2024 Volume 31 Issue 6 Pages 904-916
Aim: This study compared the effectiveness, safety, and mortality risks between cilostazol plus aspirin and clopidogrel plus aspirin treatment for patients with acute minor ischemic stroke or transient ischemic attack (TIA).
Methods: This retrospective cohort study employed a new-user design and utilized data from the nationwide Health and Welfare Database in Taiwan. Patients were included if they were discharged with newly initiated cilostazol plus aspirin or clopidogrel plus aspirin after primary acute minor ischemic stroke or TIA hospitalization between 2009 and 2018. Inverse probability of treatment weighting was applied to balance covariats between study groups. Effectiveness outcomes were the risks of acute ischemic stroke, acute myocardial infarction (AMI), TIA, and composite cardiovascular events; Safety outcomes were the risks of intracranial hemorrhage (ICH), gastrointestinal bleeding, and composite bleeding events; Mortality outcomes were the risks of fatal stroke, cardiovascular mortality, and all-cause mortality.
Results: A total of 3,403 patients were included, of which 578 were treated with cilostazol plus aspirin and 2,825 were treated with clopidogrel plus aspirin. Cilostazol plus aspirin was associated with a higher risk of ICH (HR: 1.82; 95% CI: 1.16-2.84) compared to clopidogrel plus aspirin. No significant differences in the risks of effectiveness or mortality outcomes between the two groups were found.
Conclusions: The effectiveness and mortality of the two groups were similar for patients with acute minor ischemic stroke or TIA. However, cilostazol plus aspirin was associated with a higher risk of ICH compared to clopidogrel plus aspirin. Patients treated with cilostazol plus aspirin among this population should be monitored carefully to ensure their safety.
The risk of recurrent ischemic event following an acute minor ischemic stroke or transient ischemic attack (TIA) is significant1-3). Acute minor ischemic stroke is defined as an ischemic stroke with a National Institutes of Health Stroke Scale (NIHSS) score of ≤ 4. Approximately 10% to 20% of patients with acute minor ischemic stroke or TIA may experience a recurrent stroke or TIA within 90 days2).
Dual antiplatelet therapy is recommended for patients with acute minor ischemic stroke or TIA4). Compared with monotherapy aspirin, short-term dual antiplatelet therapy of clopidogrel plus aspirin can significantly reduce the risk of ischemic stroke among patients with acute minor ischemic stroke or TIA5, 6). However, previous studies have raised concerns about decreased effectiveness of clopidogrel use for patients with specific genetic polymorphisms7-9). Clopidogrel is a prodrug that must be activated by the CYP2C19 enzyme before inhibiting platelets. In clopidogrel users with CYP2C19 enzyme poor metabolizers, the effect of clopidogrel would be unpredictable, and the benefit of dual antiplatelets may be diminished7, 10). Unlike clopidogrel, cilostazol is a phosphodiesterase III inhibitor that inhibits platelet aggregation directly by increasing cyclic adenosine monophosphate (cAMP) concentrations11). The Cilostazol for Prevention of Secondary Stroke (CSPS 2) study12) reported that cilostazol was comparable to aspirin in preventing secondary strokes. Moreover, the Cilostazol Stroke Prevention Study for Antiplatelet Combination (CSPS.com) study13) found that adding cilostazol to aspirin or clopidogrel in high-risk ischemic stroke patients reduces the incidence of recurrent ischemic stroke compared to using aspirin or clopidogrel alone. Thus, cilostazol may be a reasonable alternative to clopidogrel in the context of dual antiplatelet therapy, particularly for patients with CYP2C19 enzyme poor metabolizers14).
Evidence regarding dual antiplatelet therapy using cilostazol plus aspirin for acute minor ischemic stroke or TIA is scarce. In the Efficacy and Safety of Combination Antiplatelet Therapies in Patients with Symptomatic Intracranial Atherosclerotic Stenosis (TOSS II) study15), they demonstrated that there were no differences in the progression or regression of stenosis, new ischemic lesions, nonfatal stroke, or vascular mortality between patients with symptomatic intracranial arterial stenosis treated with cilostazol plus aspirin and those treated with clopidogrel plus aspirin. However, the TOSS II study did not focus on patients with acute minor ischemic stroke or TIA. The use of cilostazol plus aspirin for preventing recurrent ischemic events among patients in this population is not fully understood. In addition, patients recruited to clinical trials are typically distinct from those in real-world clinical settings because they are carefully selected through strict protocols15). Therefore, it is important to investigate the use of cilostazol plus aspirin following acute minor ischemic stroke or TIA by using the real-world data.
According to our review of the literature, studies have yet to compare the effectiveness and safety of cilostazol plus aspirin with that of clopidogrel plus aspirin for patients with acute minor ischemic stroke or TIA. Previous studies have shown that patients with acute minor ischemic stroke or TIA face a substantial risk of recurrent ischemic event1-3). Therefore, determining the optimal dual antiplatelet therapy for the prevention of recurrent ischemic events among patients with acute minor ischemic stroke or TIA is crucial.
This study compared the effectiveness, safety, and mortality risks between cilostazol plus aspirin and clopidogrel plus aspirin treatment for patients with acute minor ischemic stroke or TIA.
This study was conducted from the Full Population Data of the Health and Welfare Database between 2008 and 2018 16). The Health and Welfare Database contains administrative claims data including beneficiaries’ social demographics, inpatient and outpatient diagnoses, procedures, drug prescriptions, and medical expenditures. This database includes information on more than 23 million enrollees from the National Health Insurance (NHI) program in Taiwan. In addition, we used data obtained from the Cause of Death Database to assess the risk of mortality.
