Although higher plasma cholesterol concentrations have not been reported to be associated with increased stroke risk, cholesterol lowering has been reported to decrease this risk. This decrease can be achieved with statins, which are well-tolerated, provided they are not given to patients with active liver or muscle diseases. Statin treatment in addition to a healthy lifestyle is recommended for the primary prevention of ischaemic stroke in patients with pre-existing coronary heart disease or other high-risk conditions such as diabetes and hypertension. Statins with intensive lipid-lowering effects are recommended for their positive influence on reducing the risks of stroke and cardiovascular events for patients with prior ischaemic stroke or TIA presumed to be of non-cardioembolic origin, even with an LDL-C level =100 mg/dL, with or without evidence of other clinical atherosclerotic cardiovascular diseases. Despite the good safety profile of statins, further studies are clarify safety in patients with prior cerebral haemorrhage and if they may increase brain haemorrhage to a small degree. PCSK9 inhibitors are advised, as add-on therapy to statins, for patients with a high cardiac risk not able to achieve an optimal LDL-C level, though studies with longer follow-up are needed
A Review of Observational Epidemiological Studies
Whether increased serum cholesterol levels are risk factors for stroke remains controversial. A systematic review of 45 observational studies over 16 years including around 450,000 individuals reported no significant correlation between total plasma cholesterol and any stroke (after adjusting for age, gender, ethnicity, blood pressure, and history of cardiac disease), suggesting that cholesterol is not a risk factor for stroke (Prospective Studies Collaboration, 1995). Furthermore, a meta-analysis of individual data from 61 prospective studies, most of which were carried out in the USA and Europe, reported no independent positive association between total cholesterol and ischaemic or total stroke mortality (Lewington et al., 2007). Finally, a review of 14 Japanese cohort studies, including subject pools ranging from 1621 to 19,219 (mean follow-up period ranged from 7.6 to 32 years), reported no association between hypercholesterolaemia and total stroke (Tanaka and Tomonori, 2012).
Even though an association between total cholesterol and all strokes has not been confirmed, it could be that total cholesterol and all stroke dilutes associations between cholesterol and pathological subtypes of stroke and between cholesterol fractions and pathological and aetiological subtypes of stroke.
Subtypes of Stroke
A weak (positive) association between increasing plasma cholesterol concentrations and increasing risk of ischaemic stroke has been reported, which is partially offset by a weaker (negative) association between decreasing plasma cholesterol concentrations and an increasing risk of haemorrhagic stroke in both Western and Asiatic populations (Nagasawa et al., 2012; Zhang et al., 2012; Wang et al., 2013).
Increasing low-density lipoprotein-cholesterol (LDL-C) concentrations of 1.03 mmol/L (40 mg/dL) have been associated independently with a 14% (95% confidence interval [CI]: 0–26%) increase in the odds of ischaemic stroke or transient ischaemic attack (TIA) (Koran-Morag et al., 2002).
A significant, independent (negative) association between decreasing plasma high-density lipoprotein (HDL) concentrations and increasing risk of ischaemic stroke has been identified in several studies (Leppala et al., 1999; Koran-Morag et al., 2002; Pikula et al., 2015; Orozco-Beltran et al., 2017).
A significant, independent, positive association has been reported between the ratio of total cholesterol/HDL-C concentrations and ischaemic stroke (Simons et al., 2001).
Apolipoproteins make up the protein moiety of lipoproteins, with apolipoprotein B (apo B) being found mainly in LDL and apolipoprotein A1 (apo A1) in HDL. The assessment of apo B and apo A1 levels provides information on the total number of atherogenic (apo B) and antiatherogenic (apo A1) particles.
A study has suggested that apo B and the ratio of apo B/apo A1 are better predictors of ischaemic stroke than total cholesterol, LDL-C or HDL-C (hazard ratio [HR] 2.27; 95% CI: 1.14–4.50 and HR 2.86; 95% CI: 1.37–5.88, respectively) (Bhatia et al., 2004). Additionally, a collaborative analysis of 79,036 individuals reported that lipoprotein-associated phospholipase A2 (Lp-PLA2) activity and mass are associated with the risk of ischaemic stroke (Lp-PLA2 Studies Collaboration, 2010).
Increasing LDL-C concentrations have been associated independently with a 68% (95% CI: 23–230%) increase in odds of ischaemic stroke, due to large artery atherothrombosis (Koran-Morag et al., 2002). Also lacunar stroke, but not cardioembolic types, is associated with an increase in LDL-C levels (Imamura et al., 2009; Bezerra et al., 2012).
