Blood Pressure–Lowering Treatment With Candesartan in Patients With Acute Hemorrhagic Stroke
Background and Purpose—Early and intensive blood pressure–lowering treatment seems to be beneficial in patients with acute hemorrhagic stroke and high blood pressure. We wanted to see if similar benefits can be shown from a later and more gradual blood pressure lowering, using data from the Scandinavian Candesartan Acute Stroke Trial (SCAST).
Methods—SCAST was a randomized- and placebo-controlled, double-masked trial of candesartan given for 7 days, in 2029 patients with acute stroke and systolic blood pressure ≥140 mm Hg. We assessed the effects of candesartan in the 274 patients with hemorrhagic stroke, using the trial’s 2 coprimary effect variables: the composite vascular end point of vascular death, stroke or myocardial infarction, and functional outcome at 6 months, according to the modified Rankin Scale. We used Cox proportional hazards models and ordinal regression for analysis and adjusted for key, predefined prognostic variables.
Results—There was no association between treatment with candesartan and risk of vascular events (17 of 144 [11.8%] versus 13 of 130 [10.0%]; hazard ratio, 1.36; 95% confidence interval, 0.65–2.83; P=0.41). For functional outcome we found evidence of a negative effect of candesartan (common odds ratio, 1.61; 95% confidence interval, 1.03–2.50; P=0.036).
Conclusions—There was no evidence that blood pressure–lowering treatment with candesartan is beneficial during the first week of hemorrhagic stroke. Instead, there were signs that such treatment may be harmful, but this needs to be verified in larger studies.
Hemorrhagic stroke is associated with a poor prognosis, and most trials of therapeutic interventions have shown disappointing results.1 One promising target for intervention is high blood pressure, which is common in the acute phase of hemorrhagic stroke.2 High blood pressure is an independent predictor of poor prognosis, and recently the Intensive Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial 2 (INTERACT2) gave hope that early and intensive blood pressure–lowering treatment improves long-term functional outcome.3
The Scandinavian Candesartan Acute Stroke Trial (SCAST) tested the effects of a later and more gradual blood pressure lowering in patients with acute ischemic or hemorrhagic stroke.4 The main analysis of all patients showed no evidence of a beneficial effect of candesartan, but rather a trend toward poorer functional outcome in the candesartan group.4 In this prespecified analysis, we wanted to see if a different effect was present in patients with hemorrhagic stroke.
Materials and Methods
SCAST was a randomized- and placebo-controlled trial of the angiotensin receptor blocker candesartan in 2029 patients presenting <30 hours of acute ischemic or hemorrhagic stroke and systolic blood pressure ≥140 mm Hg. The methods and main results have been published previously.4,5 In brief, candesartan or placebo tablets were started at ≈18 hours after stroke onset and given for 7 days, with doses increasing from 4 to 16 mg once daily during the first 3 days, and patients were followed for 6 months.
For the present analysis, we selected all patients with intracerebral hemorrhage. All patients in the trial were examined with computed tomography or MRI before inclusion, and the diagnosis of hemorrhagic stroke was based on the evaluation of the investigators at the site, who were blinded to treatment assignment.
The trial had 2 coprimary effect variables: the composite vascular end point of vascular death, stroke or myocardial infarction, and functional outcome at 6 months’ follow-up, as measured by the modified Rankin Scale.
The statistical analyses were performed according to the intention-to-treat principle and the trial’s statistical analysis plan.5 The composite vascular end point was analyzed by time to first event, using Cox proportional hazard models, and functional outcome was analyzed by ordinal regression, after checking that the parallel lines assumption was met. All analyses were adjusted for predefined prognostic variables: age, Scandinavian Stroke Scale score, and systolic blood pressure. In the subgroup analyses, heterogeneity of the treatment effect was assessed by adding an interaction term in the unadjusted model. The effects on the secondary end points were shown with unadjusted risk ratios. All analyses were performed with the SPSS software (version 18.0; SPSS Statistics, Chicago, IL).
In total, 274 patients with hemorrhagic stroke were included in the trial. Baseline characteristics were well balanced between the 2 treatment groups, except that there were more patients with atrial fibrillation and diabetes mellitus in the candesartan group (Table), and follow-up data were 99% complete.
Mean systolic blood pressure started at 174 mm Hg in both groups and fell in both groups during the treatment period, but was lower in patients treated with candesartan already from day 2, and the mean systolic difference from day 4 onward was ≈5 mm Hg (Figure 1).
The cumulative risk of the composite end point of vascular death, stroke, or myocardial infarction is shown in Figure 2. The composite vascular end point occurred in 17 of 144 patients (11.8%) treated with candesartan and in 13 of 130 controls (10.0%; adjusted hazard ratio, 1.36; 95% confidence interval, 0.65–2.83; P=0.41; panel A).
