Development of a model-based clinical sepsis biomarker for critically ill patients

Low serum thyroid-stimulating hormone levels may be an early predictor of sepsis

OBJECTIVE

This study aimed to explore whether thyroid-stimulating hormone (TSH) plays an early warning role in detecting progression of bacterial infection to sepsis and can serve as a novel marker for the diagnosis of sepsis.

METHOD

This was a prospective study of patients treated for 'bacterial infection' in the emergency department of Beijing Chaoyang Hospital from 1 January 2021 to 31 August 2021. Subjects were divided into a sepsis group (SG) and a non-SG (NSG), according to whether their condition had progressed to sepsis within 72 hours of admission. Routine blood test results as well as biochemical and thyroid function indices (T4, FT4, T3, FT3) were recorded at the time of admission. TSH, Acute Physiology and Chronic Health Evaluation II scores and Sequential Organ Failure Assessment scores were likewise documented.

RESULTS

A total of 62 patients were enrolled, the SG and the NSG showed significant differences in their levels of TSH. The results indicate that TSH is an early warning marker for sepsis.

CONCLUSIONS

TSH plays an early warning role in the diagnosis of bacterial infection progressing to sepsis, having a strong predictive value.

Estimating Enhanced Endogenous Glucose Production in Intensive Care Unit Patients with Severe Insulin Resistance

BACKGROUND

Critically ill ICU patients frequently experience acute insulin resistance and increased endogenous glucose production, manifesting as stress-induced hyperglycemia and hyperinsulinemia. STAR (Stochastic TARgeted) is a glycemic control protocol, which directly manages inter- and intra- patient variability using model-based insulin sensitivity (SI). The model behind STAR assumes a population constant for endogenous glucose production (EGP), which is not otherwise identifiable.

OBJECTIVE

This study analyses the effect of estimating EGP for ICU patients with very low SI (severe insulin resistance) and its impact on identified, model-based insulin sensitivity identification, modeling accuracy, and model-based glycemic clinical control.

METHODS

Using clinical data from 717 STAR patients in 3 independent cohorts (Hungary, New Zealand, and Malaysia), insulin sensitivity, time of insulin resistance, and EGP values are analyzed. A method is presented to estimate EGP in the presence of non-physiologically low SI. Performance is assessed via model accuracy.

RESULTS

Results show 22%-62% of patients experience 1+ episodes of severe insulin resistance, representing 0.87%-9.00% of hours. Episodes primarily occur in the first 24 h, matching clinical expectations. The Malaysian cohort is most affected. In this subset of hours, constant model-based EGP values can bias identified SI and increase blood glucose (BG) fitting error. Using the EGP estimation method presented in these constrained hours significantly reduced BG fitting errors.

CONCLUSIONS

Patients early in ICU stay may have significantly increased EGP. Increasing modeled EGP in model-based glycemic control can improve control accuracy in these hours. The results provide new insight into the frequency and level of significantly increased EGP in critical illness.

Changes in Biomarkers and Hemodynamics According to Antibiotic Susceptibility in a Model of Bacteremia

Proper selection of susceptible antibiotics in drug-resistant bacteria is critical to treat bloodstream infection. Although biomarkers that guide antibiotic therapy have been extensively evaluated, little is known about host biomarkers targeting in vivo antibiotic susceptibility. Therefore, we aimed to evaluate the trends of hemodynamics and biomarkers in a porcine bacteremia model treated with insusceptible antibiotics compared to those in susceptible models. Extended-spectrum β-lactamase (ESBL)-producing Escherichia coli (E. coli, 5.0 * 10^9 CFU) was intravenously administered to 11 male pigs. One hour after bacterial infusion, pigs were assigned to two groups of antibiotics, ceftriaxone (n = 6) or ertapenem (n = 5). Pigs were monitored up to 7 h after bacterial injection with fluid and vasopressor support to maintain the mean arterial blood pressure over 65 mmHg. Blood sampling for blood culture and plasma acquisition was performed before and every predefined hour after E. coli injection. Cytokine (tumor necrosis factor-α, interleukin [IL]-1β, IL-6, IL-8, IL-10, C-reactive protein, procalcitonin, presepsin, heparan sulfate, syndecan, and soluble triggering receptor expressed on myeloid cells-1 [sTREM-1]) levels in plasma were analyzed using enzyme-linked immunosorbent assays. Bacteremia developed after intravenous injection of E. coli, and negative conversion was confirmed only in the ertapenem group. While trends of other biomarkers failed to show differences, the trend of sTREM-1 was significantly different between the two groups (P = 0.0001, two-way repeated measures analysis of variance). Among hemodynamics and biomarkers, the sTREM-1 level at post 2 h after antibiotics administration represented a significant difference depending on susceptibility, which can be suggested as a biomarker candidate of in vivo antibiotics susceptibility. Further clinical studies are warranted for validation.

