Although hyperlipidemia is an important risk factor for coronary heart disease and ischemic stroke, it may protect individuals from developing ICH. A study in the United States of America (USA) examined the relationship between the serum total cholesterol level and risk of death from a stroke during a 6-year follow-up in 350,977 men aged 35 to 57 years. They found that the 6-year risk of death from ICH was three times higher in men with serum total cholesterol levels <160 mg/dL [12]. A South Korean study also showed that hypocholesterolemia was associated with an increased ICH risk. This study additionally revealed that the ICH risk in subjects with hypocholesterolemia was more augmented in those with concomitant hypertension [13]. Another South Korean case-control study showed that decreased fat intake was an important factor for ICH development in men [14]. A recent meta-analysis [15] involving 12 prospective studies with 476,173 participants showed that a 10 mg/dL increase in LDL-C was associated with a 3% lower ICH risk (pooled relative risk [RR], 0.97; 95% CI, 0.95 to 0.98). The association appeared to be more pronounced in Asians (pooled RR, 0.95; 95% CI, 0.92 to 0.98) relative to Caucasians (pooled RR, 0.98; 95% CI, 0.97 to 1.00). Decreased serum lipid levels were also found to be an independent risk factor for CMBs [16]. Thus, hypolipidemia appears to be a risk factor for brain microbleeds or macrobleeds, particularly in patients with hypertension. Differences in serum cholesterol levels among patients with different types of stroke were also demonstrated in an autopsy study from Japan [17], with mean serum cholesterol levels of 164±2, 177±3, and 200±4 mg/dL for patients with ICH, SVD, and cortical infarction, respectively. These data appear to be related to the degree of underlying vascular atherosclerosis among the groups; the mean values of atherosclerosis scores (assigning grade 0 for no stenosis to 4 for ≥50% stenosis in each of the 22 sites of the basal cerebral arteries) were 25±1, 35±2, and 40±2, respectively.

Since the SPARCL study showed that patients on statin treatment had a significant increase (HR, 1.66; 95% CI, 1.08 to 2.55) in the ICH risk [4], there had been concerns regarding ICH development in patients taking statins. In 2011, a meta-analysis involving 23 randomized trials and 19 observational studies was published [18], which showed that statin use was not associated with an increased ICH risk in randomized trials (RR, 1.10; 95% CI, 0.86 to 1.41), cohort studies (RR, 0.94; 95% CI, 0.81 to 1.10), or case-control studies (RR, 0.60; 95% CI, 0.41 to 0.88). Another meta-analysis published a year later included 91,588 and 91,215 subjects in the statin treatment and control groups, respectively [19]. Although total stroke (odds ratio [OR], 0.84; 95% CI, 0.78 to 0.91; P<0.0001) and all-cause mortality (OR, 0.92; 95% CI, 0.87 to 0.96; P=0.0007) were significantly reduced in the statin therapy group, there was no significant difference in the ICH incidence between the groups (OR, 1.08; 95% CI, 0.88 to 1.32; P=0.47). ICH risk was not related to the degree of LDL-C reduction or achieved LDL-C level.

Subsequently, a large population-based Swedish study was published that included 7,696 ICH cases and 14,670 age-sex-matched stroke-free controls [20]. After adjusting for vascular risk factors and the use of antithrombotic therapy, statin use before ICH was associated with a decreased ICH risk in patients (OR, 0.68; 95% CI, 0.63 to 0.74). This protective mechanism was thought to be due to the pleiotropic effects of statins, such as anti-inflammatory and antithrombotic properties that may protect the cerebral vessels [21]. A more recent meta-analysis showed that patients who used statins before ICH had a decreased mortality risk at 3 months after symptom onset (OR, 0.47; 95% CI, 0.32 to 0.68) and an increased probability of good functional outcomes (OR, 1.49; 95% CI, 1.01 to 2.19), as compared with those who did not [22]. Therefore, although hypocholesterolemia appears to be a risk factor for ICH, particularly in those with hypertension, statin use does not appear to increase the ICH risk in the normal population and may ameliorate the severity of ICH if it develops.

A recent meta-analysis examined the relationship between statin use and the development of CMBs [23]. It included seven studies (n=3,734 participants, with an unselected general population [n=1,965], patients with ischemic stroke [n=849], patients with hemorrhagic stroke [n=252], and patients with hypertension aged >60 years [n=668]). The results showed that statin use was not associated with CMBs in either unadjusted (OR, 1.15; 95% CI, 0.76 to 1.74) or adjusted analyses (OR, 1.09; 95% CI, 0.64 to 1.86). Statin use was more strongly related to lobar CMB presence (OR, 2.01; 95% CI, 1.48 to 2.72) in the unadjusted, but not in the adjusted analysis (OR, 2.26; 95% CI, 0.86 to 5.91). Thus, statin treatment did not appear to cause CMBs overall.

