Search

Introduction: Intensive care unit (ICU) patients exhibit erratic blood glucose (BG) fluctuations, including hypoglycemic and hyperglycemic episodes, and require exogenous insulin delivery to keep their BG in healthy ranges. Glycemic control via glycemic management (GM) is associated with reduced mortality and morbidity in the ICU, but GM increases the cognitive load on clinicians. The availability of robust, accurate, and actionable clinical decision support (CDS) tools reduces this burden and assists in the decision-making process to improve health outcomes. Clinicians currently follow GM protocol flow charts for patient intravenous insulin delivery rate computations.Methods: We present a mechanistic model-based control algorithm that estimates the optimal intravenous insulin rate to keep BG within a target range; the goal is to develop this approach for eventual use within CDS systems. In this control framework, we employed a stochastic model representing BG dynamics in the ICU setting and used the linear quadratic Gaussian control methodology to develop a controller.Results: We designed two experiments, one using virtual (simulated) patients and one using a real-world retrospective dataset. Using these, we evaluated the safety and efficacy of this model-based glycemic control methodology. The presented controller avoids hypoglycemia and hyperglycemia in virtual patients, maintaining BG levels in the target range more consistently than two existing GM protocols. Moreover, this methodology could theoretically prevent a large proportion of hypoglycemic and hyperglycemic events recorded in a real-world retrospective dataset.Discussion: The current version of the methodology shows potential usefulness in GM of ICU patients. However, it is limited to a subgroup of the ICU patient population, who are fed through and enteral tube and delivered intravenous insulin. After extending to a broader ICU patient population who can consume oral nutrition and are delivered subcutaneous insulin for GM, the methodology could be tested with pilot studies and clinical trials for eventual use as a CDS tool.

Background: The protocols and therapeutic guidance established for treating traumatic brain injury (TBI) in neurointensive care focus on managing cerebral blood flow (CBF) and brain tissue oxygenation based on pressure signals. The decision support process relies on assumed relationships between cerebral perfusion pressure (CPP) and blood flow, pressure-flow relationships (PFRs), and shares this framework of assumptions with mathematical intracranial hemodynamics models. These foundational assumptions are difficult to verify, and their violation can impact clinical decision-making and model validity.Methods: A hypothesis- and model-driven method for verifying and understanding the foundational intracranial hemodynamic PFRs is developed and applied to a novel multi-modality monitoring dataset.Results: Model analysis of joint observations of CPP and CBF validates the standard PFR when autoregulatory processes are impaired as well as unmodelable cases dominated by autoregulation. However, it also identifies a dynamical regime -or behavior pattern-where the PFR assumptions are wrong in a precise, data-inferable way due to negative CPP-CBF coordination over long timescales. This regime is of both clinical and research interest: its dynamics are modelable under modified assumptions while its causal direction and mechanistic pathway remain unclear.Conclusion: Motivated by the understanding of mathematical physiology, the validity of the standard PFR can be assessed a) directly by analyzing pressure reactivity and mean flow indices (PRx and Mx) or b) indirectly through the relationship between CBF and other clinical observables. This approach could potentially help to personalize TBI care by considering intracranial pressure and CPP in relation to other data, particularly CBF. The analysis suggests a threshold using clinical indices of autoregulation jointly generalizes independently set indicators to assess CA functionality. These results support the use of increasingly data-rich environments to develop more robust hybrid physiological-machine learning models.

Acute respiratory distress syndrome (ARDS) and acute lung injury have a diverse spectrum of causative factors including sepsis, aspiration of gastric contents, and near drowning. Clinical management of severe lung injury typically includes mechanical ventilation to maintain gas exchange which can lead to ventilator-induced lung injury (VILI). The cause of respiratory failure is acknowledged to affect the degree of lung inflammation, changes in lung structure, and the mechanical function of the injured lung. However, these differential effects of injury and the role of etiology in the structure-function relationship are not fully understood. To address this knowledge gap we caused lung injury with intratracheal hydrochloric acid (HCL) or endotoxin (LPS) 2 days prior to ventilation or with an injurious lavage (LAV) immediately prior to ventilation. These injury groups were then ventilated with high inspiratory pressures and positive end expiratory pressure (PEEP) = 0 cmH2O to cause VILI and model the clinical course of ARDS followed by supportive ventilation. The effects of injury were quantified using invasive lung function measurements recorded during PEEP ladders where the end-expiratory pressure was increased from 0 to 15 cm H2O and decreased back to 0 cmH2O in steps of 3 cmH2O. Design-based stereology was used to quantify the parenchymal structure of lungs air-inflated to 2, 5, and 10 cmH2O. Pro-inflammatory gene expression was measured with real-time quantitative polymerase chain reaction and alveolocapillary leak was estimated by measuring bronchoalveolar lavage protein content. The LAV group had small, stiff lungs that were recruitable at higher pressures, but did not demonstrate substantial inflammation. The LPS group showed septal swelling and high pro-inflammatory gene expression that was exacerbated by VILI. Despite widespread alveolar collapse, elastance in LPS was only modestly elevated above healthy mice (CTL) and there was no evidence of recruitability. The HCL group showed increased elastance and some recruitability, although to a lesser degree than LAV. Pro-inflammatory gene expression was elevated, but less than LPS, and the airspace dimensions were reduced. Taken together, those data highlight how different modes of injury, in combination with a 2nd hit of VILI, yield markedly different effects.

