Mario Menk

The COVID-19 pandemic has posed a major challenge to healthcare systems globally. Millions of people have been infected, and millions of deaths have been reported worldwide. Glucocorticoids have attracted worldwide attention for their potential efficacy in the treatment of COVID-19. Various glucocorticoids with different dosages and treatment durations have been studied in patients with different severities, with a suitable dosage and treatment duration not yet defined. This study aimed to investigate whether in-hospital survival differs between critically ill patients treated with low-dose glucocorticoids, high-dose glucocorticoids or no glucocorticoids. All critically ill patients admitted to the intensive care unit of the Charité Hospital—Universitätsmedizin Berlin between February 2020 and December 2021 with COVID-19 pneumonia receiving supplemental oxygen were eligible to participate in this multicenter real-world data study. Patients were retrospectively assigned to one of three groups: the high corticosteroid dose (HighC) group (receiving 6 mg parenteral dexamethasone or an equivalent corticosteroid dosage for ten days), the low corticosteroid dose (LowC) group (receiving less than 6 mg parenteral dexamethasone or an equivalent corticosteroid dosage for ten days), or the no corticosteroid (NoC) group. Overall survival and risk effects were compared among groups within the total observation period, as well as at 35 days after the onset of COVID-19 symptoms. Adjusted multivariable Cox proportional hazard regression analysis was performed to compare the risk of death between the treatment groups. Out of 1561 critically ill COVID-19 patients, 1014 were included in the baseline analysis. In the survival study, 1009 patients were assigned to the NoC (n = 346), HighC (n = 552), or LowC group (n = 111). The baseline characteristics were balanced between groups, except for age, BMI, APACHE II score, SOFA and SAPS II. While the 35-day survival did not show any differences, a landmark analysis of the patients surviving beyond 35 days revealed differences between groups. The restricted mean survival time was 112 days in the LowC group [95% CI: 97 – 128], 133 days in the HighC group [95% CI: 124 – 141] and 144 days in the NoC group [95% CI: 121 – 167]. The multivariable-adjusted Cox proportional hazard analysis indicated that, regardless of age, sex, health status or invasive oxygenation, a low-dose treatment increased the hazard of death of critically ill COVID-19 patients by a factor of 2.09 ([95% CI: 0.99, 4.4], p = 0.05) and a high-dose corticosteroid treatment increased the risk by a factor of 1.07 ([95% CI: 0.53, 2.15], p = 0.85) compared to no treatment with glucocorticoids. The analysis reveals that corticosteroid treatment does not influence the survival of critically ill COVID-19 patients in the intensive care unit within 35 days. Our evaluations further suggest that regardless of ventilation status, the decision-making process for administering corticosteroid therapy should account for the individual severity of the illness.

Background: Anticoagulation therapy with heparin is a frequent treatment in intensive care units and is monitored by activated partial thromboplastin clotting time (aPTT). It has been demonstrated that reaching an established anticoagulation target within 24 hours is associated with favorable outcomes. However, patients respond to heparin differently and reaching the anticoagulation target can be challenging. Machine learning algorithms may potentially support clinicians with improved dosing recommendations. Objective: This study evaluates a range of machine learning algorithms on their capability of predicting the patients’ response to heparin treatment. In this analysis, we apply, for the first time, a model that considers time series. Methods: We extracted patient demographics, laboratory values, dialysis and extracorporeal membrane oxygenation treatments, and scores from the hospital information system. We predicted the numerical values of aPTT laboratory values 24 hours after continuous heparin infusion and evaluated 7 different machine learning models. The best-performing model was compared to recently published models on a classification task. We considered all data before and within the first 12 hours of continuous heparin infusion as features and predicted the aPTT value after 24 hours. Results: The distribution of aPTT in our cohort of 5926 hospital admissions was highly skewed. Most patients showed aPTT values below 75 s, while some outliers showed much higher aPTT values. A recurrent neural network that consumes a time series of features showed the highest performance on the test set. Conclusions: A recurrent neural network that uses time series of features instead of only static and aggregated features showed the highest performance in predicting aPTT after heparin treatment.