Unfortunately, there have been no randomized controlled trials examining the efficacy or hazards of lipid-lowering therapy in patients with ICH. A retrospective study from Taiwan, which used the National Health Insurance Research Database data, compared the outcomes of early statin users (n=749) and statin non-users (n=7,583). During follow-up, the incidence of recurrent ICH was similar between the two groups: 9.2% and 8.9% in the statin and control groups, respectively. However, the mortality rate was significantly lower in patients in the statin group: 90 (12.0%) and 1,519 (20.0%) in the statin and control groups (adjusted HR, 0.742; 95% CI, 0.598 to 0.919), respectively [24]. Another study from China that used data from the China National Stroke Registry analyzed 3,218 admitted patients with ICH; 220 (6.8%) were on statins during their hospitalization. Multivariate analysis showed that patients with ICH on statins had better functional outcomes at 3 months (OR, 2.24; 95% CI, 1.49 to 3.36) and at 1 year (OR, 2.04; 95% CI, 1.37 to 3.06), and lower mortality rates at 3 months (OR, 0.44; 95% CI, 0.22 to 0.87) and 1 year (OR, 0.49; 95% CI, 0.27 to 0.86) than those without statin therapy [25].

A retrospective study from Japan [26] involved 381 patients with ICH, among whom statins were used in 56 patients. Serum total cholesterol and LDL-C levels were significantly lower in the statin group than in the non-statin group. Hematoma volume was inversely correlated with total cholesterol and LDL-C levels in both groups. Multiple logistic regression analysis showed that serum cholesterol levels ≤150 mg/dL were associated with worsening of the National Institutes of Health Stroke Scale (NIHSS) score (OR, 1.4; 95% CI, 1.21 to 1.63) and increased hematoma volume (OR, 1.1; 95% CI, 1.07 to 1.13) in patients with ICH. Similar findings were observed when patients who used statins were analyzed separately. A more recent retrospective Finnish study [27] analyzed 964 patients with ICH and reported that, after adjusting for factors such as age, NIHSS, Glasgow Coma Scale, ICH volume, and intraventricular hemorrhage, lower LDL-C levels were independently associated with in-hospital mortality (OR, 0.54; 95% CI, 0.31 to 0.93; P=0.028).

Thus, statin use after ICH may not increase the risk of further ICH and may improve the functional outcome and mortality in patients with ICH. However, excessive lowering of cholesterol levels by statins may have to be avoided. However, these retrospective studies have limitations. Details were frequently unknown, such as why and when the statin was initiated, the type and dose of statin, and the duration of treatment. Prospective, controlled trials are needed to elucidate whether statin therapy is needed in patients with ICH, and if so, when and with what dosage should be initiated.

Similar to atherosclerosis in other vessels, ICAS is caused by vascular risk factors, such as age, hypertension, diabetes mellitus, cigarette smoking, and hyperlipidemia. However, studies that compared the risk factors between ICAS and ECAS have shown that hyperlipidemia is a less important risk factor for ICAS than for ECAS [46-51]. To the best of my knowledge, no studies have reported the opposite results. Because Asians have lower serum cholesterol levels than Caucasians, this difference may at least in part explain the ethnic differences in the location of cerebral atherosclerosis. Intracranial arteries may have an inherent resistance to the toxic effects of hypercholesterolemia. In monkeys [52] and rabbits [53], the development of atherosclerosis after consuming an atherogenic diet occurred significantly later in intracranial arteries than in extracranial arteries. It has been suggested that a specific glycocalyx composition on luminal endothelial cells might inhibit trapping of chylomicrons and very-low-density lipoprotein, resulting in reduced deposition of apolipoproteins in the intima of intracranial vessels [52].

Although there may be differences in the pathological features between ICAS and ECAS, few studies have explored this issue. A recent autopsy study from Hong Kong [54] examined the pathology of large intracranial arteries in 32 autopsy cases aged ≥45 years. Although atherosclerotic lesions were present in 91 (71%) of the 128 arteries, features of complicated plaques were infrequent; plaque hemorrhage was encountered in 12%, neo-vasculature in 12%, lumen thrombi in 13%, macrophage infiltration in 20%, and calcification in 25% of arteries. Luminal narrowing was the most severe in the MCA, followed by the vertebral artery (VA), while the basilar artery (BA) was the least stenotic. Thus, it appears that the lesions generally lacked features of complicated plaques in proximal ICA atherosclerotic disease. However, it should be kept in mind that the complicated plaques described in proximal ICA disease were obtained primarily in patients who underwent carotid endarterectomy, a highly selected patient population mostly with severe atherosclerosis, whereas most of the subjects in the autopsy study described above did not have a history of stroke.

