The rapid deployment of mRNA vaccines during the COVID-19 pandemic exposed limitations in traditional pharmacovigilance systems, including delayed reporting, high underreporting rates, and inability to calculate true incidence. Machine learning (ML) offers new pathways to overcome these challenges by integrating multi-source real-world data.

We systematically reviewed English-language studies from database inception to June 2026. Searches were performed in PubMed, Embase, and Web of Science. Two reviewers independently screened records. Given substantial heterogeneity across ML tasks (signal detection, text extraction, risk prediction, prognosis stratification), algorithms, data sources, and metrics, we performed narrative synthesis. Risk of bias was assessed using adapted QUADAS-2.

We identified 43 studies. For adverse-event prediction, tree-based models reported AUCs of 0.85-0.87, though estimates derive from heterogeneous settings. NLP reduced redundant signals by 17% in vaccine reporting systems. For myocarditis, ML models reached AUCs up to 0.899 in cardiovascular cohorts, but direct mRNA vaccine applications remain limited and retrospective. Emerging platforms (self-amplifying and tumor mRNA vaccines) lack post-marketing data, rendering ML applications largely conceptual.

ML-assisted pharmacovigilance enables a shift from passive to active, intelligent monitoring. Despite challenges in data quality, model interpretability, and regulatory approval, intelligent pharmacovigilance systems will become essential infrastructure for safeguarding public health.

Highly pathogenic avian influenza (HPAI) has circulated in West Africa since Nigeria's first human case in 2006, making the region a key entry point and dispersal hub for multiple virus lineages in sub-Saharan Africa. The 2.3.4.4b virus clade driving the global spread of avian influenza is now infecting mammals more often, a troubling sign that raises the risk of a future pandemic. At the same time, West Africa's ability to control HPAI is held back by fragmented surveillance systems. Animal, human and environmental health services operate in isolation, rely on incompatible data formats and are managed by separate institutions, which undermines early detection and rapid response. This paper describes the existing surveillance silos in West Africa and outlines the consequences of failing to integrate them, including delayed detection of future human or animal health threats. It highlights emerging examples of One Health surveillance efforts in Nigeria, Senegal, Guinea, Benin and Ghana. From an operational standpoint, five conditions are essential for early warning and timely response: interoperable health information systems; shared governance across sectors; sustainable financing; strong communication with poultry-keeping communities; and effective cross-border collaboration. Establishing fully integrated, system-wide national surveillance platforms is indispensable for effective preparedness. Without such integration, West African countries will continue to detect and respond to emerging health threats too late. Given the region's recurrent zoonotic spillover and expanding ecological vulnerabilities, embedding data integration at the core of preparedness strategies is essential for mitigating the risk of future pandemics.

The COVID-19 pandemic was accompanied by a widespread infodemic characterised by rapidly evolving information, misinformation, and uncertainty, particularly within low- and middle-income country contexts. Social listening emerged as a critical tool for risk communication and community engagement by enabling real-time monitoring of public concerns, misinformation, and information gaps. This study presents a systematic evidence synthesis of national COVID-19 social listening reports produced in South Africa between 2021 and 2023 to inform infodemic management in the Global South. A total of 91 publicly available social listening reports were identified through coordinated national repositories and screened according to predefined inclusion and exclusion criteria. Following PRISMA 2020 guidance for transparent reporting, misinformation signals and corresponding recommended actions were extracted, thematically coded, and synthesised. The synthesis identified 964 instances of misinformation across 12 thematic categories, with dominant themes relating to COVID-19 disease management, vaccine safety, side effects, and conspiracy narratives. In parallel, 573 recommended action items were consolidated into eight thematic response domains, emphasising community-based engagement, targeted health education, communication strategy development, political and public figure credibility, and continuous monitoring and research. Building on these findings, the study proposes a circular social listening health priority nexus that operationalises the translation of misinformation signals into adaptive public health communication responses. While this synthesis does not evaluate the effectiveness of implemented interventions, it provides structured evidence to support decision-making and priority setting for infodemic management. The findings highlight the importance of context-sensitive, adaptive communication strategies and sustained social listening systems to strengthen public health responses to future infodemics in the Global South.