(1) Background: Acute respiratory distress syndrome (ARDS) is a rare complication in multiply injured patients. Due to the rarity of ARDS development after trauma, little is known about outcomes of patients with trauma-associated ARDS compared to patients with non-trauma-associated ARDS. (2) Methods: This retrospective analysis included n = 1038 ARDS patients admitted to the ARDS center of Charité—Universitätsmedizin Berlin between 2007 and 2018. Patients with trauma-associated ARDS (n = 62) were compared to patients with non-trauma-associated ARDS (n = 976). In a secondary analysis, patients from the group with non-trauma-associated ARDS were 1:1 nearest neighbor matched to patients with trauma-associated ARDS. The primary outcomes were 28-day in-hospital mortality, 60-day in-hospital mortality, and overall in-hospital mortality. (3) Results: Overall in-hospital mortality in trauma-associated ARDS was 29.0% compared to 40.5% in all patients with non-trauma-associated ARDS (p = 0.074). The in-hospital mortality rate in matched patients with non-trauma-associated ARDS (33.9%) was comparable to the trauma-associated ARDS cohort (p = 0.701). Kaplan–Meier curves indicated time-sensitive variations in 28-day and 60-day in-hospital survival. (4) Conclusion: Mortality was not different in patients with trauma-associated ARDS compared to patients with non-trauma-associated ARDS. Survival rate in the Kaplan–Meier curves stabilized after the critical initial phase and throughout the further 60-day period in patients with trauma-associated ARDS compared to patients with non-trauma-associated ARDS. Since this divergence was less pronounced in the matched cohort, it may be related to the younger age, fewer comorbidities, and lower ARDS severity in patients with trauma-associated ARDS. Patients with trauma-associated ARDS remain a very different cohort compared to patients with non-trauma-associated ARDS. Therefore, the outcome comparison is limited, even after matching.

BACKGROUND Anticoagulation therapy with Heparin is a frequent treatment in intensive care units, which is monitored by activated partial thromboplastin clotting time (aPTT). It has been demonstrated that reaching an established anticoagulation target within 24 hours is associated with favorable outcomes. However, patients respond to heparin differently and reaching the anticoagulation target can be challenging. Machine learning algorithms may potentially support clinicians with improved dosing recommendations. OBJECTIVE This study evaluates a range of machine learning algorithms on their capability of predicting the patients’ response to heparin treatment. In the present analysis, we apply, for the first time, a model that considers time series. METHODS We extracted patient demographics, laboratory values, dialysis and ecmo treatments, and scores from the hospital information system. We set up a regression task for predicting the aPTT laboratory values 24 hours after continuous heparin infusion and evaluate seven different machine learning models. We consider all data before and within the first 12 hours of continuous heparin infusion as features and predict the aPTT value after 24 hours. RESULTS The distribution of aPTT in our cohort of 5926 hospital admissions is highly skewed. While most patients show aPTT values below 75 s, some outliers show much higher aPTT values. A recurrent neural network that consumes time series of features shows the highest performance on the test set. CONCLUSIONS A recurrent neural network that uses time series of features instead of only static and aggregated features, shows the highest performance of predicting aPTT after heparin treatment.

Background: Increased plasma concentrations of circulating cell-free hemoglobin (CFH) are supposed to contribute to the multifactorial etiology of acute kidney injury (AKI) in critically ill patients while the CFH-scavenger haptoglobin might play a protective role. We evaluated the association of CFH and haptoglobin with AKI in patients with an acute respiratory distress syndrome (ARDS) requiring therapy with VV ECMO. Methods: Patients with CFH and haptoglobin measurements before initiation of ECMO therapy were identified from a cohort of 1044 ARDS patients and grouped into three CFH concentration groups using a risk stratification. The primary objective was to assess the association of CFH and haptoglobin with KDIGO stage 3 AKI. Further objectives included the identification of a target haptoglobin concentration to protect from CFH-associated AKI. Measurements and main results: Two hundred seventy-three patients fulfilled the inclusion criteria. Of those, 154 patients (56.4%) had AKI at ECMO initiation. The incidence of AKI increased stepwise with increasing concentrations of CFH reaching a plateau at 15 mg/dl. Compared to patients with low [< 5 mg/dl] CFH concentrations, patients with moderate [5–14 mg/dl] and high [≥ 15 mg/dl] CFH concentrations had a three- and five-fold increased risk for AKI (adjusted odds ratio [OR] moderate vs. low, 2.69 [95% CI, 1.25–5.95], P = 0.012; and OR high vs. low, 5.47 [2.00–15.9], P = 0.001). Among patients with increased CFH concentrations, haptoglobin plasma levels were lower in patients with AKI compared to patients without AKI. A haptoglobin concentration greater than 2.7 g/l in the moderate and 2.4 g/l in the high CFH group was identified as clinical cutoff value to protect from CFH-associated AKI (sensitivity 89.5% [95% CI, 83–96] and 90.2% [80–97], respectively). Conclusions: In critically ill patients with ARDS requiring therapy with VV ECMO, an increased plasma concentration of CFH was identified as independent risk factor for AKI. Among patients with increased CFH concentrations, higher plasma haptoglobin concentrations might protect from CFH-associated AKI and should be subject of future research.