IMPORTANCE Early and appropriate antibiotic treatment is a keystone in treating patients with sepsis. Despite its importance, blood culture which requires a few days remains as a pillar of diagnostic method for microorganisms and their antibiotic susceptibility. Whether changes in biomarkers and hemodynamics indicate treatment response of susceptible antibiotic compared to resistant one is not well understood to date. In this study using extended-spectrum β-lactamase -producing E. coli bacteremia porcine model, we have demonstrated the comprehensive cardiovascular hemodynamics and trends of plasma biomarkers in sepsis and compared them between two groups with susceptible and resistant antibiotics. While other hemodynamics and biomarkers have failed to differ, we have identified that levels of soluble triggering receptor expressed on myeloid cells-1 (sTREM-1) significantly differed between the two groups over time. Based on the data in this study, trends of sTREM-1 obtained before the antibiotics and 2~4 h after the antibiotics could be a novel host biomarker that triggers the step-up choice of antibiotics.

Kernel density estimates for sepsis classification

OBJECTIVE

Severe sepsis is a leading cause of intensive care unit (ICU) admission, length of stay, mortality, and cost. systemic inflammatory response syndrome (SIRS) and organ failure due to infection define it, but also make it hard to diagnose. Early diagnosis reduces morbidity, mortality and cost, and diagnosis is often significantly delayed due to a lack of effective biomarkers. This research employs kernel density estimation (KDE) methods fusing a personalized, model-based insulin sensitivity (SI) metric with standard bedside measures of: temperature, heart rate, respiratory rate, systolic and diastolic blood pressure, and SIRS, as these measures are available hourly or more frequently.

METHODS

Model-based SI is a derived metric, identified using clinical data and a clinically validated metabolic model. The KDE classifier discriminates severe sepsis and septic shock from moderate sepsis using accepted consensus sepsis scores. A best case in-sample estimate, a worst case independent cross validation estimate, and an accepted .632 bootstrap estimate are calculated to assess performance using multi-level likelihood ratios, and sensitivity and specificity. Performance is assessed against clinically and statistically defined thresholds denoted for the minimum acceptable level as: "high accuracy, often providing useful information, and clinical significance," and similar definitions for greater or lesser quality.

RESULTS

The .632 bootstrap estimate performs near clinically defined levels of high accuracy, often providing useful information, and clinical significance based on sensitivity, specificity, and multilevel likelihood ratios.

CONCLUSION AND SIGNIFICANCE

The classifier created and this overall approach is useful for clinical decision making in diagnosing severe sepsis and septic shock in real time, for both case and control hours. However, improvements could be made with larger clinical data sets.

Next-generation, personalised, model-based critical care medicine: a state-of-the art review of in silico virtual patient models, methods, and cohorts, and how to validation them

