a nadir of ≈120 to 128 mm Hg systolic blood pressure and 65 to 70 mm Hg diastolic blood pressure. Future studies should evaluate the impact of excessive blood pressure reduction, especially in older populations with preexisting vascular disease.
See Editorial Commentary, pp 32–33
Recent guidelines for the treatment of high blood pressure (BP) recommend lowering BP to a specified target, with no mention of a lower limit above which pressure should be maintained.1,2 Whether a lower level for BP is also associated with increased risk of adverse events is uncertain. The majority of the literature points toward the presence of a J-shaped association between BP and cardiovascular events, whereby there is an increased risk of events at both high and low BP levels.3,4 Whether a similar J-shaped relationship exists between BP and risk of stroke is more tenuous.5 Additionally, some observational studies have suggested that low diastolic BP (DBP) in combination with high systolic BP (SBP) is associated with cardiovascular outcomes.6–8 Whether the association between low DBP and increased risk of outcomes is independent of SBP is unclear.
There are limited data on achieved BP in the setting of treatment for stroke prevention, although the existing data suggest a linear relationship between BP and stroke outcomes.5 A limitation of prior literature is that most data come from observational studies or randomized trials that were designed for a composite cardiovascular outcome, and not specifically to evaluate stroke. Low BP may result in inadequate cerebral blood flow because of decreased perfusion pressure.5 This may be exaggerated in persons with preexisting cerebrovascular disease, although no prior randomized trials of BP lowering have evaluated the association of low BP and stroke in the setting of secondary prevention.
The Secondary Prevention of Small Subcortical Strokes (SPS3) trial was a randomized study comparing higher (130–149 mm Hg) versus lower (<130 mm Hg) SBP targets among individuals with recent lacunar infarcts.9 Investigators demonstrated a nonsignificant reduction in both stroke and a composite outcome of myocardial infarction and vascular death in the lower target group compared with the higher target group. In this study, we compared the association between achieved SBP and DBP with the primary outcome of recurrent stroke, as well as major vascular events, and all-cause mortality. We examined each of the BP parameters both individually and in combination to determine the level of achieved BP associated with the lowest risk of stroke, major vascular events, and death.
SPS3 was a randomized, multicenter clinical trial designed to evaluate the effectiveness of 2 antiplatelet treatments in a blinded fashion (aspirin versus aspirin plus clopidogrel) and 2 target levels of SBP (open label) for secondary stroke prevention in participants with recent lacunar infarcts. Details of the study design have been previously published.10 Briefly, persons in North America, Latin America, and Spain aged ≥30 years with a recent symptomatic magnetic resonance imaging–determined lacunar infarct were randomized in a 2-by-2 factorial design to the antiplatelet intervention and to an intensive SBP target of <130 mm Hg or usual target of 130 to 149 mm Hg. Participants were enrolled between 2 weeks and 180 days after the qualifying event. Detailed inclusion and exclusion criteria are reported elsewhere.10 All patients signed an informed consent, and the protocol was approved by each local Institutional Review Board.
Blood Pressure Management and Measurement
Patients were randomly assigned to 1 of 2 SBP targets, irrespective of DBP. Randomization was stratified by clinical center and according to baseline hypertension status (hypertensive versus normotensive) using a permuted-block design. Because the goal was to achieve a target SBP, randomization was not blinded (open-label) and there was no washout period. All major classes of antihypertensive medicines were available to use, with an algorithm based on JNC 7 guidelines developed and distributed to sites.11,12
BP was measured in the sitting position and in the same arm after a strict validated protocol as described previously.10 BP was measured using an automated BP machine (Colin 8800C, Omron, San Antonio, TX) provided to each clinical center. Achieved BP was calculated as the mean BP from all trial readings taken after the first 6 months of follow-up; only those participants with ≥2 BP measures after this time interval are included in the present study (n=2748 of 3020). Participants with <6 months of follow-up were excluded from this analysis.
The primary end point was all recurrent stroke. Ischemic stroke was clinically defined as a focal neurological deficit persisting for >24 h, with an absence of hemorrhage confirmed by neuroimaging. Major vascular events, a secondary outcome, included acute myocardial infarction (defined by standard criteria, ie, compatible clinical history with changes on ECG or in cardiac enzyme concentrations) or need for acute admission to hospital for a major vascular event. Death was classified as vascular, nonvascular, or unknown. All outcomes were ascertained by a site examiner blinded to BP treatment group assignment and confirmed by a central adjudication committee that was blinded to treatment assignment.
Several investigations of BP and outcomes in older adults have demonstrated nonlinear associations; therefore, we used a data-guided approach to identify the relationship between achieved BP and outcomes. To identify the point of lowest risk of events, we began by fitting restricted cubic spline models with 5 knots (5th, 25th, 50th, 75th, and 95th percentiles) of achieved SBP and DBP and all recurrent stroke. Based on these splines, we stratified participants based on dichotomized SBP and DBP above or below the point of lowest risk for each component (determined to be 124 and 67 mm Hg, respectively). We stratified into 2 groups to preserve sample size within the strata.