A Systematic Review of Randomized Controlled Trials
A systematic review and meta-analysis of 58 randomized controlled trials (RCTs) investigating cholesterol lowering by any means reported that lowering the concentration of LDL-C by 1.0 mmol/L decreased the risk of all stroke by 10%, and lowering cholesterol by 1.8 mmol/L decreased the risk of stroke by 17% (95% CI: 9–25%) and the risk of coronary events by 61% (95% CI: 51–71%) (Law et al., 2003).
Non-pharmacological (e.g. diets) and non-statin lipid-lowering interventions are less effective in reducing plasma cholesterol concentrations, compared with statins (3-hydroxy-3-methylglutaryl coenzyme A [HMG-CoA] reductase inhibitors). In fact, the mean reduction in cholesterol concentrations among trials of non-statin interventions were reported to be half as great when compared with trials of statin drugs (Di Mascio et al., 2000). Indeed, a significant decrease in stroke incidence has only been observed in trials on statin drugs.
Among the five trials to date investigating the benefit of non-pharmacological interventions in lowering plasma lipid concentrations (diet, ileal bypass, LDL apheresis) in 2198 patients (1102 in intervention group and 1096 in control group), total cholesterol was reduced by 14.5% (standard deviation [SD]: 3.7) the odds of stroke was reduced by 28% (95%CI: -17% to +56%), from 3.6% (40/1096) in the control group to 2.6% (29/1102) in the intervention group P for heterogeneity 0.7 (Di Mascio et al., 2000).
In a meta-regression analysis including 26,969 participants treated with non-statin therapies that act via upregulation of LDL receptor expression to reduce LDL-C (diet, bile acid sequestrants, ileal bypass, and ezetimibe), the risk ratio (RR) for major vascular events (a composite of cardiovascular death, acute myocardial infarction [MI] or other acute coronary syndrome, coronary revascularization, or stroke) per 1 mmol/L reduction in LDL-C level was 0.77 (95% CI: 0.75–0.79; P = 0.01). No significant heterogeneity was seen (Silverman et al., 2016).
Reduction of saturated fat intake is known to reduce serum cholesterol levels, therein reducing the risk of MI, while the benefits on stroke are less clear (any stroke, RR 1.00; 95% CI: 0.89–1.12; 8 trials with 50,952 participants) (Hopper et al., 2015).
Among the 12 trials of non-statin lipid-lowering drugs (clofibrate, niacin, colestipol, cholestyramine, gemfibrozil, probucol) in 27,519 patients (12,143 in intervention group, 15,376 in control group) published before 1998, total cholesterol was reported to be reduced by 12.6% (SD: 5.9) and stroke reduced by 21%; 1.76% (270/15,376) with control and 1.39% (169/12,143) with intervention (Di Mascio et al., 2000).
Since 1998, the Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial (VAHIT) has reported that, compared with placebo, men with ischaemic heart disease allocated to gemfibrozil 1200 mg for 5 years, experienced no change in LDL-C levels (2.87 mmol/L), an increase in HDL levels by 6% (0.82–0.85 mmol/L), a reduction in triglyceride levels by 31% (1.81–1.25 mmol/L), and a reduction in the incidence of stroke (as designated by the investigators), from 6.9% (88/1267) to 5.1% (64/1264); relative risk reduction (RRR): 27% (95% CI: 1–47, P = 0.04) (Rubins et al., 1999). However, the rate of stroke (a secondary outcome event in this trial), as confirmed by the blinded adjudication committee of three neurologists, was not significantly different: 6% (76/1267) for patients allocated placebo versus 4.6% (58/1264) for patients allocated gemfibrozil; RRR: 24% (95% CI: −7% to +45%, P = 0.1). Furthermore, the BIP trial reported that patients with known coronary heart disease (CHD) allocated to bezafibrate, compared with placebo, had a 6.5% reduction (3.82 to 3.6 mmol/L) in LDL-C levels, a fall in triglyceride levels of 21% (1.63 to 1.29 mmol/L), an increase in HDL levels of 18% (0.89 to 1.03 mmol/L), but no significant difference in the rate of all stroke (placebo: 5.0% [77/1542], bezafibrate: 4.6% [72/1548]; RRR: 7% [95% CI: −27% to +32%, P = 0.66]), or ischaemic stroke (placebo: 4.5% [69/1542], bezafibrate: 3.8% [59/1548]; RRR: 15% [95% CI: −20% to +39%, P = 0.36]) (BIP Study Group, 2000).
