Several established resources provide evidence-based guidance for implementing pharmacogenomic testing in stroke care. The Clinical Pharmacogenetics Implementation Consortium (CPIC) serves as a leading authority, publishing regularly updated guidelines that translate genetic test results into actionable prescribing recommendations [9].

The Pharmacogenomics Knowledge Base (PharmGKB) complements CPIC by grading the strength of gene-drug associations from Level 1A (highest) to Level 4 (lowest), thereby enabling clinicians to assess the robustness of available evidence efficiently [10,11]. Currently, Level 1A evidence supports testing for three drug classes highly relevant to stroke management: clopidogrel, warfarin, and statins. CPIC’s genotype-based prescribing recommendations for antiplatelet and anticoagulant therapy, emphasizing efficacy, are summarized in Table 1, and recommendations for statin therapy and statin-associated muscle symptoms (SAMS) are provided in Table 2.

Recently, CPIC and PharmGKB were integrated into a unified platform, ClinPGx (https://www.clinpgx.org/), to streamline access and harmonize pharmacogenomic guidance. While CPIC guidelines primarily focus on pharmacokinetics—especially drug metabolism pathways (the “M” in Absorption, Distribution, Metabolism, Excretion [ADME])—PharmGKB has traditionally cataloged a broader range of evidence, including pharmacodynamic associations and genetic determinants of adverse drug reactions. This distinction highlights the complementary value of these resources: CPIC provides explicit prescribing recommendations for well-validated pharmacokinetic variants, whereas PharmGKB includes pharmacodynamic and safety-related variants that may be highly relevant to stroke medications but are not yet incorporated into formal guidelines.

In addition, the Dutch Pharmacogenetics Working Group (DPWG) issues prescribing recommendations that sometimes differ from those of CPIC, thereby offering alternative perspectives that may assist clinical decision-making [12]. For example, therapeutic recommendations synthesized in CPIC guidelines are largely based on expert consensus, whereas DPWG relies more heavily on quantitative pharmacokinetic data from studies of good or moderate quality [13]. Taken together, these international frameworks provide a coherent foundation for integrating pharmacogenomics into clinical practice and reinforce the growing consensus regarding its clinical significance.

The Ticagrelor or Clopidogrel with Aspirin in High-Risk Patients with Acute Nondisabling Cerebrovascular Events II (CHANCE-2) trial marked a milestone in stroke PGx, providing the first largescale validation of genotype-guided antiplatelet therapy in acute stroke [14]. In this trial, 6,412 patients with minor ischemic stroke or high-risk transient ischemic attack (TIA) in China were randomized to receive ticagrelor or clopidogrel based on their CYP2C19 LOF carrier status.

Among LOF carriers, ticagrelor reduced the risk of recurrent stroke by 34% (risk ratio 0.66; 95% confidence interval [CI], 0.48-0.99; P=0.009), confirming the clinical benefit of genotype-guided therapy. These findings strengthened the rationale for CYP2C19 testing in antiplatelet selection and directly informed updates to regional stroke guidelines in East Asia [15].

Importantly, the CHANCE-2 trial demonstrated how pharmacogenomic information can be integrated into time-sensitive clinical workflows, indicating that genetic testing can be both actionable and feasible at the point of care. This trial therefore, serves as a model for future research evaluating PGx in acute stroke management, although its genetic testing platform has not yet been widely implemented outside China.

To synthesize current knowledge, a systematic literature review was conducted focusing on pharmacogenomic studies relevant to medications commonly used in stroke care. Priority was given to high-quality evidence incorporated into CPIC guidelines or classified as Level 1 or 2 in PharmGKB, with particular emphasis on studies conducted in East Asian populations.

This review concentrated on three major drug classes—antiplatelet agents (particularly clopidogrel), anticoagulants (primarily warfarin), and statins—for which substantial pharmacogenomic evidence supports clinical utility. Additional stroke-related medications, including aspirin and direct oral anticoagulants (DOACs), were also evaluated, although robust pharmacogenomic evidence for these agents remains limited despite their central role in current stroke management.

