Early neurological deterioration (END) and recurrent ischemic stroke (RIS) are the most common conditions after acute ischemic stroke (AIS). END, defined as deterioration on the neurological scale (1), usually occurs within 24–72 hours from symptom onset and is often associated with a poor prognosis (2–4). Although antiplatelet agents (especially aspirin) are the most widely used and recommended medication in the early management of AIS, clinical neurologists often complain of their limited effect in halting symptom progression. END and RIS are still commonly observed after treatment with antiplatelet agents, meaning that not all patients benefit most from this therapy.

Despite being commonly used in clinical practice, the current guidelines do not recommend anticoagulant use in AIS (5). This conclusion was reached mainly based on two meta-analyses, which stated that anticoagulants were not associated with net short- or long-term benefits and had an increased bleeding risk (6, 7). However, we realized that the estimated effect may be subjective, as the results were primarily driven by one RCT, which accounted for approximately 80% of cases in all outcome analyses, and unfractionated heparin (UFH) was administered subcutaneously instead of routine intravenous injection; moreover, the incidence of bleeding due to anticoagulants was overestimated as placebo-controlled trials were involved. A meta-analysis comparing low-molecular-weight heparin (LMWH) with standard treatment aspirin has not been performed since 2002 (8). Recent randomized controlled trials (RCTs) have shown that direct oral anticoagulants (DOACs) could reduce ischemic lesion growth and improve recanalization with a similar risk of hemorrhagic transformation compared to that of aspirin (9, 10), indicating that anticoagulants may play a certain role in AIS. Heparin, a traditional and reliable anticoagulant, has also been found to have anti-inflammatory properties. Neuroinflammation is known to play an essential role in the pathophysiology of ischemic stroke, and these promising results prompted us to revisit the effects of LMWH in AIS as it has not been evaluated for many years.

In this meta-analysis, we aimed to provide more accurate estimates of the LMWH in the early management of AIS compared to aspirin. We paid special attention to END, as it has not been closely examined by other systematic reviews before.

We conducted this review in accordance to the PRISMA statements (64). See online Supplementary Appendix Table 3 for the completed PRISMA checklist. Our review was registered with the International Prospective Register of Systematic Reviews (PROSPERO; http://www.crd.york.ac.uk/PROSPERO), registration number: CRD42020185446.

The efficacy and safety of aspirin in AIS have been validated in large trials (26, 27); however, some patients still develop END or RIS after receiving antiplatelet therapy. Therefore, clinical neurologists consider anticoagulants to be a feasible therapy. Pharmacological studies have shown that heparin can reduce blood viscosity, promote fibrinolysis, prevent the growth of early ischemic lesions, and have better anti-inflammatory effects than aspirin in AIS (28–31). Although previous meta-analyses have provided no recommendation on anticoagulants in AIS, the conclusion may be subject to imprecision and no evaluation based on stroke etiology has been performed.

In our meta-analysis, the results showed no significant differences in short- or long-term efficacy between LMWH and aspirin among all patients, which is consistent with previous studies. However, in patients with non-cardioembolic stroke, a significant difference in short-term efficacy between LMWH and aspirin was observed. LMWH was associated with a significant reduction in END (RR: 0.44, 95% CI: 0.35–0.56; absolute risk reduction [ARR]: 9%, 95% CI: 7.94–10.06%) and RIS (RR: 0.50, 95% CI: 0.27–0.91; ARR: 1.19%, 95% CI: 0.79–1.59%). To the best of our knowledge, this is the first meta-analysis to evaluate the estimated effect of heparin in the early management of AIS with consideration of stroke etiology. The positive results were mainly driven by three studies (21–23), in which patients shared similar characteristics (non-cardioembolic stroke and most were diagnosed with large-artery stenosis), and LMWH was administered twice daily. As an important factor of poor prognosis (32), END has been clinically observed to be relevant to atherosclerotic diseases such as atherosclerosis (AT) and small artery disease (SAD) (33, 34). This explains the higher incidence of END in these three trials since all strokes were caused by large-artery occlusive disease (LAOD) in FISS-tris, and patients were diagnosed with either AT or SAD in Yi 2014/2015. Few studies have revealed the intrinsic link between END and stroke etiology. The most common reason for END could be attributed to ischemic lesion growth (35), and heparin might prevent END by reducing ischemic extension and salvage the ischemic penumbra (36). Complete or partial lysis of intraluminal thrombus has been observed in patients who received intravenous heparin (37), indicating that anticoagulants are capable of promoting innate thrombolysis and improving blood flow in the ischemic area. However, this effect could be limited in cardioembolic stroke, which is caused by emboli originating from an organized heparin-unresponsive thrombus within the heart (20). It is noteworthy that although FISS-tris showed no reduction in RIS in the LMWH group, this is possibly due to chance, as the incidence of RIS in patients receiving aspirin was significantly lower in FISS-tris than in other larger RCTs (0% vs. 3%) (21, 26).

