Categorical regression of toxicity data: a case study using aldicarb

Probabilistic methods for non-cancer health effects

Noncancer risk assessment methods and harmonization with cancer assessment methods have advanced from the simple divide a No Observed Adverse Effect Level (NOAEL) by a default safety factor or a linear extrapolation to background of the early 1980's. This advance is due in part due to groups such as the American Industrial Health Council, the National Institute of Environmental Health Sciences, the Society for Risk Analysis, the Society of Toxicology, and the U.S. Environmental Protection Agency (Bogdanffy et al., 2001), the National Academy of Sciences (1983 and 2009), the International Programme on Chemical Safety (2005, 2009), and to many independent researchers outside of and within a workshop series sponsored by the Alliance for Risk Assessment (ARA, 2023) prompted by the NAS (2009). Several of the case studies from this workshop series, and earlier work such as Bogdanffy et al. (2001), demonstrate that the dose response assessment of non-cancer toxicity and the harmonization of cancer and non-cancer methods are more than just a simple reflection of treating all non-cancer toxicity as if it has a threshold, or all cancer toxicity as if it did not. Moreover, one recommendation of NAS (2009) was to develop a problem formulation with risk managers prior to conducting any risk assessment. If the development of this problem formulation only necessitates the determination of a safe, or virtually safe dose, then the estimation of a Reference Dose (RfD) or virtually safe dose (VSD) or similar constructs should be encouraged. Not all of our environmental problems need a precise quantitative solution.

The Carbamate Aldicarb Altered the Gut Microbiome, Metabolome, and Lipidome of C57BL/6J Mice

The gut microbiome is highly involved in numerous aspects of host physiology, from energy harvest to stress response, and can confer many benefits to the host. The gut microbiome development could be affected by genetic and environmental factors, including pesticides. The carbamate insecticide aldicarb has been extensively used in agriculture, which raises serious public health concerns. However, the impact of aldicarb on the gut microbiome, host metabolome, and lipidome has not been well studied yet. Herein, we use multiomics approaches, including16S rRNA sequencing, shotgun metagenomics sequencing, metabolomics, and lipidomics, to elucidate aldicarb-induced toxicity in the gut microbiome and the host metabolic homeostasis. We demonstrated that aldicarb perturbed the gut microbiome development trajectory, enhanced gut bacterial pathogenicity, altered complex lipid profile, and induced oxidative stress, protein degradation, and DNA damage. The brain metabolism was also disturbed by the aldicarb exposure. These findings may provide a novel understanding of the toxicity of carbamate insecticides.

Dose-Related Severity Sequence, and Risk-Based Integration, of Chemically Induced Health Effects

Risk assessment of chemical hazards is typically based on single critical health effects. This work aims to expand the current approach by characterizing the dose-related sequence of the development of multiple (lower- to higher-order) toxicological health effects caused by a chemical. To this end a "reference point profile" is defined as the relation between benchmark doses for considered health effects, and a standardized severity score determined for these effects. For a given dose of a chemical or mixture the probability for exceeding the reference point profile, thereby provoking lower- to higher-order effects, can be assessed. The overall impact at the same dose can also be derived by integrating contributions across all health effects following severity-weighting. In its generalized form the new impact metric relates to the probability of response for the most severe health effects. Reference points (points of departure) corresponding to defined levels of response can also be estimated. The proposed concept, which is evaluated for dioxin-like chemicals, provides an alternative for characterizing the low-dose region below the reference point for a severe effect like cancer. The shape and variability of the reference point profile add new dimensions to risk assessment, which for example extends the characterization of chemical potency, and the concept of acceptable effect sizes for individual health effects. Based on the present data the method shows high stability at low doses/responses, and is also robust to differences in severity categorization of effects. In conclusion, the novel method proposed enables risk-based integration of multiple dose-related health effects. It provides a first step towards a more comprehensive characterization of chemical toxicity, and suggests a potential for improved low-dose risk assessment.

