Effect of Postural Hypotension on Recurrent Stroke: Secondary Prevention of Small Subcortical Strokes (SPS3) Study

Abstract

Background:

Orthostatic hypotension (OH) has been independently associated with increased risk of stroke and other cardiovascular events. We sought to investigate the relationship between OH at follow up and recurrent stroke risk in SPS3 (Secondary Prevention of Small Subcortical Strokes) trial patient cohort. This is a retrospective cohort analysis.

Methods:

We included all SPS3 trial participants with blood pressure measurements in both sitting and standing position per protocol at baseline, with at least 1 follow up visit to establish the relationship between OH at follow up and recurrent stroke risk (primary outcome). Secondary outcomes included major vascular events, myocardial infarction (MI), all-cause mortality, and, ischemic and hemorrhagic stroke subtypes. Participants were classified as having OH at baseline and at each follow up visit based on a systolic BP decline ≥ 20 mm Hg or a diastolic BP decline ≥ 10 mm Hg on position change from sitting to standing. We used Cox proportional hazards regression modeling to compare the risk of outcomes among those with and without OH.

Results:

A total of 2275 patients were included with a mean follow up time 3.2 years (standard deviation = 1.6 years). 39% (881/2275) had OH at some point during their follow up. Of these, 41% (366/881) had orthostatic symptoms accompanying the BP drop. In a fully adjusted model, those with OH had a 1.8 times higher risk of recurrent stroke than those without OH (95% CI: 1.1–3.0). The risk of ischemic stroke, major vascular events, and all-cause mortality was similarly elevated among the OH group.

Conclusion:

Orthostatic hypotension was associated with increased recurrent stroke risk, vascular events, and all-cause death in this large cohort of lacunar stroke patients. Whether minimizing OH in the management of post-stroke hypertension in patients with lacunar stroke reduces recurrent stroke risk deserves further study.

INTRODUCTION

The treatment of hypertension (HTN) is possibly the most important intervention for the secondary prevention of ischemic stroke.¹ One common side effect of anti-hypertensive medications is orthostatic hypotension (OH) or a substantial fall in blood pressure (BP) upon standing. This fall may be accompanied by symptoms or may be asymptomatic. Studies have reported that OH is associated with an increased risk of ischemic stroke in the general population² and also that orthostatic variation (both increase and decrease in BP on standing) was associated with an increased risk of incident lacunar infarcts.^{3, 4} Most studies on OH and stroke have reported on incident stroke events. Less is known about the association between OH and recurrent stroke. Since many stroke patients are on anti-hypertensive medications, it is useful and clinically relevant to understand if there is an association between OH and recurrent strokes and vascular events. This could have implications for choice of antihypertensive medications in stroke survivors with HTN since some antihypertensive medications are more likely to be associated with OH than others^{5, 6}

The Secondary Prevention of Small Subcortical Strokes (SPS3) study provides a unique opportunity to examine the relationship between OH and stroke recurrence. The SPS3 study examined the impact of two targets of SBP management on stroke recurrence in more than 3000 lacunar stroke survivors.⁷ In this manuscript, we examine the relationship between OH and risk of recurrent stroke in the SPS3 cohort. This cohort is very well-suited to answer this question since OH was measured at baseline and follow-up visits using a consistent, standard approach.

METHODS

Because of the sensitive nature of the data collected for this study, requests to access the dataset from qualified researchers trained in human subject confidentiality protocols may be sent to the SPS3 publication committee at whitec2@uthscsa.edu.

The SPS3 trial was a randomized, multi-center clinical trial conducted in 81 clinical centers in North America, Latin America and Spain. Study design details are described elsewhere⁸ and summarized here. Patients with recent (within 6 months) symptomatic lacunar infarcts proven by MRI were randomly assigned in a two-by-two factorial design to two target levels of systolic blood pressure (1:1; 130–149 mm Hg vs <130 mm Hg; open-label, PROBE design) and to a once-daily antiplatelet treatment (1:1; aspirin 325 mg plus clopidogrel 75 mg vs aspirin 325 mg plus placebo; double-blind). Patients with critical stenosis in cervical or intracranial vessels pertinent to the index event as well as those with stroke due to other etiologies such cardioembolism from atrial fibrillation were excluded in the SPS3 trial. The primary SPS3 outcome was recurrent stroke. Secondary outcomes were acute myocardial infarction (MI) and death which was classified as due to vascular, non-vascular or unknown cause.

