Introduction
Statins are effective in primary and secondary prevention of stroke [1-3]. Their long-term beneficial effects may be primarily mediated by their lipid-lowering effects. Statins may also work effectively in preventing recurrence or progression during the acute stage of stroke because they have antithrombotic, anti-inflammatory, and anti-oxidative effects [4-7].
Several observational studies have suggested that statin use before or during the acute stage of stroke was associated with better functional outcome and reduced mortality [7-9]. By contrast, intracerebral hemorrhage (ICH) developed more frequently in patients treated with high-dose statins during the subacute or chronic stage of ischemic stroke in large randomized trials [2,10]. Since the risk of hemorrhagic transformation is greater during the acute stage of stroke, statin treatment in acute stroke raises the concern of increased ICH risk. However, data from randomized trials are insufficient to establish whether statins are effective and safe in patients with acute ischemic stroke [11]. Furthermore, statins are not recommended in stroke guidelines as an agent with neuroprotective actions to improve outcome in acute stroke [12]. Therefore, we investigated the effect and safety of rosuvastatin in acute ischemic stroke patients.
Methods
This double-blind, placebo-controlled, randomized, multicenter study was approved by the Ministry of Food and Drug Safety, Korea, and the institutional review board at each study center. Written informed consent was obtained from each patient. This trial (Effects of very early use of rosuvastatin in preventing recurrence of ischemic stroke [EUREKA]) was registered on ClinicalTrials.gov (NCT01364220).
Study population
We enrolled patients over 20 years old diagnosed with acute ischemic stroke on diffusion-weighted imaging (DWI) within 48 hours after symptom onset who had been untreated with a statin for the previous 3 months. Patients also should show any degree of stenosis on the relevant artery of infarction on DWI. Patients with hemorrhagic stroke, history of symptomatic hemorrhagic stroke, high-risk cardiac sources of embolism, or stroke of other determined etiology were excluded. Other exclusion criteria are described in Supplementary data.
The modified intention-to-treat population consisted of patients who underwent a baseline magnetic resonance imaging (MRI), had triglyceride (TG) < 500 mg/dL, low-density lipoprotein cholesterol < 190 mg/dL, and took at least one dose of study medication. The per-protocol (PP) population consisted of patients who completed scheduled MRIs without a major protocol violation.
Randomization, blinding, and interventions
Patients were randomized 1:1 to receive either rosuvastatin or placebo. Permutated-block randomization with a block size of 4 was generated by an independent clinical trials center (Severance Hospital, Yonsei University Health System, Seoul, Korea). After patients were screened and completed enrollment, drugs were assigned a unique study number, selected sequentially from the central randomization list that corresponded to the treatment pack, and allocated in a double-blind manner. The drug was administered within 18 hours after baseline MRI and then daily during the 14-day treatment period. Patients received either one 20-mg tablet or a placebo tablet, once daily.
Sample size
We hypothesized that, compared to placebo, rosuvastatin would reduce the occurrence of new ischemic lesions on MRI by 30%. To test our hypothesis, assuming a type I error of 5% and a power of 80%, sample sizes were calculated as 260 in each group. The proportion in the rosuvastatin group was assumed to be 0.40 under the null hypothesis and 0.28 under the alternative hypothesis, based on a previous study that recognized new ischemic lesions on DWI in 34%-47.4% of patients during the first week after baseline DWI taken within 24 hours after symptom onset [13,14]. The test statistic used was the 2-sided Fisher’s exact test. Assuming a drop-out rate of 5%, the total number of patients needed was 547.
Imaging protocol
To be eligible, patients underwent DWI, fluid attenuated inversion recovery, gradient-recalled echo (GRE), and magnetic resonance (MR) angiography that included both the circle of Willis and neck vessels at baseline using a 1.5 T or 3.0 T MR scanner. Computed tomography (CT) angiography was also allowed. The follow-up imaging schedule included DWI, GRE, and fluid attenuated inversion recovery at 5 ± 1 days and 14 ± 2 days using the same MR scanner. The images were saved in Digital Imaging and Communications in Medicine (DICOM) format and sent to the independent clinical trials center for review of adequacy and analyses.