Benchmark dose (BMD) modeling: current practice, issues, and challenges

Benchmark dose (BMD) modeling is now the state of the science for determining the point of departure for risk assessment. Key advantages include the fact that the modeling takes account of all of the data for a particular effect from a particular experiment, increased consistency, and better accounting for statistical uncertainties. Despite these strong advantages, disagreements remain as to several specific aspects of the modeling, including differences in the recommendations of the US Environmental Protection Agency (US EPA) and the European Food Safety Authority (EFSA). Differences exist in the choice of the benchmark response (BMR) for continuous data, the use of unrestricted models, and the mathematical models used; these can lead to differences in the final BMDL. It is important to take confidence in the model into account in choosing the BMDL, rather than simply choosing the lowest value. The field is moving in the direction of model averaging, which will avoid many of the challenges of choosing a single best model when the underlying biology does not suggest one, but additional research would be useful into methods of incorporating biological considerations into the weights used in the averaging. Additional research is also needed regarding the interplay between the BMR and the UF to ensure appropriate use for studies supporting a lower BMR than default values, such as for epidemiology data. Addressing these issues will aid in harmonizing methods and moving the field of risk assessment forward.

Modeling U-Shaped Exposure-Response Relationships for Agents that Demonstrate Toxicity Due to Both Excess and Deficiency

Essential elements such as copper and manganese may demonstrate U-shaped exposure-response relationships due to toxic responses occurring as a result of both excess and deficiency. Previous work on a copper toxicity database employed CatReg, a software program for categorical regression developed by the U.S. Environmental Protection Agency, to model copper excess and deficiency exposure-response relationships separately. This analysis involved the use of a severity scoring system to place diverse toxic responses on a common severity scale, thereby allowing their inclusion in the same CatReg model. In this article, we present methods for simultaneously fitting excess and deficiency data in the form of a single U-shaped exposure-response curve, the minimum of which occurs at the exposure level that minimizes the probability of an adverse outcome due to either excess or deficiency (or both). We also present a closed-form expression for the point at which the exposure-response curves for excess and deficiency cross, corresponding to the exposure level at which the risk of an adverse outcome due to excess is equal to that for deficiency. The application of these methods is illustrated using the same copper toxicity database noted above. The use of these methods permits the analysis of all available exposure-response data from multiple studies expressing multiple endpoints due to both excess and deficiency. The exposure level corresponding to the minimum of this U-shaped curve, and the confidence limits around this exposure level, may be useful in establishing an acceptable range of exposures that minimize the overall risk associated with the agent of interest.

Exposure Estimation and Interpretation of Occupational Risk: Enhanced Information for the Occupational Risk Manager

The fundamental goal of this article is to describe, define, and analyze the components of the risk characterization process for occupational exposures. Current methods are described for the probabilistic characterization of exposure, including newer techniques that have increasing applications for assessing data from occupational exposure scenarios. In addition, since the probability of health effects reflects variability in the exposure estimate as well as the dose-response curve-the integrated considerations of variability surrounding both components of the risk characterization provide greater information to the occupational hygienist. Probabilistic tools provide a more informed view of exposure as compared to use of discrete point estimates for these inputs to the risk characterization process. Active use of such tools for exposure and risk assessment will lead to a scientifically supported worker health protection program. Understanding the bases for an occupational risk assessment, focusing on important sources of variability and uncertainty enables characterizing occupational risk in terms of a probability, rather than a binary decision of acceptable risk or unacceptable risk. A critical review of existing methods highlights several conclusions: (1) exposure estimates and the dose-response are impacted by both variability and uncertainty and a well-developed risk characterization reflects and communicates this consideration; (2) occupational risk is probabilistic in nature and most accurately considered as a distribution, not a point estimate; and (3) occupational hygienists have a variety of tools available to incorporate concepts of risk characterization into occupational health and practice.