Study Design and PatientsThis retrospective cohort study employed a new-user design. Patients were enrolled to our study if they were discharged with newly prescribed cilostazol plus aspirin or clopidogrel plus aspirin after hospitalization for primary acute minor ischemic stroke or TIA between January 1, 2009, and December 31, 2018. Using a new-user design helps mitigate potential selection bias, such as healthy-user bias, and strengthens causal inference in this retrospective observational study17). We identified acute minor ischemic stroke diagnoses on the basis of International Classification of Diseases, 9th Revision Clinical Modification (ICD-9-CM) codes 433 and 434 and International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) code I63. Patients with acute minor ischemic stroke should have an NIHSS score of ≤ 4. TIA diagnoses were based on ICD-9-CM code 435 and ICD-10-CM codes H34.0 and G45, excluding G45.4. Because we could not access the actual NIHSS scores from the administrative claims data, we used a validated proxy for NIHSS score to determine stroke severity18, 19). The proxy NIHSS score was developed using data mining techniques to explore the relationship between the in-hospital management received after acute ischemic stroke and the actual baseline NIHSS score18). Highly correlated in-hospital management factors were identified and incorporated into the calculation of the proxy NIHSS score, which closely approximates the actual baseline NIHSS score at the time of stroke onset. To determine the proxy NIHSS score, the first step is to calculate the stroke severity index (SSI) for each patient. The SSI is computed using seven in-hospital management factors that have been validated and found to be associated with the actual baseline NIHSS score18). These factors are intensive care unit stays, general ward stays, airway suctioning, bacterial sensitivity test results, nasogastric intubation, osmotherapy, and urinary catheterization. Each factor has a coefficient that represents its association with stroke severity. For instance, the coefficient for intensive care unit stay is 4.1770, while that for general ward stay, it is -5.5761. Other coefficients include airway suctioning (3.5083), bacterial sensitivity test (1.3642), nasogastric intubation (4.5809), osmotherapy (2.1448), and urinary catheterization (1.6569) 18). The SSI score is obtained by summing up the coefficients for each factor and adding a constant value of 9.6804 to the total sum of coefficients. Once the SSI was calculated, it is then converted to a proxy NIHSS score using the formula: proxy NIHSS=1.1722×SSI−0.7533 18). The resulting proxy NIHSS score provides a reliable estimate of the patient’s baseline NIHSS score and can be used to assess stroke severity.
We defined the date of the primary acute minor ischemic stroke or TIA hospitalization as that of the index admission, and the date of discharge from the index admission was defined as the index date. Each patient was assigned a pre-index period of 12 months prior to the index date.
To ensure the continuity of dual antiplatelet therapy after acute minor ischemic stroke or TIA, we implemented a 30-day gap following the index date. We selected this 30-day gap because it aligns with the Taiwan Stroke Society’s recommendation of using dual antiplatelet therapy for a duration of 21 to 30 days following acute minor stroke or TIA, depending on the choice of antiplatelet20). Patients who switched between cilostazol and clopidogrel or discontinued dual antiplatelet therapy within this 30-day gap after the index date were excluded. In other words, our study only included patients who were continuously prescribed dual antiplatelet therapy (i.e., cilostazol plus aspirin or clopidogrel plus aspirin) for at least 30 days after the index date. After this initial period, they were not required to persist with dual antiplatelet therapy during the follow-up period; they could either continue dual antiplatelet therapy or shift to monotherapy antiplatelet, depending on clinical decisions or physician/patients’ preferences. All patients were continued to follow from the index date until the first occurrence of the outcome of interests, alteration or addition of antiplatelet agents (including clopidogrel, cilostazol, prasugrel, ticagrelor, ticlopidine, aspirin plus extended-release dipyridamole, and dipyridamole), cessation of antiplatelet use, or until the end of the study period (i.e., December 31, 2018). Antiplatelet discontinuation was defined as the end of a prescription for any study drug (i.e., aspirin, cilostazol, or clopidogrel) or a 30-day gap between the end of days’ supply and the next study drug prescriptions.
Patients were excluded if they had a history of acute ischemic stroke or TIA hospitalization, or were ever prescribed with aspirin, cilostazol, or P2Y12 inhibitors (i.e., clopidogrel, ticagrelor, ticlopidine, or prasugrel) within 12 months prior to the index admission, or received a diagnosis of intermittent claudication during the pre-index period. Patients were also excluded if they had an acute myocardial infarction (AMI), received stenting procedures, switched from one to another antiplatelet, were prescribed with other antiplatelets, were aged under 20, or died during the index admission. We excluded patients with intermittent claudication, AMI, or receiving stenting procedures to ensure that all patients were treated with dual antiplatelet only for the prevention of ischemic events following acute minor ischemic stroke or TIA without other possible purposes21-24). Fig.1 shows a flowchart of patient selection.
Fig. 1 shows a flowchart of patient selection. After the study enrollment processes, a total of 3,403 patients were included in this study, of which 578 were cilostazol plus aspirin users and 2,825 were clopidogrel plus aspirin users
Patients in our study were assigned to one of two study groups, cilostazol plus aspirin and clopidogrel plus aspirin. Patients were assigned to either group according to their first prescription (cilostazol or clopidogrel) after the index admission. The cilostazol plus aspirin users served as the exposure group, and the clopidogrel plus aspirin users served as the comparison group.
