Study selection
Overall, 5396 records were identified through database searching. 125 additional records were identified through other sources. After removal of duplicates, a total of 4406 records remained and were initially screened by title and abstract. Of these, 40 were included for full-text review. Twelve studies were included in the final analysis consisting of six RCTs [10, 11, 28–31] and six observational studies [12, 13, 32–35]. The study selection is illustrated in Fig. 1.
Study characteristics
The characteristics of the included studies are summarized in Table 1. Overall, these 12 studies involved 37,557 patients. Six were RCTs (6,131 patients), while the other 6 were observational studies (31,426 patients). The mean age of study participants was 69.9 years (Standard deviation [SD], 5.1) and 24.3% of participants were female (interquartile range [IQR], 19.6% − 33.8%). Eight studies (66.7%) were performed with ACS patients [10–13, 31, 30, 33, 35], and four studies (33.3%) were performed with CAD patients [28, 29, 32, 34]. Nine studies (75.0%) examined the effects of a 3.75 mg dose of prasugrel [10, 12, 13, 28, 31–35], two studies (16.7%) examined 5 mg dose [11, 30] and one (8.3%) compared these two doses [29]. Most studies were conducted in Japan (10 studies) [11–13, 28–30, 32–35], while the other 2 studies included one conducted in Italy [31] and one that gathered data from 52 countries around the world [10]. The median duration of follow-up was 12 months (IQR, 8–12 months). Characteristics of all included studies are given in Table S2.
Table 1
Study characteristics of included studies
Study | Study design | Country | Target population | Total population | Patients | Intervention (n) | Control (n) | Outcomes | Follow-up (months) | Age (years), Mean | Female (%) |
---|
MACE | Major Bleeding | |
---|
Isshiki (2014)[28] | RCT | Japan | Asian (Japanese patients) | 742 | CAD (stable angina or prior MI) undergoing PCI | prasugrel 20/3.75 mg (370) | clopidogrel 300/75 mg (372) | Yes | Yes | 12 | 27.6 | 67.5 |
Kimura (2015)[29] | RCT | Japan | Asian (Japanese patients) | 422 | CAD (stable angina or prior MI) undergoing PCI | prasugrel 20/3.75-5 mg (207) | clopidogrel 300/75 mg (104) | Yes | Yes | 3.5 | 14.4 | 64.5 |
Roe (2013)[10] | RCT | Italy | Elderly patients (age ≥ 75 years) | 2083 | ACS (UA/NSTEMI) | prasugrel NA/5 mg (1043) | clopidogrel NA/75 mg (1040) | Yes | Yes | 30 | 50.5 | 79.5 |
Saito (2014)[11] | RCT | Japan | Asian (Japanese patients) | 1363 | ACS (UA/STEMI/NSTEMI) undergoing PCI | prasugrel 20/3.75 mg (685) | clopidogrel 300/75 mg (678) | Yes | Yes | 12 | 21.2 | 65.3 |
Savonitto (2018)[31] | RCT | 52 countries around the world | Elderly patients (age ≥ 75 years) | 1443 | ACS (STEMI/NSTEMI) undergoing PCI | Prasugrel 60/5 mg (713) | clopidogrel 300–600/75 mg (730) | Yes | Yes | 12 | 40.0 | 80.0 |
Kitano (2019)[30] | RCT | Japan | Asian (Japanese patients) | 78 | ACS undergoing PCI | prasugrel 20/3.75 mg (39) | clopidogrel 300/75 mg (39) | Yes | No | 8 | 18.0 | 64.8 |
Akita (2020)[12] | Cohort | Japan | Asian (Japanese patients) | 24032 | ACS undergoing PCI | prasugrel 20/3.75 mg (12016) | clopidogrel 300/75 mg (12016) | No | Yes | In-hospital | 25.9 | 70.6 |
Koyabu (2019)[32] | Cohort | Japan | Asian (Japanese patients) | 500 | CAD (CAD and ACS) undergoing PCI | prasugrel 20/3.75 mg (250) | clopidogrel 300/75 mg (250) | Yes | Yes | 12 | 79.6 | 67.5 |
Ohya (2018)[33] | Cohort | Japan | Asian (Japanese patients) | 992 | ACS undergoing PCI | prasugrel 20/3.75 mg (487) | clopidogrel NA/75 mg (192) | Yes | Yes | 17 | 17.0 | 71.0 |
Shoji (2020)[13] | Cohort | Japan | Asian (Japanese patients) | 1802 | ACS undergoing PCI | prasugrel 20/3.75 mg (901) | clopidogrel 300/75 mg (901) | Yes | Yes | In-hospital | 24.0 | 71.5 |
Tokimasa (2019)[34] | Cohort | Japan | Asian (Japanese patients) | 1031 | CAD undergoing PCI | prasugrel 20/3.75 mg (412) | clopidogrel 300/75 mg (619) | Yes | Yes | 6 | 23.1 | 67.9 |
Yasuda (2019)[35] | Cohort | Japan | Asian (Japanese patients) | 3069 | ACS (STEMI or NSTEMI) undergoing PCI | prasugrel 20/3.75 mg (2607) | clopidogrel 300/75 mg (462) | Yes | Yes | 12 | 23.5 | 69.8 |
