SEARCH · PubMed Health

Results for “Closed-loop control”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

9 recordsLinked to original sources

PubMed2026

Closed-loop insulin delivery for glycaemic control in hospitalised and perioperative adults: A systematic review and meta-analysis of randomised controlled trials.

We evaluated whether closed-loop insulin delivery improves glycaemic control in hospitalised and perioperative adults. PubMed/MEDLINE, Embase, CENTRAL, and ClinicalTrials.gov were searched from inception to 29 June 2026 for randomised controlled trials comparing closed-loop or automated insulin delivery with usual care or conventional insulin therapy. Random-effects meta-analyses were conducted; risk of bias was assessed using RoB 2 and certainty of evidence using GRADE. Seven trials involving 375 analysed participants were included. Closed-loop insulin delivery increased time in target glucose range by 23.91 percentage points (95% CI 19.40 to 28.43; I2 = 0%) and reduced mean glucose by 1.79 mmol/L (95% CI 1.06 to 2.53 lower; I2 = 36.3%); certainty was moderate for both outcomes. Two trials involving 69 participants reported compatible participant-level data for clinically significant hyperglycaemia, and both estimates favoured closed-loop insulin delivery, although the evidence was exploratory and imprecise. No severe hypoglycaemic events occurred in either group, precluding reliable estimation of comparative safety. Closed-loop insulin delivery may improve glycaemic process measures, but larger pragmatic trials are needed to establish clinical benefits, safety, and implementation feasibility.

Humans↗

PubMed2026

Closed-loop vasopressor systems for hemodynamic control in perioperative and critical care settings: a systematic review and meta-analysis.

Maintaining mean arterial pressure (MAP) within a predefined target is central to haemodynamic management in surgical and critically ill adults receiving vasopressors. Closed-loop vasopressor (CLV) systems automate titration to optimise blood pressure control, but their clinical effectiveness remains uncertain. We performed a systematic review and meta-analysis comparing CLV with manual titration. This PRISMA 2020-compliant review was prospectively registered in PROSPERO (CRD420250655697). MEDLINE, Embase, Scopus, Web of Science, CENTRAL, and the Cochrane Library were searched (January 2000-June 2025). Randomised controlled trials enrolling adults receiving vasopressors in perioperative or intensive care settings were included. Primary outcomes were time within the MAP target range and time spent in hypotension or hypertension. Risk of bias was assessed using RoB 2.0 and certainty of evidence using GRADE. Random- or fixed-effects models were selected according to heterogeneity. Six randomized controlled trials (215 patients) were included in the systematic review, whereas five perioperative trials contributed to the meta-analysis of haemodynamic control outcomes, and one ICU-based study was summarized narratively because it did not report comparable MAP control endpoints. CLV increased time within the MAP target range (mean difference [MD] 33.94%, 95% CI 20.41-47.46; I2 = 77%) and reduced time in hypotension (MD - 18.24%, 95% CI - 28.95 to - 7.53; I2 = 73%). There was no significant difference in time in hypertension, cumulative norepinephrine dose, or major/minor adverse events. ICU length of stay was not pooled because of clinical and methodological heterogeneity. Certainty of evidence ranged from low to high (moderate for haemodynamic control outcomes). CLV systems improved haemodynamic control, primarily in perioperative settings, but heterogeneity and small samples limit confidence in effect size and generalisability. Evidence in critically ill populations remains limited, and larger trials are needed to determine whether improvements in these physiological surrogate endpoints translate into meaningful patient-centred outcomes.

Humans↗

PubMed2026

Glycemic and safety outcomes of the insulin-only bionic pancreas in older adults and individuals with impaired awareness of Hypoglycemia: a post hoc analysis of a randomized pivotal trial.

AIMS: Evaluate the efficacy and safety of iLet Bionic Pancreas (BP) in older adults and individuals with impaired awareness of hypoglycemia (IAH). METHODS: This post hoc analysis used individual participant-level data from the Insulin-Only Bionic Pancreas Pivotal Trial (n = 440; NCT04200313). Eligible participants (n = 96) with type 1 diabetes, aged ≥ 60 years and/or had IAH (Clarke score ≥ 4), were randomized to BP with aspart/lispro (BP-Asp/Lis; n = 45), BP with fast-acting aspart configuration (BP-Fiasp; n = 31), or standard care (SC; n = 20) for 13 weeks. RESULTS: Compared with SC, time-in-range (70-180 mg/dL) significantly increased by 7.49 % (95 % CI: 2.61 to 12.38; ∼1.8 h/day) with BP-Asp/Lis and by 8.28 % (95 % CI: 3.15 to 13.41; ∼2.0 h/day) with BP-Fiasp, driven by reduced hyperglycemia. No significant differences were observed in hypoglycemia exposure. Severe hypoglycemia occurred in four participants (four events) on BP-Asp/Lis and one participant (two events) on SC. One diabetic ketoacidosis event occurred on BP-Fiasp due to an infusion set failure. CONCLUSIONS: In high-risk, clinically vulnerable populations, the BP system significantly improved glycemic control while maintaining safety parity with respect to hypoglycemia risk, providing a resilient therapeutic alternative for vulnerable cohorts.

