A Pharmacokinetic/Pharmacodynamic Comparison of SAAM II and PC/WinNonlin Modeling Software

doi:10.1021/js9603562

Journal of Pharmaceutical Sciences

Volume 87, Issue 10, October 1998, Pages 1255-1263

https://doi.org/10.1021/js9603562 Get rights and content

Abstract

This paper presents a detailed comparison of the kinetic analysis software packages SAAM II and PCNonlin/WinNonlin, based on benchmark modeling problems reported in “Pharmacokinetic and Pharmacodynamic Data Analysis: Concepts and Applications” (Gab-rielsson and Weiner, 1994) and seven additional models. For each model, both software packages were presented with identical implementations. Models were initially executed in PCNonlin or WinNonlin and automated comparisons with SAAM II made using Microsoft Test. Models investigated included one- and multicompart-ment models with nonlinearities, multiple inputs and samples, multiple simultaneous experiments, and linear equations. Maximum number of compartments, data sets, and parameters were 9, 5, and 10, respectively. We compared 88 different models, many of them in different configurations, e.g., different weighting schemes or different parameter limits. The total number of attempted comparisons between SAAM II and PCNonlin was 161, of which 142 executed without problems. Parameter estimates, their precision (standard errors), and model predictions were compared; a difference of 1% or less was considered “agreement”. Observed differences, mainly in parameter standard errors, can be accounted for in terms of different optimization algorithms, convergence criteria, and individual capabilities. In general, there was good agreement (<1% difference) between SAAM II and PCNonlin in terms of parameter estimates and model predictions. However, due to differences in the optimization procedure, parameter standard errors showed considerable differences. Additionally, there were differences when multiple data sets were fitted, indicating the importance of different fitting procedures for interpreting multiple kinetic data sets. The full results of the comparison and the model files in SAAM II and PCNonlin/WinNonlin formats are available from the authors.

References and Notes (12)

B.M. Bell et al.
Comput. Stat. Data Anal.
(1996)
L.F. Shampine
Comput. Mathematics Applications
(1986)
(1995)
P. Tanswell et al.
Int. J. Clin. Pharmacol. Ther. Toxicol.
(1993)
PCNonlin 4.2
Scientific Consulting Inc: Cary
(1993)
Mathematica
Wolfram Research Champaign
(1995)

There are more references available in the full text version of this article.

Cited by (34)