The mRS is a simplified overall assessment of disability, with scores ranging from 0 (no symptoms at all) to 6 (death), with 5 indicating severe disability (complete dependency). A functional outcome of independence as reflected by an mRS score of 0–2 was similar between LMWH- and aspirin-treated patients, which has been interpreted as heparin and aspirin sharing equal efficacy in improving patients’ disability (6). However, we realized that LMWH might enable more patients to achieve full recovery than aspirin, as one of our included trials (FISS-tris) suggested that LMWH was significantly associated with a higher likelihood of an mRS of 0–1 in patients with LAOD at 6 months. An mRS of 0–1 was usually considered as a favorable outcome in previous AIS thrombolysis trials (38, 39), and differences in the dichotomization of mRS strongly influenced the interpretation of the results; for example, although the difference in the number of patients with an mRS score of 0–2 was not significant, there was a significant difference in the number of patients with an mRS of 0–1 between the alteplase and placebo groups in ECASS III, supporting the use of thrombolysis therapy in AIS after 3–4.5 hours (recommended by current guideline) (40). Similarly, the significant association between an mRS score of 0–1 and LMWH indicates that LMWH is effective in improving 6-month independence in AIS patients with LAOD. According to the discussion above, patients with LAOD could benefit from LMWH in the short term, and the long-term efficacy of LMWH in patients with AT has also been proven by TOAST (1) (although LMWH was administered differently), strongly supporting that LMWH is capable of improving functional outcomes in stroke patients with large-artery stenosis.

We did not observe differences in the number of deaths at any time point, indicating that LMWH is relatively safe in the early management of AIS relative to aspirin. Although the incidence of extracranial hemorrhage increased significantly in patients who received LMWH, as most mild bleeding is reversible, major or severe bleeding is the main clinical concern. There was no significant difference in sICH or major extracranial hemorrhage between the LMWH and aspirin groups. Overall, in every 1000 patients who received LMWH rather than aspirin, 55 ENDs and nine RISs would be prevented; at the same time, 19 cases of extracranial hemorrhage and three cases of SICH would occur. As all complications mentioned above were “symptomatic”, we concluded that the net benefit of LMWH was observable. Furthermore, the benefits of LMWH were mainly seen in non-cardioembolic patients, and the trial (HAEST) that included only patients with cardioembolism found no difference either in RIS or independence (20). According to the consensus in the ESC Guidelines, patients with atrial fibrillation and moderate stroke should initiate or continue anticoagulation 6 days after stroke onset (41). Thus, whether the administration of LMWH in cardioembolic patients should be postponed requires further investigation.

