Comparison between clopidogrel and ticagrelor in CYP2C19 loss-of-function alleles coronary artery disease and stroke patients: a meta-analysis

Abstract

Background

It is suggested that in patients with coronary artery diseases (CAD) and stroke, the use of ticagrelor and aspirin may perform better than clopidogrel and aspirin regarding the risk of thrombosis/embolism, including recurrent myocardial infarction (MI) and cardiovascular death, especially in those carrying CYP2C19 loss-of-function (LOF) alleles. Therefore, we conducted the present systematic review and meta-analysis to investigate the effect of clopidogrel and ticagrelor in CAD and stroke patients with CYP2C19 LOF alleles (poor metabolizers of clopidogrel).

Methods

We performed the current systematic review and meta-analysis by searching for all eligible publications on PubMed, Web of Science, and Scopus from inception to November 2024. A search strategy employing three primary keywords in conjunction with their corresponding Medical Subject Headings (MeSH) terms: “Ticagrelor” AND “Clopidogrel” AND “CYP2C19” (PROSPERO ID CRD420251050533). We implemented the odds ratio (OR) as an effect estimate for the dichotomous variables. The analysis was done at 95% confidence intervals (CI), and the p-value was significant if it was less than or equal to 0.05.

Results

Using clopidogrel was associated with an increased risk of thrombosis/embolism compared with ticagrelor, showing OR = 1.78 (95%CI, 1.08, 2.95; p = 0.02). Also, clopidogrel led to an increased risk of stroke, whether when used in stroke or CAD patients with CYP2C19 LOF alleles, compared with ticagrelor, with an overall OR = 1.43 (95%CI, 1.23, 1.66; p < 0.00001) and a higher rate of MI with OR = 1.53 (95%CI, 1.22, 1.92; p = 0.0003). No significant difference was observed between the two groups (clopidogrel and ticagrelor) in stroke or CAD patients with OR = 0.98 (95%CI, 0.79, 1.22; p = 0.87). Also, no significant difference was observed between both groups regarding the risk of minor bleeding in stroke or CAD patients with OR = 0.66 (95%CI, 0.42, 1.05; p = 0.08) and any types of bleeding (major or minor bleeding) with overall OR = 0.81 (95%CI, 0.54, 1.21; p = 0.3) and I² = 88%, p < 0.00001.

Conclusion

The meta-analysis of the selected articles indicated a preference for ticagrelor over clopidogrel in patients with stroke or CAD possessing CYP2C19 LOF alleles. The reduced incidence of thrombosis/embolism and associated events, such as stroke and MI, was noted in individuals administered ticagrelor in comparison to those receiving clopidogrel. Bleeding remains a concern with ticagrelor; however, current studies indicate its safety since there are no significant changes in the risk of minor and major bleeding and ICH compared to clopidogrel.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00228-025-03860-4.

Introduction

Acute coronary syndrome (ACS) is among the most prevalent fatal conditions globally [1]. Percutaneous coronary intervention (PCI) has diminished mortality from ACS. However, numerous complications, such as ischemic events associated with drug-eluting stents (DESs) and bleeding issues, particularly at access sites, pose significant challenges [2]. Post-PCI antiplatelet therapy is a crucial treatment approach for ACS patients [1, 2]. Dual antiplatelet therapy utilizing aspirin and clopidogrel is the established protocol for preventing stent thrombosis and enhancing clinical outcomes during PCI [3].

In patients experiencing an acute mild ischemic stroke or transient ischemic attack (TIA), the likelihood of a subsequent stroke occurring within 3 months of the initial incident is between 5 and 10% [4–6]. Dual antiplatelet therapy utilizing clopidogrel and aspirin has demonstrated superior efficacy compared to aspirin monotherapy in mitigating subsequent events in patients with minor stroke or TIA, as evidenced by the CHANCE (clopidogrel in high-risk patients with acute nondisabling cerebrovascular events) [7] and POINT (platelet-oriented inhibition in new TIA and minor ischemic stroke) trials [8].

Clopidogrel is a prodrug that necessitates conversion into its active metabolite by hepatic cytochrome P450, primarily CYP2C19. Platelet responses to clopidogrel exhibit significant interindividual variability. High on-treatment platelet reactivity (HTPR) refers to patients exhibiting an insufficient response to clopidogrel therapy. In other words, despite receiving the standard dose of clopidogrel, a subset of patients (reported as 4–30% in various studies) does not achieve adequate platelet inhibition. This inadequate response results in persistently high platelet activity, which increases the risk of adverse cardiovascular events such as stent restenosis, stent thrombosis, and even clinical death. This results from CYP2C19 loss-of-function (LOF) alleles in these individuals [9, 10]. Numerous independent clinical studies have shown that individuals with HTPR face a heightened risk of stent thrombosis and other cardiac problems [11, 12]. Therefore, there is a therapeutic necessity to evaluate the effectiveness of antiplatelet medication and implement more potent ADP-receptor antagonists.

