Óbitos e internações de vítimas de eventos toxicológicos não medicamentosos no Brasil, 2009 a 2018 / Muertes y hospitalizaciones de víctimas de eventos toxicológicos no medicamentosos en Brasil, 2009 a 2018 / Deaths and hospitalizations of victims of non-drug toxicological events in Brazil, 2009 to 2018

Objetivo: determinar a taxa de internações por eventos agudos de intoxicação não medicamentosa (NMx) e analisar a mortalidade decorrente desses agravos no Brasil, de 2009 a 2018. Métodos: estudo de série temporal, no qual se analisaram registros de internações por "tratamento de intoxicação ou envenenamento por exposição a substâncias de uso não medicamentoso" no Sistema de Informações Hospitalares (SIH), por regressão de Prais-Winsten. Resultados: ocorreram 125.570 internações em virtude de intoxicação NMx. A taxa média de internações foi de 6,3/100 mil habitantes, sendo maior no sexo masculino (8,0/100 mil hab.) comparado ao feminino (4,6/100 mil hab.). A taxa de internações e a mortalidade geral de internações por intoxicação NMx diminuíram de 9,4 para 4,5/100 mil hab. e de 2,5 para 1,6/1 milhão de hab., respectivamente. Conclusões: houve redução da taxa de internações e da mortalidade por intoxicações NMx durante a década analisada.

Objetivo: determinar la tasa de hospitalizaciones por eventos agudos de intoxicación no medicamentosa (NMx) y analizar la mortalidad resultante en Brasil de 2009 a 2018. Métodos: estudio de serie temporal en el que se analizaron los registros de hospitalizaciones por "tratamiento de intoxicación o envenenamiento por exposición a sustancias de uso no farmacológico" del Sistema de Información Hospitalaria (SIH) por la regresión de Prais-Winsten. Resultados: hubo 125.570 hospitalizaciones por intoxicación NMx. La mortalidad promedio de hospitalizaciones fue de 6,3/100 mil hab., siendo más alta en el sexo masculino (8,0/100 mil hab.) en comparación con el femenino (4,6/100 mil hab.). La tasa de hospitalizaciones y la mortalidad global de las hospitalizaciones por NMx disminuyeron de 9,4 a 4,5 por 100 mil hab. y de 2,5 a 1,6 por 1 millón de hab., respectivamente. Conclusiones: hubo reducción en la tasa de hospitalizaciones y en la mortalidad por intoxicaciones NMx durante la década analizada.

Objective: to determine the rate of hospitalizations due to acute non-drug poisoning (NDP) events and to analyze mortality arising from these health conditions in Brazil from 2009 to 2018. Methods: this was a time-series study using Prais-Winsten regression to analyze records of hospitalizations for "treatment of intoxication or poisoning due to exposure to non-drug substances" held on the Hospital Information System. Results: there were 125,570 hospitalizations due to NDP. The average hospitalization rate was 6.3/100,000 inhabitants, although it was higher in males (8.0/100,000 inhab.) compared to females (4.6/100,000 inhab.). The hospitalization rate and the overall mortality rate due NDP to fell from 9.4 to 4.5/100,000 inhab. and from 2.5 to 1.6/1 million inhab., respectively. Conclusions: there was a reduction in the NDP hospitalization rate and in mortality due to NDP during the decade analyzed.

Revealing cytotoxic substructures in molecules using deep learning.

In drug development, late stage toxicity issues of a compound are the main cause of failure in clinical trials. In silico methods are therefore of high importance to guide the early design process to reduce time, costs and animal testing. Technical advances and the ever growing amount of available toxicity data enabled machine learning, especially neural networks, to impact the field of predictive toxicology. In this study, cytotoxicity prediction, one of the earliest handles in drug discovery, is investigated using a deep learning approach trained on a highly consistent in-house data set of over 34,000 compounds with a share of less than 5% of cytotoxic molecules. The model reached a balanced accuracy of over 70%, similar to previously reported studies using Random Forest. Albeit yielding good results, neural networks are often described as a black box lacking deeper mechanistic understanding of the underlying model. To overcome this absence of interpretability, a Deep Taylor Decomposition method is investigated to identify substructures that may be responsible for the cytotoxic effects, the so-called toxicophores. Furthermore, this study introduces cytotoxicity maps which provide a visual structural interpretation of the relevance of these substructures. Using this approach could be helpful in drug development to predict the potential toxicity of a compound as well as to generate new insights into the toxic mechanism. Moreover, it could also help to de-risk and optimize compounds.