Figure 2 also shows the distribution of modified Rankin Scale scores in the 2 groups at 6 months (panel B). Treatment with candesartan was associated with an increased risk of a poorer functional outcome (adjusted common odds ratio, 1.61; 95% confidence interval, 1.03–2.50; P=0.036). A formal goodness-to-fit test gave no evidence that the proportional odds assumption was violated (P=0.32).
There were no statistically significant differences in the risks of the secondary end points (Table in the online-only Data Supplement), and there was no indication of a beneficial effect of candesartan in any of the subgroups, including patients treated early after stroke onset (<6 hours) or in patients with very high systolic blood pressure at baseline (Figure in the online-only Data Supplement).
INTERACT2 indicated that early and intensive blood pressure–lowering treatment is beneficial in patients with intracerebral hemorrhage,3 but no beneficial effects could be seen in our trial. Instead, we found that treatment with candesartan was associated with a worsening in functional outcome, as in the main analysis of all patients in the trial.4 These diverging findings can possibly be explained by the fundamental differences between the 2 trials. In INTERACT2, treatment was started at ≈3.7 hours and a systolic blood pressure difference of 10 mm Hg was achieved <30 minutes.3 In SCAST, treatment was started at ≈18 hours, and the maximal difference in systolic blood pressure of 5 mm Hg was achieved on day 4.4 Clearly, the 2 interventions affect very different pathophysiological mechanisms. Although INTERACT2 was designed to reduce hematoma growth,6 SCAST aimed to protect against the effects of high blood pressure and activation of the renin–angiotensin system during the aftermath of the hemorrhage.
SCAST was based on the findings from observational studies that high blood pressure is detrimental in the acute phase of stroke,2,7 possibly because of the negative effects of high blood pressure on microcirculation, edema, and hematoma growth.6,8,9 The study was also based on findings from experimental studies and 1 clinical study indicating that candesartan has beneficial effects in the acute phase of stroke.10 The absence of any sign of beneficial effects in SCAST might imply that treatment was started too late to limit brain injury, and that blood pressure reduction after the first few hours can only reduce cerebral perfusion and increase brain injury even further. Alternatively, it might imply that angiotensin receptor blockers have unwanted properties in the acute phase of stroke. Other trials are ongoing and will show whether agents with other properties can produce beneficial effects when given after the first few hours of hemorrhagic stroke.11,12
This analysis represents a subgroup of patients included in SCAST, and it is therefore at risk of false-negative conclusions, because of the play of chance alone. Nevertheless, the analysis was prespecified and represents a group of patients randomized to blood pressure–lowering treatment, with blinded outcome assessments and complete follow-up. Furthermore, the results are consistent with the result from the main analysis of all effect variables and in all the prespecified subgroups.4
In summary, we found no beneficial effect of later and more gradual blood pressure–lowering treatment with candesartan given after the first few hours of hemorrhagic stroke. Instead, we found signs of a negative effect on functional outcome. This might imply that treatment was started too late or that angiotensin receptor blockers have unwanted properties in the acute phase of stroke. Further studies are needed to help clarify the effects of blood pressure–lowering drugs after the first few hours of hemorrhagic stroke.
Sources of Funding
The trial was funded by grants from the South Eastern Norway Regional Health Authority and Oslo University Hospital. AstraZeneca supplied the study drugs, and AstraZeneca and Takeda supported the trial with limited, unrestricted grants.
P.M.W. Bath received travel support from AstraZeneca to attend meetings in the trial steering committee. Dr Berge received payment for lectures given at meetings arranged by AstraZeneca. The other authors report no conflicts.
Guest Editor for this article was Stephen M. Davis, MD, FRACP.
Presented in part at the European Stroke Conference, Nice, France, May 6–9, 2014.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.114.006433/-/DC1.
- Received June 12, 2014.
- Accepted August 25, 2014.
- © 2014 American Heart Association, Inc.
- Morgenstern LB,
- Hemphill JC III,
- Anderson C,
- Becker K,
- Broderick JP,
- Connolly ES Jr,
- et al
- Sandset EC,
- Murray G,
- Boysen G,
- Jatuzis D,
- Kõrv J,
- Lüders S,
- et al
- Arima H,
- Huang Y,
- Wang JG,
- Heeley E,
- Delcourt C,
- Parsons M,
- et al
- Anderson CS,
- Huang Y,
- Arima H,
- Heeley E,
- Skulina C,
- Parsons MW,
- et al
- Schrader J,
- Lüders S,
- Kulschewski A,
- Berger J,
- Zidek W,
- Treib J,
- et al