https://orcid.org/0009-0006-0357-3949
Figures
Abstract
Background
Ischemic stroke poses a substantial global health burden. Reliable biomarkers for risk stratification in critically ill stroke patients are lacking. This study investigates estimated pulse wave velocity (ePWV), a non-invasive measure of arterial stiffness, as a novel prognostic indicator for mortality in this population.
Methods
This retrospective cohort study analyzed data from 3,408 adult ischemic stroke patients admitted to the ICU within the MIMIC-IV database. Patients were categorized by ePWV tertiles. The primary outcome was 28-day mortality (in-hospital and ICU). Multivariate Cox regression models were employed to assess the association between ePWV and mortality, adjusting for comprehensive clinical variables.
Results
Of the 3,408 patients, 481 (14.1%) died within 28 days of hospitalization. Non-survivors demonstrated significantly higher ePWV values (11.19 vs. 10.57, P < 0.001). Multivariate analysis revealed that ePWV was an independent predictor of both in-hospital (HR = 1.16, 95% CI: 1.05–1.28, P = 0.0033) and ICU 28-day mortality (HR = 1.31, 95% CI: 1.16–1.48, P < 0.0001). Subgroup analyses revealed significant interactions between ePWV and atrial fibrillation for in-hospital mortality (P = 0.0498) and mechanical ventilation for ICU mortality (P = 0.0294). For in-hospital mortality, the ePWV-associated risk was higher in patients with atrial fibrillation (HR 1.19, 95% CI: 1.07–1.31) compared to those without (HR 1.10, 95% CI: 0.98–1.23). For ICU mortality, the ePWV-associated risk was higher in patients without mechanical ventilation (HR 1.45, 95% CI: 1.24–1.70) compared to those with (HR 1.26, 95% CI: 1.11–1.44).
Citation: Zhao S, Liu M, Zhang M, Liu H, Chen X, Wang Y (2025) Association between Estimated Pulse Wave Velocity (ePWV) and in-hospital and ICU 28-day mortality in ischemic stroke patients: A retrospective analysis of the MIMIC-IV database. PLoS One 20(8): e0328818. https://doi.org/10.1371/journal.pone.0328818
Editor: Redoy Ranjan, Royal Holloway University of London, UNITED KINGDOM OF GREAT BRITAIN AND NORTHERN IRELAND
Received: March 9, 2025; Accepted: July 7, 2025; Published: August 12, 2025
Copyright: © 2025 Zhao et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data used in this study were obtained from the MIMIC-IV database (version 3.0). Due to restrictions in the Data Use Agreement (DUA), redistribution of the data in any form is prohibited. Researchers may access the original data by completing the required certification training and signing the DUA through the PhysioNet platform (https://physionet.org/content/mimiciv/). The authors confirm they did not have any special access privileges.
Funding: This study was supported by the Basic Research Program of Shanxi Province (Grant No. 20210302124579, received by Dr. Yu Wang).
Competing interests: The authors have declared that no competing interests exist.
Introduction
Ischemic stroke, characterized by the interruption of blood flow to the brain, is a leading cause of morbidity and mortality worldwide [1]. The pathophysiology of ischemic stroke involves a complex interplay of vascular occlusion, inflammation, and neuronal injury, which can lead to rapid brain tissue death if not promptly addressed [2]. The burden of ischemic stroke is compounded by its association with various risk factors, including hypertension, diabetes, and hyperlipidemia, which contribute to the underlying vascular pathology [3]. Given the high incidence and the profound impact of ischemic stroke on patients and healthcare systems, understanding its prognostic factors is crucial for improving clinical outcomes.
Previous studies have highlighted the utility of various clinical scoring systems and biomarkers in predicting outcomes in critically ill patients, including those with ischemic stroke [4,5]. For instance, the Sequential Organ Failure Assessment (SOFA) and Simplified Acute Physiology Score II (SAPS II) scores are commonly utilized to evaluate the severity of illness and predict mortality in critically ill patients [6,7]. However, there remains a need for more reliable and easily applicable prognostic tools that can enhance risk stratification in this vulnerable population. In contrast to these conventional scores that primarily evaluate organ dysfunction (e.g., SOFA) or acute physiological status (e.g., SAPS II), estimated pulse wave velocity (ePWV) enables more nuanced characterization of cardiovascular status, potentially refining risk stratification. The ePWV is a non-invasive measure of arterial stiffness, which is calculated using age and mean blood pressure (MBP) [8]. This estimation is based on established equations derived from the Reference Values for Arterial Stiffness Collaboration, which correlate well with the gold standard measurement of carotid-femoral pulse wave velocity (cfPWV) [9]. The calculation of ePWV allows for a practical assessment of vascular health in clinical settings, particularly in populations where direct measurement of cfPWV may not be feasible. The ePWV is increasingly recognized as a significant marker of cardiovascular health, reflecting the elasticity of arterial walls. Elevated ePWV values are associated with various cardiovascular risk factors, including hypertension, diabetes, and aging [10]. Studies have demonstrated that increased arterial stiffness, as indicated by higher ePWV, correlates with adverse cardiovascular outcomes, including myocardial infarction and stroke [11]. Furthermore, ePWV has been shown to predict cardiovascular mortality independently of traditional risk factors, highlighting its potential role in risk stratification and management of patients with cardiovascular diseases [12]. The study by Stamatelopoulos et al. revealed that incorporating ePWV enhances the prognostic value of existing clinical scoring methods through its unique capacity to provide vascular health insights [13]. Their findings demonstrated that ePWV provided significant discrimination enhancement and reclassification advantages over the 4C Mortality score and Charlson comorbidity index (CCI).
