Jennifer L. Dickson

Generalisability of a Virtual Trials Method for Glycaemic Control in Intensive Care

Background: Elevated blood glucose (BG) concentrations (Hyperglycaemia) are a common complication in critically ill patients. Insulin therapy is commonly used to treat hyperglycaemia, but metabolic variability often results in poor BG control and low BG (hypoglycaemia). Objective: This paper presents a model-based virtual trial method for glycaemic control protocol design, and evaluates its generalisability across different populations. Methods: Model-based insulin sensitivity (SI) was used to create virtual patients from clinical data from three different ICUs in New Zealand, Hungary, and Belgium. Glycaemic results from simulation of virtual patients under their original protocol (self-simulation) and protocols from other units (cross simulation) were compared. Results: Differences were found between the three cohorts in median SI and inter-patient variability in SI. However, hour-to-hour intra-patient variability in SI was found to be consistent between cohorts. Self and cross-simulation results were found to have overall similarity and consistency, though results may differ in the first 24–48 h due to different cohort starting BG and underlying SI. Conclusions and Significance: Virtual patients and the virtual trial method were found to be generalisable across different ICUs. This virtual trial method is useful for in silico protocol design and testing, given an understanding of the underlying assumptions and limitations of this method.

On the Problem of Patient-Specific Endogenous Glucose Production in Neonates on Stochastic Targeted Glycemic Control

Background: Both stress and prematurity can induce hyperglycemia in the neonatal intensive care unit, which, in turn, is associated with worsened outcomes. Endogenous glucose production (EGP) is the formation of glucose by the body from substrates and contributes to blood glucose (BG) levels. Due to the inherent fragility of the extremely low birth weight (ELBW) neonates, true fasting EGP cannot be explicitly determined, introducing uncertainty into glycemic models that rely on quantifying glucose sources. Stochastic targeting, or STAR, is one such glycemic control framework. Methods: A literature review was carried out to gather metabolic and EGP values on preterm infants with a gestational age (GA) <32 weeks and a birth weight (BW) <2 kg. The data were analyzed for EGP trends with BW, GA, BG, plasma insulin, and glucose infusion (GI) rates. Trends were modeled and compared with a literature-derived range of population constant EGP models using clinically validated virtual trials on retrospective clinical data. Results: No clear relationship was found for EGP and BW, GA, or plasma insulin. Some evidence of suppression of EGP with increasing GI or BG was seen. Virtual trial results showed that population-constant EGP models fit clinical data best and gave tighter control performance to a target band in virtual trials. Conclusions: Variation in EGP cannot easily be quantified, and EGP is sufficiently modeled as a population constant in the neonatal intensive care insulin-nutrition-glucose model. Analysis of the clinical data and fitting error suggests that ELBW hyperglycemic preterm neonates have unsuppressed EGP in the higher range than that seen in literature.

Continuous glucose monitoring in neonates: a review

Continuous glucose monitoring (CGM) is well established in the management of diabetes mellitus, but its role in neonatal glycaemic control is less clear. CGM has provided important insights about neonatal glucose metabolism, and there is increasing interest in its clinical use, particularly in preterm neonates and in those in whom glucose control is difficult. Neonatal glucose instability, including hypoglycaemia and hyperglycaemia, has been associated with poorer neurodevelopment, and CGM offers the possibility of adjusting treatment in real time to account for individual metabolic requirements while reducing the number of blood tests required, potentially improving long-term outcomes. However, current devices are optimised for use at relatively high glucose concentrations, and several technical issues need to be resolved before real-time CGM can be recommended for routine neonatal care. These include: 1) limited point accuracy, especially at low or rapidly changing glucose concentrations; 2) calibration methods that are designed for higher glucose concentrations of children and adults, and not for neonates; 3) sensor drift, which is under-recognised; and 4) the need for dynamic and integrated metrics that can be related to long-term neurodevelopmental outcomes. CGM remains an important tool for retrospective investigation of neonatal glycaemia and the effect of different treatments on glucose metabolism. However, at present CGM should be limited to research studies, and should only be introduced into routine clinical care once benefit is demonstrated in randomised trials.

