Incidence and risk factors of major bleeding following major orthopaedic surgery with fondaparinux thromboprophylaxis. A time-to-event analysis.

AIMS: Increased exposure to fondaparinux, as observed in patients with renal impairment, may increase bleeding risk. This study aims to determine the time course of major bleeding after major orthopaedic surgery, identify predictors of bleeding and simulate the effect of a reduced dose of fondaparinux on bleeding for patients with moderate renal impairment (creatinine clearance = 20-50 ml min-1 ). METHODS: Data including fondaparinux anti-Xa activities from two multicentre prospective cohorts were used. In the first cohort, patients (n = 957) received fondaparinux 2.5 mg once a day. In the second, patients with moderate renal impairment (n = 436) received 1.5 mg once per day. The time-to-major bleeding after the end of surgery was modelled using a parametric survival analysis in NONMEM. RESULTS: The observed rate of major bleeding up to day 11 was 5.2%. The time-to-event analysis indicated that the hazard of bleeding was highest in the first days following surgery and then remained low thereafter. Independent significant predictors of an increased hazard of major bleeding were male sex, lower body weight and increased drug exposure. Simulated rates of major bleeding up to day 11 in patients with moderate renal impairment were 6.5% with fondaparinux 2.5 mg once daily and 3.8% with fondaparinux 1.5 mg once daily. CONCLUSION: The hazard of major bleeding is highest in the first postoperative days and increases with fondaparinux exposure. To reduce the risk of bleeding in patients with moderate renal impairment, this study supports the use of a lower dose of fondaparinux 1.5 mg once daily.

Warfarin Dosing Algorithms Underpredict Dose Requirements in Patients Requiring ≥7 mg Daily: A Systematic Review and Meta-analysis.

There is preliminary evidence to suggest that some published warfarin dosing algorithms produce biased maintenance dose predictions in patients who require higher than average doses. We conducted a meta-analysis of warfarin dosing algorithms to determine if there exists a systematic under- or overprediction of dose requirements for patients requiring ≥7 mg/day across published algorithms. Medline and Embase databases were searched up to September 2015. We quantified the proportion of over- and underpredicted doses in patients whose observed maintenance dose was ≥7 mg/day. The meta-analysis included 47 evaluations of 22 different warfarin dosing algorithms from 16 studies. The meta-analysis included data from 1,492 patients who required warfarin doses of ≥7 mg/day. All 22 algorithms were found to underpredict warfarin dosing requirements in patients who required ≥7 mg/day by an average of 2.3 mg/day with a pooled estimate of underpredicted doses of 92.3% (95% confidence interval 90.3-94.1, I2 = 24%).

A Philosophical Framework for Integrating Systems Pharmacology Models Into Pharmacometrics.

The framework for systems pharmacology style models does not naturally sit with the usual scientific dogma of parsimony and falsifiability based on deductive reasoning. This does not invalidate the importance or need for overarching models based on pharmacology to describe and understand complicated biological systems. However, it does require some consideration on how systems pharmacology fits into the overall scientific approach.

A framework for quantifying the influence of adherence and dose individualization.

A failure to accommodate for a patient's imperfect adherence may result in therapeutic failure. Similarly, failure to accommodate a patient's individual needs via dose individualization may also result in poor patient outcomes. The property of a drug that signifies the likelihood of therapeutic success to imperfect adherence is termed "forgiveness." We introduce an extension to this concept as: (1) a priori forgiveness (forgiveness when dose individualization is not considered) and (2) a posteriori forgiveness (forgiveness when considering dose individualization). We illustrate cases when adherence is of primary importance and in which dose individualization is of primary importance. The concept of a priori forgiveness and a posteriori forgiveness provides a quantitative measure that allows the influence of adherence to be disentangled from dose individualization and could be used to provide clear guidelines about the relative importance of each in clinical practice.

Quantification of the Forgiveness of Drugs to Imperfect Adherence.

The circumstance of how sensitive therapeutic success is under imperfect adherence is driven by the property known as forgiveness. To date, no studies have considered variability in the pharmacokinetic-pharmacodynamic process in conjunction with imperfect adherence patterns in order to develop a comparative criterion to determine the forgiveness of a drug. In this study, we have proposed a criterion to quantify forgiveness; illustrated the criterion for a theoretical example and evaluated the forgiveness of a motivating example, namely warfarin. A forgiveness criterion, relative forgiveness, is defined as the number of times more likely that a target is successfully attained under perfect adherence compared to imperfect adherence; or when comparing two drugs under a standard setting of imperfect adherence. The relative forgiveness criterion may have important implications for both drug development and clinical practice since the choice of drug can account for the likely influence of its forgiveness.

