Author + information
- Received February 16, 2006
- Revision received June 29, 2006
- Accepted July 3, 2006
- Published online November 7, 2006.
- David D. Waters, MD, FACC⁎,†,‡,§∥,¶,#,1,⁎ (, )
- John C. LaRosa, MD⁎,†,‡,§∥,¶,#,2,
- Philip Barter, MD⁎,†,‡,§∥,¶,#,3,
- Jean-Charles Fruchart, PhD⁎,†,‡,§∥,¶,#,4,
- Antonio M. Gotto Jr, MD, DPhil, FACC⁎,†,‡,§∥,¶,#,5,
- Roddy Carter, MD⁎,†,‡,§∥,¶,#,6,
- Andrei Breazna, PhD⁎,†,‡,§∥,¶,#,6,
- John J.P. Kastelein, MD, PhD⁎,†,‡,§∥,¶,#,7 and
- Scott M. Grundy, MD, PhD⁎,†,‡,§∥,¶,#,8
- ↵⁎Reprint requests and correspondence:
Dr. David D. Waters, Division of Cardiology, Room 5G1, San Francisco General Hospital, 1001 Potrero Avenue, San Francisco, California 94110.
Objective We sought to assess the effects on cerebrovascular events of treating patients with stable coronary disease with low-density lipoprotein cholesterol (LDL-C) levels substantially below 100 mg/dl.
Background Lowering LDL-C with statins has been shown to reduce the risk of stroke in patients with stable coronary disease. In observational studies, naturally low cholesterol levels have been associated with an increased risk of hemorrhagic stroke. The cerebrovascular benefits of treating patients with stable coronary disease to LDL-C levels substantially below 100 mg/dl have not been previously investigated.
Methods We describe an analysis of cerebrovascular events in the Treating to New Targets study, a trial where 10,001 patients with documented coronary disease were randomized to treatment with atorvastatin at 10 mg/day or 80 mg/day and followed for a median of 4.9 years.
Results Mean LDL-C levels were 101 mg/dl on 10 mg atorvastatin and 77 mg/dl on 80 mg. In addition to the reduction in major cardiovascular events (hazard ratio 0.78, 95% confidence interval [CI] 0.69 to 0.89; p = 0.0002), the primary end point of the trial, patients in the 80-mg arm experienced a reduction in cerebrovascular events (hazard ratio 0.77, 95% CI 0.64 to 0.93; p = 0.007) and stroke (hazard ratio 0.75, 95% CI 0.59 to 0.96; p = 0.02). Each 1-mg/dl reduction in LDL-C with treatment was associated with a 0.6% relative risk reduction in cerebrovascular events (p = 0.002) and a 0.5% relative risk reduction in stroke (p = 0.041). The incidence of hemorrhagic stroke was similar in the 80-mg and 10-mg groups, 16 and 18 respectively, and the hemorrhagic strokes were distributed evenly across quintiles of achieved LDL-C during treatment.
Conclusions Among patients with established coronary disease, treating to an LDL-cholesterol substantially below 100 mg/dl with 80 mg/day atorvastatin reduces both stroke and cerebrovascular events by an additional 20% to 25% compared with the 10 mg/day dose. An increase in hemorrhagic stroke was not seen at low LDL-C levels. (Treating to New Targets; http://www.clinicaltrials.gov; NCT00327691).
The incidence of cerebrovascular events in Western countries appears to be at least as high as the incidence of coronary events (1). Furthermore, the long-term outcome after even a minor ischemic stroke or transient ischemic attack (TIA) is bleak; for example, in a recent report of a large cohort study from the Netherlands, 60% of patients died and 54% had had at least one other vascular event over a mean follow-up period of 10 years (2).
The relationship between blood cholesterol levels and stroke is not as clear as the relationship between cholesterol and coronary events; in fact, in a meta-analysis of 45 prospective observational studies involving 450,000 individuals, no association was found between cholesterol and stroke rate (3). Differences emerge when strokes are classified as to type. An increased rate of hemorrhagic stroke has been reported in subjects with naturally low cholesterol levels (4–7), but the incidence of ischemic stroke is increased at higher cholesterol levels in the same populations (7,8).