Of note, substantial knowledge gaps remain in other drug classes, such as antihypertensive and antidiabetic agents, despite their widespread use in stroke prevention. For example, only metoprolol demonstrates modest pharmacogenomic associations through CYP2D6 variants [16]. This paucity of evidence is particularly concerning given recent findings from the NIH All of Us Research Program, which identified racial differences in drug responses to commonly used antihypertensives [17]. These gaps are particularly pronounced among East Asian populations, including Koreans, highlighting the need for population-specific pharmacogenomic studies to inform clinical practice in this regional context.

To address this issue, allele frequency data from Korean populations were incorporated using the ongoing GENomic Exploration in Stroke and pharmacogenomic Innovation Study-Korea (GENESIS-K) cohort, highlighting areas where populationtailored pharmacogenomic guidance is urgently required. This approach aimed to contextualize global evidence within the East Asian clinical framework.

Clopidogrel is a widely used antiplatelet agent in stroke pre-vention, but its effectiveness varies considerably due to genetic variations that influence its metabolism. As a prodrug, clopidogrel requires two-step hepatic bioactivation—primarily mediated by CYP2C19—to be converted into its active metabolite [18]. This metabolite irreversibly inhibits the platelet P2Y12 receptor, thereby reducing platelet aggregation.

Genetic polymorphisms in CYP2C19 are strongly associated with clopidogrel’s therapeutic efficacy.6 The most clinically relevant variants include the LOF alleles (*2, *3) and the gain-offunction (GOF) allele (*17). Approximately 30% of the general population carries at least one LOF allele, whereas the prevalence is considerably higher in East Asians, reaching up to 60%, compared with a lower prevalence in other ethnic groups (11%-19% in Europeans and 11%-17% in Africans) (Table 3) [19].

The LOF alleles reduce the formation of the active metabolite by 25%-40%, resulting in diminished platelet inhibition and an increased risk of recurrent vascular events [20]. Meta-analyses have shown that carriers of *2/*2 or *2/*3 genotypes have a 1.5- to 3.5-fold higher risk of stroke or other cardiovascular events compared with non-carriers [21]. In contrast, carriers of the *17 GOF allele exhibit enhanced response to clopidogrel but face a higher risk of bleeding [22]. Consistent with these findings, a Korean multicenter prospective study reported a higher incidence of composite vascular outcomes among LOF carriers (2.7% vs. 1.6%) [23].

The CHANCE-2 trial provided definitive evidence supporting genotype-guided antiplatelet therapy [14]. In this randomized trial involving 6,412 Chinese patients with minor ischemic stroke or TIA, ticagrelor was superior to clopidogrel among CYP2C19 LOF allele carriers, resulting in a 34% reduction in stroke recurrence. These findings informed updates to East Asian stroke guidelines, including those of the Chinese Stroke Association, which now recommend genotype-guided therapy for selected patients [15].

Beyond CYP2C19, several other genes have been investigated for their potential influence on clopidogrel pharmacokinetics and pharmacodynamics, although their clinical relevance remains less well established:

- ABCB1: Affects drug absorption and distribution [24].

- CES1: Participates in hydrolysis of the prodrug [25].

- PON1: May contribute to clopidogrel activation [26].

- P2Y12: Modifies target receptor function [27].

- CYP2B6/CYP2C9: May influence alternative metabolic pathways [20].

Currently, only CYP2C19 has sufficient evidence for clinical actionability and is recommended by CPIC and PharmGKB for genotype-guided therapy (Table 1). Other candidate genes remain under investigation.

Aspirin remains a cornerstone of both primary and secondary stroke prevention. It irreversibly inhibits cyclooxygenase-1 (COX-1) and cyclooxygenase-2 (COX-2), which are encoded by the PTGS1 and PTGS2 genes, thereby reducing thromboxane A2- mediated platelet aggregation [28]. Although generally effective, substantial interindividual variability in aspirin response has been observed, and genetic factors may partially account for this variation [29].