Ticagrelor, a reversible oral antagonist that directly inhibits the platelet P2Y12 receptor and does not necessitate metabolic activation for its antiplatelet activity, may produce comparable or superior levels of platelet aggregation inhibition compared to clopidogrel [13, 14]. Ticagrelor, in conjunction with aspirin, demonstrated superiority over aspirin monotherapy in decreasing the incidence of stroke or mortality in individuals with acute mild-to-moderate ischemic stroke or high-risk TIA [15]. In the PRINCE (platelet reactivity in acute stroke or transient ischemic attack) trial, patients with minor stroke or TIA receiving ticagrelor in conjunction with aspirin exhibited reduced platelet reactivity compared to those administered clopidogrel with aspirin, especially among carriers of the CYP2C19 LOF allele [16]. These results indicate that the combination of ticagrelor and aspirin may confer a reduced risk of subsequent stroke compared to the combination of clopidogrel and aspirin in patients with mild stroke or TIA who possess CYP2C19 LOF alleles. Also, this suggests that in patients with coronary artery disease (CAD), the use of ticagrelor and aspirin may perform better than clopidogrel and aspirin regarding the risk of thrombosis/embolism, including recurrent myocardial infarction (MI) and cardiovascular death, especially in those carrying the CYP2C19 LOF allele. Despite its clinical efficacy, ticagrelor therapy has some limitations, such as an increased risk of non-procedure-related bleeding [14], major bleeding events [17], dyspnea [18], and patient adherence issues either due to its twice-daily dosing regimen or these side effects, particularly in fragile patients [19].

Earlier evidence already hints that clopidogrel is sub-optimal in patients who carry CYP2C19 LOF alleles, yet important knowledge gaps remain. Three meta-analyses confined to CAD cohorts reported that replacing clopidogrel with an alternative P2Y12 inhibitor (prasugrel or ticagrelor) lowered the composite risk of major adverse cardiovascular events (MACE) by roughly 40–50%, driven mainly by fewer myocardial infarctions and cardiovascular deaths [20–22]. A large registry study reached a similar conclusion in routine practice [23]. However, those reviews pooled prasugrel and ticagrelor together, did not analyze stroke populations, and presented only relative risk

estimates, which are difficult to translate into bedside decisions. A very recent meta-analysis limited to genotyped PCI patients confirmed a 62% higher MACE risk when LOF carriers stayed on clopidogrel, but it still grouped the two alternative drugs and excluded stroke [24].

To close these gaps, we conducted the present systematic review and meta-analysis to investigate the efficacy and safety outcomes of clopidogrel and ticagrelor in CAD and stroke patients with CYP2C19 LOF alleles and poor metabolizers of clopidogrel. We isolated ticagrelor (the agent most often recommended for switching), included both CAD and stroke cohorts of CYP2C19 LOF carriers, and calculated absolute measures such as the number needed to treat. To our knowledge, no previous analysis has combined all three of these elements.

Methods

This systematic review and meta-analysis adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) standards [25], searching for all eligible publications on PubMed, Web of Science, and Scopus from inception to November 2024. According to each database, a search strategy employing three primary keywords in conjunction with their corresponding Medical Subject Heading (MeSH) terms—”Ticagrelor” AND “Clopidogrel” AND “CYP2C19″—was initiated. The protocol of the systematic review and meta-analysis was registered in PROSPERO with registration number CRD420251050533.

Eligibility criteria and screening

After the search, articles were imported into Rayyan to streamline the filtering process [26]. The initial screening was conducted independently by two pairs of authors (YA and YN and GB and IA), who reviewed titles, abstracts, and full-text articles to determine study eligibility. Any discrepancies were resolved by the senior author (MHEM). The study examined a group of patients with coronary artery disease or stroke who were poor metabolizers due to CYP2C19 LOF alleles, utilizing clopidogrel as the intervention and ticagrelor as the comparator while assessing various efficacy and safety outcomes: thrombosis/embolism, stroke, MI, minor or major bleeding, all-cause mortality, cardiac mortality, intracerebral hemorrhage (ICH), and major adverse cardiovascular events (MACE). The design specifications emphasized observational studies (cohort and case–control) and randomized controlled trials (RCTs) without imposing a restriction based on language (foreign language studies were translated by a professional translator) or sample size. We excluded case reports, reviews, and grey literature, such as unpublished theses, conference proceedings, government reports, white papers, and technical reports.

Data extraction

The baseline characteristics of the studies, including research design, gender, age, sample size, and population, were extracted by two pairs of authors (YA and YN and GB and IA) using Microsoft Excel sheets. Any discrepancies were resolved by the senior author (MaE). Outcome data were also collected, encompassing events of thrombosis/embolism, stroke, MI, minor or major bleeding, all-cause mortality, cardiac mortality, ICH, and MACE.

Quality assessment

Two authors performed a risk of bias evaluation (MoE and MHEM), and any discrepancies were referred to the senior author (MA) for resolution. We employed various assessment instruments in alignment with the study design. The Newcastle–Ottawa Scale (NOS), which allocates a star value from 0 to 9 for each study, was employed to evaluate the quality of observational studies. All questions may be awarded one or zero stars, except for comparison questions, which may get a maximum of two stars. A study rated 0–3 stars is categorized as low quality; 4–6 stars signify intermediate quality, while 7–9 stars represent excellent quality [27]. We employed the Cochrane risk-of-bias tool (Rob-2) for RCTs, comprising five areas, each accompanied by a series of inquiries. The results are subsequently incorporated into a visual representation to ascertain one of three levels of bias: low risk, moderate concern, or severe hazard. A study is deemed to possess a low overall risk of bias if all five domains demonstrate a low risk. The study has potential bias issues if any domain presents difficulties. If any domain shows a substantial risk of bias or numerous domains present problems, the study is classified as having a high risk of bias [28].