A model-based meta-analysis of the influence of factors that impact adherence to medications.

WHAT IS KNOWN AND OBJECTIVE: Several studies have investigated factors that may influence adherence for a given disease. The influence of disease on adherence has received limited attention. Less work has been conducted to investigate the influence of other factors in conjunction with disease on adherence. The aim of this study was to determine the independent influence of disease and other factors on adherence. METHODS: A literature search was conducted to retrieve adherence studies using medication event monitoring system devices. Studies were categorized into different therapeutic areas. Only the two most commonly studied therapeutic areas were selected. Pseudopatient-level data were extracted from each study. The extracted data were analysed using a model-based meta-analysis technique. Univariate and multivariate models were developed. Model selection was based on a likelihood ratio test and visual plots. RESULTS: The most commonly studied therapeutic areas were HIV and hypertension. The most commonly recorded adherence criterion was percentage of prescribed doses taken per day. Based on this adherence criterion, ultimately, 24 HIV papers and 12 hypertension papers were included for data extraction. The statistically significant factors were disease, age and dosing regimen. The independent influences of each factor on adherence were as follows: an increase in adherence of approximately 8% per 10-year increase of age, a 15-19% reduction from once to thrice daily dosing and that patients with HIV were 5% more adherent than those with hypertension. WHAT IS NEW AND CONCLUSION: Although the influence of disease on adherence was significant, it was of limited clinical significance in the diseases studied here. Adherence appears to improve with age and decline with more frequent dosing. Additionally, the influence of dosing regimen wanes with increasing age. These results should be treated as exploratory and require prospective assessment.

Scale reduction of a systems coagulation model with an application to modeling pharmacokinetic-pharmacodynamic data.

Bridging systems biology and pharmacokinetics-pharmacodynamics has resulted in models that are highly complex and complicated. They usually contain large numbers of states and parameters and describe multiple input-output relationships. Based on any given data set relating to a specific input-output process, it is possible that some states of the system are either less important or have no influence at all. In this study, we explore a simplification of a systems pharmacology model of the coagulation network for use in describing the time course of fibrinogen recovery after a brown snake bite. The technique of proper lumping is used to simplify the 62-state systems model to a 5-state model that describes the brown snake venom-fibrinogen relationship while maintaining an appropriate mechanistic relationship. The simplified 5-state model explains the observed decline and recovery in fibrinogen concentrations well. The techniques used in this study can be applied to other multiscale models.

Predicting the requirement for N-acetylcysteine in paracetamol poisoning from reported dose.

CONTEXT: There is contention over whether reported dose correlates with toxicity in paracetamol poisoning and risk assessment is currently based on serum paracetamol concentration compared to a nomogram, irrespective of reported dose. Objective. To determine if reported dose predicts the need for N-acetylcysteine (NAC). METHODS: Data were taken from paracetamol overdoses presenting to a tertiary toxicology service. Age, sex, reported dose, ingestion time, timed paracetamol concentrations between 4 and 16 h, hepatotoxicity (peak alanine transaminase > 1000 U/L) and treatment (single dose-activated charcoal [SDAC] and NAC) were analysed. Data were analysed within a repeated measures logistic regression framework using NONMEM (ver 7.2). The primary outcome was administration of NAC, which was determined based on a serum paracetamol concentration greater than the nomogram line. RESULT: There were 1571 admissions in 1303 patients, with a median age of 27 years (12-96 years) and 1140 (73%) were females. The median dose was 10 g (1-100 g). The paracetamol concentration was above the nomogram line in 337 of 1571 (22%) patients. Patients presenting later (first paracetamol concentration between 7 and 16 h post-overdose) compared to those presenting earlier (4-7 h post-overdose) were more likely to have hepatotoxicity (5.5% vs. 0.4%; p < 0.0001), have a toxic paracetamol concentration (34% vs. 18%; p < 0.0001) and receive NAC (48% vs. 23%; p < 0.0001). SDAC reduced the probability of the paracetamol concentration being above the nomogram. Based on SDAC not being administered there was a 5% probability of requiring NAC at a dose of 6-9 g, a 10% chance of requiring NAC at a dose of 13-16 g, a 50% chance of requiring NAC at a dose of 30-34 g and a 90% chance for needing NAC at 48-50 g. CONCLUSION: Reported dose was a good predictor of a toxic paracetamol concentration and SDAC reduced the probability of the concentration being above the nomogram. These predictions may assist in determining which patients could be started on NAC immediately.

