Blood Pressure Reduction and Outcome after Endovascular Therapy: A Secondary Analysis of the BEST Study

Abstract

Background:

Elevated systolic blood pressure (SBP) in the acute phase after endovascular therapy (EVT) is associated with worse outcome. However, the association between systolic blood pressure reduction (SBPr) and the outcome of EVT is not well understood.

Methods:

This is a post-hoc analysis of the Blood Pressure after Endovascular Stroke Therapy (BEST) prospective observational cohort study. SBPr was defined as the absolute difference between admission SBP and mean SBP in the first 24 hours post EVT. Logistic regression was used to assess the association between SBPr and poor functional outcome (modified Rankin Scale score 3–6) at 90 days.

Results:

A total of 259/433 (59%) patients had poor outcome. SBPr was higher in the poor outcome group compared to the good outcome group (26.6 ± 27.4 vs. 19 ±22.3 mm Hg; p<0.001). However, in adjusted models, SBPr was not independently associated with poor outcome (OR=1.00 per 1 mm Hg increase, 95% CI, 0.99–1.01) or death (OR, 0.9 per 1 mm Hg increase; 95% CI 0.98–1.00). There remained no association when SBPr was divided into tertiles. Subgroup analyses based on history of hypertension, revascularization status and antihypertensive treatment yielded similar results.

Conclusion:

The reduction in baseline SBP following EVT was not associated with poor functional outcomes. Most of our cohort (88%) achieved successful recanalization, therefore, our results mainly apply for patients with successful recanalization.

Introduction

Hypertension is common among patients presenting with ischemic stroke and has been attributed to multiple physiological and psychological factors¹. Several studies have demonstrated a link between sustained hypertension after stroke and poor outcome, which led to randomized trials evaluating the benefit from blood pressure (BP) control in acute ischemic stroke patients^2–4. Although varying considerably in design, these trials failed to show benefit from BP control in the days after stroke^5–8.

In the endovascular treatment (EVT) era, BP has gained increasing attention due to its potential role in reperfusion injury after EVT⁹. Studies from the pre-thrombectomy era suggested a potential detrimental effect of BP reduction on the outcome of acute ischemic stroke^10–12. However, these studies included heterogenous group of patients regardless of the presence of large vessel occlusion or reperfusion status. More recently, observational studies of patients treated with EVT have demonstrated a strong correlation between elevated BP after EVT and poor outcome^2–6. However, whether SBP reduction following EVT is associated with functional outcome is still not well known. In contrast to previous reports, a recent study of patients with successful revascularization following EVT demonstrated that patients with higher systolic blood pressure reduction (SBPr) had better functional outcomes¹³. Given the biases inherent to the retrospective design, we undertook the current study to determine the association between SBPr and clinical outcomes in a prospective multicenter cohort.

Methods

Patient population:

This is a post hoc analysis of the BEST (Blood Pressure after Endovascular Therapy for Ischemic Stroke) study. The BEST study protocol and detailed methods were reported previously¹⁴. In brief, BEST was a prospective, multicenter study that enrolled patients at 12 comprehensive stroke centers across the United States from November 2017 to September 2018. BEST was approved by local Institutional Review boards at 11 sites and deemed exempt at one site. Informed patient consent was waived. BEST enrolled patients treated with EVT for acute ischemic stroke due occlusion of the internal carotid artery, or the M1 or M2 segments of the middle cerebral artery. BEST excluded patients with (1) a known prestroke mRS >2, (2) terminal medical diagnoses such as a stage IV cancer, (3) left ventricular assist device, or (4) stroke in the perioperative or inpatient setting.

Study variables and outcome measures:

BP was recorded on admission and at least at hourly in the first 24 hours post EVT. Invasive BP measurements were preferably recorded over non-invasive cuff measurements. We calculated SBPr using two formulas. SBPrmean which was defined as the difference between admission SBP and mean SBP in the first 24 hours; SBPrmin which was defined as difference between admission SBP and minimum SBP in the first 24 hour. BP management differed between centers and between operators in each center. During the period of enrolment, none of the centres had a protocol to specifically guide post-EVT BP management, rather it was managed on a case-by-case basis.

