Add‐On PCSK9 Inhibitor Therapy for Acute Stroke Attributable to Intracranial Atherosclerosis: A Randomized Clinical Trial

Abstract

Background

Intracranial atherosclerotic stenosis is a major cause of ischemic stroke with high recurrence rates despite intensive therapy. The efficacy of proprotein convertase subtilisin/kexin type 9 inhibitors (PCSK9i) as adjunctive treatment in intracranial atherosclerotic stenosis remains unclear.

Methods

We conducted a prospective, randomized, open‐label, blinded‐end point trial involving 62 patients with symptomatic intracranial atherosclerotic stenosis. Participants were randomized 1:2 to receive either a PCSK9i (alirocumab 75 mg every 2 weeks) plus high‐intensity statin or high‐intensity statin alone for 6 months. The primary outcome was the change in intracranial artery stenosis measured by high‐resolution vessel‐wall magnetic resonance imaging. Secondary outcomes included plaque enhancement volume, low‐density lipoprotein cholesterol target achievement (<55 mg/dL), recurrent stroke events, modified Rankin Scale, and safety assessments.

Results

Among 62 participants (median age, 66 years; 77% men), 60 completed the study per protocol. At 6 months, median stenosis reduction was greater in the PCSK9i group (7.1%; interquartile range, 3.6–12.8%) than in controls (−1.2%; interquartile range, −4.9–4.5%) (P<0.001). The baseline‐adjusted mean difference estimated using a generalized linear model was 8.9 percentage points (95% CI, 5.1–12.6). Both groups showed significant reduction in plaque enhancement volume, but between‐group difference was not significant (4.3 versus 4.0 mm³; P=0.79). Low‐density lipoprotein cholesterol <55 mg/dL was achieved in 85% of the PCSK9i group versus 13% of controls (P<0.01). Recurrent stroke occurred in 5% versus 13% of patients (P=0.34). No serious adverse events were reported in either group.

Conclusions

In patients with symptomatic intracranial atherosclerotic stenosis, adjunctive PCSK9i therapy significantly reduced intracranial stenosis and improved low‐density lipoprotein cholesterol control over 6 months.

Registration

URL: https://www.clinicaltrials.gov; Unique Identifier: NCT05001984.

Nonstandard Abbreviations and Acronyms

ICAS

intracranial atherosclerotic stenosis

PCSK9i

Proprotein Convertase Subtilisin/Kexin Type 9 Inhibitor

Clinical Perspectives.

What Is New?

• This randomized controlled trial shows that adding proprotein convertase subtilisin/kexin type 9 inhibitor therapy to high‐intensity statins significantly reduces intracranial stenosis and improves low‐density lipoprotein cholesterol target achievement.

• It also demonstrates, using vessel‐wall magnetic resonance imaging, a clear association between low‐density lipoprotein cholesterol reduction and plaque regression in symptomatic intracranial atherosclerotic stenosis.

What Are the Clinical Implications?

• The findings support the potential clinical benefit of proprotein convertase subtilisin/kexin type 9 inhibitor in intracranial atherosclerotic stenosis management and suggest that more intensive lipid‐lowering strategies may improve vascular outcomes; these results could inform future stroke prevention guidelines and prompt larger trials to refine low‐density lipoprotein cholesterol targets for this high‐risk population.

Intracranial atherosclerotic stenosis (ICAS) is a major cause of ischemic stroke worldwide, particularly among Asian populations, ¹ and is associated with a high risk of recurrent events despite standard secondary prevention. ² , ³ Mechanisms of stroke recurrence in ICAS include hypoperfusion, artery‐to‐artery embolism, and branch atheromatous disease, and existing therapies—namely dual antiplatelet therapy and high‐intensity statins—are often insufficient to fully mitigate this risk. ⁴ , ⁵

Intensive lipid lowering has been established as an effective strategy to reduce vascular events in atherosclerotic disease. The SPARCL (Stroke Prevention by Aggressive Reduction in Cholesterol Levels) trial demonstrated that high‐intensity statin significantly reduces stroke and cardiovascular risk in patients with prior ischemic events, ⁶ , ⁷ while the TST (Treat Stroke to Target) trial showed that targeting low‐density lipoprotein cholesterol (LDL‐C) to <70 mg/dL, compared with a target of 90–110mg/dL, was associated with a lower risk of subsequent cardiovascular events in patients with previous ischemic stroke or transient ischemic attack (TIA) with evidence of atherosclerosis. ⁸ High‐intensity statin treatment and an LDL‐C target <70 mg/dL is now recommended in patients with symptomatic ICAS. ⁹ , ¹⁰ , ¹¹

Proprotein convertase subtilisin/kexin type 9 inhibitors (PCSK9i), such as alirocumab and evolocumab, have shown substantial benefit in reducing LDL‐C and major cardiovascular events in high‐risk populations. ¹² , ¹³ , ¹⁴ Although their effect on coronary and carotid plaque burden has been validated, their impact on intracranial atherosclerosis remains less well defined. ¹⁵ Observational and imaging studies suggest that statin treatment may contribute to plaque regression and improved vessel stability in ICAS. ¹⁶ , ¹⁷

We hypothesized that the addition of PCSK9i to high‐intensity statin therapy would promote intracranial plaque stabilization and regression in patients with symptomatic ICAS. To test this, we conducted the TOPICAL‐MRI study (Trial of PCSK9 Inhibition in Patients with Acute Stroke and Symptomatic Intracranial Atherosclerosis‐Magnetic Resonance Imaging) using vessel‐wall magnetic resonance imaging (MRI) to assess changes in intracranial plaque and clinical outcomes over a 6‐month treatment period. ¹⁸

METHODS

The study protocol was approved by the Institutional Review Board of Chang Gung Memorial Hospital (approval no. 202002482A3) and was registered on ClinicalTrials.gov (NCT05001984). Written informed consent was obtained from all participants or their legally authorized representatives before study enrollment. The data that support the findings of this study are available from the corresponding author upon reasonable request.

