Endovascular therapy (EVT) is the standard treatment for acute ischemic stroke caused by large vessel occlusion (AIS-LVO) (1). The efficacy of EVT is influenced by multiple factors, including the medical capabilities of hospitals and their staff, such as the use of EVT devices, the EVT skills of neuro-interventionalists, and the establishment of emergency treatment green channels for AIS. Additionally, other factors such as traffic conditions, public education, and awareness of AIS also play important roles, which may vary significantly between developed and developing countries. The RESILIENT randomized trial in Brazil demonstrated that patients with AIS-LVO in a developing country can still benefit from EVT (2).

Differences in stroke care are evident not only between developed and developing countries but also between county and municipal-level hospitals. Given China’s vast geographical expanse, there is significant variation in the construction of stroke centers, stroke burden, and the quality of stroke care across different regions (3). To improve the quality of stroke care, the guidelines for stroke center construction in China, issued in 2015, recommended establishing two levels of stroke centers: primary stroke centers (PSCs), typically county-level hospitals, and comprehensive stroke centers (CSCs), usually municipal or provincial hospitals (4). The efficacy of reperfusion treatments, including intravenous thrombolysis (IVT) and endovascular thrombectomy (EVT) for AIS is time depend (5). Therefore, reperfusion therapy should be performed as soon as possible. The “2019 Stroke Report in China” showed that EVT was increasingly performed in PSCs. However, differences in medical care capabilities, emergency treatment green channels for AIS, and traffic conditions may exist between PSCs and CSCs. We hypothesize that the AIS workflow and medical care capabilities of hospitals and their staff are suboptimal in PSCs. To explore these differences, we conducted a prospective real-world registry study, named the “Endovascular Treatment for Acute Ischemic Stroke in Chinese Municipal and County Hospitals (the ETERNITY registry),” from November 2021 to December 2022. We hope the findings will provide valuable information for improving stroke care quality in China.

To our best knowledge, the ETERNITY registry is the first real-world nationwide database with a broad geographic distribution of endovascular capable centers with relative few restrictions for patient selection. It provides an accurate reflection of the current status of EVT in clinical practice both in CSCs and PSCs in China. The registry specifically aims to explore differences in AIS workflow, patients’ characteristic, preprocedural brain image, EVT technique, clinical and radiological outcome between CSCs and PCSs.

From ETERNITY registry, we observed significant differences in patient characteristics between CSCs and PSCs. First, regarding TOAST classification, CE stroke was more common in PSCs, while LAA was more prevalent in CSCs. Second, OPT was shorter in PSCs, with a higher proportion of patients presenting within 4.5 h. Third, rate of IVT, IVT proportion was higher in PSCs. CE strokes typically present with a sudden onset of neurological deficits that are maximal at onset, whereas LAA strokes may have a less defined, waxing and waning course (15). Due to the characteristic onset of CE, patients with CE etiology may be referred to nearby hospitals, which are often PSCs. Combined with better traffic conditions, this may contribute to shorter OPTs and a higher proportion of patients presenting to PSCs within 4.5 h, potentially eligible for IVT.

IVT is a cornerstone treatment for AIS treatment, significantly improving outcomes when administered within 4.5 h of symptom onset (16). However, its utilization in China remains suboptimal. Only 10–20% of AIS patients in China arrive at a hospital within 3 h, and fewer than 3% receive IVT (17). This underuse can be attributed to multiple factors, including delayed symptom recognition, limited awareness among healthcare providers, inefficient in-hospital AIS workflow, and inadequate medical resources. In the ETERNITY registry, approximately 60% of patients arrived at the hospital within 4.5 h, and over 50% received IVT. This higher rate of IVT use in PSCs compared to previous studies (4, 18, 19) likely reflects improvements in stroke care due to the construction of stroke centers in China. The higher proportion of IVT in PSCs may also be influenced by the characteristics of cardioembolic (CE) stroke. CE strokes often present with sudden, severe neurological deficits, prompting quicker hospital presentation. Additionally, better traffic conditions in urban areas where PSCs are often located facilitate faster hospital arrival. The underuse of IVT in eligible population may also be influenced by the perception that direct mechanical thrombectomy (MT) without prior IVT is non-inferior to bridging therapy, as suggested by the DIRECT MT trial, although with a broad non-inferiority margin (20). Moreover, the IRIS meta-analysis showed no superiority of thrombolysis in later time windows (after 2.5 h), further complicating the decision-making process for IVT administration. In summary, while IVT is a highly effective treatment for AIS, its use in China is still limited. The ETERNITY registry highlights improvements in IVT utilization in PSCs, likely due to enhanced stroke center infrastructure and the characteristics of CE stroke. However, continued efforts are needed to optimize IVT delivery and improve stroke outcomes nationwide.

