Association of temperature management strategy with fever in critically ill children after out-of-hospital cardiac arrest.

Objective: To determine whether ICU temperature management strategy is associated with fever in children with return of spontaneous circulation (ROSC) after out-of-hospital cardiac arrest (OHCA). Methods: We conducted a single-center retrospective cohort study at a quaternary Children's hospital between 1/1/2016-31/12/2020. Mechanically ventilated children (<18 y/o) admitted to Pediatric or Cardiac ICU (PICU/CICU) with ROSC after OHCA who survived at least 72 h were included. Primary exposure was initial PICU/CICU temperature management strategy of: (1) passive management; or (2) warming with an air-warming blanket; or (3) targeted temperature management with a heating/cooling (homeothermic) blanket. Primary outcome was fever (≥38°C) within 72 h of admission. Results: Over the study period, 111 children with ROSC after OHCA were admitted to PICU/CICU, received mechanical ventilation and survived at least 72 h. Median age was 31 (IQR 6-135) months, 64% (71/111) were male, and 49% (54/111) were previously healthy. Fever within 72 h of admission occurred in 51% (57/111) of patients. The choice of initial temperature management strategy was associated with occurrence of fever (χ2 = 9.36, df = 2, p = 0.009). Fever occurred in 60% (43/72) of patients managed passively, 45% (13/29) of patients managed with the air-warming blanket and 10% (1/10) of patients managed with the homeothermic blanket. Compared to passive management, use of homeothermic, but not of air-warming, blanket reduced fever risk [homeothermic: Risk Ratio (RR) = 0.17, 95%CI 0.03-0.69; air-warming: RR = 0.75, 95%CI 0.46-1.12]. To prevent fever in one child using a homeothermic blanket, number needed to treat (NNT) = 2. Conclusion: In critically ill children with ROSC after OHCA, ICU temperature management strategy is associated with fever. Use of a heating/cooling blanket with homeothermic feedback reduces fever incidence during post-arrest care.

Classification of Activated Microglia by Convolutional Neural Networks.

Microglia are the resident macrophages in the central nervous system. Brain injuries, such as traumatic brain injury, hypoxia, and stroke, can induce inflammatory responses accompanying microglial activation. The morphology of microglia is notably diverse and is one of the prominent manifestations during activation. In this study, we proposed to detect the activated microglia in immunohistochemistry images by convolutional neural networks (CNN). 2D Iba1 images (40µm) were acquired from a control and a cardiac arrest treated Sprague-Dawley rat brain by a scanning microscope using a 20X objective. The training data were a collection of 54,333 single-cell images obtained from the cortex and midbrain areas, and curated by experienced neuroscientists. Results were compared between CNNs with different architectures, including Resnet18, Resnet50, Resnet101, and support vector machine (SVM) classifiers. The highest model performance was found by Resnet18, trained after 120 epochs with a classification accuracy of 95.5%. The findings indicate a potential application for using CNN in quantitative analysis of microglial morphology over regional difference in a large brain section.

Children With Medical Complexity: A Web-Based Multimedia Curriculum Assessing Pediatric Residents Across North America.

OBJECTIVE: No standardized curricula exist for training residents in the special needs of children with medical complexity. We assessed resident satisfaction, knowledge, and behavior after implementing a novel online curriculum composed of multimedia modules on care of children with medical complexity utilizing virtual simulation. METHODS: We conducted a randomized controlled trial of residents across North America. A Web-based curriculum of 6 self-paced, interactive, multimedia modules was developed. Readings for each topic served as the control curriculum. Residents were randomized to 1 of 2 groups, each completing 3 modules and 3 sets of readings that were mutually exclusive. Outcomes included resident scores on satisfaction, knowledge-based assessments, and virtual simulation activities. RESULTS: Four hundred forty-two residents from 56 training programs enrolled in the curriculum, 229 of whom completed it and were included in the analysis. Subjects were more likely to report comfort with all topics if they reviewed modules compared to readings (P ≤ .01 for all 6 topics). Posttest knowledge scores were significantly higher than pretest scores overall (mean increase in score 17.7%; 95% confidence interval 16.0, 19.4), and the mean pre-post score increase for modules was significantly higher than readings (20.9% vs 15.4%, P < .001). Mean scores on the verbal handoff virtual simulation increased by 1.1 points (95% confidence interval 0.2, 2.0, P = .02). There were no significant differences found in pre-post performance for the device-related emergency virtual simulation. CONCLUSIONS: There was high satisfaction, significant knowledge acquisition, and specific behavior change after participating in this innovative online curriculum. This is the first multisite, randomized trial assessing satisfaction, knowledge impact, and behavior change in a virtually simulated environment with pediatric trainees.