The relationship between ePWV and cardiovascular health underscores the importance of monitoring arterial stiffness as part of comprehensive cardiovascular risk assessment [14]. However, a critical gap remains in our understanding of ePWV’s association with short-term mortality among critically ill ischemic stroke patients, particularly within the ICU setting. Previous research on ePWV in stroke has primarily focused on long-term outcomes in non-institutionalized stroke populations [15,16], largely neglecting the unique challenges and complexities presented by critically ill patients. Moreover, while studies on critically ill patients have explored ePWV in conditions such as acute kidney injury, subarachnoid hemorrhage, and coronary heart disease [17–19], this specific population of critically ill ischemic stroke patients has not been directly examined. Therefore, this study aims to evaluate the predictive value of ePWV for in-hospital and ICU 28-day mortality in patients with ischemic stroke using data from the MIMIC-IV database. We hypothesize that elevated ePWV will be independently associated with increased in-hospital and ICU 28-day mortality in this population. By conducting a retrospective analysis, we seek to determine whether ePWV can serve as an independent predictor of mortality in this population, thereby enhancing risk stratification and informing clinical management strategies.
Methods
Study design and population
The analysis utilized data extracted from the Medical Information Mart for Intensive Care (MIMIC-IV version 3.0) database, which contains comprehensive clinical information on critically ill patients [20,21]. The author of this study, Mingjie Liu (ID: 13908736), registered, completed necessary training, and obtained approval to use the dataset for this study. The database contains comprehensive medical records of patients admitted to the intensive care units at Beth Israel Deaconess Medical Center, which provides a rich resource for analyzing patient outcomes in various critical care settings [22]. The use of the MIMIC-IV database was approved by the Institutional Review Boards of the Massachusetts Institute of Technology (MIT) and Beth Israel Deaconess Medical Center (BIDMC). Since the data were anonymized, informed consent was not required.
This study included adult patients aged 18 years and older admitted to the ICU with a diagnosis of ischemic stroke based on the International Classification of Diseases, 9th and 10th Revision [23]. Inclusion criteria required patients to have their first ICU stay recorded and a stay time in the ICU of at least 24 hours. Patients were excluded if they lacked ePWV data. Finally, 3408 patients were included and divided into three groups according to the ePWV tertiles (Fig 1).
Measurement of ePWV and outcomes
In this study, the ePWV was calculated using the non-invasive arterial stiffness measurement equation published by the Reference Values for Arterial Stiffness Collaboration in 2010 [24]. This formula effectively captures the combined effects of age and MBP on pulse wave velocity. The MBP is determined by adding the diastolic blood pressure to 0.4 times the difference between the systolic and diastolic blood pressure. Then ePWV was derived from the formula: ePWV = 9.587 – (0.402 × age) + (4.5610 × 0.001 × age2) – (2.621 × 0.00001 × age2 × MBP) + (3.176 × 0.001 × age × MBP) – (1.832 × 0.01 × MBP) [25]. Previous studies have validated the robustness of this equation for arterial stiffness assessment by demonstrating comparable predictive value between calculated ePWV and measured cfPWV [26]. By applying this formula, we aimed to accurately assess arterial health in ischemic stroke patients, a population typically characterized by multiple cardiovascular risk factors. The primary outcomes were in-hospital and ICU 28-day mortality [27].
Variables selection
We extracted variables as potential covariates from the database, all measured within 24 hours after ICU admission. The selection of these covariates was guided by both clinical relevance and established literature on ischemic stroke outcomes. Clinical severity scores, including SOFA, Acute Physiology Score III (APS3), SAPS II, Oxford Acute Severity of Illness Score (OASIS), and Glasgow Coma Scale (GCS), were included to comprehensively capture patient condition at admission, as these indices have been previously demonstrated to correlate with mortality risk in critically ill patients [28]. Demographic characteristics (sex, age) and comorbidities, such as heart failure (HF), chronic kidney disease (CKD), diabetes, hypertension, chronic pulmonary disease, and atrial fibrillation (AF), were selected based on their known impact on cardiovascular risk and stroke prognosis. These factors have been consistently identified in prior epidemiological studies as significant predictors of both short-term and long-term outcomes in ischemic stroke patients [29].
Laboratory parameters, including complete blood count components [red blood cell count (RBC), white blood cell count (WBC), platelet count, red cell distribution width (RDW), hemoglobin] and metabolic markers [sodium, serum creatinine, bicarbonate, potassium, chloride, total calcium], were chosen to provide a comprehensive assessment of physiological status. These markers can reflect underlying inflammatory processes, metabolic disturbances, and organ dysfunction that may influence stroke outcomes [30]. Treatment-related variables such as intravenous tissue plasminogen activator (iv-tPA), vasopressin, and mechanical ventilation were included to account for potential therapeutic interventions that might modify patient prognosis. The CCI was incorporated to adjust for the cumulative impact of pre-existing chronic conditions [31].