In addition to BP and demographic variables, all patients had the following variables recorded: National Institutes of Health Stroke Scale (NIHSS) at presentation, the Alberta Stroke Program Early CT Score (ASPECTS), time from last known well to revascularization or end of procedure, modified Thrombolysis in Cerebral Ischemia (mTICI) score, and symptomatic (defined as that associated with a ≥4 points increase in the NIHSS from baseline) and asymptomatic intracerebral hemorrhage (ICH).

The primary outcome of this study was 90-day poor functional outcome, defined as mRS-score (modified Rankin Scale score) of 3–6. Secondary outcomes were 90-day mortality (mRS-score of 6), early neurological recovery (defined as NIHSS 0–1 or >7 points decrease in NIHSS-score from baseline to 24 hours) and hemorrhagic complications. Hemorrhagic complications included any intracerebral hemorrhage (ICH) and symptomatic ICH (any ICH associated with ≥4 points increase in NIHSS from baseline to 24 hours). Primary and secondary outcomes data was collected via telephone call or during in-person follow-up. In this analysis, we excluded patients with no functional outcome data or missing baseline covariate data.

Statistical analysis:

Data are presented as proportions for categorical variables, mean with SD for normally distributed continuous variables, and median with interquartile range for non-normally distributed continuous variables. Baseline characteristics were compared using Student’s t-test, Chi-square test and Wilcoxon Rank Sum test for continuous variables, binary variables and ordinal variables respectively. The association between SBPr measures and clinical outcomes was assessed using univariable and multivariable binary logistic regression analyses. Both adjusted and unadjusted analyses are reported. The following pre-specified predictors were included in the adjusted model: age, mTICI score, ASPECTS, time to revascularization, IV tPA, gender, admission NIHSS-score, baseline glucose, and anti-hypertensive treatment during first 24 hours post EVT. We first estimated the association between SBPr measures as a continuous variable and clinical outcomes. Then we performed a sensitivity analysis, comparing outcomes across SBPr tertiles. We performed subgroup analyses based on revascularization status, history of hypertension, treatment with antihypertensive, and admission SBP.

Results:

A total of 443 patients were included in this study, of whom 259 (58.4%) patients had poor outcome. Mean age of the study cohort was 68 years, 49% were female, 56.4% were white and 74.3% had chronic hypertension. Successful revascularization (mTICI 2b-3) was achieved in 88.5% of patients. Table 1 and 2 depicts the difference in baseline characteristics between good and poor outcome groups. Compared to the good outcome group, patients with poor outcome were older (71.7± 15 vs. 62.8±13.7 years; p<0.001), had higher median NIHSS-score (17 vs. 13; p<0.01), higher mean baseline glucose on admission (147 vs. 127 mg/dl; p<0.001) and lower median ASPECTS (8 vs. 9; p<0.001). In addition, history of hypertension, atrial fibrillation and tobacco use were more frequent in the poor outcome group. Mean and SD of SBPr are reported in Table I in Supplemental Material.

Table 1:

Baseline demographics of the entire cohort, and compared between patients with good (mRS 0–2) versus poor outcome (mRS 3–6).^*

