Background The role of dexmedetomidine in patients admitted to the intensive care unit (ICU) following non-cardiac surgery remains unclear. While dexmedetomidine is widely used for its sedative and potential delirium-sparing properties, its pooled effects on clinically relevant postoperative outcomes in this population are not well established. This systematic review and network meta-analysis evaluated the comparative effectiveness and safety of dexmedetomidine compared with other sedative strategies in postoperative non-cardiac surgical ICU patients. Methods We systematically searched PubMed, Cochrane CENTRAL, and Embase from inception to 26 May 2026 for randomised controlled trials (RCTs) comparing postoperative dexmedetomidine with normal saline or other sedatives in adults (≥18 years) admitted to the ICU after non-cardiac surgery. The primary outcome was incidence of delirium during the ICU stay. Secondary outcomes were time to extubation, ICU length of stay, duration of mechanical ventilation, and haemodynamic adverse events. Effect estimates were expressed as relative risks (RRs) or mean differences (MDs) using random-effects models. Risk of bias was assessed using the Cochrane RoB 2 tool, and certainty of evidence was evaluated using the GRADE framework. Interventions were ranked using the surface under the cumulative ranking curve (SUCRA) for each outcome. Results Twenty-four RCTs involving 2,777 patients were included in the systematic review, of which 22 RCTs involving 2,544 patients contributed to the network meta-analysis. Compared with midazolam (RR 0.11, 95% CI 0.02 to 0.80; very low certainty) and normal saline (RR 0.46, 95% CI 0.36 to 0.59; low certainty), dexmedetomidine may reduce the risk of delirium. Evidence is insufficient to determine whether propofol differs from dexmedetomidine in its effect on delirium. There were no differences in time to extubation, duration of mechanical ventilation, or ICU length of stay across sedative strategies. Dexmedetomidine probably increases the risk of bradycardia compared with propofol (RR 3.03, 95% CI 1.39–6.25; low certainty), but the evidence is insufficient to determine whether the two agents differ in their effects on hypotension. In safety rankings, dexmedetomidine performed least favourably for bradycardia (SUCRA 0.03) and hypotension (SUCRA 0.13). Conclusions Dexmedetomidine may reduce postoperative delirium in non-cardiac surgical ICU patients but does not appear to improve ventilatory recovery or ICU length of stay and is associated with higher bradycardia risk. Its principal benefit lies in delirium prevention, supporting an individualised approach to sedative selection.
Leukocyte count is widely available and included in Intensive Care Unit prognostic systems. We hypothesised that the relationship between leukocytosis and mortality risk might differ in critically ill patients admitted with infection, where leukocytosis may be an adaptive response. We performed a registry-based study using the Australian and New Zealand Intensive Care Society Adult Patient Database between 2010 and 2023, across 212 Intensive Care Units. Using descriptive statistics and mixed-effects multivariable logistic regression, we evaluated the association between early peak leukocyte count and mortality, according to whether infection was the primary diagnosis. We examined 2,016,578 patients, of whom 1,742,195 had non-infective illnesses (86.4
Zero-shot learning (ZSL) offers tremendous potential for histopathology image analysis, enabling models to generalize to unseen classes without extensive labeled data. Recent vision-language model (VLM) advancements have expanded ZSL capabilities, allowing task performance without task-specific fine-tuning. However, applying VLMs to histopathology presents considerable challenges due to the complexity of histopathological imagery and the nuanced nature of diagnostic tasks. We propose Multi-Resolution Prompt-guided Hybrid Embedding (MR-PHE), a novel framework for zero-shot histopathology image classification. MR-PHE mimics pathologists' workflow through multiresolution patch extraction to capture key cellular and tissue features. It introduces a hybrid embedding strategy that integrates global image embeddings with weighted patch embeddings, effectively combining local and global contextual information. Additionally, we develop a comprehensive prompt generation and selection framework, enriching class descriptions with domain-specific synonyms and clinically relevant features to enhance semantic understanding. A similarity-based patch weighting mechanism assigns attention-like weights to patches based on their relevance to class embeddings, emphasizing diagnostically important regions during classification. Experimental results demonstrate MR-PHE significantly improves zero-shot classification performance on histopathology datasets, often surpassing fully supervised models, showing its effectiveness and potential to advance computational pathology.