Statistical analysis

Review Manager software version 5.4 was used for the statistical analysis, including subgroup analysis [29]. We implemented the odds ratio (OR) as an effect estimate for the dichotomous variables. Heterogeneity was measured using I², and it was deemed significant if the p-value was less than or equal to 0.05. The fixed effect model was used for non-significant heterogeneity, and the random effect model was used for significant heterogeneity among the included outcomes. We calculated the Number Needed to Treat [NNT = 1/Absolute Risk Reduction (ARR)] for significant outcomes. We performed sensitivity analyses to determine the robustness of the effect size by removing one study at a time to check the strength of the evidence and ensure the overall results were not altered. The analysis was done at 95% confidence intervals (CI), and the p-value was significant if it was less than or equal to 0.05. Finally, if at least ten studies were reported in the outcome, the asymmetry analysis was performed to determine the publication bias by visual inspection of the funnel plot of the studies, and Egger’s test confirmed the results [30].

Results

Literature searching and screening

The conducted search strategy yielded 619 (224 from PubMed, 212 from Scopus, and 183 from Web of Science) entries, of which 352 were classified as duplicates. After evaluating the titles and abstracts of the remaining 267 papers, 245 were excluded (108 of different designs, including case reports, reviews, and grey literature, such as unpublished theses, conference proceedings, government reports, white papers, and technical reports, and 137 irrelevant articles including different comparisons and different populations), and 22 satisfied the criteria for an exhaustive evaluation of the full text. In conclusion, 17 publications were deemed suitable for inclusion in the final systematic review [16, 31–46] (Fig. 1).

Fig. 1.

PRISMA flow diagram of searching and screening processes

Baseline characteristics

In the present systematic review and meta-analysis, 11 cohort studies and six RCTs were included, which compared the use of clopidogrel against ticagrelor in patients with poor metabolism of clopidogrel due to CYP2C19 LOF alleles. These studies were conducted in CAD patients (13 studies), whether stable CAD or ACS, and patients with minor stroke or TIA. The mean (SD) age of patients ranged from 59.61 (9.58) to 72.44 (5.1) years old. Baseline characteristics of the included studies are fully illustrated in Table 1.

Table 1.

Baseline characteristics of the included studies

Study ID	Population	Design	Sample size, N		Age, mean (SD)		Male, n (%)
			Clopidogrel	Ticagrelor	Clopidogrel	Ticagrelor	Clopidogrel	Ticagrelor
Cavallari (2018)	CAD	Cohort	1276	539	62.7 (11.7)	62.5 (11.4)	150 (67.5)	237 (68.9)
Chen (2017)	CAD	Cohort	46	57	59.61 (9.6)	60.8 (9.2)	29 (63)	34 (59.6)
Dong (2016)	CAD	Cohort	102	64	67 (12)	67 (15)	80 (78.4)	52 (81.2)
Hu (2024)	CAD	Cohort	240	60	NR	NR	NR	NR
Lee (2018)	CAD	Cohort	1193†		63.3 (12.0)		NR	NR
Lee (2021)	CAD	Cohort	3342		64 (12)	60 (11)	308 (65)	683 (73)
Liu (2024)	Stroke	Cohort	284	88	65.6 (9.3)	67.2 (10.2)	241 (85)	71 (80)
Wallentin (2010)	CAD	RCT	5137	5148	62.5 (11.0)	62.5 (10.9)	NR	NR
Wang (2019)	Stroke	RCT	339	336	60.5 (9.0)	61.1 (8.5)	249 (73.5)	245 (72.9)
Wang (2021)	Stroke	RCT	3207	3205	64.2 (10.5)	64.3 (10.4)	2127 (66.3)	2115 (66)
Wang (2024)	CAD	Cohort	243	302	NR	NR	78 (63.9)	61 (80.3)
Xi (2020)	CAD	Cohort	977	359	60.6 (9.7)	58.3 (10.0)	729 (74.6)	281 (78.3)
Xiong (2015)	CAD	RCT	112	112	67 (9)	66 (8)	71 (87)	70 (86)
Yang (2020)	Stroke	RCT	190	197	60 (9)	61.3 (9)	137 (72.1)	140 (71)
Yu (2017)	CAD	Cohort	724	247	NR	NR	NR	NR
Zhang (2016)	CAD	RCT	90	91	71.7 (9.2)	68.8 (9.6)	49 (54.4)	42 (46.2)
Zhang (2024)	CAD	Cohort	2056	695	72.4 (5.1)	69.73 (3.42)	1182 (57.5)	442 (63.6)

CAD coronary artery disease, RCT randomized controlled trial, NR not reported, SD standard deviation

^†Among these, 233 patients were CYP2C19 intermediate or poor metabolizers (IM/PM); of these, 68 received clopidogrel, and 165 received ticagrelor or prasugrel

The most commonly used dose for clopidogrel was 75 mg once daily, except for one study, which was 150 mg once daily, while that of ticagrelor was 90 mg twice daily. The evaluation time of the outcomes or the follow-up period ranged from 1 to 12 months (most common) and 3 or 6 months in other studies. All the patients in the included studies had CYP2C19 with at least one LOF allele *2 or *3. Different treatments were used with the antiplatelets, including proton pump inhibitors, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, calcium channel blockers, and statins. The patients in all the studies had comorbid conditions that increased the severity of the cases, such as previous stroke, myocardial infarction, diabetes, atrial fibrillation, heart failure, and dyslipidemia. Reasons for antiplatelets were the prevention of thrombosis after PCI or recurrent stroke after stroke or TIA. Types of CAD included ACS, such as ST-elevated myocardial infarction (STEMI), non-ST-elevated myocardial infarction (NSTEMI), and angina, in addition to stable CAD (Table 2).