Statistical analysis
Continuous variables were presented as mean ± standard deviation (SD) for normally distributed variables, median (interquartile range, IQR) for non-normally distributed continuous variables, and categorical variables were presented as numbers and percentages. Baseline characteristics were compared using t-tests or one-way analysis of variance (ANOVA) for continuous variables and Pearson’s chi-square tests for categorical variables. Missing data imputation was conducted using the missForest R package, which implements an iterative random forest algorithm for multivariate imputation [32]. This method was applied to variables exhibiting missing data rates below 30% (S1 Table), while variables exceeding this missingness threshold were excluded. The missForest algorithm is a non-parametric method that can handle non-linear relationships and complex interactions between variables. It relies on the assumption that data are missing at random (MAR), meaning that the probability of missingness depends only on observed values and not on the unobserved values themselves. In the context of this study using the MIMIC-IV database, this assumption is considered reasonable due to the database’s standardized data collection procedures. These procedures suggest that missing data are more likely due to factors in clinical practice (e.g., certain tests not being routinely performed on all patients) rather than unobserved patient-specific factors. Furthermore, variable selection was guided by clinical relevance and established literature on ischemic stroke outcomes, allowing for inference of the likelihood of missing values based on other observed variables. However, the MAR assumption is untestable and may not perfectly hold. To mitigate potential bias from this assumption, a sensitivity analysis was performed by repeating primary analyses using the original, un-imputed dataset and comparing the results.
The Kaplan-Meier curves were employed to evaluate differences in survival outcomes among tertiles of ePWV. Univariate and multivariate Cox regression models were used to examine the association between ePWV and 28-day mortality. Variables with a P-value < 0.05 in the univariate analysis were included in the multivariate Cox regression. Two models were constructed: Model 1 (adjusted for age, APS3, SAPS2, OASIS, SOFA, GCS, CCI), and Model 2 (fully adjusted model). Multicollinearity was assessed using the variance inflation factor (VIF), with a VIF > 10 indicating significant multicollinearity. Smooth curve fitting was also utilized to explore the association between ePWV and 28-day mortality outcomes, and threshold effect analysis was conducted to confirm the linearity of this relationship [33]. Additionally, Stratified analyses were performed across subgroups defined by age, heart failure, atrial fibrillation, mechanical ventilation, iv-tPA, and vasopressin to assess the consistency of the association. All statistical analyses were performed using R version 4.4.2 and Empowerstats (version 4.2), and a two-tailed P-value < 0.05 was considered statistically significant.
Results
Baseline characteristics
The study included 3,408 patients, with 2,927 (85.9%) surviving and 481 (14.1%) dying within 28 days of hospitalization (Table 1). Of these, 1,667 (48.91%) were female and 1,741 (51.09%) were male. Among survivors, 1423 (48.62%) were female and 1504 (51.38%) were male, while among non-survivors, 244 (50.73%) were female and 237 (49.27%) were male. There was no significant difference in sex distribution between the two groups (P = 0.42). Non-survivors had significantly higher ePWV values (11.19 vs. 10.57, P < 0.001), indicating greater arterial stiffness. They were also older (median age: 74.87 vs. 70.24 years, P < 0.0001) and had higher severity scores, including SOFA (5.00 vs. 3.00, P < 0.0001), APS3 (52.00 vs. 36.00, P < 0.0001), and SAPSII (43.00 vs. 32.00, P < 0.0001). Additionally, non-survivors exhibited a higher prevalence of heart failure (34.51% vs. 24.09%, P < 0.0001), CKD (24.53% vs. 18.65%, P < 0.01), and diabetes mellitus (38.67% vs. 32.46%, P < 0.01). However, no significant differences were observed in hypertension (P = 0.09) or chronic pulmonary disease (P = 0.31).
Non-survivors demonstrated poorer laboratory profiles, including lower hemoglobin (11.00 vs. 11.40 g/dL, P < 0.0001), platelet counts (198.00 vs. 207.00 × 10³/μL, P = 0.02), and RBC counts (3.69 vs. 3.81 × 10⁶/μL, P < 0.0001), alongside higher WBC counts (13.00 vs. 10.00 × 10³/μL, P < 0.0001) and serum creatinine levels (1.12 vs. 0.90 mg/dL, P < 0.0001). Metabolic disturbances were more pronounced, with lower bicarbonate (22.00 vs. 23.00 mEq/L, P < 0.0001) and calcium levels (8.54 vs. 8.66 mg/dL, P < 0.01), but higher potassium levels (4.30 vs. 4.10 mEq/L, P < 0.0001). The use of vasopressin was significantly higher in non-survivors (26.61% vs. 8.10%, P < 0.0001), and they had a shorter median in-hospital stay (6.31 vs. 8.91 days, P < 0.0001).
Kaplan–Meier curves for In-hospital and ICU mortality
The Kaplan-Meier curves for in-hospital and ICU 28-day mortality, stratified by ePWV categories (T1, T2, T3), demonstrate significant differences in survival probabilities (P < 0.0001 for both) (Fig 2). In both analyses, patients in the T1 group exhibited the highest survival rates, while those in the T3 group had the lowest. The survival probabilities declined steadily over the 28 days, with the T3 group showing a more pronounced decrease compared to T1 and T2. These findings indicate that higher ePWV is associated with increased mortality risk, highlighting its role as a critical prognostic indicator for both in-hospital and ICU outcomes in critically ill patients.
Association of ePWV with In-Hospital and ICU 28-Day Mortality in Ischemic Stroke Patients
We evaluated all potential covariates using univariate Cox regression analysis (see Table 2). The results demonstrated that unadjusted ePWV was significantly associated with both in-hospital and ICU 28-day all-cause mortality (HR = 1.12, 95% CI: 1.08, 1.15, P < 0.0001; HR = 1.12, 95% CI: 1.07, 1.17, P < 0.0001, respectively). Subsequently, covariates with P < 0.05 in the univariate analysis were included in the multivariate Cox regression analysis. Table 3 presents the adjusted analysis of ePWV with in-hospital and ICU 28-day mortality among ischemic stroke patients.