Although earlier clinical trials did not show a reduction in stroke with cholesterol lowering (9), statins have reduced stroke or cerebrovascular events in randomized placebo-controlled trials in a variety of different patient populations (10–16). Whether statins reduce stroke by LDL-C reduction or primarily through other mechanisms is controversial (17–19).
The TNT (Treating to New Targets) study randomized 10,001 patients with documented coronary disease to atorvastatin at 10 mg/day or 80 mg/day and followed them for a median of 4.9 years (20). The primary end point of TNT, a composite of coronary death, nonfatal myocardial infarction (MI), resuscitated cardiac arrest, and fatal or nonfatal stroke, was reduced in the 80-mg compared with the 10-mg arm (hazard ratio 0.78, 95% confidence interval [CI] 0.69 to 0.89; p = 0.0002). In addition to fatal or nonfatal stroke being a component of the predefined primary end point of the trial, a cerebrovascular event, defined as fatal or nonfatal stroke or TIA, was a predefined secondary end point. Herein we report in detail the results of TNT with respect to stroke and cerebrovascular events.
The design of the TNT study has been previously described in detail (20,21). Eligible patients were men and women 35 to 75 years of age who had clinical evidence of coronary disease, either previous MI or coronary revascularization, or previous or current angina with objective evidence of coronary disease. Patients with LDL-C levels between 130 and 250 mg/dl and triglycerides ≤600 mg/dl off lipid-lowering treatment were eligible to enter an 8-week run-in period where they received 10 mg/day atorvastatin. Those with an LDL-C level <130 mg/dl were randomly assigned to double-blind therapy with either 10 mg or 80 mg atorvastatin per day. Follow-up visits occurred at week 12 and at months 6, 9, and 12, then every 6 months thereafter. Patients were questioned at each visit about interim events or hospitalizations, and hospital records were obtained.
Study end points
The primary predefined outcome was the occurrence of a major cardiovascular event, defined as death from coronary heart disease, nonfatal nonprocedure-related MI, resuscitated cardiac arrest, or fatal or nonfatal stroke. Secondary outcomes were a major coronary event (the primary end point without stroke), a cerebrovascular event, hospitalization for congestive heart failure, peripheral artery disease, death from any cause, any cardiovascular event, and any coronary event.
An independent end points committee consisting of 5 cardiologists adjudicated all potential end points in a blinded fashion. They defined stroke as the rapid onset of a persistent brain deficit thought to be due to obstruction or rupture in the arterial system and not secondary to brain trauma, tumor, or infection. The deficit must have lasted longer than 24 h unless death intervened or unless there was a persistently demonstrable lesion that was consistent with the deficit. Patients admitted with symptoms of developing stroke who were given tissue plasminogen activator (TPA) with subsequent resolution of symptoms were also classified as stroke. Strokes occurring during surgery were also included.
Stroke were categorized as ischemic, hemorrhagic, embolic, or unknown, using the criteria developed for the Systolic Hypertension in the Elderly Program (22). Strokes were classified as ischemic when one of the following criteria was met: 1) a focal brain deficit without evidence by computerized tomography (CT), magnetic resonance imaging (MRI), or lumbar puncture of blood, with the exception of a mottled cerebral pattern, either decreased density by CT in a compatible location, or a negative CT or MRI, or none done; 2) surgical or autopsy evidence of ischemic infarction; 3) symptoms of stroke followed by TPA with resolution of symptoms; or 4) symptoms lasting <24 h but with a CT or MRI showing an infarct or hemorrhage in a location to explain the findings.
Strokes were classified as hemorrhagic when there was: 1) blood in the subarachnoid space or intraparenchymal hemorrhage by CT or MRI scan; 2) bloody spinal fluid by lumbar puncture; or 3) surgical or autopsy evidence of hemorrhage as the cause of the clinical syndrome. Strokes were classified as embolic when that was the likely diagnosis (e.g., in the presence of atrial fibrillation). Strokes were classified as unknown when the available data were insufficient for classification into one of the other categories. A TIA was defined as an event that persisted for <24 h with a documented focal neurologic deficit; the symptoms had to be typical, such as transient unilateral blindness, a speech deficit, or transient hemiparesis.