Baseline characteristics of participants stratified by high (≥124 mm Hg) and low (<124 mm Hg) achieved SBP were summarized in Table 1 and compared using ANOVA or χ2 tests as appropriate.
We next fit a series of multivariable Cox proportional hazard models to evaluate the independent association of achieved SBP and DBP (per standard deviation [SD] higher BP) with outcomes; participants were censored at the time of event or last follow-up. Because the relationship of BP and outcomes appeared J-shaped, we stratified all models based on the cut points identified earlier. The first model included age, sex, ethnicity, region, smoking, alcohol use, body mass index, baseline SBP, baseline DBP, history of hypertension, diabetes mellitus, heart disease, and prior stroke/transient ischemic attack. The second model added number of medications at baseline, number of medications at 1-year follow-up, and randomization group. Finally, the third model added achieved SBP (linear and quadratic term) in the models with DBP as the primary predictor and vice versa in the models with SBP as the primary predictor.
We tested for effect modification in all models of the primary outcome of interest (all stroke) by including an interaction term between achieved SBP and DBP and age and randomization group, as well between achieved SBP and both achieved and baseline DBP, and vice versa.
After a mean follow-up of 3.7 (SD 2.0) years, participants in the higher target BP randomization group achieved a mean SBP of 137 (SD 9.2) mm Hg and DBP of 75.3 (SD 7.9) mm Hg compared with an SBP of 126 (SD 9.9) mm Hg and DBP of 69.4 (SD 7.5) mm Hg in the lower target group. The spline plots demonstrated a J-shaped association of achieved SBP and DBP with stroke (Figure 1). The lowest risk of stroke occurred at ≈124 mm Hg SBP and 67 mm Hg DBP; 27% and 26% of participants achieved BPs below these levels, respectively (Figure 2).
Participants with high achieved SBP had a different racial/ethnic distribution compared with those with lower achieved SBP. Additionally, those with high achieved SBP were more likely to be from the United States and have higher body mass index and baseline SBP and DBP (Table 1). Additionally, these participants had a higher prevalence of hypertension, diabetes mellitus, and ischemic heart disease and were more likely to be on all classes of antihypertensive medications compared with persons with low achieved SBP. Finally, participants with high achieved SBP were also more likely to be on more antihypertensive medications after 1 year, specifically calcium-channel blockers, β-blockers, and other non-thiazide, non–angiotensin-converting enzyme inhibitor/angiotensin II receptor blockers medications; not surprisingly, participants in this group were more likely to be in the higher target SBP randomization group and had fewer visits.
The associations between SBP and all stroke, ischemic stroke, major vascular events, and mortality differed above and below 124 mm Hg; P for all interaction <0.05. (Table 2) Above SBP of 124 mm Hg, higher SBP was associated with an increased risk of all stroke, ischemic stroke, major vascular events, and mortality (Table 2). In contrast, below SBP of 124 mm Hg, the associations were inverted; higher pressure was associated with lower risk. These patterns were consistent after multivariable adjustment, although the association between SBP and ischemic stroke was attenuated in the SBP ≥124 mm Hg group. The associations with ischemic stroke and major vascular events were further attenuated after the inclusion of DBP. The J-shaped association between SBP and mortality remained strong even after adjustment for DBP.
We observed a similar pattern of association between DBP and outcomes (Table 3). Above DBP of 67 mm Hg, higher DBP was associated with an increased risk of all stroke, ischemic stroke, and major vascular events. In contrast, below DBP of 67 mm Hg, the associations were significantly different and in the opposite direction. Higher DBP was associated with increased mortality above 67 mm Hg, although this association was attenuated after adjustment for SBP. The association between DBP and all stroke, ischemic stroke, and major vascular events both above and below 67 mm Hg were robust against adjustment for covariates and SBP.
There were no significant interactions between achieved SBP and DBP and age, randomization group, baseline SBP and DBP, and achieved SBP and DBP.
In this observational analysis of the SPS3 trial, there was a J-shaped association between achieved SBP and DBP and stroke, vascular events, and mortality. The lowest risk of events was at ≈120 to 128 mm Hg SBP and 65 to 70 mm Hg DBP; above these levels, higher SBP and DBP were associated with an increased risk of events. In contrast, below this level, we observed inverted associations; higher pressures were associated with a lower risk of events. Achieved DBP appeared to be more important compared with SBP for ischemic stroke and major vascular events; however, achieved SBP had a stronger association with all-cause mortality.
