Late neurological improvement during hospitalization is a predicative factor for acute ischemic stroke

Abstract

Purpose

This study aimed to investigate whether late neurological improvement (LNI) during hospitalization serves as a favorable prognostic indicator in patients with acute ischemic stroke (AIS) and to identify the predictors of LNI.

Methods

We retrospectively analyzed data from the Safe Implementation of Treatments in Stroke (SITS) registry at two stroke centers in Egypt. LNI was defined as the lack of early neurological improvement (ENI) within 24 h after admission, accompanied by neurological improvement within 7 days of hospitalization. Multivariate logistic regression was employed to analyze the factors influencing favorable prognosis (modified Rankin Scale score 0–2) and LNI.

Results

A total of 834 patients with AIS were included in this study, of which 130 (15.6%) exhibited LNI. Among these, 99 (76.2%) achieved favorable outcomes. Both ENI (OR = 6.756, 95% CI 3.355–13.603, P < 0.001) and LNI (OR = 5.760, 95% CI 3.146–10.545, P < 0.001) were independently associated with favorable prognosis after adjustment. Predictors of LNI in multivariable-adjusted logistic regression included younger age (OR = 0.973, 95% CI 0.957–0.990, P = 0.001), higher baseline National Institutes of Health Scale (NIHSS) score (OR = 1.196, 95% CI 1.146–1.248, P < 0.001), and rt-PA treatment (OR = 1.953, 95% CI 1.206–3.163, P = 0.006).

Conclusions

Approximately three-quarters of patients with LNI are expected to achieve good outcome. LNI was a favorable prognostic indicator in patients with AIS and younger age, higher baseline NIHSS score and rt-PA treatment predicted LNI.

Introduction

Early neurological improvement (ENI), initially established as an efficacy biomarker for intravenous thrombolysis in acute ischemic stroke (AIS) [1, 2], has been increasingly adopted to evaluate therapeutic outcomes following endovascular thrombectomy [3,4,5]. ENI typically indicates the presence of good collateral circulation, smaller core infarct volume, shorter reperfusion time, and successful reperfusion [6, 7]. Therefore, ENI has been shown to be strongly associated with 90-day functional independence across diverse treatments, spanning intravenous thrombolysis [8, 9], thrombectomy [3, 10, 11] and conservative medical treatment [12, 13]. However, a considerable proportion of patients do not experience ENI but exhibit relatively late neurological improvement (LNI) during hospitalization, and such patients do not necessarily indicate poor prognosis. Research indicates that any neurological improvement during hospitalization predicts favorable 90-day prognosis compared to non-responders, regardless of the treatment (thrombectomy or conservative treatment) [13, 14]. However, most studies have predominantly focused on elucidating the association between ENI and long-term outcomes in AIS, while investigating hospitalization-associated LNI remains comparatively neglected. Furthermore, the prognostic implications of LNI specifically in thrombolysis-treated populations have not been comprehensively characterized. Additionally, early AIS is characterized by instability and may be complicated by conditions such as brain edema, hemorrhagic transformation, and pulmonary infections [15, 16]. Therefore, the aim of this study is to investigate the long-term relationship between relative LNI during hospitalization and prognosis in AIS patients, as well as to identify the predictive factors for LNI.

Methods

Study population

This retrospective study utilized data from the Safe Implementation of Treatments in Stroke (SITS) International Registry, a prospective, open-label, multinational registry monitoring acute stroke management across 48 countries [17]. We extracted consecutive patients treated at two comprehensive stroke centers within Ain Shams University Hospitals (Cairo, Egypt) and there were no patients undergoing endovascular thrombectomy in this data [18]. The inclusion criteria were as follows: (1) patients diagnosed with acute ischemic stroke; (2) patients who presented within 24 h of symptom onset; (3) data that included National Institutes of Health Scale (NIHSS) scores on the day of admission, at 24 h post-admission, at 7 days post-admission, and the 90-day mRS score. Acute ischemic stroke (AIS) is defined as: an episode of neurological dysfunction caused by focal cerebral, spinal, or retinal infarction [19]. The exclusion criteria were as follows: (1) patients with transient ischemic attacks (TIA); (2) patients with hemorrhagic stroke, brain tumors, or other non-ischemic strokes; (3) patients without complete clinical data at 24 h or day 7 of hospitalization; (4) patients whose symptom onset exceeded 24 h. The dataset initially contained 1450 patients, of whom 344 were lost to follow-up, 206 had symptom onset after 24 h, 60 had incomplete NIHSS data, and 6 were diagnosed with TIA. Ultimately, 834 patients were included in the final analysis (Fig. 1).

