Efficacy of immediate anti-hypertensive treatment in patients with acute ischaemic stroke stratified by mean arterial pressure and pulse pressure: a secondary analysis of the China Antihypertensive Trial in Acute Ischemic Stroke trial

Abstract

Background

Whether mean arterial pressure (MAP) and pulse pressure (PP), two indicators of cerebral perfusion, could guide the selection of anti-hypertensive strategies after acute ischaemic stroke remains uncertain. Our study was to explore the impact of early anti-hypertensive intervention on adverse clinical outcomes following ischaemic stroke stratified by the levels of MAP and PP based on the China Antihypertensive Trial in Acute Ischemic Stroke (CATIS).

Methods

The trial randomised 4071 acute ischaemic stroke patients with elevated systolic blood pressure (SBP) to receive anti-hypertensive treatment (targeting a 10%–25% reduction in SBP during the 24 hours postrandomisation, reaching a BP level <140/90 mm Hg in 7 days, further keeping these levels throughout hospitalisation) or discontinue anti-hypertensive treatment during hospitalisation. The primary outcome was death or major disability at 14 days or hospital discharge. Study outcomes were analysed by comparing the BP-lowering intervention group and control group, stratified by tertiles of MAP or PP levels.

Results

No significant difference was observed in the primary outcome between the intervention and control groups across all MAP (p=0.69 for homogeneity) and PP (p=0.78 for homogeneity) categories. The corresponding odds ratios (95% CIs) were 1.08 (0.85–1.36), 0.92 (0.74–1.15) and 1.00 (0.81–1.25) for participants with low, intermediate, and high MAP and were 0.99 (0.79–1.25), 1.06 (0.84–1.34) and 0.95 (0.77–1.18) for participants in PP subgroups, respectively. Furthermore, early anti-hypertensive intervention was not associated with secondary outcomes (including neurological deterioration, recurrent stroke, vascular events and all-cause mortality) by MAP and PP (all p>0.05).

Conclusions

Early anti-hypertensive therapy neither decreased nor increased the odds of major disability, mortality, recurrent stroke or vascular events in patients with acute ischaemic stroke regardless of different MAP and PP levels.

Trial registration number

ClinicalTrials.gov identifier: NCT01840072.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Existed trials had yielded inconsistent effects of anti-hypertensive treatment on adverse outcomes after acute ischaemic stroke. Whether mean arterial pressure (MAP) and pulse pressure (PP), two indicators of cerebral perfusion, could guide the selection of anti-hypertensive strategies after acute ischaemic stroke remains uncertain.

WHAT THIS STUDY ADDS

• In this subgroup analysis of the China Antihypertensive Trial in Acute Ischemic Stroke, early anti-hypertensive treatment had a neutral impact on clinical outcomes among acute ischaemic stroke patients with various MAP and PP levels.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• In the setting of acute ischaemic stroke, anti-hypertensive treatment does not improve or worsen outcomes regardless of different MAP and PP levels and therefore should be based on individual clinical judgement.

Introduction

Globally, stroke is one of the leading contributors to mortality and neurological impairment in adults, with hypertension as its primary modifiable risk factor.^{1 2} The significant benefits of lowering blood pressure (BP) in reducing stroke incidence in the hypertensive and stroke recurrence have been well established.³ However, existing trials had yielded inconsistent effects of anti-hypertensive medications and BP-lowering treatments on adverse outcomes after acute ischaemic stroke.^{4 5} The Enhanced Control of Hypertension and Thrombolysis Stroke Study (ENCHANTED) and ENCHANTED-2 trials have recently revealed that intensive BP intervention during the initial few hours following treatment with intravenous thrombolysis or endovascular thrombectomy does not enhance functional outcomes and may potentially exacerbate them.^{6 7}

Increased BP is commonly observed after the onset of stroke, but it is still uncertain whether or not anti-hypertensive treatment should be administered immediately after ischaemic stroke.8,11 The main concern is that early BP-lowering therapy could diminish collateral flow and expand the size of brain infarction, especially for patients with disturbed cerebral autoregulation or cerebral hypoperfusion.¹¹ Mean arterial pressure (MAP) is a combination of systolic BP (SBP) and diastolic BP (DBP) and serves as a well-known prognostic marker for cardiovascular events across various clinical settings.^{12 13} Importantly, MAP is a key factor in determining cerebral perfusion and cerebral blood flow velocity. And both higher and lower MAP may be a manifestation of impaired cerebral haemodynamics.^{14 15} However, whether MAP can affect the efficacy of anti-hypertensive treatment on adverse clinical events after acute stroke is an urgently solved issue, which has important clinical practice significance for guiding the selection of BP management strategies for those with different cerebral perfusion status.

