Improved family-wise error rate control in multiple equivalence testing.

Equivalence testing is an important component of safety assessments, used for example by the European Food Safety Authority, to allow new food or feed products on the market. The aim of such tests is to demonstrate equivalence of characteristics of test and reference crops. Equivalence tests are typically univariate and applied to each measured analyte (characteristic) separately without multiplicity correction. This increases the probability of making false claims of equivalence (type I errors) when evaluating multiple analytes simultaneously. To solve this problem, familywise error rate (FWER) control using Hochberg's method has been proposed. This paper demonstrates that, in the context of equivalence testing, other FWER-controlling methods are more powerful than Hochberg's. Particularly, it is shown that Hommel's method is guaranteed to perform at least as well as Hochberg's and that an "adaptive" version of Bonferroni's method, which uses an estimator of the proportion of non-equivalent characteristics, often substantially outperforms Hommel's method. Adaptive Bonferroni takes better advantage of the particular context of food safety where a large proportion of true equivalences is expected, a situation where other methods are particularly conservative. The different methods are illustrated by their application to two compositional datasets and further assessed and compared using simulated data.

FAIR environmental and health registry (FAIREHR)- supporting the science to policy interface and life science research, development and innovation.

The environmental impact on health is an inevitable by-product of human activity. Environmental health sciences is a multidisciplinary field addressing complex issues on how people are exposed to hazardous chemicals that can potentially affect adversely the health of present and future generations. Exposure sciences and environmental epidemiology are becoming increasingly data-driven and their efficiency and effectiveness can significantly improve by implementing the FAIR (findable, accessible, interoperable, reusable) principles for scientific data management and stewardship. This will enable data integration, interoperability and (re)use while also facilitating the use of new and powerful analytical tools such as artificial intelligence and machine learning in the benefit of public health policy, and research, development and innovation (RDI). Early research planning is critical to ensuring data is FAIR at the outset. This entails a well-informed and planned strategy concerning the identification of appropriate data and metadata to be gathered, along with established procedures for their collection, documentation, and management. Furthermore, suitable approaches must be implemented to evaluate and ensure the quality of the data. Therefore, the 'Europe Regional Chapter of the International Society of Exposure Science' (ISES Europe) human biomonitoring working group (ISES Europe HBM WG) proposes the development of a FAIR Environment and health registry (FAIREHR) (hereafter FAIREHR). FAIR Environment and health registry offers preregistration of studies on exposure sciences and environmental epidemiology using HBM (as a starting point) across all areas of environmental and occupational health globally. The registry is proposed to receive a dedicated web-based interface, to be electronically searchable and to be available to all relevant data providers, users and stakeholders. Planned Human biomonitoring studies would ideally be registered before formal recruitment of study participants. The resulting FAIREHR would contain public records of metadata such as study design, data management, an audit trail of major changes to planned methods, details of when the study will be completed, and links to resulting publications and data repositories when provided by the authors. The FAIREHR would function as an integrated platform designed to cater to the needs of scientists, companies, publishers, and policymakers by providing user-friendly features. The implementation of FAIREHR is expected to yield significant benefits in terms of enabling more effective utilization of human biomonitoring (HBM) data.

Comparative safety assessment of genetically modified crops: focus on equivalence with reference varieties could contribute to more efficient and effective field trials.

The initial compositional analysis of plants plays an important role within the internationally harmonized comparative safety assessment approach for genetically modified plants. Current EFSA guidance prescribes two types of comparison, namely difference tests with regard to a conventional comparator or control, and equivalence tests with regard to a collection of commercial reference varieties. The experience gained so far shows that most of the statistically significant differences between the test and control can be discounted based on the fact that they are still within equivalence limits of reference varieties with a presumed history of safe use. Inclusion of a test variety and reference varieties into field trial design, and of the statistical equivalence test would already suffice for the purpose of finding relevant parameters that warrant further assessment, hence both the inclusion of a conventional counterpart and the performance of difference testing can be omitted. This would also allow for the inclusion of safety testing regimes into plant variety testing VCU (value for cultivation and use) or other, independent variety trials.

Multivariate equivalence testing for food safety assessment.

Products for food and feed derived from genetically modified (GM) crops are only allowed on the market when they are deemed to be safe for human health and the environment. The European Food Safety Authority (EFSA) performs safety assessment including a comparative approach: the compositional characteristics of a GM genotype are compared to those of reference genotypes that have a history of safe use. Statistical equivalence tests are used to carry out such a comparative assessment. These tests are univariate and therefore only consider one measured variable at a time. Phenotypic data, however, often comprise measurements on multiple variables that must be integrated to arrive at a single decision on acceptance in the regulatory process. The surge of modern molecular phenotyping platforms further challenges this integration, due to the large number of characteristics measured on the plants. This paper presents a new multivariate equivalence test that naturally extends a recently proposed univariate equivalence test and allows to assess equivalence across all variables simultaneously. The proposed test is illustrated on plant compositional data from a field study on maize grain and on untargeted metabolomic data of potato tubers, while its performance is assessed on simulated data.