Previous studies suggest that Yinaokang Capsules (YNKC) may offer therapeutic benefits for acute ischemic stroke (AIS) by reducing atherosclerosis, inhibiting inflammatory mediators, and exhibiting neuroprotective properties. Nonetheless, there is currently a lack of substantial clinical data that validates the overall efficacy and safety of YNKC. Furthermore, the systemic metabolic changes that underpin its neurovascular protective effects have not been thoroughly investigated in clinical contexts.

This study is a prospective, randomized, open-label, blinded-endpoint trial designed to enroll 80 patients diagnosed with AIS aged between 18 and 80 years at the Guangdong Provincial Hospital of Chinese Medicine. Eligible participants must present within 72 h of symptom onset and have a baseline National Institutes of Health Stroke Scale (NIHSS) score ranging from 4 to 22. Participants will be randomly assigned in a 1:1 ratio to either the YNKC group or the control group. Those in the YNKC group will receive four YNKC orally three times daily for a duration of 14 days. All participants will concurrently receive standard guideline-based medical treatment. The primary efficacy outcome will be the proportion of patients achieving a favorable functional outcome, defined as a modified Rankin Scale (mRS) score of 0 to 1 at day 90 following randomization. Secondary outcomes included the overall distribution of mRS scores, NIHSS score, Montreal Cognitive Assessment (MoCA) scores, Traditional Chinese Medicine (TCM) Syndrome Scores, Barthel Index (BI) score, the Stroke-Specific Quality of Life (SS-QoL) scores. The main safety outcome will be the incidence of adverse events (AEs) and serious adverse events (SAEs). Additionally, targeted blood metabolomics will be conducted to explore the underlying metabolic mechanisms of YNKC.

Aims to evaluate the impact of YNKC on day 90 functional prognosis of patients with AIS, and to explore its underlying metabolic mechanisms.

itmctr.ccebtcm.org.cn, ITMCTR2025001037.

We performed this systematic review and meta-analysis to assess the efficacy and safety of anticoagulation therapy versus no anticoagulation in pediatric patients with cerebral sinovenous thrombosis (CSVT). Eligible studies included pediatric patients with CSVT that compared anticoagulation to no anticoagulation and reported on at least one relevant outcome (mortality, neurological deficit, thrombus resolution, recurrence, and bleeding). Meta-analysis reported risk ratios (RR) or differences (RD) (95% CIs) and absolute effects per 1000 patients. Risk of bias was assessed using ROBINS-I, and the certainty of evidence was evaluated with GRADE. After removing duplicates, we screened 8925 records independently and in duplicate, from which 12 nonrandomized studies of intervention were included. Compared with no anticoagulation, anticoagulation may be associated with lower risk of overall mortality RR 0.14 (0.05-0.39) (4 studies, n = 448), lower risk of neurological deficits RD -0.13 (-0.39 to 0.13) (6 studies, n = 462), and higher rates of complete and partial resolution RR 1.25 (0.94-1.66) (7 studies, n = 150). Recurrence RD 0 (-0.11 to 0.11) (2 studies, n = 56) and bleeding RD 0.03 (-0.09 to 0.15) (5 studies, n = 95) may be comparable between anticoagulation versus no anticoagulation. We judged outcomes to be very low to low certainty of evidence due to risk of bias, concerns from confounding, and imprecision from small sample sizes. To conclude, we show in this systematic review that anticoagulation therapy in pediatric patients with CSVT may be associated with better outcomes (reduced mortality, neurological deficits, and higher rate of thrombus resolution) with little to no effect on bleeding and recurrence.

Thrombocytopenia in APS is often regarded primarily as a bleeding-related manifestation, but the relationship between platelet-count severity and thrombotic manifestations remains uncertain. We evaluated the association between severe thrombocytopenia and arterial thrombotic manifestations in patients with antiphospholipid syndrome (APS).

In a two-center baseline cohort of 313 patients with APS, baseline platelet count (PLT) was categorized as PLT <50×10⁹/L, PLT 50-<100×10⁹/L, or PLT ≥100×10⁹/L. Associations with prevalent baseline thrombotic manifestations were assessed using multivariable logistic regression in 300 patients with primary APS or systemic lupus erythematosus-associated APS. In the Renji follow-up cohort (n=254), 48-month thrombotic events were evaluated using Kaplan-Meier curves and Cox models.