Cholesterol inclusion/exclusion criteria were selected to achieve an average level of 100 mg/dl in the 10 mg/day atorvastatin arm. To reach an average LDL-C level in the comparator group of approximately 75 mg/dl, 80 mg/day atorvastatin was chosen. All analyses were performed on an intention-to-treat basis. End points were analyzed from the time of the first dose of study drug to the first event according to the Kaplan-Meier method. Differences between the 80-mg and 10-mg treatment groups were based on log rank analyses of the first occurrence of an end point during the 4.9-year follow-up period in each group. Relative risks, hazard ratios, and their 95% CIs were calculated in univariate and multivariate models using the Cox regression with one or more predictors in the same model. Note that Cox regression models allow for the predictor to be continuous variables (like on-treatment LDL-C) or dichotomous. Proportions were compared using the Fisher exact test.
A total of 18,469 patients were screened at 256 sites in 14 countries; of these, 15,464 patients entered the open-label run-in period and 10,001 were randomized to double-blind treatment with either 10 or 80 mg/day atorvastatin, as previously described (20,21). Of the 8,468 patients who did not enter the randomized phase of the study, 7,639 had exclusion criteria, including 96 with elevated hepatic enzymes >1.5 times the upper limit of normal while taking 10 mg/day atorvastatin and 727 with LDL-C levels >130 mg/dl or triglycerides >600 mg/dl. Among the other 827 patients who were screened but not randomized, 16 died and 195 had ischemic events during the run-in period; 197 had other adverse events, including myalgia in 35. The remaining 419 patients either did not comply with treatment or were eliminated for a variety of administrative reasons. Two patients who were randomized but never given drug were also excluded. Patients were followed for a median of 4.9 years after randomization, during which time 38 patients in the 10-mg group and 35 in the 80-mg group were lost to follow-up.
The baseline features of the 2 treatment groups were well balanced, and are listed in Table 1.Of note, 5.2% of the patients had a history of a cerebrovascular accident, blood pressure was well controlled, and 86.5% were receiving aspirin.
Changes in lipid levels and blood pressure
During the run-in period on 10 mg/day atorvastatin, LDL-C levels decreased by 35% from a mean of 152 to a mean of 98 mg/dl. Mean LDL-C levels during the study were 101 mg/dl in the 10-mg arm of the trial and 77 mg/dl in the 80-mg arm, and total cholesterol levels were 178 mg/dl and 150 mg/dl, respectively. Mean high-density lipoprotein cholesterol levels were 47 mg/dl in both treatment arms. Mean triglyceride levels were 156 mg/dl in the 10-mg arm and 132 mg/dl in the 80-mg arm.
At baseline, mean systolic and diastolic blood pressures were identical in the 2 treatment groups: systolic 131 ± 17 mm Hg and diastolic 78 ± 10 mm Hg. At last follow-up, systolic blood pressure had decreased by 1.2 ± 15.4 mm Hg in the 10-mg arm and by 1.2 ± 15.1 mm Hg in the 80-mg arm; diastolic pressure decreases were 0.7 ± 9.3 mm Hg in both groups. Blood pressures were very similar in the 2 treatment groups throughout the study.
During the follow-up period, 272 patients (2.7%) experienced a stroke, including 30 who experienced a fatal stroke. A TIA occurred in 197 patients, and a total of 448 (4.5%) experienced a cerebrovascular event (stroke or TIA). As shown in Table 2,stroke was reduced in the 80-mg compared with the 10-mg group (hazard ratio 0.75, 95% CI 0.59 to 0.86; p = 0.021). Cerebrovascular events were also reduced in the 80-mg arm (hazard ratio 0.77, 95% CI 0.64 to 0.93; p = 0.007). The Kaplan-Meier curves for these end points are depicted in Figure 1.Among patients without a history of cerebrovascular disease (who comprised 95% of the study population), fewer strokes occurred in the 80-mg compared with the 10-mg group: 101 versus 128 (2.1% vs. 2.8%, hazard ratio 0.79, 95% CI 0.61 to 1.02; p = 0.070). Fewer cerebrovascular events also occurred at the higher dose among patients without a history of cerebrovascular disease: 171 versus 212 (3.6% vs. 4.5%, hazard ratio 0.80, 95% CI 0.66 to 0.98; p = 0.032).
