Abstract
Background
Studies have been conducted to explore the potential predictive indicators of unfavorable outcomes in patients with acute ischemic stroke (AIS) caused by large vessel occlusion (LVO). However, few studies have proposed a comprehensive predictive model combined with clinical baseline data and ancillary examination before surgery.
Method
In a retrospective study, we collected data on 823 patients with AIS-LVO who had undergone endovascular therapy (EVT); 562 patients who achieved successful revascularization with complete clinical and prognostic information were incorporated into the study. Those patients with a 90-day modified Rankin Scale (mRS) score of 0–2 were defined as having a favorable outcome, while a score of 3–6 represented an unfavorable outcome or futile reperfusion. To build up a predictive model, we applied multivariate logistic regression stepwise backward selection to decide which factors are supposed to be the components of the predictive model. Final model validity was testified by the variance inflation factor test and the Hosmer-Lemeshow (HL) goodness of fit test. The ultimate efficacy was supported by an area under the curve (AUC) value in both training groups and validation groups.
Results
562 patients were enrolled in our study and divided into the training group and verification group in a ratio of 7:3. Factors of baseline data with P<0.1 in univariate logistic regression analysis were enrolled as the potential risk variables to conduct stepwise backward selection. The model was constructed by eight variables; higher mRS score (adjusted OR (aOR) 93.64, 95% CI 12.05 to 727.82, P<0.01), age >80 years (aOR 91.11, 95% CI 1.36 to 6116.36, P<0.05), National Institutes of Health Stroke Scale (NIHSS) >14 (aOR 0.15, 95% CI 0.02 to 0.99, P<0.05), operation history (aOR 8.13, 95% CI 1.32 to 50.20, P<0.05), creatinine (aOR 1.10, 95% CI 1.04 to 1.17, P<0.01), and neutrophil count (aOR 1.07, 95% CI 1.01 to 1.13, P<0.05) were associated with poor outcomes.
Conclusion
We established an estimation model for invalid reperfusion in AIS-LVO patients and constructed the nomogram for individualized predictions. The AUC of the training group and validation group were both 0.96, with excellent HL and decision curve analysis, presenting excellent clinical prediction efficiency and application potential.
Keywords: Angiography, Thrombectomy, Angioplasty, Catheter, Intracranial Thrombosis
WHAT IS ALREADY KNOWN ON THIS TOPIC
In ischemic stroke, mechanical thrombectomy achieves high reperfusion rates, yet about 50% of patients fail to regain functional independence—a scenario termed "ineffective recanalization." Prior studies have identified age, admission NIHSS score, neutrophil and etc as risk factors for ineffective recanalization, but no comprehensive clinical prediction models exist for this condition.
WHAT THIS STUDY ADDS
Based on the baseline characters of patients and the auxiliary test indicators of emergency admission, we developed a clinical prediction model for ineffective recanalization after thrombectomy, with excellent discrimination, calibration, and clinical utility. We also identified novel potential predictors, including BNP, globulin, and neutrophil percentage, which may be related to inflammation, coagulation function, and vasculopathy severity.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
We construct a nomogram of ineffective recanalization, which is beneficial to more accurately assess the prognostic risk of patients before surgery, so as to develop a more personalized treatment plan and allocate medical resources more rationally.
Introduction
Stroke has long been regarded as one of the most lethal medical conditions in the world; its morbidity and mortality are maintained at a high level, causing numerous disabilities and deaths and serious social and economic load.1 Acute ischemic stroke (AIS) accounts for about 62.4% of all incident strokes; about one third of AIS patients suffer from large vessel occlusion (LVO), which develops more serious clinical symptoms and higher rates of disability and mortality compared with non-LVO strokes.2 Therapies such as endovascular thrombectomy (EVT) or intravenous thrombolysis have been applied to attenuate symptoms in the therapeutic time window.3 With advances in technology and materials, the average rate of recanalization on angiographic imaging in acute LVO patients has reached nearly 90%.4 However, despite successful recanalization of the occluded vessel, nearly half of all patients fail to achieve expected clinical improvement and still have functional sequelae,5 defined as ‘futile reperfusion’.6 7 To investigate the potential influencing factors responsible for futile reperfusion, further research is necessary to explore more related pre-warning clues among patients’ clinical characteristics.
Certain clinical features such as baseline and ancillary data at admission are suspected to be associated with patient outcome. So far, higher age,8 National Institutes of Health Stroke Scale (NIHSS) scores,9 hyperglycemia,10 dyslipidemia,11 and abnormal systemic inflammatory response12 have been reported to be associated with unfavorable outcomes after successful recanalization. In addition, preoperative auxiliary test indexes may also influence the prognosis of patients with AIS-LVO. For example, inflammation factors such as white blood cells, neutrophils, neutrophil to lymphocyte ratio,13 interleukin-6, and chemokine ligand 17/thymus and activation-regulated chemokine levels14 have been identified as independent predictors of futile reperfusion in previous studies. At present, most of the prediction models focus on predicting the poor prognosis of patients after thrombectomy, while a few models specifically predict ineffective recanalization and explore the influencing factors of ineffective recanalization. However, doctors still need a comprehensive model to evaluate patient outcomes individually and assist clinical decision-making precisely. In our study, we continuously adopted the baseline and ancillary test data as the most convenient and effective indicators reflecting different physical conditions. Patients with anterior circulation LVO were chosen as research subjects to analyze the potential association between those elements and the 90-day outcomes. Ultimately, we aimed to develop a reliable model for accurate prognosis estimation to guide clinical therapeutic strategies.