External validation and sub-cohort analysis of stochastic forecasting models in NICU cohorts

Abstract Hyperglycaemia is a prevalent complication in the neonatal intensive care unit (NICU) and is associated with worsened outcomes. It occurs as a result of prematurity, under-developed endogenous glucose regulatory systems, and clinical stress. The stochastic targeting (STAR) framework provides patient-specific, model-based glycaemic control with a clinically proven level of confidence on the outcome of treatment interventions, thus directly managing the risk of hypo- and hyper-glycaemia. However, stochastic models that are over conservative can limit control performance. Retrospective clinical data from 61 episodes (25 retrospective to STAR, and 36 from a prospective-STAR blood glucose control study) of insulin therapy in very-low birth weight (VLBW) and extremely-low birth weight (ELBW) neonates are used to create a new stochastic model of model-based insulin sensitivity ( S I [L/mU/min]). Sub-cohort models based on gestational age (GA) and birth weight (BW) are also created. Performance is assessed by the percentage of patients who have 90% of actual intra-patient variability in S I captured by the 90% confidence bands of the cohort based (inter-patient) stochastic variability model created. This assessment measures per-patient accuracy for any given cohort model. Per-patient coverage trends were very similar between prospective and retrospective cohorts, providing a measure of external validation of cohort similarity. Per-patient coverage was improved through the use of BW and GA dependent stochastic models, which ensures that the stochastic models more accurately capture both inter- and intra-patient variability. Stochastic models based on insulin sensitivities during insulin treatment periods are tighter, and give better and safer glycaemic control. Overall it seems that inter-patient variation is more significant than intra-patient variation as a limiting factor in this stochastic forecasting model, and a small number of patients are essentially different in behaviour. More patient specific methods, particularly in the modelling of endogenous insulin and glucose production, will be required to further improve forecasting and glycaemic control.

Hyperglycaemic Preterm Babies Have Sex Differences in Insulin Secretion

Background: Hyperglycaemia is a common complication of prematurity and is associated with neonatal mortality and morbidity, yet the aetiology is incompletely understood. C-peptide has been used in adults to estimate endogenous insulin secretion due to its simple clearance kinetics. Objective: To determine insulin secretion calculated from plasma C-peptide concentrations in hyperglycaemic preterm babies. Methods: This is a retrospective analysis of a cohort of 41 very preterm babies with a median gestational age of 27.2 weeks (26.2-28.7) enrolled in a randomised controlled trial of tight glycaemic control when they developed hyperglycaemia (2 consecutive blood glucose concentrations, BGC, > 8.5 mmol·l-1). Insulin secretion was determined using a steady state analysis of a 2-compartment C-peptide kinetic model. Results: BGC, plasma insulin concentration, plasma C-peptide concentrations, and insulin secretion were higher at randomisation than 1-2 weeks following randomisation (p ≤ 0.02). Insulin secretion was higher in girls at 11.7 mU·l-1·kg-1·min-1 (5.3-18.7) vs. 4.7 mU·l-1·kg-1·min-1 (2.1-8.3; p < 0.005), with no difference in clinical characteristics, BGC, plasma insulin concentration, or nutrition between the sexes (p > 0.25). Insulin secretion was lower in samples taken during exogenous insulin delivery at 3.7 mU·l-1·kg-1·min-1 (1.8-6.9) vs. 9.8 mU·l-1·kg-1·min-1 (4.7-17.8; p = 0.02). Conclusions: Insulin secretion was higher when babies had higher BGC, indicating that endogenous insulin secretion is sensitive to BGC. Girls had higher insulin secretion, at similar blood glucose and plasma insulin concentrations, than boys.

Autoregressive Modeling of Drift and Random Error to Characterize a Continuous Intravascular Glucose Monitoring Sensor

Background: Continuous glucose monitoring (CGM) devices have been effective in managing diabetes and offer potential benefits for use in the intensive care unit (ICU). Use of CGM devices in the ICU has been limited, primarily due to the higher point accuracy errors over currently used traditional intermittent blood glucose (BG) measures. General models of CGM errors, including drift and random errors, are lacking, but would enable better design of protocols to utilize these devices. This article presents an autoregressive (AR) based modeling method that separately characterizes the drift and random noise of the GlySure CGM sensor (GlySure Limited, Oxfordshire, UK). Methods: Clinical sensor data (n = 33) and reference measurements were used to generate 2 AR models to describe sensor drift and noise. These models were used to generate 100 Monte Carlo simulations based on reference blood glucose measurements. These were then compared to the original CGM clinical data using mean absolute relative difference (MARD) and a Trend Compass. Results: The point accuracy MARD was very similar between simulated and clinical data (9.6% vs 9.9%). A Trend Compass was used to assess trend accuracy, and found simulated and clinical sensor profiles were similar (simulated trend index 11.4° vs clinical trend index 10.9°). Conclusion: The model and method accurately represents cohort sensor behavior over patients, providing a general modeling approach to any such sensor by separately characterizing each type of error that can arise in the data. Overall, it enables better protocol design based on accurate expected CGM sensor behavior, as well as enabling the analysis of what level of each type of sensor error would be necessary to obtain desired glycemic control safety and performance with a given protocol.