Pulse pressure (PP) is a clinical indicator of arterial stiffness, which is calculated by subtracting SBP and DBP values. Elevated PP may contribute to augmented pulse-wave velocity and systemic load, and further lead to a reduced tissue perfusion pressure.^{16 17} Therefore, baseline PP could be important in determining the effectiveness of early BP reduction following ischaemic stroke.

Accordingly, we performed this subgroup analysis of the China Antihypertensive Trial in Acute Ischemic Stroke (CATIS) to evaluate how immediate anti-hypertensive treatment affects the primary outcome of death or major disability, as well as several secondary outcomes based on their initial MAP and PP levels.

Methods

Study design

The CATIS was a multicentre randomised clinical trial carried out in 26 hospitals in China. We have previously described the aim and main findings of this trial.^{4 18 19} In brief, the trial recruited 4071 patients within 48 hours of symptom onset from August 2009 through May 2013. The qualified subjects aged ≥22 years, who experienced their first ischaemic stroke and an elevated SBP of 140–220 mm Hg, were enrolled. Brain MRI or CT was used to confirm the ischaemic stroke. Patients in a deep coma, or with resistant hypertension, cerebrovascular stenosis, aortic dissection, atrial fibrillation, unstable angina, acute myocardial infarction or serious heart failure, as well as BP ≥220/120 mm Hg were excluded (see trial protocol in online supplemental 1).

This clinical trial received ethical approval from the Ethics Committees at Soochow University, Tulane University and all 26 participating hospitals. All individuals provided written informed consent prior to their enrolment.

Intervention

Participants were randomised to the intervention or control group. Randomisation was centrally performed with stratification according to participating hospitals and the administration of anti-hypertensive drugs. The schedule of randomisation was generated using PROC PLAN in SAS V.9.4 and concealed until a qualified participant was prepared to be enrolled.

In the anti-hypertensive intervention group, the goal was to reduce SBP by 10%–25% within 24 hours postrandomisation, reach a BP level <140/90 mm Hg within 7 days, and sustain this target BP within their hospital stay. A predetermined treatment algorithm was used in the intervention group which was presented in online supplemental eFigure 1. Patients stopped taking all anti-hypertensive medications throughout the patient’s hospitalisation in the control group.^{4 18 19} Physicians will monitor patients with ischaemic stroke closely during the trial. Patients who encounter serious circumstances during the intervention will be withdrawn. The intervention and the control group will both get standard care based on the national guidelines of China during hospitalisation and discharge from hospital.²⁰

Data measurements

On enrolment, demographic information, medical and medication history, as well as the type of ischaemic stroke were assessed. Trained neurologists used the National Institute of Health Stroke Scale (NIHSS) to assess the magnitude of stroke severity at baseline. BP levels were measured with the participant using a standard mercury sphygmomanometer by trained nurses. After randomisation, BP was obtained three times every 2 hours within the first 24 hours, then each 4 hours for the next two and 3 days, thereafter three times daily until hospital discharge or death.^{4 18 19} In the present study, MAP was computed as SBP/3+2*DBP/3; PP was computed as SBP - DBP. Study subjects were separately categorised into three subgroups based on tertiles of baseline MAP (low, <114 mm Hg; intermediate, 114–123 mm Hg; high, ≥123 mm Hg) or PP (low, <61 mm Hg; intermediate, 61–73 mm Hg; high, ≥73 mm Hg).

