The randomized controlled trials of aspirin on primary prevention of stroke in PubMed, Cochrane Library, EMBASE, MEDLINE, web of science, CNKI, China biomedical literature database, VIP database and Wanfang database were searched by computer, and the included references were tracked and searched in the way of “snowball”. The retrieval time limit was from the establishment of the database to December 2021. A combination of subject headings and free words were used to search, and the search terms mainly included: aspirin, acetylsalicylic acid, stroke, cerebrovascular accident, cerebrovascular disease, cardio-cerebrovascular disease, cardiovascular disease, primary prevention, randomized controlled trial. Taking PubMed as an example, its specific retrieval strategy is shown in Figure 1.

Inclusion criteria: 1) Study design. The content involves a randomized controlled trial (RCT) of aspirin in the primary prevention of stroke; 2) Research subjects. Participants were ≥18 years old; 3) Interventions included in the study. Intervention group: aspirin was routinely taken every day or every other day; control group: no or placebo at the same dose every day or every other day; 4) Outcome indicators. Primary outcome measures: stroke, ischemic stroke,

hemorrhagic stroke, transient ischemic attack (TIA); Secondary outcome measures: myocardial infarction, gastrointestinal bleeding, mortality.

Exclusion criteria: 1) Taking aspirin, having a history of gastrointestinal disease, stroke or confirmed myocardial infarction; 2) Studies for which the full text was not available or incomplete data; 3) The study type was review, case report, qualitative study; 4) No relevant outcome indicators or could not be calculated; 5) Studies published in languages other than Chinese and English.

Data were extracted from all the included literatures, including: inclusion study, year, country, study type, study object, dose, follow-up years, average age, sample size, intervention measures of intervention group and control group, outcome indicators, etc.

The Cochrane Manual version 5.1.0 bias risk Assessment tool was used [7], which was completed by 3 evidence-based trained researchers, 2 of whom performed independently, and the third researcher decided in case of disagreement. The evaluation contents include: random sequence generation, allocation concealment, blinding of research subjects and implementers, blinding of outcome measurers, complete outcome data reporting, selective research reporting results, and other aspects of bias. Those who fully meet the above items indicate that various risk biases are low, and their quality grade is A; those who meet some of the items indicate that various risk biases are medium, and their quality grades are B; those who do not meet the above items at all indicate that various risk biases are high, and their quality is low. A grade of C is excluded.

Meta-analysis was performed using RevMan 5.3. The outcome indicators of this study were all dichotomous data, so the relative risk (RR) value was calculated, and the 95% confidence interval (CI) was calculated for the effect analysis. The c² test was used to determine whether there was heterogeneity among the studies, and the I² value was used to judge the size of the heterogeneity. If P ≥ 0.1 and I² ≤ 50%, a fixed-effect model was used for Meta-analysis; If P < 0.1 and I² > 50%, the source of heterogeneity should be analyzed. If there was no obvious clinical heterogeneity, a random-effects model was used for Meta-analysis. If the heterogeneity is too large and the source cannot be judged, descriptive analysis of the results is carried out.

A total of 3791 literatures, 887 in Chinese and 2904 in English were obtained through preliminary search, and 3 related literatures were manually searched, and a total of 3794 relevant literatures were obtained. By reading the titles and abstracts, excluding duplicate publications, reviews, and literatures inconsistent with the theme, and further reading the full text, 19 literatures were finally included, including 1 Chinese literature and 18 English literatures, including a total of 220,636 research subjects. The literature retrieval and screening process is shown in Figure 2.

The basic characteristics of the included studies are shown in Table 1. All included literatures were strictly assessed for quality according to the Cochrane

Note: ① Cerebral apoplexy; ② Ischemic stroke; ③ Hemorrhagic stroke; ④ TIA; ⑤ Myocardial infarction; ⑥ Gastrointestinal bleeding; ⑦ Mortality.

Handbook version 5.1.0 tool for assessing risk of bias [7] and their quality was graded. The results showed that of the 19 included papers [8] - [26], 6 of them [11] [20] [21] [24] [25] [26] were graded A, and 13 [8] [9] [10] [12] - [19] [22] [23] were rated B.