Study Design and Participants

This prospective, open‐label, blinded‐end point trial was conducted at a tertiary medical center. ¹⁸ Eligible participants had acute ischemic stroke with diffusion‐weighted imaging‐confirmed infarction within 7 days attributable to symptomatic ICAS and a pre‐stroke modified Rankin Scale ≤2. Patients with isolated transient symptoms without diffusion‐weighted imaging infarction (tissue‐negative TIA) were not eligible. All participants had symptomatic ICAS (30%–99%) involving the intracranial internal carotid artery (ICA), middle cerebral artery M1 segment, or basilar artery—confirmed by high‐resolution vessel‐wall MRI. ¹⁸ All patients enrolled in this trial were classified as having large artery atherosclerosis based on clinical presentation and imaging findings, with other stroke etiologies (eg, cardioembolism, dissection, vasculitis) excluded. ¹⁹ Stroke mechanisms associated with symptomatic ICAS included: (1) branch atheromatous disease, (2) artery‐to‐artery embolism, (3) hypoperfusion or impaired embolic clearance distal to a stenosis, and (4) in situ thrombo‐occlusion beyond a stenotic segment. ⁴ , ²⁰ A baseline LDL‐C≥100 mg/dL or ≥70 mg/dL with ongoing statin therapy was required.

Randomization and Interventions

Participants were randomized in a 1:2 ratio to receive either: (1) a PCSK9i (alirocumab 75 mg subcutaneously every 2 weeks) plus high‐intensity statin (rosuvastatin 20 mg or atorvastatin 40–80 mg daily), or (2) high‐intensity statin alone.

The 1:2 allocation was prespecified to improve precision of the control estimate while maintaining feasibility (cost/logistics of PCSK9i) and to maximize data on contemporary standard care. Although 1:1 is most statistically efficient, the chosen scheme balanced feasibility with information yield and was implemented with concealed allocation.

Randomization was performed using a computer‐generated sequence with permuted block sizes in a 1:2 ratio. Allocation was concealed using sequentially numbered, sealed opaque envelopes prepared by an independent coordinator not involved in enrollment. Investigators enrolled participants and obtained assignment after eligibility confirmation. Outcome assessors and imaging readers were blinded to group allocation.

All patients received standardized stroke care, including dual antiplatelet therapy for 90 days, management of vascular risk factors, and structured lifestyle education. Participants underwent clinical and laboratory evaluations at baseline, 2 weeks, 3 months, and 6 months. Follow‐up imaging with the same MRI protocol was performed at 6 months. LDL‐C levels were monitored, and ezetimibe was permitted to achieve LDL‐C<70 mg/dL, per guideline.

Study Outcomes

The primary end point was the absolute change in intracranial stenosis (percentage points) of the target intracranial artery from baseline to 6 months on high‐resolution MRI.

Secondary outcomes included: (1) change in plaque enhancement volume, (2) proportion achieving LDL‐C<55 mg/dL at 6 months, (3) recurrent ischemic stroke or TIA in the same territory, (4) modified Rankin Scale (mRS) at 3 and 6 months, and (5) adverse events including hepatotoxicity, myopathy, injection‐site reactions, and serious events.

MRI Protocol and Image Analysis

MRI was performed using a 3.0 Tesla Siemens Verio system with a 32‐channel head coil. Sequences included axial diffusion‐weighted imaging, axial T1‐ and T2‐weighted imaging, 3‐dimensional time‐of‐flight (TOF) magnetic resonance angiography (MRA), and dynamic susceptibility contrast perfusion imaging. High‐resolution vessel‐wall imaging was acquired using black‐blood sequences with a spatial pre‐saturation. Axial T2‐weighted images were acquired with the following parameters: TR/TE=3200/323 milliseconds, echo train length=185, slice thickness=6 mm, flip angle=180°, matrix size=320 × 266, and field of view (FOV)=149.6 mm. Pre‐ and post‐contrast axial T1 fluid‐attenuated inversion recovery images were obtained with TR/TE=700/22 ms, echo train length=55, slice thickness=6 mm, flip angle=180°, matrix=320 × 320, and FOV=190 mm. A gadolinium‐based contrast agent was administered at a dose of 0.1 mmol/kg for post‐contrast imaging. Sagittal views of T1‐weighted, T2‐weighted, and post‐contrast T1‐weighted images were reconstructed from the corresponding axial acquisitions, with the reconstruction plane‐oriented perpendicular to the long axis of the vessel at the site of the culprit stenosis to optimize visualization of the vessel wall and plaque characteristics.