CPhaMAS: An online platform for pharmacokinetic data analysis based on optimized parameter fitting algorithm
2024, Computer Methods and Programs in Biomedicine
Clinical pharmacological modeling and statistical analysis software is an essential basic tool for drug development and personalized drug therapy. The learning curve of current basic tools is steep and unfriendly to beginners. The curve is even more challenging in cases of significant individual differences or measurement errors in data, resulting in difficulties in accurately estimating pharmacokinetic parameters by existing fitting algorithms. Hence, this study aims to explore a new optimized parameter fitting algorithm that reduces the sensitivity of the model to initial values and integrate it into the CPhaMAS platform, a user-friendly online application for pharmacokinetic data analysis.
In this study, we proposed an optimized Nelder-Mead method that reinitializes simplex vertices when trapped in local solutions and integrated it into the CPhaMAS platform. The CPhaMAS, an online platform for pharmacokinetic data analysis, includes three modules: compartment model analysis, non-compartment analysis (NCA) and bioequivalence/bioavailability (BE/BA) analysis. Our proposed CPhaMAS platform was evaluated and compared with existing WinNonlin.
The platform was easy to learn and did not require code programming. The accuracy investigation found that the optimized Nelder-Mead method of the CPhaMAS platform showed better accuracy (smaller mean relative error and higher R²) in two-compartment and extravascular administration models when the initial value was set to true and abnormal values (10 times larger or smaller than the true value) compared with the WinNonlin. The mean relative error of the NCA calculation parameters of CPhaMAS and WinNonlin was <0.0001 %. When calculating BE for conventional, high-variability and narrow-therapeutic drugs. The main statistical parameters of the parameters C_max, AUC_t, and AUC_inf in CPhaMAS have a mean relative error of <0.01% compared to WinNonLin.
In summary, CPhaMAS is a user-friendly platform with relatively accurate algorithms. It is a powerful tool for analysing pharmacokinetic data for new drug development and precision medicine.
Small interfering RNA delivery to the neurons near the amyloid plaques for improved treatment of Alzheimer׳s disease
2019, Acta Pharmaceutica Sinica B
Citation Excerpt :
An average concentration of triplicate was used for data analysis. Pharmacokinetic parameters were calculated by non-compartment analysis using the WinNonlin software (Mountain View, CA, USA)22. APP/PS1 double transgenic mice (7–8 months old) were randomly divided into 9 groups (n = 9): the AD control group (AD), the CQ/NC-siRNA group (NC, 80 µg siRNA/mouse), the M/BACE1-siRNA group (M, 80 µg siRNA/mouse), the Q/BACE1-siRNA group (Q, 80 µg siRNA/mouse), the C/BACE1-siRNA middle-dose (CM, 50 µg siRNA/mouse) and high-dose (CH, 80 µg siRNA/mouse) group, the CQ/BACE1-siRNA low-dose (CQL, 30 µg siRNA/mouse), middle-dose (CQM, 50 µg siRNA/mouse) and high-dose (CQH, 80 µg siRNA/mouse) group.
Gene therapy represents a promising treatment for the Alzheimer׳s disease (AD). However, gene delivery specific to brain lesions through systemic administration remains big challenge. In our previous work, we have developed an siRNA nanocomplex able to be specifically delivered to the amyloid plaques through surface modification with both CGN peptide for the blood–brain barrier (BBB) penetration and QSH peptide for β-amyloid binding. But, whether the as-designed nanocomplex could indeed improve the gene accumulation in the impaired neuron cells and ameliorate AD-associated symptoms remains further study. Herein, we prepared the nanocomplexes with an siRNA against β-site amyloid precursor protein-cleaving enzyme 1 (BACE1), the rate-limiting enzyme of Aβ production, as the therapeutic siRNA of AD. The nanocomplexes exhibited high distribution in the Aβ deposits-enriched hippocampus, especially in the neurons near the amyloid plaques after intravenous administration. In APP/PS1 transgenic mice, the nanocomplexes down-regulated BACE1 in both mRNA and protein levels, as well as Aβ and amyloid plaques to the level of wild-type mice. Moreover, the nanocomplexes significantly increased the level of synaptophysin and rescued memory loss of the AD transgenic mice without hematological or histological toxicity. Taken together, this work presented direct evidences that the design of precise gene delivery to the AD lesions markedly improves the therapeutic outcome.
MATLAB-implemented estimation procedure for model-based assessment of hepatic insulin degradation from standard intravenous glucose tolerance test data
2013, Computer Methods and Programs in Biomedicine
Citation Excerpt :
The starting algorithm is the GNA because it is simpler and faster; and the more robust LMA takes over in case of GNA failure (Fig. 6). Thus, the alternating action of GMA and LMA allowed to find a solution for our optimization problem even when it started very far off the final minimum (no procedure failure, in our attempts on eighteen subjects), and to reduce computational time to value similar to those obtained in SAAM II, whose optimizer is based on the Gauss–Newton method alone [14]. Also for the solution of the initial value problem, a specific choice has been done, indicating the ode15s as the most suitable MATLAB solver function for the given problem.
Present study provides a novel MATLAB-based parameter estimation procedure for individual assessment of hepatic insulin degradation (HID) process from standard frequently-sampled intravenous glucose tolerance test (FSIGTT) data. Direct access to the source code, offered by MATLAB, enabled us to design an optimization procedure based on the alternating use of Gauss-Newton's and Levenberg-Marquardt's algorithms, which assures the full convergence of the process and the containment of computational time. Reliability was tested by direct comparison with the application, in eighteen non-diabetic subjects, of well-known kinetic analysis software package SAAM II, and by application on different data. Agreement between MATLAB and SAAM II was warranted by intraclass correlation coefficients ≥0.73; no significant differences between corresponding mean parameter estimates and prediction of HID rate; and consistent residual analysis. Moreover, MATLAB optimization procedure resulted in a significant 51% reduction of CV% for the worst-estimated parameter by SAAM II and in maintaining all model-parameter CV% <20%. In conclusion, our MATLAB-based procedure was suggested as a suitable tool for the individual assessment of HID process.
