SPRIN₂D nomogram: predicting 90-day unfavorable outcomes in patients with minor ischemic stroke receiving intravenous thrombolysis

Abstract

Objective:

This study aimed to construct and validate a nomogram for the intuitive prediction of 90-day unfavorable outcomes in patients with minor ischemic stroke who underwent intravenous thrombolysis.

Methods:

Data were extracted from the Multicenter Clinical Trial of Revascularization Treatment for Acute Ischemic Stroke (TRAIS) study. Multivariable logistic regression was used to identify independent factors associated with 90-day unfavorable outcomes (modified Rankin Scale [mRS] score 2–6) in patients with minor ischemic stroke (National Institutes of Health Stroke Scale [NIHSS] score 0–5) receiving intravenous thrombolysis. A nomogram (designated as SPRIN₂D) was developed based on the final prediction model, incorporating 6 key variables: NIHSS score, functional disability, systolic blood pressure, neutrophil count, international normalized ratio and renal dysfunction.

Results:

A total of 1,438 eligible patients were enrolled and randomly assigned to a derivation cohort (n = 1,007) and a validation cohort (n = 431). In the overall cohort, the median age was 65 years (interquartile range, 56–73); 67.94% of participants were male, and the median admission NIHSS score of 2 (interquartile range, 1–4). The SPRIN₂D nomogram exhibited robust performance in predicting 90-day unfavorable outcomes in minor ischemic stroke patients treated with intravenous thrombolysis. The area under the curve was 0.784 (95% confidence interval [CI], 0.744–0.822) in the derivation cohort and 0.776 (95% CI, 0.716–0.835) in the validation cohort. Moreover, the SPRIN₂D nomogram demonstrated superior discriminative ability compared with other existing prognostic models for acute ischemic stroke and minor ischemic stroke.

Conclusion:

The SPRIN₂D nomogram is a stable and promising tool for predicting 90-day unfavorable outcomes in minor ischemic stroke patients treated with intravenous thrombolysis.

1 Introduction

Acute ischemic stroke (AIS) remains a leading global cause of disability and mortality (1). Despite significant advancements in diagnostic and therapeutic strategies, the associated social and economic burdens continue to escalate (2). Minor ischemic stroke (MIS) accounts for approximately 50% of all patients with AIS (3, 4). Notably, the prognosis of MIS patients is not optimistic: 13.6% experience early neurological deterioration (END), which significantly increases the risk of stroke recurrence, disability, and death, highlighting the necessity of proactive clinical management (5, 6).

Intravenous thrombolysis (IVT) is the guideline-recommended first-line treatment for AIS, but its efficacy is highly time-dependent, requiring administration within a 4.5 h therapeutic window. However, IVT utilization among MIS patients remains suboptimal: fewer than half of MIS patients in the United States receive this therapy, with even lower rates reported in China (7, 8). This underutilization stems from clinical uncertainty. Current guidelines recommend IVT for MIS patients with disabling symptoms, but lack sufficient evidence to support its use in non-disabling cases (9). Importantly,a diagnosis of MIS does not guarantee a favorable outcome; patients with low National Institutes of Health Stroke Scale (NIHSS) scores may still develop subsequent functional impairments (10). For instance, studies have indicated that approximately one-third of MIS patients without IVT treatment have an unfavorable outcome at 3 months (11), while several clinical trials have demonstrated the potential benefits of IVT in this population (12). Thus, relying solely on neurological deficit types and NIHSS scores to guide thrombolysis decisions for MIS patients is overly restrictive. There is an urgent need to establish an integrated predictive model for unfavorable outcomes in IVT-treated MIS patients.

Previous studies have identified multiple factors associated with unfavorable outcomes in MIS patients undergoing IVT, including stroke severity, NIHSS score, age, sex, diabetes mellitus, and hypertension (13, 14). However, existing thrombolysis-related prognostic models lack a specific focus on MIS patients (15, 16). To address this gap, the present study aims to develop a stable nomogram for predicting the risk of unfavorable outcomes in MIS patients who received IVT therapy. By integrating key clinical, demographic, and laboratory variables, this tool aims to provide clinicians with an evidence-based approach for identifying patients at high risk of unfavorable outcomes at an early stage, thereby optimizing the efficient utilization of medical resources and improving patient outcomes.

2 Materials and methods

2.1 Study population

The data for this study were derived from the Multicenter Clinical Trial of Revascularization Treatment for Acute Ischemic Stroke (TRAIS) study, a multicenter cohort that enrolled patients from 14 hospital centers between June 2020 and January 2025 (17). Patients were recruited in accordance with the following inclusion criteria: (1) received IVT within the time window according to thrombolytic therapy indications; (2) clinical and radiological diagnosis of AIS; (3) age ≥18 years; Exclusion criteria were as follows: (1) cerebral hemorrhage or contraindications to thrombolytic therapy; (2) incomplete clinical data; (3) mental disorders or severe cognitive impairment. Patients with NIHSS scores ≤ 5 were included in the study.

This study was approved by the Ethics Committee of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, registered in the Chinese Clinical Trial Registry (ChiCTR2000033456), and written informed consent was obtained from all participants prior to enrollment.