The Global Landscape of Occupational Exposure Limits--Implementation of Harmonization Principles to Guide Limit Selection

Occupational exposure limits (OELs) serve as health-based benchmarks against which measured or estimated workplace exposures can be compared. In the years since the introduction of OELs to public health practice, both developed and developing countries have established processes for deriving, setting, and using OELs to protect workers exposed to hazardous chemicals. These processes vary widely, however, and have thus resulted in a confusing international landscape for identifying and applying such limits in workplaces. The occupational hygienist will encounter significant overlap in coverage among organizations for many chemicals, while other important chemicals have OELs developed by few, if any, organizations. Where multiple organizations have published an OEL, the derived value often varies considerably-reflecting differences in both risk policy and risk assessment methodology as well as access to available pertinent data. This article explores the underlying reasons for variability in OELs, and recommends the harmonization of risk-based methods used by OEL-deriving organizations. A framework is also proposed for the identification and systematic evaluation of OEL resources, which occupational hygienists can use to support risk characterization and risk management decisions in situations where multiple potentially relevant OELs exist.

Benchmark dose calculation for ordered categorical responses

The use of benchmark dose (BMD) calculations for dichotomous or continuous responses is well established in the risk assessment of cancer and noncancer endpoints. In some cases, responses to exposure are categorized in terms of ordinal severity effects such as none, mild, adverse, and severe. Such responses can be assessed using categorical regression (CATREG) analysis. However, while CATREG has been employed to compare the benchmark approach and the no-adverse-effect-level (NOAEL) approach in determining a reference dose, the utility of CATREG for risk assessment remains unclear. This study proposes a CATREG model to extend the BMD approach to ordered categorical responses by modeling severity levels as censored interval limits of a standard normal distribution. The BMD is calculated as a weighted average of the BMDs obtained at dichotomous cutoffs for each adverse severity level above the critical effect, with the weights being proportional to the reciprocal of the expected loss at the cutoff under the normal probability model. This approach provides a link between the current BMD procedures for dichotomous and continuous data. We estimate the CATREG parameters using a Markov chain Monte Carlo simulation procedure. The proposed method is demonstrated using examples of aldicarb and urethane, each with several categories of severity levels. Simulation studies comparing the BMD and BMDL (lower confidence bound on the BMD) using the proposed method to the correspondent estimates using the existing methods for dichotomous and continuous data are quite compatible; the difference is mainly dependent on the choice of cutoffs for the severity levels.

Risk-benefit analysis of mineral intakes: case studies on copper and iron

Dietary reference values for essential trace elements are designed to meet requirements with minimal risk of deficiency and toxicity. Risk-benefit analysis requires data on habitual dietary intakes, an estimate of variation and effects of deficiency and excess on health. For some nutrients, the range between the upper and lower limits may be extremely narrow and even overlap, which creates difficulties when setting safety margins. A new approach for estimating optimal intakes, taking into account several health biomarkers, has been developed and applied to selenium, but at present there are insufficient data to extend this technique to other micronutrients. The existing methods for deriving reference values for Cu and Fe are described. For Cu, there are no sensitive biomarkers of status or health relating to marginal deficiency or toxicity, despite the well-characterised genetic disorders of Menkes and Wilson's disease which, if untreated, lead to lethal deficiency and overload, respectively. For Fe, the wide variation in bioavailability confounds the relationship between intake and status and complicates risk-benefit analysis. As with Cu, health effects associated with deficiency or toxicity are not easy to quantify, therefore status is the most accessible variable for risk-benefit analysis. Serum ferritin reflects Fe stores but is affected by infection/inflammation, and therefore additional biomarkers are generally employed to measure and assess Fe status. Characterising the relationship between health and dietary intake is problematic for both these trace elements due to the confounding effects of bioavailability, inadequate biomarkers of status and a lack of sensitive and specific biomarkers for health outcomes.

