| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(Hypertension. 2004;43:18.)
© 2004 American Heart Association, Inc.
Scientific Contributions |
From the Institute of Neuroscience, University of Nottingham, Nottingham, UK.
Correspondence to Prof Philip Bath, Division of Stroke Medicine, University of Nottingham, City Hospital Campus, Nottingham, NG5 1PB UK. E-mail philip.bath{at}nottingham.ac.uk
| Abstract |
|---|
|
|
|---|
Key Words: stroke, thrombotic stroke, hemorrhagic blood pressure morbidity mortality
| Introduction |
|---|
|
|
|---|
40% of patients remain hypertensive. The causes of this pathophysiologic response are multifactorial and are related to preexisting high BP, activation of the neuroendocrine systems (sympathetic nervous system, renin-angiotensin axis, and glucocorticoid system), increased cardiac output, and "white coat hypertension."37 It has been suggested that high BP is associated with a poor outcome after acute stroke, although the results of observational studies have given conflicting results. Some authors have even demonstrated better outcomes in patients with high initial BP.8,9 Data from the International Stroke Trial (IST) confirmed that the risk of early death and late death or dependency was independently associated with increasing systolic BP (SBP) in 17 398 patients.10 We report here a systematic review of observational studies of BP and outcome and assess the relation between the two.
| Methods |
|---|
|
|
|---|
Data Extraction
Two authors (M.W. and J.L.-B.) independently extracted data; discrepancies were resolved by P.M.W.B. Studies with dichotomous data were analyzed separately from those with continuous data. Within both types, the studies were subdivided by outcome measure (death, death or dependency, death or deterioration, stroke recurrence, cerebral edema, or hematoma expansion) and by BP measurement (SBP, diastolic BP [DBP], or mean arterial BP [MABP]). The articles were then arranged by stroke type (primary intracerebral hemorrhage [PICH], ischemic stroke, or mixed). Figures 1A and 1B show typical forest plots and illustrate the layout of studies.
|
In cases where articles quoted several BP measurements, the earliest readings were used. Outcomes after the longest follow-up period were used when present at multiple time points. Some articles gave outcome data for several BP strata (eg, mortality with SBP <140 mm Hg, 141 to 180 mm Hg, and >180 mm Hg).14 In such instances, the data were dichotomized into a high-BP and a combined normal/low-BP group by using a cut point nearest to 150 mm Hg because outcome might be best at this level.10 Likewise, publications with DBP or MABP in several strata were dichotomized as close as possible to 90 mm Hg and 110 mm Hg, respectively. Additionally, some articles gave continuous BP data in >2 groups (eg, SBP for patients with "improved," "unchanged," or "worse" neurologic status). These studies were incorporated into the review after the data were transformed into 2 groups (eg, combined "improved"/"unchanged" and "worse"). This was achieved by generating pseudo-random BP data (assuming a normal distribution for BP) and then merging this before recalculating an overall combined mean BP and SD.
Analysis
Data were analyzed with the Cochrane Collaboration Review Manager (version 4.1) software. Dichotomous data are given as odds ratio (OR) and 95% confidence interval (CI) and as weighted mean difference (WMD) with 95% CI for continuous data. These were calculated with a random-effects model, and statistical heterogeneity was assessed with a
2 test. We explored the causes of heterogeneity by using sensitivity analyses based on type of stroke. Publication bias was assessed with the Egger asymmetry test15 (STATA function Metabias) on dichotomous outcome data. Significance was set at P<0.05.
| Results |
|---|
|
|
|---|
|
|
BP was recorded at admission in 18 of the included articles. The method used was given in only 9 studies; 5 used "clinic" BP measurement, 3 used both "clinic" BP measurement and ambulatory BP monitoring, and 1 used ambulatory BP monitoring alone. The criteria used to define high BP varied considerably: SBP thresholds ranged from 150 to 200 mm Hg,1618 MABP levels of 140 to 145 mm Hg,16,19 and DBP between 90 and 115 mm Hg.16,2022 One article did not specify the criteria used for defining high BP23; this study was analyzed with the publications that dichotomized according to MABP. In addition, 3 studies allocated subjects to the high-BP group when SBP, DBP, or both exceeded certain threshold values.2426 These articles were also analyzed with publications that dichotomized according to MABP.