Fig. 1

Flow chart of patient selection. ENI early neurological improvement, LNI late neurological improvement, AIS acute ischemic stroke, mRS modified Rankin Scale, NIHSS National Institutes of Health Stroke Scale, SITS Safe Implementation of Treatments in Stroke

Date collection

Data collected included baseline and demographic characteristics, acute treatment variables, baseline stroke severity within 24 h after stroke onset using the admission NIHSS score (NIHSS baseline), and follow-up NIHSS scores (NIHSS 24 h, NIHSS 7d) collected at 24 h and 7 days from the baseline examination. Stroke subtypes were classified according to the TOAST (Trial of ORG 10172 in Acute Stroke Treatment) criteria [20].

Definition of ENI and LNL

ENI was defined as the ≥ 4-point NIHSS reduction or complete neurological recovery (NIHSS = 0) within 24 h post-admission [21]. LNI was defined as the ≥ 4-point NIHSS reduction or complete recovery within 7 days after admission but without ENI. A good outcome was defined as an mRS score ≤ 2 at 3 months.

Statistical methods

SPSS Statistical Package 25.0 software was used for statistical analysis. Normally distributed variables were presented as the mean ± SD and were compared using Student’s t-tests. Non-normally distributed variables were presented as the interquartile range (IQR) and were compared using the Mann–Whitney U test. Pearson’s chi-square test was used to compare differences in categorical variables. Backward stepwise logistic regression was employed to identify independent predictors of favorable prognosis and LNI. P-value of < 0.05 was considered statistically significant.

Results

Patients baseline characteristics

A total of 834 patients with AIS were included in this study. The study flowchart is shown in Fig. 1. The median age was 62 years, 64.5% were male, and the median NIHSS score at admission was 7 points. ENI occurred in 107 (12.8%) patients, and LNI in 130 (15.6%). In the thrombolysis group (n=191), ENI and LNI occurred in 58 (30.4%) and 37 (19.4%) patients, respectively. In the conservative treatment group (n=643), ENI and LNI occurred in 49 (7.6%) and 93 (14.5%) patients, respectively. Among the patients experiencing ENI, 94 (87.9%) achieved a favorable prognosis, while the remaining 13 (12.1%) did not. In the LNI group, 99 (76.2%) patients achieved a favorable prognosis, while 31 (23.8%) did not (Fig. 2). The baseline variables between the ENI and non-ENI groups are shown in Table 1. A higher proportion of patients in the ENI group received thrombolysis treatment (54.2% vs. 18.3%, P < 0.001), had a shorter onset-to-door time (ODT) (180 [120–360] min vs. 420 [300–720] min, P < 0.001), and had fewer small vessel disease subtypes (28.0% vs. 39.5%, P = 0.023). Table 1 also presents the baseline characteristics of the LNI and non-LNI groups. Compared to the non-LNI group, the LNI group had a higher proportion of patients receiving rt-PA treatment (28.5% vs. 16.1%, P = 0.001), a greater proportion of large vessel disease subtypes (43.8% vs. 34.7%, P = 0.049), higher NIHSS scores at admission (10 [6–14] vs. 6 [4–10], P < 0.001), and fewer small vessel disease subtypes (25.4% vs. 42.5%, P < 0.001).