The types of stroke in each treatment group are listed in Table 3.Overall, 54% of the strokes were ischemic, 25% were embolic, 12% were hemorrhagic, and in 9% the cause was unknown. For each type of stroke the incidence was lower in the 80-mg group than in the 10-mg group.
Predictors of cerebrovascular events
The factors that were associated with a first cerebrovascular event are listed in Table 4.The factors predictive of stroke (not shown) are similar to those predictive of cerebrovascular events. The LDL-C level at the screening visit was not a predictor of cerebrovascular events during the study, but LDL-C level on treatment at 3 months was. Each 1-mg/dl reduction in LDL-C was associated with a 0.6% relative risk reduction in cerebrovascular events (p = 0.002) and a 0.5% relative risk reduction in stroke (p = 0.041).
Events according to attained LDL-C level
Patients in both treatment groups were divided into quintiles according to their LDL-C levels at 3 months, as depicted in Table 5.Patients in the lowest quintile had LDL-C levels of <64 mg/dl, with a mean of 54 mg/dl. A stepwise reduction in major cardiovascular events was seen from the highest to the lowest quintile of from 11.9% to 7.7% (p < 0.0001). A similar pattern was seen for cerebrovascular events (p = 0.002) and stroke (p = 0.041). The number of hemorrhagic strokes in each quintile, from lowest to highest, was 6, 5, 6, 9, and 7 (p = NS).
Mortality rate in patients with cerebrovascular events
Among the 518 patients with a history of cerebrovascular disease at baseline, the mortality rate during the 4.9 years of the study was 11.0%, compared with 5.4% in the 9,483 patients without a history of cerebrovascular disease at baseline (p = 0.0001). Among the 448 patients experiencing a cerebrovascular event during the follow-up period, 69 of them (15.4%) subsequently died.
During the run-in period, with 15,464 patients taking 10 mg/day atorvastatin, the drug was stopped because of myalgia in 35 cases and because of alanine transaminase (ALT) and/or aspartate aminotransferase (AST) elevations >1.5 times the upper limit of normal in 96 cases. During the 4.9 years of study follow-up, treatment-related adverse events occurred in 406 patients in the 80-mg group and in 289 patients in the 10-mg group (8.1% vs. 5.8%; p < 0.001). Myalgia was reported in 241 patients in the 80-mg arm and in 234 patients in the 10-mg arm (4.8% vs. 4.7%; p = 0.72). Consecutive ALT and/or AST elevations >3 times the upper limit of normal were seen in 60 patients in the 80-mg arm and in 9 patients in the 10-mg arm (1.2% vs. 0.2%; p < 0.001). No patient experienced consecutive creatine kinase elevations >10 times the upper limit of normal.
This study provides evidence that reducing LDL-C levels with 80 mg/day atorvastatin to levels well below 100 mg/dl in patients with stable coronary disease reduces the incidence of stroke and all cerebrovascular events. The relative risk reductions in cerebrovascular events and coronary events are similar. An increase in the incidence of hemorrhagic stroke was not seen either at the high dose of atorvastatin or at the lowest achieved LDL-C levels during treatment.
Randomized, placebo-controlled statin trials have shown a reduction in cerebrovascular events as well as coronary events in patients with stable coronary disease (10–12) as well as in patients with unstable coronary disease (13), hypertension (14), high coronary risk (15), and diabetes (16). The Cholesterol Treatment Trial collaborators have recently reported a prospective meta-analysis of 90,056 patients from 14 randomized statin trials (23). Overall, fatal or nonfatal stroke was reduced by 17% (95% CI 12% to 22%). There was a trend toward greater proportional reductions in stroke with greater LDL-C reduction (p = 0.009).