Humans↗

PubMed2026

Early Worsening of Diabetic Retinopathy Following Initiation of Hybrid Closed-Loop/Automated Insulin Delivery Systems in Type 1 Diabetes: A Systematic Review and Structured Study-Level Synthesis.

BACKGROUND: Hybrid closed-loop (HCL) systems achieve rapid, algorithm-driven improvements in glycaemia in type 1 diabetes (T1D). Paradoxically, rapid improvement in glycaemic control is associated with early worsening of diabetic retinopathy (EWDR), a phenomenon established in the intensive insulin therapy era. Whether HCL initiation carries a clinically meaningful EWDR risk is unknown. No systematic review has previously addressed this question. METHODS: A systematic review and structured quantitative synthesis was performed using study-level estimates only (PROSPERO CRD:420261391951). MEDLINE, SCOPUS and Web of Science were searched to 14th May 2026. Studies reporting retinal outcomes in people with T1D initiating any HCL system were eligible. Two reviewers independently screened studies and extracted data. Risk of bias was assessed using ROBINS-I and certainty of evidence using the GRADE framework. EWDR incidence was summarised using study-level proportions, and comparative studies were summarised using study-specific risk ratios for HCL versus control therapy. Given substantial heterogeneity in EWDR definitions, retinal assessment timing, follow-up duration, and comparator groups, no pooled or meta-analytic estimates were derived. RESULTS: Eight studies (n = 1487 participants; 860 HCL users) were included; all were observational and six were retrospective. EWDR varied markedly with the timing of retinal assessment. In studies assessing the retina within ≤ 12 months of HCL initiation, EWDR rates ranged from 8.9% to 26.5%. Studies with longer follow-up reported lower rates of retinal worsening or incident DR, 6.7% at 24 months and 6.1% over a mean follow-up of 4.9 years, suggesting that these studies may capture background DR progression rather than true early worsening. Three comparative studies included 177 HCL users and 315 controls; EWDR study-specific risk ratios were directionally inconsistent, ranging from 0.32 to 1.51, and were therefore not pooled. The most consistently identified risk factors were higher baseline HbA1c and older age. The magnitude of HbA1c reduction was not a consistent predictor of EWDR in the HCL context, in contrast to pre-HCL era evidence. Risk of bias ranged from moderate to critical and certainty of evidence was very low for all outcomes. CONCLUSIONS: Study-defined retinal worsening was reported in a minority of participants. The current evidence base is dominated by retrospective studies, variable retinal assessment timing, and inconsistent EWDR definitions. Well-designed prospective studies with protocol-specified retinal surveillance anchored to HCL initiation are required to generate reliable incidence estimates, identify risk factors, determine visual consequences, and inform standardised screening guidance.

Humans↗

PubMed2026

Deep generative models in biological sequence and structure analysis and design.

Deep generative models have transformed biological sequence modeling from predictive analysis toward increasingly controllable design. Early biological applications of Variational Autoencoders (VAEs) and Generative Adversarial Networks (GANs) established latent representation learning and sequence synthesis, while recent advances in transformer-based language models, discrete diffusion, flow-matching, and multimodal generative frameworks have substantially expanded the scope of biological design. This review examines generative models for DNA, RNA, and protein sequence design, emphasizing how different model classes represent biological constraints, operate over discrete and continuous spaces, and integrate sequence, structure, and function. We compare VAEs, GANs, autoregressive and masked language models, diffusion models, and flow-based approaches across genomics, transcriptomics, and proteomics, with particular attention to controllability, long-range dependency modeling, structural grounding, generalization, and experimental utility. We further examine evaluation strategies, out-of-distribution generalization, and closed-loop design-build-test-learn workflows that connect in silico generation with empirical validation. We distinguish fundamental modality-dependent constraints including sequence discreteness, context length, structural coupling, and physical or thermodynamic requirements from architecture-dependent advantages that reflect the current state of the field. Current studies suggest that long-context models are particularly useful for genome-scale representation and sequence modeling, whereas structure-aware diffusion, flow-based, and inverse-folding approaches provide better frameworks for geometry-constrained RNA and protein design. This perspective provides a critical framework for understanding the present capabilities, limitations, and convergence of generative approaches toward reliable and experimentally grounded biological design.