Variants in PTGS1 and PTGS2 may alter the antiplatelet efficacy of aspirin. For instance, the PTGS1 -842A>G polymorphism has been associated with reduced COX-1 expression and increased cardiovascular risk (hazard ratio 2.57; 95% CI, 1.87- 3.54) [30]. Similarly, the PTGS2 -765G>C variant may influence COX-2 activity and has been linked to an attenuated antiplatelet response [31]. The PEAR1 rs12041331 A allele, which encodes platelet endothelial aggregation receptor 1 [32], has also been associated with increased cardiovascular risk, although findings are inconsistent and have not yet been validated in stroke-specific cohorts [33].

Genetic factors may also influence aspirin safety. HLA-DPB1 variants have been associated with aspirin-induced asthma, with HLA-DPB1*03:01 increasing susceptibility and HLA-DPB1*04:01 conferring protection [34,35]. Additionally, UGT1A6 AA polymorphisms have been linked to altered aspirin metabolism and differential long-term colorectal cancer risk reduction [31].

Despite these associations, no genetic markers related to aspirin currently have sufficient evidence for clinical implementation. According to CPIC and PharmGKB classifications, routine pharmacogenomic-guided aspirin use is not recommended.

Warfarin remains an important agent for stroke prevention in patients at risk of cardioembolic stroke, including those with atrial fibrillation—particularly in settings where DOACs are contraindicated or unavailable [36]. However, warfarin therapy exhibits substantial interindividual variability in dose requirements and bleeding risk, much of which can be attributed to genetic variation [37].

The three principal pharmacogenes associated with warfarin response are CYP2C9, VKORC1, and CYP4F2 [38]. The CYP2C9 gene encodes the primary enzyme responsible for warfarin metabolism. The most common variants, CYP2C9*2 and CYP2C9*3 alleles, reduce enzymatic activity, resulting in slower warfarin clearance and an elevated risk of bleeding [39]. Carriers of these variants typically require lower maintenance doses and longer periods to achieve a stable international normalized ratio (INR) [39].

The VKORC1 gene encodes the warfarin target enzyme, which catalyzes the reduction of vitamin K epoxide to its active form [40]. The common -1639G>A variant (VKORC1 A allele) decreases gene expression and increases warfarin sensitivity [40]. Homozygous carriers (AA genotype) generally require approximately 50% lower doses compared with wild-type individuals [40]. Population-based studies have identified significant ethnic differences in VKORC1 allele frequencies, partially explaining interethnic variability in warfarin dose requirements across populations (Table 3) [41].

CYP4F2 variants, particularly c.1297G>A (rs2108622), affect vitamin K metabolism, with carriers typically requiring higher warfarin doses [42]. The combined effects of variants in these three genes, along with clinical factors, can lead to as much as a 20-fold difference in warfarin dose requirements among individuals [38].

Warfarin dosing algorithms that integrate CYP2C9 and VKORC1 genotypes, along with clinical factors such as age, body size, and concomitant medications, have been developed and validated. The International Warfarin Pharmacogenetics Consortium [43] and Gage algorithms [44] are among the most widely used. Several clinical trials—Clarification of Optimal Anticoagulation through Genetics (COAG), European Pharmacogenetics of Anticoagulant Therapy (EU-PACT), and Genetic Informatics Trial (GIFT)—have evaluated genotype-guided dosing [45-47]. Although results have been mixed, the GIFT trial demonstrated reduced bleeding and thromboembolic events in the genotype-guided group, supporting its clinical utility in selected populations [46]. Similarly, in the EU-PACT trial, pharmacogenetic-guided dosing was associated with a greater percentage of time within the therapeutic INR range and significantly fewer episodes of excessive anticoagulation (INR ≥4.0) compared with standard dosing [47]. Accordingly, CPIC guidelines recommend incorporating CYP2C9 and VKORC1 genotypes into initial dosing whenever available (Table 1) [7].

However, several challenges persist in translating these findings into clinical practice, particularly for Asian populations. First, most dosing algorithms have been derived from predominantly European cohorts and may overestimate warfarin dose requirements in East Asians. Second, East Asians, including Koreans, have a substantially higher frequency of the VKORC1-1639A allele and CYP2C9 *3/*3 genotypes, which are associated with markedly reduced warfarin dose needs (often <2 mg/day) [48]. Third, minor alleles in other genes, such as CYP4F2 and GGCX, although less influential, may further modulate dosing requirements in specific populations [7,49]. Finally, the risk of warfarin-associated intracranial hemorrhage is significantly higher among Asians, emphasizing the importance of precise dosing and vigilant INR monitoring [50].