Death was the most commonly reported outcome measure (survival status present in 7242 patients, 66.5%). There were fewer data for combined death or dependency (1290, 11.8%) and combined death or deterioration (1196, 11.0%). These outcomes were assessed with validated stroke scales in 7 articles (Rankin score,27 Scandinavian Neurological Stroke Scale,30,32 Canadian Neurological Score,33 or National Institutes of Health Scale29). The other studies used either nonvalidated scales9 or qualitative assessments, such as discharge disposition.18,25 Patient follow-up varied considerably between 6 days and 6 years, although most articles chose to measure outcome at discharge from hospital. Potential mechanisms of poor outcome were reported in 4 articles. Three looked at PICH (708, 6.5%) and assessed the relation between BP and hematoma expansion; the other study investigated early recurrence in ischemic stroke (1273, 11.7%). No studies assessed cerebral edema or hemorrhagic transformation.
There was no publication bias (Egger test P=0.21) in articles reporting SBP and mortality data. When stroke as a whole was assessed, patients with high SBP or DBP were at a 1.5- to 5.0-fold increased risk of dying or combined death or dependency/deterioration (Table 2). Similarly, high MABP was associated with increased odds for combined death and dependency. In patients with poor outcome, judged as death and death or dependency/deterioration, there was a trend for higher SBP/DBP levels of 5/3 mm Hg. Heterogeneity was present in several of these analyses (Table 2), so the data were further examined by stroke type.
|
Primary Intracerebral Hemorrhage
Increased odds of death and death or disability/deterioration were found in patients with high BP (Table 3). Additionally, MABP was higher in patients who died after PICH (Table 4). The odds of hematoma expansion were increased for patients with high SBP (Table 2).
|
|
Ischemic Stroke
Limited data were available for studies specifically including patients with ischemic stroke. SBP/DBP were higher by 12/6 mm Hg in patients who died or became dependent (Table 4). Similarly, ischemic stroke patients had a 2-fold increase in the risk of stroke recurrence when their DBP was elevated (Table 2).
Mixed Stroke Studies
The odds of death were doubled in patients with high DBP. SBP was higher by 6.39 mm Hg in patients who subsequently died (Table 4).
| Discussion |
|---|
|
|
|---|
This review has found evidence for mechanisms that might link high BP to poor outcome. In ischemic stroke, high DBP was associated with a 2-fold increase in risk of early recurrence. SBP was an important determinant of recurrent stroke in IST, wherein an initial SBP of 200 mm Hg or more conferred a >50% higher risk of recurrence than did an SBP of 130 mm Hg.10 No studies investigating the relation between BP and hemorrhagic transformation or cerebral edema in ischemic stroke fulfilled our inclusion criteria. Nevertheless, evidence that was excluded from the review suggests that an acutely elevated SBP is associated with increased fatal cerebral edema.10,35 Also, several studies have observed that high BP promotes hemorrhagic transformation in animal models,3638 although this was not found in the IST.10 As far as PICH is concerned, patients with high SBP were almost twice as likely to have hematoma expansion. This could support the concept that high BP in the acute phase of hemorrhagic stroke leads to a worse outcome, at least in part, by promoting continued intracerebral bleeding. However, this relation might be confounded by the timing of inclusion, because patients with severe PICH tend to present earlier and are at greater risk of rebleeding.