Three classes of agents targeted at increasing HDL levels included in a meta-analysis (niacin, fibrates, and cholesteryl ester transfer protein [CETP] inhibitors), were not associated with a significantly reduced risk of stroke in both the pre-statin era and the present era (HPS2-THRIVE Collaborative Group, 2014; Keene et al., 2014). A meta-analysis that included 39,195 participants concluded that niacin does not reduce overall mortality (RR 1.05; 95% CI: 0.97–1.12) and the number of fatal or non-fatal strokes (RR 0.93; 95% CI: 0.74–1.22). Participants randomized to niacin were more likely to discontinue treatment due to side effects than participants randomized to control group (RR 2.17; 95% CI: 1.70–2.77) (Schandelmaier et al., 2017). A randomized trial involving 30,449 adults with atherosclerotic vascular disease (6781 with history of cerebrovascular atherosclerotic disease), who were receiving intensive atorvastatin therapy and who had a mean LDL-C level of 61 mg/dL, showed that the primary outcome (composite of coronary death, MI, or coronary revascularization) occurred in significantly fewer patients in the anacetrapib (a CE inhibitor) group than in the placebo group (RR 0.91; 95% CI: 0.85–0.97; P = 0.004). There were no significant between-group differences in the risk of death, cancer, or other serious adverse events. However, the risk of presumed ischaemic stroke (secondary endpoint) was similar between the two groups (RR 0.99; 95% CI: 0.87–1.12) (HPS3/TIMI55–REVEAL Collaborative Group, 2017).
Trials have compared ezetimibe, a lipid-lowering agent that inhibits intestinal absorption of dietary cholesterol plus a lipid-lowering drug, with ezetimibe; reporting a non-significant impact in reducing the incidence of stroke (RR 0.65; 95% CI: 0.08–4.59). Furthermore, ezetimibe plus simvastatin versus simvastatin alone was not associated with a reduction in stroke (RR 2.38; 95% CI: 0.46–12.35) (Battaggia et al., 2015). Conversely, a meta-analysis including 7 trials, enrolling 31,048 patients, reported that ezetimibe versus placebo or ezetimibe plus another hypolipidaemic agent versus the same hypolipidaemic drug alone significantly reduced the risk of any stroke by 16.0% (RR 0.840, 95% CI: 0.744–0.949; P = 0.005), without any effect on all-cause and cardiovascular mortalities (RR 1.003, 95% CI: 0.954–1.055; P = 0.908; RR 0.958, 95% CI: 0.879–1.044; P = 0.330; respectively) as well as the risk of new cancer (RR 1.040, 95% CI: 0.965–1.120; P = 0.303) (Savarese et al., 2015). Another meta-analysis including 23,499 participants showed that adding ezetimibe to statins probably reduces the risk of non-fatal MI (RR 0.88; 95% CI: 0.81–0.95) and non-fatal stroke (RR 0.83; 95% CI: 0.71–0.97), but trials reporting all-cause mortality used ezetimibe with statins or fenofibrate found they have no effect on this outcome (RR 0.98; 95% CI: 0.91–1.05) (Zhan et al., 2018).
Until 1998, there were 25 secondary prevention trials including 33,000 patients (14,979 in the intervention group, 18,237 in the control group) (Di Mascio et al., 2000). Total cholesterol was reduced by 18.1% (SD: 6.6), and the odds of stroke was reduced by 20% (95% CI: 9–29%), from 3.53% (643/18,237) in the control group to 2.86% (428/14,979) in the intervention group (Di Mascio et al., 2000).
A later systematic review of lipid-lowering interventions for preventing stroke recurrence reported that among 627 patients with a history of stroke or TIA randomized in two trials (Acheson and Hutchinson, 1972; The Veterans Administration Cooperative Study Group, 1973) to clofibrate (n = 315) or control (n = 312), the odds of a recurrent stroke were non-significantly increased among patients allocated to clofibrate (14.4% control, 19.0% clofibrate; odds ratio [OR] 1.48, 95% CI: 0.94–2.30) (Manktelow et al., 2002; Wang et al., 2015).
Numerous statin trials, including patients with known CHD and primary prevention trials including high-risk populations, have declared a decrease in stroke incidence in patients treated with statins (Paciaroni et al., 2007). The Stroke Prevention by Aggressive Reduction of Cholesterol Levels (SPARCL) study evidenced the positive effects of statin therapy also in the secondary prevention of cerebrovascular diseases (SPARCL Investigators, 2006).