The SPS3 (Secondary Prevention of Small Subcortical Strokes) study took place between March 23, 2003, and April 30, 2012 and a total of 3020 patients were enrolled in the study. Detailed medical history and demographic information was collected at baseline. Study participants were seen monthly for the first 3 months after randomization and until the target blood pressure was achieved at 2 consecutive visits. Thereafter, they were seen quarterly. At each follow up visit, BP was measured using a standard protocol (Appendix).^8–12 At each visit seated BP (3 readings) as well as standing BP was measured, and a standard questionnaire assessed adverse events including orthostatic symptoms and study outcomes (Appendix). The protocol used for BP measurement and questionnaires for assessment of orthostatic symptoms and other adverse events are included in the Appendix. BP measurements were done by SPS3 affiliated personnel with special training. These measurements were carried out with the Colin electronic sphygmomanometer. In the event of readings which were unexpectedly high or low, a re-check with a recently calibrated mercury manometer was part of the protocol. Orthostatic (standing) measurements obtained after sitting BPs are measured.

The SPS3 study was approved by the institutional review boards of all participating centers, and all patients provided written informed consent. De-identified data from the study was further analyzed by authors approved for data use.

Study design and outcomes.

This is a retrospective, cohort study using the cohort of SPS3 trial participants. The primary outcome of interest is recurrent stroke. We also examined secondary outcomes, namely, ischemic and hemorrhagic strokes, MI, vascular events (MI, stroke, cardiovascular death) and all-cause mortality. Definitions of outcome events are available in the online supplement of the primary study manuscript.¹³ Our exposure of interest was OH at a follow up visit. The reason we used OH at a follow up visit, (rather than baseline OH) was that the baseline BP was sometimes measured soon after stroke, typically, within 3–4 weeks post-stroke. BP are often times elevated in the immediate aftermath after stroke due to cerebral dysregulation.¹⁴

We included all SPS3 participants who had BP measured both sitting and standing by protocol at baseline and also at least 1 follow up visit. Participants were classified as having OH at baseline and at each follow up visit based on either (1) a decline in SBP of at least 20 mm Hg, or (2) a decline in DBP of at least 10 mm Hg with posture change from sitting to standing.

We further examined the sensitivity of our results using a more liberal definition of OH which included one of the BP criteria above or the presence orthostatic symptoms on posture change from sitting to standing. These symptoms included one or more of the following symptoms upon standing: orthostatic dizziness, lightheadedness, unsteadiness, blurry vision, and, palpitations.

Analysis.

We compared baseline characteristics between those with OH at 1 or more follow up vs. those with no OH during follow up using the t-test or the Chi-Square test of association, as appropriate. Cox proportional hazards models were used to compare the risk of events (any stroke, ischemic stroke, hemorrhagic stroke, major vascular events, MI and death) between those with OH and those without. We included OH as a time-varying covariate in order to account for the strong relationship between OH and the number of follow-ups. Models were fit (1) univariately, (2) with the addition of age, sex, race and region of the study, (3) with the further addition of smoking, alcohol use, baseline SBP and baseline OH, and (4) with the addition of diabetes, HTN at baseline, the number of anti-hypertensive medications at baseline, angina, hyperlipidemia, and the BP target group (i.e. intense vs. less intense control). We adjusted for baseline OH in model 3 in order to understand whether the baseline OH had an effect on the relationship between the exposure variable and recurrent stroke risk. Model 4 tested the interaction between OH and diabetes. We also examined the interaction between OH and target BP group of the trial in Model 4. All analyses assessed the definition of OH using BP criteria alone, as well as with the definition of OH using BP criteria and symptom criteria. Finally we computed the E-value for the primary outcome to assess for the extent of unmeasured confounding.¹⁵

RESULTS

The SPS3 enrolled a total of 3020 participants. Of these, 660 participants were enrolled before the study required the collection of standing BP and an additional 85 participants did not have orthostatic BP collected at any follow up visit. After, these participants were excluded, a total of 2275 participants were included in this analysis. Mean participant follow up time was 3.2 years (SD=1.6 years).

Primary Analysis

Of the 2275 participants, 39% (881/2275) had OH at some point during their follow up. Of these, 41% (366/881) had orthostatic symptoms accompanying the BP drop. An additional 357 patients reported orthostatic symptoms without meeting the BP criteria and are not classified as having OH in the primary analysis.