Bisfenol-A: cuantificacion en orina de mujeres embarazadas por cromatografia gaseosa-espectrometria de masa / Bisfenol-A: quantification in urine of pregnant women by gas chromatography-mass spectrometry / Bisfenol-A: quantificacao em urina de gestantes por cromatografia gasosa-espectrometria de massa

El uso de bisfenol-A (BPA) a nivel de la industria global se ha venido incrementando en los ultimos anos, y fueron los mercados emergentes los impulsores de esta demanda creciente. Las aplicaciones de BPA en la industria de los alimentos y bebidas representan solo del 3 al 4% del consumo global de policarbonato, pero su uso esta siendo reexaminado debido a que se conocieron varios trabajos cientificos que indican la existencia de una relacion directa entre el BPA y los efectos adversos para la salud. La contaminacion de los alimentos y bebidas se produce por migracion del BPA desde los envases que los contienen (alimentos enlatados, vinos, etc.), y es la principal fuente de exposicion en el humano. Para evaluar dicha exposicion se desarrollo y valido un metodo analitico por cromatografia gaseosa acoplada a espectrometria de masa para la cuantificacion de BPA total en orina de mujeres embarazadas atendidas en el Hospital Italiano de Buenos Aires en el ano 2013, con un limite de cuantificacion de 2,0 ng/mL y un limite de deteccion de 0,8 ng/mL. De las 149 muestras de orina analizadas, el 66,4% fueron cuantificables, con la mediana de BPA total de 4,8 ng/mL (4,3 ng/mg de creatinina) y la media geometrica de 4,8 ng/mL (4,7 ng/mg de creatinina).

The use of bisphenol-A (BPA) at the level of the global industry has been increasing in recent years, with emerging markets being the drivers of this growing demand. BPA applications in the food and beverage industry represent only 3 to 4% of the global consumption of polycarbonate, but its use is being reexamined because several scientific works were reported indicating the existence of a direct relationship between BPA and adverse effects on health. The contamination of food and beverages is produced by the migration of BPA from the containers that hold them (canned foods, wines, etc.) and it is the main source of exposure in humans. To evaluate this exposure, an analytical method was developed by gas chromatography coupled to mass spectrometry for the quantification of total BPA in urine of pregnant women treated at the Hospital Italiano de Buenos Aires in 2013, with a limit of quantification of 2.0 ng/mL and of detection of 0.8 ng/mL. Of the 149 urine samples analyzed, 66.4% were quantifiable, with a median total BPA of 4.8 ng/mL (4.3 ng/mg creatinine) and a geometric mean of 4.8 ng/mL (4.7 ng/mg creatinine).

O uso de bisfenol-A (BPA) ao nivel da industria global foi aumentando nos ultimos anos, e foram os mercados emergentes que deram impulso a essa demanda crescente. As aplicacoes de BPA na industria de alimentos e bebidas representam apenas 3 a 4% do consumo global de policarbonato, mas seu uso esta sendo reexaminado visto que varios trabalhos cientificos indicando a existencia de uma relacao direta entre o BPA e os efeitos adversos na saude foram conhecidos. A contaminacao dos alimentos e bebidas e produzida pela migracao de BPA das embalagens que os contem (alimentos enlatados, vinhos, etc.) e e a principal fonte de exposicao em humanos. Para avaliar esta exposicao, foi desenvolvido e avaliado um metodo analitico por cromatografia gasosa acoplada a espectrometria de massas para a quantificacao do BPA total na urina de gestantes atendidas no Hospital Italiano de Buenos Aires em 2013, com um limite de quantificacao de 2,0 ng/mL e um limite de deteccao de 0,8 ng/mL. Das 149 amostras de urina analisadas, 66,4% foram quantificaveis, com uma mediana de BPA total de 4,8 ng/mL (4,3 ng/mg de creatinina) e a media geometrica de 4,8 ng/mL (4,7 ng/mg de creatinina).

Interval estimators of relative potency in toxicology and radiation countermeasure studies: comparing methods and experimental designs.

