Reduction in Early Stroke Risk in Carotid Stenosis With Transient Ischemic Attack Associated With Statin Treatment
Background and Purpose—Statins reduce stroke risk when initiated months after transient ischemic attack (TIA)/stroke and reduce early vascular events in acute coronary syndromes, possibly via pleiotropic plaque stabilization. Few data exist on acute statin use in TIA. We aimed to determine whether statin pretreatment at TIA onset modified early stroke risk in carotid stenosis.
Methods—We analyzed data from 2770 patients with TIA from 11 centers, 387 with ipsilateral carotid stenosis. ABCD2 score, abnormal diffusion weighted imaging, medication pretreatment, and early stroke were recorded.
Results—In patients with carotid stenosis, 7-day stroke risk was 8.3% (95% confidence interval [CI], 5.7–11.1) compared with 2.7% (CI, 2.0%–3.4%) without stenosis (P<0.0001; 90-day risks 17.8% and 5.7% [P<0.0001]). Among carotid stenosis patients, nonprocedural 7-day stroke risk was 3.8% (CI, 1.2%–9.7%) with statin treatment at TIA onset, compared with 13.2% (CI, 8.5%–19.8%) in those not statin pretreated (P=0.01; 90-day risks 8.9% versus 20.8% [P=0.01]). Statin pretreatment was associated with reduced stroke risk in patients with carotid stenosis (odds ratio for 90-day stroke, 0.37; CI, 0.17–0.82) but not nonstenosis patients (odds ratio, 1.3; CI, 0.8–2.24; P for interaction, 0.008). On multivariable logistic regression, the association remained after adjustment for ABCD2 score, smoking, antiplatelet treatment, recent TIA, and diffusion weighted imaging hyperintensity (adjusted P for interaction, 0.054).
Conclusions—In acute symptomatic carotid stenosis, statin pretreatment was associated with reduced stroke risk, consistent with findings from randomized trials in acute coronary syndromes. These data support the hypothesis that statins started acutely after TIA symptom onset may also be beneficial to prevent early stroke. Randomized trials addressing this question are required.
Patients with transient ischemic attack (TIA) and carotid stenosis are at high risk of early stroke, independently of abnormal acute diffusion weighted imaging (DWI), ABCD2 score, and vascular risk factors, with highest risk observed in the first days after symptom onset and in patients with hemispheric TIA, greater lumen stenosis, and ulcerated plaque on angiography.1–3 In randomized trials, carotid endarterectomy (CEA) is highly beneficial for secondary stroke prevention, with maximum benefit observed in those who underwent surgery within 2 weeks of symptom onset.4 However, despite greatest recurrent stroke risk within the first days after initial symptoms, the safety of very early CEA remains to be established. For example, in the Swedish Vascular Registry the combined rate of stroke and death in patients who underwent CEA within 48 hours was 11.5%, with a 4-fold increase in the odds of poor outcome compared with 3 to 7 days.5 Available data indicate that about half of recently symptomatic patients do not undergo CEA, and only a minority have CEA within the recommended 14-day period.6–9 For these reasons, improved medical treatments are needed within the first days after symptoms for patients before revascularization and in those not selected for revascularization.
In randomized trials, statins begun within days of acute coronary syndromes were beneficial for secondary vascular event prevention as early as 30 days, possibly because plaque stabilization independently of lipid lowering.10,11 Although statins have proven benefit for late stroke prevention when begun months after TIA or stroke,12 few data exist in patients with unstable carotid stenosis on the effects on early stroke recurrence of statin pretreatment at the time of symptom onset or statins begun acutely after symptoms. We hypothesized that statin pretreatment at TIA onset would be associated with reduced early stroke risk in patients with TIA with carotid stenosis, including those awaiting CEA. In the absence of randomized trial data, we investigated this hypothesis in an international multicenter pooled data set of individual patients with TIA.
Centers with patients with TIA large cohorts were identified from published studies and were invited to include data. Eleven centers from Europe, Asia, and North America contributed data from 2770 individual patients. Study settings were hospital-based stroke specialist services (stroke specialist treated hospitalized and TIA clinic patients; 9 centers), and population-based settings (2 centers; specialist and nonspecialist care). Cohorts were recruited and prospectively followed for recurrent stroke at each center.