Table 1 presents the baseline characteristics of participants overall and by “ever vs. never” OH during follow up visits. Orthostatic participants were more often women: 41% women in the OH group vs. 33% in the non-OH group, (p=0.0002). They were more likely to be of the white race 54% white in the OH group vs. 46% in the non-OH group, (p=0.0002). They were also more likely to be from the U.S. (Table 1). Participants in the OH group had a higher baseline prevalence of hyperlipidemia and a history of angina. Those who were orthostatic at follow up were more likely to have had OH at baseline and were also in the higher target BP group (i.e. less intense management with BP goal of 130–149 mm Hg). Participants who were orthostatic at one or more follow up visits had significantly more follow up visits.

TABLE 1.

Baseline Characteristics of Participants According to Categories of Orthostatic Blood Pressure. Orthostatic Hypotension (OH) Defined by Drop of SBP of 20 mm Hg or DBP of 10 mm Hg on Changing from Sitting to Standing Position During 1 or More Follow Up Visits.

	Overall (n=2275)	No OH (n=1394)	OH (n =881)	p-value
Demographics
Age, years (mean, SD)	63.5 (10.8)	63.4 (10.8)	63.6 (10.8)	0.70
Male	1445 (64%)	927 (67%)	518 (59%)	0.0002
Race				0.0002
White	1118 (49%)	640 (46%)	478 (54%)
Black	313 (14%)	206 (15%)	107 (12%)
Hispanic	794 (35%)	523 (38%)	271 (31%)
Other/mixed	50 (2%)	25 (2%)	25 (3%)
Region				<0.0001
US	1011 (44%)	570 (41%)	441 (50%)
Canada	232 (10%)	136 (10%)	96 (11%)
Latin America	672 (30%)	455 (33%)	217 (25%)
Spain	360 (16%)	233 (17%)	127 (14%)
Health Behaviors
Smoking				0.59
Current	455 (20%)	273 (20%)	182 (21%)
Past	892 (39%)	541 (39%)	351 (40%)
Never	928 (41%)	580 (42%)	348 (40%)
Regular Alcohol Use	305 (13%)	174 (12%)	131 (15%)	0.11
Physiologic Measures
Baseline SBP (mean, SD)	142 (19)	142 (18)	142 (19)	0.55
Baseline DBP (mean, SD)	78 (11)	78 (10)	78 (11)	0.47
Body Mass Index (mean, SD)	29.0 (6.9)	28.8 (5.6)	29.3 (8.6)	0.13
Baseline OH	192 (8%)	85 (6%)	107 (12%)	<0.0001
Baseline Orthostatic symptoms	178 (8%)	98 (7%)	80 (9%)	0.078
Health History
Diabetes Mellitus Type 2	834 (37%)	489 (35%)	345 (39%)	0.049
Hypertension History	2031 (89%)	1238 (89%)	793 (90%)	0.36
MI	98 (4%)	41 (4%)	57 (5%)	0.45
Angina	111 (5%)	39 (4%)	72 (6%)	0.024
CHF	16 (0.70%)	5 (0.5%)	11 (0.9%)	0.25
CABG/PTCA/Stent	92 (4%)	43 (4%)	48 (4%)	0.74
COPD	55 (2%)	25 (2%)	30 (2%)	0.99
Hyperlipidemia	1060 (47%)	443 (43%)	517 (50%)	0.0007
Endarterectomy, etc	16 (0.70%)	5 (0.5%)	11 (0.9%)	0.25
Pacemaker	1 (0.04%)	0 (0%)	1 (0.04%)	>0.99
Peripheral vascular disease	72 (3%)	26 (3%)	46 (4%)	0.12
Sleep apnea	80 (4%)	37 (4%)	43 (4%)	0.90
Medication Use
Mean hypertension medications at baseline (SD)	1.6 (1.2)	1.6 (1.5)	1.7 (1.6)	0.053
Statin use baseline	1602 (70%)	976 (70%)	626 (71%)	0.60
Mean hypertension medications at follow-up (SD)	2.0 (1.3)	2.0 (1.3)	2.1 (1.3)	0.062
Study Parameters
Higher BP target	1146 (50%)	669 (48%)	477 (54%)	0.0042
Number of follow-up visits	11.6 (6.4)	10.4 (6.4)	13.6 (5.9)	<0.0001
Loss to follow-up	44 (1.9%)	31 (2%)	13 (1.5%)	0.21

The unadjusted event rates are significantly higher in the non-OH group (Table 2). The incremental models (Table 3) found that OH was significantly associated with all outcome events with the exception of hemorrhagic stroke and MI, even after multivariable adjustment. In a fully adjusted model, those with OH had a 1.8 times higher risk of recurrent stroke than those without OH (95% CI: 1.1–3.0). The risk of ischemic stroke, major vascular events, and all-cause mortality was similarly elevated among the OH group. There was no significant difference in the relationship between OH and outcomes in diabetics vs. non-diabetic s (Table 4). Similarly, the interaction between OH and target BP group of the trial in fully adjusted models was non-significant (Table 5).