2.2 Collection of clinical data

Demographic characteristics (age and gender) and medical history information were collected from all patients at admission. Hypertension was defined as meeting one or more of the following criteria: blood pressure ≥ 140/90 mmHg, self-reported physician-diagnosed hypertension, or current use of antihypertensive medications (18). Hyperlipidemia was identified as elevated blood lipid levels (total cholesterol ≥ 6.22 mmol/L, triglyceride ≥ 2.26 mmol/L, low-density lipoprotein cholesterol ≥ 4.14 mmol/L) or a documented history of hyperlipidemia (19). Diabetes was defined as a fasting blood glucose (FBG) level ≥ 7.0 mmol/L, physician-diagnosed diabetes, or current use of antidiabetic drugs. Renal dysfunction was characterized by the presence of albuminuria (urine albumin-creatinine ratio ≥ 30 mg/g), abnormal urine sediment, renal tubule-related lesions, histological or structural renal abnormalities, history of kidney transplantation, reduced glomerular filtration rate (GFR < 60 mL/min/1.73 m²), or self-reported chronic kidney disease (20).

Admission clinical characteristics, including baseline systolic blood pressure (SBP), baseline diastolic blood pressure (DBP), NIHSS score, and functional disability, were assessed by experienced neurologists before IVT treatment. Functional disability was assessed before IVT administration using baseline NIHSS item scores. Patients were classified as having functional disability if they presented with any of the following potentially disabling neurological deficits: complete hemianopia (NIHSS score ≥ 2), severe aphasia (NIHSS score ≥ 2), neglect (NIHSS score ≥ 1), or limb weakness (NIHSS score ≥ 2) (21).

Laboratory parameters, including lymphocyte count, neutrophil count, platelet count, prothrombin time (PT), fibrinogen, activated partial thromboplastin time (APTT) and international normalized ratio (INR), were collected on admission. FBG was measured within 12 h after enrollment.

2.3 Outcome

The primary outcome was defined as an unfavorable outcome, which was identified as a modified Rankin Scale (mRS) score of 2–6 at 90 days post-treatment. This cutoff was selected given that an mRS score of 2 or higher is the widely accepted standard for identifying clinically relevant long-term functional dependence after AIS. The mRS score was assessed through telephone interviews or electronic communication with patients or their authorized representatives and all assessors were blinded to patients’ baseline clinical characteristics to ensure consistent and reliable outcome collection.

2.4 Statistical analysis

The Kolmogorov–Smirnov test was applied to assess the normality of data distribution. Continuous variables were expressed as the mean ± standard deviation (SD) if normally distributed or median with interquartile range (IQR) if non-normally distributed. Categorical variables were presented as frequencies (%) and analyzed using chi-square tests. Differences between groups were examined by independent samples t-tests (for normally distributed data) or Mann–Whitney U tests (for non-normally distributed data), with a two-sided p < 0.05 considered statistically significant. Intra-group differences were analyzed using independent sample two-tailed t-tests or Mann–Whitney U tests, as appropriate.

Logistic regression analyses were conducted to identify potential prognostic predictors, with results presented as odds ratios (ORs) and corresponding 95% confidence interval (CI). Variables with p < 0.05 in the univariate logistic analysis were incorporated into stepwise multivariate logistic regression to screen for independent prognostic factors. The discriminative ability of the nomogram was evaluated using the receiver operating characteristic (ROC) curves, with the area under the curve (AUC) serving as the key indicator.

Patients were randomly divided into the derivation cohort and validation cohort at a ratio of 7:3. Univariate and multivariate logistic regression analyses were conducted in the derivation cohort to construct the SPRIN₂D nomogram, which was then validated in both cohorts. Each predictor was assigned a weight based on its regression coefficient. For each individual patient, the weighted sum of all predictors, together with the model intercept, was entered into the logistic function to calculate the estimated probability of a 90-day unfavorable outcome.

The “rms” and “rmda” packages in R software were used to construct the nomogram and perform clinical decision curve analysis (DCA). Calibration curves were plotted to assess the consistency between the predicted probabilities and the observed outcomes of the logistic regression model. To evaluate the predictive performance of the nomogram for 90-day unfavorable outcomes, we calculated diagnostic indices including sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (PLR) and negative likelihood ratio (NLR) at distinct candidate score thresholds. Subsequently, we defined the risk stratification cutoffs in a data-driven manner by comprehensively analyzing the distribution of 90-day unfavorable outcome rates across nomogram scores, variations in diagnostic indices at different threshold values, and the clinical applicability of the stratified groups. Using these optimized score cutoffs, patients were categorized into low-, medium-, and high-risk groups. Cutoff-based risk estimates were separately calculated in the derivation and validation cohorts, and the incidence of 90-day unfavorable outcomes was calculated for each risk group. A linear trend test was performed to assess the stepwise increase in risk across the three strata.

Furthermore, we compared the discriminative ability of the SPRIN₂D nomogram with existing prognostic models for AIS and MIS model in predicting 90-day unfavorable outcomes (mRS score ≥ 2) in the validation cohort. For a more comprehensive comparison, calibration performance and clinical utility were also assessed for all included models. All statistical analyses were performed using R software (version 4.3.2) and SPSS statistics (version 29.0.2).

3 Result

3.1 Baseline characteristics of participants

A total of 2,336 participants were enrolled in the TRAIS study, among whom 1,595 were diagnosed with MIS. Patients were excluded if they had incomplete clinical data (n = 15), were lost to follow-up (n = 95), or underwent endovascular treatment (n = 47). Incomplete clinical data were mainly due to the absence of subsequent examinations or further treatment after thrombolysis. The term “lost to follow-up” was defined as the unavailability of 90-day mRS scores. After screening, 1,438 eligible MIS patients were randomly assigned to the derivation cohort (n = 1,007) and the validation cohort (n = 431) at a 7:3 ratio (Figure 1). Unfavorable outcomes (mRS score ≥ 2) were observed in 167 patients (16.58%) in the derivation cohort and 56 patients (12.99%) in the validation cohort.