Methods
Study strategy and data collection
This is a retrospective analysis of consecutive AIS patients with LVO who underwent EVT from March 2016 to August 2023 in the Stroke Center Department of the Second Affiliated Hospital of Zhejiang University School of Medicine. Patients with the following criteria were included in our cohort study: (1) patients >18 years of age; (2) patients with occlusion of the large artery identified through digital subtraction angiography (DSA); (3) patients with a modified Rankin Scale (mRS) score at admission ≥1; (4) EVT was administered within 6 hours after symptom onset or mechanical thrombectomy within 6–24 hours after symptom onset; (5) DSA showed full recanalization after EVT with modified Treatment In Cerebral Infarction (mTICI) score >2B/3. Clinical data were collected from the electronic medical records system in the hospital, including baseline demographic data, medical history, stroke characteristics, procedural variables, treatment variables, and hospitalization complications. Baseline characteristics encompassed age, sex, National Institutes of Health Stroke Scale (NIHSS) score, and mRS score. At admission, ancillary data included: blood routine examination such as white blood cell count, neutrophil count, eosinophil count, monocyte count, lymphocyte count, and platelet count; biochemical markers such as lactate dehydrogenase (LDH), glucose level, N-terminal pro-brain natriuretic peptide (NT-proBNP), and cardiac troponin T; and other coagulation indexes such as D-dimer, international normalized ratio, prothrombin time, prothrombin activity, and so on. Occlusion sites were identified through CT angiography, magnetic resonance angiography, and/or cerebral DSA reports, while the reperfusion situation was reassessed through DSA during the surgery after mechanical thrombectomy. All the image data were evaluated by two experienced radiologists blinded to the clinical characteristics. At discharge, the NIHSS score, mRS, complications including symptomatic intracerebral hemorrhage (sICH), and hemorrhage transformation (HT) were recorded.
Follow-up and outcomes
The 90-days outcome data were collected by a stroke neurologist during routine telephone follow-up visits at 90±14 days after stroke, including NIHSS and mRS. The primary outcome was defined as good functional outcome at 90 days, with an mRS score of 0–3.9 15 As regards complications, HT referred to a secondary intracranial bleed after vessel reperfusion in the ischemic area,16 while sICH was judged by any intracranial hemorrhage with an increase in the NIHSS score of ≥4 from baseline at admission within 7 days after EVT, according to the ECASS (European-Australasian Acute Stroke Study) II criteria.16
Statistical analysis
Patients were dichotomized into a favorable outcomes group with an mRS score at 3 months after EVT of 0–3, and an unfavorable outcomes group with an mRS score of 4–6. Outcome data of patients who were lost to follow-up were excluded, and the total missing data were <20%. A total of 562 patients were included in our cohort groups, and were then divided into a training group and a verification group in a ratio of 7:3. For partially lacking datasets in clinical characteristic data, we applied multiple interpolation to fill in the missing values.17 Characteristics were summarized into the proportions in categorical variables and mean±SD or median (25–75th percentile) in quantitative variables, as appropriate. Two-sample t-tests or Mann-Whitney U test were used to compare the dichotomous in continuous quantitative variables, while χ2 tests or Fisher’s exact test were executed for categorical variables. Through ANOVA (analysis of variance) and univariate logistic regression, potential risk factors were filtered out to be associated with outcomes of AIS patients (P<0.05); then those potential predictors were corrected by confounding factors which were filtered from clinical characters through multivariate logistic regression (P<0.05). Results were displayed as adjusted odds ratios (aOR) combined with 95% confidence intervals (95% CI). P values <0.05 were considered statistically significant. All reported P values were two-sided. The statistical analyses were performed using IBM Statistical Package for the Social Sciences (SPSS) Statistics for Windows version 26 or R language 4.0.5.
Results
Baseline characteristics of the cohort
The patients were divided into two groups—a training group and a verification group—in the ratio of 7 to 3 with no significant difference (online supplemental table 1). Among 393 patients in the training group, 223 patients exhibited favorable outcomes while 170 patients had a valid reperfusion at 3 months after EVT (table 1). Patients with favorable outcomes were younger than those with unfavorable outcomes on average (64.22 vs 72.53 years, P<0.01). The proportion of NIHSS >14 and general anesthesia were significantly higher in the unfavorable group (72.48 vs 48.15, P<0.001; 52.71 vs 37.78, P=0.05).
