Statin Treatment and Stroke Outcome in the Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) Trial
Background and Purpose— Laboratory experiments suggest statins reduce stroke severity and improve outcomes. The Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) trial was a placebo-controlled, randomized trial designed to determine whether treatment with atorvastatin reduces strokes in subjects with recent stroke or transient ischemic attack (n=4731). We analyzed SPARCL trial data to determine whether treatment favorably shifts the distribution of severities of ischemic cerebrovascular outcomes.
Methods— Severity was assessed with the National Institutes of Health Stroke Scale, Barthel Index, and modified Rankin Scale score at enrollment (1 to 6 months after the index event) and 90 days poststroke in subjects having a stroke during the trial.
Results— Over 4.9 years, strokes occurred in 576 subjects. There were reductions in fatal, severe (modified Rankin Scale score 5 or 4), moderate (modified Rankin Scale score 3 or 2), and mild (modified Rankin Scale score 1 or 0) outcome ischemic strokes and transient ischemic attacks and an increase in the proportion of event-free subjects randomized to atorvastatin (P<0.001 unadjusted and adjusted). Results were similar for all outcome events (ischemic and hemorrhagic, P<0.001 unadjusted and adjusted) with no effect on outcome hemorrhagic stroke severity (P=0.174 unadjusted, P=0.218 adjusted). If the analysis is restricted to those having an outcome ischemic stroke (ie, excluding those having a transient ischemic attack or no event), there was only a trend toward lesser severity with treatment based on the modified Rankin Scale score (P=0.0647) with no difference based on the National Institutes of Health Stroke Scale or Barthel Index.
Conclusion— The present exploratory analysis suggests that the outcome of recurrent ischemic cerebrovascular events might be improved among statin users as compared with nonusers.
The Stroke Prevention with Aggressive Reduction of Cholesterol Level (SPARCL) trial randomized 4731 subjects who had a stroke or transient ischemic attack (TIA) within 1 to 6 months before study entry, a low-density lipoprotein cholesterol (LDL-C) level of 100 to 190 mg/dL, and no known coronary heart disease to double-blind treatment with 80 mg atorvastatin per day or placebo.1 The trial’s primary end point (time to a first fatal or nonfatal stroke) was reduced by 16% in subjects who were randomized to atorvastatin.1
3-hydroxy-3-methylglutaryl coenzyme reductase inhibitors (statins) such as atorvastatin have a variety of actions in addition to their lipid-lowering properties that might not only affect the risk of stroke, but also stroke outcome.2–4⇓⇓ Results of studies evaluating the effects of statins on stroke severity and functional outcome in humans have been inconsistent.5–10⇓⇓⇓⇓⇓ The primary purpose of this exploratory analysis was to determine whether randomization to statin therapy affected the outcome of recurrent ischemic strokes in SPARCL subjects. Additional analyses were conducted to assess the effect of being randomized to statin treatment on the outcome of all recurrent strokes (ischemic plus hemorrhagic) and the outcome of hemorrhagic strokes separately.
The methods of the SPARCL study have been described in detail previously.1,11⇓ Briefly, the local research ethics committee or Institutional Review Board at each center approved the study protocol, and all patients gave written informed consent. Eligible patients were >18 years of age, had an ischemic or hemorrhagic stroke or a TIA 1 to 6 months before randomization, and had no known coronary heart disease. Subjects had to have a modified Rankin Scale (mRS) score of no more than 3 (see subsequently) and an LDL-C level of at least 100 mg/dL and no more than 190 mg/dL (15 of the 205 centers excluded otherwise suitable patients with an LDL-C level >160 mg/dL as required by their Institutional Review Boards). Patients with atrial fibrillation, mechanical prosthetic heart valves, severe mitral valve stenosis, or subarachnoid hemorrhage were excluded. Patients with hemorrhagic stroke (2% of the study population) could be included if they were deemed by the investigator to be at risk for ischemic stroke or coronary heart disease. Subjects were enrolled between September 1998 and March 2001 at 205 centers in 27 countries. An independent committee adjudicated all potential primary and secondary end points without knowledge of the patients’ treatment status or cholesterol levels. The type of outcome stroke was not adjudicated.
