Prediction mixture

Borgert, C.J., QuiU, T.F., McCarty, L.S., Mason, A.M. (2004) Can Mode of Action Predict Mixture Toxicity for Risk Assessment Toxicology and Applied Pharmacology, 210, 85-96. 

Mixed Micelles. The CMC values -for the two pure sur-factants and well de-fined mixtures thereo-f are shown in Figure 2. The experiments were run at a high added salt level (swamping electrolyte) so the counterion contributed by the dissolved sur-factant is negligible. Predicted mixture CMC values -for ideal mixing -from Equation 1 are also shown. Ideal solution theory describes mixed micelle -formation very well, as is usually the case -for similarly structured sur-factant mixtures (12.19.21—2A) ... 

Hydrate experimental conditions have been defined in large part by the needs of the natural gas transportation industry, which in turn determined that experiments be done above the ice point. Below 273.15 K there is the danger of ice as a second solid phase (in addition to hydrate) to cause fouling of transmission or processing equipment. However, since the development of the statistical theory, there has been a need to fit the hydrate formation conditions of pure components below the ice point with the objective of predicting mixtures, as suggested in Chapter 5. 

In this paper we report experimental and theoretical results on the sorption of methane and krypton on 5A zeolite. The sorption of methane in the 5A cavity is reported to be non-localized (9.), whereas that of krypton is localized at a cavity site and window site (10). The multicomponent form of the isotherm of Schirmer et al. is used to interpret the experimental data and to predict mixture equilibria at other concentrations. 

Accurate prediction of mixtures of chemicals is one of the future challenges for risk assessment and (Q)SAR modeling. Most compounds are present in the environment at concentrations far below their individual median effective concentration (EC50 or LC50) and possibly below their no-observed-adverse-effect concentrations as well, yet they may contribute to substantial effects through combination with other chemicals. It is theoretically possible to construct a (Q)SAR model to predict mixture effects however, it is generally difficult to validate their predictive power, due to a lack of experimental calibration (Altenburger et al. 2003). Semiempirical... 

Each subsystem may have various active sites for the toxic action of compounds. However, between subsystems, the same mixture is likely to cause quantitatively different responses. Because an organism is composed of an array of subsystems, the outcome of a mixture study with compounds with distinctly different modes of action on reproduction effects may demonstrate a response that is numerically similar to concentration addition, whereas the underlying responses in all subsystems (e.g., endocrine, energetic, and metabolic systems) may differ mechanistically. Although all this is theoretically plausible, the challenge still is to provide empirical evidence that the models derived accurately predict mixture effects in exposed species. 

Because these necessary concepts have not been widely applied in the existing experimental studies, the data that have been collected in the past were reviewed on the basis of existing reviews and the mathematical characteristics of mixture models. From that, it was concluded that the mathematical models that are used in the best-case studies do predict mixture responses relatively well, although the use of some models may not be mechanistically justified, and although the models have peculiar biases that need be taken into account in relation to the objective of the extrapolation. 

Borgert C J, Quill TF, McCarty LS. 2004. Can mode of action predict mixture toxicity for risk assessment Toxicol Appl Pharmacol 201 85-96. 

Rocks defined by composition may be monomineralic, like marble and quartzite, or predictable mixtures of minerals, like serpentinite and greenstone. 

In this chapter, we outline the issues and principles that are relevant to toxicity assessments of combined exposures. The scope of this overview is limited to combinations of chemicals, but excludes the topic of nonchemical stressors acting in concert with chemicals. Because the issues are of a generic nature, we draw on examples from human, environmental, and ecological toxicology. Section 3.2 briefly outlines approaches to mixture effects assessment (Chapter 4 elaborates these approaches in more detail), Section 3.3 discusses mixture effects in relation to modes and mechanisms of action, and Section 3.4 addresses the problems and possibilities of predicting mixture effects. In Sections 3.5 and 3.6, emphasis is on the predictability of synergism and on effects at low concentration or dose levels of chemicals in mixtures. Section 3.7 provides an overview of scarcely available data on mixture effects in real-world exposure scenarios. This chapter ends with an outlook to the future. 

In a more recent review, Belden et al. (2007) evaluated 45 studies dealing with 303 pesticide mixture experiments. The authors quantified the difference between predicted and observed mixture effect concentrations. In 88% of the studies that could be evaluated using CA, the predicted mixture effect concentrations differed by no more than a factor of 2 from the observed effect concentrations, again irrespective of the involved mode of action of the mixture components. 