Table 2.

Summary of characteristics of the included studies

Study ID	Dose		Time of evaluation of outcomes		CYP2C19 alleles		Co-treatment		Severity of the disease		Reason for antiplatelet		Type of CAD
	Clopidogrel	Ticagrelor	Clopidogrel	Ticagrelor	Clopidogrel	Ticagrelor	Clopidogrel	Ticagrelor	Clopidogrel	Ticagrelor	Clopidogrel	Ticagrelor	Clopidogrel	Ticagrelor
Cavallari (2018)	75 mg once daily	90 mg twice daily	12 months		CYP2C19 with at least 1 of LOF alleles *2 or *3		Aspirin, PPI, Statins, ARB, ACEi, B-blocker, Anticoagulants		Comorbidities include atrial fibrillation, heart failure, diabetes, hypertension, dyslipidemia, stroke		After PCI		STEMI, NSTEMI, UA, stable CAD
Chen (2017)	75 mg once daily	90 mg twice daily	12 months		At least 1 CYP2C19*2		Aspirin, PPI, ARB, ACEi, B-blocker, CCB		Comorbidities include diabetes, hypertension, dyslipidemia		After PCI		Not specified
Dong (2016)	75 mg once daily	90 mg twice daily	1 month		CYP2C19 with at least 1 of LOF alleles *2 or *3		NR	NR	Comorbidities include diabetes, hypertension, dyslipidemia		After PCI		Acute coronary syndrome
Hu (2024)	75 mg once daily	90 mg twice daily	12 months		CYP2C19 with at least 1 of LOF alleles *2 or *3		NR	NR	Comorbidities include diabetes, hypertension, dyslipidemia		After PCI		Not specified
Lee (2018)	NR	NR	12 months		CYP2C19 with at least 1 of LOF alleles *2 or *3		Aspirin, ARB, ACEi, B-blocker, Anticoagulants		Comorbidities include atrial fibrillation, heart failure, diabetes, hypertension, dyslipidemia, stroke		After PCI		STEMI, NSTEMI, UA, stable CAD
Lee (2021)	75 mg once daily	90 mg twice daily	12 months		CYP2C19 with at least 1 of LOF alleles *2 or *3		Aspirin, PPI, Statins, ARB, ACEi, B-blocker, Anticoagulants		Comorbidities include atrial fibrillation, heart failure, diabetes, hypertension, dyslipidemia, stroke		After PCI		STEMI, NSTEMI, UA, stable CAD
Liu (2024)	75 mg once daily	90 mg twice daily	6 months		CYP2C19 with at least 1 of LOF alleles *2 or *3		NR	NR	Comorbidities include intracranial arterial stenting, diabetes, hypertension, dyslipidemia, hyperuricemia, kidney insufficiency, pulmonary infections		Ischemic stroke patients with cerebral artery stenting		NA
Wallentin (2010)	75 mg once daily	90 mg twice daily	12 months		CYP2C19 with at least 1 of LOF alleles *2 or *3		PPI, statins, aspirin		Comorbidities include diabetes, smoking		Acute coronary syndrome with PCI or not		Acute coronary syndrome
Wang (2019)	75 mg once daily	90 mg twice daily	12 months		CYP2C19 with at least 1 of LOF alleles *2 or *3		PPI, statins, aspirin		Comorbidities include atrial fibrillation, coronary artery disease, diabetes, hypertension, dyslipidemia, stroke		For platelet reactivity in minor stroke or transient ischemic attack		NA
Wang (2021)	75 mg once daily	90 mg twice daily	12 months		CYP2C19 with at least 1 of LOF alleles *2 or *3		Aspirin		Comorbidities include myocardial infarction, diabetes, hypertension, dyslipidemia, stroke		Secondary prevention of stroke after stroke or transient ischemic attack		NA
Wang (2024)	75 mg once daily	90 mg twice daily	12 months		CYP2C19 with at least 1 of LOF alleles *2 or *3		PPI, statins, aspirin		Comorbidities include acute coronary syndrome, diabetes, hypertension, dyslipidemia, stroke		After off-pump coronary artery bypass grafting		Not specified
Xi (2020)	75 mg once daily	90 mg twice daily	12 months		Two CYP2C19 LOF Alleles		Aspirin		Comorbidities include atrial fibrillation, diabetes, hypertension, dyslipidemia, ischemic stroke, cerebral hemorrhage, myocardial infarction		After PCI		STEMI, NSTEMI, UA, stable CAD
Xiong (2015)	150 mg daily	90 mg twice daily	1 month		CYP2C19*2 homozygotes†		Aspirin		Comorbidities include diabetes, hypertension, dyslipidemia		Acute coronary syndrome		Acute coronary syndrome
Yang (2020)	75 mg once daily	90 mg twice daily	3 months		CYP2C19 with at least 1 of LOF alleles *2 or *3		PPI, statins, aspirin		Comorbidities include diabetes, hypertension, dyslipidemia, ischemic stroke, coronary artery disease		For platelet reactivity in minor stroke or transient ischemic attack		NA
Yu (2017)	75 mg once daily	90 mg twice daily	12 months		CYP2C19 with at least 1 of LOF alleles *2 or *3		Aspirin, PPI, Statins, ARB, ACEi, B-blocker, CCB		Comorbidities include diabetes, hypertension, cerebral infarction		After PCI		Not specified
Zhang (2016)	75 mg once daily	90 mg twice daily	6 months		CYP2C19 with at least 1 of LOF alleles *2 or *3		PPI, statins, aspirin		Comorbidities include diabetes, hypertension, dyslipidemia		After PCI		Acute coronary syndrome
Zhang (2024)	NR	NR	12 months		CYP2C19 with at least 1 of LOF alleles *2 or *3		Aspirin, PPI, Statins, ARB, ACEi, B-blocker, CCB, Diuretics		Comorbidities include diabetes, hypertension, dyslipidemia, stroke, myocardial infarction		After PCI		Acute coronary syndrome