Figure 1

Baseline characteristics of the participants, including demographic data, clinical characteristics, and laboratory measurements, are summarized in Table 1. The median age of the total cohort was 65 (IQR, 56–73) years, 67.94% were male, and the median baseline NIHSS score was 2 (IQR, 1–4). No statistically significant differences were observed in any baseline characteristics between the two cohorts (p = 0.072–0.994) confirming the effectiveness of random allocation (Table 1).

3.2 Univariate and multivariate analyses

In the derivation cohort, univariate logistic regression analysis indicated that patients with unfavorable outcomes had significantly higher SBP (p < 0.001), DBP (p < 0.001), FBG levels (p < 0.001), neutrophil counts (p = 0.004), INR (p < 0.001) and NIHSS scores (p < 0.001) compared with those with favorable outcomes. Additionally, medical histories of renal dysfunction (p < 0.001) and hypertension (p = 0.018) were more prevalent in MIS patients with unfavorable outcomes (Table 2). We subsequently conducted multicollinearity analysis for the above variables, and all variance inflation factor (VIF) values were below 5, indicating the absence of severe multicollinearity among these variables (Supplementary Table S1).

For multivariate analysis, continuous variables were binarized using clinically relevant cutoffs determined by routine clinical practice and laboratory reference ranges. All cutoffs were rounded to the nearest integer to enhance clinical feasibility and applicability. Given the limited range of NIHSS scores (0–5) among patients with MIS, this variable was retained as a continuous measure to preserve significant differences in neurological severity, which would be masked if it were dichotomized. Due to the strong correlation between SBP and DBP, only SBP was retained in the multivariable analysis as the representative blood pressure parameter to avoid multicollinearity and improve model stability.

Multivariate logistic regression analysis identified 6 independent factors associated with 90-day unfavorable outcomes after IVT treatment: renal dysfunction (OR = 3.19; 95% CI, 1.94–5.25; p < 0.001), NIHSS score (OR = 1.23; 95% CI, 1.07–1.42; p = 0.004), functional disability (OR = 2.03; 95% CI, 1.33–3.12; p = 0.001), SBP ≥ 140 mmHg (OR = 2.66; 95% CI, 1.71–4.16; p < 0.001), neutrophil counts ≥ 6.0 × 10⁹/L (OR = 2.32; 95% CI, 1.60–3.38; p < 0.001) and INR ≥ 1.0 (OR = 2.27; 95% CI, 1.56–3.29; p < 0.001). Hypertension (OR = 1.03; 95% CI, 0.69–1.52; p = 0.895) and FBG ≥ 7.0 mmol/L (OR = 1.43; 95% CI, 0.98–2.07; p = 0.062) were excluded due to non-significant associations (Table 3). These 6 predictors were subsequently incorporated into the SPRIN₂D nomogram, whose corresponding regression coefficients and model intercept were as follows: intercept (−4.058), NIHSS score (0.217), functional disability (0.699), SBP ≥ 140 mmHg (0.982), neutrophil counts ≥ 6.0 × 10⁹/L (0.839), INR ≥ 1.0 (0.817), renal dysfunction (1.172).

3.3 The SPRIN₂D nomogram

The SPRIN₂D nomogram was constructed based on the 6 independent prognostic factors identified in the derivation cohort. The ROC curve analysis suggested that the model exhibited good discriminative ability in both the derivation cohort (AUC = 0.784; 95% CI, 0.744–0.822) and the validation cohort (AUC = 0.776; 95% CI, 0.716–0.835; Figure 2).

Figure 2

The nomogram uses a cumulative scoring method: for each patient, the corresponding score for each of the 6 variables is obtained by vertically aligning the variable value with the topmost score axis (total score range: 0–100). The individual scores for all variables are summed to obtain a total score, which is then mapped to the “Total points” axis. A higher total score indicates an increased probability of 90-day unfavorable outcomes. For illustration, a patient with a NIHSS score of 3 (55.5 points), no functional disability (0 points), SBP of 150 mmHg (83.5 points), neutrophil count of 7.5 × 10⁹/L (71.5 points), INR of 1.22 (69 points), and no renal dysfunction (0 points) achieved a total score of 279.5, corresponding to a predicted 90-day unfavorable outcome probability of approximately 31.7% (Figure 3).

Figure 3

Calibration curves were plotted to evaluate the consistency between the predicted and observed probabilities. The SPRIN₂D nomogram showed good calibration in the derivation cohort (Hosmer-Lemeshow test p = 0.289; intercept = −0.001; slope = 1.000; Brier score = 0.115) and the validation cohort (Hosmer-Lemeshow test p = 0.196; intercept = −0.218; slope = 1.056; Brier score = 0.096). Non-significant Hosmer-Lemeshow test results, intercepts near 0, slopes close to 1, and low Brier scores collectively confirmed satisfactory consistency between predicted and actual risks (Supplementary Figure S1).

Furthermore, the DCA curve was constructed with threshold probability on the x-axis and net benefit on the y-axis. The SPRIN₂D nomogram yielded superior net benefit relative to the treat-all and treat-none approaches across threshold ranges of 0.120–0.820 (derivation cohort) and 0.188–0.820 (validation cohort). For patients with elevated predicted risk within these intervals, clinicians can implement intensified surveillance, personalized management and early preventive interventions. Collectively, DCA verified the favorable predictive performance of the SPRIN₂D nomogram model for 90-day unfavorable outcomes (Supplementary Figure S2).