Table 1. Analysis of clinical characteristics between effective perfusion and ineffective recanalization of the training group.
| Variables | Total (n=393) | Favorable group (n=258) | Unfavorable group (n=135) | P value |
|---|---|---|---|---|
| Male, n (%) | 229 (58.27) | 79 (58.52) | 150 (58.14) | 0.942 |
| Age, mean±SD, years | 69.68±12.96 | 64.22±13.02 | 72.53±12.00 | <0.001** |
| SBP, mean±SD, mmHg | 145.59±25.09 | 142.88±23.88 | 147.02±25.63 | 0.121 |
| DBP, mean±SD, mmHg | 81.38±14.54 | 82.01±15.08 | 81.05±14.27 | 0.539 |
| General anesthesia, n (%) | 187 (47.58) | 51 (37.78) | 136 (52.71) | 0.005* |
| mRS score at discharge, M (Q₁, Q₃) | 4.00 (3.00, 5.00) | 3.00 (1.00, 3.00) | 5.00 (4.00, 5.00) | <0.001** |
| Procedural variables | ||||
| OTA, M (Q₁, Q₃) | 272.00 (173.25, 373.00) | 274.00 (179.00, 383.50) | 271.00 (161.00, 370.00) | 0.478 |
| OTI, M (Q₁, Q₃) | 223.50 (119.75, 325.00) | 227.50 (129.75, 320.75) | 219.50 (115.00, 325.00) | 0.842 |
| ITP, M (Q₁, Q₃) | 103.00 (77.00, 135.00) | 96.50 (75.00, 119.89) | 107.00 (80.00, 143.50) | 0.028* |
| TPR, M (Q₁, Q₃) | 70.00 (45.00, 104.00) | 60.00 (40.00, 90.00) | 75.00 (50.00, 108.88) | 0.007* |
| OTR, M (Q₁, Q₃) | 400.00 (309.00, 510.00) | 392.00 (300.00, 494.75) | 411.87 (318.00, 520.00) | 0.197 |
| OTP, M (Q₁, Q₃) | 325.00 (231.00, 419.73) | 325.00 (231.40, 405.00) | 325.00 (233.00, 424.00) | 0.599 |
| BMI, M (Q₁, Q₃) | 23.03 (20.76, 25.39) | 23.44 (21.53, 25.48) | 22.59 (20.43, 25.39) | 0.054 |
| Medical history | ||||
| IVT, n (%) | 238 (60.56) | 87 (64.44) | 151 (58.53) | 0.254 |
| History of cerebral infarction, n (%) | 79 (20.10) | 22 (16.30) | 57 (22.09) | 0.173 |
| Hyperlipidemia, n (%) | 15 (3.82) | 6 (4.44) | 9 (3.49) | 0.639 |
| Hypertension, n (%) | 247 (62.85) | 71 (52.59) | 176 (68.22) | 0.002* |
| CHD, n (%) | 47 (11.96) | 10 (7.41) | 37 (14.34) | 0.044* |
| Diabetes, n (%) | 75 (19.08) | 14 (10.37) | 61 (23.64) | 0.001** |
| Atrial fibrillation, n (%) | 148 (37.66) | 44 (32.59) | 104 (40.31) | 0.134 |
| Pre-antiplatelet, n (%) | 43 (10.94) | 16 (11.85) | 27 (10.47) | 0.676 |
| Pre-anticoagulation, n (%) | 52 (13.23) | 17 (12.59) | 35 (13.57) | 0.787 |
| Operation history, n (%) | 129 (32.82) | 37 (27.41) | 92 (35.66) | 0.098 |
| Stroke characteristics | ||||
| *Rescue therapy, n (%) | 78 (19.85) | 26 (19.26) | 52 (20.16) | 0.833 |
| TOAST classification, n (%) | 0.167 | |||
| LAA | 118 (45.56) | 54 (52.43) | 64 (41.03) | |
| CE | 124 (47.88) | 42 (40.78) | 82 (52.56) | |
| Others or undetermined etiology | 17 (6.56) | 7 (6.80) | 10 (6.41) | |
| Site of occlusion, n (%) | 0.001** | |||
| ICA | 67 (17.05) | 20 (14.81) | 47 (18.22) | |
| MCA | 208 (52.93) | 87 (64.44) | 121 (46.90) | |
| ICA-MCA | 45 (11.45) | 13 (9.63) | 32 (12.40) | |
| BA | 46 (11.70) | 5 (3.70) | 41 (15.89) | |
| Others | 27 (6.87) | 10 (7.41) | 17 (6.59) | |
| Multistage thrombus, n (%) | 74 (18.83) | 23 (17.04) | 51 (19.77) | 0.511 |
| Distal escape, n (%) | 25 (6.36) | 12 (8.89) | 13 (5.04) | 0.138 |
| Tirofiban treatment, n (%) | 69 (17.56) | 29 (21.48) | 40 (15.50) | 0.139 |
| Number of passes >3 times, n (%) | 56 (14.25) | 10 (7.41) | 46 (17.83) | 0.005* |
| NIHSS >14, n (%) | 252 (64.12) | 65 (48.15) | 187 (72.48) | <0.001** |
| Postoperative complications | ||||
| HT, n (%) | 122 (31.04) | 22 (16.30) | 100 (38.76) | <0.001** |
| sICH, n (%) | 81 (20.61) | 0 (0.00) | 81 (31.40) | <0.001** |
| Pulmonary infection, n (%) | 178 (45.29) | 38 (28.15) | 140 (54.26) | <0.001** |
| Liver dysfunction, n (%) | 54 (13.74) | 15 (11.11) | 39 (15.12) | 0.273 |
| Respiratory failure, n (%) | 28 (7.12) | 1 (0.74) | 27 (10.47) | <0.001** |
Qualitative variables are n (%), mean±SD, or median (IQR).
*P<0.05.
**P<0.001.
Rescue therapy included balloon angioplasty and permanent stenting.