Although this review has demonstrated a positive association between high BP and subsequent events in acute stroke, the findings are limited by several factors. First, there were considerable differences in patient eligibility, case mix (including baseline BP), definition of high BP, measurement and timing of BP, and type and timing of outcome among the studies. This could be considered an advantage, because it means that the relations observed are more likely to be generalizable. However, it might also have led to statistical heterogeneity. To assess whether stroke type was a potential source of heterogeneity in the review, we analyzed ischemic stroke and PICH separately. The relation between outcome and BP tended to be stronger in patients with PICH (OR, 2.26 to 5.57) compared with those with ischemic stroke. Similarly, patients who had a poor outcome had a higher BP if they had PICH (MABP, 11 mm Hg) than ischemic stroke (SBP/DBP, 12/6 mm Hg, equivalent to an MABP of
8 mm Hg). Nevertheless, these comparisons are indirect and not precise. Other explanations for heterogeneity among the studies are likely, but we were unable to explore these owing to the paucity of data.
A second limiting factor is that we were unable to judge whether the relations that we observed were independent of other factors such as age, premorbid BP status, drug therapy, stroke severity, or timing of BP measurement. A meta-analysis of the studies, based on individual patient data, would be required to investigate this further. This issue is important, because some of the relations might have been different if potential confounding factors had been accounted for. For example, the relation between BP and rebleeding in PICH is probably mostly explained by an interaction between timing of BP measurement and severity.39,40 Likewise, high premorbid BP contributes to elevated BP in acute stroke and could contribute, in part, to poor outcome through previous cerebral vascular damage. Third, the strategy of dichotomizing BP, though clinically relevant, might create artificial strata for the analysis. For example, several articles have reported a U-shaped relation between BP and outcome, with the least poor outcome at an SBP of 140 to 160 mm Hg in the IST (judged by the nadir of the U).10,41,42 Hence, studies including a significant proportion of patients with a BP below a cut point of 160 mm Hg might be expected to miss a relation between high BP and poor outcome. Unfortunately, because only 1 article gave sufficient data on low BP in acute stroke, it was not possible to examine how it influenced outcome. Last, it is worth noting a consistent weakness in the observational studies themselves, namely, the lack of reporting on how BP was measured. All studies assessing BP should give information on equipment (manufacturer, model, technique, validation), user (how trained, assessed, and reassessed), and patient (number and site of readings, position).43
Perspectives
This article assesses data from all available observational studies and shows that there is a positive association between high BP and poor outcome in acute stroke. This relation might be mediated, at least in part, by early recurrence in patients with ischemic stroke and possibly, by early rebleeding in those with PICH. Although these relations suggest that high admission BP is directly causal, we cannot rule out that this relation was independent of other factors, such as stroke severity or timing of measurements. Nevertheless, because lowering a high BP prevents first and recurrent stroke, it can be hypothesized that moderate lowering of a high BP in acute stroke might similarly reduce early death and deterioration and late death and dependency. Aggressive lowering of BP will do harm through inducing cerebral hypoperfusion. Also, some animal data and limited clinical data support the alternative concept of induced hypertension as a treatment strategy.44,45 Until the results of ongoing randomized, controlled trials are available,46 the management of high BP in acute stroke will remain controversial.