Outcome MeasuresEffectiveness, safety, and mortality outcomes were evaluated separately during the follow-up period. The effectiveness outcomes were the occurrence of acute ischemic stroke, TIA, AMI, and composite cardiovascular events (i.e., a composite measure of acute ischemic stroke, TIA, and AMI) during the follow-up period. The safety outcomes were the occurrence of intracranial hemorrhage (ICH), gastrointestinal (GI) bleeding, and composite bleeding events (i.e., a composite measure of ICH and GI bleeding). We identified effectiveness and safety outcomes by using only primary discharge diagnoses to prevent misclassification. Mortality was also evaluated as the risks of fatal stroke, cardiovascular mortality, and all-cause mortality. The primary death diagnosis in the Cause of Death Database was used to determine both fatal stroke and cardiovascular mortality. Fatal stroke was classified as deaths resulting from ischemic or hemorrhagic strokes, while cardiovascular mortality was defined as deaths caused by ischemic heart diseases, cardiac arrhythmia, heart failure, or cerebrovascular diseases. The ICD-9-CM and ICD-10-CM codes used to identify the effectiveness, safety, and mortality outcomes were listed in Supplemental Table 1.
| Disease | ICD-9-CM codes | ICD-10-CM codes |
|---|---|---|
| Effectiveness outcomes | ||
| Acute ischemic stroke | 433.x, 434.x | I63.x |
| Transient ischemic attack | 435.x | H34.0, G45.x (excluding G45.4) |
| Acute myocardial infarction | 410.x | I21.0, I21, I21.2, I21.3, I21.4 |
| Safety outcomes | ||
| Intracranial hemorrhage | 430.x, 431.x, 432.x | I60.x, I61.x, I62.x |
| Gastrointestinal bleeding |
531.0, 531.2, 531.4, 531.6, 532.0, 532.2, 532.4, 532.6, 533.0x, 533.2, 533.4, 533.6, 534.0, 534.2, 534.4, 534.6, 562.02, 562.03, 562.12, 562.13, 569.3, 569.85, 578.1,578.9 |
K25.0, K25.2, K25.4, K25.6, K26.0, K26.2, K26.4, K26.6, K27.0, K27.2, K27.4, K27.6, K28.0, K28.2, K28.4, K28.6, K57.01, K57.11, K57.13, K57.21, K57.31, K57.33, K57.41, K57.51, K57.53, K57.81, K57.91, K57.93, K62.5, K55.21, K92.1, K92.2 |
| Mortality outcome (1): Fatal stroke | ||
| Death because of ischemic or hemorrhage stroke | 430.x, 431.x, 432.x, 433.x, 434.x | I60.x, I61.x, I62.x, I63.x |
| Mortality outcome (2): Cardiovascular mortality | ||
| Ischemic heart diseases | 410.x, 411.x, 412.x, 413.x, 414.x | I20.x, I21.x, I22.x, I23.x, I24.x, I25.x |
| Cardiac arrythmias | 426.x, 427.x | I44.x, I45.x, I46.x, I47.x, I48.x, I49.x |
| Heart failure | 428.x | I50.x |
| Cerebrovascular diseases |
430.x, 431.x, 432.x, 433.x, 434.x, 435.x, 436.x, 437.x, 438.x |
I60.x, I61.x, I62.x, I63.x, I64.x, I65.x, I66.x, I67.x, I68.x, I69.x |
Abbreviations: ICD-9-CM: The International Classification of Diseases, Ninth Revision and Clinical Modification; ICD-10-CM: International Classification of Diseases, Tenth Revision, Clinical Modification
Demographic and clinical characteristics that could affect study outcomes and antiplatelet treatment were adjusted for as covariates, including age, sex, comorbidities, and concomitant medications25-29). Comorbidities were identified from the pre-index period, including hypertension, AMI, atrial fibrillation, peripheral vascular disease, heart failure, dyslipidemia, diabetes mellitus, malignancy, renal dialysis, peptic ulcer disease, and ICH. The selected concomitant medications, namely proton pump inhibitors (PPIs), histamine-2 receptor antagonists (H2 blockers), oral anticoagulants, anti-hypertensive drugs, oral anti-diabetic drugs, insulin, statin, and non-steroidal anti-inflammatory drugs, were evaluated on the basis of prescriptions within 90 days before the index date. We categorized patients who received antihypertensive drugs into three groups based on the number of antihypertensive drug categories they used: one category, two categories, or three or more categories of antihypertensive drug. This categorization served as a surrogate measure of blood pressure control status and was included as covariates in the analysis. In addition, we adjusted for covariates associated with the index admission event severity, such as the length of hospital stay and mean proxy NIHSS score18, 19, 30).
Statistical AnalysisWe balanced the covariates between the exposure and comparison groups by using inverse probability of treatment weighting (IPTW)31). An absolute standardized mean difference (aSMD) ＜0.1 indicates that the covariates between exposure and comparison groups were balanced after IPTW adjustment32, 33). Cox-proportional regression hazard model were used to evaluate the association between the dual antiplatelet treatment (i.e., cilostazol plus aspirin or clopidogrel plus aspirin) and study outcomes. Hazard ratios (HRs) and 95% confidence intervals (CIs) were used to compare the effectiveness and safety of the treatments used in the exposure and comparison groups. We conducted all statistical analyses using SAS software version 9.4 (SAS Institute Inc., Cary, NC, USA). The Taipei Medical University Joint Institutional Review Board reviewed and approved of this study (TMU-JIRB, approval number: N202104063).
Sensitivity AnalysisWe conducted two sensitivity analyses to enhance the robustness of our study. In the first sensitivity analysis, we repeated the same methods as in the main analysis, with the only difference being the exclusion of patients with atrial fibrillation or oral anticoagulant use during or one year prior to the index date. This was done because patients requiring oral anticoagulant treatment or using oral anticoagulants may potentially affect the risk of bleeding during the follow-up period34). By performing this sensitivity analysis, we can determine whether the risk of bleeding between exposure and comparison groups remains consistent with our main analysis results, even after excluding patients with atrial fibrillation or those treated with oral anticoagulants.