The severity of the index stroke (ie, at the time of randomization) was assessed with the National Institutes of Health Stroke Scale (NIHSS; neurological impairments),12 Barthel Index (BI; disability),13 and mRS score (handicap).14 The reliability of the NIHSS has been established in several studies, and all investigators were certified in its use.15,16⇓ A score of 0 reflects no measurable impairment. The BI evaluates independence in activities of daily living. Pivotal scores correspond to severe dependence (score <40) and assisted independence (score >60).17 A score of 100 reflects no disability. The mRS assesses functional outcome with scores ranging from 0 to 6 (0, no symptoms; 1, no significant disability despite symptoms and able to carry out all usual duties and activities; 2, slight disability, unable to carry out all previous activities but able to look after own affairs without assistance; 3, moderate disability requiring some help but able to walk without assistance; 4, moderate severe disability, unable to walk without assistance, and unable to attend to own bodily needs without assistance; 5, severe disability; bedridden, incontinent, and requiring constant nursing care and attention; and 6, death). These measures were repeated 90 days after a recurrent stroke. Because the data come from a secondary prevention trial, subjects could have been evaluated and treated at a variety of nonstudy sites at the time of the recurrent outcome stroke. Therefore, obtaining data on initial severity of the recurrent stroke was not possible. Assessors were unaware of the subject’s treatment group.
The net difference in statin use between groups was 78.1% by the end of the trial (versus 100% had all subjects adhered to their randomized treatment group).1 For the purposes of the present analyses, subjects were assigned to their randomized treatment group regardless of subsequent adherence. We a priori hypothesized that atorvastatin treatment would favorably shift the distribution of ischemic outcome events (ie, a shift along a continuum from fatal to severe, moderate, or mild strokes; TIAs; and no events) based on mRS scores at 90 days after a first outcome event (mRS 5 or 4, severe; mRS 3 or 2, moderate; mRS 1 or 0, mild). Probabilities were calculated from Cochran-Mantel-Haenszel tests with modified RIDIT scores adjusting for age, gender, and baseline severity.
We performed secondary analyses for all first outcome events (ischemic and hemorrhagic) and separately for first hemorrhagic strokes (without TIA as a possible outcome). We also compared the mean and median NIHSS and BI (<60, 60 to 95, ≥95) for subjects surviving a first outcome stroke (ie, excluding those having no events, TIAs, or fatal events) and the mRS (including mortality) for each type of outcome.
In a prior exploratory analysis, we found that 96% of SPARCL subjects reaching a ≥50% decrease in LDL-C from baseline had been randomized to atorvastatin.18 In a further secondary analysis, we assessed the impact of having a ≥50% decrease in LDL-C from baseline at the visit just before an outcome event (or the last LDL-C for those not having an event) on the distribution of outcomes (ie, having an LDL-C reduction ≥50% was used as an indicator of statin use).
Probabilities were calculated from analysis of covariance models adjusting for age, gender, and baseline severity for mean NIHSS; from Wilcoxon tests for median NIHSS; and from Cochran-Mantel-Haenszel tests with modified RIDIT scores adjusted for age, gender, and baseline severity for the BI and mRS.
Details of the characteristics of the SPARCL subject population and the baseline comparability of the groups randomized to atorvastatin and placebo have been published (see referenced report; Table 1).1 Baseline stroke severity data were available for 209 atorvastatin-randomized and 246 placebo-randomized subjects who later had an outcome stroke and were similar for the 2 groups (Table 1).
Over a median follow-up of 4.9 years, 492 subjects had an outcome ischemic stroke (218 in the atorvastatin group versus 274 in the placebo group; hazard ratio, 0.79; 95% CI, 0.66 to 0.95).19 Baseline characteristics were similar for subjects having an outcome ischemic stroke who were randomized to atorvastatin versus placebo (Table 2). The Figure (middle panel) gives the distribution of outcomes for subjects randomized to atorvastatin (data available for 2291 of 2310 subjects who had an ischemic event or no stroke [99.2%]) versus placebo (data available for 2300 of 2333 subjects [98.6%]). There are proportional reductions in fatal, severe, moderate, and mild ischemic strokes and TIAs and an increase in the proportion of event-free subjects randomized to atorvastatin (P<0.001 unadjusted, P<0.001 adjusted).