Only a limited number of studies have evaluated genotypeguided warfarin dosing in Korean populations, demonstrating that existing algorithms frequently overestimate dose requirements and do not adequately account for the high prevalence of ultra-sensitive genotypes [51]. These findings underscore the necessity for validated, population-specific dosing models that integrate Korean-specific allele frequencies and clinical outcomes (Table 3).

DOACs have become a cornerstone of thromboembolic event prevention in patients with atrial fibrillation and other high-risk conditions [52,53]. In Korea, the prescription rate of DOACs for patients with acute ischemic stroke and atrial fibrillation increased from 4.6% in 2013 to 73.9% in 2019 [54]. Unlike warfarin, DOACs exhibit more predictable pharmacokinetic profiles and generally do not require routine coagulation monitoring. Their mechanisms of action include direct thrombin inhibition by dabigatran [55] and factor Xa inhibition by apixaban, edoxaban, and rivaroxaban [56-58].

Dabigatran is administered as a prodrug (dabigatran etexilate mesylate) and is converted to its active form by intestinal and hepatic carboxylesterases (CES1). In contrast, apixaban, edoxaban, and rivaroxaban are administered in their active forms. All DOACs are substrates of the efflux transporter P-glycoprotein, and their plasma concentrations can be influenced by P-glycoprotein inhibitors, which may alter both anticoagulant efficacy and bleeding risk [55-58].

Genetic variations may contribute to interindividual variability in DOAC pharmacokinetics and pharmacodynamics, although their clinical significance remains under investigation. Key genes implicated include ABCB1, which encodes P-glycoprotein; CES1, responsible for dabigatran activation; and CYP3A4/5, involved in the metabolism of factor Xa inhibitors, especially apixaban and rivaroxaban [59-65].

Polymorphisms in ABCB1, such as rs4148738 and rs1045642, have been associated with increased peak plasma concentrations of dabigatran, apixaban, and rivaroxaban, and in some studies, with an elevated bleeding risk [59,60]. However, associations with clinical outcomes, including thromboembolism and major bleeding, remain inconclusive. Similarly, CYP3A5*3, a LOF variant, has been linked to higher plasma levels of apixaban, although findings have been inconsistent across studies [61,62].

For dabigatran, the CES1 rs2244613 variant has been associated with lower plasma concentrations and a reduced bleeding risk [60]. Other CES1 variants, including rs8192935 and rs71647871, have demonstrated similar associations; however, their clinical utility has not yet been clearly established [63-65].

To date, no pharmacogenomic markers associated with DOACs have been incorporated into clinical guidelines, including those from the CPIC or the DPWG, owing to limited and inconsistent evidence. Future research should prioritize validating candidate variants in large, ethnically diverse cohorts and evaluating their additive or synergistic effects through polygenic modeling approaches.

While pharmacogenomic associations are well established for several stroke medications, clinicians should remain mindful of key limitations in translating genetic findings into clinical outcomes. First, genetic variants explain only part of the interindividual variability in treatment response: CYP2C19 polymorphisms account for approximately 12% of clopidogrel response variability [82], whereas VKORC1 and CYP2C9 variants together explain about 30%-50% of warfarin dose requirements [7,38]. Second, the pathway from genotype to clinical benefit involves several sequential steps—altered drug metabolism (pharmacokinetics), modified drug effects (pharmacodynamics), intermediate outcomes (e.g., platelet inhibition or INR control), and clinical endpoints (e.g., stroke recurrence or bleeding)—with differing levels of evidence supporting each stage [83]. For the CYP2C19-clopidogrel association, this full causal pathway has been validated in randomized clinical trials, most notably CHANCE-2, which demonstrated improved outcomes with genotype-guided therapy [14]. However, for aspirin and DOACs, current evidence remains limited to pharmacokinetic or pharmacodynamic associations, without robust outcome data confirming clinical benefit. Third, functional assays that assess drug effects in real time can complement pharmacogenomic testing. Platelet function testing may assist in evaluating antiplatelet response, particularly in intermediate metabolizers where the clinical implications of genetic results remain uncertain [84]. For warfarin, genotype-guided initial dosing should always be used alongside INR monitoring rather than as a replacement [7]. In the case of DOACs, plasma concentrations or anti-factor Xa activity may provide supplementary information in select high-risk scenarios, although routine laboratory monitoring is not currently recommended [85]. Clinicians should interpret pharmacogenomic results with an understanding that genetic data represent one important dimension of personalized therapy, not a sole determinant of drug response.