An approach for identifiability of population pharmacokinetic-pharmacodynamic models.

Mathematical models are routinely used in clinical pharmacology to study the pharmacokinetic and pharmacodynamic properties of a drug in the body. Identifiability of these models is an important requirement for the success of these clinical studies. Identifiability is classified into two types, structural identifiability related to the structure of the mathematical model and deterministic identifiability which is related to the study design. There are existing approaches for assessment of structural identifiability of fixed-effects models, although their use appears uncommon in the literature. In this study, we develop an informal unified approach for simultaneous assessment of structural and deterministic identifiability for fixed and mixed-effects pharmacokinetic or pharmacokinetic-pharmacodynamic models. This approach uses an information theoretic framework. The method is applied both to simple examples to explore known identifiability properties and to a more complex example to illustrate its utility.CPT: Pharmacometrics & Systems Pharmacology (2013) 2, e49; doi:10.1038/psp.2013.25; advance online publication 19 June 2013.

A learning-based approach for performing an in-depth literature search using MEDLINE.

WHAT IS KNOWN AND OBJECTIVE: Exhaustive literature searching is a core requirement for developing guidelines for evidence-based practice. MEDLINE is typically used. Searching requires the user to identify appropriate search terms, called Medical Subject Headings (MeSH) and refine the search to retrieve relevant articles. The objective of this study was to develop and test a learning algorithm for conducting a thorough literature search. METHODS: A learning algorithm to effectively utilize MeSH terms is presented. This algorithm creates combinations of available MeSH terms from which a search is conducted. The algorithm was applied to search MEDLINE (January 1950 to Janaury 2008) focusing on the impact of pharmaceutical care in HIV-infected patients. The number of relevant articles retrieved from the learning algorithm search was then compared against a static search with a fixed set of keywords implemented by an independent user. RESULTS AND DISCUSSION: The learning algorithm retrieved 1670 articles with six relevant articles identified. The static search retrieved a total of 49 articles, with three being relevant. These three articles were also located from the learning algorithm-based search. WHAT IS KNOWN AND CONCLUSION: Performing a literature search for retrieving evidence-based studies can be a daunting and error-prone process. The introduction of automatic, learning tools for searching is desirable and we present a possible approach.

Prediction of torsade de pointes from the QT interval: analysis of a case series of amisulpride overdoses.

Electrocardiograms (ECGs) from a case series of 86 amisulpride overdose events in 66 patients were reviewed for abnormal QT intervals and torsade de pointes (TdP). Eight patients exhibited TdP. In this investigative case series, the magnitude of prolongation of the QT interval was a stronger predictor of TdP than the mere presence of a prolongation per se.

Failure of antivenom to improve recovery in Australian snakebite coagulopathy.

BACKGROUND: Venom-induced consumption coagulopathy (VICC) is an important feature of snake envenoming. AIM: To investigate the effect of antivenom and fresh frozen plasma (FFP) on recovery of VICC in Australian elapid snake envenoming. DESIGN: Prospective cohort study. METHODS: Patients with VICC were included from the Australian Snakebite Project (ASP). Time to recovery of VICC (defined as time until INR <2) was investigated using a time to event analysis in WinBUGS. The model considered the effects of age, sex, snake type, time of antivenom after bite, antivenom dose and use of FFP within 4 h. RESULTS: The study included 167 cases of VICC, median age being 41 [interquartile range (IQR): 28-53) years, and 130 (78%) were males. Antivenom was administered at a median of 3.6 (IQR: 2.2-5.6) h after the bite at a median dose of four vials (IQR: 2-6 vials). Thirteen patients received FFP within 4 h. Recovery of VICC occurred after a median of 14.4 (IQR: 11.5-17.5) h, and only the use of FFP within 4 h influenced the time to recovery. Neither antivenom dose nor time of antivenom administration had an effect on recovery of VICC. In patients administered with FFP, 12% [credible interval (CrI): 6-21%] and 81% (CrI: 61-94%) had recovered at 6 and 12 h, respectively, vs 2.5% (CrI: 1.5-4%) and 28% (CrI: 22-34%) not receiving FFP. DISCUSSION: Antivenom did not appear to be effective for the coagulopathy in snake envenoming in Australia. FFP appeared to shorten the time of VICC recovery.