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.

Model-based glycaemic control: methodology and initial results from neonatal intensive care.

Abstract Very/extremely premature infants often experience glycaemic dysregulation, resulting in abnormally elevated (hyperglycaemia) or low (hypoglycaemia) blood glucose (BG) concentrations, due to prematurity, stress, and illness. STAR-GRYPHON is a computerised protocol that utilises a model-based insulin sensitivity parameter to directly tailor therapy for individual patients and their changing conditions, unlike other common insulin protocols in this cohort. From January 2013 to January 2015, 13 patients totalling 16 hyperglycaemic control episodes received insulin under STAR-GRYPHON. A significant improvement in control was achieved in comparison to a retrospective cohort, with a 26% absolute improvement in BG within the targeted range and no hypoglycaemia. This improvement was obtained predominantly due to the reduction of hyperglycaemia (%BG>10.0 mmol/l: 5.6 vs. 17.7%, p<0.001), and lowering of the median per-patient BG [6.9 (6.1–7.9) vs. 7.8 (6.6–9.1) mmol/l, p<0.001, Mann-Witney U test]. While cohort-wide control results show good control overall, there is high intra-patient variability in BG behaviour, resulting in overly conservative treatments for some patients. Patient insulin sensitivity differs between and within patients over time, with some patients having stable insulin sensitivity, while others change rapidly. These results demonstrate the trade-off between safety and performance in a highly variable and fragile cohort.

Performance and Safety of STAR Glycaemic Control in Neonatal Intensive Care: Further Clinical Results Including Pilot Results from a New Protocol Implementation

Abstract Elevated blood glucose concentrations (BG) (Hyperglycaemia) are a common complication of prematurity in extremely low birth weight neonates in the neonatal intensive care unit (NICU), and are associated with increased mortality and morbidity. Insulin therapy allows glucose tolerance and weight gain to be increased. However, insulin therapy is commonly associated with a significant increase in low BG events (hypoglycaemia), which is also associated with adverse outcomes. Controlling BG levels via nutrition restriction reduces infant growth and is thus undesirable. STAR (Stochastic TARgeted) is a model-based glycaemic framework that mitigates the risks of hypoglycaemia through quantification of current insulin sensitivity and future variability. From August 2008 to December 2012 40 patients totaling 61 glycaemic episodes were treated with STAR in the NICU (STAR-NICU). Percentage time in the clinically targeted 4.0-8.0 mmol/L band was 62%, a 14% increase compared to retrospective data and hyperglycaemia (BG>10.0mmol/L) was halved. Overall incidence of severe hypoglycemia (BG

Brain mass estimation by head circumference and body mass methods in neonatal glycaemic modelling and control

INTRODUCTION Hyperglycaemia is a common complication of stress and prematurity in extremely low-birth-weight infants. Model-based insulin therapy protocols have the ability to safely improve glycaemic control for this group. Estimating non-insulin-mediated brain glucose uptake by the central nervous system in these models is typically done using population-based body weight models, which may not be ideal. METHOD A head circumference-based model that separately treats small-for-gestational-age (SGA) and appropriate-for-gestational-age (AGA) infants is compared to a body weight model in a retrospective analysis of 48 patients with a median birth weight of 750g and median gestational age of 25 weeks. Estimated brain mass, model-based insulin sensitivity (SI) profiles, and projected glycaemic control outcomes are investigated. SGA infants (5) are also analyzed as a separate cohort. RESULTS Across the entire cohort, estimated brain mass deviated by a median 10% between models, with a per-patient median difference in SI of 3.5%. For the SGA group, brain mass deviation was 42%, and per-patient SI deviation 13.7%. In virtual trials, 87-93% of recommended insulin rates were equal or slightly reduced (Δ<0.16mU/h) under the head circumference method, while glycaemic control outcomes showed little change. CONCLUSION The results suggest that body weight methods are not as accurate as head circumference methods. Head circumference-based estimates may offer improved modelling accuracy and a small reduction in insulin administration, particularly for SGA infants.