Two-step interpretable modeling of ICU-AIs.

We present a novel methodology for integrating high resolution longitudinal data with the dynamic prediction capabilities of survival models. The aim is two-fold: to improve the predictive power while maintaining the interpretability of the models. To go beyond the black box paradigm of artificial neural networks, we propose a parsimonious and robust semi-parametric approach (i.e., a landmarking competing risks model) that combines routinely collected low-resolution data with predictive features extracted from a convolutional neural network, that was trained on high resolution time-dependent information. We then use saliency maps to analyze and explain the extra predictive power of this model. To illustrate our methodology, we focus on healthcare-associated infections in patients admitted to an intensive care unit.

Dynamic prediction of mortality in COVID-19 patients in the intensive care unit: A retrospective multi-center cohort study.

Background: The COVID-19 pandemic continues to overwhelm intensive care units (ICUs) worldwide, and improved prediction of mortality among COVID-19 patients could assist decision making in the ICU setting. In this work, we report on the development and validation of a dynamic mortality model specifically for critically ill COVID-19 patients and discuss its potential utility in the ICU. Methods: We collected electronic medical record (EMR) data from 3222 ICU admissions with a COVID-19 infection from 25 different ICUs in the Netherlands. We extracted daily observations of each patient and fitted both a linear (logistic regression) and non-linear (random forest) model to predict mortality within 24 h from the moment of prediction. Isotonic regression was used to re-calibrate the predictions of the fitted models. We evaluated the models in a leave-one-ICU-out (LOIO) cross-validation procedure. Results: The logistic regression and random forest model yielded an area under the receiver operating characteristic curve of 0.87 [0.85; 0.88] and 0.86 [0.84; 0.88], respectively. The recalibrated model predictions showed a calibration intercept of -0.04 [-0.12; 0.04] and slope of 0.90 [0.85; 0.95] for logistic regression model and a calibration intercept of -0.19 [-0.27; -0.10] and slope of 0.89 [0.84; 0.94] for the random forest model. Discussion: We presented a model for dynamic mortality prediction, specifically for critically ill COVID-19 patients, which predicts near-term mortality rather than in-ICU mortality. The potential clinical utility of dynamic mortality models such as benchmarking, improving resource allocation and informing family members, as well as the development of models with more causal structure, should be topics for future research.

Initiation of veno-arterial extracorporeal membrane oxygenation (VA-ECMO) for cardiogenic shock during out of hours versus working hours is not associated with increased mortality.

BACKGROUND: Initiation of veno-arterial (VA) Extracorporeal Membrane Oxygenator (ECMO) is associated with severe complications. It is unknown whether these adverse consequences occur more often after initiations during out of hours service compared to working hours. METHODS: All patients receiving VA-ECMO for cardiogenic shock between 2009 and 2020 were categorized into a working hours group (between 8 am and 5 pm on weekdays) and an out of hours service group (between 5 pm and 8 am, or between Friday 5 pm and Monday 8 am). Primary outcome was all-cause mortality at 30 days. Secondary outcomes included vascular complications (including limb ischemia and/or bleeding), bloodstream infections and length of ICU stay. Propensity scores were used to adjust for potential confounding effects. RESULTS: Among 250 patients (median (IQR) age 56 (42-64) years) receiving VA-ECMO (median duration 3.5 (1.0-9.0) days), 160 (64%) runs were initiated between 5 pm and 8 am whereas the remainder (36%) started during working hours. Characteristic did not differ between the working hours- and out of hours-group. By day 30, 37 (41.1%), and 68 (42.5%) patients in either group had died, respectively (p = 0.831). VA-ECMO support duration and length of stay on the ICU did not differ significantly in both crude and adjusted analyses. More complications occurred during out of hours service (p = 0.039). CONCLUSIONS: Out of hours- versus working hours-initiation of VA-ECMO for cardiogenic shock was not associated with higher mortality, longer VA-ECMO support duration, or longer length of stay on the intensive care. Vascular complications were more common in the out of hours group.

