Fluid Accumulation in Mechanically Ventilated, Critically Ill Children: Retrospective Cohort Study of Prevalence and Outcome

Fluid Management in Critically Ill Children: Single-Center Retrospective Comparison of Trauma and Postoperative Patients, 2020-2022

OBJECTIVE

Injury and surgery both represent well-defined starting points of a predictable inflammatory response, but the consequent response to IV fluids has not been studied. We aimed to review and compare our single-center fluid management strategies in these two populations.

DESIGN

Retrospective cohort study from January 2020 to July 2022. The primary outcome was total IV fluid volume administered. Net fluid balances and select clinical outcomes were also evaluated.

SETTING

Single tertiary academic center and level 1 pediatric trauma center in New York.

PATIENTS

A dataset of critically ill trauma and surgical patients aged 0-18 years who were admitted to the PICU, 2020-2022. Trauma patients had at least moderate traumatic injuries (Injury Severity Score ≥ 9) and surgical patients had at least a 1-hour operation time.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

We identified 25 trauma and 115 surgical patients. During the first 5 days of hospitalization, we did not identify an association between grouping and total IV fluids administered and fluid balance in the prehospital, emergency department, and operating room (p = 0.90 and p = 0.79), even when adjusted for weight (p = 0.96). Time trend graphs of net fluid balance and IV fluid administered illustrated analogous fluid requirement and response with the transition from net positive to net negative fluid balance between 48 and 72 hours. There was an association between total IV fluid and ventilator requirement (p = 0.003).

CONCLUSIONS

Critically ill pediatric trauma and postoperative patients seem to have similar fluid management and balance after injury or surgery. In our opinion, these two critically ill populations could be combined in large prospective studies on optimal fluid therapy in critically ill children.

Fluid accumulation in critically ill children: a systematic review and meta-analysis

Background

Fluids are often administered for various purposes, such as resuscitation, replacement, maintenance, nutrition, or drug infusion. However, its use is not without risks. Critically ill patients are highly susceptible to fluid accumulation (FA), which is associated with poor outcomes, including organ dysfunction, prolonged mechanical ventilation, extended hospital stays, and increased mortality. This study aimed to assess the association between FA and poor outcomes in critically ill children.

Methods

In this systematic review and meta-analysis, we searched PubMed, Embase, ClinicalTrials.gov, and Cochrane Library databases from inception to May 2024. Relevant publications were searched using the following terms: child, children, infant, infants, pediatric, pediatrics, critically ill children, critical illness, critical care, intensive care, pediatric intensive care, pediatric intensive care unit, fluid balance, fluid overload, fluid accumulation, fluid therapy, edema, respiratory failure, respiratory insufficiency, pulmonary edema, mechanical ventilation, hemodynamic instability, shock, sepsis, acute renal failure, acute kidney failure, acute kidney injury, renal replacement therapy, dialysis, mortality. Paediatric studies were considered eligible if they assessed the effect of FA on the outcomes of interest. The main outcome was all-cause mortality. Pooled analyses were performed by using random-effects models. This review was registered on PROSPERO (CRD42023432879).

Findings

A total of 120 studies (44,682 children) were included. Thirty-five FA definitions were identified. In general, FA was significantly associated with increased mortality (odds ratio [OR] 4.36; 95% confidence interval [CI] 3.53-5.38), acute kidney injury (OR 1.98; 95% CI 1.60-2.44), prolonged mechanical ventilation (weighted mean difference [WMD] 38.1 h, 95% CI 19.35-56.84), and longer stay in the intensive care unit (WMD 2.29 days; 95% CI 1.19-3.38). The percentage of FA was lower in survivors when compared to non-survivors (WMD -4.95 [95% CI, -6.03 to -3.87]). When considering only studies that controlled for potential confounding variables, the pooled analysis revealed 6% increased odds of mortality associated with each 1% increase in the percentage of FA (adjusted OR = 1.06 [95% CI, 1.04-1.09).

Interpretation

FA is significantly associated with poorer outcomes in critically ill children. Thus, clinicians should closely monitor fluid balance, especially when new-onset or worsening organ dysfunction occurs in oedematous patients, indicating potential FA syndrome. Future research should explore interventions like restrictive fluid therapy or de-resuscitation methods. Meanwhile, preventive measures should be prioritized to mitigate FA until further evidence is available.

Funding

None.

Unsupervised machine learning analysis to identify patterns of ICU medication use for fluid overload prediction

INTRODUCTION

Intravenous (IV) medications are a fundamental cause of fluid overload (FO) in the intensive care unit (ICU); however, the association between IV medication use (including volume), administration timing, and FO occurrence remains unclear.