Outcome assessment

The primary outcome in the CATIS trial was a composite of death and major disability at 14 days or hospital discharge, which was defined as a modified Rankin Scale (mRS) score 3–6. The mRS ranges from 0 to 6, and greater scores indicate higher levels of disability.²¹ The secondary outcomes included neurological deterioration, a combination of all-cause mortality and major disability, recurrent stroke and vascular events at 3 months.^{4 18 19} Neurological deterioration was defined as NIHSS_{14 Days or Hospital Discharge} - NIHSS_baseline≥2 points.²²

Experienced neurologists and nurses, who were blinded to the therapy allocation, conducted the assessment of study outcomes. Vascular events were extracted from hospital data, which were further assessed and adjudicated by a trial-wide outcomes assessment committee. And all certificates of death were obtained for deceased participants.

Statistical analysis

The baseline characteristics of patients in different MAP and PP categories were compared between the intervention and control groups. Kaplan-Meier cumulative incidence curves were used to estimate the cumulative incidence rates of mortality, vascular events and recurrent stroke across two groups. The log-rank test was used to assess differences in the study outcomes. Using categorical logistic regression and Cox proportional hazards regression models, unadjusted ORs or HRs and their 95% CIs were derived to assess the differences between the intervention and control groups. Additionally, the impact of early BP reduction on the whole mRS range was analysed using an ordinal logistic regression model.²³ Heterogeneity of the intervention effects on adverse study outcomes based on MAP or PP subgroups was evaluated by adding an interaction term in models. Furthermore, a sensitivity analysis was carried out to test the robustness of the results according to adjusting age, sex, baseline SBP and NIHSS score, time from onset to hospital, ischaemic stroke subtype, body mass index, cigarette smoking and history of diabetes, which may influence the efficacy of immediate anti-hypertensive treatment. Data analyses were performed using SAS, V.9.4.

Results

Study participants

In the CATIS trial, 4071 participants were ultimately enrolled. Of these participants, seven patients of the intervention group and six of the control group withdrew from this trial while hospitalised. Additionally, 43 participants in the intervention group and 40 in the control group were lost to follow-up at 3 months (figure 1). All subjects were involved at 14 days or hospital discharge, and 3975 (1988 in treatment group and 1987 in the control group) subjects were involved in the study at 3 months. Overall, baseline characteristics were in balance between intervention and control groups in each MAP category and PP category (table 1; online supplemental eTable 1).

Figure 1. Study participant flow chart. MAP, mean arterial pressure; PP, pulse pressure.

Table 1. Baseline characteristics of subjects according to the levels of mean arterial pressure.