Fig. 2

Scores on the modified Rankin Scale (mRS) at 90-d follow-up according to neurological improvement assessed at different times. ENI early neurological improvement, LNI late neurological improvement

Table 1 Comparison between groups with and without ENI and LNI

Relationship between LNI and good prognosis in AIS patients

As shown in Table 2, multivariate regression analysis indicated that both ENI (OR = 6.756, 95% CI 3.355–13.603, P < 0.001) and LNI (OR = 5.760, 95% CI 3.146–10.545, P < 0.001) were significantly associated with favorable prognosis in both the thrombolysis group and the medical conservative treatment group. In the thrombolysis group, patients with ENI had 9.3-fold higher odds of favorable outcomes compared to those without ENI (OR = 9.299, 95% CI 3.080–27.661, P < 0.001), while patients with LNI had 3.5-fold higher odds compared to those without LNI (OR = 3.482, 95% CI 1.110–10.927, P = 0.032). In the medical conservative treatment group, ENI and LNI were associated with 4.5-fold (OR = 4.490, 95% CI 1.613–12.496, P = 0.004) and 7.1-fold (OR = 7.138, 95% CI 3.437–14.826, P < 0.001) higher odds of favorable outcomes, respectively.

Table 2 Multivariable logistic regression analysis of ENI and LINI association with favorable outcomes

Predictors of LNI

In multivariate analysis, patients with LNI were significantly younger (OR = 0.973, 95% CI 0.957–0.990, P = 0.001), had higher baseline NIHSS score (OR = 1.196, 95% CI 1.146–1.248, P < 0.001), and were more likely to have received thrombolysis treatment (OR = 1.953, 95% CI 1.206–3.163, P = 0.006) (Table 3).

Table 3 Independent predictors of LNI

Discussion

The main finding of this study is that LNI during hospitalization serves as a critical prognostic factor for long-term favorable outcomes in AIS patients undergoing thrombolysis and conservative drug therapy, paralleling the predictive value of ENI. This observation holds substantial clinical relevance, as patients exhibiting LNI demonstrate comparable prognostic trajectories to those with ENI. Consequently, aggressive therapeutic strategies may be considered irrespective of the initial clinical progression patterns. Moreover, the identification of baseline predictors associated with LNI offers clinicians actionable insights to refine prognostic evaluations for patients lacking ENI. Such stratification may foster proactive clinical decision-making, thereby enhancing patient engagement in rehabilitation protocols and optimizing long-term functional recovery and quality of life.

In our study cohort, the prevalence of ENI in patients receiving thrombolysis and conservative drug therapy was 30.4% and 7.6%, respectively, which is consistent with the previous studies [13, 21]. The prevalence of LNI was 19.4% in thrombolysis patients and 14.5% in conservative drug therapy patients, while the latter is lower than the previously reported prevalence of 33.41% [13], a discrepancy potentially explained by prolonged onset-to-door time (ODT) observed in our conservative therapy group (10.6 ± 6.2 h vs. 4.25 ± 3.15 h). It is widely acknowledged that the shorter the ODT, the greater the likelihood of neurological improvement, regardless of the treatment received [22].