Inclusion Criteria and Definitions
Prespecified inclusion criteria were as follows: (1) TIA verified by a stroke specialist, (2) data available on presence/absence of ipsilateral carotid stenosis, (3) medication data at time of TIA onset available, (4) outcome nonprocedural stroke data available at 7 or 90 days. Patients were excluded if an alternative diagnosis other than TIA was reached, if periprocedural stroke occurred after CEA/stenting, or if medical attention and brain imaging was first sought for a stroke recurrence rather than the index TIA. Medication treatment at TIA onset (ie, prescribed before the onset of TIA symptoms) and post-TIA (including continuing treatment with pre-TIA medications, and newly prescribed medications) was recorded on the basis of data obtained at each individual site.
Data were abstracted from existing TIA registries at each individual center using a standardized electronic template, deidentified, and collated centrally. As the study was observational, patient treatment at each center was at the discretion of the treating clinician.
Standardized definitions of all variables were applied by all centers as previously described.3 Because the American Stroke Association proposed that tissue-based definition was not used uniformly in the United States, Europe, and Asia, and because all included cohorts had applied the traditional time-based definition, TIA was defined clinically as an acute loss of focal cerebral or ocular function lasting <24 hours, attributed to embolic or thrombotic vascular disease. The index TIA for study inclusion was defined as that most recently preceding stroke specialist assessment. For standardization and generalizability, stroke was defined as a new neurological deficit according to the World Health Organization definition, which occurred after complete resolution of symptoms of the preceding TIA.
ABCD2 score was trichotomized into low- (0–3), medium- (4–5), and high-risk (6–7) categories. Carotid stenosis was defined as ≥50% nonocclusive narrowing of the internal carotid artery lumen on carotid imaging (duplex ultrasound, computerized tomography, magnetic resonance, or invasive angiogram), as interpreted by the reporting physician using the North American Symptomatic Carotid Endarterectomy Trial method. Dual TIA was defined as the occurrence of ≥1 earlier TIA within 7 days of presentation with the index TIA. DWI hyperintensity (DWI<72 hours of TIA) was defined as lesion(s) consistent with acute cerebral ischemia determined by the treating neuroradiologist and stroke physician at each center, supported by apparent diffusion coefficient and fluid attenuated inversion recovery/T2 sequences.
Stroke status at 7 and 90 days was determined at each site by in-person assessment, and telephone interview and medical file review. Periprocedural stroke was defined as stroke occurring within 48 hours after the day of carotid revascularization (CEA or stenting).
Ethics committee approval was provided per local site procedures, and patients provided informed consent for participation in research into stroke prevention after TIA.
Statistical analysis was performed using Stata 9.0. Parametric and nonparametric comparisons of categorical and continuous variables were made using χ2, Fisher exact, t test, and Mann–Whitney tests, as appropriate. All significance tests were 2 sided. Analyses for interaction between statins at TIA onset, carotid stenosis, and early stroke were performed by inclusion of interaction terms in logistic regression models. Forward stepwise multivariable logistic regression was performed to adjust for other independent variables associated with outcome at the P<0.05 level on univariate analysis. A case deletion strategy was applied for patients where key data variables were unavailable. Data from patients with periprocedural stroke were excluded from analysis and all outcomes reported describe nonprocedure-related stroke.
Of 2770 patients with TIA from 11 centers, carotid imaging data were available in 98.2% (2721 patients), 387 (14.2%) of whom had ipsilateral carotid stenosis. Of the 387 patients with carotid stenosis, statin pretreatment data were available in 68% (262 patients). Of these, follow-up was complete in 95% (249 patients) at 7 days, and 94% (245 patients) at 90 days.
Patients with stenosis were more frequently men (P=0.04), and were older, with higher prevalence of hypertension, hyperlipidemia, prior stroke, recent TIA, coronary disease, and diabetes mellitus (P<0.0001 for all; Table 1). In the patients with carotid stenosis, 39.1% underwent carotid revascularization (35.8% CEA; 3.3% carotid artery stenting; median presentation-revascularization interval 7 days [interquartile range, 4–14 days]). Of the 387 patients with carotid stenosis, statin data at TIA onset were available in 262 (Figure 1; Table 2; Table I in the online-only Data Supplement). In these, 43.5% were prescribed a statin before the index TIA, rising to 87.7% after the index TIA. Antiplatelet medication was highly associated with statin treatment (P<0.0001). Despite higher rates of preventive medications in carotid stenosis patients, nonprocedural 7-day stroke risk was 8.3% compared with 2.7% in patients without stenosis (P<0.001; 17.8% versus 5.7% at 90 days; P<0.001; Table 1). Carotid revascularization was more common in patients with acute DWI hyperintensity (P=0.01), but was not associated with other vascular risk factors (Table II in the online-only Data Supplement). No interaction was observed among statin pretreatment, carotid revascularization, and early stroke risk (P for interaction, 0.8).