TABLE 2.

Unadjusted Event Rates (95% CI) According to Categories of Orthostatic Blood Pressure. Orthostatic Hypotension (OH) Defined by Drop of SBP of 20 mm Hg or DBP of 10 mm Hg on Changing from Sitting to Standing Position During 1 or More Follow Up Visits.

	No OH		OH
Outcomes	Events N (Mean Follow up, years)	Rate Percent / person-year	Events N (Mean Follow up, years)	Rate Percent / person-year
All stroke	101 (2.9)	2.5 (2.0, 3.0)	48 (3.7)	1.5 (1.1, 1.9)
Ischemic Stroke	86 (2.9)	2.1 (1.7, 2.6)	41 (3.7)	1.2 (0.89, 1.7)
Hemorrhagic Stroke	15 (2.9)	0.37 (0.21, 0.58)	7 (3.7)	0.21 (0.085, 0.40)
Major Vascular Events	121 (2.9)	3.0 (2.5, 3.5)	61 (3.7)	1.9 (1.4, 3.4)
Myocardial infarction	22 (3.0)	0.52 (0.32, 0.76)	14 (3.8)	0.42 (0.23, 0.67)
All-Cause Death	75 (3.1)	1.8 (1.4, 2.2)	39 (3.8)	1.2 (0.82, 1.6)

TABLE 3:

Hazard Ratios (95% CI) of Events by Exposure Defined as Orthostatic Hypotension (OH) Defined by Drop of SBP of 20 mm Hg or DBP of 10 mm Hg on Changing from Sitting to Standing Position During 1 or More Follow Up Visits. (Reference is no OH).

Outcomes	Model 1*	Model 2 †	Model 3 ‡	Model 4§
All stroke	2.1 (1.3, 3.3)	2.0 (1.3, 3.3)	2.0 (1.2, 3.2)	1.8 (1.1, 3.0)
Ischemic Stroke	2.1 (1.2, 3.4)	2.0 (1.2, 3.3)	1.9 (1.1, 3.2)	1.8 (1.0, 3.0)
Hemorrhagic Stroke	2.2 (0.64, 7.3)	2.6 (0.75, 8.8)	2.5 (0.71, 8.5)	2.2 (0.63, 7.6)
Major Vascular Events	2.1 (1.4, 3.3)	2.1 (1.4, 3.3)	2.1 (1.3, 3.2)	1.9 (1.2, 2.9)
Myocardial Infarction**	2.2 (0.84, 5.6)	2.1 (0.80, 5.4)	2.0 (0.78, 5.4)	1.9 (0.73, 5.1)
All-Cause Death	2.3 (1.4, 3.8)	2.4 (1.4, 4.0)	2.0 (1.2, 3.4)	1.9 (1.1, 3.2)

^*Model 1: Univariate

^†Model 2: +Age, race, gender, region

^‡Model 3: +Smoking, alcohol use, BL SBP, BL orthostatic hypotension

^§Model 4: +Diabetes Mellitus, History of hypertension, Baseline number of hypertension medications, Follow up number of hypertension medications, hyperlipidemia at baseline, baseline angina, Target BP group

^**Myocardial Infarction models omit race due to the lack of events among some race groups

TABLE 4.

Interaction between Orthostatic Hypotension and Diabetes Mellitus in Fully Adjusted Model (Model 4 from Table 3).

Outcomes	p for Interaction
All stroke	0.96
Ischemic Stroke	0.78
Hemorrhagic Stroke	0.56
Major Vascular Events	0.28
Myocardial Infarction	0.99
All-Cause Death	0.79

TABLE 5.