Data from the registry also highlight difference in AIS workflow between CSCs and PSCs. First, regarding brain imaging modalities, angiography and perfusion imaging were less frequently performed in PSCs. Second, DPT was longer in PSCs. Stroke imaging is crucial in AIS management (21). The role of brain imaging in AIS management includes identifying the infarct core (IC), tissue at risk, and arterial lesion (21). Non-contrast CT is an easily accessible and time-saving tool used in AIS management and can be performed at any time. However, it is suboptimal for identifying large vessel occlusions (LVOs) and is less sensitive in detecting extremely early ischemic changes (22). Clinician using NIHSS score ≥ 10 to identify LVO only have 48% sensitivity, potentially missing half of LVO cases (23). Therefore, invasive arterial imaging, such as CTA and MRA, is necessary in AIS management. CTA is the most accurate and efficient non-invasive method for confirming or excluding LVOs and is recommended to be performed as quickly as possible in patients with suspected LVOs (AHA Class I, Level of Evidence B) (23). Perfusion imaging is another important modality for detecting the IC and tissue at risk (24). CTA images analyzed on a PACS workstation with MIP and MPR capabilities may miss about 20% of LVOs (mostly M2 segment) during initial CTA evaluation (25, 26). Perfusion imaging with automated processing software can assist clinicians in easily identifying lesions (27). Since a complete imaging package with non-contrast CT, CTA, and CTP can be obtained in less than 6 min, published endovascular stroke intervention trials have shown that performing CTA imaging does not delay IV tPA infusion (22). Therefore, one-stop multimodal CTA and CTP should be recommended to be performed as soon as possible for better AIS management. However, less than half of AIS-LVO patients in PSCs underwent CTA/CTP. The reasons may include limited awareness among staff in PSCs, medical resource shortages, and in-hospital workflow inefficiencies. Despite the potential time-saving benefits from no advanced brain imaging, DPT was longer in PSCs. The reasons may include the lack of an unobstructed AIS in-hospital workflow, shortage of neuro-interventionists, and absence of a standardized workflow in the catheterization suite. Therefore, the AIS in-hospital workflow, including the emergency department, radiology department, and catheterization suite in PSCs, should be optimized according to the criteria proposed by the CSPPC to minimize door-to-CT, door-to-needle, and DPT time (6). Additionally, training of qualified personnel should be emphasized.

According to the registry, the successful reperfusion rate was lower in PSCs. Our registry showed that more than half of the cases were classified as CE according to the TOAST criteria. Previous studies had demonstrated that compared with no LAA, lower intraprocedural re-occlusion, higher successful reperfusion, shorter PRT (28) in CE stroke. However, data from registry did not show these promising results, on the contrary, low successful reperfusion rate was observed in PSCs. Many factors are associated with successful reperfusion, including occlusion site characteristics; EVT device use, including the use of a stent retriever or balloon guiding catheter; aspiration catheter use; EVT procedural technique (29–37); and neuro-interventionist skill. Effort should be made to improve neuro-interventionists’ capacity by training, enhance EVT success rate via virtual supervision by remote streaming support (38) or utilize remote robotic EVT technology (39). Additionally, it is crucial to ensure that a sufficient range of EVT devices is available to allow the use of appropriate devices according to lesion characteristics.

Despite the longer DPT and lower successful reperfusion rate in the PSCs, the rate of good functional outcomes was similar between PSCs and CSCs. Good functional outcome is associated with multiple factors, including the reperfusion status (13, 40), the onset-to-reperfusion time (41), procedural complications such as vessel perforation (42), and distal embolism (43). In PSCs, the short OPT may mitigate the impact of the lower successful reperfusion rate on overall clinical outcomes.

We also found that nearly one-quarter of thrombectomy cases died in PSCs, with mortality being higher in PSCs. The high mortality rate in PSCs may be attributed to the several factors: first, the high incidence of CE, which is associated with more severe complications and a higher incidence of adverse outcomes (44). Second, the low successful reperfusion rate, which is associated with lager infarct area. Third, post-procedure management. The post-procedure management of AIS, especially in patients with large infarcts is complex and includes blood pressure control, antiplatelet agent administration, brain edema control, and sICH management (42). Fourth, lack of stroke unit. Stroke unit care were associated with 15% lower odds of complications and 15% lower odds of mortality during hospitalization among patients who received thrombolytic therapy (6). Accordingly, to narrow the mortality gap between PSCs and CSCs, we must implement uniform, evidence-based post-reperfusion management protocols and guarantee equitable, round-the-clock access to dedicated stroke-unit beds.

Our nationwide real-world registry highlights several differences in EVT for AIS-LVO patients between PSCs and CSCs in China. Given that the clinical outcomes of AIS-LVO are closely associated with timely intervention and successful EVT, enhancing EVT skills through targeted training programs and standardizing stroke center construction are essential. With further improvements in in-hospital workflow and physician training at PSCs, we anticipate that clinical outcomes for AIS-LVO patients treated at PSCs will improve in the future.