Twelve years of circulatory extracorporeal life support at the University Medical Centre Utrecht.

INTRODUCTION: Circulatory extracorporeal life support (ECLS) has been performed at the University Medical Centre Utrecht for 12 years. During this time, case mix, indications, ECLS set-ups and outcomes seem to have substantially changed. We set out to describe these characteristics and their evolution over time. METHODS: All patients receiving circulatory ECLS between 2007 and 2018 were retrospectively identified and divided into six groups according to a 2-year period of time corresponding to the date of ECLS initiation. General characteristics plus data pertaining to comorbidities, indications and technical details of ECLS commencement as well as in-hospital, 30-day, 1­year and overall mortality were collected. Temporal trends in these characteristics were examined. RESULTS: A total of 347 circulatory ECLS runs were performed in 289 patients. The number of patients and ECLS runs increased from 8 till a maximum of 40 runs a year. The distribution of circulatory ECLS indications shifted from predominantly postcardiotomy to a wider set of indications. The proportion of peripheral insertions with or without application of left ventricular unloading techniques substantially increased, while in-hospital, 30-day, 1­year and overall mortality decreased over time. CONCLUSION: Circulatory ECLS was increasingly applied at the University Medical Centre Utrecht. Over time, indications as well as treatment goals broadened, and cannulation techniques shifted from central to mainly peripheral approaches. Meanwhile, weaning success increased and mortality rates diminished.

O-serotype distribution of Escherichia coli bloodstream infection isolates in critically ill patients in The Netherlands.

OBJECTIVES: Invasive infections by extra-intestinal pathogenic Escherichia coli (ExPEC) strains are increasing. We determined O-serogroups of E. coli isolates from ICU patients having bloodstream infections (BSI) and the potential coverage of a 10-valent O-polysaccharide conjugate vaccine currently in development for the prevention of invasive ExPEC disease. METHODS: We studied E. coli BSI among patients admitted to a tertiary ICU in the Netherlands between April 2011 and November 2016. O-serogroups were determined in vitro by agglutination and whole genome sequencing. RESULTS: Among 714 ICU patients having BSI, 70 (10%) had an E. coli BSI. Among 68 (97%) isolates serogrouped, the most common serogroups were O25 (n = 11; 16%), O8 (n = 5; 7%), O2 (n = 4; 6%), O6 (n = 4; 6%), and O15 (n = 4; 6%). The theoretical coverage of a 10-valent ExPEC vaccine was 54% (n = 37). CONCLUSIONS: A multi-valent ExPEC O-polysaccharide conjugate vaccine in development could potentially aid in the prevention of E. coli BSI in Dutch ICU patients.

Attitudes of Dutch intensive care unit clinicians towards oxygen therapy.

BACKGROUND: Over the last decade, there has been an increasing awareness for the potential harm of the administration of too much oxygen. We aimed to describe self-reported attitudes towards oxygen therapy by clinicians from a large representative sample of intensive care units (ICUs) in the Netherlands. METHODS: In April 2019, 36 ICUs in the Netherlands were approached and asked to send out a questionnaire (59 questions) to their nursing and medical staff (ICU clinicians) eliciting self-reported behaviour and attitudes towards oxygen therapy in general and in specific ICU case scenarios. RESULTS: In total, 1361 ICU clinicians (71% nurses, 24% physicians) from 28 ICUs returned the questionnaire. Of responding ICU clinicians, 64% considered oxygen-induced lung injury to be a major concern. The majority of respondents considered a partial pressure of oxygen (PaO2) of 6-10 kPa (45-75 mmHg) and an arterial saturation (SaO2) of 85-90% as acceptable for 15 minutes, and a PaO2 7-10 kPa (53-75 mmHg) and SaO2 90-95% as acceptable for 24-48 hours in an acute respiratory distress syndrome (ARDS) patient. In most case scenarios, respondents reported not to change the fraction of inspired oxygen (FiO2) if SaO2 was 90-95% or PaO2 was 12 kPa (90 mmHg). CONCLUSION: A representative sample of ICU clinicians from the Netherlands were concerned about oxygen-induced lung injury, and reported that they preferred PaO2 and SaO2 targets in the lower physiological range and would adjust ventilation settings accordingly.