LOF loss of function, ACEi angiotensin converting enzyme inhibitor, PPI proton pump inhibitor, ARB angiotensin receptor blocker, CCB calcium channel blocker, PCI percutaneous coronary intervention, STEMI ST-elevated myocardial infarction, NSTEMI non-ST-elevated myocardial infarction, UA unstable angina, CAD coronary artery disease, NA not applicable, NR not reported

^†No deviation from the expected proportions of genotypes in the selected population predicted by the Hardy–Weinberg equilibrium for polymorphisms

Quality and risk of bias assessment

According to NOS, the 11 cohort studies were rated high quality, as shown in Table 3. Regarding RCTs, the Rob-2 used for risk of bias assessment showed that all the included RCTs were deemed to have a low risk of bias across all the domains (Supplementary Fig. 1).

Table 3.

Quality assessment of cohort studies using the Newcastle–Ottawa scale tool

Study name	The level of representation of the affected cohort (★)	Identification of the unexposed cohort (★)	Determination of exposure (★)	Evidence that the outcome of interest was absent at the commencement of the research (★)	Comparison of cohorts based on design or assessment (max★★)	Was the follow-up duration sufficient for consequences to manifest? (★)	Evaluation of results (★)	Assessment of cohort follow-up sufficiency (★)	Quality level†
Cavallari (2018)	★	★	★	★	★★	★	★	★	High
Chen (2017)	★	★	★	★	★★	★	★	★	High
Dong (2016)	★	-	★	★	★★	★	★	★	High
Hu (2024)	★	★	★	★	★★	-	★	★	High
Lee (2018)	★	-	★	★	★★	★	★	★	High
Lee (2021)	★	★	★	★	★★	★	★	★	High
Liu (2024)	★	-	★	★	★★	★	★	★	High
Wang (2024)	★	-	★	★	★★	-	★	★	High
Xi (2020)	★	★	★	★	★★	★	★	★	High
Yu (2017)	★	-	★	★	★★	★	★	★	High
Zhang (2024)	★	-	★	★	★★	★	★	★	High

^†A study rated 0–3 stars is categorized as low quality; 4–6 stars signify intermediate quality, while 7–9 stars represent excellent quality

Meta-analysis

Using clopidogrel was associated with an increased risk of thrombosis/embolism compared with ticagrelor, showing OR = 1.78 (95%CI, 1.08, 2.95; p = 0.02) and I² = 0% [AAR = 1%, NNT = 100, with uncertainty in 95% CI]. Also, clopidogrel led to an increased risk of stroke, whether when used in stroke or CAD patients with CYP2C19 LOF alleles, compared with ticagrelor, with overall OR = 1.43 (95%CI, 1.23, 1.66; p < 0.00001) and I² = 28%, p = 0.15 [AAR = 2%, NNT = 50]. Moreover, clopidogrel was associated with a higher rate of MI with OR = 1.53 (95%CI, 1.22, 1.92; p = 0.0003) and I² = 0% [AAR = 2%, NNT = 50] (Fig. 2, 3, and 4). Therefore, ticagrelor is safer than clopidogrel in patients with CYP2C19 LOF alleles in terms of the risk of thrombosis-related events. Regarding the risk of major bleeding, no significant difference was observed between the two groups (clopidogrel and ticagrelor) in stroke or CAD patients with OR = 0.98 (95%CI, 0.79, 1.22; p = 0.87) and I² = 0% (Fig. 5).

Fig. 2.

Comparison between clopidogrel and ticagrelor in the risk of thrombosis/embolism among stroke or coronary artery disease patients with CYP2C19 loss-of-function allele using odds ratio

Fig. 3.

Comparison between clopidogrel and ticagrelor in the risk of stroke among stroke and coronary artery disease patients with CYP2C19 loss-of-function allele using odds ratio

Fig. 4.

Comparison between clopidogrel and ticagrelor in the risk of myocardial infarction among stroke and coronary artery disease patients with CYP2C19 loss-of-function allele using odds ratio

Fig. 5.

Comparison between clopidogrel and ticagrelor in the risk of major bleeding among stroke and coronary artery disease patients with CYP2C19 loss-of-function allele using odds ratio

Also, no significant difference was observed between both groups regarding the risk of minor bleeding in stroke or CAD patients with OR = 0.66 (95%CI, 0.42, 1.05; p = 0.08) and I² = 79%, p < 0.0001. Moreover, no significant differences were observed regarding any types of bleeding (major or minor bleeding) in stroke or CAD patients with overall OR = 0.81 (95%CI, 0.54; 1.21, p = 0.3) and I² = 88%, p < 0.00001. Furthermore, no significant difference was observed between both groups regarding ICH in stroke patients with OR = 1.24 (95%CI, 0.56, 2.76; p = 0.59) and I² = 0% (Supplementary Figs. 2–4).