Interaction between Orthostatic Hypotension and target blood pressure group of the trial in Fully Adjusted Model (Model 4 from Table 3)

Outcomes	p for Interaction
All stroke	0.71
Ischemic Stroke	0.54
Hemorrhagic Stroke	0.56
Major Vascular Events	0.96
Myocardial Infarction	0.20
All-Cause Death	0.75

Unmeasured Confounding

The E-value for the hazard ratio for primary outcome (all stroke) and it’s 95% confidence interval was: 3.1 (1.4, 5.5). This suggests that the observed hazard ratio of 1.8 could be attenuated by an unmeasured confounder that was associated with both the orthostatic hypertension and stroke by a relative risk of 3.1 each, beyond the confounders included in our model.

Sensitivity Analysis

When the OH definition was expanded to include either the BP criteria or the presence of orthostatic symptoms, 54% (1238/2275) were defined as having OH at some point during their follow up. Multivariable model results were similar to the more specific OH definition, although with the exception of all-cause mortality, the hazard ratios were no longer statistically significant between the two groups (Online Tables 1–3). The risk of death for those with OH was 2.2 times (95% CI: 1.5–3.4) higher than for those without OH. There was no significant interaction between OH and DM (Online Table 4) for any of the outcomes, indicating that the association between OH and outcomes did not differ in those with and without DM.

DISCUSSION

We studied the association between OH and vascular outcomes in this large, well-defined cohort of lacunar stroke survivors from the SPS3 trial and found that OH was associated with an increased risk of recurrent stroke, ischemic stroke, major vascular events, and all cause death. This association was not observed for hemorrhagic strokes and MI.

To date, this report is the first comprehensive examination of orthostatic BP drop and recurrent stroke and vascular events. Prior reports have largely examined incident events and have shown an association between OH and incident ischemic stroke³ and also silent cerebrovascular infarcts⁴. Our findings on the association between OH and recurrent ischemic stroke, major vascular events, and all-cause mortality expand what is known on this topic. The issue of recurrent strokes is particularly relevant since the majority of incident stroke survivors are hypertensive and are typically on antihypertensive medications or are started on them after the stroke. We also note that the majority of participants with an orthostatic BP drop did not report orthostatic symptoms (59%).

Physiological adjustments to orthostatic stress have been studied previously.^{5, 16, 17} Positional change from sitting or supine to standing leads to pooling of blood to the lower extremities, and, in the splanchnic and pulmonary circulations.^{16, 17} As a result, there is a decrease in venous return and ventricular filling. This causes a transient reduction in stroke volume, a slight fall in SBP and a rise in DBP. This can activate high-pressure baroreceptors in the carotid sinus and aortic arch and low-pressure receptors in the heart and lungs, and can increase sympathetic outflow and decrease parasympathetic tone.^18–20 These changes lead to an increase in heart rate and peripheral vascular resistance. Increase in sympathetic activity caused by orthostatic stress can stimulate the Renin-angiotensin-aldosterone system which contributes to vasoconstriction.^{5, 16, 17} Abnormalities disrupting this chain of responses can lead to OH.

In our sample with lacunar infarcts, the association of OH with recurrent stroke could be due to several mechanisms. Physiologically, cerebral small vessel disease is thought to occur due to arteriolar hyalinization of smaller perforating blood vessels in the brain.^{21, 22} Patients who already have these vascular changes in their small perforating brain vessels, are likely to have long standing hypertension and arteriosclerosis.²³ In such condition, there is decreased arterial wall compliance leading to a reduced baroreceptor responsiveness or subsequent pooling of blood in the peripheral vascular system which can lead to exaggerated brain hypoperfusion due to orthostatic stress.^{2, 24} Procoagulant activation of cerebral microvascular endothelium has also been described in lacunar stroke which can be exaggerated with OH.^{25, 26} Diabetes, a well-established risk factor for stroke,²⁷ may also lead to OH through autonomic dysfunction and associated hypertension as well as cardiac autonomic neuropathy.²⁸ We did not find any significant differences in relationship between OH and the risk of events by diabetes status. A different possibility is that patients with OH (especially symptomatic OH) may have had lower medication adherence due to side effects and hence higher stroke recurrence. Unfortunately, we did not collect adherence to BP medications systematically. The study did collect adherence to antiplatelet therapies and the adherence to these was 94%. Nevertheless, we are unable to answer whether a worse adherence to anti-hypertensive medications in patients with symptomatic OH played a role in the worse observed outcomes.¹³

We note that the hazard ratios in Cox models using the more liberal definition of OH (BP drop + symptoms) were attenuated for all stroke, ischemic stroke and major vascular events when compared to models using the strict definition of OH (BP drop only). This may suggest that the presence of BP drop was a key marker of the sympathetic dysregulation that led to worse outcomes. This did not hold for all-cause mortality. The HR for all outcomes with both definitions of OH were however consistent and comparable.