In the second sensitivity analysis, we once again utilized the same design as in the main analysis, except for the exclusion of patients with a prior history of ICH before the index date. This exclusion was performed because a history of ICH could potentially affect the risk of future ICH development. The objective of this sensitivity analysis was to assess whether the results remain robust compared to the main analysis.
A total of 3,403 patients met the eligible criteria, of which 578 (17%) were treated with cilostazol plus aspirin and 2,825 (83%) were treated with clopidogrel plus aspirin. The baseline demographic and clinical characteristics before and after IPTW are listed in Table 1. Before IPTW, patients treated with cilostazol plus aspirin had longer hospital stays; a greater mean proxy NIHSS score; higher rates of peripheral vascular disease, diabetes mellitus, and ICH, and more concomitant use of oral anti-diabetic medications and insulin than did patients treated with clopidogrel plus aspirin. However, patients treated with clopidogrel plus aspirin had higher rates of atrial fibrillation, peptic ulcer disease, and concomitant use of PPIs and H2 blockers than did patients treated with cilostazol plus aspirin. The distributions of baseline demographics and clinical characteristics were adequately balanced after the IPTW adjustment.
| Before IPTW | After IPTW | |||||
|---|---|---|---|---|---|---|
|
Cilostazol (N= 578) (%) |
Clopidogrel (N= 2825) (%) |
aSMD e |
Cilostazol (N= 3435.73) (%) |
Clopidogrel (N= 3402.66) (%) |
aSMD e | |
| Male | 64.36 | 62.34 | 0.042 | 59.96 | 62.60 | -0.0544 |
| Age (mean, SD) | 67.81 (13.41) | 68.69 (13.50) | -0.0649 | 68.76 (33.49) | 68.52 (14.83) | 0.0093 |
| ＜65 years | 41.7 | 37.66 | 0.0825 | 38.54 | 38.29 | 0.0052 |
| 65-69 years | 11.42 | 12.35 | -0.0289 | 12.77 | 12.22 | 0.0166 |
| 70-74 years | 11.94 | 11.29 | 0.0202 | 10.67 | 11.37 | -0.0223 |
| 75-80 years | 13.49 | 13.7 | -0.006 | 13.90 | 13.67 | 0.0067 |
| ＞80 years | 21.45 | 24.99 | -0.0839 | 24.11 | 24.45 | -0.0079 |
| Index admission a event | ||||||
| Acute minor ischemic stroke | 83.04 | 76.25 | 0.1694 | 75.90 | 77.37 | -0.0348 |
| TIA | 16.96 | 23.75 | - | 24.10 | 22.63 | - |
| Proxy NIHSS b (mean, SD) | 3.37 (1.52) | 3.09 (1.72) | 0.1693 | 3.08 (4.24) | 3.14 (1.86) | -0.0182 |
| Length of hospital stay (mean, SD) | 10.03 (12.06) | 7.50 (6.70) | 0.2594 | 8.33 (22.02) | 7.72 (7.63) | 0.0002 |
| ＜7 days | 47.4 | 57.56 | -0.2044 | 55.34 | 55.84 | -0.0100 |
| 7-14 days | 39.45 | 34.58 | 0.1008 | 36.94 | 35.46 | 0.0309 |
| ＞14 days | 13.15 | 7.86 | 0.1732 | 7.72 | 8.71 | 0.0366 |
| Comorbidities c | ||||||
| Hypertension | 79.93 | 76.5 | 0.0833 | 76.30 | 77.07 | -0.018 |
| Dyslipidemia | 46.37 | 48.18 | -0.0363 | 47.15 | 47.78 | -0.0125 |
| Acute myocardial infarction | 0.69 | 0.25 | 0.0650 | 0.32 | 0.33 | -0.0010 |
| Peripheral vascular disease | 5.88 | 2.69 | 0.1581 | 3.67 | 3.29 | 0.0208 |
| Atrial fibrillation | 2.94 | 6.37 | -0.1634 | 4.39 | 5.78 | -0.0630 |
| Heart failure | 3.81 | 5.59 | -0.0845 | 5.65 | 5.31 | 0.0152 |
| Diabetes mellitus | 44.12 | 36.39 | 0.1581 | 35.51 | 37.73 | -0.0460 |
| Cancer | 6.06 | 8 | -0.0761 | 7.85 | 7.70 | 0.0059 |
| Peptic ulcer disease | 14.53 | 32.81 | -0.4404 | 31.30 | 29.74 | 0.0338 |
| Intracranial hemorrhage | 23.18 | 17.63 | 0.1382 | 19.22 | 18.55 | 0.0170 |
| Dialysis | 1.73 | 1.88 | -0.011 | 1.95 | 1.85 | 0.0074 |
| Selected concomitant medications d | ||||||
| Proton pump inhibitors | 13.49 | 30.8 | -0.426 | 28.77 | 27.87 | 0.0200 |
| Histamine 2 blockers | 37.72 | 41.84 | -0.0843 | 42.07 | 41.22 | 0.0171 |
| NSAIDs | 40.83 | 43.01 | -0.0442 | 44.34 | 42.77 | 0.0315 |
| Statin | 47.58 | 49.06 | -0.0297 | 48.11 | 48.82 | -0.0142 |
| Oral anti-diabetes drugs | 36.85 | 29.77 | 0.1507 | 29.63 | 30.98 | -0.0294 |
| Insulin | 19.55 | 13.42 | 0.1659 | 13.87 | 14.48 | -0.0173 |
| Oral anticoagulants | 2.94 | 2.87 | 0.0044 | 2.42 | 2.90 | -0.0302 |
| Beta-blocking agents | 33.74 | 29.77 | 0.0853 | 31.45 | 30.49 | 0.0209 |
| Calcium channel blockers | 48.79 | 49.91 | -0.0225 | 54.14 | 49.85 | 0.0860 |
| ACEI/ARB | 50.69 | 49.2 | 0.0298 | 50.97 | 49.53 | 0.0288 |
| Diuretics | 14.53 | 14.34 | 0.0056 | 17.25 | 14.52 | 0.0748 |
| Used a single category of antihypertensive | 1.04 | 1.88 | -0.0700 | 1.97 | 1.74 | 0.0174 |
| Used two categories of antihypertensive | 27.85 | 25.06 | 0.0633 | 28.81 | 25.65 | 0.0711 |
| Used ≥ three categories of antihypertensive | 21.63 | 20.25 | 0.0339 | 21.74 | 20.54 | 0.0295 |
Abbreviations: ACEI: angiotensin-converting enzyme inhibitor, ARB: angiotensin receptor blocker, aSMD: absolute standardized median difference, IPTW: inverse probability of treatment weighting, NIHSS: National Institutes of Health Stroke Scale, NSAIDs: nonsteroidal anti- inflammatory drugs, SD: standard deviation, TIA: transient ischemic attack.
