Antithrombotic agents inhibit thrombosis by blocking platelet activation or blood coagulation, thereby averting a variety of cardiovascular diseases, but this process often exposes patients to an increased risk of hemorrhage(3). In this study, we compared the risk of bleeding induced by different antithrombotic agents in a wide cohort of patients receiving antithrombotic therapy, by using a large and comprehensive spontaneous reporting system database. Furthermore, we performed a TTO analysis of hemorrhage related to antithrombotic therapy to clarify how the risk of bleeding changes over time during administration. In summary, our study provided relevant insights to support the pharmacological management of patients in clinical practice.
In the signal detection, Crude LMWHs-hemorrhagic ADR was found as a positive signal, implying a stronger causal relationship between this medication and hemorrhagic ADR compared with the other 20 medications. In other words, our results showed that the patients taking Crude LMWHs were more likely to experience a hemorrhagic ADR, which may be attributed to irrational use due to improper management. LMWHs are obtained by cleaving heparin with different chemical or enzymatic methods and have an average molecular weight of 12–16,000 Da(15, 16). Because of the better bioavailability and pharmacokinetics with more predictable dose responses, they have been widely used for preventing blood clots, treating deep vein thrombosis, pulmonary embolism, and myocardial infarction since the 1980s(17–19). LMWHs were classified into many types depending on their molecular weight and antithrombotic activity. Each LMWH is a unique agent with its own biological, chemical, biophysical, biochemical, pharmacodynamic, and pharmacokinetic characteristics and has been studied independently in preclinical trials or clinical trials(20, 21). Therefore, different LMWHs were approved separately and had different indications in Europe and the United States(22). However, in China, LMWHs can be approved under the generic name of low molecular weight heparins without distinguishing between specific molecular weight and antithrombotic activity, which is the Crude LMWHs in this study. According to the National Medical Products Administration (NMPA), there were more than thirty varieties of Crude LMWHs registered in China. Studies have confirmed that treatment interchange between different LMWHs is considered inappropriate and potentially harmful to patients(23, 24). Since many Crude LMWHs marketed in China were not categorized in detail, they are susceptible to therapeutic interchange and off-label use, thus posing medical safety risks, which is a possible reason why Crude LMWHs-hemorrhagic ADR is a positive signal in our study. A survey of a Chinese hospital found that the proportion of irrational use of Crude LMWHs was 80.4 percent, which further supports our point(25). We recommend that the NMPA strengthen the regulation of LMWHs and use them more carefully in clinical practice.
To examine the onset profiles of antithrombotic agents induced hemorrhagic ADRs, time-to-onset analyses were performed. Our results showed that the hemorrhagic ADRs for most antithrombotic agents were random failure profiles, which means that the hazard was constantly occurring over time. In particular, the hazard of hemorrhagic ADRs decreased over time for warfarin and clopidogrel and increased alteplase, nadroparin, and dipyridamole. Although there are few published results of TTO analysis for antithrombotic agents, we could evaluate the failure type from the cumulative incidence curve of bleeding events in clinical trials and real-world research. When the slope of the curve decreases over time, the event might be the early type, when it becomes larger with time, it might be the wear-out type, and when it is essentially constant, it might be the random type.
As in our study, several clinical trials confirmed the decreasing risk of warfarin-induced bleeding events over time(26–28). Similar results were also obtained by some real-world research, which demonstrates the importance of coagulation monitoring early in the administration(29, 30). Clopidogrel is an inactive prodrug requiring metabolized and activated by cytochrome P450 hepatic enzymes, of which the active metabolite exerts platelet inhibition activity by binding irreversibly to the platelet adenosine diphosphate P2Y12 receptor(31, 32). Although some research has demonstrated that the bleeding effects related to dual antiplatelet therapy (DAPT) with clopidogrel and aspirin are constant over time, the failure type of clopidogrel monotherapy-induced hemorrhage is still unclear(11, 33). In a multicenter, double-blind, randomized trial, the investigators compared clopidogrel and ticagrelor for the prevention of cardiovascular events in 18,624 patients with the acute coronary syndrome. In line with our study, their analysis of the cumulative incidence of major bleeding indicated that the hazard of hemorrhage decreased over time for clopidogrel(34). Another recent study evaluated the benefit of clopidogrel monotherapy after one 1-month DAPT compared with 12-month DAPT in acute coronary syndrome and chronic coronary syndrome patients. The results of safety analysis showed that the risk of bleeding with clopidogrel monotherapy after one 1-month DAPT was lower in the later stages of treatment, which also support our findings(35). However, the failure type of hemorrhage induced by clopidogrel monotherapy remains understudied and further research is warranted.