In Model 1, after adjusting for age, APS3, SAPSII, OASIS, SOFA, GCS, and CCI, ePWV remained significantly associated with in-hospital 28-day mortality (HR = 1.14, 95% CI: 1.03, 1.25, P = 0.0086) and ICU 28-day mortality (HR = 1.25, 95% CI: 1.11, 1.40, P = 0.0002). Furthermore, in Model 2, after additional adjustments for RDW, WBC, Bicarbonate, Chloride, Serum Creatinine, Sodium, Potassium, HF, AF, Mechanical Ventilation, iv-tPA, and Vasopressin, the ePWV continued to serve as an independent predictor for in-hospital 28-day mortality (HR = 1.16, 95% CI: 1.05, 1.28, P = 0.0033) and ICU 28-day mortality (HR = 1.31, 95% CI: 1.16, 1.48, P < 0.0001).
Analysis of ePWV: Smooth curve fitting and threshold effects on mortality outcomes.
The results of the curve fitting analysis illustrate the linear relationships between ePWV and both in-hospital and ICU mortality (Fig 3). The log relative risk for mortality outcomes increases with higher ePWV values, indicating a greater mortality risk for individuals with elevated ePWV. The threshold effect analysis results in Table 4 assess the associations between ePWV and hospital 28-day mortality as well as ICU 28-day mortality using standard linear and two-piecewise linear models. The log-likelihood ratio test P-values for hospital mortality and ICU mortality are 0.118 and 0.202, respectively, suggesting no significant deviation from linearity in the relationships between ePWV and mortality outcomes. Overall, these findings suggest linear correlations between ePWV and both hospital 28-day mortality and ICU 28-day mortality.
Adjusted for covariates in Model 2.
Subgroup analyses
The subgroup analyses for hospital and ICU mortality at 28 days reveal differential impacts of clinical factors on mortality risk (Fig 4). For hospital mortality, no significant interactions were detected for age, HF, mechanical ventilation, iv-tPA, or vasopressin. However, atrial fibrillation significantly influenced hospital mortality (P = 0.0498), with higher risk observed in patients with AF. For ICU mortality, no significant interactions were detected for age, HF, AF, iv-tPA, or vasopressin. Notably, mechanical ventilation demonstrated a significant interaction effect (P = 0.0294), with higher risk observed in patients not on mechanical ventilation. These findings highlight atrial fibrillation as a significant modifier of hospital mortality and mechanical ventilation as a significant modifier of ICU mortality, while other factors did not significantly alter the relationship between ePWV and mortality outcomes.
Sensitivity analysis
To evaluate the influence of random forest imputation on our results, we conducted a sensitivity analysis utilizing complete case analysis, which included only patients without any missing data. We subsequently re-performed the multivariate Cox regression analysis. Table 5 displays the relationship between ePWV and both in-hospital 28-day mortality and ICU 28-day mortality among patients with ischemic stroke in the complete case analysis. Consistent with the initial findings, continuous ePWV remained significantly associated with both outcomes: in-hospital 28-day mortality (Model 2: HR = 1.15, 95% CI: 1.02–1.30, P = 0.0194) and ICU 28-day mortality (Model 2: HR = 1.26, 95% CI: 1.10–1.46, P = 0.0013). Additionally, categorical ePWV continued to show significant associations with ICU 28-day mortality (T2 versus T1: P = 0.0335; T3 versus T1: P = 0.0094), with a notable trend across categories (P for trend = 0.0106). These findings indicate that our primary findings are relatively robust to different approaches in handling missing data.
Discussion
The findings of this study underscore the significant association between ePWV and in-hospital and ICU 28-day mortality in ischemic stroke patients, providing novel insights into the role of arterial stiffness in critical care outcomes. The baseline demographic and clinical characteristics highlight the heterogeneity of the study population, with ePWV emerging as a key marker of vascular health. The multivariable Cox regression model confirms ePWV as an independent predictor of mortality, even after adjusting for traditional risk factors such as age, heart failure, atrial fibrillation, and mechanical ventilation. Notably, the subgroup analyses reveal that ePWV has a significant predictive value in patients with AF for hospital mortality and those not receiving timely mechanical ventilation for ICU mortality. The results highlight the potential of ePWV as a non-invasive prognostic marker in ischemic stroke.
Our findings regarding the relationship between ePWV and mortality in ischemic stroke patients align with recent evidence connecting ePWV to adverse cardiovascular outcomes. Patients with elevated ePWV face a significantly higher risk of cardiovascular mortality and major adverse cardiovascular events [34]. High ePWV levels have been linked to worse early clinical outcomes post-stroke, underscoring its importance in risk assessment for this patient group [35]. Additionally, PWV correlates with the severity of cerebral arterial damage, further underscoring its role in predicting outcomes in stroke patients [36]. While previous studies have investigated PWV as a prognostic marker, our research introduces ePWV as a non-invasive predictive measure, specifically highlighting its direct correlation with mortality risk in critically ill ischemic stroke patients. We discovered a direct correlation between ePWV and mortality risk in critically ill ischemic stroke patients, suggesting that ePWV could be a more effective prognostic tool compared to conventional risk factors.