BA, basilar artery; BMI, body mass index; CE, cardioembolism; CHD, coronary heart disease; DBP, diastolic blood pressure; HT, hemorrhagic transformation; ICA, Internal carotid artery; ITP, time from image to puncture; IVT, intravenous thrombolysis; LAA, large artery atherosclerosis; M, monocytes; MCA, middle cerebral artery; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; OTA, time from onset to admission; OTI, time from onset to image; OTP, time from onset to groin puncture; OTR, time from onset to reperfusion; SBP, systolic blood pressure; sICH, symptomatic intracranial hemorrhage; TOAST, Trial of Org 10172 in Acute Stroke Treatment; TPR, time from puncture to reperfusion.
Many other complications appeared to be significantly different in the good outcomes and bad outcomes groups, such as pulmonary infection (54.26% vs 28.16%, P<0.01) and diabetes (23.64% vs 10.37%, P<0.01). Patients with bad outcomes seemed to have higher mean blood pressure at emergency presentation (68.22% vs 52.59%, P<0.01). The distribution of occlusion sites differed remarkably in the two groups, and the time of thrombectomy could be a potential influencing element on outcomes. Hemorrhage secondary to EVT was more often seen in the unfavorable outcome groups including HT (median 38.76% vs 16.3%, P<0.01) and sICH (median 31.4% vs 0%, P<0.01).
Characters of ancillary examination at emergency of cohort study
There were significant differences in serum biochemical detection between the two groups (table 2). Compared with the poor prognosis group, the patients with a good prognosis had lower BNP (847.0 vs 337.5 pg/mL, P<0.01), creatine kinase-MB (CKMB) (P<0.05), LDH (222.0 vs 210.5 U/L, P<0.01), urea (6.29 vs 5.64 mmo1/L, P<0.01), glucose (7.55 vs 6.83 mmo1/L, P<0.01), and globulin (28.7 vs 27.1 g/L, P<0.05). Notably, CK and troponin may rise even higher in patients with good outcomes (P<0.01), and also higher levels of albumin/globulin ratio (AGR) may predict effective reperfusion with better 90-days outcome (1.23 vs 1.34, P<0.01).
Table 2. Analysis of peripheral blood auxiliary test indexes between effective perfusion and ineffective recanalization of the training group.
| Variables | Total (n=393) | Favorable group (n=135) | Unfavorable group (n=258) | P value |
|---|---|---|---|---|
| Coagulation function index | ||||
| INR, M (Q₁, Q₃) | 1.04 (0.98, 1.11) | 1.03 (0.96, 1.10) | 1.04 (0.98, 1.11) | 0.277 |
| APTT, M (Q₁, Q₃) | 33.80 (31.20, 37.20) | 34.20 (30.95, 37.00) | 33.60 (31.20, 37.27) | 0.918 |
| TT, M (Q₁, Q₃) | 17.20 (16.40, 18.40) | 17.20 (16.40, 18.50) | 17.20 (16.40, 18.28) | 0.554 |
| PT, M (Q₁, Q₃) | 13.50 (12.90, 14.30) | 13.40 (12.80, 14.10) | 13.50 (13.00, 14.30) | 0.275 |
| PTA, M (Q₁, Q₃) | 93.00 (83.00, 103.00) | 95.00 (86.00, 106.00) | 93.00 (82.25, 102.75) | 0.219 |
| FIB, M (Q₁, Q₃) | 3.13 (2.63, 3.75) | 2.96 (2.55, 3.55) | 3.19 (2.67, 3.83) | 0.013* |
| Serum biochemical index | ||||
| BNP, M (Q₁, Q₃) | 665.77 (188.00, 1496.00) | 337.50 (116.25, 1097.50) | 847.00 (268.00, 1773.50) | <0.001** |
| CK, M (Q₁, Q₃) | 83.50 (57.25, 138.50) | 96.00 (64.00, 143.00) | 78.00 (53.00, 127.00) | 0.038* |
| CKMB, M (Q₁, Q₃) | 15.00 (11.00, 19.00) | 14.00 (11.00, 18.00) | 15.00 (12.00, 19.00) | 0.026* |
| Troponin, M (Q₁, Q₃) | 0.01 (0.01, 0.02) | 0.01 (0.01, 0.02) | 0.01 (0.01, 0.03) | <0.001** |
| LDH, M (Q₁, Q₃) | 217.00 (187.00, 251.50) | 210.50 (181.25, 241.25) | 222.00 (193.00, 260.00) | 0.008* |
| Urea, M (Q₁, Q₃) | 6.06 (5.04, 7.55) | 5.64 (4.78, 7.10) | 6.29 (5.20, 7.66) | 0.001* |
| Creatinine, M (Q₁, Q₃) | 70.00 (58.00, 86.00) | 69.00 (58.00, 80.50) | 71.00 (58.25, 89.00) | 0.135 |