Biological sequence analysis↗

PubMed2026

Alarms and alarm management with automated versus conventional ventilation in neurocritical care patients.

INTRODUCTION: False or clinically irrelevant alarms are a major driver of ICU alarm fatigue and nursing workload. Ventilator alarms make up a large share, and although automated ventilation modes can reduce manual adjustments, their effect on alarm burden is still unclear. This issue can be particularly relevant in neurocritical care patients, where precise ventilator and alarm management is imperative for patient safety. OBJECTIVES: This explorative post hoc analysis of a randomized clinical trial compared alarm frequency and management between automated ventilation and conventional ventilation in neurocritical care patients. METHODS: Ventilator alarms and manual ventilator changes were captured continuously from the ventilator for up to 24 h per patient. The primary endpoint was a composite of workload-relevant alarms; with alarm management interventions at the ventilator as a key secondary outcome. Additional endpoints included redundant alarms, alarm duration and ventilator management. RESULTS: 13 patients received automated ventilation and 24 received conventional ventilation. No difference was observed in workload-relevant alarm frequency between automated and conventional ventilation (3.28 [2.87 to 4.30] vs 3.73 [1.66 to 7.33] alarms per hour; P = 0.81), while alarm management interventions at the ventilator were lower with automated ventilation (0.14 [0.10 to 0.15] vs 0.21 [0.17 to 0.31] interventions per hour; P = 0.01). Other alarm frequencies, duration of alarms and ventilator management were similar. CONCLUSIONS: In this exploratory post hoc analysis of a randomized clinical trial in neurocritical care patients during the early phase of mechanical ventilation, automated ventilation did not reduce the frequency of total or workload-relevant alarms, nor their duration, but was associated with fewer alarm management interventions compared to conventional ventilation. IMPLICATIONS FOR CLINICAL PRACTICE: Automated ventilation may not reduce alarm frequency in neurocritical care patients, but the observed reduction in alarm-related bedside interventions suggests a potential benefit for nursing workload.

Humans↗

PubMed2025

Deep learning guided programmable design of Escherichia coli core promoters from sequence architecture to strength control.

Core promoters are essential regulatory elements that control transcription initiation, but accurately predicting and designing their strength remains challenging due to complex sequence-function relationships and the limited generalizability of existing AI-based approaches. To address this, we developed a modular platform integrating rational library design, predictive modelling, and generative optimization into a closed-loop workflow for end-to-end core promoter engineering. Conserved and spacer region of core promoters exert distinct effects on transcriptional strength, with the former driving large-scale variation and the latter enabling finer gradation. Based on this insight, Mutation-Barcoding-Reverse Sequencing approach was used and constructed a synthetic promoter library comprising 112 955 variants with minimal redundancy and a 16 226-fold expression range. A Transformer-based model trained on this dataset achieved a Pearson correlation of 0.87 with experimentally measured promoter strengths. When combined with a conditional diffusion model, the system enabled de novo generation of promoter sequences with defined strengths, achieving a design-to-measurement correlation of 0.95 and maintaining high accuracy (R = 0.93) across varied sequence contexts. The designed promoters consistently preserved their intended strength gradients, demonstrating robust plug-and-play functionality. This work establishes a scalable and extensible platform (www.yudenglab.com) for deep learning-guided programmable design of Escherichia coli core promoters, enabling precise transcriptional control.

Promoter Regions, Genetic↗

PubMed2026

Beyond Glycaemia: Fear of Hypoglycaemia, Cognition and Functional Mobility After Advanced Hybrid Closed-Loop Therapy in Older Adults With Type 1 Diabetes: A Prespecified Secondary Analysis of a Randomised, Single-Centre Study.