Early Stroke Risk Factors Stratified by Carotid Stenosis
Among patients with carotid stenosis, nonprocedural 7-day stroke risk was 3.8% (4/105) in those pretreated with statins compared with 13.2% (19/144) in nonstatin pretreated patients (P=0.01). Corresponding risks at 90 days were 8.9% (statin treated, 9/101) and 20.8% (statin untreated, 30/144; P=0.01; Figure 2). In patients with TIA with carotid stenosis, the odds ratio of stroke at 7 days associated with statin pretreatment was 0.26 (95% confidence interval, 0.09–0.79), indicating reduced stroke risk. The odds ratio for 90-day stroke was 0.37 (confidence interval, 0.17–0.82). No such benefit was observed in nonstenosis TIA (Figure 3).
A significant interaction was observed between statin pretreatment at TIA onset and carotid stenosis for early recurrent stroke (P for interaction 0.03 for 7-day stroke, 0.008 for 90-day stroke; Figure 3; Table III in the online-only Data Supplement). Smoking at symptom onset was associated with higher risk of 7-day stroke in patients with carotid stenosis (P for interaction, 0.04) with a protective influence of antiplatelet treatment at TIA onset (P for interaction, 0.02; Figure 3). Higher stroke risk was observed with higher ABCD2 score, acute DWI hyperintensity, and dual TIA in carotid stenosis and nonstenosis patients, and with atrial fibrillation in nonstenosis patients only (Figure 3; Table III in the online-only Data Supplement).
On multivariable logistic regression analysis (adjusting for ABCD2 score, smoking, and antiplatelet treatment), the statin/carotid stenosis interaction remained significant for early stroke (adjusted P for interaction, 0.01 for stroke at 90 days; n=1339). When DWI hyperintensity and dual TIA were added to the model, the statin/carotid stenosis adjusted P value for interaction was 0.054 (n=438 because of unavailable DWI or dual TIA data in 901 patients).
Few randomized trials investigating the benefit of acute statin therapy for early stroke prevention after TIA have been performed and none have selected patients with recently symptomatic carotid stenosis deemed unsuitable for, or who are awaiting, carotid revascularization. In the absence of gold-standard randomized data, we examined early stroke risk in patients with carotid stenosis who were on statin treatment at TIA onset (ie, statin pretreatment) in our large observational TIA database. We sought to provide data which might support the rationale for randomized trials of acute statin therapy begun after symptoms have occurred.
Our main finding was of a substantial reduction in nonprocedural early stroke risk in patients with carotid stenosis who were pretreated with statins at the time of TIA onset. In contrast, no such risk reduction associated with statin pretreatment was observed in patients with TIA without carotid stenosis. A significant interaction was observed between the protective association with statins for early recurrent stroke in patients with carotid stenosis, which remained after adjustment for confounding variables.
Evidence indicates that an unmet need exists for improved treatment to prevent early recurrent stroke in patients with recently symptomatic carotid stenosis. First, in population studies, carotid stenosis is associated with 3-fold increase in early stroke recurrence risk compared with other subtypes.6,13 When combined with acute DWI hyperintensity, recent earlier TIA (dual TIA), and ABCD2 score in the ABCD3-I score, the addition of carotid stenosis significantly improved the c-statistic for early stroke discrimination after TIA, compared with the ABCD2 score.3
Second, although recurrent stroke risk after symptomatic carotid stenosis is highest in the first days, the safety of very early CEA remains unclear. In a large study from the Swedish Vascular Registry, the combined stroke and death rate after CEA within 2 days of symptoms was 11.5%, greatly exceeding procedural event rates associated with benefit in randomized trials.5 On multivariate analysis, time was an independent risk factor, with a 4-fold increase in adjusted odds ratio for perioperative complications for CEA within 48 hours compared with 3 to 7 days. Supporting these data, a recent meta-analysis has reported a pooled odds ratio of stroke and death of 4.6 with emergency CEA in unstable patients (stroke-in-evolution and crescendo TIA).14
Third, in clinical practice many patients with symptomatic carotid stenosis are not selected for carotid revascularization because of disabling stroke or other factors resulting in a perceived unfavorable risk–benefit assessment by treating clinicians. In the Canadian Stroke Network Registry, only 17.5% of patients with symptomatic carotid stenosis had CEA within 6 months.7 In the Californian Kaiser-Permanente network, only 36% of women and 54% of men with TIA and severe carotid stenosis had CEA.8 Similar rates have been reported from population studies in Ireland and United Kingdom.6,9 In our pooled analysis, the rate of carotid revascularization was 39.1%, consistent with earlier reports. CEA was performed more frequently in patients with TIA with acute DWI hyperintensity, suggesting that early DWI findings may influence physicians’ perception of surgical risk–benefit. For patients who do not have CEA or in those for whom revascularization is delayed, improved acute medical treatments are needed.