Profile of the SeptiCyte™ LAB gene expression assay to diagnose infection in critically ill patients.

Sepsis is a severe and frequently occurring clinical syndrome, caused by the inflammatory response to infections. Recent studies on the human transcriptome during sepsis have yielded several gene-expression assays that might assist physicians during clinical assessment of patients suspected of sepsis. SeptiCyte™ LAB (Immunexpress, Seattle, WA) is the first gene expression assay that was cleared by the FDA in the United States to distinguish infectious from non-infectious causes of systemic inflammation in critically ill patients. The test consists of the simultaneous amplification of four RNA transcripts (CEACAM4, LAMP1, PLAC8, and PLA2G7) in whole blood using a quantitative real-time PCR reaction. This review provides an overview of the challenges in the diagnosis of sepsis, the development of gene expression signatures, and a detailed description of available clinical performance studies evaluating SeptiCyte™ LAB.

Respiratory syncytial virus in critically ill adult patients with community-acquired respiratory failure: a prospective observational study.

The incidence of respiratory syncytial virus (RSV) and influenza virus infection was determined during three RSV seasons in 158 adult patients consecutively admitted to the intensive care unit with community-acquired respiratory failure. Nasopharyngeal swabs were tested for the presence of RSV and influenza virus by real-time polymerase chain reaction. Six patients (4%) were positive for RSV and all recovered. This finding was in sharp contrast to influenza (23 (15%) patients, 4 (17%) deaths). In conclusion, even in the midst of the RSV season, RSV is an infrequent cause of respiratory failure in adults admitted to the intensive care unit.

Does ICP monitoring make a difference in neurocritical care?

Raised intracranial pressure and low cerebral perfusion pressure are associated with ischaemia and poor outcome after brain injury. Therefore, many management protocols target these variables. However, there are no randomized controlled trials that have demonstrated the effectiveness of intracranial pressure-guided care in severely head-injured patients. Observational studies of such therapy have yielded inconsistent results, ranging from decreased mortality to no effect or increased morbidity or mortality. A recent cohort study supports the notion that the possible benefits of intracranial pressure monitoring after traumatic brain injury are small - if present - and would exceed a number needed for the treatment of 16. Furthermore, intracranial pressure monitoring and aggressive management of intracranial pressure and cerebral perfusion pressure have been associated with increased lengths of stay in the neurocritical care unit, conceivable costs and possibly an increased rate of complications. Against this background, there is sufficient clinical equipoise to warrant an adequately powered randomized controlled trial to compare intracranial pressure-guided care with supportive critical care without intracranial pressure monitoring in patients with severe traumatic brain injury. However, the realization of such a trial is likely to be problematic for a number of reasons, not least of which the firmly held biases of many clinicians.

Long-term propofol infusion and cardiac failure in adult head-injured patients.

Five adult patients with head injuries inexplicably had fatal cardiac arrests In our neurosurgical intensive-care unit after the introduction of a sedation formulation containing an increased concentration of propofol. To examine the possible relation further, we did a retrospective cohort analysis of head-injured adults admitted to our unit between 1996 and 1999 who were sedated and mechanically ventilated. 67 patients met the inclusion criteria, of whom seven were judged to have died from propofol-infusion syndrome. The odds ratio for the occurrence of the syndrome was 1.93 (95% CI 1.12-3.32, p=0.018) for every mg/kg per h increase in mean propofol dose above 5 mg/kg per h. We suggest that propofol infusion at rates higher than 5 mg/kg per h should be discouraged for long-term sedation in the intensive-care unit.