Fluid Overload Phenotypes in Critical Illness—A Machine Learning Approach

Background: The detrimental impact of fluid overload (FO) on intensive care unit (ICU) morbidity and mortality is well known. However, research to identify subgroups of patients particularly prone to fluid overload is scarce. The aim of this cohort study was to derive “FO phenotypes” in the critically ill by using machine learning techniques. Methods: Retrospective single center study including adult intensive care patients with a length of stay of ≥3 days and sufficient data to compute FO. Data was analyzed by multivariable logistic regression, fast and frugal trees (FFT), classification decision trees (DT), and a random forest (RF) model. Results: Out of 1772 included patients, 387 (21.8%) met the FO definition. The random forest model had the highest area under the curve (AUC) (0.84, 95% CI 0.79–0.86), followed by multivariable logistic regression (0.81, 95% CI 0.77–0.86), FFT (0.75, 95% CI 0.69–0.79) and DT (0.73, 95% CI 0.68–0.78) to predict FO. The most important predictors identified in all models were lactate and bicarbonate at admission and postsurgical ICU admission. Sepsis/septic shock was identified as a risk factor in the MV and RF analysis. Conclusion: The FO phenotypes consist of patients admitted after surgery or with sepsis/septic shock with high lactate and low bicarbonate.

Rhabdomyolysis: An evidence-based approach

A 76-year-old lady was found on the floor following a fall at home. She was uninjured, but unable to get up, and had been lying on the floor for roughly 18 hours before her son arrived. She had been unwell for the past 3 days with a cough and shortness of breath. She had a past medical history of diabetes, hypertension, hypercholesterolaemia and atrial fibrillation (AF). On examination, she was alert but distressed, clinically dehydrated, febrile and tachycardic. She was treated for community acquired pneumonia with co-amoxiclav and was fluid resuscitated with Hartmann’s solution. Her hyperkalaemia was treated with 50 mL of 50% glucose containing 10 units of rapid-acting insulin. Her creatinine kinase (CK) on admission was 200,000, and she had an acute kidney injury (AKI). Urine dipstick was positive for blood. However, her renal function continued to deteriorate over the succeeding 48 h, when she required renal replacement therapy (RRT) due to fluid overload and anuria.

Bio-electric impedance as a tool to assess hydration in critically ill patients: an integrative review

Introduction: Assessing the hydration status of critically ill patients has been a difficult task over the decades. Determining how much fluid overload a patient has often helped in choosing a therapy. Methods such as bioelectrical impedance have been approached as a useful tool for this purpose. Objective: This study proposes to verify, through research in the literature, what is the real importance of the clinical use of bioelectrical impedance in the diagnosis of fluid overload in critically ill patients hospitalized in intensive care units. Methods: bibliographic search in the main scientific information databases: Scielo, PubMed, Cochrane, and Lilacs from January 2000 to July 2018. The selected languages were Spanish, Portuguese, and English. The keywords used were bioelectrical impedance, hydration, Intensive Care Unit, Intensive Care, bioelectrical impedance analysis, fluid balance, hydration overload. Results and Conclusion: The analysis of fluid overload in critically ill patients can be performed using multifrequency bioelectrical impedance. It is a useful tool in the diagnosis as well as in the quantification of water overload and, therefore, a corroborative method for clinical decision-making.