Insight into the mechanism of action of ORG27569 at the cannabinoid type one receptor utilising a unified mathematical model.

Allosteric modulation of CB1 is therapeutically advantageous compared to orthosteric activation as it potentially offers reduced on-target adverse effects. ORG27569 is an allosteric modulator that increases orthosteric agonist binding to CB1 but decreases functional signalling. ORG27569 is characterised by a delay in disinhibition of agonist-induced cAMP inhibition (lag); however, the mechanism behind this kinetic lag is yet to be identified. We aimed to utilise a mathematical model to predict data and design in vitro experiments to elucidate mechanisms behind the unique signalling profile of ORG27569. The established kinetic ternary complex model includes the existence of a transitional state of CB1 bound to ORG27569 and CP55940 and was used to simulate kinetic cAMP data using NONMEM 7.4 and Matlab R2020b. These data were compared with empirical cAMP BRET data in HEK293 cells stably expressing hCB1. The pharmacometric model suggested that the kinetic lag in cAMP disinhibition by ORG27569 is caused by signal amplification in the cAMP assay and can be reduced by decreasing receptor number. This was confirmed experimentally, as reducing receptor number through agonist-induced internalisation resulted in a decreased kinetic lag by ORG27569. ORG27569 was found to have a similar interaction with CP55940 and the high efficacy agonist WIN55,212-2, and was suggested to have lower affinity for CB1 bound by the partial agonist THC compared to CP55940. Allosteric modulators have unique signalling profiles that are often difficult to interrogate exclusively in vitro. We have used a combined mathematical and in vitro approach to prove that ORG27569 causes a delay in disinhibition of agonist-induced cAMP inhibition due to large receptor reserve in this pathway. We also used the pharmacometric model to investigate the common phenomenon of probe dependence, to propose that ORG27569 binds with higher affinity to CB1 bound by high efficacy orthosteric agonists.

Effects of variations in atmospheric temperature and humidity on the estimation of exclusive breastfeeding status using the deuterium oxide dose-to-mother technique.

Background: The deuterium dose-to-mother (DTM) method measures the human milk intake of breastfed children. Recently, the use of this method has been expanded to classify babies objectively as exclusively breast fed (EBF) or not (non-EBF) based on quantification of non-milk oral water intake (NMOI). However, the calculation of NMOI estimates involves atmospheric temperature and humidity. Objective: To evaluate the effects of atmospheric temperature and humidity on NMOI calculation and the classification of exclusive breastfeeding. Methods: The effect of indoor temperature and relative humidity on NMOI and the estimated prevalence of non-EBF were examined in two existing data sets of DTM in children by varying temperature in the range of 15 to 35°C and relative humidity in the range of 20 to 80% representing the maximum span of indoor conditions expected. Population-level estimates of NMOI distributions were derived using the rstan package for R v2.21.2. Results: The NMOI decreased at a rate of -1.15 g/day per °C increase and at a rate of -1.01 g/day per percent increase in relative humidity; this was due to variations in non-oral water intake from the atmosphere, a component of the calculation of NMOI, which is dependent on temperature and humidity. For the various locations considered, the mean calculated NMOI varied between 24.6 and 53.3 g/day using the same input data. In the mixed-fed sample of babies, the prevalence of non-EBF based on the earlier defined NMOI cut-off of 86.6 g/day was reduced by 19% when relative humidity was increased by 60%. Conclusions: Atmospheric conditions are essential factors in the computation of NMOI, used in the objective classification of babies as exclusively breast fed or not, and should be considered when the DTM method is used to classify exclusive breastfeeding.

What do pharmacy users think of free pharmacy services? Investigating pharmacy users' perceptions, attitudes and willingness to pay for free healthcare from pharmacies.

Background: Evidence exists of pharmacists providing free or partially subsidised clinical services in order to meet patient healthcare needs. Little is known about how patients perceive the quality and importance of such unfunded services to their healthcare. Objectives: To explore pharmacy user perspectives about unfunded services such as their valuation, why they chose to access these services from the pharmacy as well as their willingness to pay should pharmacies need to start charging for the provision of such services owing to budgetary constraints. Methods: This study was nested in a larger nationwide study where 51 pharmacies were recruited across fourteen locations across New Zealand . Semi-structured interviews were conducted with patients who had accessed unfunded services in community pharmacies. Patients were followed up to identify their percieved health outcomes resulting from accessing the unfunded service. Results: A total of 253 patient interviews were conducted on-site across 51 pharmacies in New Zealand. Two main themes were identified pertaining to (1) patient-provider relationship and (2) Willingness to pay. A total of fifteen different considerations were found to influence pharmacy users' decisions to access health services from the pharmacy. It was found that 62.8% of patients were willing to pay for unfunded services and the majority paying NZD$10. Conclusion: Patients positively rate these services and largely deem them important for their healthcare. Willingness to pay for services were also variable between patients and were dependant on the type of service accessed.

