Pediatric Critical Care Medicine最新文献

Objectives: Chylothorax after vascular ring (VR) repair is likely secondary to direct injury to the lymphatic collection system and may benefit from tailored management. We have reviewed the management of patients with chylothorax following VR repair and compared the approaches to chylothorax care after all other cardiac surgeries.

Design: Retrospective multicenter cohort.

Setting: Ten pediatric cardiac surgical centers in the United States.

Patients: Children with chylothorax after VR repair at participating pediatric cardiac surgical centers from 1/1/2021-12/31/2023. Analysis was stratified by chylothorax patients with VR surgery vs all other cardiac surgeries.

Results: Overall, 45 patients had chylothorax following VR repair (45/340, 13.2%) in comparison to 384 patients who had chylothorax following all cardiac surgeries (including VR repair and Fontan procedure) (384/8,494, 4.5%). Compared with all other cardiac surgeries, a higher proportion of VR patients, than other cardiac surgical cases, underwent invasive lymphatic intervention including thoracic duct ligation, embolization, and pleurodesis (10/45 vs. 19/314. p = 0.001) and at an earlier median interval from index surgery (3.5 vs. 28 d, p < 0.001). The 10 patients with VR who underwent a lymphatic intervention had higher chest tube output on postoperative days (PODs) 2 (9.2 vs. 2.6 mL/kg/d, p = 0.005) and 3 (4.9 vs. 0.6 mL/kg/d, p = 0.002) compared with those who resolved without an invasive lymphatic intervention. Following lymphatic intervention, chylothorax resolved a median of 1 day later and there were 0 of 10 procedure-related complications.

Conclusions: Patients with chylothorax after VR more commonly undergo invasive intervention when compared with those with other surgeries. Chest tube output on PODs 2 and 3 may help differentiate VR patients who will need invasive intervention. Chylothorax resolves soon after invasive intervention in VR patients, and in our cohort of 10 there were no procedural-related complications.

Objectives: To improve communication with caregivers who prefer a language other than English (LOE) by increasing interpreting encounters with an emphasis on face-to-face (in-person or video) modalities.

Design: Single-center quality improvement initiative.

Setting: Seventy-five-bed PICU in a quaternary children's hospital.

Patients: Patients whose caregivers preferred a LOE during the period from March 2023 to December 2024.

Interventions: Based on clinician and nursing input about barriers to using interpreting services, we enacted a bundle of interventions collectively known as the Speak Easy program which included: 1) multidisciplinary education; 2) preferred language signs; 3) a novel, standardized opt-out in-person interpreting program for caregivers whose preferred language was Spanish; and 4) increased number of video interpreting devices and presence of devices in patient rooms.

Measurements and main results: We analyzed total and face-to-face interpreting encounters during the intervention period (from March 2023 to December 2024) compared with the pre-intervention period (from November 2021 to February 2023) using statistical process control charts. Total interpreting encounters nearly tripled (from 43 to 121 encounters per 100 LOE patient-days) and face-to-face interpreting encounters increased by more than four times (from 23 to 104 encounters per 100 LOE patient-days). Significant shifts were driven by increases in the number of in-person Spanish interpreting encounters and video interpreting encounters following interventions 3 and 4. Pre-intervention surveys revealed that time constraints, unpredictability, and competing priorities represent barriers to using interpreting services that may be particularly relevant in high-acuity settings; post-intervention surveys showed that clinicians, nurses, and social workers viewed the changes that were made favorably.

Conclusions: A multidisciplinary approach emphasizing interventions to decrease the time and planning necessary to coordinate in-person and video interpreting can effectively engender cultural change and promote the delivery of language-concordant care.

Objectives: High-dose insulin (HDI) is a unique therapy for beta-blocker (BB) and calcium channel-blocker (CCB) poisonings. We have examined pediatric patients with BB and/or CCB poisonings who received HDI therapy with the purpose of describing the clinical characteristics associated with these poisonings and the treatment.

Design: Retrospective database study using our regional, three-state poison center at the Minnesota Regional Poison Center. We identified all children treated with HDI for BB and/or CCB poisonings between the years 2000 and 2024.

Setting: Regional poison center data.

Patients: Pediatric patients 18 years old or younger.

Interventions: None.

Measurements and main results: We identified 36 patients with a median age of 16 years (range 7 mo-18 yr). There were 24 of 36 females, and 14 of 36 patients were poisoned with BBs, 16 of 36 patients by CCBs, and 6 of 36 patients by both drugs. The median peak insulin infusion rate was 1 unit/kg/hr (range 0.5-11 unit/kg/hr); the median insulin infusion duration was 23 hours (range 1-136 hr). The mean dextrose infusion concentration was 37% (range 5-70%). Vasopressors were used in 23 of 36 cases; median vasopressor duration was 38 hours (range 1-199 hr). Cardiac arrest occurred in 4 of 36 patients. Life support with extracorporeal membrane oxygenation (ECMO) was used in one patient. Three patients died as a result of poisoning.