Variable	Entire cohort (n=443)	Good outcome (n=184)	Poor outcome (N=259)	p value
Age (mean ± SD)	68.0±15.1	62.8±13.7	71.7±15.0	<0.001
Male sex (n, %)	217, 49.0	99, 53.8	118, 45.6	0.087
Caucasian (n, %)	250, 56.4	116, 63.0	134, 51.7	0.018
NIHSS (median, IQR)	16, 11–20	13, 8–19	17, 13–21	<0.001
ASPECTS (median, IQR)	8, 7–9	9, 8–10	8, 7–9	<0.001
Baseline glucose level (mg/dL, mean ± SD)	138±56	127±45	147±62	<0.001
Baseline platelet count (mean ± SD)	229±75	230±79	228±72	0.790
History of hypertension (n, %)	329, 74.3	121, 65.8	208, 80.2	0.001
History of DM (n, %)	123, 27.8	43, 23.4	80, 30.9	0.082
History of stroke (n, %)	83, 18.7	28, 15.2	55, 21.2	0.110
History of AFIB (n, %)	159, 35.9	54, 29.4	105, 40.5	0.016
Tobacco use (n, %)	102, 23.0	55, 29.9	47, 18.2	0.004
Antiplatelet use (n, %)	152, 34.3	67, 36.4	85, 32.8	0.432
Anticoagulant use (n, %)	77, 17.4	24, 13.0	53, 20.5	0.042
Isolated M1 MCA occlusion (n, %)	219, 49.4	100, 54.4	119, 46.0	0.081
tPA administered (n, %)	212, 48.6	103, 56.0	109, 43.3	0.009
General anesthesia (n, %)	164, 37.6	56, 30.4	108, 42.9	0.019
Minutes from stroke onset to recanalization (mean ± SD)	432±418	349±271	492±490	<0.001
Recanalization (TICI 2b-3) (n, %)	392, 88.5	174, 94.6	218, 84.2	0.001
Symptomatic ICH (n, %)	18, 4.1	0, 0	18, 7.0	0.001

^*Difference between continuous variables tested with Student’s t-test, binary variables with the chi-squared test, and for ordinal variables with the Wilcoxon Rank Sum test.

Table 2.

Difference in blood pressure parameters between good outcome and poor outcome groups

Variable	Entire cohort (n=443)	Good outcome (n=184)	Poor outcome (N=259)	p value
Number of blood pressure measurements (mean ± SD)	37.1±12.0	39.1±12.5	35.7±11.5	0.003
Continuous infusion of antihypertensive medication in first 24 hours (n, %)	165, 37.3	43, 23.4	122, 47.1	<0.001
>2 single doses of an antihypertensive in first 24 hours (n, %)	55, 12.4	22, 12.0	33, 12.7	0.805
Continuous infusion of vasopressor medication in first 24 hours (n, %)	73, 16.5	30, 16.3	43, 16.6	0.934
SBPr (mean, +-SD)	23.7 (+/−25.5)	19 (22.3)	26.6 (27.4)	<0.001
SBPrmin (mean, +-SD)	53.3 (28.1	46.9 (24.6)	57.8 (29.5)	<0.001

SBPrmean and clinical outcomes

The distributions of SBPrmean is shown in Figure I in Supplemental Material. SBPrmean was higher in the poor outcome group compared to the good outcome group (26.6 ±27.4 vs. 19.0 ±22.3 mm Hg; p <0.001). Correspondingly, more patients in the poor outcome group received continuous antihypertensive treatment than the good outcome group (47.1% vs. 23.4%, p<0.001), but there was no difference in the use of as-needed antihypertensive treatment between the two groups.

In unadjusted analysis, there was an association between higher SBPrmean and poor outcome (OR, 1.0; 95% CI 1.00–1.01; p<0.005) (Table 3). In addition, the highest tertile group was more likely to have poor outcome compared to lowest tertile (OR 1.9; 1.19–3.08; p=0.007). This association did not retain statistical significance after adjusting for confounders (Table 3). Similarly, there was no association between SBPrmean and 90-day mortality or early neurologic recovery (Table 3). Subgroup analyses for SBPrmean based on admission SBP, use of antihypertensive treatment, revascularization status and history of hypertension revealed similar results (Figure 1).

Table 3.