Subsequently, we evaluated the nomogram’s diagnostic performance across various score thresholds (Table 4). Using total nomogram scores, participants were stratified into low-risk (< 100 points), medium-risk (100–299 points), and high-risk (≥ 300 points) subgroups to facilitate clinical implementation. As summarized in Table 5 and Figure 4, the incidence of 90-day unfavorable outcomes rose substantially from 4.1% (8/194) in the low-risk group to 14.2% (100/702) in the medium-risk group and 53.2% (59/111) in the high-risk group, with a statistically significant increasing linear trend (p < 0.001). A consistent progressive risk gradient was also identified in the validation cohort (Supplementary Table S2). Such risk stratification facilitates intuitive nomogram interpretation and helps clinicians readily distinguish patients at varying prognostic risks.

Figure 4

To further explore the model’s applicability across distinct stroke subtypes, we excluded 65 patients with undetermined stroke territories, leaving 1,373 eligible participants categorized into anterior circulation stroke (n = 1,113) and posterior circulation stroke subgroups (n = 260). Subgroup validation via ROC analysis produced an AUC of 0.779 (95% CI, 0.742–0.817) for anterior circulation stroke and 0.765 (95% CI, 0.681–0.843) for posterior circulation stroke (Supplementary Figure S3). These results imply favorable predictive efficacy of the nomogram for posterior circulation stroke. Although the underlying mechanisms responsible for this discrepancy remain unclear, our findings provide supportive evidence regarding this unaddressed question. Additional large-scale cohort validation is required, particularly for the smaller posterior circulation subgroup.

3.4 Comparing the SPRIN₂D nomogram with existing predictive models and individual predictors

To further validate the discriminative performance of the SPRIN₂D nomogram, we compared its predictive performance with that of three published prognostic tools (the THRIVE score, Stroke-TPI score, and Lei’s MIS nomogram) within the validation cohort (n = 431) (22–24). Supplementary Table S3 summarizing the key characteristics of the compared models. For 90-day unfavorable outcomes, the corresponding AUC value was 0.631 (95% CI, 0.584–0.677) for the MIS nomogram, 0.567 (95% CI, 0.519–0.614) for the THRIVE score, 0.656 (95% CI, 0.609–0.701) for the Stroke-TPI score, and 0.776 (95% CI, 0.716–0.835) for the SPRIN₂D nomogram (Figure 5; Table 6). The SPRIN₂D nomogram yielded the largest AUC value across all assessed models.

Figure 5

Model calibration was quantified using calibration intercept, slope, and Brier score. The SPRIN₂D nomogram had a calibration intercept of −0.218 and a slope of 1.056, close to ideal values of 0 and 1, alongside a low Brier score of 0.096 relative to competing models, suggesting minimal prediction error and well-balanced calibration (Supplementary Figure S4; Supplementary Table S4). DCA further confirmed superior clinical utility, as the SPRIN₂D nomogram generated greater net benefit than the three alternative models over the threshold probability interval of 0.188–0.820 (Supplementary Figure S5).

In the overall cohort (n = 1,438), the AUC values of individual independent predictors ranged from 0.590 to 0.667 (Supplementary Table S5). SBP ≥ 140 mmHg yielded the lowest AUC of 0.590, followed by renal dysfunction (0.591), neutrophil count ≥ 6.0 × 10⁹/L (0.607), INR ≥ 1.0 (0.617), functional disability (0.643), and NIHSS score (0.667). In contrast, the SPRIN₂D nomogram attained an AUC of 0.779, indicating superior discriminative performance than all standalone predictors.

4 Discussion

In clinical practice, over 50% of patients with MIS do not receive IVT treatment, primarily due to their non-disabling symptoms, uncertainty regarding thrombolytic benefit, and concerns about hemorrhagic risk (25). Whether IVT improves outcomes in MIS remains a critical, unresolved clinical question. Accumulating evidence suggests that stroke is a systemic disease with complex pathophysiology, affecting not only the nervous system but also multiple distal organs (26). To address this gap and predict the risk of unfavorable outcomes following thrombolysis in patients with MIS, we developed a nomogram model based on logistic regression analysis. The results identified 6 key predictors of unfavorable outcomes: admission NIHSS score, functional disability, admission SBP ≥ 140 mmHg, INR ≥ 1.0, neutrophil count ≥ 6.0 × 10⁹/L and concurrent renal dysfunction. Validation of the model confirmed its robust discriminative ability and high accuracy.

The admission NIHSS score is a well-established indicator of post-stroke neurological damage and prognosis. It is widely accepted that higher NIHSS scores correlate with an greater likelihood of unfavorable outcomes (27). Findings from the SOCRATES trial further verified the NIHSS score as an independent predictor of subsequent disability in patients with MIS (28). Leira et al. reported an inverse association between baseline NIHSS score (0–6) and 90-day unfavorable outcomes in MIS (29). Similarly, previous studies have linked baseline NIHSS score to IVT efficacy in MIS patients (30). However, the NIHSS score focuses primarily on motor and language function, with limited assessment of higher-level neurological functions. After eliminating multicollinearity, we incorporated functional disability into the model to overcome this limitation, which can comprehensively reflect the severity of stroke. A previous analysis showed that disabled AIS patients treated with IVT were more likely to recover to their pre-stroke functional status (31), supporting the clinical relevance of this variable. Nevertheless, as the definition of functional disability is partially constructed from individual NIHSS items, potential overlapping information between total NIHSS scores and this functional disability metric merits careful consideration. In our predictive model, the total NIHSS score reflects the overall severity of neurological impairment, whereas the functional disability variable specifically captures persistent disabling manifestations that interfere with patients’ daily functional capacity. Accordingly, the independent prognostic value of functional disability should be interpreted against this clinical context.