Conclusions: In our three-state poison center, over a 25-year period (2000-2024), HDI was predominantly used in adolescents with intentional BB/CCB overdoses. No adverse events required early discontinuation of HDI. Escalation to ECMO support was rare. More experience is needed to evaluate the safety and effectiveness of HDI in small children.

Objectives: To derive and externally validate supervised machine learning (ML) models predictive of cardiac surgery-associated acute kidney injury (CS-AKI).

Design: Retrospective cohort analysis.

Setting: Multicenter (4), cardiac surgical centers from January 2019 to February 2022.

Patients: Seven days to 18 years old who had undergone cardiac surgery.

Interventions: None.

Measurements and main results: CS-AKI was defined using Kidney Disease: Improving Global Outcomes criteria, with stages 2/3 classified as severe, during the first 7 postoperative days. Data analysis followed two approaches: 1) combining three centers for derivation and using a fourth for external validation and 2) randomly dividing the entire dataset into derivation and validation cohorts in a 4:1 ratio. Forty ML models were developed across five derivation-validation pairs using four ML algorithms (light gradient-boosting machine, extreme gradient boosting, categorical boosting, and histogram gradient boosting) to predict two outcomes (any and severe CS-AKI) utilizing preoperative, intraoperative, and immediate postoperative variables. SHapley Additive exPlanations was used for input variable importance analysis. A cohort of 1100 patients was analyzed. Any CS-AKI and severe CS-AKI occurred in 49.1% and 23.1% patients, respectively. Wide range of variations in external validation of model performance were observed among all 40 ML models. For any CS-AKI, the range in metrics were: area under the receiver operating characteristic curve (AUROC) 0.64-0.83, sensitivity 0.29-0.86, specificity 0.46-0.95, positive predictive value (PPV) 0.50-0.85, and negative predictive value (NPV) 0.60-0.86. For severe CS-AKI, we found the range in metrics with AUROC 0.65-0.77, sensitivity 0.04-0.58, specificity 0.77-0.99, PPV 0.32-0.75, and NPV 0.78-0.90. Preoperative serum creatinine, cardiopulmonary bypass, aortic cross-clamp duration, weight, and age at surgery were the most important predictors associated with CS-AKI.

Conclusions: This analysis of a retrospective multicenter dataset shows that external performance of ML models vary, highlighting challenges in generalizability, which may be due to center-based differences in practice.

Objective: Critically ill children are at risk for preventable morbidities due to immobility. Early mobility, a key component of the ICU Liberation Bundle, improves outcomes and reduces mortality. Internationally, adherence to early-mobility protocols is low and nurses are pivotal for success. This quality improvement (QI) initiative aimed to assess whether an in situ early-mobility simulation for PICU nurses increased nurse-led mobilization of critically ill children.

Design: QI initiative with an observational, pre-post design.

Setting: PICU in a tertiary academic hospital in the United States.

Patients: Critically ill pediatric patients admitted to the PICU.

Intervention: An in situ early-mobility simulation session for PICU nurses.

Measurements and main results: Data were collected and analyzed from February to October 2024 on randomly selected shifts in the pre-intervention (n = 22 day shifts) and post-intervention (n = 26 day shifts) phases. One hundred and one children 1-17 years old who were admitted to the PICU and had length of stay greater than or equal to 72 hours were included. Eighty percent of all critical care nurses (80/100) participated in the simulation session. Post-intervention, the median number of nurse-led mobilizations per patient in a 12-hour shift increased from 5 to 6 (p = 0.02). Participation in the simulation was associated with an increase of 1.9 mobilizations per patient in a 12-hour shift after adjusting for age, illness severity, functional status and mobility level. (p = 0.004). Nursing knowledge of patients' mobility levels improved (p = 0.004), and self-efficacy in mobilizing critically ill children increased from 67% to 93% (p < 0.001). No significant increase in safety events was observed.

Conclusions: In situ early-mobility simulations for PICU nurses increased nurse-led mobilizations of critically ill children without compromising safety. Further research is needed to explore the long-term impact and generalizability of this curriculum.

Objectives: During sepsis, acute brain dysfunction is associated with death and morbidity. However, there are limited data on the pathophysiology of sepsis brain dysfunction.

Design: Retrospective cohort study.

Setting: Two quaternary free-standing children's hospitals (University of Pittsburgh Medical Center [UPMC] Children's Hospital of Pittsburgh and Children's Hospital of Philadelphia [CHOP]).

Subjects: Sepsis patients with a neuropathological autopsy from 2009 to 2018 at UPMC and 2011-2021 at CHOP.

Interventions: None.