BACKGROUND: Evidence on psychological, cognitive and functional outcomes of advanced diabetes technologies in older adults with long-standing type 1 diabetes (T1D) remains limited. We evaluated whether initiation of advanced hybrid closed-loop (AHCL) therapy was associated with changes in fear of hypoglycaemia, diabetes distress, psychological well-being, cognition, frailty-related measures and mobility-related function in adults aged ≥ 65 years with T1D. METHODS: This prespecified, exploratory secondary analysis was conducted within a single-centre, open-label, randomised, controlled, parallel-group trial including adults aged ≥ 65 years with long-standing T1D. Participants were randomly assigned (1:1) to initiate AHCL therapy using the MiniMed 780G system or to continue standard diabetes treatment. The secondary outcomes included WHO-5, the 17-item Diabetes Distress Scale (DDS), Hypoglycemia Fear Survey-II (HFS-II), Montreal Cognitive Assessment, Digit Symbol Substitution Test, Fried frailty phenotype and performance-based functional measures. No formal sample-size calculation was performed for these secondary outcomes. RESULTS: Thirty-one participants were randomised and 29 completed 12 months of follow-up and were included in the treatment-effect analyses. In the baseline-adjusted primary analysis, AHCL therapy was associated with a lower HFS-II score than standard treatment (adjusted mean difference -18.9; 95% CI: -32.4 to -5.4; nominal p = 0.008), although this finding did not remain statistically significant after Holm correction (adjusted p = 0.104) or in an exploratory model additionally adjusted for sex (difference -13.6; 95% CI: -32.2 to 5.0; p = 0.145). Diabetes distress, psychological well-being, global cognition and processing speed did not differ between groups. In sex-adjusted sensitivity analyses, the between-group differences remained statistically significant for 6-min walk distance (92.6 m; 95% CI: 36.8 to 148.3; p = 0.002) and Timed Up and Go performance (-2.27 s; 95% CI: -4.28 to -0.27; p = 0.028), but not for gait speed (0.27 m/s; 95% CI: -0.05 to 0.59; p = 0.099). At 12 months, 12 of 14 AHCL participants were robust and 2 were pre-frail; in the control group, 11 of 15 were robust and 4 were pre-frail. No participant was classified as frail at follow-up. CONCLUSIONS: In this small, selected cohort, AHCL therapy was associated with a nominally lower fear-of-hypoglycaemia score and better performance on selected mobility-related tests over 12 months. The fear-of-hypoglycaemia finding did not remain statistically significant after correction for multiple comparisons or additional adjustment for sex. Six-minute walk distance and Timed Up and Go remained statistically significant in the exploratory sex-adjusted sensitivity analyses, whereas the gait-speed difference did not. No measurable between-group deterioration in global cognition or processing speed was observed. These exploratory findings require confirmation in larger studies with balanced representation by sex and direct measurement of physical activity. These findings also support a person-centred clinical message: older age alone should not be regarded as a barrier to AHCL when treatment is introduced with individualised education and appropriate ongoing support.

Humans↗

PubMed2026

Upcycling Vegetable Waste Into Functional Food Ingredients via Synergistic Microbial Engineering and Artificial Intelligence.

The escalating generation of global vegetable waste represents a critical loss of bioactive resources, necessitating a paradigm shift from passive disposal to active nutrient upcycling. However, the industrial conversion of this heterogeneous biomass into standardized functional food ingredients is currently impeded by significant techno-economic barriers, primarily structural recalcitrance, compositional inconsistency, and the presence of toxic fermentation inhibitors. This review provides a comprehensive analysis of the synergistic application of microbial engineering and artificial intelligence (AI) to resolve these bioprocessing bottlenecks within a food-to-food closed-loop framework (as shown in the graphical abstract). We evaluate recent advances in engineering food-grade microbial chassis (e.g., Saccharomyces cerevisiae and Escherichia coli) to enhance lignocellulose degradation and stress tolerance. Concurrently, we examine the integration of AI across the entire value chain, covering deep learning-based rational enzyme design, genome-scale metabolic modeling, and intelligent process control for precision fermentation. Current evidence demonstrates that the hardware-software coupling of engineered strains and AI algorithms significantly enhances conversion efficiency and process robustness. Key findings highlight that AI-driven Design-Build-Test-Learn cycles facilitate the de novo creation of enzymes with superior kinetics and strains with adaptive stress response capabilities against toxins. Moreover, dynamic digital twin models effectively mitigate the impact of substrate variability, ensuring the batch-to-batch consistency required for food applications. We conclude that this data-driven synergistic paradigm is pivotal for establishing a resilient circular bioeconomy, enabling the reliable bioconversion of waste into high-value single-cell proteins, natural flavor additives, and sustainable packaging materials.

Artificial Intelligence↗