Several mechanisms may explain the observed associations. Arterial stiffness, quantified by ePWV, disrupts cerebrovascular dynamics through complex pathophysiological mechanisms. Elevated arterial stiffness compromises blood vessels’ ability to accommodate pulsatile blood flow, resulting in heightened pulsatility within cerebral arteries [37,38]. This hemodynamic alteration substantially increases the hydraulic load on delicate microvessels, directly contributing to small vessel disease progression and escalating cerebral ischemia risk [39]. In addition to directly impacting cerebral vasculature, the elevated pulse pressure resulting from increased arterial stiffness may lead to injury in multiple organs, including the kidneys and heart. Previous studies have shown that elevated arterial stiffness reduces coronary perfusion pressure, increases left ventricular afterload, and promotes further ventricular remodeling and diastolic dysfunction [40,41]. Another investigation revealed that preoperative vascular dysfunction, characterized by increased aortic stiffness measured via PWV, is associated with a higher incidence of acute kidney injury (AKI) in patients undergoing coronary artery bypass graft surgery [42]. Mechanistically, increased arterial stiffness leads to elevated renal vascular resistance and impaired renal perfusion, both of which are critical factors in the pathogenesis of AKI [43]. Moreover, elevated ePWV may also be related to changes in inflammation and oxidative stress. Jun et al. emphasized that biochemical factors such as oxidative stress and inflammation could further exacerbate arterial stiffness, thereby impacting renal injury, especially in acute scenarios [44].
The pathological impact of arterial stiffness extends beyond hemodynamic alterations to profound structural and functional vascular changes. Higher ePWV levels are intrinsically linked to cerebral arterial calcification, serving as a robust marker of vascular health decline [45]. This calcification process not only reduces cerebral perfusion but also triggers a cascade of endothelial dysfunction and microvascular damage. The pulsatility generated by stiffened arteries initiates inflammatory responses and structural vascular remodeling, compounding the inherent risks faced by ischemic stroke patients [46]. Studies have consistently shown that elevated arterial stiffness contributes to progressive vascular damage, cognitive decline, and increased vulnerability to microvascular diseases, with particularly significant implications for patients recovering from transient ischemic attacks or strokes [47,48].
The significant interaction between ePWV and AF in-hospital mortality is particularly noteworthy. AF significantly influenced hospital mortality (P = 0.0498), with AF patients demonstrating a notably higher mortality risk. This association can be attributed to multiple pathophysiological mechanisms, including enhanced thrombotic potential, compromised cardiac hemodynamics, and systemic inflammatory responses [49–52]. This finding implies that patients with both elevated ePWV and AF may represent a high-risk subgroup requiring more intensive monitoring and targeted interventions. The interaction between mechanical ventilation and mortality in ischemic stroke patients reveals a critical protective mechanism of respiratory support. Providing careful ventilatory assistance can bring about significant physiological advantages, such as enhanced oxygen levels and prevention of respiratory muscle fatigue, ultimately leading to a notable decrease in mortality risk [53]. Non-ventilated patients are at a higher risk of mortality due to inadequate respiratory compensation and progressive respiratory muscle exhaustion [54]. The significant interaction effect (P = 0.0294) highlights the importance of early respiratory support in managing critically ill patients. These findings underscore mechanical ventilation as a pivotal life-saving intervention in severe ischemic stroke patients, serving as a critical therapeutic strategy rather than merely a supportive measure.
Strengths and limitations
Our research has several notable strengths. Firstly, it is the first to establish a clear linear correlation between ePWV and 28-day mortality in critically ill ischemic stroke patients, suggesting a promising approach to personalized patient management. This finding complements existing literature that has linked arterial stiffness to adverse clinical outcomes, further elucidating the role of ePWV in risk stratification and management of critically ill ischemic stroke patients. The consistent association between elevated ePWV and increased mortality suggests that this non-invasive marker could serve as a valuable prognostic tool in clinical decision-making. Specifically, patients with higher ePWV values might benefit from more intensive monitoring, aggressive management of cardiovascular risk factors, and tailored interventional strategies. Secondly, our subgroup analyses revealed particularly significant predictive value in patients with atrial fibrillation and those not receiving timely mechanical ventilation, suggesting that ePWV may serve as a more refined prognostic indicator in specific high-risk patient subgroups. In patients with atrial fibrillation, where cardiovascular complications are frequent, ePWV demonstrated enhanced predictive capabilities for in-hospital mortality. Similarly, among patients who did not receive mechanical ventilation within the critical time window, ePWV emerged as a potentially crucial marker for identifying increased ICU mortality risk. These findings underscore the marker’s potential to provide more granular risk stratification beyond traditional clinical parameters, potentially guiding more personalized and targeted clinical interventions.
However, our study has several important limitations. While ePWV provides a non-invasive assessment of arterial stiffness, it may not comprehensively capture the complex pathophysiological mechanisms underlying vascular changes. Our research was conducted using the MIMIC database, a single-center cohort from an academic medical center in Boston, which potentially constrains the generalizability of our findings. The specific patient population represented in this database may not fully reflect the demographic and clinical diversity across different healthcare systems. Despite our rigorous covariate selection, the potential for unobserved confounding factors remains. The retrospective nature of the study inherently limits our ability to control for all potential confounding variables, and some clinically relevant information may not have been captured in the original medical records. Moreover, it is important to emphasize that this study is observational, and the results only indicate a correlation between ePWV and 28-day mortality, without proving causality. To further address the robustness of our findings, we performed a sensitivity analysis using complete case analysis, which included only patients with no missing data. We found that the associations between ePWV and mortality outcomes remained largely consistent with the main analysis results, suggesting that our primary findings are reasonably robust to the method used for handling missing data. However, complete case analysis also has its limitations, such as potentially reducing the statistical power of the study and introducing selection bias. Therefore, we cannot completely rule out the influence of the missing data handling method on our results.