| Glucose, M (Q₁, Q₃) | 7.22 (6.19, 8.46) | 6.83 (5.92, 7.68) | 7.55 (6.44, 9.08) | <0.001** |
| Albumin, M (Q₁, Q₃) | 35.80 (33.40, 38.70) | 36.00 (34.02, 39.00) | 35.70 (32.75, 38.60) | 0.102 |
| Globulin, M (Q₁, Q₃) | 28.20 (25.10, 31.40) | 27.05 (24.52, 29.98) | 28.70 (25.30, 32.00) | 0.019* |
| AGR, M (Q₁, Q₃) | 1.28 (1.14, 1.44) | 1.34 (1.18, 1.48) | 1.23 (1.11, 1.40) | <0.001** |
| Inflammatory index | ||||
| W, M (Q₁, Q₃) | 8.50 (6.50, 11.00) | 8.00 (6.40, 10.80) | 8.80 (6.60, 11.20) | 0.221 |
| N, M (Q₁, Q₃) | 6.40 (4.36, 9.11) | 5.70 (4.06, 8.68) | 6.66 (4.51, 9.35) | 0.095 |
| M, M (Q₁, Q₃) | 0.42 (0.32, 0.56) | 0.41 (0.30, 0.57) | 0.44 (0.34, 0.56) | 0.383 |
| L, M (Q₁, Q₃) | 1.30 (0.92, 1.75) | 1.31 (0.98, 1.78) | 1.29 (0.86, 1.73) | 0.188 |
| PLT, M (Q₁, Q₃) | 183.00 (151.00, 221.00) | 190.00 (159.00, 223.50) | 180.00 (148.00, 218.00) | 0.35 |
| TBT, M (Q₁, Q₃) | 0.18 (0.15, 0.22) | 0.19 (0.16, 0.23) | 0.18 (0.15, 0.22) | 0.233 |
| N%, M (Q₁, Q₃) | 77.61 (68.27, 85.71) | 76.25 (67.71, 84.34) | 78.85 (68.45, 86.25) | 0.084 |
| M%, M (Q₁, Q₃) | 5.21 (3.72, 6.67) | 5.29 (3.62, 6.54) | 5.18 (3.78, 6.72) | 0.915 |
| L%, M (Q₁, Q₃) | 15.36 (9.51, 23.25) | 17.29 (10.41, 23.87) | 14.68 (8.94, 23.16) | 0.105 |
| NLR, M (Q₁, Q₃) | 4.91 (2.95, 8.97) | 4.42 (2.94, 8.20) | 5.29 (2.97, 9.64) | 0.103 |
| PWR, M (Q₁, Q₃) | 21.53 (16.61, 28.52) | 22.26 (17.16, 29.65) | 21.21 (16.46, 27.74) | 0.305 |
| PLR, M (Q₁, Q₃) | 141.18 (100.00, 205.19) | 137.24 (97.58, 192.33) | 148.10 (101.37, 214.15) | 0.173 |
| PNR, M (Q₁, Q₃) | 27.80 (20.27, 42.29) | 29.36 (21.13, 43.63) | 27.19 (20.05, 41.86) | 0.249 |
| PMR, M (Q₁, Q₃) | 425.00 (311.87, 579.66) | 438.93 (321.51, 571.93) | 422.22 (307.14, 581.25) | 0.599 |
| LMR, M (Q₁, Q₃) | 3.06 (2.07, 4.52) | 3.16 (2.31, 4.82) | 3.00 (1.89, 4.32) | 0.064 |
| MNR, M (Q₁, Q₃) | 0.07 (0.04, 0.10) | 0.07 (0.04, 0.10) | 0.07 (0.04, 0.10) | 0.725 |
| SSI, M (Q₁, Q₃) | 920.10 (497.09, 1677.93) | 859.62 (471.09, 1532.07) | 963.84 (508.77, 1750.65) | 0.136 |
| NLR vs. PLT, M (Q₁, Q₃) | 2.76 (1.53, 5.40) | 2.54 (1.45, 4.58) | 2.85 (1.60, 5.62) | 0.205 |
Values are median (IQR).
*P<0.05.
**P<0.001.
AGR, albumin/globulin ratio ; APTT, activated partial thromboplastin time; BNP, brain natriuretic peptide; CK, creatine kinase; CKMB, creatine kinase-MB; FIB, Fibrinogen; INR, international normalized ratio; L, lymphocytes; LDH, lactate dehydrogenase; LMR, lymphocyte to monocyte ratio; M, monocytes; MNR, monocyte to neutrophil ratio; N, neutrophils; NLR, neutrophil to lymphocyte ratio; PLR, platelet to lymphocyte ratio; PLT, platelet; PMR, platelet to monocyte ratio; PNR, platelet to neutrophil ratio; P/PLT, platelets; PT, Prothrombin time; PTA, Prothrombin activity; PWR, platelet to white blood cell ratio; SII, systemic immune inflammation index; TBT, thrombocytocrit; TT, Thrombin time; W, white blood cells.
Model building and calibration
Univariate logistic regression analysis of the training group is shown in online supplemental table 2 and online supplemental table 3. Factors of baseline data with P<0.1 in univariate logistic regression analysis were enrolled as the potential risk variables to conduct stepwise backward selection in table 3. The model was constructed by eight variables: higher mRS score (aOR 93.64, 95% CI 12.05 to 727.82, P<0.01), age >80 years (aOR 91.11, 95% CI 1.36 to 6116.36, P<0.05), and NIHSS >14 (aOR 0.15, 95% CI 0.02 to 0.99, P<0.05) were associated with poor outcomes; noticeably, also were operation history (aOR 8.13, 95% CI 1.32 to 50.20, P<0.05), creatinine (aOR 1.10, 95% CI 1.04 to 1.17, P<0.01), and neutrophil count (aOR 1.07, 95% CI 1.01 to 1.13, P<0.05).