Unadjusted and adjusted logistic regression for association of SBPrmean with the outcomes of mechanical thrombectomy

	SBPrmean in tertiles			SBPrmean as continuous variable
	Tertile 1	Tertile 2	Tertile 3	ORc	P valued
Poor outcome
OR (95%CI)a	1 (ref)	1 (0.63–1.57)	1.9 (1.19–3.08)	1.0 (1.00–1.01)	0.005
P value	1 (ref)	1	0.007
Adjusted OR (95%CI)b	1 (ref)	1.10 (0.63–1.91)	1.30 (0.73–2.31)	1.0 (0.99–1.01)	0.27
P value	1 (ref)	0.71	0.36
Death
OR (95%CI)a	1 (ref)	0.64 (0.33–1.24)	1.46 (0.83–2.58)	1.0 (0.99–1.01)	0.25
P value	1 (ref)	0.19	0.18
Adjusted OR (95%CI)b	1 (ref)	0.61 (0.29–1.27)	0.87 (0.44–1.72)	0.9 (0.98–1.00)	0.50
P value	1 (ref)	0.19	0.70
Early neurologic recovery
OR (95%CI)a	1 (ref)	1.08 (0.68–1.72)	0.98 (0.61–1.56)	0.99 (0.99–1.00)	0.66
P value	1 (ref)	0.68	0.94
Adjusted OR (95%CI)b	1 (ref)	1.19 (0.70–2.02)	1.22 (0.71–2.09)	1.00 (0.99–1.00)	0.86
P value	1 (ref)	0.49	0.45

^aunadjusted analysis;

^badjusted for age, TICI score, ASPECTS, time to recanalization, IV tPA, gender, NIHSS, baseline glucose, antihypertensive;

^cOR calculated per 1 mmHg increase in SBPr.

^dp-values calculated using a logistic regression model with SBPr as continuous variable

Figure 1:

Subanalysis according to admission systolic blood pressure, antihypertensive treatment, recanalization status and history of hypertension. Squares represent odds ratio and error bars represent 95% confidence interval. Outcome is 90-day modified Rankin Scale 0–2

Abbreviation: HTN, hypertension; mTICI, modified treatment in cerebral ischemia; SBP, systolic blood pressure.

ICH occurred in 110/443 (24.83%). In logistic regression model fit to the outcome of ICH, adjusted for ASPECTS, TICI, and tPA, we did not find that SBPrmean was associated with ICH (OR 1; 95% CI 0.99–1.01; p=0.750). Symptomatic ICH (sICH) occurred in 18/443 (4.1%) and was also not associated with sICH (OR 1.00, 95% CI 0.98–1.02, p=0.752).

SBPrmin and clinical outcomes

SBPrmin distribution is shown in Figure II in Supplemental Material, respectively. SBPrmin was higher in the poor outcome group compared to the good outcome group (57.8 +/− 29.5 vs. 46.9+/− 24.6 mm Hg; p<0.001).

In addition, SBPrmin was associated with poor outcome in unadjusted but not in adjusted analysis (Table II in Supplemental Material). Furthermore, patients in the highest tertile were more likely to have poor outcome and die than the lowest tertile group. However, this association did not retain statistical significance in the fully adjusted model (Table II in Supplemental Material).

Discussion:

In this prospective multicenter study of patients treated with EVT, SBPr in the first 24 hours was not associated with poor functional outcome after adjusting for potential confounders.

Despite considerable evidence suggesting an association between elevated SBP and poor outcome in acute ischemic stroke^14–19, the literature regarding the association between SBP reduction and functional outcome is less consistent. In a prospective study of consecutive patients presenting with acute ischemic stroke, SBP drop >20 mm Hg in the first 24 hours was associated with poor long-term functional outcome and an early neurological deterioration¹¹. These findings were reproduced in another cohort of 115 patients with acute ischemic stroke¹⁰. Interestingly, the association between BP reduction and outcome was independent of antihypertensive treatment suggesting that even spontaneous BP reduction after acute ischemic stroke could be detrimental. In contrast to previous reports, another study suggested that BP reduction was associated with early neurological improvement after intravenous thrombolysis². Notably these studies predated the widespread use of EVT and did not account for the presence of large vessel occlusion or revascularization status, therefore their results cannot be applied to EVT-treated patients.