PLT <50×10⁹/L was associated with a higher frequency of arterial thrombosis (66.7%), particularly cerebral infarction (62.7%), whereas venous thrombosis was most frequent in the PLT 50-<100×10⁹/L group (68.4%). After multivariable adjustment, PLT <50×10⁹/L remained associated with a higher prevalence of arterial thrombosis (aOR 2.16, 95% CI 1.07-4.35, P = 0.032) and cerebral infarction (aOR 2.37, 95% CI 1.18-4.73, P = 0.015). During 48-month follow-up, PLT <50×10⁹/L was associated with increased risks of any thrombotic event (HR 3.28, 95% CI 1.54-6.97, P = 0.002), arterial thrombotic event (HR 3.95, 95% CI 1.39-11.21, P = 0.010), and cerebral infarction (HR 4.11, 95% CI 1.44-11.73, P = 0.008), but not venous-related events.

Severe thrombocytopenia was associated with arterial thrombotic manifestations in APS, particularly cerebral infarction. PLT <50×10⁹/L may provide additional clinical risk information but may also reflect cumulative disease burden and overall APS severity.

To evaluate the exploratory "Collapsed Furnace" hypothesis that nutritional depletion and sarcopenia-related frailty may alter responses to ambient cold in older adults with stroke.

We integrated Global Burden of Disease (GBD) 2023 estimates for China (1990-2021), exploratory physiological data from the China Health and Retirement Longitudinal Study (CHARLS) 2015, and a retrospective clinical stroke cohort. The primary clinical analysis included patients aged ≥70 years; the full cohort aged ≥67 years was retained for sensitivity analysis. Nutritional risk was assessed using admission serum albumin as a nutritional-inflammatory marker, with the body mass index (BMI)-adjusted modified Geriatric Nutritional Risk Index (mGNRI) used as a sensitivity measure. Multivariable models examined nutritional risk, cold exposure, and their interaction in relation to admission National Institutes of Health Stroke Scale (NIHSS) score, adjusting for prespecified demographic, vascular, lifestyle, stroke-type, BMI, and C-reactive protein (CRP) covariates.

Sodium-attributable stroke burden peaked at ages 70-79 years, whereas cold-attributable burden increased in the oldest groups. In the GBD analysis, the malnutrition × low-temperature interaction was negative (β = -12.20, 95% confidence interval [CI], -18.52 to -5.88; p < 0.001). In the clinical cohort, nutritional risk was strongly associated with higher admission NIHSS scores. Cold exposure was associated with only a small additional NIHSS increase among patients with normal albumin and no measurable increase among those with low albumin; however, the interaction was not statistically significant. mGNRI- and CRP-based sensitivity analyses were directionally similar. CHARLS suggested different seasonal blood-pressure patterns by physiological status, but interpretation was limited by sparse cold-season interviews. Only 7 cold-exposed patients with low albumin aged ≥70 years and 4 CHARLS participants aged ≥70 years with cold-season interviews were available.

Nutritional depletion was strongly associated with stroke severity, but its role in modifying cold-related vulnerability remains uncertain. Because the clinical interaction was not significant and mortality, pre-stroke albumin, individual cold exposure, and core temperature were unavailable, the proposed ceiling-like effect should be considered hypothesis-generating and requires prospective validation.

The use of antiplatelet agents is essential in stent-assisted coiling (SAC) for the treatment of intracranial aneurysms, and the preoperative assessment of antiplatelet efficacy is important for reducing the risk of ischemic complications. Light transmission aggregometry (LTA) is used to assess antiplatelet drug efficacy, and we analyzed the association between patient background factors and perioperative complications in patients who underwent SAC using a retrospective analysis.

Patients who underwent SAC for unruptured cerebral aneurysms from March 2017 to June 2024 were included. Dual antiplatelet therapy was administered at least 7 days before the procedure, and LTA was measured on the day of the procedure in all included patients and included in the main analysis. An additional preprocedural LTA measurement was available in 215 patients and was used to guide adjustment of antiplatelet therapy. The association between patient characteristics, ischemic complications, and postoperative diffusion-weighted imaging (DWI) positivity was evaluated. Propensity score matching, with the treatment period included as a covariate, was performed to investigate the association between adenosine diphosphate (ADP)-induced LTA values and DWI positivity.