A plaque was defined as wall thickening relative to adjacent or contralateral normal vessel segments. A culprit plaque was defined as either solitary or the most stenotic in the vascular territory corresponding to the infarct. Measurements were obtained from cross‐sectional images of the culprit plaque. The site of maximal stenosis was first identified using 3‐dimensional time‐of‐flight MRA. Axial T2‐weighted images were used to measure the reference vessel area at a proximal or distal, non‐stenotic segment adjacent to the plaque. On pre‐contrast T1‐weighted images, the lumen area and plaque wall area at the site of maximal stenosis were manually delineated. The wall area was calculated as the difference between the total vessel area and lumen area (Figure 1). Degree of stenosis was calculated using the following formula:

$$\text{Stenosis}\left(\%\right)=\left(1-\frac{\text{Lesion lumen area}}{\text{Reference lumen area}}\right)\times 100\%$$

Figure 1. Illustration of high‐resolution MRI measurements for intracranial atherosclerotic stenosis before and after treatment.

A, Baseline imaging of the maximal stenotic lesion. The site of maximal stenosis was identified at the left middle cerebral artery on 3‐dimensional time‐of‐flight angiography (red arrow). Corresponding axial T2‐weighted, T1‐weighted, and T1‐weighted contrast‐enhanced images illustrate vessel‐wall morphology. The vessel area (red outline, red hollow arrow) was measured at the proximal normal middle cerebral artery segment adjacent to the culprit plaque on the T2‐weighted image. The lumen area (red circle, red solid arrowhead) and plaque enhancement area (red circle, red hollow arrowhead) were semi‐automatically measured at the site of maximal stenosis on the T1‐weighted image. B, Follow‐up imaging at 6 months shows a reduction in stenosis severity and plaque enhancement. Follow‐up time‐of‐flight angiography (red arrow) demonstrates improved lumen diameter. Corresponding axial MRI sequences show the vessel area (red outline, red hollow arrow, T2‐weighted image), increased lumen area (red circle, red solid arrowhead, T1‐weighted image), and reduced plaque enhancement volume (red circle, red hollow arrowhead, T1‐weighted contrast‐enhanced image). All measurements were performed on sagittal images perpendicular to the long axis of the vessel at the site of maximal stenosis. MRI indicates magnetic resonance imaging.

Plaque enhancement volume was quantified by comparing pre‐ and post‐contrast T1‐weighted images. Two independent, blinded readers performed all measurements; discrepancies were resolved by consensus.

Sample Size

The sample size was calculated using a 2‐sided t test to detect an effect size of 0.7 between groups, assuming 80% power and α=0.05. This moderate‐to‐large difference was based on prior ICAS data showing plaque regression with statins. ¹⁶ Although vessel‐wall imaging data were unavailable during study design, the estimate was considered biologically plausible. A total of 60 patients (20 intervention, 40 control) was required.

Statistical Analysis

All statistical analyses were performed using SPSS version 27 (IBM Corp., Armonk, NY). Normality of continuous variables was assessed via the Kolmogorov–Smirnov test. Non‐normally distributed data were summarized as medians with interquartile ranges and compared using the Mann–Whitney U test. Categorical variables were analyzed with Chi‐square or Fisher's exact tests.

The primary end point, absolute change in intracranial stenosis, was analyzed using a rank‐based analysis of covariance because of its non‐normal distribution. Given the limited sample size, the primary model included only treatment group and baseline stenosis to minimize the risk of model overfitting. The change in plaque enhancement volume was analyzed similarly, with adjustment for baseline enhancement volume.

Because rank‐based analysis yields treatment effects in rank units that are not directly interpretable in the original measurement scale, generalized linear models with a normal distribution and identity link were additionally fitted to estimate adjusted mean differences with 95% CI. These models included only the treatment group and the corresponding baseline MRI measure and were used solely to provide adjusted mean differences and 95% CIs in the original measurement scale, whereas statistical inference was based on the rank‐based analysis of covariance. As a sensitivity analysis, MRI outcomes were reanalyzed using the original multivariable rank‐based analysis of covariance, additionally adjusting for age, sex, baseline National Institutes of Health Stroke Scale score, baseline LDL‐C level, and ezetimibe use.

Binary secondary outcomes were analyzed using generalized linear models with a Poisson distribution, log link function, and robust standard errors to estimate adjusted risk ratios. No correction for multiple comparisons was performed for secondary outcomes, which are hypothesis‐generating.

Within‐group laboratory changes used the Wilcoxon signed‐rank test. Inter‐rater reliability for stenosis measurement was assessed by average‐measures intraclass correlation coefficient using a 2‐way random effects model with absolute agreement definition. As the final stenosis values were averaged across 2 independent raters, the average‐measures intraclass correlation coefficient was reported to reflect the reliability of these aggregated scores. To assess inter‐reader variability in stenosis measurements, a Bland–Altman analysis was performed between the 2 independent readers. For each patient, the mean stenosis value of the 2 readers and their difference were calculated. The mean difference (bias) and the 95% limits of agreement (mean difference ± 1.96 standard deviation) were used to quantify inter‐observer variability in the interpretation of stenosis degree. All stenosis measurements used in the primary analysis were based on the average of the 2 blinded readers.

Spearman rank correlation coefficient was used to assess associations between LDL‐C levels (or change) and stenosis reduction. Scatterplots with overlaid linear regression lines and 95% CI were used for visual presentation of the trends. All statistical tests were 2‐tailed, and a P value <0.05 was considered statistically significant.

RESULTS

Patient Enrollment and Baseline Characteristics

Between July 2021 and October 2024, 1859 patients were screened, of whom 1797 were excluded (Figure 2). A total of 62 patients met eligibility criteria and were enrolled in the intention‐to‐treat population—20 in the PCSK9i group and 42 in the control group. Two patients in the control group did not undergo follow‐up MRI because their family members were unable to bring them back for the imaging study, resulting in a per‐protocol population of 60 patients (20 PCSK9i, 40 control).