Critical care, like many healthcare areas, is under a dual assault from significantly increasing demographic and economic pressures. Intensive care unit (ICU) patients are highly variable in response to treatment, and increasingly aging populations mean ICUs are under increasing demand and their cohorts are increasingly ill. Equally, patient expectations are growing, while the economic ability to deliver care to all is declining. Better, more productive care is thus the big challenge. One means to that end is personalised care designed to manage the significant inter- and intra-patient variability that makes the ICU patient difficult. Thus, moving from current "one size fits all" protocolised care to adaptive, model-based "one method fits all" personalised care could deliver the required step change in the quality, and simultaneously the productivity and cost, of care. Computer models of human physiology are a unique tool to personalise care, as they can couple clinical data with mathematical methods to create subject-specific models and virtual patients to design new, personalised and more optimal protocols, as well as to guide care in real-time. They rely on identifying time varying patient-specific parameters in the model that capture inter- and intra-patient variability, the difference between patients and the evolution of patient condition. Properly validated, virtual patients represent the real patients, and can be used in silico to test different protocols or interventions, or in real-time to guide care. Hence, the underlying models and methods create the foundation for next generation care, as well as a tool for safely and rapidly developing personalised treatment protocols over large virtual cohorts using virtual trials. This review examines the models and methods used to create virtual patients. Specifically, it presents the models types and structures used and the data required. It then covers how to validate the resulting virtual patients and trials, and how these virtual trials can help design and optimise clinical trial. Links between these models and higher order, more complex physiome models are also discussed. In each section, it explores the progress reported up to date, especially on core ICU therapies in glycemic, circulatory and mechanical ventilation management, where high cost and frequency of occurrence provide a significant opportunity for model-based methods to have measurable clinical and economic impact. The outcomes are readily generalised to other areas of medical care.

Impact of haemodialysis on insulin sensitivity of acute renal failure (ARF) patients with sepsis in critical care

Critically ill patients often develop renal failure in addition to their primary diagnosis. However, the effect and impact of haemodialysis (HD) on insulin sensitivity n critically ill patients remains unclear. Specifically, this study investigates insulin sensitivity of acute renal failure (ARF) patients with sepsis who underwent HD and glycaemic control. Model-based insulin sensitivity (SI) profiles were identified for 20 critically ill ARF patients on Specialized Relative Insulin Nutrition Titration (SPRINT) glycaemic control during intervals onto HD (OFF/ON), and after HD (ON/OFF). Patients exhibited a median -18% (IQR -36% to -5% p<0.05) reduction in measured SI after the OFF/ON dialysis transition, and a median 9% (IQR -5% to 37%, p<0.05) rise after the ON/OFF transition. Almost 80% of patients exhibited decreased SI at the OFF/ON interval, and 60% exhibited increased SI at the ON/OFF transition. Results indicate that HD commencement has significant effect on insulin pharmacokinetics at a cohort and per-patient level. These results provide the data to design conclusive studies of HD effects on SI, and to inform glycaemic control protocol development and implementation for this specific group of critically ill patients with ARF-sepsis.

Impact of Haemodialysis on Insulin Kinetics of Acute Kidney Injury Patients in Critical Care

Critically ill patients are occasionally associated with an abrupt decline in renal function secondary to their primary diagnosis. The effect and impact of haemodialysis (HD) on insulin kinetics and endogenous insulin secretion in critically ill patients remains unclear. This study investigates the insulin kinetics of patients with severe acute kidney injury (AKI) who required HD treatment and glycaemic control (GC). Evidence shows that tight GC benefits the onset and progression of renal involvement in precocious phases of diabetic nephropathy for type 2 diabetes. The main objective of GC is to reduce hyperglycaemia while determining insulin sensitivity. Insulin sensitivity (S_I) is defined as the body response to the effects of insulin by lowering blood glucose levels. Particularly, this study used S_I to track changes in insulin levels during HD therapy. Model-based insulin sensitivity profiles were identified for 51 critically ill patients with severe AKI on specialized relative insulin nutrition titration GC during intervals on HD (OFF/ON) and after HD (ON/OFF). The metabolic effects of HD were observed through changes in S_I over the ON/OFF and OFF/ON transitions. Changes in model-based S_I at the OFF/ON and ON/OFF transitions indicate changes in endogenous insulin secretion and/or changes in effective insulin clearance. Patients exhibited a median reduction of −29 % (interquartile range (IQR): [−58, 6 %], p = 0.02) in measured S_I after the OFF/ON dialysis transition, and a median increase of +9 % (IQR −15 to 28 %, p = 0.7) after the ON/OFF transition. Almost 90 % of patients exhibited decreased S_I at the OFF/ON transition, and 55 % exhibited increased S_I at the ON/OFF transition. Results indicate that HD commencement has a significant effect on insulin pharmacokinetics at a cohort and per-patient level. These changes in metabolic behaviour are most likely caused by changes in insulin clearance or/and endogenous insulin secretion.