In recent years, the critical role of inflammation in stroke pathophysiology has been increasingly recognized. After stroke, neutrophils infiltrate the central nervous system, releasing matrix metalloproteinase-9, reactive oxygen species, and other proteases that induce further blood–brain barrier damage and impede vascular recanalization (32). Thrombus analysis in AIS patients revealed that neutrophil content was significantly correlated with thrombolysis resistance (33), and the pre-thrombolysis neutrophil count is an independent predictor of 90-day unfavorable outcomes in AIS patients (34). The neutrophil-to-lymphocyte ratio, which is derived from neutrophils, has also been shown to be associated with END following IVT (35). Furthermore, increased neutrophil counts and ratios have also been associated with an elevated risk of stroke recurrence in patients with MIS (36). Consistent with these findings, our study confirmed a significant correlation between neutrophil counts and 90-day unfavorable outcomes in MIS patients receiving IVT therapy, highlighting the prognostic value of inflammatory markers.

Inflammation and the coagulation system exhibit extensive crosstalk across multiple stages of stroke pathology (37). Earlier research focused on prothrombin time (PT), but INR—an adjusted index of PT—can more accurately predict unfavorable outcomes of patients with AIS by adjusting the PT ratio and laboratory-specific differences (38). Elevated admission INR has been linked to unfavorable functional outcomes in Chinese AIS patients and to END in AIS patients undergoing (38, 39). Although the underlying mechanism remains unclear, researchers have hypothesized that decreased coagulation factors, an enhanced inflammatory response, and increased bleeding risk collectively contribute to this outcome (37, 40). Our findings further emphasize the prognostic significance of INR in MIS treatment.

Renal dysfunction was also identified as an independent predictor. A meta-analysis illustrated the connection between renal dysfunction and increased stroke incidence in the general population, attributed to shared pathogenic mechanisms including vascular damage and endothelial injury (41). In particular, evidence exhibited that AIS patients with renal insufficiency have higher rates of disability and hemorrhagic transformation after IVT therapy (42), partly due to impaired endothelial tPA release and reduced fibrinolytic activity (43). Additionally, platelet and coagulation abnormalities associated with renal insufficiency increase bleeding risk following thrombolysis (44), justifying the inclusion of renal dysfunction in our nomogram.

Finally, elevated admission SBP, a traditional stroke prognostic factor, was incorporated into the nomogram. High SBP reflects abnormal vascular autoregulation, and wider blood pressure fluctuations after IVT treatment aggravate reperfusion injury to the blood–brain barrier, leading to severe complications such as cerebral edema and intracranial hemorrhage (45, 46). The ENCHANTED trial further indicated that lower SBP levels are associated with better outcomes in AIS patients receiving IVT treatment (47). In addition, validation of our model in anterior and posterior circulation stroke subgroups provides deeper insights into the study of MIS, with superior performance observed in posterior circulation stroke patients—though the underlying mechanism requires further investigation.

Admission FBG was significantly correlated with 90-day unfavorable outcomes in univariate analysis but lost independent prognostic value after multivariate adjustment and was not incorporated into the final SPRIN₂D nomogram. Multiple prior investigations have validated admission hyperglycemia as an unfavorable prognostic indicator for AIS, which exacerbates cerebrovascular injury via vascular dysfunction and oxidative stress-triggered pathological processes (48, 49). Accumulated clinical evidence further demonstrates that elevated admission glucose is closely associated with aggravated neurological impairment, increased in-hospital mortality and a higher risk of END in heterogeneous AIS populations (50, 51). Nevertheless, this independent prognostic effect of admission hyperglycemia was absent in our MIS cohort. Such inconsistency is likely attributable to population characteristics: all included patients suffered only mild neurological deficits and fewer systemic comorbidities, thereby attenuating the independent prognostic potency of admission FBG. Further large-scale multicenter prospective trials adopting standardized blood glucose testing protocols are warranted to verify our results.

Clinically, the SPRIN₂D nomogram was developed for risk stratification rather than treatment decision-making. It should not be used to determine eligibility for IVT, which must be prescribed in accordance with established clinical criteria and current guidelines. Instead, identifying patients at high risk of unfavorable outcomes enables clinicians to conduct closer neurological monitoring, formulate targeted rehabilitation plans, and deliver individualized post-discharge care. Notably, patients in the low-risk group should receive standard care, rehabilitation, and secondary prevention as recommended by guidelines. Collectively, this nomogram mainly serves as a tool for early risk evaluation and optimization of post-thrombolysis management.

Compared with existing prediction models, the SPRIN₂D nomogram has several strengths. Firstly, the model was derived from a large, multicenter cohort, enhancing the generalizability of the findings. Secondly, all independent predictors can be quickly obtained before thrombolysis, facilitating individualized follow-up planning. Thirdly, these variables are widely available in primary hospitals, enabling broad implementation of the nomogram. Fourthly, the study enrolled both disabled and non-disabled patients with MIS receiving IVT therapy, ensuring greater representativeness and providing a more comprehensive reference for MIS prognosis.