Notes: a Index admission was the period of hospitalization for the primary acute minor ischemic stroke or TIA;
b The proxy NIHSS score was validated from using the data from the National Health Insurance Research Database in Taiwan. It was converted by the stroke severity index (SSI). The SSI was computed using seven in-hospital management factors that have been validated and found to be associated with the actual baseline NIHSS score, including intensive care unit stays, general ward stays, airway suctioning, bacterial sensitivity test results, nasogastric intubation, osmotherapy, and urinary catheterization.
c Comorbidities within 1 y prior to the index date were considered.
d Selected concomitant medications within 90 d prior to the index date were considered.
e aSMD ＜ 0.1 indicates covariates balance between cilostazol and clopidogrel groups.
Regarding the effectiveness outcomes, the average follow-up duration was 1.67 years for patients treated with cilostazol plus aspirin and 2.08 years for those treated with clopidogrel plus aspirin. Table 2 shows the incidence rates and HRs of the cilostazol plus aspirin and clopidogrel plus aspirin groups after IPTW. No significant differences were observed in the risks of acute ischemic stroke, TIA, AMI, or the composite cardiovascular events between the cilostazol plus aspirin and clopidogrel plus aspirin users.
| Incidences (per 1000-person year) (95%CI) | HR (95% CI) c | P-value | ||
|---|---|---|---|---|
| Cilostazol group | Clopidogrel group | |||
| Effectiveness | ||||
| Acute ischemic stroke | 51.92 (51.73-52.10) | 44.89 (44.74-45.05) | 1.06 (0.90-1.24) | 0.48 |
| TIA | 17.63 (17.53-17.74) | 12.83 (12.75-12.91) | 1.21 (0.91-1.61) | 0.18 |
| AMI | 1.01 (0.99-1.04) | 0.48 (0.46-0.49) | 1.81 (0.47-6.90) | 0.39 |
| Composite cardiovascular events a | 70.57 (70.35-70.78) | 58.20 (58.02-58.38) | 1.10 (0.96-1.26) | 0.18 |
| Safety | ||||
| ICH | 6.95 (6.89-7.02) | 4.14 (4.10-4.19) | 1.65 (1.05-2.60) | 0.03 |
| GI bleeding | 8.43 (8.36-8.50) | 8.76 (8.69-8.83) | 0.90 (0.63-1.29) | 0.57 |
| Composite major bleeding events b | 15.38 (15.29-15.48) | 12.90 (12.82-12.98) | 1.14 (0.86-1.50) | 0.36 |
| Mortality | ||||
| Fatal stroke | 0.61 (0.60-0.63) | 1.37 (1.35-1.39) | 0.47 (0.20-1.11) | 0.09 |
| CV mortality | 9.05 (9.00-9.11) | 10.13 (10.07-10.18) | 0.89 (0.69-1.15) | 0.38 |
| All-cause mortality | 56.64 (56.50-56.77) | 60.12 (59.99-60.26) | 0.95 (0.85-1.05) | 0.29 |
Abbreviation: AMI: Acute myocardial infarction, CI: Confidence interval, CV: Cardiovascular, GI: Gastrointestinal, HR: Hazard ratio, ICH: Intracranial hemorrhage, IPTW: Inverse probability of treatment weighting, TIA: Transient ischemic attack.
Notes: a The composite cardiovascular events of either acute ischemic stroke, TIA, or AMI.
b The composite major bleeding events of either ICH or GI bleeding.
c Hazard ratios were estimated by the Cox proportional hazards model with the clopidogrel group serving as a reference. Inverse probability of treatment weighting incidence rate and hazard ratio were controlled by age, sex, index admission event (i.e., acute minor ischemic stroke or TIA), mean NIHSS score, length of hospital stay, hypertension, dyslipidemia, acute myocardial infarction, peripheral vascular disease, atrial fibrillation, heart failure, diabetes mellitus, cancer, peptic ulcer disease, intracranial hemorrhage, dialysis, concomitant medication use of proton pump inhibitors, histamine 2 blockers, nonsteroidal anti-inflammatory drugs, antihypertensive drugs, statin, oral anti-diabetes drugs, insulin, oral anticoagulants, and diuretics.
Regarding safety outcomes (Table 2), the average follow-up duration was 1.87 years for the cilostazol plus aspirin group and 2.26 years for the clopidogrel plus aspirin group. Cilostazol plus aspirin was associated with a higher risk of ICH (HR: 1.65; 95% CI: 1.05-2.60), compared with clopidogrel plus aspirin. However, the risks of GI bleeding and the composite bleeding events did not differ between the two groups.
Regarding mortality outcomes (Table 2), the average follow-up duration was 3.43 years for the cilostazol plus aspirin group and 3.79 years for the clopidogrel plus aspirin group. We discovered no significant differences in the risks of fatal stroke, cardiovascular mortality, or all-cause mortality between the two groups.