An additional challenge arises when pharmacogenomic evidence and conventional clinical trial data do not align perfectly. Although CYP2C19 LOF alleles clearly predict reduced clopidogrel responsiveness, alternative P2Y12 inhibitors such as ticagrelor and prasugrel have been evaluated primarily in non-genotype-selected cohorts, often with higher bleeding risk [86-88]. The CHANCE-2 trial confirmed the clinical benefit of ticagrelor in carriers of LOF alleles; however, clinicians must balance this benefit against individual bleeding risk and comorbid conditions [14]. Patients with high thrombotic risk and low bleeding risk are likely to benefit most from genotype-guided selection of alternative antiplatelet therapy, whereas those with a history of bleeding may remain better candidates for clopidogrel, even when harboring LOF alleles [89]. This scenario illustrates a broader principle: pharmacogenomic data should inform, not replace, clinical judgment, emphasizing the need for integration with conventional risk assessment and individualized patient characteristics.

The integration of pharmacogenomic testing into stroke care presents both substantial opportunities and notable challenges. To achieve successful clinical implementation, several interdependent components must be addressed, including cost-effectiveness, testing logistics, clinician education, and healthcare infrastructure development.

Cost-effectiveness remains a key determinant, particularly within resource-limited healthcare systems. Although some studies have demonstrated the economic value of pharmacogenomic testing—such as CYP2C19 genotyping to guide antiplatelet therapy in high-risk patients [90]—the overall evidence remains heterogeneous across healthcare settings [91]. However, a growing body of research suggests that CYP2C19 genotype-guided therapy for clopidogrel is cost-effective and, in some analyses, represents a dominant strategy (both less costly and more effective) [92]. Evidence indicates that the expenses associated with pharmacogenomic testing may be offset by reductions in downstream healthcare costs through the prevention of adverse cardiovascular events such as stroke, myocardial infarction, and stent thrombosis among poor metabolizers [93,94].

Timeliness of testing is another key factor, especially in acute stroke management where therapeutic decisions must be made rapidly. Conventional genotyping, which often requires several days to weeks for results, is incompatible with real-time clinical decision-making. Point-of-care (POC) genotyping platforms have emerged as a potential solution, enabling same-day results that can guide clopidogrel selection. However, their widespread adoption necessitates integration into clinical workflows, structured staff training, and rigorous quality control procedures [95,96]. Despite these advantages, the clinical accessibility of such platforms remains substantially limited. Currently, laboratory-based CYP2C19 genotyping is commercially available from multiple vendors with turnaround times ranging from 1 to 7 days. In contrast, POC systems validated in research contexts—such as the GMEX platform used in the CHANCE-2 trial [95,96]—have yet to achieve broad commercial distribution due to regulatory, logistical, and market-related constraints. Reimbursement policies further restrict implementation. In the United States, Medicare provides coverage for CYP2C19 testing in patients undergoing percutaneous coronary intervention; however, coverage for strokerelated indications varies across regional contractors and private insurers [97]. In Korea, the National Health Insurance reimburses CYP2C19 testing for patients with acute coronary syndrome or stroke requiring antiplatelet therapy, although specific eligibility criteria apply [98]. While pharmacogenomic testing for warfarin (VKORC1, CYP2C9) is reimbursed in some healthcare systems, its clinical relevance has declined with the increasing preference for DOACs [54,98]. Conversely, statin testing targeting SLCO1B1 variants has yet to be incorporated into reimbursement frameworks, despite CPIC recommendations supporting its clinical utility [99]. The persistent discrepancy between favorable cost-effectiveness data—particularly for CYP2C19 testing in populations with high frequencies of LOF alleles—and limited reimbursement policies represents a major implementation barrier that will require coordinated policy reform in parallel with technological progress.