Inflammation is a central biological process involved in atherosclerotic plaque rupture and thromboembolic stroke. In the Pravastatin Or Atorvastatin Evaluation and Infection Therapy (PROVE-IT) and Myocardial Ischemia Reduction with Acute Cholesterol Lowering randomized trials in patients with acute coronary syndromes, early statin treatment (within 10 and 4 days of symptoms) was associated with reduced rates of early recurrent vascular events, possibly mediated by stabilization of symptomatic plaque.10,11 In PROVE-IT, benefit of intensive statin therapy (atorvastatin 80 mg) was observed as early as 15 days, becoming statistically significant at 30 days, despite coronary revascularization in 70% of included patients.10
In cerebrovascular disease, no large studies have investigated the benefit of early statins in patients with acutely symptomatic carotid atherosclerosis. A positron-emission tomography study showed that inflammation-related carotid plaque flurodeoxyglucose uptake predicted early stroke recurrence, independently of stenosis severity.15 Others have reported dose-dependent reduction of carotid flurodeoxyglucose uptake with statins. Statins may also improve cerebral vasomotor reactivity and collateralization.16 In the Fast Assessment of Stroke and Transient ischaemic attack to prevent Early Recurrence (FASTER) trial, patients with TIA/minor stroke were randomized to clopidogrel/placebo and simvastatin/placebo.17 However, FASTER was discontinued because of suboptimal recruitment and no clear benefit of simvastatin was apparent. The nonrandomized SOS-TIA, Effect of urgent treatment of transient ischaemic attack to prevent early recurrence (EXPRESS), and Transient ischaemic attack work-up as Outpatient Assessment of Clinical Evaluation and Safety (TWO-ACES) studies have reported reduced early stroke risk with intensive acute treatment, including statins in patients with all TIA mechanisms.18–20 The SPARCL trial demonstrated the benefit of atorvastatin begun weeks or months after TIA/stroke in patients with carotid stenosis for reduction of late recurrent stroke and coronary events.12,21
Strengths of our analysis include the availability of DWI and outcome data in a large patient sample assessed acutely after TIA. We adjusted for known confounders and measured the clinically important outcome of recurrent stroke in the early phase after TIA, when stroke risk is known to be highest. We acknowledge several limitations. Although data were available on whether patients were treated with statins at discharge from hospital evaluation for their TIA, the time interval between TIA onset and prescribing of new statin was unavailable. Therefore, we were unable to distinguish patients who received emergency statin treatment immediately (within 24 hours) after TIA onset from those that received new statin treatment days or weeks later. Statin dose was unavailable, precluding analysis of the relationship of dose–response on early stroke events. Therefore, we conducted our primary analysis in patients prescribed statins before and at the time of TIA onset. Low stroke event rates (reflecting early stroke specialist treatment) limited statistical power for multivariable analysis within the first week after TIA. As pre-event statin data were unavailable for some patients and centers, we cannot exclude the possibility that this may have influenced our findings. Patients with carotid stenosis not only were more likely to be treated with statins, antiplatelet, or antihypertensive therapy at TIA onset but also were older and had higher rates of risk factors for recurrent stroke (acute DWI hyperintensity, diabetes mellitus, and greater ABCD2 score). Thus, it is possible that residual confounding by these factors may have contributed to our findings. Also, we cannot exclude the possibility that statin pretreatment at TIA onset may have been a marker for unmeasured factors associated with lower risk of early stroke (confounding by indication), or that residual confounding may exist from plaque-related factors (such as degree of stenosis or plaque morphology) or other unmeasured variables. However, our findings remained after adjusting for other significant predictors of early stroke after TIA, including smoking, ABCD2 score, and antiplatelet treatment.
