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.

Purpose of the review This review highlights recent evidence describing the outcomes associated with fluid overload in critically ill patients and provides an overview of fluid management strategies aimed at preventing fluid overload during the resuscitation of patients with shock. Recent findings Fluid overload is a common complication of fluid resuscitation and is associated with increased hospital costs, morbidity and mortality. Summary Fluid management goals differ during the resuscitation, optimization, stabilization and evacuation phases of fluid resuscitation. To prevent fluid overload, strategies that reduce excessive fluid infusions and emphasize the removal of accumulated fluids should be implemented.

Background Coronavirus disease (COVID-19) can cause severe illness and death. Predictors of poor outcome collected on hospital admission may inform clinical and public health decisions. Methods We conducted a retrospective observational cohort investigation of 297 adults admitted to eight academic and community hospitals in Georgia, United States, during March 2020. Using standardized medical record abstraction, we collected data on predictors including admission demographics, underlying medical conditions, outpatient antihypertensive medications, recorded symptoms, vital signs, radiographic findings, and laboratory values. We used random forest models to calculate adjusted odds ratios (aORs) and 95% confidence intervals (CI) for predictors of invasive mechanical ventilation (IMV) and death. Results Compared with age <45 years, ages 65–74 years and ≥75 years were predictors of IMV (aOR 3.12, CI 1.47–6.60; aOR 2.79, CI 1.23–6.33) and the strongest predictors for death (aOR 12.92, CI 3.26–51.25; aOR 18.06, CI 4.43–73.63). Comorbidities associated with death (aORs from 2.4 to 3.8, p <0.05) included end-stage renal disease, coronary artery disease, and neurologic disorders, but not pulmonary disease, immunocompromise, or hypertension. Pre-hospital use vs. non-use of angiotensin receptor blockers (aOR 2.02, CI 1.03–3.96) and dihydropyridine calcium channel blockers (aOR 1.91, CI 1.03–3.55) were associated with death. Conclusions After adjustment for patient and clinical characteristics, older age was the strongest predictor of death, exceeding comorbidities, abnormal vital signs, and laboratory test abnormalities. That coronary artery disease, but not chronic lung disease, was associated with death among hospitalized patients warrants further investigation, as do associations between certain antihypertensive medications and death.

OBJECTIVES: To determine the association between time period of hospitalization and hospital mortality among critically ill adults with coronavirus disease 2019. DESIGN: Observational cohort study from March 6, 2020, to January 31, 2021. SETTING: ICUs at four hospitals within an academic health center network in Atlanta, GA. PATIENTS: Adults greater than or equal to 18 years with coronavirus disease 2019 admitted to an ICU during the study period (i.e., Surge 1: March to April, Lull 1: May to June, Surge 2: July to August, Lull 2: September to November, Surge 3: December to January). MEASUREMENTS AND MAIN RESULTS: Among 1,686 patients with coronavirus disease 2019 admitted to an ICU during the study period, all-cause hospital mortality was 29.7%. Mortality differed significantly over time: 28.7% in Surge 1, 21.3% in Lull 1, 25.2% in Surge 2, 30.2% in Lull 2, 34.7% in Surge 3 (p = 0.007). Mortality was significantly associated with 1) preexisting risk factors (older age, race, ethnicity, lower body mass index, higher Elixhauser Comorbidity Index, admission from a nursing home); 2) clinical status at ICU admission (higher Sequential Organ Failure Assessment score, higher d-dimer, higher C-reactive protein); and 3) ICU interventions (receipt of mechanical ventilation, vasopressors, renal replacement therapy, inhaled vasodilators). After adjusting for baseline and clinical variables, there was a significantly increased risk of mortality associated with admission during Lull 2 (relative risk, 1.37 [95% CI = 1.03–1.81]) and Surge 3 (relative risk, 1.35 [95% CI = 1.04–1.77]) as compared to Surge 1. CONCLUSIONS: Despite increased experience and evidence-based treatments, the risk of death for patients admitted to the ICU with coronavirus disease 2019 was highest during the fall and winter of 2020. Reasons for this increased mortality are not clear.