The effect of decontamination procedures on the pharmacokinetics of venlafaxine in overdose.

The aim of this work was to investigate the pharmacokinetics (PK) of venlafaxine in overdose and the effects of single-dose activated charcoal (SDAC) and whole-bowel irrigation (WBI), alone or in combination, as methods of decontamination. The data included 339 concentration-time points from 76 venlafaxine overdose events (median dose 2,625 (150-13,500 mg)); 69 were slow-release doses. SDAC, WBI, a combination of both, or no decontamination were administered to patients as decided by the treating clinician. The data were modeled using WinBUGS (Windows Bayesian Inference Using Gibbs Sampling). A one-compartment model with first-order input and elimination provided an adequate description of the data. SDAC increased clearance (CL) of venlafaxine by 35%, and SDAC and WBI combined reduced the fraction absorbed by 29%. However, the latter produced a greater reduction in maximum plasma concentration (C(max)) for a similar drop in area under the plasma concentration-time curve (AUC). Both SDAC alone, and a combination of SDAC and WBI, decreased the AUC after venlafaxine overdose, but the combination may be more beneficial because it reduces peak concentrations to a greater extent.

A comprehensive model for the humoral coagulation network in humans.

Coagulation is an important process in hemostasis and comprises a complicated interaction of multiple enzymes and proteins. We have developed a mechanistic quantitative model of the coagulation network. The model accurately describes the time courses of coagulation factors following in vivo activation as well as in vitro blood coagulation tests of prothrombin time (PT, often reported as international normalized ratio (INR)) and activated partial thromboplastin time (aPTT). The model predicts the concentration-time and time-effect profiles of warfarin, heparins, and vitamin K in humans. The model can be applied to predict the time courses of coagulation kinetics in clinical situations (e.g., hemophilia) and for biomarker identification during drug development. The model developed in this study is the first quantitative description of the comprehensive coagulation network.

A population pharmacokinetic and pharmacodynamic evaluation of pralatrexate in patients with relapsed or refractory non-Hodgkin's or Hodgkin's lymphoma.

In a pralatrexate phase I study, patients displayed a high incidence of mucositis of grades 3 and 4. Preliminary evaluations of the pharmacokinetics of the drug and its association with mucositis suggested that pralatrexate exposure (area under the concentration-time curve (AUC)) could be controlled with body size (e.g., weight or body surface area)-based dosing and that pretreatment with folic acid and vitamin B(12) might diminish the incidence and severity of mucositis. The study was amended, with revised dosing and vitamin B(12) administration. Data from 47 patients were evaluated using NONMEM. Weight and methylmalonic acid (MMA) level were predictive of pharmacokinetic (PK) variability. AUC and MMA level were positively correlated with the risk of developing mucositis. A lower AUC schedule with vitamin B(12) pretreatment may control mucositis without compromising efficacy. The covariates identified in this study are comparable with other antifolate analogs. The application of modeling was a critical step in the development of pralatrexate, yielding important suggestions for dose, scheduling, and pretreatment modifications.

Simultaneous versus sequential optimal design for pharmacokinetic-pharmacodynamic models with FO and FOCE considerations.

We consider nested multiple response models which are used extensively in the area of pharmacometrics. Given the conditional nature of such models, differences in predicted responses are a consequence of different assumptions about how the models interact. As such, sequential versus simultaneous and First Order (FO) versus First Order Conditional Estimation (FOCE) techniques have been explored in the literature where it was found that the sequential and FO approaches can produce biased results. It is therefore of interest to determine any design consequences between the various methods and approximations. As optimal design for nonlinear mixed effects models is dependent upon initial parameter estimates and an approximation to the expected Fisher information matrix, it is necessary to incorporate any influence of nonlinearity (or parameter-effects curvature) into our exploration. Hence, sequential versus simultaneous design with FO and FOCE considerations are compared under low, typical and high degrees of nonlinearity. Additionally, predicted standard errors of parameters are also compared to empirical estimates formed via a simulation/estimation study in NONMEM. Initially, design theory for nested multiple response models is developed and approaches mentioned above are investigated by considering a pharmacokinetic-pharmacodynamic model found in the literature. We consider design for situations where all responses are continuous and extend this methodology to the case where a response may be a discrete random variable. In particular, for a binary response pharmacodynamic model, it is conjectured that such responses will offer little information about all parameters and hence a sequential optimization, in the form of product design optimality, may yield near optimal designs.