Another typical source of uncertainty in mixture assessment is the potential interaction between substances. Interactions may occur in the environment (e.g., precipitation after emission in water), during absorption, transportation, and transformation in the organism, or at the site of toxic action. Interactions can be either direct, for example, a chemical reaction between 2 or more mixture components, or indirect, for example, if 1 mixture component blocks an enzyme that metabolizes another mixture component (see Chapters 1 and 2). Direct interactions between mixture components are relatively easy to predict based on physical-chemical data, but prediction of indirect interactions is much more difficult because it requires detailed information about the processes involved in the toxic mechanisms of action. One of the main challenges in mixture risk assessment is the development of a method to predict mixture interactions. A first step toward such a method could be the setup of a database, which contains the results of mixture toxicity tests. Provided such a database would contain sufficient data, it could be used to predict the likelihood and magnitude of potential interaction effects, that is, deviations for CA and RA. This information could subsequently be used to decide whether application of an extra safety factor for potential interaction effects is warranted, and to determine the size of such a factor. The mixture toxicity database could also support the search for predictive parameters of interaction effects, for example, determine which modes of action are involved in typical interactions. 

TSP Two-step prediction model. Model that predicts mixture toxicity by grouping of chemicals according to similar chemistry based on mode of action, and applying CA for chemicals with similar modes of action and IA or RA for (groups of) chemicals with different modes of action. 

Other effects of toxic chemical mixtures on soil are not predictable. Mixtures of fertilizers and pesticides produce enhanced toxic effects. The additions of urea, superphosphate, and potash enhance the toxicities of carbaryl and carbofuran insecticides to nitrogen-fixing bacteria in soilJ26 Soil co-contaminated with arsenic and DDT does not break down DDT as rapidly as soil contaminated with DDT alone. This results in a persistence of DDT in the environment. I27 ... 

It is well known that the toxicity of chemicals in a mixture does not necessarily correspond to that predicted from data on pure compounds [105]. Consequently different QSAR approaches have been developed for predicting mixture toxicity (e.g., see [106-110] and also Chapter 8). 

The development of the equation was targeted primarily at improving the accuracy of VLE calculations. The underlying reasoning in developing the equations was that a necessary condition for an equation of state to predict mixture VLE properties was that it accurately predict pure component VLE properties, namely pure component vapor pressures. 

The accuracy of predicting mixture flash points depends strongly on the validity of the pure component flash point data which are used in the calculation procedure. There is, for example, little published data for TOC flash points of hydrocarbons. Some of the accepted TOC flash point data for oxygenated solvents were unreliable, e.g., the TOC flash point acetone is often quoted as 15°F this was redetermined in the present work at — 20°F. Component flash points shown in Table III were generally taken from published flash point data and were not checked experimentally. 

As with the viscosity, simplistic mole-fraction averaging of pure-component values is not advisable for gas-phase thermal conductivities. The most common method for predicting mixture values has the form... 

Leinonen and Mackay (5) used this method with the Redlich-Kister equation for the hydrocarbon phase activity coefficients to predict mixture solubilities. Agreement with experiment was good for mixtures of benzene with n-hexane, 1-hexane, and 2-methylpentane and for cyclohexane mixtures with the same substances. 

The modified VDW one-fluid mixing rules for ( X> Cx and Sx in Equations 6, 7, and 16 were used to determine the ability of this formulation of the conformal solution model for predicting mixture behavior. The following relations were used for cry, ey, and Sy where i /,... 

As with the selection of a suitable single-component adsorption isotherm model, a final decision about the accuracy and validity of this attractive approach to predict mixture isotherms has to be based on the results of a critical comparison between simulated and measured concentration profiles. 

If no binary experimental data are available, powerful predictive models (group contribution methods and group contribution equations of state) can be applied today to reliably predict the missing pure component properties (see Chapter 3) and the phase equilibrium behavior (see Chapters 5 and 7). These predicted mixture data can be used, for example, to fit the missing binary parameters for a multicomponent system. 

In this form, the Pedersen et al. model can be extended only to well-defined mixtures for which critical properties of the constituents are available. They tested their models for 419 data points of seven binary hydrocarbon mixtures and were able to predict mixture viscosities with an average error of 7.4%. 

Because of the limited success of the Langmuir model in predicting mixture equilibria, several authors have modified the equations by the introduction of a power law expression of Freuncllich form ... 

---