Objective: Clinical protocols may decrease unnecessary variation in care and improve compliance with desirable therapies. We evaluated whether highly protocolized ICUs have superior patient outcomes compared with less highly protocolized ICUs. Design: Observational study in which participating ICUs completed a general assessment and enrolled new patients 1 day each week. Patients: A total of 6,179 critically ill patients. Setting: Fifty-nine ICUs in the United States Critical Illness and Injury Trials Group Critical Illness Outcomes Study. Interventions: None. Measurements and Main Results: The primary exposure was the number of ICU protocols; the primary outcome was hospital mortality. A total of 5,809 participants were followed prospectively, and 5,454 patients in 57 ICUs had complete outcome data. The median number of protocols per ICU was 19 (interquartile range, 15-21.5). In single-variable analyses, there were no differences in ICU and hospital mortality, length of stay, use of mechanical ventilation, vasopressors, or continuous sedation among individuals in ICUs with a high versus low number of protocols. The lack of association was confirmed in adjusted multivariable analysis (p = 0.70). Protocol compliance with two ventilator management protocols was moderate and did not differ between ICUs with high versus low numbers of protocols for lung protective ventilation in acute respiratory distress syndrome (47% vs 52%; p = 0.28) and for spontaneous breathing trials (55% vs 51%; p = 0.27). Conclusions: Clinical protocols are highly prevalent in U.S. ICUs. The presence of a greater number of protocols was not associated with protocol compliance or patient mortality.

Objectives: Sepsis-3 defines organ dysfunction as an increase in the Sequential Organ Failure Assessment score by greater than or equal to 2 points. However, some Sequential Organ Failure Assessment score components are not routinely recorded in all hospitals' electronic health record systems, limiting its utility for wide-scale sepsis surveillance. The Centers for Disease Control and Prevention recently released the Adult Sepsis Event surveillance definition that includes simplified organ dysfunction criteria optimized for electronic health records (eSOFA). We compared eSOFA versus Sequential Organ Failure Assessment with regard to sepsis prevalence, overlap, and outcomes. Design: Retrospective cohort study. Setting: One hundred eleven U.S. hospitals in the Cerner HealthFacts dataset. Patients: Adults hospitalized in 2013-2015. Interventions: None. Measurements and Main Results: We identified clinical indicators of presumed infection (blood cultures and antibiotics) concurrent with either: 1) an increase in Sequential Organ Failure Assessment score by 2 or more points (Sepsis-3) or 2) 1 or more eSOFA criteria: vasopressor initiation, mechanical ventilation initiation, lactate greater than or equal to 2.0 mmol/L, doubling in creatinine, doubling in bilirubin to greater than or equal to 2.0 mg/dL, or greater than or equal to 50% decrease in platelet count to less than 100 cells/μL (Centers for Disease Control and Prevention Adult Sepsis Event). We compared area under the receiver operating characteristic curves for discriminating in-hospital mortality, adjusting for baseline characteristics. Of 942,360 patients in the cohort, 57,242 (6.1%) had sepsis by Sequential Organ Failure Assessment versus 41,618 (4.4%) by eSOFA. Agreement between sepsis by Sequential Organ Failure Assessment and eSOFA was good (Cronbach's alpha 0.81). Baseline characteristics and infectious diagnoses were similar, but mortality was higher with eSOFA (17.1%) versus Sequential Organ Failure Assessment (14.4%; p < 0.001) as was discrimination for mortality (area under the receiver operating characteristic curve, 0.774 vs 0.759; p < 0.001). Comparisons were consistent across subgroups of age, infectious diagnoses, and comorbidities. Conclusions: The Adult Sepsis Event's eSOFA organ dysfunction criteria identify a smaller, more severely ill sepsis cohort compared with the Sequential Organ Failure Assessment score, but with good overlap and similar clinical characteristics. Adult Sepsis Events may facilitate wide-scale automated sepsis surveillance that tracks closely with the more complex Sepsis-3 criteria.