Measurement uncertainty for detection of visual impurities in granular feed and food materials in relation to the investigated amount of material.

The presence is regulated of visually detectable seeds from a selection of toxic plants and fungi mycelium bodies (sclerotia) in feed (Directive 2002/32/EC) and in food (Regulation (EC) 1881/2006). Homogenisation as typical for chemical analyses is not applicable, and dedicated approaches are needed for visual examination methods. Visual methods require two parameters to characterise measurement uncertainties for both unit counts and unit weights. A new approach is to divide approximately 2 kg of sample material into four subsamples of approximately 500 g and to separately examine the four subsamples for numbers and particle weights of seeds or sclerotia. This study is the first to produce datasets on inhomogeneity among subsamples of a sample for visually detectable undesirable substances. Analytical thresholds were calculated from a simulation model and bootstrap procedures based on our data. The analytical thresholds assuring a controlled false-negative rate of 5% for decisions in compliance with legal limits depend on the diversity of the unit counts and weights, the level of the legal limit and the amount of material examined initially in the step-wise approach, either one or two subsamples. A procedure is proposed for examination in practice where only two subsamples, or alternatively even only one subsample, would be examined. If the resulting level of contamination exceeds the relevant threshold additional subsamples need to be examined as well. In most of the investigated cases, analytical thresholds could be established for the examination of just one subsample (500 g) taken from a sample of 2 kg. However, for ergot sclerotia in food with a legal limit of 200 mg kg-1, at least two subsamples (1000 g) need to be examined in the first step. Other groups of visually detectable undesirable substances exist which need further attention.

A human biomonitoring (HBM) Global Registry Framework: Further advancement of HBM research following the FAIR principles.

Data generated by the rapidly evolving human biomonitoring (HBM) programmes are providing invaluable opportunities to support and advance regulatory risk assessment and management of chemicals in occupational and environmental health domains. However, heterogeneity across studies, in terms of design, terminology, biomarker nomenclature, and data formats, limits our capacity to compare and integrate data sets retrospectively (reuse). Registration of HBM studies is common for clinical trials; however, the study designs and resulting data collections cannot be traced easily. We argue that an HBM Global Registry Framework (HBM GRF) could be the solution to several of challenges hampering the (re)use of HBM (meta)data. The aim is to develop a global, host-independent HBM registry framework based on the use of harmonised open-access protocol templates from designing, undertaking of an HBM study to the use and possible reuse of the resulting HBM (meta)data. This framework should apply FAIR (Findable, Accessible, Interoperable and Reusable) principles as a core data management strategy to enable the (re)use of HBM (meta)data to its full potential through the data value chain. Moreover, we believe that implementation of FAIR principles is a fundamental enabler for digital transformation within environmental health. The HBM GRF would encompass internationally harmonised and agreed open access templates for HBM study protocols, structured web-based functionalities to deposit, find, and access harmonised protocols of HBM studies. Registration of HBM studies using the HBM GRF is anticipated to increase FAIRness of the resulting (meta)data. It is also considered that harmonisation of existing data sets could be performed retrospectively. As a consequence, data wrangling activities to make data ready for analysis will be minimised. In addition, this framework would enable the HBM (inter)national community to trace new HBM studies already in the planning phase and their results once finalised. The HBM GRF could also serve as a platform enhancing communication between scientists, risk assessors, and risk managers/policy makers. The planned European Partnership for the Assessment of Risk from Chemicals (PARC) work along these lines, based on the experience obtained in previous joint European initiatives. Therefore, PARC could very well bring a first demonstration of first essential functionalities within the development of the HBM GRF.

Equivalence tests for safety assessment of genetically modified crops using plant composition data.

The evaluation of compositional characteristics of plants harvested from field trials is an important step in the safety assessment of a genetically modified crop and its derived products for food and feed. The European Food Safety Authority (EFSA) evaluates safety by testing for equivalence between the GM genotype and other genotypes, typically with a history of safe use. Here, a new equivalence test is proposed, which addresses issues with the EFSA test. The method is motivated by a recently proposed equivalence test for analysis of data from animal feeding trials. In order to be suitable for practical safety assessment, the new method has a statistical power set to a desired value, e.g. 95%, by construction. In addition, we assess distributions rather than average values. This way, equivalence limits can also be established when there is limited genotypic variation. The original EFSA equivalence test breaks down in this case. The method is illustrated by its application to data from a field study on maize grain. Simulation studies indicate that the proposed test has appropriate performance characteristics and is competitive with respect to recently proposed alternatives, including the EFSA/EU equivalence test.