The TNT study differs from previous statin trials because 2 doses of the same statin were compared and because a lower LDL-C level, a mean of 77 mg/dl, was maintained over a median of 4.9 years in the 80-mg atorvastatin arm. The TNT study demonstrates that additional benefit for stroke and other important vascular end points can be achieved with the highest dose of a potent statin, reducing LDL-C well below the traditional target of 100 mg/dl.
The TNT study does not address the issue of whether statin therapy reduces the risk of cerebrovascular events in patients with previous cerebrovascular events but without known coronary disease. The SPARCL (Stroke Prevention by Aggressive Reduction in Cholesterol Levels) study, comparing placebo to 80 mg/day atorvastatin in 4,731 patients with a previous stroke or TIA, recently demonstrated that the high dose of atorvastatin significantly reduced both stroke and coronary events in this population (24).
Mechanism of benefit
The reductions in stroke and cerebrovascular events in the 80-mg arm were not due to a reduction in blood pressure, because both systolic and diastolic pressures were identical in the treatment arms throughout the study. However, each 1-mg/dl change in on-treatment LDL-C was associated with a 0.6% change in the risk of a cerebrovascular event (p = 0.002) and a 0.5% change in the risk of stroke (p = 0.041). Expressed in terms of treatment groups, LDL-C was 24% lower in the 80-mg arm, stroke was 25% lower, and cerebrovascular events were 23% lower. The relationship of LDL-C changes to stroke is similar to what has been reported in meta-analyses of placebo-controlled statin trials (23).
Cholesterol lowering with statins has been shown to stabilize carotid artery plaques in patients undergoing carotid endarterectomy (25) and to slow the progression of carotid atherosclerosis as assessed by B-mode ultrasound (26). Most ischemic strokes are caused by thromboemboli from the carotid arteries or the aortic arch. Statins may interfere with the thrombotic component of ischemic stroke by inducing beneficial effects on the coagulation cascade (27) and by correcting the increased thrombogenic potential of platelets that is associated with hyperlipidemia (28).
Statins also induce an impressive variety of beneficial effects that are independent of cholesterol lowering. A limitation of many studies in this area is the difficulty in distinguishing between causation and just a statistical correlation. In the cerebral endothelium and parenchyma, statins regulate nitric oxide synthase to reduce platelet and leukocyte adhesion and activation, increase vasodilation and reduce postischemic permeability (18). Statins also exert antiinflammatory and antioxidant effects that are neuroprotective in cerebral ischemia and stroke (18). In different rodent models statins have been shown to reduce the severity of stroke through these mechanisms (18).
Low cholesterol and hemorrhagic stroke
Naturally low blood cholesterol levels have been associated with an increased risk of hemorrhagic stroke in different geographic populations (4–7). In the Multiple Risk Factor Intervention Trial, a study of 350,977 middle-aged men with no history of MI or diabetes at baseline, death from intracranial hemorrhage was 3-fold higher in those with total cholesterol levels below 160 mg/dl at baseline (7). Most of the patients in the 80-mg arm of the TNT study, where the mean total cholesterol throughout the study was 150 mg/dl (compared with 178 mg/dl in the 10 mg arm), would fall into this low-cholesterol high-risk group. An increased risk of hemorrhagic stroke was not seen, however, either in the 80-mg arm or in the lowest quintile of on-treatment LDL-C (<64 mg/dl).
The relative risk of hemorrhagic stroke for statin-treated versus placebo patients in the Cholesterol Treatment Trialists Collaborators meta-analysis is 1.05 (95% CI 0.78 to 1.41) (23). In the SPARCL study (24), high-dose atorvastatin reduced stroke overall; however, hemorrhagic stroke, albeit uncommon, occurred more frequently in the atorvastatin group (p = 0.02). Taken together, these data suggest that it is unlikely that high-dose statins increase the risk of hemorrhagic stroke in most subsets of patients. Whether some patients with previous stroke are at increased risk for hemorrhagic stroke with statin therapy requires further study.
The criteria used to define stroke and the categories of stroke selected by the end points committee differ from the generally accepted criteria, and the end points committee did not include a neurologist. The findings of this study apply to patients with stable coronary disease. The study population was overwhelmingly Caucasian, and <20% were women. Extrapolation of these results to other populations should be done with caution. Although the incidence of hemorrhagic stroke was not increased at lower LDL-C levels, the number of these events is not sufficient to detect a small increase in risk.