Nevertheless, Our research provides novel insights into the relationship between arterial stiffness and outcomes in critically ill ischemic stroke patients. To address the current study’s limitations, future research should prioritize multi-center, prospective studies with more diverse patient cohorts. Validation of our findings across different healthcare settings and populations is crucial for establishing the broader clinical relevance of ePWV as a prognostic marker. Additionally, mechanistic studies exploring the complex pathways linking arterial stiffness to mortality in ischemic stroke patients would substantially contribute to our understanding.
Conclusion
The ePWV can serve as a useful biomarker for predicting mortality risk in ischemic stroke patients, especially in patients with AF or lacking timely mechanical ventilation. These findings highlight the importance of assessing arterial stiffness in the clinical management of ischemic stroke patients. Future research should focus on validating these findings in prospective, multi-center studies, exploring the potential benefits of interventions targeting arterial stiffness, and investigating the underlying mechanisms linking ePWV to stroke outcomes. Clinically, these results suggest that ePWV assessment could be integrated into risk stratification tools to identify high-risk patients who may benefit from more intensive monitoring and targeted therapies.
References
- 1. Al Alawi AM, Al Busaidi I, Al Shibli E, Al-Senaidi A-R, Al Manwari S, Al Busaidi I, et al. Health outcomes after acute ischemic stroke:retrospective and survival analysis from Oman. Ann Saudi Med. 2022;42(4):269–75. pmid:35933604
- 2. Veltkamp R, Pearce LA, Korompoki E, Sharma M, Kasner SE, Toni D, et al. Characteristics of Recurrent Ischemic Stroke After Embolic Stroke of Undetermined Source: Secondary Analysis of a Randomized Clinical Trial. JAMA Neurol. 2020;77(10):1233–40. pmid:32628266
- 3. Moore P, Esmail F, Qin S, Nand S, Berg S. Hypercoagulability of COVID-19 and Neurological Complications: A Review. J Stroke Cerebrovasc Dis. 2022;31(1):106163.
- 4. Garza-Alatorre G, Carrion-Garcia AL, Falcon-Delgado A, Garza-Davila EC, Martinez-Ponce de Leon AR, Botello-Hernandez E. Characteristics of Pediatric Stroke and Association of Delayed Diagnosis with Mortality in a Mexican Tertiary Care Hospital. Neuropediatrics. 2021;52(6):499–503. pmid:34261144
- 5. Liu S, Li M, Yang Y, Chen Y, Wang W, Zheng X. A novel risk model based on white blood cell-related biomarkers for acute kidney injury prediction in patients with ischemic stroke admitted to the intensive care unit. Front Med (Lausanne). 2022;9:1043396. pmid:36579155
- 6. Sudarsanan S, Sivadasan P, Chandra P, Omar AS, Gaviola Atuel KL, Ulla Lone H, et al. Comparison of Four Intensive Care Scores in Predicting Outcomes After Venoarterial Extracorporeal Membrane Oxygenation: A Single-center Retrospective Study. J Cardiothorac Vasc Anesth. 2025;39(1):131–42. pmid:39550342
- 7. Liu P, Ren Y. Prognostic value of SAPS II score for 28-day mortality in ICU patients with acute pulmonary embolism. Int J Cardiol. 2025;430:133201.
- 8. Cui X, Shi H, Hu Y, Zhang Z, Lu M, Wu J, et al. Association between estimated pulse wave velocity and in-hospital and one-year mortality of patients with chronic kidney disease and atherosclerotic heart disease: a retrospective cohort analysis of the MIMIC-IV database. Ren Fail. 2024;46(2):2387932. pmid:39120152
- 9. Greve SV, Blicher MK, Kruger R, Sehestedt T, Gram-Kampmann E, Rasmussen S. Estimated carotid–femoral pulse wave velocity has similar predictive value as measured carotid–femoral pulse wave velocity. J Hypertens. 2016;34(7).
- 10. Kim J, Song TJ, Song D, Lee KJ, Kim EH, Lee HS, et al. Brachial-ankle pulse wave velocity is a strong predictor for mortality in patients with acute stroke. Hypertension. 2014;64(2):240–6.