A kinetic model for positive allosteric modulator (PAM)-antagonists for the type 1 cannabinoid (CB1 ) receptor.

BACKGROUND AND PURPOSE: The cannabinoid (CB1 ) receptor is among the most abundant G protein-coupled receptors in brain. Allosteric ligands bind to a different site on receptors than the orthosteric ligand can have effects that are unique to the allosteric ligand and modulate orthosteric ligand activity. We propose a unified mathematical model describing the interaction effects of the allosteric ligand Org27569 and the orthosteric agonist CP55940 on CB1 receptor. EXPERIMENTAL APPROACH: A ternary complex model was constructed, which incorporated kinetic properties to describe the time course of effects of Org27569 and CP55940 reported in the literature: (i) enhanced receptor binding of CP55940, (ii) reduced internalisation and (iii), time-dependent modulation of cAMP. Underlying mechanisms of time-dependent modulation by Org27569 were evaluated by simulation. KEY RESULTS: A hypothetical transitional state of CP55940-CB1 -Org27569, which can internalise but cannot inhibit cAMP, was shown to be necessary and was sufficient to describe the allosteric modulation by Org27569, prior to receptors adopting an inactive conformation. The model indicated that the formation of this transitional CP55940-CB1 -Org27569 state and final inactive CP55940-CB1 -Org27569 state contributes to the enhanced CP55940 binding. The inactive CP55940-CB1 -Org27569 cannot internalise or inhibit cAMP, leading to reduced internalisation and cessation of cAMP inhibition. CONCLUSIONS AND IMPLICATIONS: In conclusion, a kinetic mathematical model for CB1 receptor allosteric modulation was developed. However, a standard ternary complex model was not sufficient to capture the data and a hypothetical transitional state was required to describe the allosteric modulation properties of Org27569.

Predicting serotonin toxicity in serotonin reuptake inhibitor overdose.

AIMS: We aimed to investigate the frequency of serotonin toxicity following overdose of antidepressants that inhibit serotonin reuptake and the factors that influence the probability of serotonin toxicity occurring. METHODS: This was a retrospective cohort study of overdoses that included selective serotonin reuptake inhibitors (SSRIs) (70%) and serotonin norepinephrine reuptake inhibitors (SNRIs) (30%) admitted to a tertiary toxicology unit over 23 years. A multivariate mixed effects logistic regression model using NONMEM (7.2.0) was used to determine factors that influenced the probability of serotonin toxicity occurring. RESULTS: There were 1978 overdoses in 1520 patients; median age 33 y (range: 13-86 years) and 64% female. Median defined daily dose equivalent (DDD) was 15 (1-420). Co-ingestants were taken in 1678/1978 (85%) overdoses: 11 co-ingested the monoamine oxidase-A inhibitor (MAOI) moclobemide, 99 (5%) co-ingested olanzapine, 58 (3%) co-ingested risperidone and 417 co-ingested a benzodiazepine (21%). Serotonin toxicity occurred in 269 overdoses (13.6%). The probability of serotonin toxicity increased slightly per 10 DDD units dose [OR, 1.01; 95% confidence intervals (CIs): 0.93-1.10], increased for an SNRI vs. an SSRI [OR, 1.07; 95% CI: 0.99-1.15], and markedly increased with co-ingestion of moclobemide [OR, 33.12; 95% CI: 7.5-147]. The probability decreased per 10 y age [OR, 0.84; 95% CI: 0.74-0.95], and with co-ingestion of the serotonin 2 A receptor (5-HT2A) antagonists olanzapine [OR, 0.40; 95% CI: 0.17-0.94] or risperidone [OR, 0.13; 95% CI: 0.02-0.99]. The probability of serotonin toxicity was 12.5% at 1 DDD (therapeutic), 12.7% at 15 DDDs and 19% at 420 DDDs. In overdoses of the median dose of 15 DDDs, co-ingestion of moclobemide increased the probability to 83%, and co-ingestion of olanzapine or risperidone decreased it to 5.5% and 1.8%, respectively. CONCLUSIONS: Serotonin toxicity is common following SSRI/SNRI overdose. Although dose increases probability, this was only a small effect. Co-ingestion with olanzapine or risperidone reduced the risk 2-6-fold, and moclobemide increased the risk 5-fold.

