Identifying the need for infection-related consultations in intensive care patients using machine learning models.

Infection-related consultations on intensive care units (ICU) have a positive impact on quality of care and clinical outcome. However, timing of these consultations is essential and to date they are typically event-triggered and reactive. Here, we investigate a proactive approach to identify patients in need for infection-related consultations by machine learning models using routine electronic health records. Data was retrieved from a mixed ICU at a large academic tertiary care hospital including 9684 admissions. Infection-related consultations were predicted using logistic regression, random forest, gradient boosting machines, and long short-term memory neural networks (LSTM). Overall, 7.8% of admitted patients received an infection-related consultation. Time-sensitive modelling approaches performed better than static approaches. Using LSTM resulted in the prediction of infection-related consultations in the next clinical shift (up to eight hours in advance) with an area under the receiver operating curve (AUROC) of 0.921 and an area under the precision recall curve (AUPRC) of 0.541. The successful prediction of infection-related consultations for ICU patients was done without the use of classical triggers, such as (interim) microbiology reports. Predicting this key event can potentially streamline ICU and consultant workflows and improve care as well as outcome for critically ill patients with (suspected) infections.

Adverse drug events caused by three high-risk drug-drug interactions in patients admitted to intensive care units: A multicentre retrospective observational study.

AIMS: Knowledge about adverse drug events caused by drug-drug interactions (DDI-ADEs) is limited. We aimed to provide detailed insights about DDI-ADEs related to three frequent, high-risk potential DDIs (pDDIs) in the critical care setting: pDDIs with international normalized ratio increase (INR+ ) potential, pDDIs with acute kidney injury (AKI) potential, and pDDIs with QTc prolongation potential. METHODS: We extracted routinely collected retrospective data from electronic health records of intensive care units (ICUs) patients (≥18 years), admitted to ten hospitals in the Netherlands between January 2010 and September 2019. We used computerized triggers (e-triggers) to preselect patients with potential DDI-ADEs. Between September 2020 and October 2021, clinical experts conducted a retrospective manual patient chart review on a subset of preselected patients, and assessed causality, severity, preventability, and contribution to ICU length of stay of DDI-ADEs using internationally prevailing standards. RESULTS: In total 85 422 patients with ≥1 pDDI were included. Of these patients, 32 820 (38.4%) have been exposed to one of the three pDDIs. In the exposed group, 1141 (3.5%) patients were preselected using e-triggers. Of 237 patients (21%) assessed, 155 (65.4%) experienced an actual DDI-ADE; 52.9% had severity level of serious or higher, 75.5% were preventable, and 19.3% contributed to a longer ICU length of stay. The positive predictive value was the highest for DDI-INR+ e-trigger (0.76), followed by DDI-AKI e-trigger (0.57). CONCLUSION: The highly preventable nature and severity of DDI-ADEs, calls for action to optimize ICU patient safety. Use of e-triggers proved to be a promising preselection strategy.

Low Cancer Incidence in Naked Mole-Rats May Be Related to Their Inability to Express the Warburg Effect.

Metabolic flexibility in mammals enables stressed tissues to generate additional ATP by converting large amounts of glucose into lactic acid; however, this process can cause transient local or systemic acidosis. Certain mammals are adapted to extreme environments and are capable of enhanced metabolic flexibility as a specialized adaptation to challenging habitat niches. For example, naked mole-rats (NMRs) are a fossorial and hypoxia-tolerant mammal whose metabolic responses to environmental stressors markedly differ from most other mammals. When exposed to hypoxia, NMRs exhibit robust hypometabolism but develop minimal acidosis. Furthermore, and despite a very long lifespan relative to other rodents, NMRs have a remarkably low cancer incidence. Most advanced cancers in mammals display increased production of lactic acid from glucose, irrespective of oxygen availability. This hallmark of cancer is known as the Warburg effect (WE). Most malignancies acquire this metabolic phenotype during their somatic evolution, as the WE benefits tumor growth in several ways. We propose that the peculiar metabolism of the NMR makes development of the WE inherently difficult, which might contribute to the extraordinarily low cancer rate in NMRs. Such an adaptation of NMRs to their subterranean environment may have been facilitated by modified biochemical responses with a stronger inhibition of the production of CO2 and lactic acid by a decreased extracellular pH. Since this pH-inhibition could be deeply hard-wired in their metabolic make-up, it may be difficult for malignant cells in NMRs to acquire the WE-phenotype that facilitates cancer growth in other mammals. In the present commentary, we discuss this idea and propose experimental tests of our hypothesis.