There have been few studies on lipid-lowering therapies that exclusively focus on ICAS. In a recent study [55], the investigators enrolled statin-naïve patients with ischemic stroke who had symptomatic ICAS (>50% stenosis) in the proximal portion of the MCA, BA, or intracranial portion of the ICA. Atorvastatin (40 to 80 mg) or rosuvastatin (20 mg) was administered for 6 months, and patients underwent HR-MRI before and after statin treatment. Pre-specified endpoints that were measured after 6 months of statin treatment were: (1) degree of stenosis, (2) remodeling index, (3) wall area index, and (4) enhancement volume of atherosclerotic plaque on HR-MRI. The LDL-C concentrations of the 77 enrolled patients at the initial and follow-up assessments were 125.8±35.7 and 61.0±19.3 mg/dL, respectively. Overall, statin treatment significantly reduced plaque accumulation (32.07±39.15 mm³ vs. 17.06±34.53 mm³, P=0.013), wall area index (7.5±4.3 vs. 5.9±4.1, P=0.016), and degree of stenosis (76.5%±20.2% vs. 64.1%±21.3%, P<0.001), but not the remodeling index (P=0.195). This suggests that high-dose statins may benefit patients with symptomatic ICAS. However, this study has important limitations. First, it was uncertain whether such MRI changes could translate into clinical benefits. Second, because there was no control group, it remains unclear whether the MRI changes were indeed due to statin administration. Third, it is unknown whether the dose used in this study was the most appropriate for patients with ICAS. It is noteworthy that despite intensive statin therapy, 35% of the participants showed no beneficial changes; they showed unchanged or worsened vessel pathologies.

To the best of my knowledge, there is only one controlled study on statins for asymptomatic ICAS. In a single-center prospective study [56], 71 asymptomatic patients with ICAS were treated with rosuvastatin, and their vascular stenoses were evaluated using transcranial color-coded sonography (TCCS) before and after treatment. The therapeutic target was a reduction in LDL-C concentration of ≤1.8 mmol/L or a ≥50% reduction from baseline over 2 years. The participants were allocated to an intensive statin treatment (IST) or standard statin treatment (SST) group. One hundred and four ICASs were identified in 51 participants and 47 in 20 in the IST and SST groups, respectively. After 1 year, there was a larger decrease in LDL-C concentration in the IST than in the SST group (1.5±0.3 mmol/L vs. 2.2±0.6 mmol/L, respectively; P≤0.001). The percentage regression of the ICAS lesions in the IST group tended to be greater than that in the SST group (26.3% vs. 5.9%, P=0.052). Forty-nine lesions in 25 participants and 16 lesions in seven in the IST and SST groups, respectively, were reassessed 2 years later. At this time, the LDL-C concentration had decreased more in the IST group than in the SST group (1.6±0.3 mmol/L vs. 2.46±0.8 mmol/L, P=0.048), and the percentage regression of ICAS lesions in the IST group was significantly greater than that in the SST group (34.7% vs. 6.3%, P=0.017). This result suggests that the degree of stenosis in ICAS can be ameliorated using intensive statin therapy. However, this study had limitations: the number of participants who were followed for 2 years was too small, and the stenosis was assessed by TCCS rather than by standard methods, such as MRA, CTA, or conventional angiography.

Thus, although statins probably benefit patients with ICAS, further studies including a control group, assessment of clinical endpoints, and comparison of the effects of statin therapy in patients with ICAS and ECAS are required. Changes in vessel stenosis or plaque volume should also be assessed more meticulously. Considering the heterogeneity of ICAS, other issues should be considered in future trials. First, strokes in patients with ICAS are caused by various mechanisms, including atherosclerotic steno-occlusion, BAD leading to perforator occlusion, artery-to-artery embolism, and hemodynamic impairment [37,57]. Recent studies have shown that risk factors and vessel wall pathology may differ among different mechanisms [39,58-62]. Specifically, rupture-prone positive remodeling and enhancing plaques are more frequently observed in non-BAD-type ICAS than in BAD-type ICAS [62]. Therefore, it may be assumed that lipid-lowering agents might be more effective in patients with non-BAD type ICAS. Second, as imaging techniques, such as MRA and catheter angiography, can examine luminal stenosis but not vascular wall pathology, non-atherosclerotic etiologies, such as Moyamoya disease, dissection, and vasculitis can be miscategorized as ICAS. One diagnostic tool that can be used to examine the vessel wall is HR-MRI [63]. A study from South Korea [64] assessed HR-MRI findings in 95 young (≤55 years) patients who were diagnosed with presumable ICAS. They found that only 26 (27.4%) had HR-MRI findings consistent with atherosclerosis, whereas others showed findings consistent with other vascular pathologies. These data suggest that diagnosis based solely on angiogram findings may not be accurate, which may explain why the effect of statins was not consistent in patients with ICAS [55].