Characteristics	Low MAP (<114 mm Hg)			Intermediate MAP (114–123 mm Hg)			High MAP (≥123 mm Hg)
	Treatment (n=624)	Control (n=676)	P value	Treatment (n=703)	Control (n=695)	P value	Treatment (n=711)	Control (n=662)	P value
Age, y	64.3 (10.3)	64.2 (10.8)	0.91	62.1 (10.7)	61.6 (10.5)	0.36	60.1 (10.8)	59.7 (11.2)	0.50
Male sex	401 (64.3)	395 (58.4)	0.03	461 (65.6)	452 (65.0)	0.83	455 (64.0)	440 (66.5)	0.34
Time from onset to hospital, h	15.8 (13.1)	15.3 (13.5)	0.48	15.6 (12.7)	15.2 (12.8)	0.58	14.6 (12.8)	14.1 (12.5)	0.52
SBP at entry, mm Hg	152.2 (8.9)	153.0 (8.8)	0.12	163.4 (10.5)	162.4 (9.9)	0.09	182.6 (15.0)	181.8 (14.9)	0.33
DBP at entry, mm Hg	86.6 (6.4)	86.0 (6.8)	0.09	95.9 (5.4)	96.3 (5.0)	0.13	106.7 (9.0)	107.6 (9.3)	0.07
BMI, kg/m2	24.9 (3.0)	24.8 (3.2)	0.77	24.7 (3.1)	25.0 (3.2)	0.13	25.2 (3.4)	25.2 (3.0)	0.92
History of hypertension	481 (77.1)	528 (78.1)	0.66	554 (78.8)	537 (77.3)	0.49	575 (80.9)	534 (80.7)	0.92
Use of anti-hypertension medicine	298 (47.8)	331 (49.0)	0.66	359 (51.1)	322 (46.3)	0.08	357 (50.2)	330 (49.9)	0.89
History of hyperlipidaemia	42 (6.7)	51 (7.5)	0.57	52 (7.4)	55 (7.9)	0.72	43 (6.1)	34 (5.1)	0.46
History of diabetes	144 (23.1)	132 (19.5)	0.12	113 (16.1)	117 (16.8)	0.70	112 (15.8)	101 (15.3)	0.80
History of CHD	71 (11.4)	93 (13.8)	0.20	85 (12.1)	82 (11.8)	0.87	60 (8.4)	53 (8.0)	0.77
Cigarette smoke	218 (34.9)	225 (33.3)	0.53	237 (33.7)	272 (39.1)	0.04	270 (38.0)	263 (39.7)	0.51
Alcohol consumption	177 (28.4)	181 (26.8)	0.52	214 (30.4)	228 (32.8)	0.34	223 (31.4)	230 (34.7)	0.18
NIHSS score at baseline	4.0 (2.0–7.0)	4.0 (2.0–7.0)	0.59	4.0 (3.0–7.0)	5.0 (3.0–8.0)	0.02	5.0 (2.0–8.0)	5.0 (3.0–8.0)	0.83
Ischaemic stroke subtype*
Thrombotic	491 (78.7)	548 (81.1)	0.28	541 (77.0)	559 (80.4)	0.11	543 (76.4)	488 (73.7)	0.26
Embolic	27 (4.3)	30 (4.4)	0.92	36 (5.1)	38 (5.5)	0.77	36 (5.1)	35 (5.3)	0.85
Lacunar	125 (20.0)	116 (17.2)	0.18	137 (19.5)	117 (16.8)	0.20	155 (21.8)	152 (23.0)	0.61

^*Eleven patients with both thrombotic and embolic, 92 with thrombotic and lacunar, five with embolic and lacunar, and one with all three subtypes.

BMI, body mass index; CHD, coronary heart disease; DBP, diastolic blood pressure; MAP, mean arterial pressure; IQR, interquartile range; NIHSS, NIH Stroke Scale; and SBP, systolic blood pressure.

Blood pressure intervention

Within 24 hours following randomization, SBP levels differed significantly between the intervention group and control group in each MAP category. The corresponding differences in SBP were −8.9 (95% CI −10.3 to −7.4) mm Hg (p<0.001) among low MAP patients, −8.6 (95% CI −10.1 to −7.1) mm Hg (p<0.001) among intermediate MAP patients and −8.0 (95% CI −9.6 to −6.3) mm Hg (p<0.001) among high MAP patients (online supplemental eTable 2). At 7 and 14 days after randomisation, the SBP of the anti-hypertensive treatment group was also lower than the control group in each of the MAP categories (all p<0.001) (online supplemental eTable 2; figure 2). Meanwhile, similar BP decreases were observed for PP subgroups (online supplemental eTable 3; online supplemental eFigure 2).

Figure 2. Mean and 95% CI of systolic blood pressure since randomisation by baseline mean arterial pressure (MAP). (A) Patients with low MAP. (B) Patients with intermediate MAP. (C) Patients with high MAP. Solid lines indicate mean systolic blood pressure; shading, 95% CI.

Effects on primary outcome

No significant interaction of anti-hypertensive intervention with baseline MAP (p=0.69 for homogeneity) or PP (p=0.78 for homogeneity) was found on the primary outcome of death or major disability (figure 3). Across all MAP and PP categories, the intervention group and the control group had no significant differences in the primary outcome. For patients with low, intermediate and high baseline MAP, ORs connected to the BP decreasing intervention were 1.08 (95% CI 0.85 to 1.36), 0.92 (95% CI 0.74 to 1.15) and 1.00 (95% CI 0.81 to 1.25), respectively. Similarly, ORs associated with intervention were 0.99 (95% CI 0.79 to 1.25), 1.06 (95% CI 0.84 to 1.34) and 0.95 (95% CI 0.77 to 1.18) among each of the PP categories, respectively (figure 4). Furthermore, anti-hypertensive treatment did not show a significant impact on mRS scores and death among all three MAP subgroups and three PP subgroups (all p>0.05) (online supplemental eFigure 3).