Workforce education remains essential [100,101]. Clinicians must be adequately trained to interpret pharmacogenomic results and apply them appropriately, which underscores the importance of integrating PGx into medical education and continuing professional development programs.

Effective integration also requires adaptation of electronic health record systems to store and retrieve pharmacogenomic data and to link this information with CDSS [102]. These tools must provide clear, actionable recommendations while minimizing alert fatigue to ensure clinical usability [103]. Standardized reporting formats and transparent, interpretable algorithms will be essential to maintain consistency and scalability across institutions.

Equally important, shared decision-making between clinicians and patients is critical when implementing pharmacogenomic testing—especially in situations where genetic associations are well established but clinical outcome evidence remains limited [104,105]. Patients should be informed of the potential benefits and limitations of genotype-guided therapy, alternative treatment options, and associated costs to ensure that decisions align with individual values and preferences.

Quality assurance frameworks are also vital. Institutions should establish standardized protocols for assay validation, specimen handling, and result interpretation [106,107]. Continuous evaluation of testing accuracy, clinical outcomes, and adverse drug reactions will facilitate ongoing improvement in PGx-guided care [108,109].

Despite technological and guideline advancements, real-world adoption in stroke care remains modest. Barriers include fragmented infrastructure, variable reimbursement, and limited clinician familiarity. A stepwise implementation strategy—starting with high-yield use cases such as CYP2C19 testing—may represent the most practical path forward. Strong institutional leadership, sustainable funding mechanisms, and interdisciplinary collaboration will be critical to transitioning PGx from pilot programs into routine clinical care.

Polygenic risk scores (PRS) are emerging as promising tools in precision medicine, and their extension to PGx-PRS has garnered increasing attention [110,111]. Unlike single-gene approaches, PGx-PRS combine multiple variants across pharmacokinetic, pharmacodynamic, and target-response pathways, thereby enhancing the prediction of interindividual variability in both drug efficacy and toxicity.

In the context of stroke care, PGx-PRS may be particularly useful for therapies where drug response is multifactorial and cannot be fully explained by individual genetic variants. For instance, in antiplatelet therapy, integrating variants in CYP2C19, ABCB1, and PON1 into a composite score may improve the prediction of clopidogrel resistance compared with CYP2C19 genotyping alone [112,113]. Preliminary analyses from CHANCE and CHANCE-2 indicate that PGx-PRS can stratify clopidogrel response more effectively than single-gene testing.

Similarly, PGx-PRS could refine warfarin dosing algorithms by accounting for cumulative genetic contributions from CYP2C9, VKORC1, CYP4F2, and additional loci involved in vitamin K metabolism and coagulation. Although current dosing models incorporate some of these genes, PRS-based frameworks may further improve dose accuracy by capturing gene-gene interactions and subtle additive effects [43,44,114].

However, the development of clinically robust PGx-PRS remains constrained by methodological and translational challenges [110,111,115]. Large-scale, multiethnic genome-wide studies are required to identify predictive variants, optimize score calibration, and validate models across diverse populations, considering significant interethnic differences in allele frequencies and linkage disequilibrium structures. Incorporating deep learning approaches may further enhance PGx-PRS by uncovering nonlinear genetic interactions and complex multi-locus effects [116-118].

A major limitation is the absence of reliable biomarkers for drug response in certain therapeutic classes. For example, while low-density lipoprotein cholesterol levels or blood pressure can serve as pharmacodynamic markers for statins and antihypertensive agents, antiplatelet response lacks standardized and clinically validated surrogate endpoints. Platelet function assays—such as P2Y12 reaction units—are available but not universally accepted, which complicates the development and calibration of PGx-PRS models [119]. In such cases, validation based directly on clinical outcomes may be required to ensure clinical relevance.

In summary, PGx-PRS represent a scalable and promising framework for advancing personalized pharmacotherapy in stroke. Although the field remains in its early stages, ongoing research and emerging analytic technologies are likely to broaden their applicability across multiple drug classes. Crucially, validation in underrepresented populations remains indispensable. Large-scale East Asian cohorts—such as the GENESIS-K study in Korea—can play a pivotal role in developing and validating population-specific PRS, thereby ensuring both regional applicability and equity in precision stroke medicine.