Risk of bias
The risk of bias in five RCTs (83.3%) was rated as low risk [10, 11, 28, 29, 31]. We judged the risk of bias as high in one RCT (16.7%) due to inadequate information about allocation concealment in the randomization process [30]. The risk of bias in three observational studies (50.0%) was considered moderate [12, 13, 35], and the other three studies (50.0%) were considered to have serious risk of bias due to the existence of possible confounds (Figure S1) [32–34],
MACE
MACE was reported in six RCTs (n = 6020) [10, 11, 28–31], and five observational studies (n = 7081) [13, 32–35]. There was no difference in the risk of MACE between patients receiving low-dose prasugrel and patients receiving clopidogrel (RR 1.02, 95%CI 0.91 to 1.14; Fig. 2) with low heterogeneity (I2 = 46.7%). Treating study design as a factor in the model yielded no significant interaction with medication (RCTs: RR 0.98, 95%CI 0.87 to 1.11; I2 = 7.8%); observational studies: RR 1.19, 95%CI 0.93 to 1.53; I2 = 65.1%); P = 0.171 for interaction). The null result did not change when studies at serious risk of bias were excluded (RR 1.01, 95%CI 0.91 to 1.14; I2 = 12.2%; n = 10,891) (Figure S2). The GRADE confidence in this estimate is low. No publication bias was found on inspection of funnel plots or on Egger’s regression test.
Major bleeding
Major bleeding was reported in five RCT (n = 5,919) [10, 11, 28, 29, 31] and six observational studies (n = 31,113) [12, 13, 32–35]. Overall, patients receiving low-dose prasugrel exhibited an increased risk of major bleeding relative to patients receiving clopidogrel (RR 1.35, 95%CI 1.10 to 1.67; I2 = 61.4%; n = 37,110; Fig. 3). Separate analysis of observational and RCT studies yielded an increased risk of major bleeding for patients receiving prasugrel in observational studies (RR 1.57, 95%CI 1.23 to 2.01; I2 = 66.3); however, there was no difference in the risk of major bleeding between drug groups in RCTs (RR 0.91, 95%CI 0.61 to 1.36; I2 = 32.1%). To check for possible impact of bias, we repeated the main analysis with studies at serious risk of bias excluded. This analysis replicated the main finding with patients receiving prasugrel exhibiting an increased risk of major bleeding (RR 1.37, 95%CI 1.08 to 1.73; I2 = 64.3%; n = 34,900) (Figure S3). The GRADE confidence in this estimate is low. No publication biases were found either on inspection of funnel plots or on Egger’s regression test.
Minor bleeding
Minor bleeding was reported in five RCTs (n = 5,919) [10, 11, 28, 29, 31] and four observational studies (n = 4,012) [13, 32–34]. Overall, patients receiving low-dose prasugrel exhibited an increased risk of minor bleeding relative to patients receiving clopidogrel (RR 1.47, 95%CI 1.09 to 1.99; I2 53.0%). Separate analysis of observational and RCT studies yielded nonsignificant trends for both types of study design (Figure S4). The GRADE confidence in this estimate is low. No publication bias was found either on inspection of funnel plots or on Egger’s regression test.
Additional outcomes
We found no difference between patients given low-dose prasugrel and standard-dose clopidogrel in the risk of all-cause mortality, cardiovascular death, myocardial infarction, stroke, or stent thrombosis (Figure S5-S9).
Subgroup analyses
Several subgroup analyses were conducted to examine the in different characteristics such as clinical events, aged groups, types of PCI, study country (Table S4). Overall, these subgroups analyses in different characteristics did not change the findings of the main analysis. However, when we separate analyses according to time to measurement outcome, low-dose prasugrel was significant increased risk of major bleeding in short-term (RR 1.97, 95%CI 1.43 to 2.71; I2 = 59.6%) with trended to increase long-term major bleeding (RR 1.14, 95%CI 0.78 to 1.67; I2 = 61.5%) (Figure S10). No publication biases were found either on inspection of funnel plots or on Egger’s regression test in all analyses (Figure S11).