- 11. Gąsecki D, Rojek A, Kwarciany M, Kowalczyk K, Boutouyrie P, Nyka W, et al. Pulse wave velocity is associated with early clinical outcome after ischemic stroke. Atherosclerosis. 2012;225(2):348–52. pmid:23083680
- 12. Heffernan KS, Wilmoth JM, London AS. Estimated Pulse Wave Velocity and All-Cause Mortality: Findings From the Health and Retirement Study. Innov Aging. 2022;6(7):igac056. pmid:36284701
- 13. Stamatelopoulos K, Georgiopoulos G, Baker KF, Tiseo G, Delialis D, Lazaridis C, et al. Estimated pulse wave velocity improves risk stratification for all-cause mortality in patients with COVID-19. Sci Rep. 2021;11(1):20239. pmid:34642385
- 14. Kim BS, Lee Y, Park J-K, Lim Y-H, Shin J-H. Association of the Estimated Pulse Wave Velocity with Cardio-Vascular Disease Outcomes among Men and Women Aged 40-69 Years in the Korean Population: An 18-Year Follow-Up Report on the Ansung-Ansan Cohort in the Korean Genome Environment Study. J Pers Med. 2022;12(10):1611. pmid:36294750
- 15. Huang H, Bu X, Pan H, Yang S, Cheng W, Shubhra QTH, et al. Estimated pulse wave velocity is associated with all-cause and cardio-cerebrovascular disease mortality in stroke population: Results from NHANES (2003-2014). Front Cardiovasc Med. 2023;10:1140160. pmid:37153456
- 16. Li J, Jiang C, Ma J, Bai F, Yang X, Zou Q, et al. Estimated pulse wave velocity is associated with all-cause and cardiovascular mortality in individuals with stroke: A national-based prospective cohort study. Medicine (Baltimore). 2025;104(7):e41608. pmid:39960927
- 17. Li J, Zhang M, Ye B, Lu M, Liao G. Association between estimation of pulse wave velocity and all-cause mortality in critically ill patients with non-traumatic subarachnoid hemorrhage: an analysis based on the MIMIC-IV database. Front Neurol. 2024;15:1451116. pmid:39148699
- 18. Cui X, Hu Y, Li D, Lu M, Zhang Z, Kan D, et al. Association between estimated pulse wave velocity and in-hospital mortality of patients with acute kidney injury: a retrospective cohort analysis of the MIMIC-IV database. Ren Fail. 2024;46(1):2313172. pmid:38357758
- 19. Gu Y, Han X, Liu J, Li Y, Li Z, Zhang W, et al. Prognostic significance of the estimated pulse wave velocity in critically ill patients with coronary heart disease: analysis from the MIMIC‑IV database. Eur Heart J Qual Care Clin Outcomes. 2024.
- 20. Peng S, Huang J, Liu X, Deng J, Sun C, Tang J, et al. Interpretable machine learning for 28-day all-cause in-hospital mortality prediction in critically ill patients with heart failure combined with hypertension: A retrospective cohort study based on medical information mart for intensive care database-IV and eICU databases. Front Cardiovasc Med. 2022;9:994359. pmid:36312291
- 21. Johnson AEW, Bulgarelli L, Shen L, Gayles A, Shammout A, Horng S, et al. MIMIC-IV, a freely accessible electronic health record dataset. Sci Data. 2023;10(1):1. pmid:36596836
- 22. Jiang L, Wang Z, Wang L, Liu Y, Chen D, Zhang D, et al. Predictive value of the serum anion gap for 28-day in-hospital all-cause mortality in sepsis patients with acute kidney injury: a retrospective analysis of the MIMIC-IV database. Ann Transl Med. 2022;10(24):1373. pmid:36660703
- 23. Li W, Huang Q, Zhan K. Association of Serum Blood Urea Nitrogen to Albumin Ratio with in-Hospital Mortality in Patients with Acute Ischemic Stroke: A Retrospective Cohort Study of the eICU Database. Balkan Med J. 2024;41(6):458–68. pmid:39324419
- 24. Reference Values for Arterial Stiffness’ Collaboration. Determinants of pulse wave velocity in healthy people and in the presence of cardiovascular risk factors: “establishing normal and reference values”. Eur Heart J. 2010;31(19):2338–50. pmid:20530030
- 25. Chen M, Fan H, Xie L, Zhou L, Chen Y. Association between estimated pulse wave velocity and the risk of mortality in patients with subarachnoid hemorrhage: a retrospective cohort study based on the MIMIC database. BMC Neurol. 2024;24(1):408. pmid:39438839
- 26. Greve SV, Blicher MK, Kruger R, Sehestedt T, Gram-Kampmann E, Rasmussen S, et al. Estimated carotid-femoral pulse wave velocity has similar predictive value as measured carotid-femoral pulse wave velocity. J Hypertens. 2016;34(7):1279–89. pmid:27088638
- 27. Liu Q, Zheng H-L, Wu M-M, Wang Q-Z, Yan S-J, Wang M, et al. Association between lactate-to-albumin ratio and 28-days all-cause mortality in patients with acute pancreatitis: A retrospective analysis of the MIMIC-IV database. Front Immunol. 2022;13:1076121. pmid:36591285
- 28. Kozak HH, Buğrul A, Tol F. In the early period of acute ischemic stroke mortality scores and inflammatory markers can predict the need for intubation. Turkish Journal of Cerebrovascular Diseases. 2022.
- 29. Han E, Kim TH, Koo H, Yoo J, Heo JH, Nam HS. Heterogeneity in costs and prognosis for acute ischemic stroke treatment by comorbidities. J Neurol. 2019;266(6):1429–38. pmid:30879136
- 30. Frøyshov HM, Bjørnerem Å, Engstad T, Halvorsen DS. Elevated inflammatory markers predict mortality in long-term ischemic stroke-survivors: a population-based prospective study. Aging Clin Exp Res. 2017;29(3):379–85. pmid:27146666
- 31. Phan TG, Clissold BB, Ma H, Ly JV, Srikanth V. Predicting disability after ischemic stroke based on comorbidity index and stroke severity—From the virtual international stroke trials archive-acute collaboration. Frontiers in Neurology. 2017;8.
- 32. Stekhoven DJ, Bühlmann P. MissForest--non-parametric missing value imputation for mixed-type data. Bioinformatics. 2012;28(1):112–8.