Nonetheless, this study has certain limitations. (1) Participating hospitals in the TRAIS study were municipal hospitals with relatively advanced medical resources, potentially introducing selection bias that may limit the generalizability of the results to resource-limited settings. (2) The high prevalence of mRS scores of 0–1 at 90 days among MIS patients may lead to “ceiling effects,” potentially influencing outcome assessment (52). Future studies based on larger multicenter cohorts may adopt alternative outcome definitions, such as mRS 1–6 and ordinal outcome models, to evaluate their incremental prognostic value and further improve model performance. (3) Laboratory variables may fluctuate during hospitalization, and future studies should consider dynamic monitoring to capture temporal changes in these markers. (4) Key clinical metrics including thrombolytic dose and door-to-needle time were not incorporated into the regression analysis, possibly resulting in omitted variable bias and compromised model robustness. Well-designed prospective cohorts adopting standardized data collection protocols are warranted to integrate these covariates and improve the nomogram’s clinical practicability. (5) While clinically meaningful categorization improves the practicality and interpretability of the model, it inevitably leads to information loss from the original continuous variables. Accordingly, larger multicenter cohorts are needed to validate our results. Further studies could also adopt continuous-variable approaches, such as restricted cubic splines, to explore whether these methods can enhance model calibration, discrimination and clinical utility. (6) Independent external validation across geographically and clinically heterogeneous cohorts is required prior to widespread clinical adoption of the SPRIN₂D nomogram.

5 Conclusion

We successfully constructed and validated the SPRIN₂D nomogram, a stable clinical tool designed to predict 90-day unfavorable outcomes in patients with MIS undergoing IVT treatment. As a user-friendly and accessible tool, the SPRIN₂D nomogram addresses the clinical gap caused by the lack of insufficient MIS-specific prognostic models, offering clinicians actionable risk stratification to support individualized follow-up planning. It also contributes to the efficient utilization of medical resources by identifying high-risk patients for closer monitoring and targeted interventions, thereby advancing the clinical management of MIS.

References

• 1.

  MartinSSAdayAWAllenNB. Heart disease and stroke statistics: a report of US and global data from the American Heart Association. Circulation. (2025) 151:e41-e660. doi: 10.1161/CIR.0000000000001303

  • CrossRef
  • Google Scholar

• 2.

  LiXYKongXMYangCHChengZFLvJJGuoHet al. Global, regional, and national burden of ischemic stroke, 1990-2021: an analysis of data from the global burden of disease study 2021. EClinicalMedicine. (2024) 75:102758. doi: 10.1016/j.eclinm.2024.102758,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 3.

  SaberHSaverJL. Distributional validity and prognostic power of the National Institutes of Health stroke scale in US administrative claims data. JAMA Neurol. (2020) 77:606–12. doi: 10.1001/jamaneurol.2019.5061,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 4.

  XiongYGuHZhaoXQYangXWangCWangYLet al. Clinical characteristics and in-hospital outcomes of varying definitions of minor stroke: from a large-scale nation-wide longitudinal registry. Stroke. (2021) 52:1253–8. doi: 10.1161/STROKEAHA.120.031329,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 5.

  RomanoJGSmithEELiangLGardenerHCampo-BustilloIKhatriPet al. Distinct short-term outcomes in patients with mild versus rapidly improving stroke not treated with Thrombolytics. Stroke. (2016) 47:1278–85. doi: 10.1161/STROKEAHA.115.011528,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 6.

  HurfordRLiLLovettNKubiakMKukerWRothwellPMet al. Prognostic value of ‘tissue-based’ definitions of TIA and minor stroke: population-based study. Neurology. (2019) 92:e2455–61. doi: 10.1212/WNL.0000000000007531,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 7.

  SaberHKhatibiKSzederVTateshimaSColbyGPNourMet al. Reperfusion therapy frequency and outcomes in mild ischemic stroke in the United States. Stroke. (2020) 51:3241–9. doi: 10.1161/STROKEAHA.120.030898,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 8.

  XiongYYanRGuHWangSFisherMZhaoXet al. Intravenous thrombolysis in Chinese patients with mild acute ischemic stroke. Ann Transl Med. (2021) 9:767. doi: 10.21037/atm-21-40,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 9.

  LiuLLiZZhouHDuanWHuoXXuWet al. Chinese Stroke Association guidelines for clinical management of ischaemic cerebrovascular diseases: executive summary and 2023 update. Stroke Vasc Neurol. (2023) 8:e3. doi: 10.1136/svn-2023-002998,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 10.

  SmithEEFonarowGCReevesMJCoxMOlsonDMHernandezAFet al. Outcomes in mild or rapidly improving stroke not treated with intravenous recombinant tissue-type plasminogen activator: findings from get with the guidelines-stroke. Stroke. (2011) 42:3110–5. doi: 10.1161/STROKEAHA.111.613208,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 11.

  KhatriPConawayMRJohnstonKCAcute Stroke Accurate Prediction Study (ASAP) Investigators. Ninety-day outcome rates of a prospective cohort of consecutive patients with mild ischemic stroke. Stroke. (2012) 43:560–2. doi: 10.1161/STROKEAHA.110.593897,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 12.

  KaindlLFerrariJSykoraM. Minor strokes: addressing the challenges of recanalization therapies and secondary prevention. J Neurol Sci. (2024) 466:123270. doi: 10.1016/j.jns.2024.123270,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 13.