Data regarding the association between SBPr and outcome after EVT is scarce. In a previous multicenter study of patients with successful revascularization, there was a negative correlation between SBPr and poor outcome suggesting a potential benefit from SBP reduction¹³. In the present study, SBPr was associated with worse outcome in univariate analysis. However, the association did not retain statistical significance in adjusted models suggesting that the observed association was an association between SBPr and stroke severity and other baseline comorbidities, rather than a direct association between SBPr and functional outcome. Despite comparable cohort, the differences in study design and definition of SBPr may have contributed to variation in the results. Future, randomized studies to investigate the association between SBPr and outcome after MT are critically needed.

The relationship between BP and outcome after EVT is modulated by multiple factors such as revascularization status and underlying history and severity of hypertension²⁰. Patients with persistent vessel occlusion could be dependent on blood pressure to perfuse penumbral tissue. In this study, subgroup analyses based on revascularization status did not significantly alter the results. However, it should be noted that only 11.2% of our cohort failed to achieve successful revascularization. Therefore, our study is likely underpowered to detect a significant effect of SBPr in this subgroup of patients. Another group of patients that could theoretically be harmed from BP reduction is patients with chronic underlying hypertension. Chronic hypertension results in a rightward shift of autoregulation curve, therefore, a significant BP drop may have a more significant effect in patients with chronic hypertension than those without²⁰. In this study, we did not find a difference in the results between these groups. It should be noted, however, that SBPrmean was modest in our group (mean±SD, 27.3 ±25.5 mmHg).

The use of anti-hypertensive treatment was common in our cohort, consistent with previous reports¹³. To understand the effect of antihypertensive treatment on the relationship between SBPrmean and outcome, we divided patients based on antihypertensive treatment. We did not find a significant difference between groups and SBPrmean was not associated with poor outcome in the subgroups of patients treated with either continuous intravenous antihypertensives, >2 PRN intravenous antihypertensives, or neither. It should be noted that continuous antihypertensives were used more in the poor outcome than the good outcome group. This finding was also observed in the intensive blood pressure reduction with intravenous thrombolysis therapy for acute ischaemic stroke (ENCHANTED) trial and likely secondary to the fact that patients with poor outcome had more severe stroke and higher BP requiring more continuous antihypertensives.⁵

The management of BP post EVT differs between centers and between operators in each center. The American Heart Association/American Stroke Association (AHA/ASA) guidelines recommend maintaining SBP less than 180 mm Hg after EVT. However, lower SBP goals have been widely adapted due to concern of reperfusion injury^{22, 22}. The safety and benefit of lowering BP post EVT remains unknown. In an experimental model of transient middle cerebral artery occlusion, the use of hydralazine and candesartan to lower mean arterial pressure in the reperfusion period was associated with reduced infarct size²³. Moreover, a recent study comparing different SBP targets used in clinical practice after successful revascularization found a potential benefit from lower SBP goals (i.e <140 mm Hg and <160 mm Hg) compared to the American Heart AHA/ASA guideline recommended SBP goal (<180 mm Hg)²². Our study, along with previous studies, suggest that lowering SBP post EVT may be safe.

Despite the prospective design of the present study, there are limitations that should be considered when interpreting the results of this study. Data regarding collateral score and stroke etiology was not available for this study, so we could not study the association between SBPr and outcome across different stroke etiologies and collateral statuses. It is reasonable to assume that BP reduction can have detrimental effect in patients with severe atherosclerotic disease and/or poor collateral circulation. Furthermore, time-stamped BP data was not available, therefore we could not assess the effect of SBPr during different time periods on outcome. It is possible that acute BP reduction in the first 1–2 hours post EVT has a different effect on outcome than more gradual BP reduction. Furthermore, our cohort consisted mostly of recanalized patients, therefore, our results mainly apply to successfully recanalized patients. Future studies are needed to further assess the association between SBPr and outcome in nonrecanalized patients. Finally, due to heterogeneous imaging modalities across centers, it was not possible to calculate infarct extension.

Conclusion:

The present study documented no supporting evidence that SBP reduction may be independently associated with poor outcome after EVT. Subgroup analyses based on history of hypertension and antihypertensive treatment yielded similar results. Future randomized controlled clinical trials are needed to truly establish safety, and ultimately, efficacy of drug-induced SBPr in EVT-treated patients.