OH is strongly associated with risk of incident cardiovascular complications, including myocardial infarction, heart failure, stroke, and all-cause mortality.^{2, 5, 16} However, OH was not associated with MI in our study. This could reflect smaller number of MI events related to more aggressive management of vascular risk factors in study group compared to patients with OH who never had stroke. Concurrent use of multiple antihypertensive agents is also linked with OH, however, impact of individual drug class on OH is unclear with inconsistency in findings between different studies.^{5, 6, 29} In our study sample, though not statistically significant (p=0.053), higher number of anti-hypertensives were used in OH group.

The strengths of our study are the large number of well-defined lacunar strokes (using clinical and neuroimaging criteria), the standardized measurement of BP, and the standardized assessment of outcomes. We acknowledge that this is a retrospective cohort. The selection bias inherent to clinical trials (i.e. healthier patients are more likely to participate) applies, and furthermore, the trial was not designed to address the study question specifically. One consequence of this is that the available data does not allow analysis of outcomes by the presence or absence of heart rate increase as an additional aspect of the orthostatic response. We also cannot comment on any differences in outcomes between non-neurogenic and traditional neurogenic OH such as those caused by pure autonomic failure or multisystem atrophy in our analyses as patients with these significant illnesses were excluded from the SPS3 study. We acknowledge that there is a potential for unmeasured confounding, as is true with any analysis examining observational data. In order to assess the extent of unmeasured confounders, we computed the E-value for the hazard ratio for primary outcome (all stroke) and it’s 95% confidence interval: 3.1 (1.4, 5.5). This suggests that the observed hazard ratio of 1.8 could be attenuated by an unmeasured confounder that was associated with both the orthostatic hypertension and stroke by a relative risk of 3.1 each, beyond the confounders included in our model. Thus, results should be interpreted in this context.

Future work should address whether certain classes of anti-hypertensive medications were associated with OH and the impact of specific medication classes on the association between OH and outcomes, which could have important implications for clinical practice. We are unable to answer this question since medications were adjusted at each visit to achieve assigned target. Anti-hypertensive drugs have differential effects on BP variability. We acknowledge that this study did not examine the extent of BP variability and whether BP variability modified the relationship between OH and vascular outcomes in this cohort. Another limitation is that the number of follow-up visits were significantly higher in the OH group. Consequently, this could lead to ascertainment of more outcomes in the OH group. In order to mitigate this, we used OH as a time varying covariate in our analysis. We believe that this contributes to the different results between the rates vs. Cox models. Nevertheless, we do not think that outcomes were necessarily missed in the non-OH group. The SPS3 trial was a carefully monitored trial and there was close supervision of study sites. The outcomes were ascertained systematically and were hard outcomes, i.e. stroke, MI, mortality that were centrally adjudicated by a committee external to the study team. Furthermore, the loss to follow up was 2% and the loss to follow up was similar between the two groups (Table 1).

While SPS3 measured OH from a seated to standing position, it is sometimes measured by comparing supine vs. standing BP in a hospital (in-patient) setting. The latest ACC hypertension guidelines define OH as measured from a seated vs. standing position.³⁰ The SPS3 was a multi-center international trial. OH was measured from a seated to upright position to achieve consistency in measurement across clinics. A number of clinics were not equipped with an examination table to have the patient supine so rather than having OH measured in different ways across clinic, the measurement was standardized from seated to upright for all clinics. This measurement of OH as from a seated to standing position was therefore more specific (and less sensitive) and also identified the more severe cases of OH. We considered practice implications of our research. Since this is a retrospective study based on secondary analysis of clinical trial data, we do not know if evaluation and management of OH will attenuate the observed increased risk of cardiovascular events in stroke survivors with OH. We infer that our results provide Level 3 evidence (retrospective cohort) that OH maybe a prognostic factor in the outcomes of patients with lacunar infarcts. A similar association is found in the general population.^2–4

CONCLUSION

Orthostatic hypotension was associated with increased recurrent stroke risk, vascular events, and all-cause death in this large cohort of lacunar stroke patients in the SPS3 study, (Level 3 evidence). Our results raise the question as to whether clinical practice should routinely measure orthostatic BP during outpatient visits in stroke survivors since many participants with OH did not report symptoms. Furthermore, one should consider whether future clinical trials on HTN should not only treat BP to target but also treat to avoid both asymptomatic and symptomatic OH.