A total of 1,021 unruptured aneurysms received endovascular treatment, of which 548 underwent SAC [453 Neuroform Atlas, 17 Enterprise, and 78 low-profile visualized intraluminal support (LVIS)]. The mean ADP-induced LTA value was 44.2 (range, 20-74). Symptomatic ischemic complications were present in two patients (0.4%). The overall rate of DWI-positive lesions was 52.7% (289/548). Propensity score matching with treatment period included as a covariate resulted in 152 matched pairs. Covariate balance improved after matching, with most covariates showing absolute standardized mean differences of <0.1, although some residual imbalance remained. ADP-induced LTA value was significantly higher in the DWI-positive group than in the DWI-negative group (45.89vs 43.09, p = 0.019).

Higher preoperative ADP-induced LTA values were associated with postoperative DWI-positive lesions after SAC. However, given the modest discriminative performance, ADP-induced LTA values should be interpreted as a supplementary reference rather than a standalone criterion for clinical decision-making.

Inflammation plays a critical role in acute ischemic stroke (AIS) pathophysiology and may influence outcomes after endovascular treatment (EVT). However, the incremental prognostic value of inflammatory cytokines beyond conventional predictors remains unclear. This study aimed to investigate the association of cytokines with 90-day functional outcomes in EVT-treated AIS patients and evaluate their value in prognostic modeling.

Based on the prospective DETECT2-China registry, we included consecutive large vessel occlusion patients undergoing EVT between March 2022 and March 2025. Plasma cytokines were measured within 24 h post-EVT. The primary outcome was 90-day poor functional outcome (modified Rankin Scale [mRS] 3-6). Multivariable logistic regression identified independent predictors. Clinical, cytokine, and combined models were constructed and evaluated using receiver operating characteristic (ROC) curves, calibration, decision curve analysis (DCA), and reclassification metrics.

Among 139 included patients, 77 (55.4%) had 90-day poor outcomes. Serum IL-6 and IL-10 levels were significantly higher in these patients. In multivariable analysis, log₂-IL-6 was an independent predictor (OR = 1.43 per doubling, 95% CI: 1.14-1.79, p = 0.002), and log₂-IL-10 was significant in Adjusted Model 2 (OR = 1.86 per doubling, 95% CI: 1.02-3.36, p = 0.041). The combined model achieved superior discrimination (AUC: 0.841, 95% CI: 0.770-0.905) compared with the clinical model (AUC: 0.791, 95% CI: 0.717-0.865; p = 0.015), with good calibration and clinical net benefit. Reclassification was also improved (continuous NRI: 43.32%, p = 0.009; IDI: 7.27%, p < 0.001).

Serum IL-6 is independently associated with 90-day poor outcome While incorporating cytokines into clinical models may improve prognostic estimation, these exploratory findings require validation in independent cohorts before clinical application.

Spasticity is a common motor disorder that develops after stroke, negatively impacting motor control and functional capacity. Non-pharmacological methods, including transcranial direct current stimulation (tDCS), are increasingly used in its treatment; however, the optimal stimulation protocol remains unclear. Therefore, the aim of this study was to conduct a systematic review of the literature regarding the efficacy of tDCS in treating upper limb spasticity in post-stroke patients.

A systematic review of randomized controlled trials was conducted in accordance with the PRISMA guidelines. PubMed, Scopus, PEDro, Cochrane CENTRAL, ScienceDirect, Web of Science, and Google Scholar were searched using the following keywords and their combinations: transcranial direct current stimulation, stroke, muscle spasticity and upper limb. The eligibility criteria were defined according to the PICOTS framework. The risk of bias in the included studies was assessed using the Cochrane Risk of Bias 2 tool. The review was registered in PROSPERO (CRD420261363944).

Of the 2,797 records identified, nine publications met the eligibility criteria and were included in the analysis. These studies differed in terms of participant characteristics. In most cases, the intervention involved anodal stimulation of the primary motor cortex; however, the treatment parameters, including stimulation duration, current intensity, and number of sessions, were heterogeneous. A clear beneficial effect of tDCS on upper limb spasticity in patients after stroke was observed in three articles.

The available evidence regarding the effectiveness of tDCS in reducing upper limb spasticity after stroke remains inconclusive. Further high-quality studies using standardized research protocols are required.

https://www.crd.york.ac.uk/PROSPERO/view/CRD420261363944, PROSPERO; registration number CRD420261363944.