Multi-infusion with integrated multiple pressure sensing allows earlier detection of line occlusions.

BACKGROUND: Occlusions of intravenous (IV) tubing can prevent vital and time-critical medication or solutions from being delivered into the bloodstream of patients receiving IV therapy. At low flow rates (≤ 1 ml/h) the alarm delay (time to an alert to the user) can be up to 2 h using conventional pressure threshold algorithms. In order to reduce alarm delays we developed and evaluated the performance of two new real-time occlusion detection algorithms and one co-occlusion detector that determines the correlation in trends in pressure changes for multiple pumps. METHODS: Bench-tested experimental runs were recorded in triplicate at rates of 1, 2, 4, 8, 16, and 32 ml/h. Each run consisted of 10 min of non-occluded infusion followed by a period of occluded infusion of 10 min or until a conventional occlusion alarm at 400 mmHg occurred. The first algorithm based on binary logistic regression attempts to detect occlusions based on the pump's administration rate Q(t) and pressure sensor readings P(t). The second algorithm continuously monitored whether the actual variation in the pressure exceeded a threshold of 2 standard deviations (SD) above the baseline pressure. When a pump detected an occlusion using the SD algorithm, a third algorithm correlated the pressures of multiple pumps to detect the presence of a shared occlusion. The algorithms were evaluated using 6 bench-tested baseline single-pump occlusion scenarios, 9 single-pump validation scenarios and 7 multi-pump co-occlusion scenarios (i.e. with flow rates of 1 + 1, 1 + 2, 1 + 4, 1 + 8, 1 + 16, and 1 + 32 ml/h respectively). Alarm delay was the primary performance measure. RESULTS: In the baseline single-pump occlusion scenarios, the overall mean ± SD alarm delay of the regression and SD algorithms were 1.8 ± 0.8 min and 0.4 ± 0.2 min, respectively. Compared to the delay of the conventional alarm this corresponds to a mean time reduction of 76% (P = 0.003) and 95% (P = 0.001), respectively. In the validation scenarios the overall mean ± SD alarm delay of the regression and SD algorithms were respectively 1.8 ± 1.6 min and 0.3 ± 0.2 min, corresponding to a mean time reduction of 77% and 95%. In the multi-pump scenarios a correlation > 0.8 between multiple pump pressures after initial occlusion detection by the SD algorithm had a mean ± SD alarm delay of 0.4 ± 0.2 min. In 2 out of the 9 validation scenarios an occlusion was not detected by the regression algorithm before a conventional occlusion alarm occurred. Otherwise no occlusions were missed. CONCLUSIONS: In single pumps, both the regression and SD algorithm considerably reduced alarm delay compared to conventional pressure limit-based detection. The SD algorithm appeared to be more robust than the regression algorithm. For multiple pumps the correlation algorithm reliably detected co-occlusions. The latter may be used to localize the segment of tubing in which the occlusion occurs. Trial registration Not applicable.

Towards more efficient use of intravenous lumens in multi-infusion settings: development and evaluation of a multiplex infusion scheduling algorithm.

BACKGROUND: Multi-drug intravenous (IV) therapy is one of the most common medical procedures used in intensive care units (ICUs), operating rooms, oncology wards and many other hospital departments worldwide. As drugs or their solvents are frequently chemically incompatible, many solutions must be administered through separate lumens. When the number of available lumens is too low to facilitate the safe administration of these solutions, additional (peripheral) IV catheters are often required, causing physical discomfort and increasing the risk for catheter related complications. Our objective was to develop and evaluate an algorithm designed to reduce the number of intravenous lumens required in multi-infusion settings by multiplexing the administration of various parenteral drugs and solutions. METHODS: A multiplex algorithm was developed that schedules the alternating IV administration of multiple incompatible IV solutions through a single lumen, taking compatibility-related, pharmacokinetic and pharmacodynamic constraints of the relevant drugs into account. The conventional scheduling procedure executed by ICU nurses was used for comparison. The number of lumens required by the conventional procedure (LCONV) and multiplex algorithm (LMX) were compared. RESULTS: We used data from 175,993 ICU drug combinations, with 2251 unique combinations received by 2715 consecutive ICU patients. The mean ± SD number of simultaneous IV solutions was 2.8 ± 1.6. In 27% of all drug combinations, and 61% of the unique combinations the multiplex algorithm required fewer lumens (p < 0.001). With increasing LCONV, the reduction in number of lumens by the multiplex algorithm further increased (p < 0.001). In only 1% of cases multiplexing required > 3 lm, versus 12% using the conventional procedure. CONCLUSION: The multiplex algorithm addresses a major issue that occurs in ICUs, operating rooms, oncology wards, and many other hospital departments where several incompatible drugs are infused through a restricted number of lumens. The multiplex algorithm allows for more efficient use of IV lumens compared to the conventional multi-infusion strategy.