The relationship between ENI and long-term favorable prognosis in AIS has been widely recognized, which aligns with our findings, regardless of whether patients received thrombolysis or conservative drug therapy. The possible mechanisms include successful reperfusion, good collateral circulation, smaller core infarct volume, and shorter reperfusion time to successful reperfusion [6, 7]. However, a considerable portion of patients do not show ENI but instead exhibit LNI during hospitalization. This does not necessarily imply a poor prognosis. A retrospective study found that patients who did not show significant neurological improvement after thrombectomy but achieved clinical neurological improvement within 7 days post-thrombectomy were associated with long-term favorable prognosis [14]. Another study on conservative drug therapy patients demonstrated that those who experienced neurological improvement at different times during hospitalization had a higher rate of favorable prognosis than those who showed no neurological improvement during hospitalization [13]. Our study is the first to report that LNI during hospitalization in intravenous thrombolysis patients without ENI is associated with long-term favorable prognosis. This suggests that intravenous thrombolysis plays an important role in neurological improvement both in the early and late stages post-treatment and further supports the necessity of intravenous thrombolysis. Subgroup analysis in our study showed that, in the thrombolysis group, the probability of ENI occurrence was higher than that of LNI (30.4% vs. 19.4%). Compared to patients without ENI, those with ENI had a ninefold higher likelihood of favorable prognosis; compared to patients without LNI, those with LNI had a threefold higher likelihood of favorable prognosis. This may indicates that the earlier the neurological improvement after thrombolysis, the greater the association with a favorable prognosis, which aligns with previous findings from a thrombectomy study, where earlier clinical improvement was closely related to a higher chance of long-term favorable outcomes [14]. However, this does not appear to be the case in the conservative drug therapy group. The probability of ENI occurrence was lower than that of LNI (7.6% vs. 14.5%). Compared to patients without ENI, those with ENI had a fourfold higher likelihood of favorable prognosis; compared to patients without LNI, those with LNI had a sevenfold higher likelihood of favorable prognosis. These differences may be largely attributed to the treatment method and potential pathophysiological mechanisms. Thrombectomy and intravenous thrombolysis are typically associated with faster and more significant neurological improvement, as well as higher rates of vascular recanalization, compared to conservative drug therapy [23, 24]. As a result, the probability of ENI occurring in the thrombolysis group is greater than that in the conservative therapy group. Higher and earlier vascular recanalization can salvage more ischemic penumbra, making earlier clinical improvement in thrombolysis patients more closely associated with better long-term outcomes. In contrast, the rate and speed of spontaneous vascular recanalization in conservative drug therapy patients are lower than those in the thrombolysis group, which may explain the higher prevalence of LNI. Currently, the mechanisms underlying the association between LNI and favorable prognosis are still not fully understood. While delayed vascular recanalization, microcirculation improvement in ischemic tissue, recovery of cerebral autoregulation, and neuronal remodeling are plausible hypotheses based on prior studies [25, 26]. Our study lacks direct imaging evidence (e.g., perfusion imaging or angiographic reperfusion) to confirm these pathways. Future studies incorporating multimodal neuroimaging could elucidate whether LNI reflects delayed reperfusion, neuroplasticity, or other compensatory processes. Finally, our study found that younger age, higher baseline NIHSS score, and thrombolysis treatment are predictive factors of LNI. A higher baseline NIHSS score indicates a more severe stroke, which may lead to a more significant response after acute treatment (such as intravenous thrombolysis or thrombectomy). This may allow more room for neurological improvement with successful acute interventions. Therefore, patients with higher baseline NIHSS scores at admission have a greater chance of experiencing LNI, but this is specific to the characteristics of our study population. The baseline NIHSS of our study population was lower than that in other studies (7 points vs. 16 points) [16], and if the NIHSS score were higher, the chance of a favorable prognosis could be diminished due to the already high baseline NIHSS [11]. Therefore, the relationship between baseline NIHSS and LNI warrants further investigation. We hypothesize that patients demonstrating LNI during hospitalization may still achieve superior long-term functional outcomes compared to non-LNI counterparts, despite presenting with higher NIHSS scores at admission. The fact that thrombolysis is a predictor of LNI is also not difficult to understand. Compared to patients who did not receive thrombolysis, those who received thrombolysis have a higher rate of vascular recanalization and more rapid and significant neurological improvement. Additionally, it is well established that older age is associated with poorer neurological recovery [27].

This study had several limitations. Firstly, this is a retrospective study. Second, the dataset lacks certain clinical information, such as complications during hospitalization (e.g., infections, hemorrhages), imaging data, blood pressure, and blood glucose levels, which could potentially influence the results and may play a role in the mechanistic explanation. Third, since our dataset includes intravenous thrombolysis patients, we excluded those who presented more than 24 h after symptom onset. This approach helps standardize the definitions of ENI and LNI. However, after including patients with onset times exceeding 24 h in our study, we still found that LNI was associated with a favorable 3-month prognosis (OR = 6.414, 95% CI 3.504–11.740, P < 0.001). Finally, since we did not include patients who underwent endovascular thrombectomy and the average baseline NIHSS score of the patients in our study was relatively low, the predictive factors for LNI may not be fully applicable to patients in the thrombectomy group or those with a high baseline NIHSS score at admission.

Conclusions

In conclusion, LNI in acute ischemic stroke patients is significantly associated with a favorable 3-month prognosis. It is important not to focus solely on ENI while overlooking LNI, in order to avoid misjudging the prognostic assessment of stroke patients.