Table 3. Multivariate logistic regression analysis of clinical data selected through univariate regression analysis.
| Variables | β | OR (95% CI) | P value |
|---|---|---|---|
| mRS score at admission | 4.54 | 93.64 (12.05 to 727.82) | <0.001 |
| Age >80 years | 4.51 | 91.11 (1.36 to 6116.36) | 0.036 |
| NIHSS >14 | −1.88 | 0.15 (0.02 to 0.99) | 0.048 |
| Operation history | 2.1 | 8.13 (1.32 to 50.20) | 0.024 |
| BNP | 0 | 1.00 (1.00 to 1.00) | 0.083 |
| Creatinine | 0.1 | 1.10 (1.04 to 1.17) | 0.001 |
| Globulin | −0.17 | 0.85 (0.72 to 1.00) | 0.055 |
| N% | 0.06 | 1.07 (1.01 to 1.13) | 0.027 |
Evaluation and visualization of the model through logistic regression
To evaluate model effectiveness and efficiency, we applied the receiver operating characteristic (ROC) curve to verify those models. Whether in the training or the verification groups, the model has a higher diagnostic efficiency with an area under the curve (AUC) value of 0.96 (figure 1). The decision curve analysis (DCA) presented that the model got better net benefits compared with all-treated groups while significantly preceding the none-treated groups both in the training and validation groups. The P value of the model in the Hosmer-Lemeshow test was 0.309>0.05 in the training groups and 0.344>0.05 in the validation groups (figure 2).
Figure 1. The receiver operating characteristic (ROC) curve of the model with its area under the curve (AUC) were both 0.96 in the training groups and external validation groups, indicating the model has good prediction efficacy and stability in both data groups.
Figure 2. Model evaluation in the training groups (A), (B) and valid groups (C), (D). Hosmer-Lemeshow goodness of fit tests were both conducted in training groups (A) and valid groups (C) with P value >0.05. Clinical decision curve analysis (DCA) of prediction model executed in training groups (B) and validation groups (D).
Based on the above multivariate logistic regression model, a nomogram of the clinical prediction model for invalid recanalization after complete reperfusion following thrombectomy in AIS patients was constructed (figure 3). The scores corresponding to the above eight indicators in the nomogram were added up to calculate the total score, and the probability value corresponding to the total score was the predictive value of invalid recanalization after complete reperfusion following thrombectomy.
Figure 3. A nomogram for prognostic prediction of anterior circulation LVO patients after thrombectomy. For any variables, their distributions are reflected by the size of the box. A positive value in category variates like age and any kind of ICH is marked 1, while multistage thrombus counted equal to or up to two sites are marked 2. We can draw an upward vertical line to the ‘Points’ bar to calculate points respectively and sum up the total points, then draw a downward vertical line from the ‘Total points’ line to calculate risk. BNP, brain natriuretic peptide; LVO, large vessel occlusion; ICH, intracranial hemorrhage; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale.
Discussion
We retrospectively collected clinical baseline data of stroke patients after mechanical thrombectomy and assessed ineffective revascularization based on patients’ vascular revascularization status and 3-month postoperative prognosis. Based on the preoperative clinical baseline indexes and auxiliary test indexes through logistic analysis, we constructed the clinical model including mRS score at admission, age >80 years, NIHSS >14, history of previous surgical procedures, BNP, creatinine, globulin, and neutrophilic granulocyte percentage. The AUC of the training group and validation group are both 0.96, with excellent Hosmer-Lemeshow (HL) and DCA curve, presenting excellent clinical prediction efficiency and application potential.
In clinical practice, treatment of AIS focuses on rapid restoration of blood supply to ischemic brain sites and salvage of ischemic penumbra.18 As the advancement of all kinds of retriever stent techniques continues, the successful rate of recanalization keeps increasing, while the total release rate still remains at a relatively low level. To study the potential influencing factors further, we conducted this research and cast light on the potential predictive factors leading to the poor results unequal to the reperfusion. Primary factors like mRS and NIHSS scores were first suspected to reflect the outcomes across different groups. Also, age is a crucial factor that impacts the potency of a full recovery. Similarly, operation times that reflect the previous health status may also affect the stroke prognosis by causing systemic complications. Age, NIHSS score, and ICH have been proved in previous studies to be closely related to patient outcomes. Monocytes, LDH, and multistage thrombus might be first proposed. Circulating monocytes are an essential component of innate immunity. The main physiological functions of circulating monocytes are phagocytosis, removing damaged or senescent cells, and presenting the antigen to T cells. In our study, monocytes were proved to indicate cerebral stroke patient outcomes, and it's worth mentioning that monocytes are reported to be enrolled in the pathophysiological processes of inflammation and thrombosis.19 Monocytes have been proved to participate in atherogenesis through monocyte-endothelial recruitment, adhesion, migration, and pro-inflammatory macrophage foam cell formation in atherosclerotic lesions.20 Not only that, researchers have revealed that monocytes participate in regulating the immune reaction in thrombosis and thrombotic diseases through the formation of monocyte-platelet aggregates (MPAs).21 Platelet aggregation with monocytes could help the polarization of monocytes toward the M1 phenotype in a murine sepsis model, which plays a pro-inflammatory role and impacts thrombosis.22 Circulating MPAs, especially formed by intermediate monocytes, were found to be higher in patients with cardiovascular disease and those combined with other severe complications than in healthy controls,23 and circulating miRNAs of monocytes could be regarded as a biomarker for coronary artery disease diagnosis, prevention, and treatment.24 Given the facts mentioned above, we wondered whether a higher level of circulating monocytes could aggravate severe vessel atherosclerosis or promote inflammatory activation and thrombosis through MPAs in cerebral stroke patients, causing a poorer outcome ultimately. It will take further molecular studies to testify against those suspects in the future.