Sensitivity AnalysisIn the sensitivity analysis that excluded patients with atrial fibrillation or those treated with oral anticoagulants, there were 551 (17%) users of cilostazol plus aspirin and 2,610 (83%) users of clopidogrel plus aspirin. After IPTW adjustment, the baseline characteristics were well balanced between the two groups. Supplemental Table 2 presents the incidence rates and HRs of the cilostazol plus aspirin and clopidogrel plus aspirin groups in this sensitivity analysis. Consistent with the main analysis results, cilostazol plus aspirin was associated with a higher risk of ICH (HR: 1.87; 95% CI: 1.17-2.99) compared to clopidogrel plus aspirin. However, in contrast to the main analysis, this sensitivity analysis revealed an increased risk of TIA (HR: 1.41; 95% CI: 1.05-1.90) in the cilostazol plus aspirin group compared to the clopidogrel plus aspirin group.
| Incidences (per 1000-person year) (95%CI) | HR (95% CI) c | P-value | ||
|---|---|---|---|---|
| Cilostazol group | Clopidogrel group | |||
| Effectiveness | ||||
| Acute ischemic stroke | 48.70 (48.51-48.89) | 43.46 (43.30-43.62) | 1.00 (0.85-1.19) | 0.97 |
| TIA | 19.66 (19.55-19.79) | 12.02 (11.94-12.11) | 1.41 (1.05-1.90) | 0.02 |
| AMI | 1.09 (1.07-1.12) | 0.35 (0.33-0.36) | 2.57 (0.56-11.87) | 0.23 |
| Composite cardiovascular events a | 69.46 (69.24-69.69) | 55.83 (55.65-56.01) | 1.10 (0.95-1.28) | 0.18 |
| Safety | ||||
| ICH | 7.61 (7.54-7.68) | 3.98 (3.93-4.02) | 1.87 (1.17-2.99) | ＜0.01 |
| GI bleeding | 8.58 (8.50-8.65) | 8.41 (8.34-8.47) | 0.94 (0.65-1.37) | 0.94 |
| Composite major bleeding events b | 16.19 (16.09-16.29) | 12.38 (12.30-12.46) | 1.23 (0.92-1.65) | 0.16 |
| Mortality | ||||
| Fatal stroke | 0.68 (0.67-0.70) | 1.37 (1.35-1.40) | 0.52 (0.22-1.24) | 0.14 |
| CV mortality | 9.27 (9.22-9.33) | 9.12 (9.16-9.27) | 1.01 (0.77-1.33) | 0.93 |
| All-cause mortality | 54.44 (54.30-54.57) | 57.14 (57.01-57.28) | 0.96 (0.86-1.07) | 0.46 |
Abbreviation:
AMI: Acute myocardial infarction, CI: Confidence interval, CV: Cardiovascular, GI: Gastrointestinal, HR: Hazard ratio, ICH: Intracranial hemorrhage, IPTW: Inverse probability of treatment weighting, TIA: Transient ischemic attack.
Notes:
aThe composite cardiovascular events of either acute ischemic stroke, TIA, or AMI.
b The composite major bleeding events of either ICH or GI bleeding.
c Hazard ratios were estimated by the Cox proportional hazards model with the clopidogrel group serving as a reference. Inverse probability of treatment weighting incidence rate and hazard ratio were controlled by age, sex, index admission event (i.e., acute minor ischemic stroke or TIA), mean NIHSS score, length of hospital stay, hypertension, dyslipidemia, acute myocardial infarction, peripheral vascular disease, heart failure, diabetes mellitus, cancer, peptic ulcer disease, intracranial hemorrhage, dialysis, concomitant medication use of proton pump inhibitors, histamine 2 blockers, nonsteroidal anti-inflammatory drugs, antihypertensive drugs, statin, oral anti-diabetes drugs, insulin, and diuretics.
In another sensitivity analysis excluding patients with a prior history of ICH before the index date, we identified 444 (16%) users of cilostazol plus aspirin and 2327 (84%) users of clopidogrel plus aspirin. After IPTW adjustment, baseline characteristics between the two groups were well balanced. Similar to the main analysis, cilostazol plus aspirin users had a significantly higher risk of ICH (HR: 1.02; 95% CI: 1.19-3.11) compared to clopidogrel plus aspirin users (Supplemental Table 3). However, in this sensitivity analysis, cilostazol plus aspirin was associated with a higher risk of TIA (HR: 1.44; 95% CI: 1.08-1.92) and composite cardiovascular events (HR: 1.17; 95% CI: 1.02-1.36) compared to clopidogrel plus aspirin.