Investigating the provision, nature and associated costs of unfunded pharmacy services: A nationwide study.

BACKGROUND: There is increasing evidence of pharmacists providing free or partially subsidised patient-focused services in order to meet healthcare needs. Limited information exists about the types of unfunded services and their value. OBJECTIVES: (1) Identify the types of unfunded services provided nationally in New Zealand (NZ) and (2) Determine the costs associated with service provision. METHODS: A continuous observation time-motion study was conducted across New Zealand to characterise the provision of unfunded pharmacy services and the labour costs associated with their provision. The time-motion study spanned one business day (between seven to eight hours) in each participating pharmacy. The primary investigator (YA) spent one business day in each participating pharmacy (n = 51) and recorded details about the patient-focused services that were provided. Details included the type of service provided, approximate duration of the service and out-of-pocket costs borne by the patient. RESULTS: A total of 660 observations of unfunded services were recorded across the 51 pharmacies where 360 observation hours were carried out. Twenty-three types of unfunded services were identified, where minor ailments accounted for over half of the total observations. Labour costs associated with service provision were variable. CONCLUSION: Pharmacies across New Zealand are providing patient-focused services for which no funding is being provided.

In Vitro Characterization of 6-Methyl-3-(2-nitro-1-(thiophen-2-yl)ethyl)-2-phenyl-1H-indole (ZCZ011) at the Type 1 Cannabinoid Receptor: Allosteric Agonist or Allosteric Modulator?

Orthosteric activation of CB1 is known to cause a plethora of adverse side effects in vivo. Allosteric modulation is an exciting therapeutic approach and is hoped to offer improved therapeutic potential and a reduced on-target side effect profile compared to orthosteric agonists. This study aimed to systematically characterize the in vitro activity of the positive allosteric modulator ZCZ011, explicitly considering its effects on receptor regulation. HEK293 cells expressing hCB1 receptors were used to characterize ZCZ011 alone and in combination with orthosteric agonists. Real-time BRET approaches were employed for G protein dissociation, cAMP signaling, and ß-arrestin translocation. Characterization also included ERK1/2 phosphorylation (PerkinElmer AlphaLISA) and receptor internalization. ZCZ011 is an allosteric agonist of CB1 in all pathways tested, with a similar signaling profile to that of the partial orthosteric agonist Δ9-tetrahydrocannabinol. ZCZ011 also showed limited positive allosteric modulation in increasing the potency and efficacy of THC-induced ERK1/2 phosphorylation, ß-arrestin translocation, and receptor internalization. However, no positive allosteric modulation was observed for ZCZ011 in combination with either CP55940 or AMB-FUBINACA, in G protein dissociation, nor cAMP inhibition. Our study suggests that ZCZ011 is an allosteric agonist, with effects that are often difficult to differentiate from those of orthosteric agonists. Together with its pronounced agonist activity, the limited extent of ZCZ011 positive allosteric modulation suggests that further investigation into the differences between allosteric and orthosteric agonism is required, especially in receptor regulation end points.

Exploring Inductive Linearization for simulation and estimation with an application to the Michaelis-Menten model.

Nonlinear ordinary differential equations (ODEs) are common in pharmacokinetic-pharmacodynamic systems. Although their exact solutions cannot generally be determined via algebraic methods, their rapid and accurate solutions are desirable. Thus, numerical methods have a critical role. Inductive Linearization was proposed as a method to solve systems of nonlinear ODEs. It is an iterative approach that converts a nonlinear ODE into a linear time-varying (LTV) ODE, for which a range of standard integration techniques can then be used to solve (e.g., eigenvalue decomposition [EVD]). This study explores the properties of Inductive Linearization when coupled with EVD for integration of the LTV ODE and illustrates how the efficiency of the method can be improved. Improvements were based on three approaches, (1) incorporation of a convergence criterion for the iterative linearization process (for simulation and estimation), (2) creating more efficient step sizes for EVD (for simulation and estimation), and (3) updating the initial conditions of the Inductive Linearization (for estimation). The performance of these improvements were evaluated using single subject stochastic simulation-estimation with an application to a simple pharmacokinetic model with Michaelis-Menten elimination. The reference comparison was a standard non-stiff Runge-Kutta method with variable step size (ode45, MATLAB). Each of the approaches improved the speed of the Inductive Linearization technique without diminishing accuracy which, in this simple case, was faster than ode45 with comparable accuracy in the parameter estimates. The methods described here can easily be implemented in standard software programme such as R or MATLAB. Further work is needed to explore this technique for estimation in a population approach setting.