Adjudication of images
Two stroke neurologists blinded to clinical and group information reviewed angiographic images and determined the presence of relevant artery stenosis in ischemic lesions on DWI. In cases of discrepancy between the reviewers, the decision was made by a third reviewer (a neuroradiologist). The reviewers also measured the degree of stenosis of the relevant artery based on methods used in the North American Symptomatic Carotid Endarterectomy Trial for extracranial arteries and the Warfarin and Aspirin for Symptomatic Intracranial Arterial Stenosis trial for intracranial arteries [15,16] and categorized stenosis as ≥ 50% or < 50%.
Outcomes
Imaging outcomes were assessed by 2 reviewers blinded to clinical and group information. The primary outcome was occurrence of a new ischemic lesion on DWI or fluid attenuated inversion recovery at 5 or 14 days. The secondary outcomes were the volume change of ischemic lesions and the percent improvement in National Institute of Health Stroke Scale (NIHSS) at 5 days and 14 days. The safety outcome included adverse events, laboratory results, and the presence of any intracranial hemorrhagic transformation on GRE, which included hemorrhagic infarction (HI) and parenchymal hemorrhage based on European Cooperative Acute Stroke Study definitions [17]. In case of a discrepancy in the presence of new lesions or hemorrhagic transformation, conclusions were reached by consensus. The volume of infarctions was measured on DWI in a semi-automatic manner using Xelis software (Infinitt, Seoul, Korea). The intraclass correlation coefficient between the reviewers for log-transformed volume measurements was 0.99. The percent improvement was defined as ([NIHSS at 5 days or 14 days - NIHSS at baseline]/NIHSS at baseline) × 100 [18].
Details of the conduction of study are provided in Supplemental data.
Statistical analysis
The efficacy outcome was compared based on the modified intention-to-treat and PP population. Safety was assessed in all patients who took at least one dose of study medication. We used the χ2 test with continuity correction to compare the occurrence of newly developed DWI lesion, the independent sample t test to compare the percent improvement in NIHSS, and the mixed-effect model to compare the change in log-transformed DWI lesion volume between the rosuvastatin and placebo groups. The χ2 test or Fisher’s exact test was performed to compare safety outcomes. Statistical analyses were performed using SAS statistical software, version 9.2 (SAS institute Inc., Cary, NC). Data are presented as number (%) or mean ± standard deviation. Two-sided P values < 0.05 were considered statistically significant.
Results
Baseline characteristics
A total of 28 centers in Korea participated in this study, and 25 centers enrolled at least 1 patient. Among 318 patients enrolled, 316 patients met inclusion criteria and were randomized, and 314 took at least one dose of study medication (155 in the rosuvastatin group and 159 in the placebo group). Among them, the primary outcome was assessed in 289 patients (Figure 1). Baseline demographic characteristics were similar between the groups except total cholesterol and low-density lipoprotein cholesterol levels, which were higher in the rosuvastatin group (Table 1, Supplemental Table 1). Characteristics of the patients whose primary outcome was not assessed are provided in as a Supplemental Table 2.
Efficacy outcomes
Efficacy was compared in the modified intention-to-treat population (137 patients in the rosuvastatin group and 152 patients in the placebo group). New ischemic lesions on DWI were observed less frequently in the rosuvastatin group (27 patients [19.7%]) than in the placebo group (36 patients [23.6%]), but the difference was not statistically significant (absolute difference 3.9%, relative risk [RR] 0.83, 95% confidence interval [CI] 0.53-1.30, P=0.500) (Figure 2A). Infarction volumes on DWI increased at 5 days and then decreased at 14 days in both groups. Infarct volume growth at 5 days (log-transformed volume change, rosuvastatin: 0.2 ± 1.0 mm3 vs. placebo: 0.3 ± 1.3 mm3; P=0.784) and percent improvement in NIHSS (rosuvastatin vs. placebo: 36.6 ± 56.7 vs. 27.1 ± 90.8 at 5 days, P=0.282 and 51.4 ± 51.6 vs. 42.7 ± 91.5 at 14 days, P=0.315) were not different.
