Despite recent advances in providing neurological care, stroke remains one of the major causes of mortality and patient disability worldwide (1). Endovascular thrombectomy (EVT) has significantly improved the treatment of acute ischemic stroke (AIS) and has become the standard of care in many neurological and critical care centers globally (2, 3).

As it has been reported in multiple studies, early neurological improvements can be used as a surrogate marker for predicting the functional outcome of AIS patients undergoing EVT (4–10). For instance, a surrogate for long-term outcome after EVT can be a reduction in the National Institute of Health Stroke Scale (NIHSS) score of larger than 4 points or NIHSS of 0 or 1 at 24 h (11). Therefore, identification of surrogates for early neurological improvement can be crucial in predicting longer-term outcomes of AIS patients with large vessel occlusion undergoing EVT.

In the current study, we aimed to evaluate the predictors of Ultra Early Neurological Improvement (UENI) after EVT and the relationship between UENI and outcome. We defined the Ultra Early Neurological Improvement as post-EVT NIHSS reduced by 4 points or more or NIHSS as 0–1 within 1 h after EVT (12).

There were 485 patients initially, four of whom were excluded due to loss at follow-up, four patients were excluded due to pre-treatment ASPECTS score < 6, and 12 patients were excluded due to missing NIHSS data within the first hour following EVT. After the application of the inclusion and the exclusion criteria, 465 patients were enrolled in our study. Patients were divided into two groups: No-UENI (n = 343, 73.8%) and UENI (n = 122, 26.2%). There were statistically significant differences between age (p = 0.001), pre-treatment ASPECTS (p = 0.0001), vessel occlusion site (p = 0.030), DRT (p = 0.025), LKNPT (p = 0.002), mTICI of 3 (p = 0.027), number of EVT passes (p = 0.001), and symptomatic intracranial hemorrhage (sICH) (p = 0.003) between groups (Table 1). Internal carotid artery occlusion (isolated or in tandem with middle cerebral artery) were less common in patients with UENI: 35 (28.68%) vs. 135 (39.35%), (p = 0.036). In summary, patients with UENI were younger had higher pre-treatment APECTS scores, shorter DRT and LKNPT, required a lesser number of EVT passes and achieved more frequently mTICI ≥2b or 3. They also had less frequently tandem ICA + MCA-M1 occlusions and sICH. Sex and risk factors for ischemic stroke did not differ between groups.

After the multivariate regression analysis, younger age, large vessel occlusion site, shorter LKNPT, higher pre-treatment ASPECTS, mTICI final≥2b, and fewer EVT attempts were independently associated with UENI (Table 2). The 90-day mRS scores distribution is shown in Supplementary Table 1.

Our analysis demonstrated that 82.79% of patients achieved favorable outcomes in the UENI group, while only 32.36% of patients achieved favorable outcomes in the No-UENI group. In addition, the UENI group had fewer hospitalization costs compared with the No-UENI group (median 84171.150 RMB vs. 91244.665 RMB) (Table 3). The mRS scores distribution between groups after 3 months is shown in Figure 1. In the multivariate regression analysis, age, diabetes mellitus, NIHSS score at presentation, pre-treatment APECTS, intravenous thrombolysis, final mTICI ≥2band sICH were included in the final model, along with UENI (OR: 7.999, 95% C.I. 4.415–14.495) (Table 4). Suggesting that UENI is independently associated with a better 3-month prognosis in stroke patients undergoing thrombectomy.

In this multi-center retrospective study of 465 AIS patients, we aimed to evaluate the NIHSS score shortly following the EVT procedure in order to define a group of patients with UENI and assessed if this variable associated with a better functional outcome of AIS at follow-up. In our study, we defined UENI as a reduction of 4 or more points in post-EVT NIHSS or NIHSS as 0–1 within 1 h following EVT. We found that 26.2% of patients fulfilled the criteria. No differences in sex distribution or stroke risk factors were observed between patients with UENI and No-UENI. Moreover, no differences in the functional status (mRS at presentation) before the AIS event were observed between these groups.