Packed red blood cells (PRBCs), stored for prolonged intervals, might contribute to adverse clinical outcomes in critically ill patients. In this study, short-term outcome after transfusion of PRBCs of two storage duration periods was analyzed in patients with Acute Respiratory Distress Syndrome (ARDS). Patients who received transfusions of PRBCs were identified from a cohort of 1044 ARDS patients. Patients were grouped according to the mean storage age of all transfused units. Patients transfused with PRBCs of a mean storage age ≤ 28 days were compared to patients transfused with PRBCs of a mean storage age > 28 days. The primary endpoint was 28-day mortality. Secondary endpoints included failure-free days composites. Two hundred and eighty-three patients were eligible for analysis. Patients in the short-term storage group had similar baseline characteristics and received a similar amount of PRBC units compared with patients in the long-term storage group (five units (IQR, 3–10) vs. four units (2–8), p = 0.14). The mean storage age in the short-term storage group was 20 (±5.4) days compared with 32 (±3.1) days in the long-term storage group (mean difference 12 days (95%-CI, 11–13)). There was no difference in 28-day mortality between the short-term storage group compared with the long-term storage group (hazard ratio, 1.36 (95%-CI, 0.84–2.21), p = 0.21). While there were no differences in ventilator-free, sedation-free, and vasopressor-free days composites, patients in the long-term storage group compared with patients in the short-term storage group had a 75% lower chance for successful weaning from renal replacement therapy (RRT) within 28 days after ARDS onset (subdistribution hazard ratio, 0.24 (95%-CI, 0.1–0.55), p < 0.001). Further analysis indicated that even a single PRBC unit stored for more than 28 days decreased the chance for successful weaning from RRT. Prolonged storage of PRBCs was not associated with a higher mortality in adults with ARDS. However, transfusion of long-term stored PRBCs was associated with prolonged dependence of RRT in critically ill patients with an ARDS.

Background: Etomidate is typically used as an induction agent in cardiac surgery because it has little impact on hemodynamics. It is a known suppressor of adrenocortical function and may increase the risk for post-operative infections, sepsis, and mortality. The aim of this study was to evaluate whether etomidate increases the risk of postoperative sepsis (primary outcome) and infections (secondary outcome) compared to propofol. Methods: This was a retrospective before–after trial (IRB EA1/143/20) performed at a tertiary medical center in Berlin, Germany, between 10/2012 and 01/2015. Patients undergoing cardiac surgery were investigated within two observation intervals, during which etomidate and propofol were the sole induction agents. Results: One-thousand, four-hundred, and sixty-two patients, and 622 matched pairs, after caliper propensity-score matching, were included in the final analysis. Sepsis rates did not differ in the matched cohort (etomidate: 11.5% vs. propofol: 8.2%, p = 0.052). Patients in the etomidate interval were more likely to develop hospital-acquired pneumonia (etomidate: 18.6% vs. propofol: 14.0%, p = 0.031). Conclusion: Our study showed that a single-dose of etomidate is not statistically associated with higher postoperative sepsis rates after cardiac surgery, but is associated with a higher incidence of hospital-acquired pneumonia. However, there is a notable trend towards a higher sepsis rate.