The relative potency of one agent to another is commonly represented by the ratio of two quantal response parameters; for example, the LD50 of animals receiving a treatment to the LD50 of control animals, where LD50 is the dose of toxin that is lethal to 50% of animals. Though others have considered interval estimators of LD50, here, we extend Bayesian, bootstrap, likelihood ratio, Fieller's and Wald's methods to estimate intervals for relative potency in a parallel-line assay context. In addition to comparing their coverage probabilities, we also consider their power in two types of dose designs: one assigning treatment and control the same doses vs. one choosing doses for treatment and control to achieve same lethality targets. We explore these methods in realistic contexts of relative potency of radiation countermeasures. For larger experiments (e.g., ≥100 animals), the methods return similar results regardless of the interval estimation method or experiment design. For smaller experiments (e.g., < 60 animals), Wald's method stands out among the others, producing intervals that hold closely to nominal levels and providing more power than the other methods in statistically efficient designs. Using this simple statistical method within a statistically efficient design, researchers can reduce animal numbers.

Experiencing a probabilistic approach to clarify and disclose uncertainties when setting occupational exposure limits.

OBJECTIVES: Assessment factors (AFs) are commonly used for deriving reference concentrations for chemicals. These factors take into account variabilities as well as uncertainties in the dataset, such as inter-species and intra-species variabilities or exposure duration extrapolation or extrapolation from the lowest-observed-adverse-effect level (LOAEL) to the noobserved- adverse-effect level (NOAEL). In a deterministic approach, the value of an AF is the result of a debate among experts and, often a conservative value is used as a default choice. A probabilistic framework to better take into account uncertainties and/or variability when setting occupational exposure limits (OELs) is presented and discussed in this paper. MATERIAL AND METHODS: Each AF is considered as a random variable with a probabilistic distribution. A short literature was conducted before setting default distributions ranges and shapes for each AF commonly used. A random sampling, using Monte Carlo techniques, is then used for propagating the identified uncertainties and computing the final OEL distribution. RESULTS: Starting from the broad default distributions obtained, experts narrow it to its most likely range, according to the scientific knowledge available for a specific chemical. Introducing distribution rather than single deterministic values allows disclosing and clarifying variability and/or uncertainties inherent to the OEL construction process. CONCLUSIONS: This probabilistic approach yields quantitative insight into both the possible range and the relative likelihood of values for model outputs. It thereby provides a better support in decision-making and improves transparency. Int J Occup Med Environ Health 2018;31(4):475-489.

Robust ridge regression estimators for nonlinear models with applications to high throughput screening assay data.

Nonlinear regression is often used to evaluate the toxicity of a chemical or a drug by fitting data from a dose-response study. Toxicologists and pharmacologists may draw a conclusion about whether a chemical is toxic by testing the significance of the estimated parameters. However, sometimes the null hypothesis cannot be rejected even though the fit is quite good. One possible reason for such cases is that the estimated standard errors of the parameter estimates are extremely large. In this paper, we propose robust ridge regression estimation procedures for nonlinear models to solve this problem. The asymptotic properties of the proposed estimators are investigated; in particular, their mean squared errors are derived. The performances of the proposed estimators are compared with several standard estimators using simulation studies. The proposed methodology is also illustrated using high throughput screening assay data obtained from the National Toxicology Program.

Statistical evaluation of mortality in long-term carcinogenicity bioassays using a Williams-type procedure.

Several doses and a control group can be compared under order restriction using the Williams procedure for normally distributed endpoints assuming variance homogeneity. Comparison of the survival functions represents a secondary endpoint in long-term in vivo bioassays of carcinogenicity. Therefore, a Williams-type procedure for the comparison of survival functions is proposed for the assumption of the Cox proportional hazards model or the general frailty Cox model to allow a joint analysis over sex and strains. Interpretation according to both statistical significance and biological relevance is possible with simultaneous confidence intervals for hazard ratios. Related survival data can be analyzed using the R packages survival, coxme, and multcomp. Together with the R packages MCPAN and nparcomp, Dunnett- or Williams-type procedures are now available for the statistical analysis of the following endpoint types in toxicology: (i) normally distributed, (ii) non-normally distributed, (iii) score (ordered categorical) data, (iv) crude proportions, (v) survival functions, and (vi) time-to-tumor data with and without cause-of-death information.

Exact logistic models for nested binary data.

The use of logistic models for independent binary data has relied first on asymptotic theory and later on exact distributions for small samples. However, the use of logistic models for dependent analysis based on exact analysis is not as common. Moreover, attention is usually given to one-stage clustering. In this paper, we extend the exact techniques to address hypothesis testing (estimation is not addressed) for data with second-stage and probably higher levels of clustering. The methods are demonstrated through a somewhat generic example using C+ + program.