Figure 3. Clinical outcomes at 14 days or hospital discharge and 3 months according to baseline mean arterial pressure. MAP, mean arterial pressure. *P value for homogeneity.

Figure 4. Clinical outcomes at 14 days or hospital discharge and 3 months according to baseline pulse pressure. PP, pulse pressure. *P value for homogeneity.

Effects on clinical outcomes at 3 months

In all MAP and PP subgroups, the anti-hypertensive treatment group’s mean SBPs as well as DBPs were substantially lower than those of the control group at 3 months. However, Kaplan-Meier curves indicated that patients of the treatment group have no difference from the control group between either the MAP or PP subgroups about outcomes of death, as well as recurrent stroke or vascular events (log-rank p>0.05) (figure 5; online supplemental eFigure 4 and 5). For patients in both MAP subgroups (p=0.18) and PP subgroups (p=0.77), no significant differences in the composite outcome of mortality and major disability were observed between the intervention and the control group (figure 3). The ORs (95% CIs) were 1.10 (0.85–1.42) in participants with low MAP (<114 mm Hg), 1.04 (0.81–1.33) in participants with intermediate MAP (114–123 mm Hg) and 0.86 (0.68–1.10) in participants with high MAP (≥123 mm Hg). According to PP levels, the ORs (95% CIs) were 0.96 (0.74–1.23), 1.01 (0.78–1.30) and 1.01 (0.80–1.28) in each of the PP categories, respectively (figure 4). Likewise, there was not a significant distinction between the intervention and control groups with regard to the mRS score, mortality, vascular events or recurrent stroke in MAP subgroups and PP subgroups (p>0.05 for all homogeneity) (online supplemental eFigure 3; figures3 4).

Figure 5. Kaplan–Meier survival curves of death according to the baseline mean arterial pressure or pulse pressure.

Sensitivity analysis

We performed a sensitivity analysis by adjusting several covariates in the multivariable model. We found that the associations between anti-hypertensive treatment and study outcomes did not significantly change in MAP subgroups, as well as PP subgroups (online supplemental eTable 4 in online supplemental 2). For instance, patients with MAP values of less than 114, 114–123 and more than 123 mm Hg had ORs for the primary outcome of 1.17 (95% CI 0.87 to 1.58), 1.12 (95% CI 0.85 to 1.48) and 0.98 (95% CI 0.74 to 1.30) for anti-hypertensive therapy (p=0.46 for homogeneity).

Discussion

The current subgroup analysis was the first to evaluate the modification of cerebral perfusion, assessed by MAP and PP, on the effectiveness of early anti-hypertensive treatment after ischaemic stroke. First, the findings showed that compared with control group, both SBP and DBP levels were lower in the intervention group at 14 days or hospital discharge and 3 months after randomisation, among patients with various MAP and PP levels at baseline. Nonetheless, we found that anti-hypertensive intervention was not associated with the primary outcome, regardless of patients with low, intermediate and high levels of MAP or PP. Additionally, within the post-treatment follow-up of 3 months after randomisation, there was also no significant impact of immediate BP reduction on all secondary outcomes in different MAP and PP subgroups. Our study suggested that baseline MAP and PP levels did not modify the anti-hypertensive intervention effects after acute ischaemic stroke, and anti-hypertensive intervention present neutral impacts on adverse clinical outcomes for patients with different MAP and PP levels.