Background: Identifying patterns within ICU medication regimens may help artificial intelligence algorithms to better predict patient outcomes; however, machine learning methods incorporating medications require further development, including standardized terminology. The Common Data Model for Intensive Care Unit (ICU) Medications (CDM-ICURx) may provide important infrastructure to clinicians and researchers to support artificial intelligence analysis of medication-related outcomes and healthcare costs. Using an unsupervised cluster analysis approach in combination with this common data model, the objective of this evaluation was to identify novel patterns of medication clusters (termed ‘pharmacophenotypes’) correlated with ICU adverse events (e.g., fluid overload) and patient-centered outcomes (e.g., mortality). Methods: This was a retrospective, observational cohort study of 991 critically ill adults. To identify pharmacophenotypes, unsupervised machine learning analysis with automated feature learning using restricted Boltzmann machine and hierarchical clustering was performed on the medication administration records of each patient during the first 24 h of their ICU stay. Hierarchical agglomerative clustering was applied to identify unique patient clusters. Distributions of medications across pharmacophenotypes were described, and differences among patient clusters were compared using signed rank tests and Fisher's exact tests, as appropriate. Results: A total of 30,550 medication orders for the 991 patients were analyzed; five unique patient clusters and six unique pharmacophenotypes were identified. For patient outcomes, compared to patients in Clusters 1 and 3, patients in Cluster 5 had a significantly shorter duration of mechanical ventilation and ICU length of stay (p < 0.05); for medications, Cluster 5 had a higher distribution of Pharmacophenotype 1 and a smaller distribution of Pharmacophenotype 2, compared to Clusters 1 and 3. For outcomes, patients in Cluster 2, despite having the highest severity of illness and greatest medication regimen complexity, had the lowest overall mortality; for medications, Cluster 2 also had a comparably higher distribution of Pharmacophenotype 6. Conclusion: The results of this evaluation suggest that patterns among patient clusters and medication regimens may be observed using empiric methods of unsupervised machine learning in combination with a common data model. These results have potential because while phenotyping approaches have been used to classify heterogenous syndromes in critical illness to better define treatment response, the entire medication administration record has not been incorporated in those analyses. Applying knowledge of these patterns at the bedside requires further algorithm development and clinical application but may have the future potential to be leveraged in guiding medication-related decision making to improve treatment outcomes.

While medication regimen complexity, as measured by a novel medication regimen complexity-intensive care unit (MRC-ICU) score, correlates with baseline severity of illness and mortality, whether the MRC-ICU improves hospital mortality prediction is not known. After characterizing the association between MRC-ICU, severity of illness and hospital mortality we sought to evaluate the incremental benefit of adding MRC-ICU to illness severity-based hospital mortality prediction models. This was a single-center, observational cohort study of adult intensive care units (ICUs). A random sample of 991 adults admitted ≥ 24 h to the ICU from 10/2015 to 10/2020 were included. The logistic regression models for the primary outcome of mortality were assessed via area under the receiver operating characteristic (AUROC). Medication regimen complexity was evaluated daily using the MRC-ICU. This previously validated index is a weighted summation of medications prescribed in the first 24 h of ICU stay [e.g., a patient prescribed insulin (1 point) and vancomycin (3 points) has a MRC-ICU = 4 points]. Baseline demographic features (e.g., age, sex, ICU type) were collected and severity of illness (based on worst values within the first 24 h of ICU admission) was characterized using both the Acute Physiology and Chronic Health Evaluation (APACHE II) and the Sequential Organ Failure Assessment (SOFA) score. Univariate analysis of 991 patients revealed every one-point increase in the average 24-h MRC-ICU score was associated with a 5% increase in hospital mortality [Odds Ratio (OR) 1.05, 95% confidence interval 1.02–1.08, p = 0.002]. The model including MRC-ICU, APACHE II and SOFA had a AUROC for mortality of 0.81 whereas the model including only APACHE-II and SOFA had a AUROC for mortality of 0.76. Medication regimen complexity is associated with increased hospital mortality. A prediction model including medication regimen complexity only modestly improves hospital mortality prediction.