METHODS

This retrospective cohort study included consecutive adults admitted to an ICU ≥72 hours with available fluid balance data. FO was defined as a positive fluid balance ≥7% of admission body weight within 72 hours of ICU admission. After reviewing medication administration record (MAR) data in three-hour periods, IV medication exposure was categorized into clusters using principal component analysis (PCA) and Restricted Boltzmann Machine (RBM). Medication regimens of patients with and without FO were compared within clusters to assess for temporal clusters associated with FO using the Wilcoxon rank sum test. Exploratory analyses of the medication cluster most associated with FO for medications frequently appearing and used in the first 24 hours was conducted.

RESULTS

FO occurred in 127/927 (13.7%) of the patients enrolled. Patients received a median (IQR) of 31 (13-65) discrete IV medication administrations over the 72-hour period. Across all 47,803 IV medication administrations, ten unique IV medication clusters were identified with 121-130 medications in each cluster. Among the ten clusters, cluster 7 had the greatest association with FO; the mean number of cluster 7 medications received was significantly greater in patients in the FO cohort compared to patients who did not experience FO (25.6 vs.10.9. p<0.0001). 51 of the 127 medications in cluster 7 (40.2%) appeared in > 5 separate 3-hour periods during the 72-hour study window. The most common cluster 7 medications included continuous infusions, antibiotics, and sedatives/analgesics. Addition of cluster 7 medications to a prediction model with APACHE II score and receipt of diuretics improved the ability for the model to predict fluid overload (AUROC 5.65, p =0.0004).

CONCLUSIONS

Using ML approaches, a unique IV medication cluster was strongly associated with FO. Incorporation of this cluster improved the ability to predict development of fluid overload in ICU patients compared with traditional prediction models. This method may be further developed into real-time clinical applications to improve early detection of adverse outcomes.

Machine learning vs. traditional regression analysis for fluid overload prediction in the ICU

Fluid overload, while common in the ICU and associated with serious sequelae, is hard to predict and may be influenced by ICU medication use. Machine learning (ML) approaches may offer advantages over traditional regression techniques to predict it. We compared the ability of traditional regression techniques and different ML-based modeling approaches to identify clinically meaningful fluid overload predictors. This was a retrospective, observational cohort study of adult patients admitted to an ICU ≥ 72 h between 10/1/2015 and 10/31/2020 with available fluid balance data. Models to predict fluid overload (a positive fluid balance ≥ 10% of the admission body weight) in the 48-72 h after ICU admission were created. Potential patient and medication fluid overload predictor variables (n = 28) were collected at either baseline or 24 h after ICU admission. The optimal traditional logistic regression model was created using backward selection. Supervised, classification-based ML models were trained and optimized, including a meta-modeling approach. Area under the receiver operating characteristic (AUROC), positive predictive value (PPV), and negative predictive value (NPV) were compared between the traditional and ML fluid prediction models. A total of 49 of the 391 (12.5%) patients developed fluid overload. Among the ML models, the XGBoost model had the highest performance (AUROC 0.78, PPV 0.27, NPV 0.94) for fluid overload prediction. The XGBoost model performed similarly to the final traditional logistic regression model (AUROC 0.70; PPV 0.20, NPV 0.94). Feature importance analysis revealed severity of illness scores and medication-related data were the most important predictors of fluid overload. In the context of our study, ML and traditional models appear to perform similarly to predict fluid overload in the ICU. Baseline severity of illness and ICU medication regimen complexity are important predictors of fluid overload.

A Clinical Score for Quantifying Edema in Mechanically Ventilated Children With Congenital Heart Disease in Intensive Care

Standardized clinical measurements of edema do not exist.

OBJECTIVES

To describe a 19-point clinical edema score (CES), investigate its interobserver agreement, and compare changes between such CES and body weight.

DESIGN SETTING AND PARTICIPANTS

Prospective observational study in a tertiary PICU of mechanically ventilated children with congenital heart disease.

MAIN OUTCOMES AND MEASURES

Differences in the median CES between observer groups.

RESULTS

We studied 61 children, with a median age of 8.0 days (interquartile range, 1.0-14.0 d). A total of 539 CES were performed by three observer groups (medical 1 [reference], medical 2, and bedside nurse) at 0, 24, and 48 hours from enrollment. Overall, there was close agreement between observer groups in mean, median, and upper quartile of CES scores, with least agreement observed in the lower quartile of scores. Across all quartiles of CES, after adjusting for baseline weight, cardiac surgical risk, duration of cardiopulmonary bypass, or peritoneal dialysis during the study, observer groups returned similar mean scores (medical 2: 25th centile +0.1 [95% CI, -0.2 to 0.5], median +0.6 [95% CI, -0.4 to 1.5], 75th centile +0.1 [95% CI, -1.1 to 1.4] and nurse: 25th centile +0.5 [95% CI, 0.0-0.9], median +0.7 [95% CI, 0.0-1.5], 75th centile -0.2 [95% CI, -1.3 to 1.0]) Within a multivariable mixed-effects linear regression model, including adjustment for baseline CES, each 1 point increase in CES was associated with a 12.1 grams (95% CI, 3.2-21 grams) increase in body weight.