Evaluation of incidence rates in pre-clinical studies using a Williams-type procedure.

The analysis of dose-response relationships is a common problem in pre-clinical studies. For example, proportions such as mortality rates and histopathological findings are of particular interest in repeated toxicity studies. Commonly applied designs consist of an untreated control group and several, possibly unequally spaced, dosage groups. The Williams test can be formulated as a multiple contrast test and is a powerful option to evaluate such data. In this paper, we consider simultaneous inference for Williams-type multiple contrasts when the response variable is binomial and sample sizes are only moderate. Approximate simultaneous confidence limits can be constructed using the quantiles of a multivariate normal distribution taking the correlation into account. Alternatively, multiplicity-adjusted p-values can be calculated as well. A simulation study shows that a simple correction based on adding pseudo observations leads to acceptable performance for moderate sample sizes, such as 40 per group. In addition, the calculation of adjusted p-values and approximate power is presented. Finally, the proposed methods are applied to example data from two toxicological studies; the methods are available in an R-package.

The Threshold of Toxicological Concern (TTC) in risk assessment.

The Threshold of Toxicological Concern (TTC) is a level of human intake or exposure that is considered to be of negligible risk, despite the absence of chemical-specific toxicity data. The TTC approach is a form of risk characterisation in which uncertainties arising from the use of data on other compounds are balanced against the low level of exposure. The approach was initially developed by the FDA for packaging migrants, and used a single threshold value of 1.5 microg/day (called the threshold of regulation). Subsequent analyses of chronic toxicity data resulted in the development of TTC values for three structural classes with different potentials for toxicity (1,800, 540 and 90 microg/day). These TTC values have been incorporated into the procedure that is used internationally for the evaluation of flavouring substances. Further developments included additional TTC values for certain structural alerts for genotoxicity (0.15 microg/day), and for the presence of an organophosphate group (18 microg/day). All of these TTC values were incorporated into an extended decision tree for chemicals, such as contaminants, which might be present in human foods. The TTC approach has been shown to have potential applications to risk assessments of cosmetic ingredients, household products and impurities in therapeutic drugs.

A method to integrate benchmark dose estimates with genomic data to assess the functional effects of chemical exposure.

The use of genomic technology for assessing health risks associated with chemical exposure has significant potential, but its direct application has proven to be challenging for the toxicology and risk assessment communities. In this study, a method was established for analyzing dose-response microarray data using benchmark dose (BMD) calculations and gene ontology (GO) classification. Gene expression changes in the rat nasal epithelium following acute formaldehyde exposure were used as a case study. The gene expression data were first analyzed using a one-way ANOVA to identify genes that showed significant dose-response behavior. These genes were then fit to a series of four statistical models (linear, second-degree polynomial, third-degree polynomial, and power models) and the least complex model that best described the data was selected. The genes were matched to their associated GO categories, and the average BMD and benchmark dose lower confidence limit (BMDL) were calculated for each GO category. The results were used to identify doses at which individual cellular processes were altered. For the formaldehyde exposures, the BMD estimates for the GO categories related to cell proliferation and DNA damage were similar to those measured in previous studies using cell labeling indices and DNA-protein cross-links and consistent with the BMD estimated for rat nasal tumors. The method represents a significant advance in applying genomic information to risk assessment by allowing a comprehensive survey of molecular changes associated with chemical exposure and providing the capability to identify reference doses at which particular cellular processes are altered.

Can exposure characterization explain concurrence or discordance between toxicology and epidemiology?

The importance of reliable exposure assessment, as a key component of the overall risk assessment process, has been well described for some considerable time. Yet, despite this widely accepted tenet, many studies conclude significant adverse health effects, with associated public policy implications, in the absence of adequate or, in some cases, even rudimentary, exposure quantification. Moreover, it appears that epidemiological studies in humans and toxicological studies in experimental animals may both suffer from inadequate exposure assessment. In this review, we discuss the nature and quality of the exposure assessment in both epidemiologic and toxicologic studies using examples from the pesticides and phthalate literature. Each type of study has its strengths and weaknesses in how exposure is assessed and often the strength of one is also a weakness. It would appear that insufficient or incomplete information about differences in exposure assessment could explain, at least in some cases, the differences in outcome between toxicological and epidemiological studies. Research efforts should focus on improving the feasibility of including biomonitoring in both animal and human studies to facilitate comparisons between animal and human models and improve exposure assessment in epidemiologic studies. Animal and human studies should measure the same biomarkers, where possible, to facilitate human health risk assessment.