Current clinical trials are investigating emergency combination antiplatelet treatment after TIA, and carotid revascularization compared with best medical therapy in symptomatic patients stratified by the Oxford risk prediction score. We emphasize that we do not advocate delaying CEA or replacing CEA with statin therapy in patients eligible for revascularization. Neither does our article provide data to support routine emergency treatment with statins in TIA. Rather, our study provides preliminary hypothesis-generating data suggesting that emergency treatment with statins may further reduce stroke risk in symptomatic patients with carotid disease after TIA, who are particularly high risk of recurrent stroke. Future randomized trials should investigate this question.
Dr Kelly is the recipient of a Clinician Scientist Award from the Health Research Board (HRB) of Ireland. Á. Merwick was supported by the HRB. Funding was also provided by the Irish Heart Foundation, and National Lottery. P.M. Rothwell and M. Giles received funding from the National Institute of Health Research (NIHR) and the Oxford Vascular Study is funded by the Stroke Association, Medical Research Council, Dunhill Medical Trust and the NIHR Biomedical Research Center, Oxford. Drs Coutts and Demchuk and the vascular imaging of acute stroke for identifying predictors of clinical outcome and recurrent ischemic events (VISION) study were supported by grant funding from the Canadian Institutes for Health Research (MOP-118096) and Heart and Stroke Foundation of Alberta, Northwest Territories and Nunavut, Canada Foundation for Innovation, and Alberta Foundation for Health Research. Dr Coutts received the Heart and Stroke Foundation of Canada Distinguished Clinician Scientist award, supported by the Canadian Institutes of Health Research Institute of Circulatory and Respiratory Health and AstraZeneca Canada Inc, and salary support from the Alberta Heritage Foundation for Medical Research. Dr Purroy was supported by Ministerio de Sanidad y Consumo, Instituto de Salud Carlos III: FIS Number PIO81398. Dr Ay was supported by National Institutes of Health (NIH) grant ROI-N5059710. Dr Tsivgoulis has been supported by European Regional Development Fund - Project FNUSA-ICRC (No. CZ.1.05/1.1.00/02.0123), St. Annes University Hospital in Brno – International Clinical Research Centre. Dr Cucchiara was supported by an American Heart Association Fellow-to-Faculty Award and NIH grant ROI- NS61572. Dr Sharma is recipient of Clinician Scientist Award from National Medical Research Council, Singapore. Á. Merwick has received unrestricted educational grants from Lundbeck, Pfizer, and Boehringer Ingelheim to attend conferences. Dr Kelly has received speakers’ honoraria from Boehringer Ingelheim, Pfizer, and Bristol Myer Squibb. Pfizer and Servier have provided unrestricted educational grants to the Neurovascular Unit for Translational and Therapeutics Research toward stroke education and research. Dr Saver has been an advisory board member of AGA Medical, Boehringer Ingelheim, Bristol Myer Squibb, CoAxia, Fibro Gen, and Sanofi Aventis, and he has performed consultancy work for ImaRx and received honoraria from Concentric Medical. Dr Coutts has been an advisory board member for Bristol Myer Squibb. Dr Mas has been an advisory board member of Bayer, Boehringer Ingelheim, Bristol Myer Squibb, Daiichi-Sankyo, and Sanofi Aventis and has received speaker fees from AstraZeneca, Pfizer, and Takeda. The other authors report no conflicts.
Guest Editor for this article was Bo Norrving, MD.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.113.001576/-/DC1.
- Received April 8, 2013.
- Revision received June 27, 2013.
- Accepted July 2, 2013.
- © 2013 American Heart Association, Inc.
- Rerkasem K,
- Rothwell PM
- Strömberg S,
- Gelin J,
- Osterberg T,
- Bergström GM,
- Karlström L,
- Osterberg K
- Sheehan O,
- Kyne L,
- Kelly LA,
- Hannon N,
- Marnane M,
- Merwick A,
- et al
- Gladstone DJ,
- Oh J,
- Fang J,
- Lindsay P,
- Tu JV,
- Silver FL,
- et al
- Poisson SN,
- Johnston SC,
- Sidney S,
- Klingman JG,
- Nguyen-Huynh MN
- Rerkasem K,
- Rothwell PM
- Rothwell PM,
- Giles MF,
- Chandratheva A,
- Marquardt L,
- Geraghty O,
- Redgrave JN,
- et al
- Olivot JM,
- Wolford C,
- Castle J,
- Mlynash M,
- Schwartz NE,
- Lansberg MG,
- et al
- Amarenco P,
- Benavente O,
- Goldstein LB,
- Callahan A III.,
- Sillesen H,
- Hennerici MG,
- et al