Association of Timing for Starting Dual Antiplatelet Treatment With Cilostazol and Recurrent Stroke: A CSPS.com Trial Post Hoc Analysis.

BACKGROUND AND OBJECTIVES: Long-term treatment with the combination of cilostazol with aspirin or clopidogrel showed a lower risk of stroke recurrence compared to aspirin or clopidogrel alone after high-risk noncardioembolic ischemic stroke in a randomized trial. We aimed to determine whether the effect of the dual medication compared to monotherapy on risk of recurrent ischemic stroke differs according to timing of starting medication after stroke onset.
METHODS: In a subanalysis of the randomized controlled trial, patients between 8 and 180 days after stroke onset were randomly assigned to receive aspirin or clopidogrel alone or a combination of cilostazol with aspirin or clopidogrel. They were divided into 3 groups according to the timing of starting trial treatment: between 8 and 14 days after stroke onset (8-14 days group), between 15 and 28 days after stroke onset (15-28 days group), and between 29 and 180 days after stroke onset (29-180 days group). The primary efficacy outcome was the first recurrence of ischemic stroke. Safety outcomes included severe or life-threatening bleeding.
RESULTS: Of 1,879 patients, 498 belonged to the 8-14 days group, 467 to the 15-28 days group, and 914 to the 29-180 days group. There was a significant treatment-by-subgroup interaction for the recurrence of ischemic stroke between trial treatment and trichotomized groups. The recurrence of ischemic stroke was less common with dual therapy than with monotherapy in the 15-28 days group (annualized rate 1.5% vs 4.9%, respectively; adjusted hazard ratio 0.34 [95% CI 0.12-0.95]) and the 29-180 days group (1.9% vs 4.4%, respectively; 0.27 [0.12-0.63]) and similarly common in the 8-14 days group (4.5% for both; 1.02 [0.51-2.04]). Severe or life-threatening bleeding occurred similarly between patients on dual therapy and those on monotherapy in any of the trichotomized groups (crude hazard ratio 0.22 [95% CI 0.03-1.88] in the 8-14 days group, 1.07 [0.15-7.60] in the 15-28 days group, and 0.76 [0.24-2.39] in the 29-180 days group). DISCUSSION: Long-term dual antiplatelet therapy using cilostazol starting 15-180 days after stroke onset, compared to therapy started 8-14 days after onset, was more effective for secondary stroke prevention than monotherapy without increasing hemorrhage risk. TRIAL REGISTRATION INFORMATION: ClinicalTrials.gov NCT01995370; UMIN Clinical Trials Registry 000012180. CLASSIFICATION OF EVIDENCE: This study provides Class II evidence that for patients with acute noncardioembolic stroke taking either aspirin or clopidogrel, the addition of cilostazol 15-180 days after stroke onset decreases the risk of recurrent ischemic stroke.

Although dual antiplatelet therapy (DAPT) using aspirin and clopidogrel (ASA+CLO) decreases the risk of recurrent stroke early after minor ischemic stroke or high-risk TIA compared to aspirin monotherapy,[1,2] its effect on risk reduction of stroke is attenuated after the first month, when it also starts to have a higher risk of major bleeding than monotherapy.[3-6] A meta-analysis showed that DAPT with ASA+CLO started mainly within the first 24 hours of stroke onset significantly reduced the risk of recurrent ischemic stroke within the initial 3 months, but significantly increased the risk of major bleeding more than a month after stroke onset.[7] Early initiation of DAPT, within 30 days of stroke, using ticagrelor and aspirin, compared to aspirin alone, was proven to decrease the risk of the composite of stroke or death, but this combination increased severe bleeding.[8] In addition, the effect at the second month or later has not been explored.

An alternative DAPT, cilostazol, a phosphodiesterase 3 inhibitor, coupled with aspirin (CLZ+ASA) or clopidogrel (CLZ+CLO) reduced the risk of recurrent ischemic stroke and had a similar risk of severe or life-threatening bleeding compared to aspirin or clopidogrel alone in patients at high risk for recurrent ischemic stroke in long-term use of a median 1.4 years in the Cilostazol Stroke Prevention Study for Antiplatelet Combination (CSPS.com), a randomized controlled trial (RCT).[9-11] If DAPT using cilostazol is effective when started within 1 month after stroke onset, we may be able to obtain a promising therapeutic strategy by switching the aspirin or clopidogrel to cilostazol and continuing DAPT from shortly after stroke onset for years.