Type 2 diabetes mellitus is a complex and under-treated disorder closely intertwined with obesity. Adolescents with severe obesity and type 2 diabetes have a more aggressive disease compared to adults, with a rapid decline in pancreatic β cell function and increased incidence of comorbidities. Given the relative paucity of pharmacotherapies, bariatric surgery has become increasingly used as a therapeutic option. However, subsets of this population have sub-optimal outcomes with either inadequate weight loss or little improvement in disease. Predicting which patients will benefit from surgery is a difficult task and detailed physiological characteristics of patients who do not respond to treatment are generally unknown. Identifying physiological predictors of surgical response therefore has the potential to reveal both novel phenotypes of disease as well as therapeutic targets. We leverage data assimilation paired with mechanistic models of glucose metabolism to estimate pre-operative physiological states of bariatric surgery patients, thereby identifying latent phenotypes of impaired glucose metabolism. Specifically, maximal insulin secretion capacity, σ, and insulin sensitivity, SI, differentiate aberrations in glucose metabolism underlying an individual’s disease. Using multivariable logistic regression, we combine clinical data with data assimilation to predict post-operative glycemic outcomes at 12 months. Models using data assimilation sans insulin had comparable performance to models using oral glucose tolerance test glucose and insulin. Our best performing models used data assimilation and had an area under the receiver operating characteristic curve of 0.77 (95% confidence interval 0.7665, 0.7734) and mean average precision of 0.6258 (0.6206, 0.6311). We show that data assimilation extracts knowledge from mechanistic models of glucose metabolism to infer future glycemic states from limited clinical data. This method can provide a pathway to predict long-term, post-surgical glycemic states by estimating the contributions of insulin resistance and limitations of insulin secretion to pre-operative glucose metabolism.

The insulin secretion rate (ISR) contains information that can provide a personal, quantitative understanding of endocrine function. If the ISR can be reliably inferred from measurements, it could be used for understanding and clinically diagnosing problems with the glucose regulation system.Objective: This study aims to develop a model-based method for inferring a parametrization of the ISR and related physiological information among people with different glycemic conditions in a robust manner. The developed algorithm is applicable for both dense or sparsely sampled plasma glucose/insulin measurements, where sparseness is defined in terms of sampling time with respect to the fastest time scale of the dynamics.Methods: An algorithm for parametrizing and validating a functional form of the ISR for different compartmental models with unknown but estimable ISR function and absorption/decay rates describing the dynamics of insulin accumulation was developed. The method and modeling applies equally to c-peptide secretion rate (CSR) when c-peptide is measured. Accuracy of fit is reliant on reconstruction error of the measured trajectories, and when c-peptide is measured the relationship between CSR and ISR. The algorithm was applied to data from 17 subjects with normal glucose regulatory systems and 9 subjects with cystic fibrosis related diabetes (CFRD) in which glucose, insulin and c-peptide were measured in course of oral glucose tolerance tests (OGTT).Results: This model-based algorithm inferred parametrization of the ISR and CSR functional with relatively low reconstruction error for 12 of 17 control and 7 of 9 CFRD subjects. We demonstrate that when there are suspect measurements points, the validity of excluding them may be interrogated with this method.Significance: A new estimation method is available to infer the ISR and CSR functional profile along with plasma insulin and c-peptide absorption rates from sparse measurements of insulin, c-peptide, and plasma glucose concentrations. We propose a method to interrogate and exclude potentially erroneous OGTT measurement points based on reconstruction errors.

Motivated by a desire to understand pulmonary physiology, scientists have developed physiological lung models of varying complexity. However, pathophysiology and interactions between human lungs and ventilators, e.g., ventilator-induced lung injury (VILI), present challenges for modeling efforts. This is because the real-world pressure and volume signals may be too complex for simple models to capture, and while complex models tend not to be estimable with clinical data, limiting clinical utility. To address this gap, in this manuscript we developed a new damaged-informed lung ventilator (DILV) model. This approach relies on mathematizing ventilator pressure and volume waveforms, including lung physiology, mechanical ventilation, and their interaction. The model begins with nominal waveforms and adds limited, clinically relevant, hypothesis-driven features to the waveform corresponding to pulmonary pathophysiology, patient-ventilator interaction, and ventilator settings. The DILV model parameters uniquely and reliably recapitulate these features while having enough flexibility to reproduce commonly observed variability in clinical (human) and laboratory (mouse) waveform data. We evaluate the proof-in-principle capabilities of our modeling approach by estimating 399 breaths collected for differently damaged lungs for tightly controlled measurements in mice and uncontrolled human intensive care unit data in the absence and presence of ventilator dyssynchrony. The cumulative value of mean squares error for the DILV model is, on average, ≈12 times less than the single compartment lung model for all the waveforms considered. Moreover, changes in the estimated parameters correctly correlate with known measures of lung physiology, including lung compliance as a baseline evaluation. Our long-term goal is to use the DILV model for clinical monitoring and research studies by providing high fidelity estimates of lung state and sources of VILI with an end goal of improving management of VILI and acute respiratory distress syndrome.