BACKGROUND:Approximately 10% of hospitalized patients globally report a penicillin allergy, leading to inappropriate antibiotic prescribing and inferior healthcare outcomes. Evidence supporting the effectiveness and widespread implementation of inpatient penicillin direct oral challenge (DOC) is limited. METHODS:A prospective, multicenter, international type 2 hybrid effectiveness-implementation study was conducted in 40 hospitals across 8 countries between November 2022 and May 2025. Adult inpatients with a penicillin allergy underwent assessment using a digital penicillin allergy toolkit (National Antibiotic Allergy Network [NAAN] App). According to site clinical practice, participants received penicillin DOC (effectiveness intervention) or assessment only. Participating sites received bimonthly audit and feedback at least 3 months after site activation (implementation strategy). Observational data were used to emulate a target trial to examine secondary effectiveness outcomes, including antibiotic prescribing at 90 days in DOC versus non-DOC participants. The primary implementation outcome was adoption of the NAAN App within 6 months of site activation. RESULTS:Among 5121 participants assessed, 1573 (30.7%) underwent DOC, of which 1502 (95.5%) were delabeled. Of 71 (4.5%) participants with a positive DOC, 6 (0.4%) had a serious adverse event. Of 1852 inpatients in the target trial analysis, 892 underwent DOC and 960 underwent assessment only. Participants who underwent DOC were more likely to be prescribed penicillin (risk ratio [RR], 13.25 [95% confidence interval {CI}, 7.82-22.46]), and less likely to be prescribed World Health Organization (WHO) "Watch" or "Reserve" antibiotics (RR, 0.73 [95% CI, .60-.89]) at 90 days post evaluation. Within 6 months of site activation, 77 clinicians adopted the NAAN App. CONCLUSIONS:Multidisciplinary adoption of inpatient penicillin DOC was safe and significantly improved penicillin prescribing and reduced use of WHO restricted antibiotics.
Background: Robotic-assisted hysterectomy is increasingly performed using modular platforms such as the Hugo™ robotic-assisted surgery (RAS) system, but optimal or personalised docking strategies remain undefined. Objectives: To establish expert consensus on port placement and docking configurations for hysterectomy with the Hugo™ RAS system and to identify patient anthropometric factors requiring modification of standard setups. Methods: A modified Delphi consensus was conducted involving two iterative rounds of anonymous, structured questionnaires distributed to an international panel of gynaecological robotic surgeons experienced with the Hugo™ RAS system. Survey items addressed preferred docking configurations, the influence of patient anthropometry on docking strategy, and specific technical adjustments in non-standard scenarios. Consensus was predefined as ≥66.7% agreement. Main Outcome Measures: Expert agreement on docking setups, port placement modifications, and anthropometric variables influencing technical adjustments. Results: Seventeen experts completed round one and 16 completed round two. No single docking configuration reached consensus as universally optimal for standard hysterectomy. Ranking exercises identified the “standard” hysterectomy setup as the most preferred configuration, followed by the “alternate” and the “three-arm” setups. All experts agreed that patient anthropometry requires modification of port placement. Elevated body mass index (BMI), large uterine size and small pelvis were identified as key variables: increasing inter-port distance was recommended for BMI >30, cranial port displacement for large uteri, while no consensus emerged for patients with a small pelvis. A modified bridge configuration was proposed, and achieved strong expert agreement. Conclusions: No single docking configuration is deemed to be universally optimal for Hugo™ RAS hysterectomy. Expert practice combines a limited number of preferred setups with patient-tailored adjustments. What is New? This study provides the first Delphi-based expert consensus on Hugo™ RAS docking strategies, emphasizing patient-specific adjustments and flexible preoperative planning.