  ChenSDYouJYangXMGuHQHuangXYLiuHet al. Machine learning is an effective method to predict the 90-day prognosis of patients with transient ischemic attack and minor stroke. BMC Med Res Methodol. (2022) 22:195. doi: 10.1186/s12874-022-01672-z,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 14.

  WuLWangAWangXZhaoXWangCLiuLet al. Factors for short-term outcomes in patients with a minor stroke: results from China National Stroke Registry. BMC Neurol. (2015) 15:253. doi: 10.1186/s12883-015-0505-z,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 15.

  SunKLiHDongYCaoLLiDLiJet al. The use of identified hypoxia-related genes to generate models for predicting the prognosis of cerebral ischemia–reperfusion injury and developing treatment strategies. Mol Neurobiol. (2024) 62:3098–3124. doi: 10.1007/s12035-024-04433-9,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 16.

  ZhouYFengYXinNLuJXuX. Assessing stroke recurrence risk by using a lipoprotein-associated phospholipase A2 and platelet count-based nomogram. Mol Neurobiol. (2024) 62:2835–45. doi: 10.1007/s12035-024-04439-3,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 17.

  JinHPengQLiMSunSZhouJHuJet al. Supra-Blan2 t score as a multisystem-based risk score to predict poor 3-month outcome in acute ischemic stroke patients with intravenous thrombolysis. CNS Neurosci Ther. (2024) 30:e14381. doi: 10.1111/cns.14381,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 18.

  YinRYinLLiLSilva-NashJTanJPanZet al. Hypertension in China: burdens, guidelines and policy responses: a state-of-the-art review. J Hum Hypertens. (2022) 36:126–34. doi: 10.1038/s41371-021-00570-z,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 19.

  CatapanoALGrahamIDe BackerGWiklundOChapmanMJDrexelHet al. 2016 ESC/EAS guidelines for the management of dyslipidaemias. Eur Heart J. (2016) 37:2999–3058. doi: 10.1093/eurheartj/ehw272,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 20.

  KDIGO 2024. Clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int. (2024) 105:S117-S314. doi: 10.1016/j.kint.2023.10.018

  • CrossRef
  • Google Scholar

• 21.

  QiuSXuY. Guidelines for acute ischemic stroke treatment. Neurosci Bull. (2020) 36:1229–32. doi: 10.1007/s12264-020-00534-2,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 22.

  FlintACFaigelesBSCullenSPKamelHRaoVAGuptaRet al. THRIVE score predicts ischemic stroke outcomes and thrombolytic hemorrhage risk in VISTA. Stroke. (2013) 44:3365–9. doi: 10.1161/STROKEAHA.113.002794,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 23.

  KentDMSelkerHPRuthazerRBluhmkiEHackeW. The stroke-thrombolytic predictive instrument: a predictive instrument for intravenous thrombolysis in acute ischemic stroke. Stroke. (2006) 37:2957–62. doi: 10.1161/01.STR.0000249054.96644.c6,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 24.

  LeiZLiSFengHLaiYZhouYLiCet al. Prognostic nomogram for patients with minor stroke and transient ischaemic attack. Postgrad Med J. (2021) 97:644–9. doi: 10.1136/postgradmedj-2020-137680,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 25.

  AsdaghiNWangKCiliberti-VargasMAGutierrezCMKochSGardenerHet al. Predictors of thrombolysis Administration in Mild Stroke. Stroke. (2018) 49:638–45. doi: 10.1161/STROKEAHA.117.019341,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 26.

  BalchMHHNimjeeSMRinkCHannawiY. Beyond the brain: the systemic pathophysiological response to acute ischemic stroke. J Stroke. (2020) 22:159–72. doi: 10.5853/jos.2019.02978,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 27.

  WangJ. Construction and evaluation of a nomogram prediction model for aspiration pneumonia in patients with acute ischemic stroke. Heliyon. (2023) 9:e22048. doi: 10.1016/j.heliyon.2023.e22048

  • CrossRef
  • Google Scholar

• 28.

  CucchiaraBGeorgeDKKasnerSE. Disability after minor stroke and TIA: a secondary analysis of the SOCRATES trial. Neurology. (2019) 93:e708-e716. doi: 10.1212/WNL.0000000000007936

  • CrossRef
  • Google Scholar

• 29.

  LeiraECLudwigBRGurolMETornerJCAdamsHP. The types of neurological deficits might not justify withholding treatment in patients with low total National Institutes of Health stroke scale scores. Stroke. (2012) 43:782–6. doi: 10.1161/STROKEAHA.111.620674,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 30.

  LeiZLiSFengHWuXHuSLiJet al. Effects of intravenous rtPA in patients with minor stroke. Ann Med. (2024) 56:2304653. doi: 10.1080/07853890.2024.2304653,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 31.

  BelandBBalaFGaneshA. Thrombolysis for acute ischemic stroke in patients with premorbid disability: a Meta-analysis. Stroke. (2022) 53:3055–63. doi: 10.1161/STROKEAHA.121.038374,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 32.

  ChouMLBabamaleAOWalkerTLCognasseFBlumDBurnoufT. Blood-brain crosstalk: the roles of neutrophils, platelets, and neutrophil extracellular traps in neuropathologies. Trends Neurosci. (2023) 46:764–79. doi: 10.1016/j.tins.2023.06.005,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 33.

  Perez-de-PuigIMiró-MurFFerrer-FerrerMGelpiEPedragosaJJusticiaCet al. Neutrophil recruitment to the brain in mouse and human ischemic stroke. Acta Neuropathol. (2015) 129:239–57. doi: 10.1007/s00401-014-1381-0,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 34.

