1 INTRODUCTION

Stroke is the leading cause of adult death and disability in China (Wang, Abajobir, et al., 2017). With the continuous promotion of stroke care and prevention, the standardized mortality rate shows a downward trend in China (Wang, Hu, Sang, Luo, & Yu, 2017), while there is a significant upward trend for disability-adjusted life years (DALY; Report on Stroke Prevention & Treatment in China, 2018) Early reperfusion therapy, including intravenous recombinant tissue plasminogen activator (rt-PA) thrombolysis within 4.5 hr and mechanical thrombectomy (MT) within six hours of symptom onset, is the most effective evidence-based treatment for acute ischemic stroke (AIS; Powers et al., 2018). The outcomes for thrombolysis guided by perfusion imaging up to nine hours (Ma et al., 2019) and MT for selected patients up to 24 hr (Albers et al., 2018; Nogueira et al., 2018) after the onset of stroke are extremely time-dependent (Wardlaw, Murray, Berge, & del Zoppo, 2014). Previous research indicates that for each minute earlier that treatment is initiated, there are up to 12.5 fewer disability-adjusted days (Meretoja et al., 2014; Meretoja, Keshtkaran, Tatlisumak, Donnan, & Churilov, 2017; Schlemm, 2018). However, in China, only about 10%–20% of patients arrive at hospitals within three hours, and less than 3% of patients with AIS receive rt-PA (Wu et al., 2019). A poor understanding and ability to identify stroke symptoms, prehospital delays, and nonstandardized emergency workflows are the top three factors for delays in timely treatment (Jiang et al., 2016; Pulvers & Watson, 2017). To improve rapid access to rt-PA, the first stroke emergency map (SEM) was launched in Shenzhen in November 2016, and as a result, the rate of patients receiving rt-PA thrombolysis and endovascular thrombectomy has increased significantly (Ye et al., 2019). Subsequently, the SEM was designed to increase the likelihood that treatment would occur in the golden first-aid window of “three one hour” (no more than one hour from onset to call for ambulance, no more than one hour from prehospital transfer, and no more than one hour from admission to thrombolysis) and was released in many cities by Chinese healthcare administrations.

Stroke incidence and mortality rates are highest in the northeast, the lowest in the south, and prominent in the middle of China (Report on Stroke Prevention & Treatment in China, 2018; Wang, Jiang, et al., 2017). The Shanxi Province is located in Loess Plateau, the north area of China, where economic development and medical technology are currently lagging behind the eastern and coastal regions. In this region, 274.74 new strokes and 56.9 stroke-related deaths occurred per 100,000 individuals in 2016. An epidemiological survey of stroke patients in the Shanxi province involved 714,893 individuals (aged ≥ 40 years) and revealed that Taiyuan had the highest proportion of stroke hospitalization (3.61%) (Luo, Jiang, Yu, Luo, & Nan, 2019). Based on that survey, a stroke emergency map was developed in July 2017 in Taiyuan, as part of a rapid regional stroke treatment system. This study evaluates the time delays and functional prognosis of patients who received rt-PA thrombolysis over the two years following the implementation of the SEM. The purpose of this approach is to facilitate optimization of the stroke-related health service system and provide more stroke patients with the best opportunity to arrive at hospitals within the optimal time frame to maximize patients’ prognosis.

2 METHODS

2.1 Patient selection

Data for 2,114 consecutive suspected stroke patients acquired from the SEM registry who arrived at 19 hospitals within a 4.5-hr time window were collected between August 2017 and July 2019. We excluded stroke mimic and TIA patients (n = 60) and patients with no information on baseline demographics and timeline if they received rt-PA thrombolysis (n = 218). After the exclusions, an initial cohort of 1836 patients was assembled. To evaluate the functional outcome at 90 days, we further excluded patients with no clinical outcome available at 90 days (n = 50). The final analysis sample consisted of 1,786 patients (Figure 1).

This retrospective study was approved by the Institutional Review Board of Taiyuan Central Hospital, Shanxi, China, in accordance with the Declaration of Helsinki.

3 RESULTS

A total of 1,786 patients from August 2017 to July 2019 met the inclusion criteria. Among them, 17 (1.0%) were in-hospital stroke patients, 641 (35.9%) were EMS-transported, and 1,308 (73.2%) were directly transferred to CSC. The total number of stroke patients within the time window increased from 43.9% in phase 1 to 56.1% in phase 2. The baseline characteristics of the patients are listed in Table 1. Four more new hospitals, including 62 (6.2%) patients, were added in phase 2, so we also compared the baseline characteristics within the initial 15 hospitals between phases 1 and 2 (Table S1). Compared with patients in phase 1, patients in phase 2 were more likely to be transferred to PSC and to use EMS transport to receive rt-PA or endovascular treatment.