The use of categorical regression in modeling copper exposure-response relationships

Characterization of the exposure-response relationship for copper (Cu) is an essential step in identifying a range of exposures that can prevent against toxicity from either excess or deficiency. Categorical regression is a exposure-response modeling technique that can be used to model data from multiple studies with diverse endpoints simultaneously by organizing the toxicity data into ordered categories of severity. This study describes how categorical regression can be used to model the exposure-response relationship for Cu and presents a preliminary analysis of the comprehensive database on Cu-induced toxicity due to either excess or deficiency. Categorical regression provides a useful tool for summarizing and describing the available data on Cu excess and deficiency, as well as in identifying data gaps in Cu exposure-response. This methodology also allows for a diverse database with considerable variability in animal species, strain, age, and study design to be analyzed in its entirety. The present application of the Cu toxicity database suggests that there is a lack of information on the potential adverse health effects from chronic exposure to Cu; there are also a limited number of studies using marginally excess and deficient levels of Cu. The database presently includes insufficient data to create a complex model that accounts for a large proportion of the heterogeneity in toxicity seen among the available studies on Cu-induced toxicity. The current Cu database is presently being updated in order to permit more comprehensive categorical regression analyses with finer stratification options. The resulting exposure-response model could be used to provide information in the determination of an acceptable range of oral intake for Cu.

A novel logic-based approach for quantitative toxicology prediction

There is a pressing need for accurate in silico methods to predict the toxicity of molecules that are being introduced into the environment or are being developed into new pharmaceuticals. Predictive toxicology is in the realm of structure activity relationships (SAR), and many approaches have been used to derive such SAR. Previous work has shown that inductive logic programming (ILP) is a powerful approach that circumvents several major difficulties, such as molecular superposition, faced by some other SAR methods. The ILP approach reasons with chemical substructures within a relational framework and yields chemically understandable rules. Here, we report a general new approach, support vector inductive logic programming (SVILP), which extends the essentially qualitative ILP-based SAR to quantitative modeling. First, ILP is used to learn rules, the predictions of which are then used within a novel kernel to derive a support-vector generalization model. For a highly heterogeneous dataset of 576 molecules with known fathead minnow fish toxicity, the cross-validated correlation coefficients (R2CV) from a chemical descriptor method (CHEM) and SVILP are 0.52 and 0.66, respectively. The ILP, CHEM, and SVILP approaches correctly predict 55, 58, and 73%, respectively, of toxic molecules. In a set of 165 unseen molecules, the R2 values from the commercial software TOPKAT and SVILP are 0.26 and 0.57, respectively. In all calculations, SVILP showed significant improvements in comparison with the other methods. The SVILP approach has a major advantage in that it uses ILP automatically and consistently to derive rules, mostly novel, describing fragments that are toxicity alerts. The SVILP is a general machine-learning approach and has the potential of tackling many problems relevant to chemoinformatics including in silico drug design.

Copper and human health: biochemistry, genetics, and strategies for modeling dose-response relationships

Copper (Cu) and its alloys are used extensively in domestic and industrial applications. Cu is also an essential element in mammalian nutrition. Since both copper deficiency and copper excess produce adverse health effects, the dose-response curve is U-shaped, although the precise form has not yet been well characterized. Many animal and human studies were conducted on copper to provide a rich database from which data suitable for modeling the dose-response relationship for copper may be extracted. Possible dose-response modeling strategies are considered in this review, including those based on the benchmark dose and categorical regression. The usefulness of biologically based dose-response modeling techniques in understanding copper toxicity was difficult to assess at this time since the mechanisms underlying copper-induced toxicity have yet to be fully elucidated. A dose-response modeling strategy for copper toxicity was proposed associated with both deficiency and excess. This modeling strategy was applied to multiple studies of copper-induced toxicity, standardized with respect to severity of adverse health outcomes and selected on the basis of criteria reflecting the quality and relevance of individual studies. The use of a comprehensive database on copper-induced toxicity is essential for dose-response modeling since there is insufficient information in any single study to adequately characterize copper dose-response relationships. The dose-response modeling strategy envisioned here is designed to determine whether the existing toxicity data for copper excess or deficiency may be effectively utilized in defining the limits of the homeostatic range in humans and other species. By considering alternative techniques for determining a point of departure and low-dose extrapolation (including categorical regression, the benchmark dose, and identification of observed no-effect levels) this strategy will identify which techniques are most suitable for this purpose. This analysis also serves to identify areas in which additional data are needed to better define the characteristics of dose-response relationships for copper-induced toxicity in relation to excess or deficiency.