Background: Efficacy and safety of different hemoglobin thresholds for transfusion of red blood cells (RBCs) in adults with an acute respiratory distress syndrome (ARDS) are unknown. We therefore assessed the effect of two transfusion thresholds on short-term outcome in patients with ARDS. Methods: Patients who received transfusions of RBCs were identified from a cohort of 1044 ARDS patients. After propensity score matching, patients transfused at a hemoglobin concentration of 8 g/dl or less (lower-threshold) were compared to patients transfused at a hemoglobin concentration of 10 g/dl or less (higher-threshold). The primary endpoint was 28-day mortality. Secondary endpoints included ECMO-free, ventilator-free, sedation-free, and organ dysfunction-free composites. Measurements and main results: One hundred ninety-two patients were eligible for analysis of the matched cohort. Patients in the lower-threshold group had similar baseline characteristics and hemoglobin levels at ARDS onset but received fewer RBC units and had lower hemoglobin levels compared with the higher-threshold group during the course on the ICU (9.1 [IQR, 8.7–9.7] vs. 10.4 [10–11] g/dl, P < 0.001). There was no difference in 28-day mortality between the lower-threshold group compared with the higher-threshold group (hazard ratio, 0.94 [95%-CI, 0.59–1.48], P = 0.78). Within 28 days, 36.5% (95%-CI, 27.0–46.9) of the patients in the lower-threshold group compared with 39.5% (29.9–50.1) of the patients in the higher-threshold group had died. While there were no differences in ECMO-free, sedation-free, and organ dysfunction-free composites, the chance for successful weaning from mechanical ventilation within 28 days after ARDS onset was lower in the lower-threshold group (subdistribution hazard ratio, 0.36 [95%-CI, 0.15–0.86], P = 0.02). Conclusions: Transfusion at a hemoglobin concentration of 8 g/dl, as compared with a hemoglobin concentration of 10 g/dl, was not associated with an increase in 28-day mortality in adults with ARDS. However, a transfusion at a hemoglobin concentration of 8 g/dl was associated with a lower chance for successful weaning from the ventilator during the first 28 days after ARDS onset. Trial Registration: ClinicalTrials.gov NCT03871166.

Objectives: Intracranial hemorrhage is a serious complication in patients receiving venovenous extracorporeal membrane oxygenation during treatment of the acute respiratory distress syndrome. We analyzed timing, outcome, and risk factors of intracranial hemorrhage in patients on venovenous extracorporeal membrane oxygenation. Design: Retrospective cohort study. Setting: Single acute respiratory distress syndrome referral center. Patients: Patients receiving venovenous extracorporeal membrane oxygenation were identified from a cohort of 1,044 patients with acute respiratory distress syndrome. Patients developing an intracranial hemorrhage during venovenous extracorporeal membrane oxygenation therapy were compared with patients without evidence for intracranial hemorrhage. The primary objective was to assess the association of intracranial hemorrhage with 60-day mortality. Further objectives included the identification of risk factors for intracranial hemorrhage and the evaluation of clinical cutoff values. Interventions: None. Measurements and Main Results: Among 444 patients treated with venovenous extracorporeal membrane oxygenation, 49 patients (11.0% [95% CI, 8.3–14.4%]) developed an intracranial hemorrhage. The median time to intracranial hemorrhage occurrence was 4 days (95% CI, 2–7 d). Patients who developed an intracranial hemorrhage had a higher 60-day mortality compared with patients without intracranial hemorrhage (69.4% [54.4–81.3%] vs 44.6% [39.6–49.6%]; odds ratio 3.05 [95% CI, 1.54–6.32%]; p = 0.001). A low platelet count, a high positive end expiratory pressure, and a major initial decrease of Paco ₂ were identified as independent risk factors for the occurrence of intracranial hemorrhage. A platelet count greater than 100/nL and a positive end expiratory pressure less than or equal to 14 cm H₂O during the first 7 days of venovenous extracorporeal membrane oxygenation therapy as well as a decrease of Paco ₂ less than 24 mm Hg during venovenous extracorporeal membrane oxygenation initiation were identified as clinical cutoff values to prevent intracranial hemorrhage (sensitivity 91% [95% CI, 82–99%], 94% [85–99%], and 67% [48–81%], respectively). Conclusions: Intracranial hemorrhage occurs early during venovenous extracorporeal membrane oxygenation and is a determinant for 60-day mortality. Appropriate adjustment of identified modifiable risk factors might lower the prevalence of intracranial hemorrhage during venovenous extracorporeal membrane oxygenation therapy.