Proximal ICA atherosclerosis has many similarities to coronary artery disease, and hypercholesterolemia is an important risk factor. Of the 4,731 SPARCL participants, a subset of 1,007 participants had ICA disease. In this subgroup, atorvastatin therapy was associated with a 33% reduction in the risk of recurrent stroke, which was far greater than the 16% reduction observed in all SPARCL participants [11]. Thus, among patients with stroke, those with ICA disease appear to benefit most from a high dose of statin.

Statins appear to influence the composition of the carotid plaques. A recent meta-analysis of seven studies (361 patients) [65] included prospective studies using serial HR-MRI in patients on statin therapy to document changes in carotid plaque composition. Outcomes included changes in lipid-rich-necrotic-core (LRNC) volume, wall volume, and lumen volume at varying time points after initiation of statin therapy. There was no significant change in carotid wall volume or lumen volume at any time point. While there was no significant difference in LRNC before 7 to 12 months after statin initiation, there was a significant decrease in LRNC volume (weighted mean difference –9.9 mm³; 95% CI, –8.9 to –2.3) at >12 months.

Subsequently, a population-based Rotterdam Study [66] enrolled 1,740 participants with ICA disease who underwent carotid artery MRI to determine the presence of calcification, lipid core, and intra-plaque hemorrhage. Statin treatment was associated with a higher presence of calcification (OR, 1.73; 95% CI, 1.22 to 2.44), which was more apparent in patients with longer statin use. Current statin treatment was also associated with a lower presence of lipid core (OR, 0.66; 95% CI, 0.42 to 1.04), but only when statins were used for ≤10 months. Although there are subtle differences in the results, the two studies suggest that statin therapy favorably influences the composition of carotid atherosclerosis by shifting the composition from vulnerable plaque with a lipid core to a more stable calcified plaque.

A more recent Treat Stroke to Target–Plaque Ultrasound Study study [67] assessed the serial ultrasonic carotid findings for a median follow-up of 3.1 years. Patients were randomized to the lower LDL-C target (mean, 64 mg/dL; n=201) or higher LDL-C target group (mean, 106 mg/dL; n=212). Compared with the higher-target group, the lower-target group had a similar incidence of newly diagnosed carotid plaque: 46/201 (5-year rate, 26.1%) versus 45/212 (5-year rate, 29.7%). The change in common carotid arteries intima-media thickness was significantly (P=0.004) better in the lower-target group (−10.53 μm; 95% CI, –14.21 to –6.85) than in the higher target group (−2.69 μm; 95% CI, –6.55 to 1.18). Thus, in this study, statin use with an LDL-C target of <70 mg/dL did not reduce the incidence of new carotid plaques but produced greater regression of carotid atherosclerosis than an LDL-C target of 90 to 110 mg/dL.

The benefits of statins in patients who underwent carotid revascularization have also been investigated. A population-based, retrospective cohort study in Ontario, Canada, evaluated the statin effect in individuals aged ≥66 years who underwent carotid endarterectomy or stenting during 2002-2014 [68]. The study primarily assessed the effect of pre-procedural statin for the composite of stroke, myocardial infarction, or death at 1 year. Additionally, the benefit of persistent long-term statin use was assessed 5 years after the procedure. Of the 10,723 patients identified, 7,893 (73.6%) were pre-procedural statin users. Pre-procedural statin use was associated with a 24% reduction (adjusted HR, 0.76; 95% CI, 0.70 to 0.83) in the risk of the 1-year composite of stroke, myocardial infarction, or death. At 5 years, persistent long-term statin use was associated with a 25% risk reduction (adjusted HR, 0.75; 95% CI, 0.71 to 0.80) for the composite outcome. These results suggest that statins should be used in patients undergoing carotid revascularization.

Unlike ICA, VA atherosclerosis has rarely been examined in clinical studies of lipid-lowering agents. Similar to ICA, VA atherosclerotic disease produces artery-to-artery embolism, leading to posterior circulation ischemia. Therefore, the plaque-stabilizing effect of statins may be equally important in patients with VA disease. Thus, statins have been recommended in patients with atherosclerotic VA disease [69]. However, although pathological studies are scarce owing to the fact that endarterectomy is rarely performed in VA disease, plaque ulceration and hemorrhage are considered to be less common in VA atherosclerosis than in ICA disease; the lesions are more often fibrous and smooth [70]. Thus, it remains uncertain whether statins are really effective, and if so, what the optimal LDL-C target would be for patients with VA atherosclerotic disease.