Risk assessment for benefits analysis: framework for analysis of a thyroid-disrupting chemical

Benefit-cost analysis is of growing importance in developing policies to reduce exposures to environmental contaminants. To quantify health benefits of reduced exposures, economists generally rely on dose-response relationships estimated by risk assessors. Further, to be useful for benefits analysis, the endpoints that are quantified must be expressed as changes in incidence of illnesses or symptoms that are readily understood by and perceptible to the layperson. For most noncancer health effects and for nonlinear carcinogens, risk assessments generally do not provide the dose-response functions necessary for economic benefits analysis. This article presents the framework for a case study that addresses these issues through a combination of toxicology, epidemiology, statistics, and economics. The case study assesses a chemical that disrupts proper functioning of the thyroid gland, and considers the benefits of reducing exposures in terms of both noncancer health effects (hypothyroidism) and thyroid cancers. The effects are presumed to be due to a mode of action involving interference with thyroid-pituitary functioning that would lead to nonlinear dose response. The framework integrates data from animal testing, statistical modeling, human data from the medical and epidemiological literature, and economic methodologies and valuation studies. This interdisciplinary collaboration differs from the more typical approach in which risk assessments and economic analyses are prepared independently of one another. This framework illustrates particular approaches that may be useful for expanded quantification of adverse health effects, and demonstrates the potential of such interdisciplinary approaches. Detailed implementation of the case study framework will be presented in future publications.

Assessing biomarker use in risk assessment--a survey of practitioners

Advances in molecular epidemiology and mechanistic toxicology have provided increased opportunities for incorporating biomarkers in the human health risk assessment process. For years, the published literature has lauded the concept of incorporating biomarkers into risk assessments as a means to reduce uncertainty in estimating health risk. For all the potential benefits, one would think that markers of effective dose, markers of early biological effects, and markers of human susceptibility are frequently selected as the basis for quantitative human health risk assessments. For this article, we sought to determine the degree to which this evolution in risk assessment has come to pass. The extent to which biomarkers are being used in current human health risk assessment was determined through an informal survey of leading risk assessment practitioners. Case studies highlighting the evolution of risk assessment methods to include biomarkers are also described. The goal of this review was to enhance the implementation of biomarker technology in risk assessment by (1) highlighting successes in biomarker implementation, (2) identifying key barriers to overcome, and (3) describing evolutions in risk assessment methods.

Mathematical modelling and quantitative methods

The present review reports on the mathematical methods and statistical techniques presently available for hazard characterisation. The state of the art of mathematical modelling and quantitative methods used currently for regulatory decision-making in Europe and additional potential methods for risk assessment of chemicals in food and diet are described. Existing practices of JECFA, FDA, EPA, etc., are examined for their similarities and differences. A framework is established for the development of new and improved quantitative methodologies. Areas for refinement, improvement and increase of efficiency of each method are identified in a gap analysis. Based on this critical evaluation, needs for future research are defined. It is concluded from our work that mathematical modelling of the dose-response relationship would improve the risk assessment process. An adequate characterisation of the dose-response relationship by mathematical modelling clearly requires the use of a sufficient number of dose groups to achieve a range of different response levels. This need not necessarily lead to an increase in the total number of animals in the study if an appropriate design is used. Chemical-specific data relating to the mode or mechanism of action and/or the toxicokinetics of the chemical should be used for dose-response characterisation whenever possible. It is concluded that a single method of hazard characterisation would not be suitable for all kinds of risk assessments, and that a range of different approaches is necessary so that the method used is the most appropriate for the data available and for the risk characterisation issue. Future refinements to dose-response characterisation should incorporate more clearly the extent of uncertainty and variability in the resulting output.

[Assessment of the chronic risk for ingestion of pesticide residues in the Brazilian diet]

OBJECTIVE

To conduct a chronic dietary risk assessment of the pesticides registered in Brazil up until 1999.