Creatinine was associated with many complicated biochemistry reactions, and the elevation of creatinine has been proven to affect prognosis in different situations25 through cardiorenal pathways.26 Besides, the inflammatory factor of a high neutrophil proportion also exhibited its predicting efficacy, which stayed consistent with previous reports.13 Understanding the potential influencing factors will be beneficial for guiding clinical treatment decisions and making personalized treatment plans according to the different statuses of patients and their families. Moreover, whether some other specific therapies and nursing treatments targeting and improving those aspects would help the patients get a better recovery still remains elusive. If possible, a prospective study is needed to better testify to the hypothesis and provide more advice to complete clinical treatment guidelines.
There are some limitations in our study. First, our data were collected in a single medical center and the sample number was not adequate enough to avoid deviations. Second, we used a retrospective study as our research strategy, which determined that our data would be affected by some confounding factors, and data accuracy could not be guaranteed completely. Third, in our study, we took the 90-day situation as the final outcome, as done by previous work. However, whether the 90-day outcome could present the final prognosis after ischemic stroke remains unclear, and we believe a follow-up of 6 months, 12 months or longer will further help to clarify the long-term effect of different predictive indicators. Moreover, the follow-up in this study was primarily conducted via telephone, supplemented by face-to-face assessments during outpatient visits. However, telephone follow-ups depend heavily on patients’ subjective reports, potentially misleading the follow-up staff. To mitigate errors, we provided specialized training for the follow-up personnel and created a detailed telephone follow-up process chart. Finally, our image data were not adequate enough to support further studies, and we are still keeping and collecting data to refine our work. Imaging data may be able to reflect more intuitively important information such as collateral circulation and infarct vessels, which may provide new insights into the prediction and explanation of the mechanism of futile recanalization.
Conclusion
The clinical model used here includes the mRS score at admission, age >80 years, NIHSS >14, operation history, BNP, creatinine, globulin, and rate of neutrophil. The AUC of the training group and validation group are both 0.96, with an excellent HL and DCA curve, presenting excellent clinical prediction efficiency and application potential.
Supplementary material
Footnotes
Funding: This study was supported by Science and Technology Program Projects in Medicine and Health of Zhejiang Province (2025KY844), Science and Technology Program Projects in Traditional Chinese Medicine of Zhejiang Province (2023ZL490)
Provenance and peer review: Not commissioned; externally peer reviewed.
Patient consent for publication: Not applicable.
Ethics approval: This study involves human participants and was approved by the Ethics Committee for Human Research, The Second Affiliated Hospital, School of Medicine, Zhejiang University, Approval number: (2024), Ethical Research No. (1379). Participants gave informed consent to participate in the study before taking part.
Data availability statement
Data are available upon reasonable request.
References
- 1.GBD 2019 Stroke Collaborators Global, regional, and national burden of stroke and its risk factors, 1990–2019: a systematic analysis for the global burden of disease study 2019. Lancet Neurol. 2021;20:795–820. doi: 10.1016/S1474-4422(21)00252-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Malhotra K, Gornbein J, Saver JL. Ischemic Strokes Due to Large-Vessel Occlusions Contribute Disproportionately to Stroke-Related Dependence and Death: A Review. Front Neurol. 2017;8:651. doi: 10.3389/fneur.2017.00651. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Campbell BCV, Mitchell PJ, Kleinig TJ, et al. Endovascular therapy for ischemic stroke with perfusion-imaging selection. N Engl J Med. 2015;372:1009–18. doi: 10.1056/NEJMoa1414792. [DOI] [PubMed] [Google Scholar]
- 4.Turc G, Kim AS. First-Line Use of Contact Aspiration or Stent Retriever Thrombectomy for Large Vessel Occlusion Stroke. Stroke. 2019;50:2634–6. doi: 10.1161/STROKEAHA.119.025753. [DOI] [PubMed] [Google Scholar]
- 5.Berkhemer OA, Fransen PSS, Beumer D, et al. A randomized trial of intraarterial treatment for acute ischemic stroke. N Engl J Med. 2015;372:11–20. doi: 10.1056/NEJMoa1411587. [DOI] [PubMed] [Google Scholar]
- 6.Widimsky P, Snyder K, Sulzenko J, et al. Acute ischaemic stroke: recent advances in reperfusion treatment. Eur Heart J. 2023;44:1205–15. doi: 10.1093/eurheartj/ehac684. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Yang J, Jin Z, Song J, et al. Futile Recanalization After Endovascular Treatment in Patients With Acute Basilar Artery Occlusion. Neurosurgery. 2023;92:1006–12. doi: 10.1227/neu.0000000000002313. [DOI] [PubMed] [Google Scholar]