Figure 2. Study flowchart of TOPICAL‐MRI trial.

ICA indicates internal carotid artery; LDL‐C, low‐density lipoprotein cholesterol; PCSK9i, proprotein convertase subtilisin/kexin type 9 inhibitor; and TOPICAL‐MRI, Trial of PCSK9 Inhibition in Patients With Acute Stroke and Symptomatic Intracranial Atherosclerosis–Magnetic Resonance Imaging.

Baseline characteristics were generally well balanced between groups in the intention‐to‐treat population (Table 1), including age (median 66 years, interquartile range [IQR] 52–72), vascular risk factors, baseline National Institutes of Health Stroke Scale score (median 3, IQR 1–5), time from stroke onset to randomization (median 4 days, IQR 3–6), and interval to follow‐up MRI (median 184 days, IQR 180–193). Although the trial was initially designed to include patients with stenosis ≥30%, nearly all enrolled participants had stenosis ≥50%, with only one patient having 49% stenosis. Culprit plaques were located in the ICA in 5 patients (8%), the middle cerebral artery in 33 patients (53%), and the basilar artery in 24 patients (39%). The most common stroke mechanism was branch atheromatous disease (44%), followed by hypoperfusion or impaired embolic clearance (27%). Initial stenosis was numerically greater in the PCSK9i group than in the control group (85.6% versus 77.5%). Ezetimibe was used in 17 of 42 patients (41%) in the control group, whereas no patients in the PCSK9i group received ezetimibe (P<0.001).

Table 1.

Baseline Characteristics of the Study Groups

Characteristic	PCSK9i (N=20)	Control (N=42)	P value
Age, median (IQR), y	66 (52–72)	69 (58–79)	0.163
Female, n (%)	6 (30)	8 (19)	0.335
Past medical history, n (%)
Hypertension	18 (90)	35 (83)	0.486
Diabetes	13 (65)	22 (52)	0.348
Hypercholesterolemia	15 (75)	33 (79)	0.753
Coronary artery disease	1 (5)	0 (0)	0.323
Congestive heart failure	1 (5)	1 (2.4)	0.545
Prior stroke	3 (15)	8 (19)	1.000
Current smoking	8 (40)	16 (38)	0.886
Baseline antiplatelet use	4 (20)	6 (14)	0.714
Baseline statin use	5 (25)	10 (24)	0.919
Vital signs and laboratory data, median (IQR)
Systolic BP, mm Hg	167 (144–188)	159 (148–196)	0.940
Diastolic BP, mm Hg	101 (93–116)	103 (86–122)	0.940
Total cholesterol, mg/dL	200 (171–236)	197 (171–219)	0.418
LDL‐C, mg/dL	142 (106–167)	131 (112–152)	0.407
Triglyceride, mg/dL	142 (100–191)	126 (103–161)	0.512
Glycohemoglobin, %	6.3 (5.7–7.3)	6.3 (5.7–7.2)	0.774
Stroke information
Onset to randomization, median (IQR), d	5 (3–6)	4 (3–6)	0.861
NIHSS score on arrival, median (IQR)	3 (1–5)	3 (1–5)	0.857
Statin types, n (%)			0.368
Atorvastatin 40 mg	8 (40)	12 (29)
Rosuvastatin 20 mg	12 (60)	30 (71)
Add‐on Ezetimibe, n (%)	0 (0)	17 (41)	<0.001
Vascular territory of infarction, n (%)			0.178
Middle cerebral artery	12 (60)	21 (50)
Internal carotid artery	3 (15)	2 (5)
Basilar artery	5 (25)	19 (45)
Stenosis, median (IQR), %	85.6 (73.4–88.5)	77.5 (65.3–87.8)	0.134
Infarct mechanisms, n (%)			0.213
Artery‐to‐artery embolization	1 (5)	6 (14)
Branch atheromatous disease	7 (35)	20 (48)
Hypoperfusion	8 (40)	9 (21)
In situ thrombo‐occlusion	0 (0)	3 (7)
Mixed	4 (20)	4 (10)
Time interval of follow‐up imaging, median (IQR), d	183 (181–192)	183 (179–195)	0.707

Continuous variables are presented as median (interquartile range) and were compared using the Mann–Whitney U test. Categorical variables are presented as number (%) and were compared using the Chi‐square test or Fisher exact test, as appropriate. BP indicates blood pressure; HDL‐C, high‐density lipoprotein cholesterol; IQR, interquartile range; LDL‐C, low‐density lipoprotein cholesterol; NIHSS, National Institutes of Health Stroke Scale; and PCSK9i, proprotein convertase subtilisin/kexin type 9 inhibitor.

Primary Outcome

To ensure measurement consistency, inter‐rater reliability was assessed. The intraclass correlation coefficient for pre‐treatment stenosis measurements was 0.72 (95% CI, 0.49–0.84), and for post‐treatment stenosis was 0.70 (95% CI, 0.49–0.82), indicating moderate to good agreement between raters. Bland–Altman analysis demonstrated small mean inter‐reader differences but relatively wide limits of agreement (Figure S1). At baseline, the mean difference between readers was −3.0%, with 95% limits of agreement ranging from −15.7% to 9.7%. At follow‐up, the mean difference was −2.4%, with limits of agreement ranging from −18.7% to 13.8%. These findings indicate minimal systematic bias between readers.