Population-specific genetic variation plays a pivotal role in PGx and must be incorporated into the implementation of precision stroke therapy. The frequency and functional consequences of pharmacogenomic variants differ markedly across ancestral groups (Table 3) [120,121], and neglecting these differences may result in suboptimal therapeutic efficacy or increased risk of adverse drug reactions.

A prominent example is the high prevalence of CYP2C19 LOF alleles among East Asian populations, which underpins the demonstrated benefit of genotype-guided antiplatelet therapy in the CHANCE-2 trial. These findings prompted updates to regional clinical guidelines—such as those issued by the Chinese Stroke Association—recommending CYP2C19 genotyping to guide clopidogrel use in appropriately selected patients [15]. In contrast, such genotype-based strategies have not been widely adopted in Western populations, where LOF allele frequencies are lower and ethnicity-specific evidence remains limited.

Beyond antiplatelet therapy, population-tailored pharmacogenomic data remain sparse for other stroke-related drug classes. For instance, warfarin dosing algorithms specifically optimized for patients of African ancestry have been developed [7], yet comparable ethnicity-specific frameworks are lacking for statins, aspirin, and DOACs, despite clear interethnic differences in allele frequencies and pharmacologic response patterns.

This paucity of population-representative data reflects a broader inequity in pharmacogenomic research participation. Recent analyses reveal that over 60% of participants in PharmGKB-referenced studies are of European ancestry, compared with only 28% of East Asian ancestry, and an even smaller proportion of African, South Asian, or Indigenous descent [122]. Such underrepresentation undermines the generalizability of existing evidence and risks reinforcing disparities in precision medicine implementation across global populations.

To address this imbalance, future pharmacogenomic research must prioritize underrepresented populations [110,111,123]. Research efforts should emphasize allele frequency mapping, clinical validation of gene-drug associations, and the development of population-specific prescribing guidelines. Large-scale biobanks—such as the GENESIS-K cohort in Korea—are essential for generating regionally relevant evidence and informing locally appropriate implementation strategies. GENESIS-K is a government-funded, nationwide, multicenter prospective registry designed to enroll ≥10,000 patients with acute ischemic stroke. It includes comprehensive genomic profiling through genome-wide genotyping arrays supplemented with pharmacogene variants supported by CPIC guidelines or high-level PharmGKB evidence. The registry integrates clinical, radiologic, genomic, and long-term outcome data to enable the validation of established pharmacogenomic associations in Korean populations, the identification of population-specific variants, the development of Korean-tailored dosing algorithms, and the exploration of polygenic risk scores for stroke prognosis and medication response. Genomic results for key pharmacogenes (e.g., CYP2C19, SLCO1B1, VKORC1, CYP2C9) are reported to participating physicians, accompanied by structured education programs to support clinical implementation. As shown in Table 3, preliminary data from GENESIS-K participants have already provided population-specific allele frequency information that informs genotype-guided therapeutic strategies for Korean stroke patients, serving as a model for similar initiatives in other underrepresented populations worldwide.

Ultimately, equitable precision medicine depends on the deliberate inclusion of diverse populations in both discovery and translational research. Without such inclusion, the clinical benefits of PGx risk remaining inaccessible to substantial segments of the global population.

Beyond CYP2C19-guided clopidogrel therapy, several additional candidates represent the next frontier for pharmacogenomic integration into stroke care, albeit with differing levels of supporting evidence and clinical feasibility. Statin PGx—particularly involving SLCO1B1 and ABCG2 variants—appears to be among the most promising targets for implementation. This promise is supported by strong recommendations from both CPIC and DPWG guidelines, the widespread prescription of statins in stroke populations for both primary and secondary prevention, and the high frequency of ABCG2 variants in East Asians (approximately 50% in Koreans) (Table 3) [8,124]. However, the clinical utility of statin pharmacogenomic testing is currently constrained by its primary focus on adverse event risk rather than therapeutic efficacy. Because SAMS can often be effectively managed through dose adjustments or statin switching, payers have demonstrated limited interest in reimbursement, despite robust scientific evidence supporting genetic risk stratification [99]. Warfarin PGx represents one of the most mature fields in this domain, supported by wellvalidated dosing algorithms and clinical trials demonstrating improved anticoagulation control through genotype-guided therapy. Nonetheless, its practical relevance has diminished markedly with the growing preference for DOACs as first-line agents for stroke prevention in atrial fibrillation [7,54].