We examined the hypothesis that starting DAPT using cilostazol within the initial several weeks was more effective and as safe as monotherapy using aspirin or clopidogrel by analyzing the data from the CSPS.com dataset. The primary research question was to determine the optimal time for starting long-term medication with the combination of cilostazol with aspirin or clopidogrel after onset of high-risk noncardioembolic ischemic stroke.

Methods

Standard Protocol Approvals, Registrations, and Patient Consents

CSPS.com was registered at ClinicalTrials.gov (NCT01995370) and the University Hospital Medical Information Network clinical trial registry in Japan (UMIN 000012180) and approved by the ethics committee at each participating site. All patients gave written informed consent before randomization. This study followed the Consolidated Standards of Reporting Trials (CONSORT) guidelines.

Patients

CSPS.com was a multicenter, randomized, open-label, parallel-group trial, involving participants from 292 sites across Japan registered from December 2013 through March 2017. The trial protocol, statistical analysis plan, and design and main results of CSPS.com were described previously.[9,12]

Eligible patients were between 20 and 85 years of age and had a noncardioembolic ischemic stroke identified on MRI between 8 and 180 days before the start of the protocol treatment and were taking either aspirin or clopidogrel alone as antiplatelet therapy when providing informed consent. The patients were required to meet at least 1 of the following 3 criteria indicating a high risk for stroke recurrence: (1) ≥50% stenosis of a major intracranial artery; (2) ≥50% stenosis of an extracranial artery; and (3) 2 or more of the following risk factors: age ≥65 years, hypertension, diabetes mellitus, chronic kidney disease, peripheral arterial disease, history of ischemic stroke other than the qualifying one for this trial, history of ischemic heart disease, and current smoking. Additional information regarding the inclusion and exclusion criteria is provided elsewhere.[9,12] For example, patients with emboligenic heart disease were excluded from the study.

In this substudy, the patients were trichotomized according to the time of initiation of the trial medication: between 8 and 14 days (the 8–14 days group), between 15 and 28 days (the 15–28 days group), and between 29 and 180 days (the 29–180 days group) after stroke onset. Fourteen and 28 days are meaningful time points to estimate when DAPT using cilostazol could replace DAPT using ASA+CLO, which is effective within the initial 3 to 4 weeks.[1-6]

Patients were randomly assigned in a 1:1 ratio to receive either monotherapy with aspirin (81 or 100 mg) or clopidogrel (50 or 75 mg) once daily or dual therapy with Pletal, a brand-name product of cilostazol (100 mg, twice daily, the recommended dose for stroke prevention in Japan) and either aspirin (81 or 100 mg) or clopidogrel (50 or 75 mg), once daily. Trial medication was continued for half a year or longer, for a maximum of 3.5 years. To prevent adverse drug reactions such as headache and tachycardia, cilostazol treatment could be started at 100 mg/d and increased to 200 mg/d within 15 days. Changes in these 3 antiplatelet medications were not permitted after informed consent was obtained.

Outcomes

The primary efficacy outcome was the first recurrence of ischemic stroke. The secondary efficacy outcomes were (1) any stroke (ischemic or hemorrhagic); (2) a composite of stroke, myocardial infarction (MI), and vascular death; (3) all vascular events, including stroke, MI, and other vascular events (e.g., aortic dissection; aortic rupture; pulmonary embolism; heart failure, angina pectoris, or peripheral artery disease requiring hospitalization; revascularization of a coronary artery, aorta, cephalocervical artery, or peripheral arteries); and (4) death from any cause.