| Incidences (per 1000-person year) (95%CI) | HR (95% CI) c | P-value | ||
|---|---|---|---|---|
| Cilostazol group | Clopidogrel group | |||
| Effectiveness | ||||
| Acute ischemic stroke | 49.72 (49.53-49.91) | 42.30 (42.15-42.46) | 1.09 (0.92-1.29) | 0.35 |
| TIA | 20.53 (20.40-20.65) | 12.34 (12.25-12.42) | 1.44 (1.08-1.92) | 0.01 |
| AMI | 1.12 (1.09-1.15) | 0.53 (0.51-0.55) | 1.76 (0.46-6.72) | 0.41 |
| Composite cardiovascular events a | 71.37 (71.14-71.59) | 55.17 (54.99-55.35) | 1.17 (1.02-1.36) | 0.03 |
| Safety | ||||
| ICH | 7.38 (7.31-7.45) | 3.80 (3.76-3.85) | 1.92 (1.19-3.11) | ＜0.01 |
| GI bleeding | 9.41 (9.34-9.49) | 8.85 (8.78-8.92) | 0.99 (0.69-1.42) | 0.95 |
| Composite major bleeding events b | 16.79 (16.69-16.90) | 12.65 (12.57-12.74) | 1.26 (0.95-1.68) | 0.11 |
| Mortality | ||||
| Fatal stroke | 0.67 (0.65-0.69) | 1.49 (1.47-1.52) | 0.47 (0.20-1.12) | 0.09 |
| CV mortality | 8.40 (8.34-8.45) | 10.49 (10.43-10.55) | 0.80 (0.61-1.05) | 0.10 |
| All-cause mortality | 55.90 (55.76-56.04) | 60.81 (60.67-60.95) | 0.92 (0.83-1.03) | 0.15 |
Abbreviation:
AMI: Acute myocardial infarction, CI: Confidence interval, CV: Cardiovascular, GI: Gastrointestinal, HR: Hazard ratio, ICH: Intracranial hemorrhage, IPTW: Inverse probability of treatment weighting, TIA: Transient ischemic attack.
Notes:
a The composite cardiovascular events of either acute ischemic stroke, TIA, or AMI.
b The composite major bleeding events of either ICH or GI bleeding.
c Hazard ratios were estimated by the Cox proportional hazards model with the clopidogrel group serving as a reference. Inverse probability of treatment weighting incidence rate and hazard ratio were controlled by age, sex, index admission event (i.e., acute minor ischemic stroke or TIA), mean NIHSS score, length of hospital stay, hypertension, dyslipidemia, acute myocardial infarction, peripheral vascular disease, heart failure, diabetes mellitus, cancer, peptic ulcer disease, intracranial hemorrhage, dialysis, concomitant medication use of proton pump inhibitors, histamine 2 blockers, nonsteroidal anti-inflammatory drugs, antihypertensive drugs, statin, oral anti-diabetes drugs, insulin, and diuretics.
To the best of our knowledge, this is the first study to compare the effectiveness, safety, and mortality risks between cilostazol plus aspirin and clopidogrel plus aspirin use among patients with acute minor ischemic stroke or TIA. We observed similar risks in the effectiveness outcomes and mortality between the two study groups. However, compared to clopidogrel plus aspirin, cilostazol plus aspirin was associated with a higher risk of ICH.
Our finding added the reference of the increased ICH risk of cilostazol plus aspirin use to the previous studies. The TOSS II study did not find significant differences in the risk of major hemorrhagic complications (i.e., life-threatening hemorrhage or other major hemorrhage) between cilostazol plus aspirin and clopidogrel plus aspirin15). On the other hand, a network meta-analysis found a relatively lower incidence of major bleeding risks in cilostazol plus aspirin compared to clopidogrel plus aspirin35). However, neither of the two aforementioned studies did not specifically evaluate the risk of ICH15, 35). A previous observational study found no difference in ICH risk between cilostazol and clopidogrel use by patients with chronic ischemic stroke36). However, they did not focus on studying patients with acute minor ischemic stroke or TIA. While both the MATCH study37) and CHARISMA study38) demonstrated that clopidogrel plus aspirin had a higher major bleeding risk compared to monotherapy antiplatelet, their primary focus was to assess whether dual antiplatelet therapy would have more benefit than monotherapy antiplatelet alone in high-risk patients with ischemic stroke or TIA. They did not directly compare different dual antiplatelet therapy regimens to find out which combinations of antiplatelet would be more suitable for this population. In contrast, our study directly compared two DAPT regimens (i.e., cilostazol plus aspirin and clopidogrel plus aspirin) and contributed to the existing evidence regarding the optimal DAPT regimen for patients with acute minor stroke or TIA, especially in terms of the bleeding risks associated with different DAPT regimens. This study is the first to demonstrate the association of cilostazol plus aspirin use with an increased risk of ICH to patients with acute minor ischemic stroke or TIA.
Our findings of no difference in the risks of acute recurrent ischemic stroke and AMI between cilostazol plus aspirin and clopidogrel plus aspirin are consistent with those of previous evidence. The TOSS II study and observational study both found similar risks of acute ischemic stroke and AMI for these dual antiplatelet therapies among patients with intracranial arterial stenosis15, 36). The present study population comprised predominantly Asian individuals, among which CYP2C19 poor metabolizers are especially common39). Although a previous study discovered that the benefit of clopidogrel plus aspirin was reduced for CYP2C19 poor metabolizers7), we found no increased risk of ischemic events for clopidogrel plus aspirin use by our predominantly Asian population. However, it is important to note that the CSPS.com study13) indicated that adding cilostazol to aspirin or clopidogrel could reduce the risk of ischemic stroke compared to using aspirin or clopidogrel alone. Given the disparities in patient populations and treatment regimens, it is possible that the risk of ischemic stroke associated with cilostazol use could vary between the CSPS.com study and this present study. Unlike CSPS.com study, which was designed for patients who were already treated with aspirin or clopidogrel before trial enrollment, our study focused on patients who newly initiated dual antiplatelet therapy following primary acute minor stroke or TIA events, with no prior use of antiplatelets. Furthermore, there is a difference in the antiplatelet regimen of interest between CSPS.com study and our study. CSPS.com study included both cilostazol plus aspirin and cilostazol plus clopidogrel as dual antiplatelet therapy regimens, comparing them to monotherapy aspirin or clopidogrel, whereas our study directly compared cilostazol plus aspirin to clopidogrel plus aspirin. Our study offers a more direct insight into the choice between cilostazol and clopidogrel as part of a dual antiplatelet therapy regimen for patients with acute minor ischemic stroke or TIA.