Protocol of the sepsivit study: a prospective observational study to determine whether continuous heart rate variability measurement during the first 48 hours of hospitalisation provides an early warning for deterioration in patients presenting with infection or sepsis to the emergency department of a Dutch academic teaching hospital.

INTRODUCTION: One in five patients with sepsis deteriorates within 48 hours after hospital admission. Regrettably, a clear tool for the early detection of deterioration is still lacking. The SepsiVit study aims to determine whether continuous heart rate variability (HRV) measurement can provide an early warning for deterioration in patients presenting with suspected infection or sepsis to the emergency department (ED). METHODS AND ANALYSIS: The protocol of a prospective observational study in the ED. We will include 171 adult medical patients presenting with suspected infection or sepsis and at least two systemic inflammatory response syndrome criteria. Patients with known pregnancy, cardiac transplantation or not admitted to our hospital are excluded.High sample frequency ECG signals (500 Hz), respiratory rate, blood pressure and peripheral oxygen saturation will be recorded continuously during the first 48 hours of hospitalisation using a bedside patient monitor (Philips IntelliVue MP70). Primary endpoint is patient deterioration, defined as the development of organ dysfunction, unplanned intensive care unit admission or in-hospital mortality. The ECG data will be used for offline HRV analysis. We will compare the HRV between two groups (deterioration/no deterioration) and analyse whether HRV provides an early warning for deterioration. Furthermore, we will create a multivariate predictive model for deterioration based on heart rate, respiratory rate and HRV. As planned secondary analyses, we (1) perform a subgroup analysis for patients with pneumosepsis and urosepsis and (2) determine whether HRV using lower sample frequencies (1 Hz or less) suffices to predict deterioration. ETHICS AND DISSEMINATION: The Institutional Review Board of the University Medical Center Groningen granted a waiver for the study (METc 2015/164). Results will be disseminated through international peer-reviewed publications and conference presentations. A lay summary of the results will be provided to the study participants. TRIAL REGISTRATION NUMBER: NTR6168; Pre-results.

Postoperative fluid retention after heart surgery is accompanied by a strongly positive sodium balance and a negative potassium balance.

The conventional model on the distribution of electrolyte infusions states that water will distribute proportionally over both the intracellular (ICV) and extracellular (ECV) volumes, while potassium homes to the ICV and sodium to the ECV Therefore, total body potassium is the most accurate measure of ICV and thus potassium balances can be used to quantify changes in ICV In cardiothoracic patients admitted to the ICU we performed complementary balance studies to measure changes in ICV and ECV In 39 patients, fluid, sodium, potassium, and electrolyte-free water (EFW) balances were determined to detect changes in ICV and ECV Cumulatively over 4 days, these patients received a mean ± SE infusion of 14.0 ± 0.6 L containing 1465 ± 79 mmol sodium, 196 ± 11 mmol potassium and 2.1 ± 0.1 L EFW This resulted in strongly positive fluid (4.0 ± 0.6 L) and sodium (814 ± 75 mmol) balances but in negative potassium (-101 ± 14 mmol) and EFW (-1.1 ± 0.2 L) balances. We subsequently compared potassium balances (528 patients) and fluid balances (117 patients) between patients who were assigned to either a 4.0 or 4.5 mmol/L blood potassium target. Although fluid balances were similar in both groups, the additionally administered potassium (76 ± 23 mmol) in the higher target group was fully excreted by the kidneys (70 ± 23 mmol). These findings indicate that even in the context of rapid and profound volume expansion neither water nor potassium moves into the ICV.