Safety outcomes were (1) severe or life-threatening bleeding as defined in the Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries (GUSTO) classification,[13] which includes symptomatic intracranial hemorrhage (hemorrhagic stroke, subdural or epidural hemorrhage) and bleeding resulting in substantial hemodynamic compromise requiring treatment; and (2) symptomatic intracranial hemorrhage. In addition, as early side effects specific to cilostazol, a composite of occurrence of headache, palpitations, and tachycardia was assessed.[14-16]

Statistical Analysis

Efficacy and safety analyses were performed on the intention-to-treat population. A Cox proportional hazards model with a forward-backward stepwise selection algorithm based on Akaike information criterion (AIC) was applied to calculate adjusted hazard ratios (HRs) and 95% CIs for the dual therapy relative to the monotherapy. Multivariate analysis was not performed when the number of events was small. Moreover, the interaction between the treatment groups and the time of initiation of the trial medication was evaluated using the Cox proportional hazards model, as well as the exploratory evaluation of the optimal time to initiation of trial treatment.

The risk of ischemic stroke with dual therapy compared to monotherapy according to the time of initiation of the trial medication is plotted in the following manner: the point estimates and 95% CIs of crude HRs of ischemic stroke in the subgroup were calculated for each day for starting medication (day X), where the subgroup was defined as patients starting medication between X and 180 days after stroke onset.

A value of p < 0.1 was considered significant for assessment of the treatment by cohort interaction. In the other comparisons, a value of p < 0.05 based on a 2-sided test was considered significant. Statistical analysis was performed using R (version 4.0.2, Microsoft).

Data Availability

Deidentified individual participant data of CSPS.com may be available upon request to Japan Cardiovascular Research Foundation if the request is intended to contribute to the improvement of people's health and welfare.

Results

Of the 1,879 randomized patients, 932 were assigned to dual therapy and 947 to monotherapy. Of these, 498 (26.5% [240 assigned to dual therapy and 258 to monotherapy]) started trial medication between 8 and 14 days, 467 (24.9% [229 and 238, respectively]) between 15 and 28 days, and the remaining 914 (48.6% [463 and 451, respectively]) between 29 and 180 days after stroke onset (Figure 1). Their baseline characteristics are shown in Table 1.

Figure 1

Trial Profile

Table 1

Baseline Characteristics

There was a significant treatment by subgroup interaction for the primary efficacy outcome of ischemic stroke between assigned treatment and trichotomized groups (p = 0.07 for the 15–28 days group, p = 0.022 for the 29–180 days group, and p = 0.012 for the combined group of the above two [15–180 days group] compared with the 8–14 days group).

Patients Starting Trial Medication Between 8 and 14 Days After Stroke Onset

The median duration of follow-up was 1.3 years overall (interquartile range [IQR] 0.5–2.1 years), resulting in 713.8 person-years of follow-up. Ischemic stroke occurred in 15 patients (annualized rate 4.5%) during follow-up in the dual therapy group and 17 (4.5%) in the monotherapy group (crude HR 1.02 [95% CI 0.51–2.04]) (Table 2 and Figure 2A); assigned treatment was not chosen after stepwise selection by AIC. There were no significant differences in any secondary efficacy outcomes between the 2 groups. Severe or life-threatening bleeding occurred in 1 (0.4%) and 5 (1.7%), respectively (crude HR 0.22 [95% CI 0.03–1.88]; Table 3). All 6 events were intracranial hemorrhages.

Table 2

Efficacy Outcomes

Figure 2

Kaplan-Meier Analysis for Time to the First Event of the Primary Efficacy Outcome, Defined as Ischemic Stroke

(A) 8–14 days group. (B) 15–28 days group. (C) 29–180 days group. (D) 15–180 days group.

Table 3

Safety Outcomes

Patients Starting Trial Medication Between 15 and 28 Days After Stroke Onset

The median duration of follow-up was 1.3 years overall (IQR 0.6–2.1 years), resulting in 658.1 patient-years of follow-up. Ischemic stroke occurred in 5 patients (annualized rate 1.5%) on dual therapy and 16 (4.9%) on monotherapy (adjusted HR 0.34 [95% CI 0.12–0.95]) (Table 2 and Figure 2B). All vascular events occurred less commonly in patients on dual therapy than on monotherapy. Severe or life-threatening bleeding occurred in 2 (0.9%) and 2 (1.0%), respectively (crude HR 0.95 [95% CI 0.13–6.74]; Table 3). All 4 events were intracranial hemorrhages.