Despite these limitations, the results of TNT demonstrate that lowering LDL-C considerably below 100 mg/dl using 80 mg/day atorvastatin provides additional clinical benefit, including a reduction in stroke and cerebrovascular events. The reduction in cerebrovascular events and overall stroke, without an observed increase in hemorrhagic stroke, provides an additional reason to treat coronary patients more aggressively.
↵1 Dr. Waters has received consulting fees from Anthera, CSL, and Pfizer, lecture fees from Merck, Pfizer, and Novartis, and grant support from Johnson & Johnson and Merck.
↵2 Dr. LaRosa has received consulting fees from Pfizer, Merck, Bristol-Myers Squibb, and AstraZeneca and lecture fees from Pfizer.
↵3 Dr. Fruchart has received consulting fees from Pfizer and Fournier and lecture fees from Merck, Fournier, Pierre Fabrie, and AstraZeneca.
↵4 Dr. Kastelein has received consulting fees, lecture fees, and grant support from Pfizer, Merck, Schering Plough, AstraZeneca, Bristol-Myers Squibb, and Sankyo.
↵5 Dr. Barter has received consulting fees from Pfizer, AstraZeneca, and Sanofi-Aventis, lecture fees from Pfizer, AstraZeneca, Fournier-Pharma, and Sanofi-Aventis, and grant support from Pfizer.
↵6 Dr. Gotto has received consulting fees from AstraZeneca, Bristol-Myers Squibb, Merck, Schering Plough, Pfizer, Novartis, and Reliant and lecture fees from AstraZeneca, Merck, Schering Plough, Pfizer, and Reliant and has testified before the FDA on behalf of Johnson & Johnson–Merck.
↵7 Drs. Carter and Breazna are Pfizer employees.
↵8 Dr. Grundy has received consultation fees from Merck and Merck Schering Plough, Pfizer, Kos Pharmaceutical, Abbott, Sanofi-Aventis, Lilly, and AstraZeneca and grant support from Kos Pharmaceutical, Abbott, and Merck.
Pfizer funded the TNT Study.
- Abbreviations and Acronyms
- low-density lipoprotein cholesterol
- myocardial infarction
- transient ischemic attack
- Treating to New Targets
- Received February 16, 2006.
- Revision received June 29, 2006.
- Accepted July 3, 2006.
- American College of Cardiology Foundation
- Kagan A.,
- Popper J.S.,
- Rhoads G.G.,
- Yano K.
- Iribarren C.,
- Jacobs D.R.,
- Sadler M.,
- Claxton A.J.,
- Sidney S.
- Benfante R.,
- Yano K.,
- Hwang L.J.,
- Curb J.D.,
- Kagan A.,
- Ross W.
- Plehn J.F.,
- Davis B.R.,
- Sacks F.M.,
- et al.,
- Care Investigators
- Waters D.D.,
- Schwartz G.G.,
- Olsson A.G.,
- et al.
- Sever P.S.,
- Dahlof B.,
- Poulter N.R.,
- et al.
- Vaughan C.J.,
- Delanty N.
- Di Napoli P.,
- Taccardi A.A.,
- Oliver M.,
- De Caterina R.
- Waters D.D.,
- Guyton J.R.,
- Herrington D.M.,
- McGowan M.P.,
- Wenger N.K.,
- Shear C.
- Crisby M.,
- Nordin-Fredriksson G.,
- Shah P.K.,
- Yano J.,
- Zhu J.,
- Nilsson J.
- Furberg C.D.,
- Adams H.P. Jr..,
- Applegate W.B.,
- et al.,
- Asymptomatic Carotid Artery Progression Study (ACAPS) Research Group
- Undas A.,
- Brummel K.E.,
- Musial J.,
- Mann K.G.,
- Szczeklik A.
- Lacoste L.,
- Lam J.Y.,
- Hung J.,
- Letchacovski G.,
- Solymoss C.B.,
- Waters D.