Facilitating pharmacometric workflow with the metrumrg package for R
2013, Computer Methods and Programs in Biomedicine
Citation Excerpt :
Options for data handling and visualization include R, SPLUS, SYSTAT, SPSS, MATLAB, SAS, etc. Options for modeling and simulation include mixed-effects tools such as NONMEM [1], NLME [2], WinNonLin [3], Monolix [4], PDx-MCPEM [5], S-ADAPT [6] and ADAPT [7] or Bayesian tools such as WinBUGS [8]. Reports are typically authored using Latex tools or Microsoft Word®.
metrumrg is an R package that facilitates workflow for the discipline of pharmacometrics. Support is provided for data preparation, modeling, simulation, diagnostics, and reporting. Existing tools and techniques are emphasized where available; original solutions are provided for otherwise unmet needs. In particular, metrumrg implements an R interface for the NONMEM^® modeling software, optionally run in a distributed computing environment. The paradigm allows start-to-finish analyses in a single scripting language. Emphasis on text-based formats promotes traceability of results.
Multi-compartmental modelling and experimental designs for glucose transport study in insulin-resistant rat
2002, Comptes Rendus - Biologies
Many pathologies are associated with abnormalities of glucose metabolism or with perturbations of its transport (type 2 diabetes or insulin-resistance). The pre-diabetic state is characterised by a state of insulin-resistance, in others words a defect of glucose transport in insulin-sensible tissues, such as muscles and adipose tissues. The mathematical modelling of experimental data can be an excellent method to explore the mechanisms implied in the studied biological phenomenon. Thus, starting from a symbolic formulation like the compartmental modelling, it can be possible to develop a theoretical basis for the observation and to consider the best-adapted experiments for the study. We showed with mathematical models that 〚¹²³I〛-6-deoxy-6-iodo-d-glucose (6-DIG), shown as a tracer of glucose transport in vitro, could point out this transport abnormality. To quantify the insulin resistance, we estimated the fractional transfer coefficients of 6-DIG from the blood to the organs. We realised many studies to lead to a satisfying model; special attention has been paid to the precision of the parameter to select the best model. The results showed that by associating experimental data obtained with 6-DIG activities and an adapted mathematical model, discriminating parameters (in and out fractional transfer coefficients) between the two groups (control and insulin-resistant rats) could be pointed out.
De nombreuses pathologies sont associées aux anomalies du métabolisme du glucose ou aux perturbations de son transport (diabète de type 2 ou insulino-résistance). L’état pré-diabétique est caractérisé par un état d’insulino-résistance ou, en d’autres termes, par un défaut de transport de glucose vers les tissus insulino-sensibles, tels que les muscles et le tissu adipeux. La modélisation mathématique des données expérimentales peut être une excellente méthode pour étudier les mécanismes impliqués dans le phénomène biologique concerné. Ainsi, à partir d’une représentation symbolique comme la modélisation compartimentale, il est possible de développer une base théorique pour l’observation et de considérer les expériences les mieux adaptées pour l’étude. Nous avons montré, avec des modèles compartimentaux, que le 〚¹²³I〛-6-déoxy-6-iodo-d-glucose (6-DIG), validé comme traceur du transport du glucose in vitro, peut permettre la vérification de cette anomalie de transport. Pour mesurer cette insulino-résistance, nous avons estimé les coefficients de transfert fractionnel du 6-DIG du sang vers les différents organes de l’étude. Nous avons réalisé plusieurs études avant d’arriver à un modèle satisfaisant et une attention toute particulière a été portée à la précision des paramètres estimés, afin de choisir le meilleur modèle. Les résultats ont montré qu’en associant les données expérimentales obtenues avec les activités du 6-DIG et un modèle mathématique adapté, des paramètres discriminants (coefficients de transfert fractionnel) entre les deux groupes, rats témoins et insulinorésistants, peuvent être dégagés.
An evaluation of extended vs weighted least squares for parameter estimation in physiological modeling
2001, Journal of Biomedical Informatics
Weighted least squares (WLS) is the technique of choice for parameter estimation from noisy data in physiological modeling. WLS can be derived from maximum likelihood theory, provided that the measurement error variance is known and independent of the model parameters and the weights are calculated as the inverse of the measurement error variance. However, using measured values in lieu of predicted values to quantify the measurement error variance is approximately valid only when the noise in the data is relatively low. This practice may thus introduce sampling variation in the resulting estimates, as weights can be seriously misspecified. To avoid this, extended least squares (ELS) has been used, especially in pharmacokinetics. ELS uses an augmented objective function where the measurement error variance depends explicitly on the model parameters. Although it is more complex, ELS accounts for the Gaussian maximum likelihood statistical model of the data better than WLS, yet its usage is not as widespread. The use of ELS in high data noise situations will result in more accurate parameter estimates than WLS (when the underlying model is correct). To support this claim, we have undertaken a simulation study using four different models with varying amounts of noise in the data and further assuming that the measurement error standard deviation is proportional to the model prediction. We also motivate this in terms of maximum likelihood and comment on the practical consequences of using WLS and ELS as well as give practical guidelines for choosing one method over the other.

View all citing articles on Scopus

View full text

Research ArticlesA Pharmacokinetic/Pharmacodynamic Comparison of SAAM II and PC/WinNonlin Modeling Software

Abstract

Comput. Stat. Data Anal.

Comput. Mathematics Applications

Int. J. Clin. Pharmacol. Ther. Toxicol.

Scientific Consulting Inc: Cary

Wolfram Research Champaign

Research Articles
A Pharmacokinetic/Pharmacodynamic Comparison of SAAM II and PC/WinNonlin Modeling Software