Some limitations of the study are worth noting. Although we applied careful propensity score matching, there is still the possibility of selection bias and residual confounding, particularly due to va

The gap

Consensus across the literature

Supporting evidence — 4 representative gaps

• Machine learning-based prediction of ischemic cardio-cerebrovascular events after endovascular or microsurgical treatment of unruptured intracranial aneurysms and risk stratification by the early post-treatment triglyceride-glucose index (2026) · doi

  Author contributions First, despite a registry-based design, data were collected mainly from two centers, which may limit generalizability and introduce selection bias. Second, the model was only validated internally; external validation is needed to confirm clinical appli- cability. Third, TyG was measured only at postoperative day 3 and may be affected by perioperative stress, so it should be inter- preted as a prognostic marker rather than a causal factor. Fourth, detailed information regarding antiplatelet therapy timing and adherence was incomplete, representing a potential confounder. Fifth, as an observational study, residual confounding and treat- ment-selection bias could not be fully eliminated. Although ACS and cerebral infarction represent distinct clinical entities, the consistent directionality and robust predictive performance observed in component-specific analyses support the clinical validity of the composite ICCE endpoint for capturing overall post-treatment vascular vulnerability. YH: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing, Software. SC: Data curation, Investigation, Writing  – review & editing. DL: Data curation, Investigation, Writing – review & editing. ZZ: Methodology, Validation, Writing – review & editing. YL: Methodology, Validation, Writing – review & editing. YW: Conceptualization, Funding acquisition, Project admin- istration, Supervision, Writing – review & editing.

  Keywords: writing review editing validation clinical curation investigation methodology selection bias conceptualization author contributions first despite
• Efficacy and Safety of Dual Antiplatelet Therapy in Acute Ischemic Stroke With Recurrent Symptoms Post-thrombolysis: A Retrospective Cohort Study From the TriNetX US Collaborative Network (2026) · doi

  Some limitations of the study are worth noting. Although we applied careful propensity score matching, there is still the possibility of selection bias and residual confounding, particularly due to variations in antiplatelet dosing and administration protocols across hospitals. The TriNetX platform’s reliance on ICD-10 codes also limited our ability to assess certain clinical details, such as serial NIHSS scores during hospitalization or functional outcomes at 90 days using the modified Rankin Scale. These measures are critical for evaluating clinical progress and functional recovery, potentially omitting early complications or standardized benefits of therapy. Additionally, reliance on ICD-10 codes precluded direct identification of recurrent strokes as a distinct clinical event, as there is no specific code to distinguish stroke recurrence from previous cerebral ischemia, thereby introducing misclassification bias into the 30- to 90-day recurrence risk of ischemic stroke outcome.

  Keywords: clinical there bias reliance codes functional stroke recurrence limitations worth noting applied careful propensity score
• Associations of Blood Soluble Fas and IGF-I with Risk of Hemorrhagic Stroke: The Circulatory Risk Communities Study (CIRCS) (2026) · doi

  The strength of this study was that it was conducted within a large cohort of the general population in Japan, where the incidence of hemorrhagic stroke is relatively high. Our study was based on systematic surveillance for incident stroke and employed standardized epidemiological methods for measuring biomarkers. This study has several potential limitations. First, only single-point measurements of biomarkers and risk factors at baseline were available, which could lead to regression dilution bias. Second, statistical power was limited, particu- larly in subtype analyses (SAH and ICH), location-specific analyses (deep and lobar ICH), and tertile-based analyses; therefore, these findings should be interpreted with caution. Nevertheless, these analyses may still provide meaning- ful insights into the association between IGF-I and hem- orrhagic stroke, and possibly its underlying mechanisms. Third, although we adjusted for potential confounders, we cannot rule out residual confounding from unmeasured fac- tors, such as antithrombotic medication use, which was not available in this study. Fourth, the study population was limited to Japanese residents, so generalizability should be considered with caution. Nevertheless, because biomarkers and risk factors for hemorrhagic stroke may be common across racial and ethnic groups, the evidence may pro- vide a basis for extrapolation to other populations. Lastly, we used frozen serum samples to estimate sFas and IGF-I and did not examine long-term changes in the samples. A previous report showed that these biomarkers remained stable in serum samples stored for up to nine years at -70 °C in another Japanese cohort, where sample management methods were nearly identical to those in our study [32]. 1 3Translational Stroke Research (2026) 17:57  57 Page 6 of 7 As additional analyses stratified by storage duration yielded similar results, it is unlikely that long-term storage affected the findings.

  Keywords: stroke analyses biomarkers samples cohort population hemorrhagic based potential risk factors available limited caution nevertheless
• Differential impact of cumulative cerebral small-vessel disease burden on ischemic versus hemorrhagic stroke recurrence in hemorrhage-prone patients (2026) · doi

  There are some methodological limitations to be consid- ered. First, this study is an exploratory post hoc analysis. While our results demonstrate strong associations between m-cSVD burden and recurrent stroke risk in hemorrhage- prone patients, they should be interpreted as suggesting potential relationships rather than establishing causality. Furthermore, as most prospective clinical trials including PICASSO largely excluded patients who are in poor medi- cal condition or in a functionally dependent state, our study population might not reflect a real-world setting. Second, significant baseline imbalances existed across m-cSVD groups; patients with higher scores were older and had greater clinical and vascular frailty. This disparity, poten- tially reflecting clinical caution regarding intensive therapy in patients with high hemorrhagic burden, is noteworthy given the established role of statins in secondary preven- tion. Although we performed multivariable adjustments to mitigate these imbalances, the possibility of residual con- founding remains, necessitating a cautious interpretation of the independent prognostic value attributed to the m-cSVD score. Third, the relatively low incidence of hemorrhagic stroke compared to ischemic events might have resulted in insufficient statistical power to provide precise estimates of the association with m-cSVD scores. Consequently, the find- ings regarding hemorrhagic risk should be interpreted with caution, and larger-scale studies are warranted to further validate these associations. Forth, given that the phenotypes of cSVD markers are heterogeneous and vary according to their numbers, simple scoring of total cSVD by adding a point may jeopardizes the precision of estimates for assess- ing recurrent stroke risk. A detailed cSVD scoring approach that quantitatively incorporated the severity of each pheno- type may be more informative. Fifth, our study did not dis- tinguish between the anatomical locations (deep vs. lobar) of ICH or CMBs, which could reflect different underlying etiologies such as hypertensive arteriopathy and CAA. However, since approximately nine out of ten participants had hypertension, our results may reflect the outcomes of a population with hypertensive arteriopathy. Lastly, the use of a m-cSVD score that excludes EPVS was necessitated by data availability within the PICASSO database, and this modified score has not undergone full external validation. However, WMH, CMBs/ICH, and lacune are considered the core imaging markers in predicting vascular outcomes, which may mitigate the potential impact of excluding EPVS on our primary findings regarding stroke recurrence risk. 1 3Neurological Sciences (2026) 47:511511 Page 8 of 9

  Keywords: csvd stroke risk patients clinical reflect regarding hemorrhagic score associations burden recurrent interpreted potential picasso