Statins and Their Benefits in Stroke Prevention for Both CHD and High-risk Vascular Disease Patients (mainly diabetic and hypertensive patients without CHD)
The Cholesterol Treatment Trialists’ Collaborators reported on the results of a prospective meta-analysis examining data from 90,056 individuals in 14 randomized trials on the benefits from statin use (Cholesterol Treatment Trialists’ [CTT] Collaboration, 2005): 42,131 (47%) had pre-existing CHD, 21,575 (24%) were women, 18,686 (21%) had a history of diabetes, and 49,689 (55%) had a history of hypertension. Regarding cerebrovascular events, data were available for a total of 2,957 first-ever strokes after randomization. Overall, 2282 strokes were reported among 65,138 patients in the 9 trials that collected information on stroke type: 204 (9%) haemorrhagic, 1565 (69%) ischaemic, and 513 (22%) unknown type. A significant 17% proportional reduction in the incidence of first stroke of any type (1340 [3.0%] was reported in the statin group versus 1617 [3.7%] for the control group: RR 0.83, 95% CI: 0.78–0.88; P < 0.0001) per 1 mmol/L LDL-C reduction. This overall reduction in stroke was associated with highly significant 19% proportional reduction (RR 0.81, 99% CI: 0.74–0.89; P < 0.0001) in stroke not attributed to haemorrhage (i.e. presumed ischaemic) per 1 mmol/L LDL-C reduction, and no apparent difference in haemorrhagic stroke. The total reduction in presumed ischaemic stroke was associated with a significant 22% proportional reduction in confirmed ischaemic stroke (RR 0.78, 99% CI: 0.70–0.87; P < 0.0001) per 1 mmol/L LDL-C reduction and a 12% proportional reduction in stroke of unknown type (RR 0.88, 99% CI: 0.7–1.02; P = 0.03). There was no significant reduction in stroke during the first year after randomization (RR 0.96, 99% CI: 0.79–1.17; P = 0.6), yet there were significant reductions ranging from 20–25% over the 3 subsequent years, and thereafter favourable trends were recorded. During an average 5-year treatment period, the reduction in the overall incidence of stroke was roughly one-sixth for each 1 mmol/L LDL-C decrease, predicting 8/1000 fewer subjects would have any stroke in the pre-existing CHD disease group at baseline, compared with 5/1000 fewer subjects with no such history.
The beneficial effect of statins in stroke prevention for high-risk patients, with or without CHD, has been confirmed also by other meta-analyses. In a Cochrane review including 18 randomized control trials (19 trial arms; 56,934 subjects; 14 trials recruited patients with specific conditions such as elevated lipid levels, diabetes, hypertension, and microalbuminuria), all-cause mortality was reduced by statins (OR 0.86, 95% CI: 0.79–0.94), as was combined fatal and non-fatal stroke (RR 0.78, 95% CI: 0.68–0.89) (Taylor et al., 2013). Another meta-analysis of randomized trials on statins, in combination with other preventive strategies, included 165,792 subjects, reporting that each 1 mmol/L (39 mg/dL) decrease in LDL-C equated to a reduction in relative risk for stroke of 21.1% (95% CI: 6.3–33.5; P = 0.009) (Amarenco and Labreuche, 2009).
In men and women at an equivalent risk of cardiovascular disease, statin therapy is of similar effectiveness for the prevention of major vascular events (Cholesterol Treatment Trialists’ [CTT] Collaboration, 2015).
A meta-analysis including data from 22 trials on statin use versus controls (n = 134,537 subjects; mean LDL-C difference 1.08 mmol/L; median follow-up 4.8 years) and 5 trials of higher-dose versus lower-dose statin use (n = 39,612 subjects, difference 0.51 mmol/L; 5.1 years) reported that for stroke, the reduction in risk for subjects with 5-year risk of major vascular events lower than 10% (RR per 1 mmol/L LDL-C reduction 0.76, 99% CI: 0.61–0.95; P = 0.0012) was also similar to that seen in higher-risk categories (trend P = 0.3) (Cholesterol Treatment Trialists’ [CTT] Collaboration, 2012).
In a meta-analysis of 8 RCTs (n = 25,952) comparing any statins with placebo or usual care for primary prevention of CVD in subjects aged 65 years or older, statins significantly reduced the risks of composite major adverse cardiovascular events (RR 0.82, 95% CI: 0.74–0.92), nonfatal MI (0.75, 95% CI: 0.59–0.94), and total MI (0.74, 95% CI: 0.61–0.90). Treatment effects of statins were statistically insignificant in fatal MI (0.43, 95% CI: 0.09–2.01), stroke (fatal: 0.76, 95% CI: 0.24–2.45; nonfatal: 0.76, 95% CI: 0.53–1.11; total: 0.85, 95% CI: 0.68–1.06) and all-cause mortality (0.96, 95% CI: 0.88–1.04) (Teng et al., 2015).
