Background: This in-vitro and in-vivo study investigates the Gibbs-Donnan effect across the filter during continuous veno-venous hemofiltration (CVVH). In particular, we assessed its acid–base implications, applying the physical–chemical approach. Methods: A prospective, single-center study was conducted using the PrismaMax machine (Baxter). Two sets of in-vitro CVVH experiments (with and without albumin) were performed to quantify the Gibbs-Donnan effect. Electrolytes, glucose, and osmolarity changes were measured across the filter and in the ultrafiltrate. Strong ion difference and sieving coefficients of the main solutes were calculated. Similar measurements were performed in oligo-anuric critically ill patients undergoing CVVH. Results: In-vitro experiments without albumin showed a sieving coefficient of 1 for both positive and negative ions. On the contrary, when albumin was added, the sieving coefficient for sodium and chloride changed linearly with albumin concentration ( r = −0.94, p < 0.001 for sodium, r = 0.88, p < 0.001 for chloride), resulting in a progressive linear increase in post-filter strong ion difference ( β = 1.1, r = 0.77, p = 0.003). In 22 studied patients, calculated albumin concentration increased across the filter (2.2 ± 0.5 g/dL vs. 3.1 ± 0.7 g/dL), leading to sodium retention (138 ± 6 vs. 141 ± 6 mmol/L, p < 0.001) and chloride excretion (100 ± 5 vs . 97 ± 5 mmol/L, p < 0.001), thus resulting in a higher post-filter strong ion difference (46 ± 4 vs. 40 ± 4 mmol/L, p < 0.001). Conclusions: These in-vitro and in-vivo studies demonstrate that albumin linearly affects the sieving coefficient of ions, increasing the strong ion difference of plasma water during its passage through the filter and thus having a systemic alkalizing effect. Graphic

Background/Objectives: Effective pain management in polytrauma patients with rib fractures is essential, particularly in the critical care setting. While epidural analgesia is considered the gold standard, it is not always feasible, necessitating alternative locoregional approaches. We present the case of a polytrauma patient with multiple, bilateral rib fractures and severe chest pain that hindered weaning from mechanical ventilation. A bilateral Serratus Anterior Plane Block (SAPB) was performed, with catheters placed for continuous administration of local anesthetics. Pain relief was immediate, enabling a rapid weaning from mechanical ventilation, safe extubation, and subsequent discharge to rehabilitation. A review of the literature on this technique in critically ill patients with thoracic trauma and multiple rib fractures is also presented. Methods: We conducted a literature search up to November 2024, identifying studies evaluating the use of SAPB in critically ill patients with chest trauma and rib fractures. Results: Eight studies were identified, including a total of 197 cases, of which only 3 involved a bilateral SAPB. Studies and published case reports demonstrated significant variability in analgesic protocols and reported outcomes. Notably, only two papers addressed specifically its role in facilitating weaning from mechanical ventilation. Conclusions: Pain control is fundamental in managing severe chest trauma. This case and the reviewed literature suggest that the SAPB is a promising option when epidural analgesia is contraindicated or impractical. However, further studies are needed to define its place in clinical practice and optimize its use in critically ill patients.

Background: Linezolid standard dosing is typically applied in ICU without adjustments, even in renal impairment. This study examines serum concentration variability by renal function or CRRT administration in patients receiving 1200 mg/day of linezolid. Methods: This retrospective, single-center, non-randomized observational study was conducted at Niguarda Hospital (Milan, Italy) on data from the two-year period 2023–2024. ICU patients receiving linezolid, with a renal function determination and trough TDM performed at steady-state were included and stratified by renal function or CRRT status. Results: 54 patients were included, with 18 (33%) undergoing CRRT (CVVH). CRRT patients presented higher median linezolid concentrations (4.6 mg/L) than non-CRRT patients (3.2 mg/L), and a lower risk of underdosing (17% vs. 39%). CRRT patients showed significantly lower concentrations (4.6 mg/L vs. 10 mg/L, p = 0.007) than non-CRRT patients with renal function ≤ 30 mL/min, with fewer out-of-range levels (39% vs. 91%, p = 0.008) and overdosing (22% vs. 73%, p = 0.018). A significant inverse correlation was found between renal function and linezolid levels (Spearman’s rho = −0.61, p < 0.001), with CRRT patients exhibiting concentrations comparable to those of individuals with moderately impaired renal function. Continuous infusion resulted in significantly higher median concentrations (7.2 mg/L) than extended infusion (2.7 mg/L), with an increased risk of overdosing (47% vs. 17%; p = 0.018). Conclusions: After standard-dosing administration, linezolid levels vary widely in critically ill patients. Renal function significantly affects pharmacokinetics: severe impairment increases overdose risk, while ARC may cause underdosing. Standard-dosing appears adequate in CRRT patients, with levels similar to moderate-impairment. Continuous infusion aids target attainment in normal or ARC patients but raises overdose risk in severe impairment. TDM-based personalized dosing seems crucial to optimize therapy and reduce toxicity or failure.

Continuous venovenous hemofiltration (CVVH) is frequently performed in critically ill patients using diluted citrate for regional anticoagulation. The impact of this renal replacement strategy on plasma sodium has not been evaluated yet. Our aim was therefore to assess the period prevalence of hyponatremia (sodium <135 mmol/L) during CVVH and discuss possible underlying mechanisms. After 48 hours of treatment, 70% of the 27 oligo-anuric critically ill patients were hyponatremic, despite the use of dialysis fluid bags (Regiocit 18/0, Phoxilium by Baxter, Deerfield, IL, and Multibic K2 by Fresenius Medical Care AG & Co. KGaA, Bad Homburg, Germany) with sodium content of 140 mmol/L. Indeed, sodium decreased from 142 ± 7 to 135 ± 3 mmol/L, p < 0.001. Sodium concentrations of employed dialysis bags were confirmed using ion chromatography. However, ionized sodium of Regiocit measured with a direct-ion selective electrode (ISE) resulted lower (~118 mmol/L), suggesting the presence of sodium-to-citrate complexes. Possible mechanisms explaining the hyponatremia development could therefore include: i) plasma water dilution; ii) a reduced Gibbs-Donnan effect, given the low albumin concentration (2.6 ± 0.8 g/dl) of our critically ill patients; iii) a negative sodium balance due to the loss of sodium-to-citrate complexes across the filter. The clinical implications of the described hyponatremia and the different contributions of the hypothesized mechanisms need to be addressed in future studies.

Accurate prognostication in the intensive care unit (ICU) is essential for delivering personalized and ethically sound care, yet it remains a challenge for high-risk patients. This study aimed to develop and validate an exploratory machine learning model for predicting ICU mortality in critically ill patients requiring both mechanical ventilation and vasoactive support. We conducted a retrospective analysis of 4816 patient records from the AmsterdamUMCdb database. The XGBoost model was developed using 66 clinical and laboratory features collected within the first 24 h of ICU admission. The model demonstrated robust discriminatory performance in predicting in-ICU mortality at 24h from admission, achieving in internal validation an area under the receiver operating characteristic curve (AUROC) of 0.831 (95% CI [0.799–0.860]), a macro F1 score of 0.735, and a Brier score of 0.130, indicating reliable calibration. Explainable AI methods identified the minimum Glasgow Coma Scale score, patient age, and minimum platelet count as the most significant predictors. In this exploratory study, a machine learning model using single-center early ICU data showed promising performance for mortality prediction in a high-risk population. Given the lack of external or prospective validation, the model should be considered investigational and requires further validation before clinical application.