  MaestriniIStrbianDGautierSHaapaniemiEMoulinSSairanenTet al. Higher neutrophil counts before thrombolysis for cerebral ischemia predict worse outcomes. Neurology. (2015) 85:1408–16. doi: 10.1212/WNL.0000000000002029,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 35.

  TianTWangLXuJJiaYXueKHuangSet al. Prediction of early neurological deterioration in acute ischemic stroke patients treated with intravenous thrombolysis. J Cereb Blood Flow Metab. (2023) 43:2049–59. doi: 10.1177/0271678X231200117,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 36.

  ZhuBPanYJingJMengXZhaoXLiuLet al. Neutrophil counts, neutrophil ratio, and new stroke in minor ischemic stroke or TIA. Neurology. (2018) 90:e1870–8. doi: 10.1212/WNL.0000000000005554,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 37.

  LeviMvan der PollT. Inflammation and coagulation. Crit Care Med. (2010) 38:S26–34. doi: 10.1097/CCM.0b013e3181c98d21,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 38.

  YouSHanQDongXZhongCDuHSunYet al. Prognostic significance of international normalised ratio and prothrombin time in Chinese acute ischaemic stroke patients. Postgrad Med J. (2023) 99:333–9. doi: 10.1136/postgradmedj-2021-141204,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 39.

  JinHBiRZhouYXiaoQLiMSunSet al. CNS-LAND score: predicting early neurological deterioration after intravenous thrombolysis based on systemic responses and injury. Front Neurol. (2023) 14:14. doi: 10.3389/fneur.2023.1266526,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 40.

  ScheitzJFTurcGKujalaLPolymerisAAHeldnerMRZonneveldTPet al. Intracerebral hemorrhage and outcome after thrombolysis in stroke patients using selective serotonin-reuptake inhibitors. Stroke. (2017) 48:3239–44. doi: 10.1161/STROKEAHA.117.018377,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 41.

  LeeMSaverJLChangKHLiaoHWChangSCOvbiageleB. Low glomerular filtration rate and risk of stroke: meta-analysis. BMJ. (2010) 341:c4249. doi: 10.1136/bmj.c4249,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 42.

  ZambergIAssouline-ReinmannMCarreraESoodMMSozioSMMartinPYet al. Epidemiology, thrombolytic management, and outcomes of acute stroke among patients with chronic kidney disease: a systematic review and meta-analysis. Nephrol Dial Transplant. (2022) 37:1289–301. doi: 10.1093/ndt/gfab197,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 43.

  HrafnkelsdóttirTOttossonPGudnasonTSamuelssonOJernS. Impaired endothelial release of tissue-type plasminogen activator in patients with chronic kidney disease and hypertension. Hypertension. (2004) 44:300–4. doi: 10.1161/01.HYP.0000137380.91476.fb,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 44.

  MolshatzkiNOrionDTsabariRSchwammenthalYMerzeliakOToashiMet al. Chronic kidney disease in patients with acute intracerebral hemorrhage: association with large hematoma volume and poor outcome. Cerebrovasc Dis. (2011) 31:271–7. doi: 10.1159/000322155,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 45.

  VilionskisAGaigalaiteVSalaseviciusLJatuzisD. Association between systolic blood pressure parameters and unexplained early neurological deterioration (UnND) in acute ischemic stroke patients treated with mechanical thrombectomy. Ther Adv Neurol Disord. (2022) 15:17562864221093524. doi: 10.1177/17562864221093524,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 46.

  ZhuLXieJXieQXuYChenYLiYet al. Association of Early Blood Pressure Levels and Outcomes in ischemic stroke treated with intravenous thrombolysis: a prospective cohort study. CNS Neuroscience & Therapeutics. (2025) 31:e70318. doi: 10.1111/cns.70318

  • CrossRef
  • Google Scholar

• 47.

  WXMJMTDtGLRCXet al. Associations of early systolic blood pressure control and outcome after thrombolysis-eligible acute ischemic stroke: results from the ENCHANTED study. Stroke. (2022) 53:779–87. doi: 10.1161/STROKEAHA.121.034580

  • CrossRef
  • Google Scholar

• 48.

  SarkarRMeinbergEGStanleyJCGordonDWebbRC. Nitric oxide reversibly inhibits the migration of cultured vascular smooth muscle cells. Circ Res. (1996) 78:225–30.

  • Google Scholar

• 49.

  MaidaCDDaidoneMPacinellaGNorritoRLPintoATuttolomondoA. Diabetes and ischemic stroke: an old and new relationship an overview of the close interaction between these diseases. Int J Mol Sci. (2022) 23:4. doi: 10.3390/ijms23042397,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 50.

  CapesSEHuntDMalmbergKPathakPGersteinHC. Stress hyperglycemia and prognosis of stroke in nondiabetic and diabetic patients: a systematic overview. Stroke. (2001) 32:2426–32. doi: 10.1161/hs1001.096194,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 51.

  KruytNDBiesselsGJDevriesJHRoosYB. Hyperglycemia in acute ischemic stroke: pathophysiology and clinical management. Nat Rev Neurol. (2010) 6:145–55. doi: 10.1038/nrneurol.2009.231,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 52.

  WangXDongYDongQWangD. Should patients with minor strokes be given thrombolytics?Stroke Vasc Neurol. (2024) 10:418–21. doi: 10.1136/svn-2024-003451,

  • Pubmed Abstract
  • CrossRef
  • Google Scholar