Identifying a therapeutic target for vancomycin against staphylococci in young infants.

OBJECTIVES: To determine the therapeutic target of vancomycin in young infants with staphylococcal infections. METHODS: Retrospective data were collected for infants aged 0 to 90 days with CoNS or MRSA bacteraemia over a 4 year period at the Royal Children's Hospital Melbourne, Australia. Vancomycin broth microdilution MICs were determined. A published pharmacokinetic model was externally validated using the study dataset and a time-to-event (TTE) pharmacodynamic model developed to link the AUC of vancomycin with the event being the first negative blood culture. Simulations were performed to determine the trough vancomycin concentration that correlates with a 90% PTA of the target AUC24. RESULTS: Thirty infants, 28 with CoNS and 2 with MRSA bacteraemia, who had 165 vancomycin concentrations determined were included. The vancomycin broth microdilution MIC was determined for 24 CoNS and 1 MRSA isolate, both with a median MIC of 1 mg/L (CoNS range = 0.5-4.0). An AUC0-24 target of ≥300 mg/L·h or AUC24-48 of ≥424 mg/L·h. increased the chance of bacteriological cure by 7.8- and 7.3-fold, respectively. However, AUC0-24 performed best in the pharmacokinetic-pharmacodynamic model. This correlates with 24 to 48 h trough concentrations of >15-18 mg/L and >10-15 mg/L for 6- and 12-hourly dosing, respectively, and can be used to guide vancomycin therapy in this population. CONCLUSIONS: An AUC0-24 ≥300 mg/L·h or AUC24-48 ≥424 mg/L·h was associated with an increase in bacteriological cure in young infants with staphylococcal bloodstream infections.

A framework for simplification of quantitative systems pharmacology models in clinical pharmacology.

Quantitative systems pharmacology (QSP) is a relatively new discipline within modelling and simulation that has gained wide attention over the past few years. The application of QSP models spans drug-target identification and validation, through all drug development phases as well as clinical applications. Due to their detailed mechanistic nature, QSP models are capable of extrapolating knowledge to predict outcomes in scenarios that have not been tested experimentally, making them an important resource in experimental and clinical pharmacology. However, these models are complicated to work with due to their size and inherent complexity. This makes many applications of QSP models for simulation, parameter estimation and trial design computationally intractable. A number of techniques have been developed to simplify QSP models into smaller models that are more amenable to further analyses while retaining their accurate predictive capabilities. Different simplification techniques have different strengths and weaknesses and hence different utilities. Understanding the utilities of different methods is essential for selection of the best method for a particular situation. In this paper, we have created an overall framework for model simplification techniques that allows a natural categorisation of methods based on their utility. We provide a brief description of the concept underpinning the different methods and example applications. A summary of the utilities of methods is intended to provide a guide to modellers in their model endeavours to simplify these complicated models.

Population pharmacokinetics of immediate-release and modified-release paracetamol and its major metabolites in a supratherapeutic dosing study.

OBJECTIVES: Overdose with paracetamol modified-release (MR) formulation, a bilayer tablet containing 69% slow-release component, has been increasing since its introduction to the market. However, little evidence exists for the management of MR paracetamol overdose. We aimed to develop a population pharmacokinetic (PK) model for immediate-release (IR) and MR paracetamol and its major metabolism, and quantitatively understand the formulation difference in toxicity assessment based on the nomogram line. METHODS: Data from a cross-over study design in nine healthy volunteers administered a single supratherapeutic oral dose (80 mg/kg) of either IR and MR paracetamol were available from a published study. Plasma concentrations for paracetamol and its metabolites glucuronide (APAPG) and sulfate conjugate (APAPS) for both formulations were measured and analysed with population pharmacokinetic (PK) method using NONMEM. Toxicity in both formulations was assessed by comparing the simulated paracetamol concentrations under different paracetamol dose levels with the 150 mg/L nomograms. The difference in the assessment was compared between the two formulations. RESULTS: Paracetamol concentrations for the IR formulation were described with a two-compartment model with first-order input and a lag time. The delayed time-course of MR paracetamol concentrations was best captured by a parallel absorption model in which the slow-release component was a serial zero-order then the first-order process. The formation of APAPG was linear, while APAPS concentrations were best fitted by a Michaelis-Menten process. The relative bioavailability of MR paracetamol compared to IR (FMR/IR) was estimated as 0.81. The simulated probability of making different toxicity assessments based on nomogram line was increased with dose levels and was as high as 14.6% after 22 g IR or MR paracetamol ingested. CONCLUSIONS: A joint parent-metabolite model to describe time-course profiles of both IR and MR paracetamol and its metabolites APAPG and APAPS concentrations was developed. Simulations from the model showed that toxicity assessment based on the 150 mg/L nomograms is not suitable in MR paracetamol overdoses.