Optimal design criteria for discrimination and estimation in nonlinear models.

Nonlinear models are common in pharmacokinetics and pharmacodynamics. To date, most work in design in this area has concentrated on parameter estimation. Here, we introduce the idea of optimization of both estimation and model selection. However, experimental designs that provide powerful discrimination between a pair of competing model structures are rarely efficient in terms of estimating the parameters under each model. Conversely, designs which are efficient for parameter estimation may not provide suitable power to discriminate between the models. Several different methods of addressing both of these objectives simultaneously are introduced in this paper and are compared to an existing optimality criterion.

Promethazine overdose: clinical effects, predicting delirium and the effect of charcoal.

OBJECTIVE: The aim of this study was to describe the clinical effects of promethazine in overdose and explore the relationship between delirium and possible predictor variables. METHODS: A case series of promethazine poisonings was identified from a prospective database of poisoning admissions to a regional toxicology service. Data were extracted including demographics, details of ingestion, clinical features including delirium, complications and medical outcomes. In addition to descriptive statistics, a fully Bayesian approach using logistic regression was undertaken to investigate the relationship between predictor variables and delirium. RESULTS: There were 199 patients with 237 presentations, including 57 patients with 78 promethazine alone overdoses. Of these 57 patients who ingested promethazine alone the median age was 22 years [interquartile range (IQR): 17-31] and 42 were female (74%). The median dose ingested was 625 mg (IQR: 350-1250 mg). Median length of stay was 19 h (IQR: 13-27 h), ten were admitted to the intensive care unit (ICU) and four were ventilated. Delirium occurred in 33 patients (42%), tachycardia (HR>100) occurred on 44 occasions (56%) and hypotension only twice. There were no seizures, dysrhythmias or deaths. Multivariate analysis of 215 presentations (in 181 patients) where dose of promethazine ingested was known demonstrated that dose, administration of charcoal within 2 h and co-ingestants predicted whether patients developed delirium. No relationship was shown for sex and age. A plot of probability that a patient will develop delirium vs. dose was constructed which showed the probability of delirium for 250 mg was 31%, 500 mg was 42% and for 1 g was 55% for promethazine alone overdoses. CONCLUSION: The main feature of promethazine toxicity is delirium, the probability of which can be predicted from the dose ingested. The administration of charcoal and the presence of co-ingestants appears to reduce the probability of delirium in a predictable manner.

A model for venom-induced consumptive coagulopathy in snake bite.

Many snake venoms contain procoagulant toxins that activate the coagulation cascade and cause venom-induced consumptive coagulopathy (VICC). We developed a semi-mechanistic model of the clotting cascade in order to explore the effects of the procoagulant toxin from taipan venom on this system as well as the effects of antivenom. Simulations of the time course in the change of clotting factors were compared to data collected from taipan envenomed patients. The model accurately predicted the observed concentration of clotting factors over time following taipan envenomation. Investigations from the model indicated that the upper limit of the half-life of the procoagulant toxin was 1h. Simulations from the model also suggest that antivenom for Australasian elapids has negligible effect on reducing the recovery time of the coagulation profile unless administered almost immediately after envenomation. The model has generality to be expanded to describe the effects of other venoms and drugs on the clotting cascade.

Compound optimal design criteria for nonlinear models.

Three approaches for combining parameter estimation with opposing design criteria are proposed for nonlinear models. The first method discussed is the technique found in the literature and as such is the reference method for this paper. The compound crtierion is formed by maximizing a weighted product of efficiencies. The second criterion involves maximizing an opposing criterion while minimizing a defined loss function. The third method simultaneously maximizes both efficiencies with respect to parameter estimation and an opposing criterion with a multiple objective simulated annealing algorithm. The examples presented are based on a PK-model and a generalized linear model found in the literature.