Cilostazol represents a particularly important candidate for pharmacogenomic research in Asian populations. The Lacunar Intervention Trial-2 (LACI-2) trial demonstrated cilostazol’s superiority over aspirin in preventing recurrent stroke among patients with lacunar infarction, thereby establishing its clinical efficacy in secondary prevention [125]. Given its widespread use in Korea, Japan, and China, cilostazol offers a significant opportunity for pharmacogenomic application [126]. Genetic polymorphisms in CYP2C19 and CYP3A5 have been shown to significantly influence plasma concentrations of cilostazol and its active metabolites [127,128]. Moreover, studies have reported associations between CYP genotypes and platelet responsiveness to cilostazol in patients with cerebral infarction [129]. However, robust clinical outcome data linking these genotypes to stroke recurrence or major adverse cardiovascular events remain limited. Considering cilostazol’s established benefit in secondary prevention of lacunar stroke and its prevalent use in Asian populations, targeted pharmacogenomic research in this domain represents a key future priority. Such efforts could follow a developmental trajectory analogous to that of the clopidogrel-CYP2C19 model, progressing from pharmacokinetic and pharmacodynamic validation to genotype-guided clinical application.

Pharmacogenomics constitutes a central pillar of the broader precision medicine paradigm, which aims to individualize therapy based on each patient’s genetic, clinical, and environmental characteristics [130,131]. As testing technologies advance and sequencing costs continue to decline, the integration of pharmacogenomics with complementary omics platforms—such as transcriptomics, proteomics, and metabolomics—holds considerable promise for generating more comprehensive and clinically actionable insights.

Panel-based genetic testing, enabling the simultaneous analysis of multiple pharmacogenes, is becoming increasingly feasible in clinical settings. Compared with single-gene assays, multigene panel approaches improve cost-effectiveness and provide information applicable to a wider range of medications. Commercially available and institution-specific panels are already established in cardiology and oncology, and their extension to stroke pharmacotherapy represents a logical next step [132,133].

In addition, advanced bioinformatics pipelines and artificial intelligence (AI)-driven methods, including machine learning and deep learning, are enhancing the interpretive power of pharmacogenomic data. These approaches facilitate the identification of complex gene-drug interactions, uncover novel associations, and support clinical decision-making through predictive modeling and real-time alert systems [134,135]. Notably, AI now enables the integration of pharmacogenomic data with high-dimensional clinical information, encompassing EHRs, neuroimaging, laboratory findings, and multi-omics profiles [136]. This multimodal integration, illustrated in Figure 1, offers a robust framework for developing individualized, context-specific prediction models that may elucidate previously unrecognized gene-environment and gene-phenotype interactions in stroke management.

Real-world implementation programs are essential to translate recent pharmacogenomic advances into routine clinical practice. Large-scale initiatives enable the evaluation of feasibility, optimization of workflows, and generation of evidence on clinical outcomes, safety, and cost-effectiveness across diverse healthcare systems [137]. Ultimately, the successful integration of pharmacogenomics into precision stroke medicine will depend on the establishment of robust data infrastructure, sustained interdisciplinary collaboration, comprehensive workforce education, and supportive health policy. Efforts to harmonize pharmacogenomic data with EHRs, standardize testing and reporting protocols, and engage key stakeholders—including patients—will be crucial for long-term sustainability.

Recent national initiatives exemplify this trajectory. The Estonian Biobank, for instance, integrates genomic data with EHRs, multi-omics datasets, and digital health information for more than 200,000 participants [138]. This population-based model demonstrates that pharmacogenomics can be implemented not only in disease-specific contexts such as cardiovascular or stroke medicine, but also as a component of broader preventive and public health strategies.