The primary outcome, change in intracranial stenosis, showed a significant between‐group difference at 6 months (Table 2). The PCSK9i group showed a median reduction in stenosis of 7.1 percentage points (IQR, 3.6–12.8), whereas the control group exhibited no meaningful change (−1.2 percentage points, IQR, −4.9 to 4.5). The unadjusted difference between the groups medians was 8.3 percentage points. In the primary rank‐based analysis of covariance adjusted for baseline stenosis, the between‐group difference remained significant (P<0.001). The generalized linear model estimated an adjusted mean difference of 8.9 percentage points (95% CI, 5.1–12.6). As a sensitivity analysis, the original multivariable rank‐based analysis of covariance, additionally adjusting for age, sex, baseline National Institutes of Health Stroke Scale score, baseline LDL‐C level, and ezetimibe use, yielded consistent results, with the treatment effect remaining statistically significant (P<0.001).

Table 2.

Primary and Secondary Efficacy Outcomes, Per‐Protocol Analysis

End point	PCSK9i (N=20)	Control (N=40)	Adjusted effect estimate (95% CI)	P value
Primary efficacy outcome
Reduction in stenosis, median (IQR), %	7.1 (3.6–12.8)	−1.2 (−4.9–4.5)	8.9 percentage points (5.1–12.6)	<0.001
Secondary efficacy outcomes
Enhancement volume reduction, median (IQR), mm3	4.3 (0.5–15.5)	4.0 (0–8.4)	−0.04 (−0.33–0.25) mm3	0.79
Major cardiovascular events within 6 mo, n (%)	2 (10)	5 (13)	0.72 (0.18–2.86)*	0.65
Recurrent stroke within 6 mo, n (%)	1 (5)	5 (13)	0.39 (0.05–2.70)*	0.34
Achieved LDL‐C goals (<70 mg/dL), n (%)	19 (95)	24 (60)	1.61 (1.22–2.13)†	<0.01
Achieved LDL‐C goals (<55 mg/dL), n (%)	17 (85)	5 (13)	6.51 (3.03–16.7)†	<0.01
Good outcomes (mRS≤2) at 6 mo, n (%)‡	16 (80)	35 (88)	0.89 (0.60–1.49)*	0.42

Adjusted effect estimates are reported as adjusted mean differences (95% CI) for continuous MRI outcomes and adjusted relative risks (95% CI) for binary outcomes. Continuous MRI outcomes are presented as median (IQR) for descriptive purposes. Between‐group comparisons (p values) were performed using rank‐based analysis of covariance. Because rank‐based analysis of covariance does not provide treatment effects that are directly interpretable in the original measurement scale, adjusted mean differences and 95% CIs were additionally estimated using generalized linear models with a normal distribution and identity link, adjusted for the corresponding baseline MRI measure (baseline stenosis for stenosis reduction and baseline enhancement volume for enhancement volume reduction). Statistical significance was determined using the rank‐based analysis of covariance. IQR indicates interquartile range; LDL‐C, low‐density lipoprotein cholesterol; MRI, magnetic resonance imaging; mRS: modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; and PCSK9i, proprotein convertase subtilisin/kexin type 9 inhibitor.

^*Adjusted relative risk for PCSK9i vs control, adjusted for age and initial National Institutes of Health Stroke Scale score.

^† Adjusted relative risk for PCSK9i vs control, adjusted for baseline LDL‐C level.

^‡ Scores on the mRS of functional disability range from 0 (no symptoms) to 6 (death).

In addition, a sensitivity analysis using the WASID (Warfarin–Aspirin Symptomatic Intracranial Disease) diameter method produced consistent results, supporting the robustness of the primary findings (Tables [Link], [Link]). A sensitivity analysis in patients with baseline area stenosis ≥70% showed similar trends, as detailed in Table S2.

Secondary Outcomes and Safety Events

The median reduction in plaque enhancement volume was similar between groups (4.3 versus 4.0 mm³; P=0.79; Table 2). A significantly higher proportion of patients in the PCSK9i group achieved the target LDL‐C level of <55 mg/dL (85% versus 13%; adjusted RR 6.51, 95% CI 3.03–16.7). In contrast, there were no significant between‐group differences in the rates of major cardiovascular events (10% versus 13%, P=0. 65), recurrent stroke (5% versus 13%, P=0.34), or favorable functional outcomes (modified Rankin Scale ≤2: 80% versus 88%, P=0.42). No deaths, major bleeding, or severe adverse events occurred. Liver function abnormalities were observed in 2 patients (10%) in the PCSK9i group and 6 patients (15%) in the control group, defined as elevations in aspartate transaminase or alanine aminotransferase exceeding 1.5 times the upper limit of normal (Table S3). A reduced statin dose, attributable to either side effects or LDL‐C levels <25 mg/dL at 3 months, was required in 8 patients (40%) in the PCSK9i group and 11 patients (27.5%) in the control group. In the PCSK9i group, 5 patients required statin dose reduction because their LDL‐C levels fell below 25 mg/dL; nevertheless, all patients completed the full course of PCSK9i treatment.