Safety outcomes
Of 314 patients, 3 in the placebo group (1.9%) and none in the rosuvastatin group demonstrated clinical recurrence of ischemic stroke (P=0.248). Progression or clinical recurrence of stroke was reported as a serious adverse event in 7 patients (4.4%) in the placebo group, but in only 1 patient (0.6%) in the rosuvastatin group (P=0.067). The frequency of adverse events did not differ between the groups (Table 2).
On GRE, HI was observed in 6 patients at baseline (5/155 [3.2%] in the rosuvastatin and 1/159 [0.65] in the placebo groups). Occurrence of any new intracranial hemorrhagic transformation (HI, parenchymal hemorrhage, or subarachnoid hemorrhage) or aggravation of pre-existing HI1 at base-line (defined as conversion to HI2 or parenchymal hemorrhage) was assessed in 289 patients with available GRE at 5 or 14 days. Any new HI was observed less frequently in the rosuvastatin group (6/137, 4.4%) than in the placebo group (22/152, 14.5%) (P=0.007). In the rosuvastatin group, 1 patient developed parenchymal hemorrhage (P=0.478), and 1 patient developed focal cortical subarachnoid hemorrhage on GRE (P=0.478), both of whom were asymptomatic (Table 3).
Post-hoc subgroup analysis
We compared the occurrence of a new ischemic lesion on DWI in patients with relevant artery stenosis ≥ 50%. New ischemic lesions were found in 19 of 83 patients (22.9%) in the rosuvastatin group and 25 of 87 patients (28.7%) in the placebo group (RR, 0.80; 95% CI, 0.48-1.33; P=0.387). Twenty of 78 patients (25.6%) in the rosuvastatin group and 26 of 80 patients (32.5%) in the placebo group with multiple lesions on baseline DWI had new lesions (RR, 0.79; 95% CI, 0.48-1.29; P=0.346) (Figure 2A).
PP Population
A total of 118 patients in the rosuvastatin group and 129 patients in the placebo group were included for the PP population (Figure 1). New ischemic lesions were found in 24 patients (20.3%) in the rosuvastatin group and 34 patients (26.3%) in the placebo group (RR 0.77; 95% CI, 0.49-1.22, P=0.335) (Figure 2B). In the subgroup with relevant artery stenosis ≥ 50%, new ischemic lesions were detected in 17 of 79 patients (21.5%) in the rosuvastatin group and 24 of 78 patients (30.8%) in the placebo group (RR, 0.70; 95% CI 0.41-1.20; P=0.192). In the subgroup with multiple lesions, new ischemic lesions were observed in 17 of 69 (24.8%) patients in the rosuvastatin group and 25 of 69 (36.2%) patients in the placebo group (RR 0.68; 95% CI, 0.41-1.14; P=0.145) (Figure 2B).
Discussion
This study was inconclusive to prove the hypothesis that rosuvastatin may effectively reduce early recurrence of new ischemic lesions, probably due to insufficient sample size. We could not include a sufficient number of patients because of slow enrollment. Despite insufficient evidence that use of a statin is effective or safe in acute ischemic stroke, investigators were reluctant to administer placebo, since statin use on discharge became a performance measure for primary stroke center certification [19]. However, a non-significant trend for less frequent appearance of new ischemic lesions on DWI or progression or clinical recurrence of stroke was observed in the rosuvastatin group. Thus, our findings support previous observational studies showing a benefit of statins in the acute stage of stroke [7,8].