Investigating myotoxicity following Australian red-bellied black snake (Pseudechis porphyriacus) envenomation.

BACKGROUND: Myotoxicity is one of the common clinical manifestations of red-bellied black snake (Pseudechis porphyriacus) envenomation characterised by elevated creatine kinase (CK) concentrations of greater than 1000 U/L. This study aimed to investigate the occurrence of myotoxicity in patients following envenomation. METHODS/PRINCIPAL FINDINGS: Patient characteristics and serial blood samples (timed venom concentrations and CK concentrations, pre- and post- antivenom) from 114 patients (median age 41, 2-90y; 80 male) were extracted from the Australian Snakebite Project database. Patients were categorised into three groups based on peak CK concentrations [no myotoxicity (<1000 U/L), mild (1000-10,000 U/L) and severe (>10,000 U/L)]. The odds of (mild or severe) myotoxicity was lower in patients that received early antivenom (within 6 hours post-bite) compared to those that received late or no antivenom (odd ratio was 0.186; 95% confidence interval, 0.052-0.664). A population pharmacokinetic-pharmacodynamic (PKPD) model was developed to describe the relationship between the time course of venom (a mixture of toxins) and effect (elevated CK). In addition, a kinetic-pharmacodynamic (KPD) model was developed to describe the relationship between time course of a theoretical toxin and effect. Model development and parameter estimation was performed using NONMEM v7.3. No single set of parameter values from either the PKPD or KPD models were found that could accurately describe the time course of different levels of severity of myotoxicity. The predicted theoretical toxin half-life from the KPD model was 11 ± 3.9 hours compared to the half-life of venom of 5.3 ± 0.36 hours. This indicates that the putative causative toxin's concentration-time profile does not parallel that of venom. CONCLUSION: Early antivenom administration reduces the incidence of myotoxicity. The venom concentration profile does not appear to be the driver for myotoxicity following envenomation. Additional factors that affect the sensitivity of the patient to snake venom/toxins must be explored to understand the relationship with myotoxicity.

Are professional pharmacy services being offered for free in pharmacies? A feasibility study exploring the use of a time motion study in New Zealand.

BACKGROUND: Pharmacists report to be providing patient-focused clinical services for which they receive no remuneration. Limited literature exists about unfunded services leading to difficulties in ascertaining an appropriate study design for such research. OBJECTIVE: This study aims to assess the appropriateness of a proposed study design before launching a nationwide study to investigate the provision of unfunded patient care services. METHODS: A multi-methods approach was utilised consisting of (1) continuous time motion study in community pharmacies (2) semi structured patient interviews (3) patient follow up (4) semi structured interviews with pharmacy owners/managers. All observations of unfunded patient care services were recorded, numerically coded and descriptively analysed. Semi structured interviews were audio recorded and transcribed verbatim. A semantic thematic analysis was carried out. Appropriateness of study design was dictated by the ability to characterise services and obtain patient perceptions. RESULTS: Ten pharmacies took part in the feasibility study, across the city of Dunedin, New Zealand, representing a range of different practice settings and demographics. Ten patients were interviewed and six responded to follow up. Both pharmacy and patient recruitment proved challenging due to concerns around disruption to workflow and patient privacy. A continuous observation time motion study was found to be appropriate as it minimises disruption to workflow with no additional work required from the pharmacy teams. CONCLUSIONS: A continuous observation time motion study proved to be an appropriate method to investigate the provision of unfunded services on a national scale. The findings of the study suggest design changes such as length of observation time, increasing patient recruitment and additional patient questions to enhance the nationwide study.