A total of 265 subjects randomized to atorvastatin and 311 randomized to placebo had an outcome ischemic or hemorrhagic stroke.1 The Figure (top panel) gives the distribution of outcomes for subjects randomized to atorvastatin (data available for 2334 of 2365 subjects who had an outcome event or no stroke [98.7%]) versus placebo (data available for 2333 of 2366 subjects [98.6%]). There are proportional reductions in fatal, severe, moderate, and mild strokes and TIAs and an increase in the proportion of event-free subjects randomized to atorvastatin (P<0.001 unadjusted, P<0.001 adjusted). The results were similar using achievement of a ≥50% reduction in LDL-C from baseline instead of randomization group (data not shown; P<0.039 unadjusted; P<0.013 adjusted).
A total of 55 subjects randomized to atorvastatin and 33 randomized to placebo had an outcome hemorrhagic stroke.1 Although the incidence of hemorrhage was higher in atorvastatin-randomized subjects, there was no difference in outcome between those randomized to atorvastatin (data available for 2030 of 2147 subjects who had a hemorrhagic stroke or no event [94.6%]) versus placebo (data available for 1917 of 2092 subjects [91.6%]; Figure, bottom panel; P=0.218), including no difference in the incidence of fatal hemorrhagic strokes (17 in the atorvastatin- and 18 in the placebo-randomized groups).
Table 3 compares the mean and median NIHSS and BI (<60, 60 to 95, ≥95) for subjects surviving any outcome stroke, ischemic stroke, or hemorrhagic stroke (ie, excluding those having no events, TIAs, or fatal events) and the mRS (including mortality) for each type of outcome. Although there was a trend toward lesser severity with atorvastatin treatment based on the mRS in subjects having an outcome ischemic stroke (P=0.0647), none of the differences were significant.
We found only a trend favoring a reduction in the severity of outcome ischemic strokes with statin treatment based on the categorized mRS with no difference based on the NIHSS or BI. Because this analysis was based on data collected as part of a prospective, randomized trial assessing the impact of a statin on the occurrence of strokes in subjects with a prior stroke or TIA, however, it provides a unique opportunity for an assessment of the potential effects of these drugs on the entire range of possible stroke outcomes (ie, a shift along a continuum from fatal to severe, moderate, or mild strokes; TIAs; and no events). The primary finding of this analysis was that SPARCL subjects randomized to atorvastatin had a favorable shift in the distribution of outcome ischemic events assessed after 90 days as compared with those randomized to placebo (Figure, middle panel). The results were similar for all outcome events (Figure, top panel) and when using achievement of a 50% or greater reduction in LDL-C from baseline at the last assessment before an outcome event as an index of stain adherence.
Observational clinical studies also provide data reflecting the potential impact of statins on ischemic stroke severity and outcome, although the results have been inconsistent. Most evaluated the effects of statins taken before the index stroke and found no clear impact on initial severity.6,7,9⇓⇓ One study reported a reduction in mortality among statin users 1 month after ischemic stroke and over a mean follow-up period of 2.4 years but did not assess severity.8 Another found a reduction in stroke progression,7 a result not found in a second.5 Two of the studies found some improvement in functional outcome,5,6⇓ at least for some measures, whereas another did not.7 A third found that statin pretreatment was associated with better neurological outcomes in whites but not blacks.10 One of the studies also evaluated the impact of poststroke statins on the outcome after a second stroke, finding a functional benefit assessed after 12 weeks.6 These studies were necessarily constrained by their retrospective designs. The subjects took different types of statins at varying doses both within and between studies. This variability and other design issues may underlay the disparate results. In SPARCL, all statin-treated patients were randomized to receive a single statin at a single dose thereby limiting one source of variability. In addition, the previous studies could not evaluate the potential impact of statins on the entire continuum of outcomes because they could not account for subjects receiving statins who had outcome TIAs or no events. As noted, we found no significant effect of being randomized to statin treatment on stroke-related impairments, disabilities, or handicaps if the analysis was restricted to those having an outcome stroke (Table 3).