Patients Starting Trial Medication Between 29 and 180 Days After Stroke Onset

The median duration of follow-up was 1.4 years overall (IQR 0.5–2.2 years), resulting in 1,352.2 patient-years of follow-up. Ischemic stroke occurred in 9 patients (annualized rate 1.9%) on dual therapy and 31 (4.4%) on monotherapy (adjusted HR 0.27 [95% CI 0.12–0.63]) (Table 2 and Figure 2C). The incidence of ischemic stroke between patients on dual therapy and those on monotherapy was also different in the combined 29–180 days plus 15–28 days groups (15–180 days group, adjusted HR 0.34 [95% CI 0.18–0.65]) (Figure 2D). Any secondary efficacy outcomes other than death occurred less commonly in the 29–180 days patients on dual therapy than on monotherapy. Proportionality of the hazards in the multivariable models was not balanced regarding any stroke (p = 0.044; p values for the other outcomes are described in the footnote of Table 2). Severe or life-threatening bleeding occurred in 5 (1.1%) and 7 (1.6%), respectively (crude HR 0.78 [95% CI 0.25–2.45]; Table 3). All 12 events were intracranial hemorrhages.

Figure 3 demonstrates that the adjusted HR and 95% CI for the risk of ischemic stroke in the dual therapy group compared to the monotherapy group tended to be lowest when the trial medication was started 17 days after stroke onset. A composite of headache, palpitations, and tachycardia was seen in 22 patients (9.1%) assigned to dual therapy in the 8–14 days group, 10 (4.4%) in the 15–28 days group, and 45 (9.7%) in the 29–180 days group (p = 0.047).

Figure 3

Risk of Ischemic Stroke During Dual Therapy Compared to Monotherapy According to Time of Initiation of the Trial Medication

Adjusted hazard ratio and 95% CI for the patients who started trial medication on the day shown on the X axis or later.

This study provides Class II evidence that for patients with acute noncardioembolic stroke taking either aspirin or clopidogrel, the addition of cilostazol 15–180 days after stroke onset decreases the risk of recurrent ischemic stroke.

Discussion

CSPS.com shows the efficacy and safety of long-term DAPT for secondary prevention in high-risk, noncardioembolic ischemic stroke by using CLZ+ASA or CLZ+CLO. In this subanalysis of the trial, DAPT with this regimen was more effective and similarly safe in patients who started trial medication 15 days or later after stroke onset compared to aspirin or clopidogrel monotherapy, but not in those starting medication between 8 and 14 days after stroke onset.

Cilostazol inhibits phosphodiesterase activity and suppresses cyclic adenosine monophosphate (cAMP) degradation, increases intracellular cAMP concentrations, activates the cAMP-dependent protein kinase A, and inhibits platelet aggregation.[14-16] The same mechanism causes a vasodilatory effect on smooth muscle cells. In addition, cilostazol has pleiotropic effects, such as vasoprotection, neuroprotection, antiproliferation, anti-inflammation, lipid-lowering, and protection against ischemia-reperfusion injury.[16] These long-lasting effects seem to be beneficial for long-term stroke prevention. Cilostazol was recommended as the first-line antiplatelet agent for secondary stroke prevention in Japan[17] and several Asian countries and was weakly recommended for stroke or TIA attributable to moderate to severe intracranial artery stenosis in the United States.[18]

Cilostazol has potential strengths for use during acute stroke, including a rapid onset of action and a low risk of bleeding.[14-16] Some RCTs showed similar or somewhat better efficacy and safety after early initiation of cilostazol alone or CLZ+ASA within 48 hours after stroke onset than aspirin alone.[19,20] However, in a relatively large RCT involving 1,201 participants, initiation of CLZ+ASA within 48 hours of stroke onset did not show a decrease in the rate of a composite of neurologic deterioration, symptomatic stroke recurrence, and TIA within 14 days, and showed an insignificant tendency (p = 0.086) for an increased rate of modified Rankin Scale score of 0–1 at 3 months and a similar incidence of hemorrhagic stroke relative to aspirin monotherapy.[21] In a systematic review and meta-analysis involving these RCTs, initiation of cilostazol within 2 weeks of stroke onset did not show a better outcome at 1–4 months than the control treatment.[14] Thus, CLZ+ASA would not be an optimal alternative choice to ASA+CLO during an acute stroke. Accordingly, it seemed to be important to clarify whether DAPT including cilostazol is effective for patients at 3–4 weeks after onset, the time when ASA+CLO starts to lose its advantage.[1,7]