Interface design and human factors considerations for model-based tight glycemic control in critical care

INTRODUCTION

Tight glycemic control (TGC) has shown benefits but has been difficult to implement. Model-based methods and computerized protocols offer the opportunity to improve TGC quality and compliance. This research presents an interface design to maximize compliance, minimize real and perceived clinical effort, and minimize error based on simple human factors and end user input.

METHOD

The graphical user interface (GUI) design is presented by construction based on a series of simple, short design criteria based on fundamental human factors engineering and includes the use of user feedback and focus groups comprising nursing staff at Christchurch Hospital. The overall design maximizes ease of use and minimizes (unnecessary) interaction and use. It is coupled to a protocol that allows nurse staff to select measurement intervals and thus self-manage workload.

RESULTS

The overall GUI design is presented and requires only one data entry point per intervention cycle. The design and main interface are heavily focused on the nurse end users who are the predominant users, while additional detailed and longitudinal data, which are of interest to doctors guiding overall patient care, are available via tabs. This dichotomy of needs and interests based on the end user's immediate focus and goals shows how interfaces must adapt to offer different information to multiple types of users.

CONCLUSIONS

The interface is designed to minimize real and perceived clinical effort, and ongoing pilot trials have reported high levels of acceptance. The overall design principles, approach, and testing methods are based on fundamental human factors principles designed to reduce user effort and error and are readily generalizable.

Data entry errors and design for model-based tight glycemic control in critical care

INTRODUCTION

Tight glycemic control (TGC) has shown benefits but has been difficult to achieve consistently. Model-based methods and computerized protocols offer the opportunity to improve TGC quality but require human data entry, particularly of blood glucose (BG) values, which can be significantly prone to error. This study presents the design and optimization of data entry methods to minimize error for a computerized and model-based TGC method prior to pilot clinical trials.

METHOD

To minimize data entry error, two tests were carried out to optimize a method with errors less than the 5%-plus reported in other studies. Four initial methods were tested on 40 subjects in random order, and the best two were tested more rigorously on 34 subjects. The tests measured entry speed and accuracy. Errors were reported as corrected and uncorrected errors, with the sum comprising a total error rate. The first set of tests used randomly selected values, while the second set used the same values for all subjects to allow comparisons across users and direct assessment of the magnitude of errors. These research tests were approved by the University of Canterbury Ethics Committee.

RESULTS

The final data entry method tested reduced errors to less than 1-2%, a 60-80% reduction from reported values. The magnitude of errors was clinically significant and was typically by 10.0 mmol/liter or an order of magnitude but only for extreme values of BG < 2.0 mmol/liter or BG > 15.0-20.0 mmol/liter, both of which could be easily corrected with automated checking of extreme values for safety.

CONCLUSIONS

The data entry method selected significantly reduced data entry errors in the limited design tests presented, and is in use on a clinical pilot TGC study. The overall approach and testing methods are easily performed and generalizable to other applications and protocols.

Physiological modeling, tight glycemic control, and the ICU clinician: what are models and how can they affect practice?

Critically ill patients are highly variable in their response to care and treatment. This variability and the search for improved outcomes have led to a significant increase in the use of protocolized care to reduce variability in care. However, protocolized care does not address the variability of outcome due to inter- and intra-patient variability, both in physiological state, and the response to disease and treatment. This lack of patient-specificity defines the opportunity for patient-specific approaches to diagnosis, care, and patient management, which are complementary to, and fit within, protocolized approaches.Computational models of human physiology offer the potential, with clinical data, to create patient-specific models that capture a patient's physiological status. Such models can provide new insights into patient condition by turning a series of sometimes confusing clinical data into a clear physiological picture. More directly, they can track patient-specific conditions and thus provide new means of diagnosis and opportunities for optimising therapy.This article presents the concept of model-based therapeutics, the use of computational models in clinical medicine and critical care in specific, as well as its potential clinical advantages, in a format designed for the clinical perspective. The review is presented in terms of a series of questions and answers. These aspects directly address questions concerning what makes a model, how it is made patient-specific, what it can be used for, its limitations and, importantly, what constitutes sufficient validation.To provide a concrete foundation, the concepts are presented broadly, but the details are given in terms of a specific case example. Specifically, tight glycemic control (TGC) is an area where inter- and intra-patient variability can dominate the quality of care control and care received from any given protocol. The overall review clearly shows the concept and significant clinical potential of using computational models in critical care medicine.