Regarding the risk of mortality, no significant difference was observed between both groups regarding all-cause mortality and cardiac mortality, showing OR = 1.09 (95%CI, 0.76, 1.54; p = 0.65) and I² = 0%, and OR = 1.1 (95%CI, 0.73, 1.64; p = 0.65) and I² = 5%, p = 0.39 (Fig. 6).

Fig. 6.

Comparison between clopidogrel and ticagrelor in the risk of all-cause mortality (top) and cardiac mortality (bottom) among stroke and coronary artery disease patients with CYP2C19 loss-of-function allele using odds ratio

Clopidogrel showed an increased risk of occurrence of MACE compared with ticagrelor with OR = 1.98 (95%CI, 1.29, 3.04; p = 0.002) and I² = 70%, p = 0.0009 [AAR = 7%, NNT = 14] (Supplementary Fig. 5).

Publication bias

Egger’s regression intercepts did not suggest significant small-study effects for stroke with intercept = 0.357 (95%CI, − 0.33, 1.07; p = 0.296), MI with intercept = 0.668 (95%CI, − 0.25, 1.58; p = 0.153), or major bleeding with intercept = 0.340 (95%CI, − 0.504, 1.89; p = 0.430). However, there was evidence of publication bias for any bleeding with intercept = 2.31 (95%CI, 1.32, 3.30; p < 0.001) (Supplementary Fig. 6).

Discussion

Summary of findings

Among CYP2C19 LOF allele carriers with stroke or coronary artery disease, ticagrelor demonstrated superior efficacy to clopidogrel by significantly lowering thrombotic risks, namely thrombosis/embolism, stroke, and myocardial infarction. Clopidogrel was associated with a 7% absolute increase in MACE risk with OR = 1.98 (95%CI, 1.29, 3.04; p = 0.002), corresponding to an NNT of 14 to prevent one additional event. Safety profiles were comparable: there were no significant differences in minor bleeding, major bleeding, or intracranial hemorrhage between the two agents. All-cause mortality also did not differ. These findings support using ticagrelor over clopidogrel in CYP2C19 LOF carriers, underscoring the value of genotype-guided antiplatelet therapy.

Interpretation of findings

Antiplatelet medication is consistently administered to individuals following a mild stroke or TIA to avert stroke recurrence and other thrombotic incidents. The prognosis of small stroke or TIA is highly diverse and is affected by the presence and intensity of prognostic variables. Extensive epidemiological and clinical investigations have revealed risk variables linked to adverse clinical outcomes in patients with stroke or TIA, including advanced age, tobacco use, hypertension, diabetes mellitus, cardiovascular disease, and peripheral artery disease [47–51]. Moreover, these risk factors may affect the choice of antiplatelet medication. The POPular AGE study [52] demonstrated that in individuals aged 70 years or older with NSTEMI, clopidogrel was a preferable alternative to ticagrelor. A meta-analysis of randomized studies [53] indicated that among smokers, individuals treated with clopidogrel experienced superior therapeutic benefits in mitigating cardiovascular events compared to those administered ticagrelor. The PLATO trial [14] showed that in patients with ACS, with or without ST-segment elevation, ticagrelor significantly decreased the occurrence of cardiovascular death compared to clopidogrel. The Clopidogrel vs. Aspirin in individuals at Risk of Ischemic Events (CAPRIE) research [54] indicated that the absolute advantage of clopidogrel over aspirin for subsequent combined vascular events was enhanced in individuals with a history of diabetes or ischemic events.

Ticagrelor may serve as a clinically beneficial alternative antiplatelet agent for stroke patients with CYP2C19 LOF alleles, in whom clopidogrel’s efficacy may be diminished [55], particularly within East Asian populations where both stroke recurrence and the prevalence of CYP2C19 LOF alleles are elevated [56]. Nonetheless, the clinical utility of pharmacogenomics-informed antiplatelet medication selection is constrained by the accessibility of quick CYP2C19 genotyping methodologies and toolkits, and the cost-effectiveness of a genotype-guided approach requires more examination. Nonetheless, disagreement exists concerning the safety of ticagrelor and its effects on long-term prognosis. The application of ticagrelor in clinical practice is often limited by several variables, including adverse medication reactions (such as dyspnea, ventricular pauses of > 3 s, and increased blood creatinine and uric acid levels), advanced age, impaired liver and kidney function, and patient non-compliance [57]. Most patients preferred a combination of clopidogrel and aspirin for antiplatelet treatment.

It was indicated [58] that CYP2C19 genetic testing could assist patients undergoing PCI in selecting the most effective antiplatelet therapy, perhaps decreasing the incidence of MACE. Additionally, post hoc analysis of the CHANCE study [59] indicated that the efficacy of clopidogrel in Chinese patients with mild stroke or TIA was contingent upon both CYP2C19 genotype and risk profile. CYP2C19 LOF carriers do not derive advantages from dual antiplatelet therapy with clopidogrel and aspirin; nevertheless, high-risk LOF carriers experience substantial benefits.