Laboratory and Imaging Analysis

Lipid profile improvements were significantly greater in the PCSK9i group. At 6 months, LDL‐C decreased from 142 mg/dL to 25 mg/dL in the PCSK9i group, and from 131 mg/dL to 69 mg/dL in the control group (25 versus 69 mg/dL, P<0.01). Total cholesterol and LDL‐C levels decreased significantly in both groups, while high‐density lipoprotein cholesterol increased (Table 3). Triglyceride levels also decreased, with significance reached only in the control group. There were no significant between‐group differences in aspartate transaminase, alanine aminotransferase, creatine phosphokinase or glycohemoglobin.

Table 3.

Changes in Laboratory Values and Vessel‐Wall MRI Parameters (Per‐Protocol Analysis)

Laboratory and MRI parameters	PCSK9i			Control
	Initial	Follow‐up	P value	Initial	Follow‐up	P value
Total cholesterol, median (IQR), mg/dL	200 (171–236)	89 (82–118)	<0.01	197 (171–214)	136 (123–150)	<0.01
Triglyceride, median (IQR), mg/dL	142 (100–191)	102 (70–131)	0.06	131 (102–168)	116 (83–151)	0.02
HDL‐C, median (IQR), mg/dL	39 (33–42)	48 (45–52)	<0.01	37 (34–46)	44 (39–51)	<0.01
LDL‐C, median (IQR), mg/dL	142 (106–167)	25 (20–38)	<0.01	131 (113–153)	69 (61–84)	<0.01
AST, median (IQR), IU/L	25 (19–29)	25 (21–34)	0.28	23 (19–32)	22 (19–26)	0.67
ALT, median (IQR), IU/L	21 (18–31)	25 (19–41)	0.30	21 (16–28)	21 (16–33)	0.59
Glycohemoglobin, median (IQR), %	6.3 (5.7–7.3)	6.0 (5.7–7.1)	0.75	6.3 (5.7–6.9)	6.1 (5.8–6.5)	0.07
Creatine phosphokinase, median (IQR), U/L	104 (49–155)	107 (76–181)	0.18	64 (40–79)	78 (58–107)	0.06
Vessel‐wall MRI, median (IQR)
Stenosis degrees, %	85.6 (73.4–88.5)	75.7 (63.0–82.4)	<0.01	77.5 (65.3–87.8)	75.8 (63.4–91.3)	0.63
Enhanced volume, mm3	25.1 (9.8–41.8)	10.1 (4.9–23.6)	0.01	11.3 (4.4–21.6)	6.8 (1.9–12.2)	<0.01

Values are presented as median (interquartile range). Within‐group changes were assessed using the Wilcoxon signed‐rank test. ALT indicates alanine aminotransferase; AST, aspartate transaminase; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipoprotein cholesterol; MRI, magnetic resonance imaging; and PCSK9i, proprotein convertase subtilisin/kexin type 9 inhibitor.

Stenosis was significantly reduced after 6 months in the PCSK9i group (85.6% to 75.7%, P<0.01), while no significant change was observed in the control group (77.5% to 75.8%, P=0.63). Plaque enhancement volume decreased significantly in both groups: from 25.1 to 10.1 mm³ in the PCSK9i group (P=0.01), and from 11.3 to 6. 8 mm³ in the control group (P<0.01). As shown in Figure 3, LDL‐C levels were significantly lower in the PCSK9i group compared with controls at both 3 and 6 months.

Figure 3. Longitudinal changes in LDL‐C levels over time in the PCSK9 inhibitor and control groups.

Boxplots illustrate the distribution of LDL‐C levels at baseline, 3 months, and 6 months for each group. The PCSK9i group showed a significant and sustained reduction in LDL‐C at both 3 and 6 months compared with the control group (*p < 0.01 for both time points). Boxes represent the IQR, the horizontal line within each box indicates the median, and whiskers extend to the minimum and maximum values within 1.5× IQR. IQR indicate interquartile range; LDL‐C, low‐density lipoprotein cholesterol; and PCSK9i, proprotein convertase subtilisin/kexin type 9 inhibitor.

Correlation Between LDL‐C and Stenosis Improvement

A significant inverse correlation was observed between LDL‐C levels at 3 months and stenosis reduction at 6 months (Spearman ρ=−0.47, P<0.01), indicating that lower LDL‐C levels were associated with greater stenosis improvement (Figure 4). Additionally, absolute LDL‐C reduction (baseline to 3 months) was positively correlated with stenosis reduction (Spearman ρ=0.45, P<0.01), further supporting the impact of lipid lowering on plaque regression.

Figure 4. Association Between LDL‐C levels and intracranial stenosis reduction at 6 months.

A, Correlation between post‐treatment LDL‐C levels at 3 months and absolute stenosis reduction (percentage points). B, Correlation between absolute LDL‐C reduction and absolute stenosis reduction (percentage points). Each dot represents an individual patient, with red indicating the intervention group (PCSK9i plus statin) and blue indicating the control group (statin alone). Linear regression lines are displayed for both groups, with shaded areas indicating the 95% CIs of the regression estimates. A significant inverse correlation was observed between post‐treatment LDL‐C levels and stenosis reduction (Spearman ρ=−0.47, P<0.01), suggesting that lower LDL‐C is associated with greater plaque regression. A significant positive correlation was found between absolute LDL‐C reduction and stenosis improvement (Spearman ρ=0.45, P<0.01), indicating that greater LDL‐C reduction is associated with more pronounced vascular benefit. LDL‐C indicates low‐density lipoprotein cholesterol; and PCSK9i, proprotein convertase subtilisin/kexin type 9 inhibitor.