- 33. Lin L, Chen C, Yu X. The analysis of threshold effect using Empower Stats software. Zhonghua Liu Xing Bing Xue Za Zhi. 2013;34(11):1139–41. pmid:24517951
- 34. Kim HL, Joh HS, Lim WH, Seo JB, Kim SH, Zo JH, et al. Estimated pulse wave velocity in the prediction of clinical outcomes in patients undergoing drug-eluting stent implantation. J Clin Med. 2023;12(18).
- 35. Park K-Y, Kim YB, Moon H-S, Suh B-C, Chung P-W. Association between cerebral arterial calcification and brachial-ankle pulse wave velocity in patients with acute ischemic stroke. Eur Neurol. 2009;61(6):364–70. pmid:19365129
- 36. Webb AJS, Simoni M, Mazzucco S, Kuker W, Schulz U, Rothwell PM. Increased cerebral arterial pulsatility in patients with leukoaraiosis: arterial stiffness enhances transmission of aortic pulsatility. Stroke. 2012;43(10):2631–6. pmid:22923446
- 37. Hendrickx JO, van Gastel J, Leysen H, Santos-Otte P, Premont RT, Martin B, et al. GRK5 - A functional bridge between cardiovascular and neurodegenerative disorders. Front Pharmacol. 2018;9:1484.
- 38. Moore EE, Jefferson AL. Impact of cardiovascular hemodynamics on cognitive aging. Arterioscler Thromb Vasc Biol. 2021;41(4):1255–64.
- 39. Khanna S, Parikh NS, VanWagner LB. Fatty liver and cerebrovascular disease: plausible association and possible mechanisms. Curr Opin Lipidol. 2022;33(1):31–8.
- 40. Xue R, Zhang J, Zhen Z, Liang W, Li Y, Zhang L. Estimated pulse wave velocity predicts mortality in patients with heart failure with preserved ejection fraction. Hellenic Journal of Cardiology. 2024.
- 41. Chen M, Fan H, Xie L, Zhou L, Chen Y. Association between estimated pulse wave velocity and the risk of mortality in patients with subarachnoid hemorrhage: a retrospective cohort study based on the MIMIC database. BMC Neurol. 2024;24(1):408. pmid:39438839
- 42. Greenwood SA, Mangahis E, Castle EM, Wang J, Campbell J, Deshpande R, et al. Arterial stiffness is a predictor for acute kidney injury following coronary artery bypass graft surgery. J Cardiothorac Surg. 2019;14(1):51. pmid:30845970
- 43. Pengshung MH, Bispham NZ, You Z, Oh E, Supiano MA, Chonchol MB, et al. Arterial Stiffness Is Independently Associated with Acute Kidney Injury in SPRINT. Clin J Am Soc Nephrol. 2021;16(10):1560–1.
- 44. Fassett RG, D’Intini V, Healy H, Gowardman J, Gan J-S, Sharman JE, et al. Assessment of arterial stiffness, oxidative stress and inflammation in acute kidney injury. BMC Nephrol. 2009;10:15. pmid:19538714
- 45. Meyer T, Kreft B, Bergs J, Antes E, Anders MS, Wellge B, et al. Stiffness pulsation of the human brain detected by non-invasive time-harmonic elastography. Front Bioeng Biotechnol. 2023;11:1140734. pmid:37650041
- 46. Reeve EH, Kronquist EK, Wolf JR, Lee B, Khurana A, Pham H, et al. Pyridoxamine treatment ameliorates large artery stiffening and cerebral artery endothelial dysfunction in old mice. J Cereb Blood Flow Metab. 2023;43(2):281–95. pmid:36189840
- 47. Webb AJ, Lawson A, Li L, Mazzucco S, Rothwell PM, Oxford Vascular Study Phenotyped Cohort. Physiological determinants of residual cerebral arterial pulsatility on best medical treatment after TIA or minor stroke. J Cereb Blood Flow Metab. 2021;41(6):1463–71. pmid:33153374
- 48. Baradaran H, Gupta A. Carotid artery stiffness: imaging techniques and impact on cerebrovascular disease. Front Cardiovasc Med. 2022;9.
- 49. Shaver CM, Chen W, Janz DR, May AK, Darbar D, Bernard GR, et al. Atrial Fibrillation Is an Independent Predictor of Mortality in Critically Ill Patients. Crit Care Med. 2015;43(10):2104–11. pmid:26154932
- 50. Murtaza M, Baig MMA, Ahmed J, Serbanoiu LI, Busnatu SS. Higher Mortality Associated With New-Onset Atrial Fibrillation in Cancer Patients: A Systematic Review and Meta-Analysis. Front Cardiovasc Med. 2022;9:867002. pmid:35498001
- 51. Jackson I, Etuk A, Jackson N. Impact of atrial fibrillation on inpatient outcomes among hospitalized patients with multiple myeloma. Cureus. 2022.
- 52. Ip RJ, Ali A, Baloch ZQ, Al-Abcha A, Jacob C, Arnautovic J, et al. Atrial Fibrillation as a Predictor of Mortality in High Risk COVID-19 Patients: A Multicentre Study of 171 Patients. Heart Lung Circ. 2021;30(8):1151–6. pmid:33781697
- 53. Telias I, Brochard LJ, Gattarello S, Wunsch H, Junhasavasdikul D, Bosma KJ, et al. The physiological underpinnings of life-saving respiratory support. Intensive Care Med. 2022;48(10):1274–86. pmid:35690953
- 54. Fattori S, Reitano E, Chiara O, Cimbanassi S. Predictive Factors of Ventilatory Support in Chest Trauma. Life. 2021;11(11).