Current guidelines advocate for the administration of a P2Y12 receptor inhibitor in conjunction with aspirin as the foundational treatment for ACS patients post-PCI, significantly reducing ischemic episodes in individuals with DESs [60]. The PLATO research [14] indicated that ticagrelor therapy significantly decreased thrombotic events, such as vascular mortality, MI, and stroke, in patients with ACS without elevating severe bleeding rates compared to clopidogrel. Consequently, current guidelines promote ticagrelor over clopidogrel. The validity of ticagrelor therapy benefits for various populations of ACS patients post-PCI remains contentious. An observational trial conducted in Sweden indicated that ticagrelor therapy was not superior to clopidogrel in unselected ACS patients following PCI, exhibiting heightened bleeding risks and similar ischemic events [61]. The variance observed in the PLATO trial was attributed to the distinct characteristics of the recruited patients, which included a higher percentage of PCI, an increased number of elderly individuals, and perhaps diminished compliance with ticagrelor in real-world scenarios [61]. East Asian populations have relative resistance to ischemic events but increased susceptibility to bleeding hazards compared to Caucasian populations [62]. Numerous studies have concentrated on the safety and efficacy of ticagrelor therapy in East Asian patients. The TICAKOREA study, focusing on East Asian/Korean ACS patients, showed that standard-dose ticagrelor was associated with heightened clinically significant bleeding without advantages in decreasing ischemic episodes compared to clopidogrel [63]. The TALOS-AMI RCT [64] demonstrated that unguided de-escalation of dual antiplatelet therapy from ticagrelor to clopidogrel diminished bleeding risks while maintaining equivalent ischemic events in stable MI patients in Korea. An additional observational study [65] examining Chinese ACS patients following PCI indicated that ticagrelor elevated BARC Type 2 bleeding compared to clopidogrel. The POPular Genetics trial [66] revealed that a CYP2C19 genotype-guided approach for P2Y12 receptor inhibitors (CYP2C19 LOF carriers received ticagrelor or prasugrel, while noncarriers received clopidogrel) resulted in diminished bleeding risks without an elevation in thrombotic events compared to the ticagrelor or prasugrel treatment group. The TAILOR-PCI trial [67] showed that ticagrelor treatment did not significantly alter the rates of cardiovascular death, MI, stroke, stent thrombosis, or severe recurrent ischemia in CYP2C19 LOF carriers with ACS and stable CAD compared to standard clopidogrel therapy. However, its recent prespecified investigation indicated that ticagrelor medication markedly diminished cumulative ischemic events without elevating bleeding at 12 months post-PCI compared to clopidogrel treatment in CYP2C19 LOF carriers [68]. Parcha et al. [69] conducted a Bayesian reanalysis of the TAILOR-PCI trial, revealing a 99.9% posterior probability that the risk ratio for MACE is less than 1 in patients receiving genotype-guided P2Y12 inhibitor therapy post-PCI, utilizing informative priors from four prior clinical trials. A systematic review and meta-analysis by Galli et al. [58] demonstrated that the guided selection of antiplatelet therapy through platelet function or genetic testing in patients undergoing PCI enhanced efficacy outcomes, including MACE and specific endpoints, while significantly reducing minor bleeding. Consequently, customized antiplatelet medication that incorporates genetic and platelet function testing, alongside clinical features and procedural variables of patients, may provide a potential strategy to enhance safety and efficacy outcomes in individuals following PCI.

In addition to the CYP2C19 genotype, additional genetic variants, such as ABCB1, which encodes the P-glycoprotein efflux transporter and influences clopidogrel absorption, may affect the response to clopidogrel medication [70], although contradictory findings have been reported [71]. Consequently, integrating data from diverse genetic variants may enhance the customization of antiplatelet therapies for specific individuals. The potential impact of variant alleles of ABCB1 on ticagrelor responsiveness is still debated and requires additional investigation [32, 72]. Additional factors, including clinical comorbidities, medication adherence, and coronary anatomical and procedural characteristics, are also linked to ischemia and bleeding events in ACS patients receiving PCI. The aging process correlates with diminished cytochrome P450 levels and renal function, alongside increased comorbidities, resulting in more intricate pharmacokinetics and pharmacodynamics than those of younger individuals [73, 74]. Zhang et al. [43] suggested that clopidogrel may be more suitable for elderly ACS patients with CYP2C19 LOF mutations undergoing PCI than ticagrelor due to fewer clinically significant bleeding events and similar ischemic risks. The customized selection of P2Y12 receptor inhibitors requires the incorporation of CYP2C19 genotypes, platelet function assessments, and clinical risk factors to enhance results for older patients with elevated bleeding and ischemic risks post-PCI. Integrating additional genetic polymorphism assessments, such as ABCB1, may improve the efficacy and safety of antiplatelet medication in the future; however, further clinical data is required.

Clinical implications

U.S. and European regulators have formally recognized the impact of CYP2C19 LOF alleles on clopidogrel efficacy. The FDA’s 2022 label for Plavix carries a Boxed Warning stating that “genetic variations in CYP2C19 can impair conversion of clopidogrel to its active metabolite,” and that “tests are available to identify individuals who are CYP2C19 poor metabolizers; consider the use of an alternative P2Y₁₂ inhibitor in these patients” [75]. Likewise, the EMA’s summaries of product characteristics for clopidogrel note under “Pharmacogenetics” that “in poor CYP2C19 metabolizers, clopidogrel at recommended doses forms less active metabolite and has reduced antiplatelet effect.”