Among 821 patients with a history of stroke or TIA (and CHD), who were randomized in the CARE and LIPID trials (Plehn et al., 1999; White et al., 2000) to pravastatin (n = 436) or placebo (n = 385), the odds of a recurrent stroke were reduced by 33% (95% CI: −1% to +56%) from 15.1% (placebo) to 10.6% (pravastatin) (Manktelow et al., 2002). Patients with a history of only stroke had similar results.
Among the 3289 subjects in the HPS trial with a history of symptomatic ischaemic cerebrovascular disease a mean of 4.3 (standard error [SE]: 0.1) years previously, 1820 had a past history of stroke only and 1460 had a past history of CHD and stroke. For these 1820 with previous stroke only, allocation to simvastatin was associated with a significant reduction in major vascular events: 23.6% (placebo) to 18.7% (simvastatin). This resulted in an RRR of 21% (95% CI: 5–34%, P < 0.001), and an absolute risk reduction (ARR) of 49 major vascular events (4.9%) per 1000 stroke patients allocated to simvastatin over 5 years. Among the 1460 subjects with known CHD and a past history of stroke, allocation to simvastatin was associated with a similar reduction in any major vascular event: 37.4% (placebo) to 32.4% (simvastatin) that resulted in an RRR of 14% (95% CI: 0.5–25%), and an ARR of 5% over 5 years.
A retrospective subgroup analysis of HPS data suggested that in subjects with a history of cerebrovascular disease, simvastatin did not reduce the overall rate of recurrent stroke compared with placebo (simvastatin: 10.3%, placebo: 10.4%; RR: 0.98, 95% CI: 0.79–1.22), in contrast to other high-risk subjects who had a highly significant reduction in stroke (simvastatin: 3.2%, placebo: 4.8%; heterogeneity P = 0.002). Those with a history of cerebrovascular disease who were assigned to simvastatin had a 19% (SE: 12, P = 0.1) reduction in the RR of ischaemic stroke (simvastatin: 6.1%, placebo: 7.5%) but a nearly 2-fold increase in RR for haemorrhagic stroke (placebo: 0.7%, simvastatin: 1.3%). This latter result was in contrast with other high-risk subjects who had a non-significantly lower risk of haemorrhagic stroke (heterogeneity P = 0.03).
The SPARCL study was the first to investigate the effects of statins on the risk of cerebrovascular events in patients without a history of CHD. This double-blind, randomized, placebo-controlled, multicentre trial examined the effect of aggressive atorvastatin therapy (80 mg/day) on specified cerebrovascular endpoints. Patients were eligible for the study if they had had a previous TIA or stroke and an LDL level between 100 mg/dL (2.58 mmol/L) and 190 mg/dL (4.91 mmol/L), without any evidence of CHD. The primary clinical endpoint was the time to first occurrence of a fatal or nonfatal stroke. In this study, 4731 patients who had had a stroke or TIA within the past 6 months were randomized. After 6 years of follow-up, 265 patients in the atorvastatin group had a fatal or nonfatal stroke compared with 311 in the control group. This resulted in a 16% risk reduction in time to first occurrence of stroke for atorvastatin (adjusted hazard ratio 0.84, 95% CI: 0.71–0.99; number needed to treat 46). For the secondary endpoint of time to stroke or TIA, a 23% risk reduction was reported (HR 0.77, 95% CI: 0.67–0.88) with 375 events in the atorvastatin group and 476 in controls. Moreover, there was a 35% reduction in coronary events (HR 0.65, 95% CI: 0.49–0.87) (SPARCL Investigators, 2006). The treatment effect did not differ between men and women, in individuals aged younger than 65 years and those older than 65 years, in those with carotid stenosis at entry compared with those with no carotid stenosis, in patients with diabetes compared with those without, and across ischaemic stroke subtype at entry.
In Figure 16.1, the forest plot shows the effect of statins compared with controls, in patients with and without a history of either stroke or TIA, on the incidence of subsequent stroke (fatal and not-fatal) of any pathological type. In the secondary prevention of non-cardioembolic stroke, a significant reduction in LDL-C levels due to statin use significantly reduced the risks of recurrent stroke (RR 0.84, 0.71–0.99, P = 0.03) and major cardiovascular events (0.80, 0.69–0.92, P = 0.002) (Amarenco and Labreuche, 2009).