Management of hypertension is a crucial therapy for primary and secondary prevention of stroke.²⁴ But the optimal strategies for BP management after the onset of ischaemic stroke are still up for debate, especially for those with different cerebral perfusion status. The clinical guidelines for acute ischaemic stroke patients did not provide a definite recommendation to guide the treatment of early raised BP. SBP is an indirect index for early haemodynamic changes, and a comparatively regular SBP may help to repair haemodynamic abnormalities and keep adequate tissue perfusion.²⁵ According to our prior subgroup analysis, early anti-hypertensive intervention had no influence on death or major disability among patients with different baseline SBP levels. But SBP significantly modified the effect of anti-hypertensive intervention on mortality; the findings indicated that early anti-hypertensive intervention was in relation to a higher reduction in in-hospital mortality among patients whose SBP was 180–220 mm Hg.²⁶

Both MAP and PP are composed of SBP and DBP, reflecting the steady and pulsatile components of BP, respectively.²⁷ MAP reflects left ventricular contractility, cardiac output and arterial resistance.²⁸ Available evidence indicates that MAP provides a better predictive ability than SBP and DBP when it comes to forecasting the likelihood of cardiovascular diseases.^{13 29} Furthermore, MAP acts as a drive of the perfusion for a range of vital organs, including the brain and heart, and is considered a more accurate indicator of cerebral perfusion than SBP.^{30 31} Lowering BP immediately may lessen haemorrhagic transformation, brain oedema and vascular damage, but it may also increase the extent of the brain infarct and impair cerebral perfusion of the ischaemic region.¹¹ So we hypothesised that acute ischaemic stroke patients with lower MAP may not be safe when they received anti-hypertensive treatment. However, in the present study, we found that the early anti-hypertensive intervention did not increase or decrease the likelihood of adverse study outcomes, regardless of baseline MAP levels in the present study.

PP is determined by the pressure from reflected waves and the interaction of expelled ventricular blood with main arteries.³² A widened PP may indicate the stiffening of the conduit vessels, which increases the afterload and simultaneously reduces coronary perfusion.^{33 34} Ultimately, it may affect the ability of the heart to supply blood to the brain, resulting in inadequate cerebral perfusion. Similar to the findings of MAP, our subgroup analysis also found that the effectiveness of the intervention was unaffected by baseline PP levels, and the neutral effect on the primary outcome was noticed regardless of different PP levels at baseline.

The secondary analysis of the Systolic Blood Pressure Intervention Trial and the Action to Control Cardiovascular Risk in Diabetes BP trial indicated that in patients with hypertension, intensive SBP reduction did not heighten the stroke risk, even for patients with achieved extremely low MAP or PP values.³⁵ Our findings also indicated that even for acute ischaemic stroke patients with lower MAP and higher PP, the decision to lower BP does not enhance or worsen outcomes. As such, the selection of anti-hypertensive therapy should be made according to individual clinical discretion. So, the findings of our study offer valuable insights for enhancing BP control after acute ischaemic stroke. The current clinical guidelines and recent clinical trials indicated that administration of intervention within the first 24 hours or 48–72 hours following acute ischaemic stroke did not effectively improve the adverse outcomes.36,38 The optimal timing for starting or restarting anti-hypertensive treatment still needs to be resolved in future prospective clinical studies.

In spite of this, given that our study was a post hoc analysis of CATIS, some limitations should be recognised. First, these were subgroup analyses of the CATIS and small samples in each subgroup, and multiple comparisons could reduce the statistical power. And the comparability of the intervention and control groups may also be affected in the subgroup analysis. Second, due to different requirements for BP reduction, our study excluded participants whose BP ≥220/120 mm Hg, those undergoing endovascular thrombectomy and those undergoing intravenous thrombolytic treatment. Third, we lacked direct data on blood flow in the brain, collateral blood circulation or the existence of penumbral tissue during our study, which may limit the ability to precisely estimate the effect of cerebral perfusion on early anti-hypertensive intervention and also did not collect outcomes like haemorrhagic transformation and imaging data, which may affect us evaluating the safety of anti-hypertension. Finally, only Chinese ischaemic stroke patients participated in our study, so the generalisability of these findings to other populations warrants cautious interpretation and requires further validation.

Conclusions

The subgroup analyses of the CATIS trial found that early anti-hypertensive therapy neither decreased nor increased the odds of major disability, mortality, recurrent stroke or vascular events, regardless of different MAP and PP levels. Additional well-designed prospective clinical studies are necessary to validate our findings.

Data availability statement

Data are available upon reasonable request.