The results of this study showed that long-term preventive use of aspirin can effectively reduce the incidence of stroke, ischemic stroke and TIA. There may be two reasons for this result: 1) Aspirin is an anticoagulant widely used in clinical practice. Its mechanism of action is the irreversible binding of aspirin and cyclooxygenase-1 to inhibit the synthesis of thrombin A₂, thus blocking the platelet aggregation mediated by thrombin A₂, preventing and controlling thrombosis, and reducing the risk of ischemic cardiovascular and cerebrovascular diseases. Ischemic stroke accounts for about 68% - 80% of all strokes and is one of the common types of cerebrovascular diseases. Therefore, long-term prophylactic use of aspirin has become a major reason for the benefits of this group [27] [28] [29]. 2) Among the 19 included studies, except for Peto et al. [8] and Ridker et al. [17], who were healthy male doctors and healthy women respectively, most of the other included subjects were people with cardiovascular and cerebrovascular diseases or who were at high risk of cardiovascular and cerebrovascular diseases, so they could benefit from the preventive use of aspirin. In addition, Guirguis-Blak et al. [29] once again verified that people with high cardiovascular risk may benefit from prophylactic aspirin use. Further research is needed to determine whether healthy people at risk of cardiovascular disease can benefit from long-term prophylactic use of aspirin.

Results of this meta-analysis showed that long-term prophylactic use of aspirin increased the incidence of hemorrhagic stroke compared with the control group. Bleeding is a major risk and a major limitation of aspirin’s widespread clinical use. In addition, age is not only a major risk factor for stroke, but also an important factor to be considered in the process of drug application. With the increase of age, the organ function and immune function of the body show a trend of gradual decline, so the tolerance of the elderly population to drugs is worse, more prone to adverse reactions. In this study, the average age of the included subjects > 50 years old, which may be an important reason for the increased incidence of hemorrhagic stroke. Therefore, in the latest guidelines, the applicable age of aspirin is reduced from the original 45 - 79 years old to 50 - 69 years old, in order to improve the safety of clinical medication [30].

1) Aspirin can significantly reduce the incidence of myocardial infarction. Some domestic and foreign studies [31] [32] [33] have confirmed the safety and effectiveness of aspirin for primary cardiovascular prevention, and the results showed that aspirin can significantly reduce the incidence of myocardial infarction, which is consistent with the results of this study. 2) Increased incidence of gastrointestinal bleeding. Aspirin is a class of nonsteroidal drugs, and the main risk of long-term use is bleeding, the most common of which is gastrointestinal bleeding. In this study, 12 studies [9] [11] [12] [14] [17] [19] [20] [21] [23] [24] [25] [26] reported the impact on the incidence of gastrointestinal bleeding, and some studies [9] [11] [14] [17] [21] [23] [26] reported bleeding in other parts of the nose, retina, urinary system and so on. There is high heterogeneity among studies. On the one hand, due to different definitions of massive bleeding in the tests and differences in assessment tools, measurement results are inconsistent. On the other hand, the dose, dosage form and duration of aspirin were also a cause of heterogeneity. 3) No effect on mortality. The results of this study showed that long-term prophylactic aspirin use did not reduce the risk of death, which was different from the study of Li Wei et al. [4]. Since the follow-up years of the studies included in this meta-analysis ranged from 0.5 years to 10.1 years, the results may be affected to some extent, suggesting that researchers should continue to pay attention to the long-term effects of aspirin.

This study only searched Chinese and English databases, and there may be incomplete literature inclusion. The research subjects included in the literature are quite different in terms of regions, races, cultural backgrounds, genetic factors, and economic levels, which may lead to heterogeneity among studies. Since the follow-up periods of the included studies vary (the shortest is 0.5 years, the longest is 10.1 years), the time span is long, the number of participants is large, and there are many confounding factors, which may have a certain impact on the results of the study. Only 1 domestic literature was included in this study [18], and the sample size was relatively small. Due to differences in regions, races, and physiques, the safety and efficacy of aspirin in the domestic population needs to be further verified.

This suggests that it is necessary for Chinese researchers to carry out high-quality, multi-center clinical randomized controlled trials and extend the follow-up time to further verify the safety and efficacy of long-term preventive use of aspirin in the population.