Exploring group size for statistical analysis of real-time signalling experiments.

BACKGROUND AND PURPOSE: Classical pharmacological bioassays generally use observed effects from a concentration series, at a single equilibrium time point to construct a concentration-effect curve, representing one experiment. However, if the full kinetic profile of the effect data for each concentration was evaluated simultaneously, then the analysis would be more powerful. In this work, we explore if more precise parameters can be achieved by using the full kinetic method. EXPERIMENTAL APPROACH: We used a simulation estimation study to explore the influence of kinetic analysis on the precision of the Emax model parameter estimates (Emax and C50 ). We compared a full kinetic approach in which all effect versus time data from a theoretical real-time signalling experiment were analysed simultaneously with a 'reference' approach. The theoretical real-time signalling experiment was based on a previously published CB2 receptor-binding experiment. KEY RESULTS: The reference method with a group size (n) of 5 provided highly precise parameter estimates (coefficient of variation [CV] 3.4% for Emax and 0.72% for C50 ). A full kinetic method provided more precise estimates than the reference with equal or smaller group sizes. Note that group size 'n' here refers to the number of technical replicates rather than the number of biological replicates. CONCLUSION AND IMPLICATIONS: A full kinetic method can yield more precise parameter estimates than the equilibrium method. Such an approach may be more useful for researchers.

Does the intact nephron hypothesis provide a reasonable model for metformin dosing in chronic kidney disease?

This research explored the intact nephron hypothesis (INH) as a model for metformin dosing in patients with chronic kidney disease (CKD). The INH assumes that glomerular filtration rate (GFR) will account for all kidney drug handling even for drugs eliminated by tubular secretion like metformin. We conducted two studies: (1) a regression analysis to explore the relationship between metformin clearance and eGFR metrics, and (2) a joint population pharmacokinetic analysis to test the relationship between metformin renal clearance and gentamicin clearance. The relationship between metformin renal clearance and eGFR metrics and gentamicin clearance was found to be linear, suggesting that a proportional dose reduction based on GFR in patients with CKD is reasonable.

Population pharmacokinetics of Pseudechis porphyriacus (red-bellied black snake) venom in snakebite patients.

OBJECTIVES: Understanding the time course of venom exposure in snakebite patients is important for the optimisation of treatment including antivenom dose and timing. We aimed to investigate the pharmacokinetics of red-bellied black snake (RBBS; Pseudechis porphyriacus) venom in envenomed patients. METHODS: Timed venom concentration data were obtained from patients with RBBS envenomation recruited to the Australian Snakebite Project (ASP), including demographics and antivenom treatment. Venom concentrations were measured using an enzyme immunoassay. Data were modelled using NONMEM version 7.3. Uncertainty in venom "dose" was accounted for by arbitrarily fixing the average amount to 1 mg and incorporating between-subject variability on relative bioavailability. A scale parameter for venom clearance was implemented to account for the rapid venom clearance following antivenom dosing. A sensitivity analysis was performed to determine the magnitude of venom clearance amplification. RESULTS: There were 457 venom concentrations in 114 patients (median age 41, 2-90 y; 80 male). Antivenom was administered to 54 patients a median of 4.2 h post-bite (0.67 to 32 h). A one-compartment model with first-order absorption and elimination provided the best description of the data. The estimated clearance and volume of distribution were 5.21 L/h and 39.9 L, respectively. The calculated elimination half-life of P. porphyriacus venom from the final pharmacokinetic model was 5.35 ± 0.36 h. The variability in the relative dose of injected venom was 140%. Antivenom administration increased venom clearance by 40-fold. Ten patients showed evidence of a double peak in the absorption profile. CONCLUSION: The information on the exposure time of venom in the body following envenomation will help improve treatment and the timing of antivenom.

Reduction of quantitative systems pharmacology models using artificial neural networks.

Quantitative systems pharmacology models are often highly complex and not amenable to further simulation and/or estimation analyses. Model-order reduction can be used to derive a mechanistically sound yet simpler model of the desired input-output relationship. In this study, we explore the use of artificial neural networks for approximating an input-output relationship within highly dimensional systems models. We illustrate this approach using a model of blood coagulation. The model consists of two components linked together through a highly dimensional discontinuous interface, which creates a difficulty for model reduction techniques. The proposed approach enables the development of an efficient approximation to complex models with the desired level of accuracy. The technique is applicable to a wide variety of models and provides substantial speed boost for use of such models in simulation and control purposes.