| Acknowledgments |
|---|
Received August 26, 2003; first decision September 15, 2003; accepted October 24, 2003.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
R. Delgado-Mederos, M. Ribo, A. Rovira, M. Rubiera, J. Munuera, E. Santamarina, P. Delgado, O. Maisterra, J. Alvarez-Sabin, and C. A. Molina Prognostic significance of blood pressure variability after thrombolysis in acute stroke Neurology, August 19, 2008; 71(8): 552 - 558. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. C. Fric-Shamji, M. F. Shamji, J. Cole, and B. G. Benoit Modifiable risk factors for intracerebral hemorrhage: Study of anticoagulated patients Can Fam Physician, August 1, 2008; 54(8): 1138 - 1139.e4. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. I. Qureshi Acute Hypertensive Response in Patients With Stroke: Pathophysiology and Management Circulation, July 8, 2008; 118(2): 176 - 187. [Full Text] [PDF] |
||||
![]() |
H. K. Shin, M. Nishimura, P. B. Jones, H. Ay, D. A. Boas, M. A. Moskowitz, and C. Ayata Mild Induced Hypertension Improves Blood Flow and Oxygen Metabolism in Transient Focal Cerebral Ischemia Stroke, May 1, 2008; 39(5): 1548 - 1555. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Yong and M. Kaste Association of Characteristics of Blood Pressure Profiles and Stroke Outcomes in the ECASS-II Trial Stroke, February 1, 2008; 39(2): 366 - 372. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. Ay, E. M. Arsava, W. J. Koroshetz, and A. G. Sorensen Middle Cerebral Artery Infarcts Encompassing the Insula Are More Prone to Growth Stroke, February 1, 2008; 39(2): 373 - 378. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. B. van der Worp and J. van Gijn Acute Ischemic Stroke N. Engl. J. Med., August 9, 2007; 357(6): 572 - 579. [Full Text] [PDF] |
||||
![]() |
G. K.C. Wong, W. S. Poon, A. R. Zazulia, and W. J. Powers SYMPTOMATIC AUTOREGULATORY FAILURE IN ACUTE ISCHEMIC STROKE Neurology, July 10, 2007; 69(2): 222 - 222. [Full Text] [PDF] |
||||
![]() |
M. Rodriguez-Yanez, M. Castellanos, M. Blanco, M. M. Garcia, F. Nombela, J. Serena, R. Leira, I. Lizasoain, A. Davalos, and J. Castillo New-onset hypertension and inflammatory response/poor outcome in acute ischemic stroke Neurology, December 12, 2006; 67(11): 1973 - 1978. [Abstract] [Full Text] [PDF] |
||||
![]() |
O. Takahashi, E.F. Cook, T. Nakamura, J. Saito, F. Ikawa, and T. Fukui Risk stratification for in-hospital mortality in spontaneous intracerebral haemorrhage: A Classification and Regression Tree Analysis QJM, November 1, 2006; 99(11): 743 - 750. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Willmot, A. Ghadami, B. Whysall, W. Clarke, J. Wardlaw, and P. M.W. Bath Transdermal Glyceryl Trinitrate Lowers Blood Pressure and Maintains Cerebral Blood Flow in Recent Stroke Hypertension, June 1, 2006; 47(6): 1209 - 1215. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. Saloheimo, M. Ahonen, S. Juvela, J. Pyhtinen, E.-R. Savolainen, and M. Hillbom Regular Aspirin-Use Preceding the Onset of Primary Intracerebral Hemorrhage is an Independent Predictor for Death Stroke, January 1, 2006; 37(1): 129 - 133. [Abstract] [Full Text] [PDF] |
||||
![]() |
G Tsivgoulis, K Spengos, N Zakopoulos, E Manios, K Xinos, D Vassilopoulos, and K N Vemmos Twenty four hour pulse pressure predicts long term recurrence in acute stroke patients J. Neurol. Neurosurg. Psychiatry, October 1, 2005; 76(10): 1360 - 1365. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. P. Mattle, L. Kappeler, M. Arnold, U. Fischer, K. Nedeltchev, L. Remonda, S. M. Jakob, and G. Schroth Blood Pressure and Vessel Recanalization in the First Hours After Ischemic Stroke Stroke, February 1, 2005; 36(2): 264 - 268. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. N. Vemmos, G. Tsivgoulis, and K. Spengos Association Between Pulse Pressure Values During the Acute Stroke Stage and Stroke Outcome Stroke, November 1, 2004; 35(11): 2436 - 2436. [Full Text] [PDF] |
||||
![]() |
H. Christensen, M. Willmot, J. Leonardi-Bee, and P. M.W. Bath The Timing of the Blood Pressure Measurement May Affect the Result in Patients With Acute Stroke Hypertension, June 1, 2004; 43(6): e36 - e36. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Hypertension Home | Subscriptions | Archives | Feedback | Authors | Help | AHA Journals Home | Search Copyright © 2004 American Heart Association, Inc. All rights reserved. Unauthorized use prohibited. |