There has been increasing evidence over the past decade that peripheral innate and adaptive immune cells, such as neutrophils and natural killer cells, may play an important role in the pathophysiology of ischemic stroke. The influx of neutrophils may cause increased vascular resistance due to increased blood viscosity and cellular obstruction, leading to cerebral infarction associated with collateral failure (42–44). A recent study showed that toll-like receptors (TLRs) are involved in the activation of inflammatory responses during the acute phase after ischemic stroke (45, 46). TLR-2 and -4 persist for at least 7 days after reperfusion, promoting exacerbation of acute inflammation and impeding neurological recovery (45) Histones released by dead cells can induce thrombin generation by activating platelets via TLR-2 and TLR-4 (47), which may lead to immunothrombosis after stroke (48). Heparin could lower TLR-4 protein expression and prevent histone interactions with platelets, which has been shown to reduce the risk of immunothrombosis (49). At the same time, secondary neuroinflammation in AIS leads to compromised integrity of the blood-brain barrier (BBB), allowing water molecules and blood components to enter the extracellular space of the brain, resulting in serious clinical consequences (50–53). The extent of BBB is associated with stroke severity and progression (54, 55). Thus, preventing BBB disruption is considered a potential therapeutic strategy for improving END. Glycocalyx, a polysaccharide protein complex that covers the surface of vascular endothelial cells, is abundantly expressed on the endothelial cells of the BBB and has been shown to regulate BBB permeability (51). It plays a key role in the inflammatory process by interrupting the cycle of endothelial dysfunction and inflammation (56). After stroke, endothelial cells are exposed to neuroinflammation and elicit degradation of the glycocalyx (57–59). In atherosclerosis, one of the main pathogeneses of stroke, glycocalyx degradation promotes lipid deposition in the vessel walls and reduces endothelial cell expression of endothelial nitric oxide synthase (eNOS), causing loss of vasodilation (60). LWMH could suppress glycocalyx shedding in a dose-dependent manner as an inhibitor of heparanase activity (61, 62), thereby improving BBB leakage, brain edema, decreasing the expression of inflammatory factors, and improving neurologic outcomes (63).

The strengths of our study were that the meta-analysis was conducted and reported according to the PRISMA methodology; the risk of bias of each included study was carefully screened using the Cochrane Collaboration’s tool, finding that all trials were adequately randomized and accessor-blinded; and we analyzed the heterogeneity between trials based on stroke etiology, which provides information on how to achieve precision treatment. The limitations were that we did not obtain original data for each trial; hence, no evaluation based on the individual participant level was performed, and two trials might be not sufficiently normative as they were not registered on NCT or ISRCTN (the methodology of two trials was considered appropriate) (22, 23). In addition, our positive results are mainly based on three Asian studies; as racial differences in stroke etiology and pathogenesis are essential factors underlying drug selection, the risk of bias cannot be neglected. Furthermore, only one RCT study used the mRS score of 0-1 to evaluate the effect of LMWH on cerebral ischemic injury, which could not provide the enough evidence to draw the conclusion. However, mRS score of 0-1 is an important measurement in clinical practice, we emphasized this score in our article. It is promising that LMWH benefits certain patients with non-cardioembolic stroke. We believe that further basic and clinical studies should be conducted on the application of LMWH to elucidate the population most likely to benefit and reveal the mechanisms of LMWH in the treatment of AIS.

In patients with acute non-cardioembolic ischemic stroke, especially those with large-artery stenosis, LMWH significantly reduced the incidence of END and RIS, and improved independence (mRS 0–1) at 6 months relative to those with aspirin. We found no significant difference between LMWH and aspirin in improving patients’ long-term functional outcomes measured as reflected by an mRS score of 0–2. LMWH treatments was related to increased risk of extracranial hemorrhage among all patients; however, the difference in major extracranial hemorrhage and sICH was not significant. LMWH was associated with a net benefit in non-cardioembolic stroke. The current situation is that anticoagulation therapy has not been recommended for non-cardioembolic ischemic stroke; and the increased risk for bleeding, instead of not reducing the risk of neurological deterioration and recurrent ischemic stroke in early acute ischemic stroke, is the main reason. Therefore, choosing the appropriate patients and paying attention to the start time and duration of treatment are very important. Nowadays, lots of physician still use the individualized LMWH in the treatment of noncardiogenic acute ischemic stroke. Hope our study can trigger more experts to re-recognize.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding authors.

ZL, ZW, and HX designed the systematic review. ZW, HX, and MT performed the literature search the literature search, reviewed all publications, extracted the information and data from the included studies, performed the data analysis, and produced the figures and tables. ZW, HX, and MT wrote the manuscript. ZL and ZZ revised the manuscript. All authors have read and agreed to the published version of the manuscript.

This study was supported by grants from National Natural Science Foundation of China (No. 52073310) and Elite Medical Professionals project of China-Japan Friendship Hospital (NO. ZRJY2021-BJ03) and Technical innovation project in major clinical fields (CX2019LC103).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.