Previous randomized trials using statins in acute stroke included small numbers of patients and showed no clinical effect or even worse outcomes after statin treatment [20-22]. Serial MRIs were used to determine outcomes in this study. Defining clinical recurrence of stroke is sometimes difficult during the acute stage because pre-existing symptoms often fluctuate or progress. Many ischemic lesions that are recognized on DWI during the acute stage of stroke are clinically silent, but they are direct surrogate markers of recurrence. Ischemic injury progresses during the acute stage of infarctions [23] and may cause growth of infarct volume. By using MRI surrogate markers such as DWI and GRE, subclinical occurrence of ischemic and expansion of ischemic lesions as well as hemorrhagic outcomes might be assessed accurately.
In this study, new ischemic lesions in the placebo group developed less often than expectation which was assumed based on previous reports [13,14]. This might be partly ascribed to the high frequency of use with dual or triple antiplatelet agents (about 75%) in this study population. We hypothesized that the use of statins could reduce the risk of early recurrence of either symptomatic or asymptomatic ischemic lesions and expansion of ischemic lesions in acute stroke. This was because antithrombotic and anti-inflammatory effects of statins have been demonstrated in many experimental studies. Statins also play a beneficial role in stabilizing atherosclerotic plaques [6,24]. In the subgroup analysis of our trial, the RR reduction of new DWI lesion occurrence in the rosuvastatin group was greater in patients who were more likely to have had an atherothrombotic infarction, such as a 20% reduction in patients with relevant artery stenosis ≥ 50% and 21% reduction in those with multiple lesions in the territory of relevant artery stenosis. In the PP population (excluding patients without relevant artery stenosis after assessment), the difference was much greater. Although it is possible that statins are more effective in the prevention of stroke with an atherothrombotic mechanism, this hypothesis was inconclusive in our study.
In this study, rosuvastatin 20 mg was safe in that there were no differences in the development of adverse events. Notably, the occurrence of HI on GRE was remarkably reduced in the rosuvastatin group. In the Stroke Prevention by Aggressive Reduction in Cholesterol Levels trial, atorvastatin 80 mg increased the risk of ICH (hazard ratio, 1.68; 95% CI 1.09-2.59) [2], and in the Heart Protection Study, simvastatin 40 mg was associated with approximately 2-fold increased risk of ICH in patients with prior stroke [10]. However, statin use was not associated with ICH in a large cohort study in patients with recent ischemic stroke [25], and in a meta-analysis of primary and secondary prevention studies of statins [26,27]. Our findings suggest that statin use in the acute stage of stroke may protect against microvascular (capillary) damage and prevent HI. Matrix metalloproteinase-9 is a key proteinase that mediates HI by disrupting microvascular integrity [28]. Statins reduce expression of matrix metalloproteinase-9 in endothelial cells, astrocytes, and human plasma [29,30]. Thus, the protective effect of statins against occurrence of HI might be in part mediated by inhibiting matrix metalloproteinase-9.
This study has several limitations. First, this study did not include patients with cardioembolic sources and was conducted in one Asian country. Therefore, our results should be interpreted with caution. Second, the median NIHSS of the study population at screening was 3, which suggests that enrolled patients had rather milder stroke. Although we assessed the outcome in the subgroup with significant stenosis of the relevant artery, this might affect the outcome such as the frequency of progression or recurrence of stroke. Furthermore, the primary outcome of this trial was not clinical stroke recurrence, but imaging-based recurrent ischemic lesions. As a result, the data of clinical stroke recurrence was captured based on the investigators’ reports of adverse events. Finally, this study was stopped early due to slow enrollment, which resulted in underpowered results.
In conclusion, by using MRI surrogate markers such as DWI and GRE, subclinical occurrence of ischemic as well as hemorrhagic outcomes may be assessed very sensitively and accurately with a relatively smaller sample size. Further studies are required to elucidate the potential benefit of statins in acute stroke patients to conclusively support the routine use of statins.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement
Print