In the present study, CLZ+ASA or CLZ+CLO significantly decreased the risk of ischemic stroke compared to monotherapy in both the 15–28 days and 29–180 days groups. The event curves of ischemic stroke during dual therapy and during monotherapy separated gradually from 3 to 4 months after randomization in the 15–28 days group. A similar separation of the event curves was observed in other RCTs that compared cilostazol monotherapy or CLZ+ASA with aspirin monotherapy.[22,23] In the meta-analysis, cilostazol reduced recurrent ischemic stroke more when given more than 6 months than when given short term.[14] The delayed preventive effect against stroke suggests the contribution of the above-mentioned long-lasting pleiotropic mechanisms unique to cilostazol.[14-16] A potential antidementia effect of cilostazol might maintain quality of life and indirectly prevent stroke.[14,24] The mechanisms would be commonly shared by phosphodiesterase inhibitors, as adding dipyridamole, a phosphodiesterase 5 inhibitor, to aspirin reduced the risk of recurrent stroke only after 12 weeks in a meta-analysis.[25] A particularly desirable time for starting DAPT using cilostazol between 15 and 180 days after stroke onset would be around 17 days based on the data shown in Figure 3.

In contrast, CLZ+ASA or CLZ+CLO did not decrease the risk of ischemic stroke relative to monotherapy in the 8–14 days group. This would be partly due to milder antiplatelet effects presumed by the lower risk of bleeding than aspirin or clopidogrel, which is a critical defect in the acute to subacute stage when the patient is at high risk for recurrent thromboembolism.[15,26] The event curves in this group were almost identical between dual therapy and monotherapy not only during the subacute stage, but also during the long-term chronic stage. A reason why dual therapy did not show superiority during the chronic stage would be the difference in the baseline characteristics of the patients between the 8–14 days group and the others; intracranial artery stenosis and clopidogrel usage at randomization, both regarded to be suitable conditions for dual therapy in CSPS.com and other studies,[9-11,27] were less common in the 8–14 days group.

Secondary efficacy outcomes except for any death were less common or marginally less common during dual therapy than monotherapy in both the 15–28 days and 29–180 days groups. Bleeding events were similarly common between dual therapy and monotherapy in any of the 3 groups divided by timing. A low risk of bleeding is a strength of cilostazol that enables safe long-term DAPT including cilostazol. The composite of headache, palpitations, and tachycardia was similarly common between the 8–14 days and 29–180 days groups and somewhat less common in the 15–28 days group. At least these events should be prevented regardless of the timing. The optional choice of cilostazol 100 mg once per day (half a regular dose) as an initial dose for 15 days adopted in CSPS.com seemed to decrease such early side effects.[8]

The limitations of CSPS.com were described elsewhere, including the smaller number of enrolled patients than planned, relatively low occurrence of the primary efficacy endpoint, uncertainty of the generalizability of the findings to non-Japanese populations, and relatively frequent discontinuation of trial medication due to headache and palpitations, known early side effects of cilostazol.[9] Because this was a subanalysis that divided the overall participants into 3 groups, efficacy endpoints and, in particular, safety endpoints in each group were even fewer, which might cause statistical bias. In addition, there were differences in the baseline characteristics of the patients among the 3 groups divided by timing, because timing was not randomized. This made the interpretation of intergroup differences in efficacy outcomes complicated. Finally, the present findings might not be generalizable to women given that only three-tenths enrolled were women.

Long-term DAPT using cilostazol was more effective for secondary prevention of noncardioembolic stroke than monotherapy in high-risk patients who started the medication 15 days or later after stroke onset without increasing hemorrhage risk. The finding suggests the feasibility of a seamless DAPT strategy after stroke, switching from ASA+CLO in the acute to subacute stage to CLZ+ASA or CLZ+CLO at 15 days or later. Clinical studies to prove this strategy are needed.