DISCUSSION

This randomized controlled trial provides compelling evidence that adjunctive therapy with a PCSK9i significantly reduces intracranial stenosis and LDL‐C levels in patients with symptomatic ICAS relative to standard lipid‐lowering therapy alone. These findings align with prior studies, underscoring that PCSK9i achieves substantial LDL‐C reductions and may promote plaque stabilization, thereby reducing the risk of early stroke recurrence.

Compelling evidence from recent coronary imaging trials examining PCSK9i combined with statin therapy demonstrates a consistent, linear association between achieved LDL‐C levels and favorable coronary plaque changes. ²¹ , ²² , ²³ These changes include reductions in percent atheroma volume and increases in fibrous cap thickness. ²¹ Our results are further corroborated by recent intracranial‐specific studies. In a prospective single‐arm study, Zeng et al. reported significant reductions in both stenosis severity and plaque enhancement after 24 weeks of combined high‐intensity statin and evolocumab therapy. ²⁴ Importantly, a strong correlation was observed between LDL‐C reductions and plaque contrast volume, reinforcing the concept that PCSK9i promotes intracranial plaque stabilization and vascular remodeling. Similarly, Wu et al. demonstrated that combined statins and PCSK9i significantly reduced plaque enhancement volume and stenosis degree in the observation study. ²⁵ More recently, a cohort study showed that, despite plaque regression after intensive lipid‐lowering therapy, shrinkage of the outer‐wall boundary area was associated with a higher risk of recurrent ipsilateral stroke, suggesting prognostic relevance beyond conventional stenosis measurements. ²⁶ Overall, these data reinforce the potential of aggressive lipid‐lowering therapy to stabilize intracranial plaques and reduce recurrent cerebrovascular events.

It is important to contextualize our findings within the current framework of lipid management guidelines. Current stroke‐prevention guidelines recommend LDL‐C < 70 mg/dL for patients with ischemic stroke of atherosclerotic origin, yet do not specify thresholds for ICAS. ¹⁰ , ²⁷ Notably, the 2022 European Stroke Organization guidelines on treatment of patients with ICAS suggest that isolated symptomatic ICAS should be considered a very‐high‐risk condition—targeting LDL‐C levels below 55 mg/dL. ²⁸ Our study provides valuable preliminary evidence that such aggressive lipid targets may be particularly beneficial in this population and warrant further investigation.

Both treatment groups in our trial showed significant reductions in plaque enhancement volume; however, between‐group differences were not significant. Plaque enhancement reflects neovascularization, inflammation, and endothelial dysfunction ²⁹ —features linked to plaque instability and heightened stroke risk. ³⁰ , ³¹ , ³² These findings suggest that statins alone may confer strong anti‐inflammatory effects sufficient to reduce enhancement volume. ¹⁵ , ¹⁷ Prior research has shown that statins can stabilize plaque independent of LDL‐C lowering. ³³ Therefore, the similar reductions in enhancement volume between groups likely reflect the robust anti‐inflammatory activity of statins, reinforcing their foundational role in ICAS treatment.

In a previous symptomatic ICAS trial, PCSK9i add‐on therapy further lowered LDL‐C levels and reduced early recurrent stroke risk in patients with symptomatic ICAS, particularly within the first month poststroke. ³⁴ A recent randomized subgroup analysis showed that evolocumab plus atorvastatin reduced early neurological deterioration and improved 90‐day functional outcomes in large artery atherosclerosis, but not small vessel occlusion, suggesting that anti‐inflammatory effects, reflected by greater IL‐6 reduction, may contribute beyond LDL‐C lowering. ³⁵ However, our study did not detect significant differences in recurrent stroke or major cardiovascular events between groups, likely attributable to limited sample size and follow‐up. The relatively high recurrent stroke rate in the control group may reflect the high‐risk nature of our study population. Unlike broader observational cohorts, ² which included both TIA and ischemic stroke, our trial enrolled only patients with diffusion‐weighted imaging‐confirmed acute infarction within 7 days attributable to symptomatic ICAS, thereby enriching for more active disease. This is clinically relevant because prior data showed a higher 12‐month recurrent stroke rate in patients presenting with acute infarction than with TIA (15% versus 4%). ³⁶ In addition, all participants had symptomatic ICAS confirmed by high‐resolution vessel‐wall MRI, and both groups had substantial baseline plaque enhancement, suggesting active culprit plaques at high short‐term risk of recurrence. Thus, the higher recurrence rate in our control group likely reflects enrichment for patients with poorer short‐term prognosis rather than an unexpectedly unfavorable sample.

It remains uncertain whether PCSK9i exerts plaque‐stabilizing effects independent of LDL‐C lowering. In our study, the number of PCSK9i‐treated patients who did not achieve LDL‐C < 55 mg/dL was too small to permit a reliable subgroup analysis. Instead, correlation analyses across the full cohort showed that lower LDL‐C levels at 3 months and greater absolute LDL‐C reduction were associated with greater stenosis improvement, with a similar directional trend observed in both the control and PCSK9i groups, supporting the interpretation that LDL lowering is likely the main driver of the imaging effect.