As shown in Table 2, in patients received rt-PA, history of cerebral infarction, the mode of transportation, and the type of arrived hospital were significantly different between the two groups.

The median onset-to-door time decreased from 118 min (IQR: 75–160 min) in phase 1 to 105 min (IQR: 68–150 min; p = .002) in phase 2. Furthermore, 1,207 (67.6%) patients were treated with intravenous rt-PA, which is increase from 63.9% in phase 1 to 70.5% in phase 2. Among them, the onset-to-door time and onset-to-needle time were reduced by five minutes (100 [IQR: 62–135] vs. 105 [IQR: 70–145], p = .005) and nine minutes (158 [IQR: 124–197] vs. 167 [IQR: 132–214], p = .001), respectively. However, the door-to-needle time was not observed to be significantly shorter in phase 2 (p = .312) (Table 3).

We performed a further comparison of patients treated with rt-PA between the two phases according to the time stratification. The proportion of patients with an onset-to-door time ≤ 2 hr and ≤ 3.5 hr among phase 2 was higher than in phase 1 (82.9% vs. 77.4%, p = .019; 97.6% vs. 93.6%, p = .001). Similar results were observed for onset-to-needle time ≤ 2.5 hr, ≤ 3 hr, and ≤ 3.5 hr (45.2% vs. 38.5%, p = .021; 66.0% vs. 60.3%, p = .042; 81.2% vs. 72.7%, p < .001), respectively (Table 4).

Additionally, we observed the timelines for rt-PA patients with and without EMS usage. The onset-to-door time and onset-to-needle time were shorter for patients who were EMS-transported (p < .001); however, we did not observe similar superiority of EMS transportation in the door-to-needle time (p = .763). The patients with atrial fibrillation and higher baseline NIHSS scores and who were included in phase 2 were more likely to use EMS (Tables S2 and S3).

We documented higher rates of independent functional outcome following rt-PA in phase 2 compared to phase 1 (79.9% vs. 72.1%, p = .002). The rate of favorable clinical outcome at 90 days was slightly, but not significantly better in phase 2 than in phase 1 (68.4% vs. 65.3%, p = .252). No differences in the rates of sICHs (1.3% vs. 2.1%, p = .142) and mortality at 90 days (2.0% vs. 2.8%, p = .356) were observed between the two phases (Figure 2). The percentages of patients with mRS scores at 90 days are presented in Figure S1, and after adjustment for age, sex, admission NIHSS scores, the mode of transportation, and the type of arrived hospital, the significance persisted (mRS score 0–2 at 90 days, adjusted odds ratio, 2.010; 95% CI, 1.444–2.798).

Among the 925 patients who achieved an independent outcome at 90 days, age, gender, diabetes, coronary heart disease, atrial fibrillation, history of cerebral infarction, admission NIHSS score, transportation means, onset-to-needle time, and period of presentation were significantly related to 90-day functional independence in the initial univariable analysis. In the multivariable regression analysis, young in age (OR: 1.811; 95% CI: 1.304–2.516; p < .001), lower admission NIHSS score (OR: 39.120; 95% CI: 23.477–65.188; p < .001 and OR: 18.324; 95% CI: 11.425–29.388; p < .001 and OR: 3.123; 95% CI: 2.044–4.773; p < .001), shorter onset-to-needle time (OR: 0.994; 95% CI: 0.992–0.997; p < .001), and treatment with rt-PA in phase 2 (OR: 1.965; 95% CI: 1.413–2.731; p < .001) were identified as independent predictors of 90-days mRS 0–2 (Table 5).

4 DISCUSSION

The Chinese stroke emergency map not only provides patients with the locations of nearby available thrombolytic hospitals, but also a quality management system, from early stroke recognition, timely arrival, and rapid treatment to minimize the prehospital and in-hospital delays for thrombolysis. In our study, the significance of the implementation of the SEM is that the proportion of patients who arrived at hospitals within a 4.5-hr time window and received rt-PA increased. The median onset-to-door time was reduced from 118 min in phase 1 to 105 min in phase 2. Furthermore, the onset-to-door time and onset-to-needle time were reduced by five minutes and nine minutes for patients treated with rt-PA, respectively. Although the door-to-needle time was not shortened significantly in phase 2, the proportion of patients with DNT ≤ 1 hr increased from 62.9% to 67.6%. Moreover, decreased onset-to-needle time was independently related to the independent functional outcome at 90 days, without increased rates of hemorrhagic complications or mortality. This phenomenon is likely due to the promotion of SEM, effectively increasing the timely recognition of stroke and awareness of intravenous thrombolysis, all while guiding more patients to the closest qualified hospitals, effectively avoiding the prehospital delays previously caused by referral systems (Huang et al., 2015).