Recent guidance from the American Heart Association (AHA), as outlined by Pereira et al. [76], reinforces the clinical significance of CYP2C19 genetic testing in patients receiving oral P2Y12 inhibitors, especially clopidogrel. The statement emphasizes that patients with CYP2C19 LOF alleles treated with clopidogrel are at increased risk for adverse ischemic outcomes and recommends genotype-guided therapy, preferably using ticagrelor or prasugrel in LOF carriers. Pereira et al. [76] also highlighted implementation barriers to routine genetic testing, such as result turnaround time, clinician education, cost-effectiveness, and integration with electronic health records. Addressing these issues is crucial for translating evidence into widespread practice. They advocate for either rapid or preemptive testing to facilitate timely decision-making, which could significantly improve outcomes in patients undergoing PCI or with ACS.

Routine CYP2C19 genotyping is now practicable in many clinical settings. On-site point-of-care assays such as the Spartan RX device yield a median of 2 h (range, 0.5–13.5 h) post-sample, enabling same-day therapeutic decisions [77]. Cost analyses in a community-pharmacy setting estimate net genotyping costs of approximately €43 per patient, which may be offset by reduced adverse events and medication optimization [78]. The RAPID-GENE trial further demonstrated that bedside genotyping of CYP2C19*2 carriers can be completed within 1 h, guiding prasugrel therapy to lower on-treatment platelet reactivity effectively [78]. In light of these data, and consistent with CPIC and AHA/ESC consensus statements advocating genotype-guided P2Y₁₂ inhibitor selection [76], we recommend the following:

• For CYP2C19 intermediate or poor metabolizers, prescribe prasugrel or ticagrelor to mitigate thrombotic risk.

• For extensive or rapid metabolizers, clopidogrel remains a safe, cost-effective option.

It is also important to highlight that adherence was not systematically assessed in the included studies, so we cannot directly account for it in our meta-analysis. In controlled trials such as PLATO [14], treatment discontinuation rates at 12 months were similar for ticagrelor (≈21%) and clopidogrel (≈24%). However, real-world data indicate that twice-daily regimens generally achieve 10–20% lower adherence than once-daily dosing [79–81]. Therefore, in routine practice, poorer adherence to ticagrelor’s twice-daily schedule could attenuate its effectiveness relative to clopidogrel and may have influenced the observed outcomes [19].

Limitations

Among the limitations of this meta-analysis is the predominance of observational studies, which inherently introduce bias. Key factors remain inadequately addressed in the literature, including variability in CYP2C19 allele definitions, clopidogrel dosing strategies, and the cost-effectiveness of genotyping. Most outcomes were reported at approximately 12 months, with only a handful at 1, 3, or 6 months, precluding subgroup analyses by follow-up interval or other study characteristics (Table 2). We could only conduct indication-specific analyses for stroke versus CAD, where outcome definitions overlapped.

Additionally, although stratification by heterozygous versus homozygous CYP2C19 LOF carriers would be ideal, only two of the 17 included studies (Xiong 2015 [*2 homozygotes only] and Xi 2020 [two LOF alleles]) provided genotype-specific outcomes. No trial reported outcomes for heterozygotes alone. Consequently, we were unable to conduct subgroup meta-analyses by genotype.

Finally, despite sensitivity analyses, heterogeneity remained high for MACE (I² = 70%), minor bleeding (I² = 79%), and any bleeding (I² = 88%). This reflects irreducible study-level differences: follow-up ranged from 1 to 3 months (e.g., Dong 2016; Xiong 2015; Yang 2020) to 12 months in most trials; antiplatelet dosing was largely uniform (clopidogrel 75 mg vs. ticagrelor 90 mg BID), aside from Xiong (2015) (clopidogrel 150 mg) and two studies lacking dose data; CYP2C19 definitions varied from “ ≥ 1 LOF allele” to “two LOF alleles” or *2 homozygotes; concomitant therapies (aspirin, PPIs, statins) were inconsistently reported; and patient cohorts mixed post-PCI/ACS and cerebrovascular populations. Moreover, Egger’s test detected small-study effects for any bleeding (intercept = 2.31; 95% CI, 1.32–3.30; p < 0.001), suggesting publication bias. Consequently, results should be interpreted cautiously, especially those with high heterogeneity or bias risk.

Conclusion

Among CYP2C19 LOF allele carriers with stroke or coronary artery disease, ticagrelor consistently outperformed clopidogrel—reducing thrombosis/embolism, stroke, and MI without increasing minor or major bleeding, intracranial hemorrhage, or mortality—and clopidogrel users faced a 7% higher MACE rate (NNT = 14). To build on these findings, future studies should employ prospective, genotype-guided randomized designs with harmonized allele panels and dosing regimens; systematically monitor adherence; extend follow-up; distinguish ischemic versus hemorrhagic stroke outcomes; conduct cost-effectiveness analyses of point-of-care genotyping; and include diverse, real-world populations to enhance generalizability.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

MaE, MHEM, and MoE: conceptualization and methodology. YA, YN, GB, and IA: investigation and data curation. MHEM: formal analysis. MaE, MHEM, and MoE: Writing—Original Draft. RMW: Supervision. MHEM: Project administration. MA: Writing—Review & Editing. All authors read and approved the final content.

Data Availability

Data is provided within the manuscript or supplementary information files.

Declarations

Ethics approval and consent to participate

Not applicable.

Competing interests

The authors declare no competing interests.