- 8.Martini M, Mocco J, Turk A, et al. An international multicenter retrospective study to survey the landscape of thrombectomy in the treatment of anterior circulation acute ischemic stroke: outcomes with respect to age. J Neurointerv Surg. 2020;12:115–21. doi: 10.1136/neurintsurg-2019-015093. [DOI] [PubMed] [Google Scholar]
- 9.Johnston KC, Connors AF, Wagner DP, et al. Predicting Outcome in Ischemic Stroke. Stroke. 2003;34:200–2. doi: 10.1161/01.STR.0000047102.61863.E3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Perez-Vega C, Domingo RA, Tripathi S, et al. Influence of glucose levels on clinical outcome after mechanical thrombectomy for large-vessel occlusion: a systematic review and meta-analysis. J Neurointerv Surg. 2022;14:17–21. doi: 10.1136/neurintsurg-2021-017771. [DOI] [PubMed] [Google Scholar]
- 11.Ruan H, Ran X, Li S-S, et al. Dyslipidemia versus obesity as predictors of ischemic stroke prognosis: a multi-center study in China. Lipids Health Dis. 2024;23:72. doi: 10.1186/s12944-024-02061-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Lattanzi S, Norata D, Divani AA, et al. Systemic Inflammatory Response Index and Futile Recanalization in Patients with Ischemic Stroke Undergoing Endovascular Treatment. Brain Sci. 2021;11:1164. doi: 10.3390/brainsci11091164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Zhang P, Xu P, Duan Z, et al. Effects of admission systemic inflammatory indicators on clinical outcomes in patients with vertebrobasilar artery occlusion: insight from the PERSIST registry. J Neurointerv Surg. 2023;15:e270–6. doi: 10.1136/jnis-2022-019437. [DOI] [PubMed] [Google Scholar]
- 14.Dargazanli C, Blaquière M, Moynier M, et al. Inflammation biomarkers in the intracranial blood are associated with outcome in patients with ischemic stroke. J Neurointerv Surg. 2025;17:159–66. doi: 10.1136/jnis-2023-021365. [DOI] [PubMed] [Google Scholar]
- 15.Kasner SE. Clinical interpretation and use of stroke scales. Lancet Neurol. 2006;5:603–12. doi: 10.1016/S1474-4422(06)70495-1. [DOI] [PubMed] [Google Scholar]
- 16.Larrue V, von Kummer R R, Müller A, et al. Risk factors for severe hemorrhagic transformation in ischemic stroke patients treated with recombinant tissue plasminogen activator: a secondary analysis of the European-Australasian Acute Stroke Study (ECASS II) Stroke. 2001;32:438–41. doi: 10.1161/01.str.32.2.438. [DOI] [PubMed] [Google Scholar]
- 17.Pásztor L, Szabó KZ, Szatmári G, et al. Mapping geogenic radon potential by regression kriging. Sci Total Environ. 2016;544:883–91. doi: 10.1016/j.scitotenv.2015.11.175. [DOI] [PubMed] [Google Scholar]
- 18.Broocks G, Kemmling A, Kniep H, et al. Edema Reduction versus Penumbra Salvage: Investigating Treatment Effects of Mechanical Thrombectomy in Ischemic Stroke. Ann Neurol. 2023 doi: 10.1002/ana.26802. [DOI] [PubMed] [Google Scholar]
- 19.Han Z, Liu Q, Li H, et al. The role of monocytes in thrombotic diseases: a review. Front Cardiovasc Med. 2023;10:1113827. doi: 10.3389/fcvm.2023.1113827. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Wu W, Bao W, Chen X, et al. Endothelial Gata6 deletion reduces monocyte recruitment and proinflammatory macrophage formation and attenuates atherosclerosis through Cmpk2-Nlrp3 pathways. Redox Biol. 2023;64:102775. doi: 10.1016/j.redox.2023.102775. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Allen N, Barrett TJ, Guo Y, et al. Circulating monocyte-platelet aggregates are a robust marker of platelet activity in cardiovascular disease. Atherosclerosis. 2019;282:11–8. doi: 10.1016/j.atherosclerosis.2018.12.029. [DOI] [PubMed] [Google Scholar]
- 22.Carestia A, Mena HA, Olexen CM, et al. Platelets Promote Macrophage Polarization toward Pro-inflammatory Phenotype and Increase Survival of Septic Mice. Cell Rep. 2019;28:896–908. doi: 10.1016/j.celrep.2019.06.062. [DOI] [PubMed] [Google Scholar]
- 23.Loguinova M, Pinegina N, Kogan V, et al. Monocytes of Different Subsets in Complexes with Platelets in Patients with Myocardial Infarction. Thromb Haemost. 2018;118:1969–81. doi: 10.1055/s-0038-1673342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Witten A, Martens L, Schäfer A-C, et al. Monocyte subpopulation profiling indicates CDK6-derived cell differentiation and identifies subpopulation-specific miRNA expression sets in acute and stable coronary artery disease. Sci Rep. 2022;12:5589. doi: 10.1038/s41598-022-08600-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Yang F, Wang M, Chen Y, et al. Association of cardio-renal biomarkers and mortality in the U.S.: a prospective cohort study. Cardiovasc Diabetol. 2023;22:265. doi: 10.1186/s12933-023-01986-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Schmidt M, Mansfield KE, Bhaskaran K, et al. Serum creatinine elevation after renin-angiotensin system blockade and long term cardiorenal risks: cohort study. BMJ. 2017;356 doi: 10.1136/bmj.j791. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data are available upon reasonable request.