Bayesian dose-finding in phase I/II clinical trials using toxicity and efficacy odds ratios.

A Bayesian adaptive design is proposed for dose-finding in phase I/II clinical trials to incorporate the bivariate outcomes, toxicity and efficacy, of a new treatment. Without specifying any parametric functional form for the drug dose-response curve, we jointly model the bivariate binary data to account for the correlation between toxicity and efficacy. After observing all the responses of each cohort of patients, the dosage for the next cohort is escalated, deescalated, or unchanged according to the proposed odds ratio criteria constructed from the posterior toxicity and efficacy probabilities. A novel class of prior distributions is proposed through logit transformations which implicitly imposes a monotonic constraint on dose toxicity probabilities and correlates the probabilities of the bivariate outcomes. We conduct simulation studies to evaluate the operating characteristics of the proposed method. Under various scenarios, the new Bayesian design based on the toxicity-efficacy odds ratio trade-offs exhibits good properties and treats most patients at the desirable dose levels. The method is illustrated with a real trial design for a breast medical oncology study.

On the multivariate probit model for exchangeable binary data with covariates.

This paper considers the use of a multivariate binomial probit model for the analysis of correlated exchangeable binary data. The model can naturally accommodate both cluster and individual level covariates, while keeping a fairly flexible intracluster association structure. We discuss Bayesian estimation when a sample of independent clusters of varying sizes are available, and show how Gibbs sampling may be used to derive the posterior densities of parameters. The methodology is illustrated with two examples: the first involves epidemiological data from a study of familial disease aggregation; the second uses teratological data from a developmental toxicity application.

Overview: Using mode of action and life stage information to evaluate the human relevance of animal toxicity data.

A complete mode of action human relevance analysis--as distinct from mode of action (MOA) analysis alone--depends on robust information on the animal MOA, as well as systematic comparison of the animal data with corresponding information from humans. In November 2003, the International Life Sciences Institute's Risk Science Institute (ILSI RSI) published a 2-year study using animal and human MOA information to generate a four-part Human Relevance Framework (HRF) for systematic and transparent analysis of MOA data and information. Based mainly on non-DNA-reactive carcinogens, the HRF features a "concordance" analysis of MOA information from both animal and human sources, with a focus on determining the appropriate role for each MOA data set in human risk assessment. With MOA information increasingly available for risk assessment purposes, this article illustrates the further applicability of the HRF for reproductive, developmental, neurologic, and renal endpoints, as well as cancer. Based on qualitative and quantitative MOA considerations, the MOA/human relevance analysis also contributes to identifying data needs and issues essential for the dose-response and exposure assessment steps in the overall risk assessment.

The stability of historical control data for common neoplasms in laboratory rats: adrenal gland (medulla), mammary gland, liver, endocrine pancreas, and pituitary gland.

Time-related changes in the incidences of spontaneous neoplasms in adrenals (medulla), mamma, liver, pituitary, and (endocrine) pancreas were assessed statistically in Wistar, Sprague-Dawley, and F344 rats employed by BASF, Germany and major European contract research organizations over the last 20 years. Negative trends (7 of 80 cases) were observed for pituitary pars distalis adenomas in Sprague-Dawley males and females, for pancreas islet cell adenomas in BASF Wistar males and females, for benign adrenal pheochromocytomas in Sprague-Dawley males, for malignant pheochromocytomas in F344 males, and for mammary gland fibroadenomas in BASF Wistar females. Positive trends (13 of 80 cases) were observed for benign pheochromocytomas, mammary gland adenocarcinomas, and pancreas islet cell carcinomas in HanWistar females, for malignant pheochromocytomas and islet cell carcinomas in BASF Wistar males, for benign pheochromocytomas and islet cell adenomas in F344 males, for mammary gland fibroadenomas in Sprague-Dawley females, and for benign hepatocellular tumors in HanWistar males and females, and in BASF Wistar and Sprague-Dawley females. In 60 of 80 cases there were no statistically significant trends. These results indicate that the majority of tumor types showed no time trends and that, in each rat strain, certain tumor types are susceptible to slight positive or negative time trends. Accordingly, the validity and use of historical control data should be based on an organ- and strain-specific statistical analysis of tumor incidence over time.