METHODS

The Theoretical Maximum Daily Intake (TMDI) for each pesticide was calculated using the Brazilian maximum residue limits and food consumption data from IBGE, the Brazilian Statistical Institute. The risk characterization was done comparing the TMDI with the acceptable daily intakes (ADI) from other countries and from the Codex Alimentarius.

RESULTS

The TMDI was higher than the ADI (%ADI>100) at least in one Brazilian metropolitan region for 23 pesticides. Sixteen compounds are organophosphate insecticides, with methyl parathion having the TMDI exceeding the most toxicological parameter (%ADI N=9,300). Rice, beans, citrus and tomato were the commodities which most contributed to the ingestion. From the compounds under higher risk, only 6 were registered according to the Law 98.816/90, which concerns the use of pesticides in the country.

CONCLUSIONS

The compounds identified in the study as presenting a potential health concern to the Brazilian consumers, and the commodities which most contributed to the ingestion, should be prioritized by the government in pesticide residue monitoring programs and in the re-registration process. In addition, residue data in food as consumed, processing factors and appropriate consumption data should be generated to allow further studies.

Haber's rule: a special case in a family of curves relating concentration and duration of exposure to a fixed level of response for a given endpoint

The concept that the product of the concentration (C) of a substance and the length of time (t) it is administered produces a fixed level of effect for a given endpoint has been ascribed to Fritz Haber, who was a German scientist in the early 1900s. He contended that the acute lethality of war gases could be assessed by the amount of the gas in a cubic meter of air (i.e. the concentration) multiplied by the time in min that the animal had to breathe the air before death ensued (i.e. C x t=k). While Haber recognized that C x t=k was applicable only under certain conditions, many toxicologists have used his rule to analyze experimental data whether or not their chemicals, biological endpoints, and exposure scenarios were suitable candidates for the rule. The fact that the relationship between C and t is linear on a log-log scale and could easily be solved by hand, led to early acceptance among toxicologists, particularly in the field of entomology. In 1940, a statistician named Bliss provided an elegant treatment on the relationships among exposure time, concentration, and the toxicity of insecticides. He proposed solutions for when the log-log plot of C and t was composed of two or more rectilinear segments, for when the log-log plot was curvilinear, and for when the slope of the dosage-mortality curve was a function of C. Despite the fact that Haber's rule can underestimate or overestimate effects (and consequently risks), it has been used in various settings by regulatory bodies. Examples are presented from the literature of data sets that follow Haber's rule as well as those that do not. Haber's rule is put into perspective by showing that it is simply a special case in a family of power law curves relating concentration and duration of exposure to a fixed level of response for a given endpoint. Also shown is how this power law family can be used to examine the three-dimensional surface relating C, t, and varying levels of response. The time has come to move beyond the limited view of C and t relationships inferred by Haber's rule to the use of the broader family of curves of which this rule is a special case.

Health risk above the reference dose for multiple chemicals

Recent work indicates that the regression of toxicity data viewed as categories of pathological staging is useful for exploring the likely health risk at doses above a Reference Dose (RfD), which is an estimate (with uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human population (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a lifetime. Toxic effects, which may include both quantal and continuous data, are classified into ordered categories of total toxic severity (e.g., none, mild, adverse, severe). These severity categories are regressed on explanatory variables, such as dose or exposure duration, to estimate the probability of observing an adverse or severe effect. In this paper, categorical regression has been expanded to compare the likely risks across multiple chemicals when exposures are above their RfDs. Existing health risk data for diazinon, disulfoton, S-ethyl dipropylthiocarbamate, fenamiphos, and lindane were analyzed. As expected, the estimated risks of adverse effects above the RfD varied among the chemicals. For example, at 10-fold above the RfD these risks were modeled to be 0.002, 0.0001, 0.0007, 0.002, and 0.02, respectively. The results and impacts of this analysis indicate that categorical regression is a useful screening tool to analyze risks above the RfD for specific chemicals and suggest its application in evaluating comparative risks where multiple chemical exposures exist.