Our data revealed that in the UENI group, 82.79% of patients achieved favorable outcomes, whereas favorable outcomes were only achieved in 32.36% of patients in the No-UENI group. In addition, there was less hospitalization costs associated with patients who had UENI (p = 0.003). Furthermore, UENI patients achieved favorable outcomes more frequently at 3 months, compared with the No-UENI group (p < 0.0011).

Making appropriate medical and surgical decisions in the acute phase of AIS is important, as patients’ outcomes often rely on the usage of medium and/or long-term prognosis indicators (13). Baseline factors and scores, such as NIHSS have been routinely used as outcome predictors for patients with AIS (14).

It is important to acknowledge that since the advent of EVT, post-treatment measures have gained significance and have become as important as baseline measures for predicting functional outcomes in patients with AIS (14). Therefore, it is reasonable to hypothesize that post-EVT NIHSS score can be used as an appropriate surrogate for the assessment of AIS patient prognosis (15–19).

There have been slight disagreements between different studies regarding the exact definition of early neurological improvements. However, it has been generally defined as 24-h NIHSS improvement of ≥8 or 10 points or a 24-h NIHSS score of 0–1 (2, 18, 19). In addition, other studies reported that early trajectory of the post-EVT NIHSS score within two days can be an accurate predictor of functional outcome (14). However, the significance of ultra-early post-EVT indicators, namely post-EVT NIHSS, has remained understudied (20).

A major advantage of using an NIHSS score shortly after EVT for functional outcomes is that it summarizes the effect of other variables with potential prognostic weight into a single score in order to predict the functional outcome few months later, providing a potential frame to guide clinical and surgical decisions (20). In our study, younger age at AIS presentation, less parenchymal involvement assessed with ASPECTS, shorter ischemic time determined with LKNPT, greater recanalization flow by the mTICI score and less thrombotic-occlusion burden reflected by the lower number of EVT passes, related to a UENI.

Different studies have reported more discriminative power in predicting desirable functional outcomes could be achieved by incorporating immediate post-EVT NIHSS (21). In addition, constructing several multivariable regression analysis models revealed that those models which incorporated post-EVT NIHSS were more successful in predicting the patients’ outcomes (21). Similarly, another study revealed that functional outcomes can be predicted with high accuracy using early trajectory of NIHSS score within 48 h post-EVT (14). Our study is subject to some limitations. This is a retrospective study of prospectively collected data with a relatively small sample size, which can compromise the external validity of the findings. As a retrospective study, patient data collection and recording could be at risk of recall and classification bias, despite all the efforts for standardization between all 4 centers involved in the study. Patients with an ASPECTS score below 6 were excluded. This exclusion criterion removed cases with greater parenchymal involvement, which may limit the generalizability of the results to patients with more severe AIS. Future studies with more inclusive criteria are required to generalize these findings. It is also important to consider that the current study was conducted in China, which has a significant stroke burden. Findings from this study can differ from results observed in different healthcare settings using different stroke care protocols or covering patients with different demographics. Future large-scale multi-center international studies are required to address these limitations. Despite such limitations, disseminating these findings from the current study can prompt other stroke centers across the globe to use UENI to predict functional patient outcomes. In addition, findings from the current study can be used in combination with other factors to build predictive models for the prognosis of AIS patients undergoing EVT (22–29).

Appropriate medical and surgical decisions in the acute phase of AIS rely on the usage of medium and/or long-term indicators of prognosis. We showed that younger age, higher pre-treatment ASPECTS, shorter LKNPT, mTICI 2b-3 and, fewer numbers of EVT attempts were associated with UENI. In addition, UENI related to fewer hospitalization costs and more frequent favorable outcomes at 3 months. Data obtained from our study underscores that UENI can be associated to functional outcome, but further larger studies should confirm UENI as a reliable independent surrogate of favorable functional outcome in patients with AIS undergoing EVT.