Our findings demonstrate that the mean DNT was 52 min, which is consistent with the international recommendation of DNT < 60 min (Jauch et al., 2013). However, there is still a large gap compared with more developed countries. In the Helsinki Stroke Thrombolysis Registry, the median DNT has been cut to 20 min, 94% of patients were treated within 60 min of arrival (Meretoja et al., 2012), and the Canadian guidelines recommend a 30-min median DNT target for patients treated with rt-PA (Casaubon et al., 2015). These differences may be due to multiple factors. The majority of patients in our study were self-transported (64.1%), which to some extent can reduce the delay in onset call EMS arrival time, but it may also fail to ensure timely activation of the in-hospital emergency management program for stroke, which is triggered by referral or triage in the emergency room.

Several previous papers have reported the beneficial effects of the EMS on the reduction of the prehospital pathway and an increase in the rate of thrombolytic treatment by prehospital hospital notification (Gu et al., 2019). However, here, the use of EMS was slightly increased in phase 2. This corrective factor did not improve in-hospital delay in our study, possibly because only 35.9% of stroke patients in the time window were transferred to the hospital by EMS, which was much lower than the 63.7% observed in the “Get With The Guidelines ― Stroke” program (Ekundayo et al., 2013). A survey of three cities in China indicated that only 18.8% of residents called for emergency services (Jiang et al., 2016). This is a reminder that future work should focus on culture-adapted stroke education, and recent “stroke-1-2-0” or “FAST” public awareness campaigns are likely to promote awareness of stroke and the EMS activation. Additionally, we observed that patients with atrial fibrillation or greater stroke severity were more likely to use the EMS. For these patients, the neurologist evaluation (Vahidy, Rahbar, Lal, Grotta, & Savitz, 2015) and the patients’ caregivers’ concerns about the risk of thrombolysis may cause delays for treatment. Standardized training has been implemented to teach clinicians to be more persuasive when educating patients and caregivers regarding the importance of timely intervention for stroke.

The clinical outcome benefit associated with decreased time to treatment with rt-PA and without increased risk of intracranial hemorrhage and death (Fonarow et al., 2014; Lees et al., 2010). Our findings are similar to those of previous research. In the “Get with The Guidelines ― Stroke” registry, the patients treated with rt-PA within 60 min were more likely to be discharged to their home and capable of independent ambulation at discharge than the patients treated within 61–270 min (Tsivgoulis et al., 2017). These findings indicate that intensive efforts to improve the timeliness of thrombolytic therapy are vital. The EMS rapid arrival on the scene, prenotification, and timely activation of the emergency services are important to improve the clinical outcomes.

Due to a poor understanding of stroke, patients typically prefer to select the CSCs that are representative of the relatively high-level hospitals. However, the unequal geographical distribution of CSCs and PSCs results in limited access to thrombolysis therapy for some regions. Most CSCs are located in densely populated and congested areas with numerous self-transport patients in emergency departments, and wait times for triage are extensive. This is another possible explanation for in-hospital delays. In addition, the prevalence of stroke in rural areas is higher than in urban areas. Therefore, the other significance of promoting the SEM is that it results in training more PSCs located in rural areas eligible for thrombolysis. Patients choosing a qualified PSC might be a better option for expedited treatment. Although the proportion of patients transferred to PSCs has increased since the promotion of the SEMs, it is still significantly lower than those transported to CSCs. Therefore, actively correcting misconceptions, utilizing the emergency service system to select an optimal stroke center, and establishing a proper referral system are likely to benefit stroke patients.

5 CONCLUSIONS

In conclusion, the implementation of the stroke emergency map is more likely to reduce the prehospital delay and improve the likelihood of function outcomes. Future efforts are necessary to encourage other thrombolysis capable hospitals to participate in the SEM alliance, continually improve public awareness of stroke, and increase the usage of EMS.

CONFLICTS OF INTEREST

We have no conflicts of interest to declare.

AUTHOR CONTRIBUTIONS

WL contributed to conceive and revise the manuscript. TZ performed the data collection, and wrote and revised the manuscript. XZ designed the study and revised the manuscript. TZ and XZ are co-correspondence authors. HS contributed to the conception and design of study. FZ and SL contributed to perform the analysis with constructive discussions. HS, ZZ, NZ, and JS performed the data collection and analysis.

ETHICAL APPROVAL

The ethics committee of the Taiyuan Central Hospital, Shanxi, China, approved the study protocol. Patient consent was waived by the ethics committee of the Taiyuan Central Hospital due to the retrospective design of the study.