This analysis was primarily aimed at assessing the potential effect of statin therapy on the outcomes of ischemic events, but a secondary analysis was conducted for outcome hemorrhagic strokes. Experimentally, statin administration after intracerebral hemorrhage reduces acute inflammation and improves neurological outcome.20,21⇓ A retrospective study found that prior statin use was associated with reduced perihematomal edema and improved survival in a series of patients with nontraumatic intracerebral hemorrhage.22,23⇓ Analysis of SPARCL data found that statin treatment was associated with an increased risk of hemorrhagic stroke in subjects with prior stroke that was independent of other factors.1,19⇓ Although limited by small numbers and therefore low statistical power, the current analysis found no difference in the outcomes of hemorrhagic strokes (Figure, bottom panel). Similarly, a retrospective analysis of 90-day outcome data from 629 consecutive patients from a single-center longitudinal cohort study of primary intracerebral hemorrhage found no effect of prior statin use on functional independence or mortality.24 We found that the overall benefit of randomization to atorvastatin on outcome was maintained if the analysis was carried out for all events (ie, ischemic and hemorrhagic events; Figure, top panel).
If statin treatment does reduce ischemic stroke severity, the effect is unlikely to be related to the potent lipid-lowering properties of these drugs. Statins have anti-inflammatory, antithrombotic, and vasoactive effects as well as other actions that might be neuroprotective or affect infarct size, leading to a reduction in the severity of the acute stroke. For example, 2 to 4 days of statin pretreatment protects against glutamate excitotoxicity in primary cortical neuronal cell cultures, independent of the drug’s effects on 3-hydroxy-3-methylglutaryl coenzyme reductase.25 Induction of heat shock proteins enhances survival of axotomized retinal ganglion cells.26 Statins reduce superoxide production and stroke-associated infiltration of inflammatory cells in stroke-prone spontaneously hypertensive mice.27 Statins inhibit nuclear factor κB activity and cytokine gene expression thereby potentially reducing the poststroke inflammatory response.28 Abrupt termination of statin administration results in rapid loss of protection in mouse models of cerebral ischemia further suggesting that the protective effect is independent of cholesterol-lowering.29 Statin pretreatment also upregulates endothelial nitric oxide synthase, increases cerebral blood flow, reduces infarct size, and improves neurological outcome without affecting serum cholesterol levels in mice subjected to experimental stroke, effects absent in endothelial nitric oxide synthase-deficient mice.30 Statin-related upregulation of endothelial nitric oxide synthase also reduces platelet activation independent of the drug’s lipid-lowering properties.31 Furthermore, pre-exposure to statins upregulates endogenous tissue plasminogen activator and enhances clot lysis after experimental embolic stroke.32 In addition, statins promote angiogenesis, potentially improving blood supply to ischemic tissue.4,33⇓ Interestingly, preischemic stroke statin users have more extensive arterial collaterals than nonstatin users.34 More extensive arterial collaterals might also be expected to lead to lower stroke severity, although a functional benefit was not found in this observational study. Despite these effects, as noted previously, none of the observational clinical studies have found lower initial ischemic stroke severity among prestroke statin users as compared with nonusers.
In addition to their potential effects in decreasing initial stroke severity, statins might also affect the brain’s response to injury. Statins induce poststroke neurogenesis in laboratory animals.4 They also increase neurogenesis in the dentate gyrus, reduce delayed neuronal death, and improve spatial learning in rats after traumatic brain injury.35 Our study measured functional outcome 90 days after an outcome stroke. Data reflecting stroke severity within the first days after the outcome stroke were not available. Therefore, it is not possible to determine to what degree an effect of treatment on overall outcome in SPARCL subjects might be due to a reduction in acute stroke severity related to “neuroprotective” properties, enhanced poststroke recovery, or a combination of the 2 effects.