The present study fills a gap in knowledge regarding the association of achieved BP and outcomes; no previous investigation has studied a population with symptomatic lacunar infarcts. The majority of previous literature has reported a J-shaped association between BP and cardiovascular events and a linear association with stroke outcomes. In a post hoc analysis of 22 576 patients with hypertension and coronary artery disease enrolled in the International Verapamil-Trandoloapril Study (INVEST), the association between both SBP and DBP and a composite cardiovascular outcome was J-shaped, although the J-shape was not apparent for stroke outcomes alone.4 A meta-analysis from the Individual Data Analysis of Antihypertensive Intervention (INDANA) database reported a J-shaped association of both DBP and SBP with mortality; this association was only present for CVD mortality in the treatment group.13 In the Treating to New Targets (TNT) trial, which enrolled over 10 000 participants with a history of coronary artery disease, the mortality rate was the lowest at a BP of 146/81 mm Hg.14 A nonlinear J-shaped association was found for SBP and vascular events, but not stroke. In a pooled analysis of the European Carotid Surgery Trial, the North American Carotid Endarterectomy Trial, and the United Kingdom Transient Ischemic Attack Aspirin Trial, investigators found evidence of a J-shaped association of BP and recurrent stroke among participants with bilateral carotid stenosis, suggesting that participants with advanced carotid disease may be more susceptible to stroke risk at lower BP.15
An important strength of the present study is that the BP lowering was primarily because of medication, which allows us to distinguish the effect from observational studies in which low BP may be because of comorbid conditions, such as heart failure. However, the mechanisms mediating the associations of low BP with poor outcomes remain to be determined. Diastolic arteriolar tone seems to be an important determinant of cardiovascular risk. Some investigators have argued that low DBP may have a negative effect on coronary events because the heart is perfused during diastole; these effects may be accentuated in patients with preexisting coronary artery disease.4 Similarly, we theorize that low BP could result in inadequate perfusion of the brain in persons with preexisting small vessel disease, leading to stroke, thus, leading to a cerebral infarct. The strong association of both low and high SBP with mortality suggests that both extremes of BP are associated with severe events. Finally, there are challenges with accurate measurement of BP in older adults, including the presence of pseudohypertension, postprandial hypertension, and cuff artifact,16–18 which could have contributed to some overtreatment of older adults who have normal intra-arterial BP or low DBP.
Although this is the first study to investigate a J-shaped association of achieved BP and outcomes in the setting of intensive BP lowering for secondary prevention, this study also has limitations which must be considered. The primary limitation of the present study is that it is a post hoc analysis of a randomized trial. Although achieved BP was highly influenced by pharmacotherapy, there were other factors that impacted achieved SBP and DBP. We adjusted for a large number of potential confounders, but residual and unmeasured confounding may remain. Subclinical cardiovascular disease could affect both BP and risk of events. Second, the mechanism explaining a J-shaped association of BP and outcomes remains uncertain. Future studies designed to better understand the physiological effects of low BP may better describe the mediating pathways.
In summary, the lowest risk of events occurred at a nadir of ≈120 to 128 mm Hg systolic blood pressure and 65 to 70 mm Hg diastolic blood pressure. The present study highlights the need to evaluate the impact of excessive BP reduction in persons treated with antihypertensives. This may be more important in older patients with preexisting vascular disease or in those who are vulnerable to adverse effects of hypotension.
Sources of Funding
This research was supported by the National Institute on Aging (K01AG039387, R01AG46206). The SPS3 trial was supported by the National Institute of Neurological Disorders and Stroke (U01NS038529). The funding agencies had no role in the design of the study, analysis, interpretation of results, writing of the article, or approval of the article.
Current affiliation for L.A.M.: Department of Epidemiology and Biostatistics, Dornsife School of Public Health, Drexel University, Philadelphia, PA.
- Received September 10, 2015.
- Revision received September 21, 2015.
- Accepted October 13, 2015.
- © 2015 American Heart Association, Inc.
- Mancia G,
- Fagard R,
- Narkiewicz K,
- et al
- Franklin SS,
- Khan SA,
- Wong ND,
- Larson MG,
- Levy D.
- Benavente OR,
- White CL,
- Pearce L,
- et al
- Chobanian AV,
- Bakris GL,
- Black HR,
- Cushman WC,
- Green LA,
- Izzo JL Jr.,
- Jones DW,
- Materson BJ,
- Oparil S,
- Wright JT Jr.,
- Roccella EJ
- Pergola PE,
- White CL,
- Szychowski JM,
- Talbert R,
- Brutto OD,
- Castellanos M,
- Graves JW,
- Matamala G,
- Pretell EJ,
- Yee J,
- Rebello R,
- Zhang Y,
- Benavente OR
- Boutitie F,
- Gueyffier F,
- Pocock S,
- Fagard R,
- Boissel JP
- Bangalore S,
- Messerli FH,
- Wun CC,
- Zuckerman AL,
- DeMicco D,
- Kostis JB,
- LaRosa JC
- Rothwell PM,
- Howard SC,
- Spence JD
Novelty and Significance
What Is New?
In a trial of blood pressure control in participants with a history of lacunar stroke, there was a J-shaped association between blood pressure and stroke, vascular events, and mortality.
What Is Relevant?
The present study suggests the need to evaluate whether a lower bound for blood pressure reduction should be considered for prevention of cerebrovascular events.
In the Secondary Prevention of Small Subcortical Strokes (SPS3) trial of systolic blood pressure lowering in participants with a history of lacunar stroke, the lowest risk of events was at ≈120 to 128 mm Hg systolic blood pressure and 65 to 70 mm Hg diastolic blood pressure; above these levels, higher blood pressures were associated with an increased risk of events. In contrast, below this level, we observed inverted associations; higher pressures were associated with a lower risk of events.