Measurements and main results: We identified 57 patients with a median (interquartile range [IQR]) age at admission of 34.5 months (IQR, 9.2-118.0 mo). Overall, 48 of 57 (84%) patients had preexisting comorbidity including: metabolic/genetic disorders (19/57, 33%), immunocompromise (12/57, 21%), and prematurity (12/57, 21%). Almost all patients required inotropes (53/57, 93%) or invasive mechanical ventilation (54/57, 95%). Forty-four percent of patients (25/57) had a cardiac arrest either before or during admission. Almost half of patients had multisite infections (25/57, 44%), with 23 of 25 having a component of bacteremia and four of 25 having a CNS infection. Gross neuropathological findings included edema in 39% (22/57), ventriculomegaly in 39% (22/57), and hemorrhage in 30% (17/57). Microscopic CNS examination results were available for 51 of 57 patients (89%) and the acute/subacute microscopic findings included: neuronal injury in 45% (23/51), acute infarction in 25% (13/51), and white matter necrosis in 25% (13/51). Most patients (49/57, 86%) had at least one acute gross or microscopic finding. In univariate analyses, premortem CNS infections, was associated with edema (p = 0.001). Also, absence of comorbidities was associated with gross edema (p = 0.009) and microscopic acute neuronal injury (p = 0.008), acute infarction (p = 0.01), and meningitis (p = 0.004).

Conclusions: In our retrospective neuropathological study of sepsis cases, we found that pathologies were common. While correlations to sepsis survivors cannot be determined, further studies are needed to develop strategies that prevent and treat neuropathological injury.

Objective: Unfractionated heparin (UFH) is the most widely used anticoagulant during extracorporeal membrane oxygenation (ECMO). Optimal dosing of UFH for children undergoing ECMO is unknown, leading to suboptimal exposure. We aimed to develop a population pharmacokinetic (PK) model for UFH in children undergoing venoarterial (VA) ECMO, using anti-Xa activity, to improve the initial dosing regimen.

Design: A retrospective observational study (March 2015 to May 2021) using nonlinear, mixed-effect modeling software (Monolix) and Monte Carlo simulations to optimize the dosing.

Setting: Single-center study in a PICU in Paris, France.

Patients: All children under 18 years old who needed VA-ECMO and received continuous UFH.

Interventions: None.

Measurements and main results: We included 59 children and 1305 anti-Xa activity results, 13% of which were outside the limits of quantification. Median (range) initial UFH bolus and infusion rate were 50 international units/kg (20-100) and 20 international units/kg/hr (9-40), respectively. Among the 1305 anti-Xa activity results, 875 (67%) were in the therapeutic range (0.3-0.7 international units/mL), whereas 263 (20%) and 167 (13%) were below and above the PK target, respectively. A one-compartment model with first-order elimination and time-varying clearance best fitted the data. Body weight according to allometric scale on clearance and volume of distribution was the selected covariate. Initial clearance was low and increased until steady state was reached after 16 hours. Simulations showed that initial bolus of 40 international units/kg followed by continuous infusion of 25, 20, and 15 international units/kg/hr for patients under 10, between 10 and 15 and above 15 kg, respectively, improved exposure.

Conclusions: In children undergoing VA-ECMO, body weight, and time-varying clearance explained the variability for UFH between subjects. This model can be used to optimize initial UFH dosing.

Objectives: We evaluated the contributions of fluid input and output to fluid overload (FO) and examined associations with renal acute kidney injury (AKI) and mortality. Associations between FO, fluid creep (i.e., from medication dilutions, continuous infusions of sedatives and vasoactive agents), and the Renal Angina Index (RAI) were also evaluated.

Design: Retrospective cohort of critically ill children (1 mo-18 yr old) admitted to the PICU for greater than or equal to 72 hours between 2020 and 2024. Patients with chronic kidney disease or congenital heart disease were excluded. Fluid balance, fluid type and category, renal function, and longitudinal laboratory data were analyzed. FO was defined as cumulative fluid balance greater than 5% or greater than 10% of body weight on day 3. The primary outcome was PICU mortality; secondary outcomes included AKI, mechanical ventilation duration, and PICU length of stay.

Setting: Single-tertiary PICU in a University center.

Patients: Patients hospitalized for greater than or equal to 72 hours.

Interventions: None.

Measurements and main results: Among 217 patients, cumulative FO greater than 5% and fluid creep were each associated with greater odds ratio (OR and 95% CI) of mortality, respectively: OR 4.01 (95% CI, 1.1-14.8) and OR 1.01 (95% CI, 1.00-1.02). Cumulative fluid input, output, and RAI were also associated with FO (p < 0.001). Using area under the receiver operating characteristic curve (AUC) analysis we found that only the pediatric index of mortality score had an excellent discriminating power (AUC 0.92 [95% CI, 0.84-0.99], p < 0.001). All other parameters associated with mortality (i.e., FO, cumulative fluid balance, day-1 fluid creep, RAI, and AKI) were only of "acceptable" discriminating power (AUC 0.70-0.80).

Conclusions: In our 2020-2024 cohort, we have found that FO, including from fluid creep, was associated with greater odds of mortality in critically ill children with admission greater than or equal to 3 days. RAI may also aid early risk-stratification of such patients.