This analysis has several limitations. As noted, not all subjects adhered to their randomized treatments, but the results were not substantially different using LDL-C reduction as an indicator of statin use. We did not have data for the subject’s socioeconomic status, immediate poststroke complications, detailed cognitive function, family support, insurance status, duration of hospitalization, or poststroke rehabilitation, so we could not adjust for these factors, which can affect poststroke functional outcome. The analysis was, however, based on a subjects enrolled in a prospective trial who were randomized to one of 2 groups (ie, statin treatment versus placebo), the independent variable for this analysis. The 2 groups were otherwise comparable (prior publication Table 1,1 current report; Tables 1 and 2⇑), and there is no reason to suspect that they would differ for other characteristics that might affect outcome. Although the primary hypothesis and statistical plan for this analysis were determined a priori, SPARCL was not primarily designed to assess the effects of atorvastatin on stroke severity, and this study should be viewed as exploratory.
The primary results of the SPARCL trial showed that treatment with atorvastatin reduced the rate of recurrent stroke in patients with recent stroke or TIA. The present exploratory analysis needs to be interpreted with caution but suggests that the outcome of recurrent ischemic cerebrovascular events may be improved among statin users as compared with nonusers if the full range of possible outcomes is considered. There is, however, only a nonsignificant trend favoring functional improvement (based on the mRS) if the analysis is limited to those who have a second stroke.
Source of Funding
The SPARCL trial was funded by Pfizer. Employees of Pfizer contributed to the design and conduct of the SPARCL trial; the collection, management, analysis, and interpretation of the data; and reviewed the manuscript.
L.B.G. received honoraria from Pfizer during the course of this study and has served as a consultant for Pfizer. The honoraria and consulting fees did not exceed $10 000/year. P.A. has received grants from Pfizer for other research or activities not reported in this research exceeding $10 000/year and honoraria from Pfizer in excess of $10 000/year during the course of this study. J.Z. received honoraria from Pfizer during the course of this study and has served as a consultant for Pfizer. The honoraria and consulting fees did not exceed $10 000/year. M.M. is an employee of Pfizer and owns Pfizer stock. I.A. served as a site investigator in the SPARCL trial. He owns Pfizer stock and has received honoraria from Pfizer. The amount of honoraria and all other income from Pfizer is less than $10 000/year. A.C. has received research support and honoraria from Pfizer in excess of $10 000/year during the course of this study. M.H. has received grants from Pfizer for other research or activities not reported in this research/article and honoraria from Pfizer during the course of the study. Neither the grants nor the honoraria exceeded $10 000/year. M.J.M. served as a site investigator in the SPARCL trial. She has received honoraria totaling less than $1000 from Pfizer. H.S. has received grants from Pfizer for other research or activities not reported in this research/article in excess of $10 000/year and honoraria exceeding $10 000/year during the course of this study. K.M.A.W. received honoraria from Pfizer during the course of this study and has served as a consultant for Pfizer in excess of $10 000/year.
- Received May 7, 2009.
- Revision received June 5, 2009.
- Accepted June 25, 2009.
- ↵Marti-Fabregas J, Gomis M, Arboix A, Aleu A, Pagonabarraga J, Belvis R, Cocho D, Roquer J, Rodriguez A, Garcia MD, Molina-Porcel L, Diaz-Manera J, Marti-Vilalta JL. Favorable outcome of ischemic stroke in patients pretreated with statins. Stroke. 2004; 35: 1117–1121.
- ↵Moonis M, Kane K, Schwiderski U, Sandage BW, Fisher M. HMG-CoA reductase inhibitors improve acute ischemic stroke outcome. Stroke. 2005; 36: 1298–1300.
- ↵Elkind MS, Flint AC, Sciacca RR, Sacco RL. Lipid-lowering agent use at ischemic stroke onset is associated with decreased mortality. Neurology. 2005; 65: 253–258.
- ↵Bushnell CD, Griffin J, Newby LK, Goldstein LB, Mahaffey KW, Graffagnino CA, Harrington RA, White HD, Simes RJ, Califf RM, Topol EJ, Easton JD. Statin use and sex-specific stroke outcomes in patients with vascular disease. Stroke. 2006; 37: 1427–1431.
