Statin Therapy After Acute Ischemic Stroke in the Heart Protection Study
Is the Role in Recurrent Stroke Prevention Now Defined?
The Heart Protection Study (HPS) recently reported that for individuals at high risk for experiencing vascular occlusive events, simvastatin taken 40 mg/d reduces the relative risk of ischemic stroke by 28% (P<0.001) without consequent increased risk of hemorrhagic stroke.1 The results of this study are important and should impact practice because HPS is a well-designed, large, prospective, randomized clinical trial with prespecified and adjudicated endpoints. HPS enrolled men and women between the ages of 40 and 80 years with high risk of coronary heart disease, treated hypertension, diabetes mellitus, or occlusive noncoronary artery disease to receive either simvastatin 40 mg per day or placebo. Inclusion criteria required a total cholesterol concentration of at least 135 mg/dL (3.5 mL/L).2 Subjects with a history of nondisabling stroke or transient ischemic attack or carotid endarterectomy were also eligible for enrollment. However, excluded were stroke, myocardial infarction, or hospital admission for angina within 6 months of randomization, as well as hepatic disease, severe renal disease, inflammatory muscle diseases, and a variety of other life-threatening conditions or contraindications for the use of statins. The design was a 2×2 factorial using simvastatin 40 mg daily with a matched placebo or antioxidant vitamins with matching placebo capsules. For statin therapy, the study was powered to detect a reduction of 25% in 5-year coronary heart disease mortality, along with a 15% to 20% reduction in all-cause mortality (>90% power, P<0.01). Prespecified secondary analyses included the effects of simvastatin therapy on 10 noncoronary causes of death including stroke and hemorrhagic stroke.
In the total cohort, HPS found that individuals randomized to receive simvastatin had a robust overall 24% (95% CI, 19 to 28; P<0.0001) relative risk reduction in nonfatal myocardial infarction, coronary death, stroke of any type, or revascularization. Of these endpoints, first nonfatal or fatal stroke was reduced by 25% (95% CI, 15 to 34; P<0.0001) in the simvastatin group, with the greatest benefit accruing for a reduction in ischemic stroke with a relative risk reduction of 30%. This positive effect on the reduction of stroke began within the first year of treatment with ≈20% reduction and continued throughout the course of the trial. The absolute risk reduction for all types of stroke was a modest 1.4% (4.3% simvastatin versus 5.7% placebo).
Although the results indicate a very significant reduction in stroke events for the entire group of 10 269 individuals who were randomized to receive simvastatin, in a secondary analysis, which was not prespecified, of subjects who were entered into the trial with previous stroke or transient ischemic attack, there appeared to be a surprising lack of beneficial treatment effect for stroke prevention by simvastatin. Individuals in this subgroup were typical for a population at risk for stroke and on average were older and more often hypertensive but had less coronary heart disease than others in the trial. Therefore, this deviant result in the “cerebrovascular disease” group is even more surprising because the same subgroup experienced essentially an identical reduction in other endpoints including major coronary events and revascularizations as all others receiving simvastatin. The authors suggest that this apparent discrepant result may be explained simply by a “play of chance,” and although such a statistical assertion is reasonable, other factors should also be considered. As emphasized by the HPS investigators, the analysis was not prespecified and the analysis combined ischemic, hemorrhagic, and nonclassified types of stroke. Because there is no evidence that statins reduce the incidence of hemorrhagic stroke, the inclusion of brain bleeds would bias the results toward lack of efficacy. Data supporting this possibility are provided by the analysis of the treatment effect on stroke subtypes. There were 22 (100 versus 122) fewer ischemic strokes and twice as many hemorrhages (21 versus 11) in the cerebrovascular disease subgroup receiving simvastatin. In this group, the reduction in ischemic stroke neared significance with a 19% (SE, 12; P=0.1) proportional reduction that was not statistically different from the highly significant 34% (SE, 7; P<0.0001) proportional reduction in ischemic stroke observed in other high-risk individuals in HPS treated with simvastatin. Furthermore, the absolute risk reduction that simvastatin treatment afforded for ischemic stroke was greater in the cerebrovascular group (absolute RR=1.4%) than that among other individuals (absolute RR=1.1%) at high risk. The magnitude of the absolute RR for stroke is approximately the same as observed in Scandinavian Simvastatin Survival Study (4S) and Cholesterol And Recurrent Events (CARE) trials (1.6% and 1.2%, respectively) and is appreciably larger than that reported in 5 other trials.3 Although the issue of a potential contribution to brain hemorrhages by statin drug therapy remains open, the group with no previous cerebrovascular disease randomized to simvastatin had fewer hemorrhagic events than the placebo group (30 [0.03%] simvastatin versus 42 [0.5%] placebo) and overall in HPS the number of hemorrhages was small (0.5%). Finally, in addition to the relatively small numbers of stroke events available for secondary analysis, fully 20% of adjudicated stroke endpoints could not be classified as either ischemic or hemorrhagic types. This is unfortunate because an increase in the total number of subjects with defined stroke types in the cerebrovascular group would influence the statistical power.
Despite the impressive reduction in vascular endpoints reported by HPS, ultimately, the trial does not answer the important question about the immediate and near-term benefits or lack thereof of statin treatment for individuals with recent ischemic stroke. In HPS, subjects with stroke within 6 months of enrollment were specifically excluded from the trial and the average duration between cerebrovascular symptoms and enrollment for subjects with stroke or transient ischemic attack was 4.3 years. Recurrent stroke incidence declines with time after the initial stroke to approach the population background rate by 3 to 5 years, whereas the risk of myocardial infarction and related cardiovascular events gradually increases with time.4 The hiatus in entry in HPS is long after the peak incidence of recurrent stroke and probably accounts for the relatively low annual stroke rate of <2% per year that was observed in the cerebrovascular group. An analysis of the principal first endpoint from 4 major antiplatelet trials showed that over a mean follow-up of up to 3.25 years after enrollment, recurrent stroke occurred on average 4 times more frequently than myocardial infarction.5 Thus, the major insight provided by HPS is the powerful beneficial effect of statin therapy for prevention of vascular endpoints of all types, including stroke in subjects at high risk for such events. As acknowledged by the authors, further clarification of the role of statin treatment for recurrent stroke prevention in individuals with recent ischemic stroke awaits the results of an ongoing clinical trial.6
Section Editors: Marc Fisher, MD, and Antoni Dávalos, MD
- Received June 23, 2004.
- Accepted June 23, 2004.
MRC/BHF Heart Protection Study Colloraborative Group. MRC/BHF Heart Protection Study of cholesterol-lowering therapy and of antioxidant vitamin supplementation in a wide range of patients at increased risk of coronary heart disease death: early safety and efficacy experience. Eur Heart J. 1999; 20: 725–741.
Hankey GJ, Jamrozik K, Broadhurst RJ, Forbes S, Burvill PW, Anderson CS, Stewart-Wynne EG. Five-year survival after first-ever stroke and related prognostic factors in the Perth Community Stroke Study. Stroke. 2000; 31: 2080–2086.
Albers GW. Choice of endpoints in antiplatelet trials. Neurology. 2000; 54: 1022–1025.