IV thrombolysis with alteplase (recombinant tissue plasminogen activator, rtPA) improves the outcome of patients with ischemic stroke treated within a 4.5-hour time window(36). However, symptomatic intracranial hemorrhage occurred in 4.7–11.4% of patients treated with alteplase, and major systemic bleeding events occurred in an additional 1.6–3.6%(37, 38).In our analysis, the median TTO of hemorrhage after initiation of alteplase therapy was one day and the majority occurred within the first 24 hours, which is consistent with the earlier findings(39, 40). The mechanism of alteplase thrombolysis is the conversion of plasminogen to plasmin in the presence of fibrin. Plasmin then breaks fibrin into fibrin degradation products and its fibrinolytic activity also induces depletion of coagulation factors, which finally decreases the fibrinogen level and prolongs both prothrombin time and activated partial thromboplastin time (41). The process also leads to hemorrhagic ADRs and continued fibrinogen reduction and coagulation factors depletion under prolonged administration may explain the increased risk of hemorrhagic ADRs with alteplase over time. With an average molecular weight of 4300 Da and an Anti-Xa/IIa ratio of 3.3, nadroparin is derived from unfractionated heparin by deaminating cleavage with nitrous acid and is widely used for prophylaxis of thromboembolic disorders and general orthopedic surgery, treatment of deep venous thrombosis, and prevention of clotting during hemodialysis(19). Our results indicated that the hazard of nadroparin-related hemorrhage increases over time and the median time of hemorrhage after therapy is 4 days. These results are in line with a cohort study, which found that the hemorrhage complications occurred mainly in the later stages of drug administration for patients using nadroparin after pancreatic surgery(42). As a platelet inhibitor, dipyridamole works by multiple biochemical routes, including blocking adenosine reuptake and inhibiting cyclic adenosine monophosphate, and enhancing prostaglandin I2-mediated activity(43). The Second European Stroke Prevention Study (ESPS-2) showed that the risk of bleeding persisted throughout treatment exposure for the acetylsalicylic acid group and the combination of acetylsalicylic acid and modified-release dipyridamole group. Moreover, this study also found that the headache and gastrointestinal disorders induced by dipyridamole generally occurred early in the trial(44). Different from the findings of ESPS-2, our results indicated that the hazard of hemorrhagic ADRs increased over time for dipyridamole. It’s difficult to explain this difference between the two studies within the context of missing further research.
Previous studies have shown that hemorrhage following thrombolytic therapy is associated with poor clinical outcomes(45–47). We also analyzed the effect of hemorrhagic ADRs on the severity of ADRs, and the results showed there was no difference in the severity of ADRs between hemorrhagic and non-hemorrhagic cases after balancing the effects of confounding factors, which may be related to the underreporting of slight non-bleeding ADRs. In our study, most patients in both case and non-case groups improved or cured after treatment and intervention. But all deaths experienced hemorrhagic ADRs, indicating that these adverse effects can be fatal for patients.
This study has potential limitations. Firstly, due to the restrictions of the spontaneous reporting system, we only analyzed the bleeding risk of individual drugs without multidrug combination therapy. Secondly, the inevitable under-reporting and false-reporting in SRS may introduce bias into the results. Finally, since the time-to-onset data in our study is measured in days, we could not analyze the change in hemorrhage risk on more precise time units, which may be inappropriate for some medications. However, our research still was important. The analysis of individual antithrombotic agents can extend the understanding of the characteristics of hemorrhage. More importantly, our research helps clarify how the risk of bleeding changes over time during administration, which could lead to the development of treatment strategies aimed at reducing hemorrhage risk and promote the rational use of antithrombotic agents.