- ↵Reeves MJ, Gargano JW, Luo Z, Mullard AJ, Jacobs BS, Majid A. Effect of pretreatment with statins on ischemic stroke outcomes. Stroke. 2008; 39: 1779–1785.
- ↵Brott T, Adams HPJ, Olinger CP, Marler JR, Barsan WG, Biller J, Spilker J, Holleran R, Eberle R, Hertzberg V, Rorick M, Moomaw CJ, Walker M. Measurements of acute cerebral infarction: a clinical examination scale. Stroke. 1989; 20: 864–870.
- ↵Goldstein LB, Samsa GP. Reliability of the National Institutes of Health Stroke Scale: extension to non-neurologists in the context of a clinical trial. Stroke. 1997; 28: 307–310.
- ↵Amarenco P, Goldstein LB, Szarek M, Sillesen H, Rudolph AE, Callahan A, Hennerici M, Simunovic L, Zivin JA, Welch KMA. Effects of intense low-density lipoprotein cholesterol reduction in patients with stroke or transient ischemic attack: the Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) trial. Stroke. 2007; 38: 3198–3204.
- ↵Goldstein LB, Amarenco P, Szarek M, Callahan A, Hennerici M, Sillesen H, Zivin J, Welch KMA. Secondary analysis of hemorrhagic stroke in the Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) Study. Neurology. 2008; 70: 2364–2370.
- ↵Jung KH, Chu K, Jeong SW, Han SY, Lee ST, Kim JY, Kim M, Roh JK. HMG-CoA reductase inhibitor, atorvastatin, promotes sensorimotor recovery, suppressing acute inflammatory reaction after experimental intracerebral hemorrhage. Stroke. 2004; 35: 1744–1749.
- ↵FitzMaurice E, Wendell L, Snider R, Schwab K, Chanderraj R, Kinnecom C, Nandigam K, Rost NS, Viswanathan A, Rosand J, Greenberg SM, Smith EE. Effect of statins on intracerebral hemorrhage outcome and recurrence. Stroke. 2008; 39: 2151–2154.
- ↵Kawashima S, Yamashita T, Miwa Y, Ozaki M, Namiki M, Hirase T, Inoue N, Hirata K, Yokoyama M. HMG-CoA reductase inhibitor has protective effects against stroke events in stroke-prone spontaneously hypertensive rats. Stroke. 2003; 34: 157–163.
- ↵Gertz K, Laufs U, Lindauer U, Nickenig G, Bohm M, Dirnagl U, Endres M. Withdrawal of statin treatment abrogates stroke protection in mice. Stroke. 2003; 34: 551–557.
- ↵Endres M, Laufs U, Huang Z, Nakamura T, Huang P, Moskowitz MA, Liao JK. Stroke protection by 3-hydroxy-3-methylglutaryl (HMG)-CoA reductase inhibitors mediated by endothelial nitric oxide synthase. Proc Natl Acad Sci U S A. 1998; 95: 8880–8885.
- ↵Laufs U, Gertz K, Huang P, Nickenig G, Bohm M, Dirnagl U, Endres M. Atorvastatin upregulates type III nitric oxide synthase in thrombocytes, decreases platelet activation, and protects from cerebral ischemia in normocholesterolemic mice. Stroke. 2000; 31: 2442–2449.
- ↵Ovbiagele B, Saver JL, Starkman S, Kim D, Ali LK, Jahan R, Duckwiler GR, Vinuela F, Pineda S, Liebeskind DS. Statin enhancement of collateralization in acute stroke. Neurology. 2007; 68: 2129–2131.
- ↵Lu D, Qu C, Goussev A, Jiang H, Lu C, Schallert T, Mahmood A, Chen J, Li Y, Chopp M. Statins increase neurogenesis in the dentate gyrus, reduce delayed neuronal death in the hippocampal CA3 region, and improve spatial learning in rat after traumatic brain injury. J Neurotrauma. 2007; 24: 1132–1146.