Nevertheless, the safety profile of PCSK9i was reaffirmed, with low adverse event rates and no serious safety concerns, consistent with prior cardiovascular outcome trials. Trials such as SAMMPRIS (Stenting and Aggressive Medical Management for Preventing Recurrent Stroke in Intracranial Stenosis) and CASSISS (China Angioplasty & Stenting for Symptomatic Intracranial Severe Stenosis) underscore the challenges in managing ICAS, marked by high stroke recurrence and uncertain benefit from invasive interventions like stenting. ¹¹ , ³⁷ , ³⁸ These findings support the value of optimized medical therapy, including dual antiplatelet and intensive lipid‐lowering regimens. Our study adds evidence for PCSK9i as adjunctive therapy to statins in comprehensive ICAS care. Larger trials with extended follow‐up are needed to assess long‐term outcomes and define optimal treatment targets.

Limitations

Several limitations warrant careful consideration. First, the small sample size and six‐month follow‐up limit power to detect differences in clinical outcomes. Although we observed significant imaging changes, larger studies with longer follow‐up are required to confirm the clinical significance of plaque regression. Prior ICAS trials (eg, SAMMPRIS, CASSISS) highlight the importance of sufficient sample sizes and extended follow‐up to capture infrequent or delayed events. ¹¹ , ³⁷ , ³⁸

Second, the open‐label design may have introduced bias. Although imaging assessors were blinded, the lack of patient and clinician blinding could have influenced treatment adherence or clinical management. Future double‐blind, placebo‐controlled studies would reduce this risk. In addition, baseline stenosis was numerically higher in the intervention group; although the primary analysis adjusted for baseline stenosis and sensitivity analyses were consistent, some influence of baseline imbalance or regression to the mean cannot be completely excluded. Ezetimibe use was also imbalanced between groups because more patients in the control group required add‐on lipid‐lowering therapy to achieve LDL‐C targets. Although a sensitivity analysis adjusting for ezetimibe use showed that the treatment effect of PCSK9i on stenosis reduction remained significant, ezetimibe use was not stratified at randomization, and residual confounding cannot be fully excluded. Furthermore, although imaging readers were blinded to treatment allocation and clinical information, blinding to imaging time point was not formally prespecified, which may have introduced assessment bias. Because the primary outcome was expressed as change in stenosis, the same absolute luminal change may correspond to different percentage changes across vessels or arterial segments with different baseline calibers. Although we used baseline‐adjusted rank‐based analysis of covariance and normalized stenosis to a reference lumen area, this mathematical sensitivity remains a limitation.

Third, although high‐resolution vessel‐wall MRI is a well‐established method for evaluating intracranial plaque morphology, it has inherent limitations. Vessel‐wall imaging lacks direct histopathological validation; therefore, MRI findings such as plaque enhancement or regression cannot be conclusively linked to specific histological processes such as neovascularization or inflammation. ³² , ³⁹ Multimodal imaging and pathological validation studies would improve interpretation of these MRI‐based findings. ⁴⁰ In addition, measurement of luminal stenosis is subject to spatial resolution constraints and inter‐observer variability. Although stenosis measurements were performed independently by 2 readers blinded to treatment allocation and were averaged to reduce random measurement error, Bland–Altman analysis showed relatively wide limits of agreement for individual measurements. Therefore, the observed between‐group difference in stenosis reduction should be interpreted cautiously.

Furthermore, generalizability may be limited by the trial's selective eligibility criteria and enrollment pattern. Patients with renal or hepatic dysfunction, prior intolerance to statins or PCSK9i, or contraindications to MRI were excluded. Importantly, many screened patients were also excluded because baseline LDL‐C levels were below 100 mg/dL, which may limit the applicability of these findings to broader ICAS populations with less severe dyslipidemia. Larger ongoing studies like PISTIAS, which employ broader eligibility criteria, may provide further insights into the generalizability of PCSK9i therapy in ICAS. ⁴¹ Finally, the optimal LDL‐C target in ICAS remains uncertain. While our findings support targeting LDL‐C below 55 mg/dL, current stroke‐prevention guidelines generally recommend an LDL‐C target <70 mg/dL for patients with ischemic stroke of atherosclerotic origin, ¹⁰ , ²⁷ although more recent dyslipidemia guidelines have increasingly supported lower targets in selected very‐high‐risk atherosclerotic cardiovascular disease populations. ²⁷ , ⁴² , ⁴³ , ⁴⁴ Future trials designed to compare different LDL‐C targets in ICAS may provide important guidance for clinical practice.

In summary, while this trial provides promising evidence for PCSK9i therapy in symptomatic ICAS, these limitations highlight important areas for future research. Larger, rigorously designed randomized trials incorporating multimodal imaging, longer follow‐up, and broader populations are essential to validate and expand upon our findings.

CONCLUSIONS

The addition of PCSK9i to high‐intensity statin therapy significantly reduced intracranial stenosis and improved LDL‐C goal achievement in patients with symptomatic ICAS, with encouraging signs of plaque stabilization. While both regimens appeared effective in reducing plaque enhancement, PCSK9i conferred added benefit in stenosis regression. These findings support the potential of PCSK9i in ICAS management and highlight the need for large‐scale trials to confirm long‐term efficacy and clarify LDL‐C targets in this high‐risk population.

Sources of Funding

The TOPICAL‐MRI trial was funded by National Science and Technology Council (Taiwan) (MOST 110‐2314‐B‐182‐056, 111‐2314‐B‐182‐055, 112‐2314‐B‐182‐041, 113‐2314‐B‐182‐061) and Chang Gung Memorial Hospital (CMRPG6P0231).

Disclosures

None.

Supporting information

CONSORT 2025 Checklist

Tables S1–S3

Figure S1