Similar to the main analysis, both sensitivity analyses, one excluding patients with atrial fibrillation or those treated with oral anticoagulants, and the other excluding patients with a prior history of ICH, also showed that cilostazol plus aspirin was associated with a higher risk of ICH. These results indicate that even after excluding patients with a potentially higher risk of future ICH development, we still observed an increased risk of bleeding in the cilostazol plus aspirin group. This finding reinforces the association between cilostazol plus aspirin and elevated risk of ICH. Additionally, we observed a potential finding that the risk of TIA was higher in the cilostazol plus aspirin group. However, the increased risk of TIA in the cilostazol plus aspirin group may be a chance finding and requires further confirmation in future studies with a larger population.
Our study has several strengths. First, this study employed a new-user design with an active comparator, as done extensively in previous pharmacoepidemiology studies40). This approach can minimize potential biases, such as confounding by indication, immortal-time bias, and healthy-user bias, and ensure the robustness of study results40). Second, we used data from a nationwide full population database from Taiwan, thereby achieving high generalizability and reliability of the clinical implications of the study results. Finally, this was the first study to evaluate cilostazol plus aspirin use among acute minor stroke or TIA in an Asian population. Asian individuals have higher risks of bleeding and thrombotic events associated with the use of anti-thrombotic agents compared with non-Asian individuals41, 42). However, the use of anti-thrombotic agents in Asians was rarely evaluated directly in previous study. Our study can be used as a reference for prescribing cilostazol plus aspirin to Asian patients with acute minor ischemic stroke or TIA.
Our study has several limitations. First, we could not access actual NIHSS scores and cerebrovascular angiography data because we used the data from an administrative claims database. However, we applied a proxy NIHSS score validated on the Taiwan NHIRD18, 19) and was used extensively in previous database studies43-46). Although the proxy NIHSS was validated from claims data, it remains an estimated approach based on the surrogate measures of stroke severity. We acknowledge that potential selection bias may still exist in this observational study despite using a validated approach. Second, several lifestyle-related factors, such as tobacco and alcohol consumption, and body mass index, that might confound our results could not be assessed. Third, the patients’ CYP2C19 genotypes were not available; thus, the proportion of patients who were poor clopidogrel metabolizers was difficult to evaluate. Fourth, as the data used in this study is secondary administrative claims data intended for reimbursement purposes, the precise timing of the initiation of dual antiplatelet therapy and whether the recurrent event occurred during the same hospitalization period are unknown. We were unable to identify recurrent events that occurred shortly before discharge. Fifth, we excluded TIA patients with a history of hospitalization but did not exclude those diagnosed in outpatient clinics or emergency rooms. It is possible that some TIA patients were not hospitalized. However, we attempted to address this issue by excluding patients who received antiplatelet treatment, which is commonly used for secondary stroke or TIA prevention. Thus, patients who did not use these antiplatelet medications at baseline are less likely to have a history of ischemic stroke or TIA. Sixth, we could not obtain laboratory data related to the ischemic stroke or TIA prognosis from the Health and Welfare Database, such as blood sugar, cholesterol, serum creatinine, or serum uric acid levels. Seventh, we were unable to track blood pressure values or changes during the follow-up period because blood pressure data were not available in our database. Given that blood pressure status can impact the risk of future ICH, we used the number of antihypertensive medications taken at baseline as a proxy to help balance the baseline blood pressure condition. However, it is important to note that the absence of data on blood pressure control during the follow-up period remains a limitation of our study. Moreover, stroke subtypes identification is challenging when using administrative claims database because we did not have access to image reports from Computed Tomography (CT) or Magnetic Resonance Imaging (MRI) scans, while subtype of lacunar infarction may bear relative higher risk of ICH. Although we made efforts to use a validated proxy NIHSS score as a surrogate to identify possible lacunar infarction patients, it remains a limitation of our study. In addition, detecting cerebral microbleeds is also challenging due to the absence of CT or MRI information. While we attempted to address this by conducting a sensitivity analysis by excluding patients with ICH history as a proxy, we acknowledge that residual confounding related to the presence of cerebral microbleeds remains a potential limitation of our study. Future studies could explore ways to access information on image reports to further validate the results of our study.
In conclusion, the risks of composite cardiovascular events (i.e., a composite measure of acute ischemic stroke, TIA, or AMI) and mortality were similar between cilostazol plus aspirin and clopidogrel plus aspirin use in patients with acute minor ischemic stroke or TIA. However, cilostazol plus aspirin was associated with a higher risk of ICH than was clopidogrel plus aspirin. Patients with acute minor ischemic stroke or TIA treated with cilostazol plus aspirin should be monitored closely to ensure their safety.
The data for this study were obtained from the Health and Welfare Database, which is provided by the Health and Welfare Data Science Center of the Department of Statistics under the Ministry of Health and Welfare (HWDC, MOHW), Taiwan. The interpretation and conclusions contained in this study do not represent those of the Health and Welfare Data Science Center.
This study was funded by Shuang-Ho hospital, New Taipei city, Taiwan (W109YSR-05). The funder had no role in the study design, data analysis, interpretation of results, manuscript preparation, or publication decisions.
Dr. Hsin-Yi Huang, Dr. Jia-Hung Chen, Dr. Nai-Fang Chi, and Dr. You-Chia Chen declare that there is no conflict of interest to the content of this study.
The datasets analyzed during the current study are not publicly available due to legal restrictions by Taiwan’s policy. Data are available from the corresponding author on reasonable request and with permission from the Health and Welfare Data Science Center, Ministry of Health and Welfare in Taiwan.
This study was reviewed and approved by the Taipei Medical University Joint Institutional Review Board (TMU-JIRB, approval number: N202104063).
All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Dr. Hsin-Yi Huang. The first draft of the manuscript was written by Dr. Hsin-Yi Huang and Dr. You-Chia Chen and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