CONCLUSIONS AND RELEVANCE

In mechanically ventilated children with congenital heart disease, three groups of observers tended to agree when assessing overall edema using an ordinal clinical score assessed in six body regions, with agreement least at low edema scores. An increase in CES was associated with an increase in body weight, suggesting some validity for quantifying edema. Further exploration of the CES as a rapid clinical tool is indicated.

Fluid balance in critically ill children with lower respiratory tract viral infection: a cohort study

BACKGROUND

Increasing evidence has associated positive fluid balance of critically ill patients with poor outcomes. The aim of this study was to explore the pattern of daily fluid balances and their association with outcomes in critically ill children with lower respiratory tract viral infection.

METHODS

A retrospective single-center study was conducted, in children supported with high-flow nasal cannula, non-invasive ventilation, or invasive ventilation. Median (interquartile range) daily fluid balances, cumulative fluid overload (FO) and peak FO variation, indexed as the % of admission body weight, over the first week of Pediatric Intensive Care Unit admission, and their association with the duration of respiratory support were assessed.

RESULTS

Overall, 94 patients with a median age of 6.9 (1.9-18) months, and a respiratory support duration of 4 (2-7) days, showed a median (interquartile range) daily fluid balance of 18 (4.5-19.5) ml/kg at day 1, which decreased up to day 3 to 5.9 (- 14 to 24.9) ml/kg and increased to 13 (- 11 to 29.9) ml/kg at day 7 (p = 0.001). Median cumulative FO% was 4.6 (- 0.8 to 11) and peak FO% was 5.7 (1.9-12.4). Daily fluid balances, once patients were stratified according to the respiratory support, were significantly lower in those requiring mechanical ventilation (p = 0.003). No correlation was found between all examined fluid balances and respiratory support duration or oxygen saturation, even after subgroup analysis of patients with invasive mechanical ventilation, or respiratory comorbidities, or bacterial coinfection, or of patients under 1 year old.

CONCLUSIONS

In a cohort of children with bronchiolitis, fluid balance was not associated with duration of respiratory support or other parameters of pulmonary function.

Epidemiology, risk factors and outcomes of prolonged mechanical ventilation with different cut-points in a PICU

Background

A consensus on the definition of prolonged mechanical ventilation (PMV) for children does not exist. There is still lack of published work presenting the epidemiology, risk factors and outcomes at different cut-points for PMV patients. These are important for planning the goals of treatment and counseling of the prognosis for patient families. We aimed to determine the incidence, baseline characteristics, risk factors and outcomes of PMV in pediatric patients at various cut-points (>14, >21 or >30days).

Methods

A retrospective cohort study among children <18-years-old who were PMV > 14 days in the PICU of King Chulalongkorn Memorial Hospital was conducted. The primary outcomes were incidence of PMV with various cut-points. We stratified patients into three groups (Group 1; PMV > 14-21, Group 2; >21-30, Group 3; >30 days) for evaluating the baseline characteristics, risk factors, and outcomes of PMV (extubation success, tracheostomy status and death). Factors associated with PMV and deaths were analyzed using univariate and multivariate logistic regression.

Results

From January 2018 to August 2022, 1,050 patients were screened. Of these, 114 patients were enrolled. The incidence of PMV > 14, >21 and >30 days were 10.9%, 7.3% and 5.0% respectively. Extubation success was significantly lower in Group 3 than in Groups 1 & 2 (15.4% vs. 62.2% & 56.0%, P < 0.001). Consequently, the tracheostomy rate (63.5% vs. 16.2% & 12.0%, P < 0.001), VAP rate (98.1% vs. 59.5% & 80.0%, P < 0.001), mortality rate by disease (34.6% vs. 5.4% & 20.0%, P = 0.003), median PICU LOS (50.5 vs. 22.0 & 28.0 days, P < 0.001) and median hospital LOS (124.5 vs. 55.0 & 62.0 days, P < 0.001) were also significantly higher for Group 3 compared with Groups 1 & 2. The factor associated with PMV > 30 days was VAP (aOR: 19.53, 95% CI: 2.38-160.34, P = 0.01). Factors associated with non-surviving patients were 3rd degree PEM (aOR: 5.14, 95% CI: 1.57-16.88, P = 0.01), PIM3 score ≥14 (aOR: 6.75, 95% CI: 2.26-20.15, P < 0.001) and muscle relaxant usage (aOR: 5.58, 95% CI: 1.65-18.86, P = 0.01).

Conclusion

Extubation failure, tracheostomy rate, VAP rate, mortality rate by disease, PICU LOS and hospital LOS were significantly higher for PMV >30 days. Consequently, we suggest that a 30-day duration as a cut-point for PMV in PICUs might be more appropriate.