OBJECTIVES: Increasing time to mechanical ventilation and high-flow nasal cannula use may be associated with mortality in coronavirus disease 2019. We examined the impact of time to intubation and use of high-flow nasal cannula on clinical outcomes in patients with coronavirus disease 2019. DESIGN: Retrospective cohort study. SETTING: Six coronavirus disease 2019-specific ICUs across four university-affiliated hospitals in Atlanta, Georgia. PATIENTS: Adults with laboratory-confirmed severe acute respiratory syndrome coronavirus 2 infection who received high-flow nasal cannula or mechanical ventilation. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: Among 231 patients admitted to the ICU, 109 (47.2%) were treated with high-flow nasal cannula and 97 (42.0%) were intubated without preceding high-flow nasal cannula use. Of those managed with high-flow nasal cannula, 78 (71.6%) ultimately received mechanical ventilation. In total, 175 patients received mechanical ventilation; 44.6% were female, 66.3% were Black, and the median age was 66 years (interquartile range, 56-75 yr). Seventy-six patients (43.4%) were intubated within 8 hours of ICU admission, 57 (32.6%) between 8 and 24 hours of admission, and 42 (24.0%) greater than or equal to 24 hours after admission. Patients intubated within 8 hours were more likely to have diabetes, chronic comorbidities, and higher admission Sequential Organ Failure Assessment scores. Mortality did not differ by time to intubation (≤ 8 hr: 38.2%; 8-24 hr: 31.6%; ≥ 24 hr: 38.1%; p = 0.7), and there was no association between time to intubation and mortality in adjusted analysis. Similarly, there was no difference in initial static compliance, duration of mechanical ventilation, or ICU length of stay by timing of intubation. High-flow nasal cannula use prior to intubation was not associated with mortality. CONCLUSIONS: In this cohort of critically ill patients with coronavirus disease 2019, neither time from ICU admission to intubation nor high-flow nasal cannula use were associated with increased mortality. This study provides evidence that coronavirus disease 2019 respiratory failure can be managed similarly to hypoxic respiratory failure of other etiologies.

Objectives: To identify circumstances in which repeated measures of organ failure would improve mortality prediction in ICU patients. Design: Retrospective cohort study, with external validation in a deidentified ICU database. Setting: Eleven ICUs in three university hospitals within an academic healthcare system in 2014. Patients: Adults (18 yr old or older) who satisfied the following criteria: 1) two of four systemic inflammatory response syndrome criteria plus an ordered blood culture, all within 24 hours of hospital admission; and 2) ICU admission for at least 2 calendar days, within 72 hours of emergency department presentation. Intervention: None Measurements and Main Results: Data were collected until death, ICU discharge, or the seventh ICU day, whichever came first. The highest Sequential Organ Failure Assessment score from the ICU admission day (ICU day 1) was included in a multivariable model controlling for other covariates. The worst Sequential Organ Failure Assessment scores from the first 7 days after ICU admission were incrementally added and retained if they obtained statistical significance (p < 0.05). The cohort was divided into seven subcohorts to facilitate statistical comparison using the integrated discriminatory index. Of the 1,290 derivation cohort patients, 83 patients (6.4%) died in the ICU, compared with 949 of the 8,441 patients (11.2%) in the validation cohort. Incremental addition of Sequential Organ Failure Assessment data up to ICU day 5 improved the integrated discriminatory index in the validation cohort. Adding ICU day 6 or 7 Sequential Organ Failure Assessment data did not further improve model performance. Conclusions: Serial organ failure data improve prediction of ICU mortality, but a point exists after which further data no longer improve ICU mortality prediction of early sepsis.