Methodology for health risk assessment of combined exposures to multiple chemicals.

Focus on risks to human health and the environment from combined exposure to multiple chemicals ("mixture risk assessment") has increased in the last couple of decades. There has been a rise in awareness and concern in the community, especially concerning unintentional environmental exposure to unknown chemical mixtures. The Horizon 2020 project EuroMix has developed methodologies and tools for mixture risk assessment with a focus on component-based approach where substances are grouped based on toxicological considerations. Dose addition is used as the model for calculating the combined toxicity of mixture components. The methodology is anchored in the Adverse Outcome Pathway (AOP) concept, which provides a structured basis for e.g. grouping substances into assessment groups and identifying and collecting relevant toxicity data. The aim of this paper is to describes development of the methodology for mixture risk assessment and to provide detailed methodology for problem formulation, use of AOP networks for development of tiered testing strategies and grouping of substances, as well as considerations for use of dose addition methodology.

A generic PBTK model implemented in the MCRA platform: Predictive performance and uses in risk assessment of chemicals.

Physiologically-based toxicokinetic (PBTK) models are important tools for in vitro to in vivo or inter-species extrapolations in health risk assessment of foodborne and non-foodborne chemicals. Here we present a generic PBTK model implemented in the EuroMix toolbox, MCRA 9 and predict internal kinetics of nine chemicals (three endocrine disrupters, three liver steatosis inducers, and three developmental toxicants), in data-rich and data-poor conditions, when increasingly complex levels of parametrization are applied. At the first stage, only QSAR models were used to determine substance-specific parameters, then some parameter values were refined by estimates from substance-specific or high-throughput in vitro experiments. At the last stage, elimination or absorption parameters were calibrated based on available in vivo kinetic data. The results illustrate that parametrization plays a capital role in the output of the PBTK model, as it can change how chemicals are prioritized based on internal concentration factors. In data-poor situations, estimates can be far from observed values. In many cases of chronic exposure, the PBTK model can be summarized by an external to internal dose factor, and interspecies concentration factors can be used to perform interspecies extrapolation. We finally discuss the implementation and use of the model in the MCRA risk assessment platform.

Correction to: Lack of adverse effects in subchronic and chronic toxicity/carcinogenicity studies on the glyphosate-resistant genetically modified maize NK603 in Wistar Han RCC rats.

In the original publication, the starting point in time for the three feeding trials.

The MCRA toolbox of models and data to support chemical mixture risk assessment.

A model and data toolbox is presented to assess risks from combined exposure to multiple chemicals using probabilistic methods. The Monte Carlo Risk Assessment (MCRA) toolbox, also known as the EuroMix toolbox, has more than 40 modules addressing all areas of risk assessment, and includes a data repository with data collected in the EuroMix project. This paper gives an introduction to the toolbox and illustrates its use with examples from the EuroMix project. The toolbox can be used for hazard identification, hazard characterisation, exposure assessment and risk characterisation. Examples for hazard identification are selection of substances relevant for a specific adverse outcome based on adverse outcome pathways and QSAR models. Examples for hazard characterisation are calculation of benchmark doses and relative potency factors with uncertainty from dose response data, and use of kinetic models to perform in vitro to in vivo extrapolation. Examples for exposure assessment are assessing cumulative exposure at external or internal level, where the latter option is needed when dietary and non-dietary routes have to be aggregated. Finally, risk characterisation is illustrated by calculation and display of the margin of exposure for single substances and for the cumulation, including uncertainties derived from exposure and hazard characterisation estimates.

A retain and refine approach to cumulative risk assessment.

Mixtures of substances to which humans are exposed may lead to cumulative exposure and health effects. To study their effects, it is first necessary to identify a cumulative assessment group (CAG) of substances for risk assessment or hazard testing. Excluding substances from consideration before there is sufficient evidence may underestimate the risk. Conversely, including everything and treating the inevitable uncertainties using conservative assumptions is inefficient and may overestimate the risk, with an unknown level of protection. An efficient, transparent strategy is described to retain a large group, quantifying the uncertainty of group membership and other uncertainties. Iterative refinement of the CAG then focuses on adding information for the substances with high probability of contributing significantly to the risk. Probabilities can be estimated using expert opinion or derived from data on substance properties. An example is presented with 100 pesticides, in which the retain step identified a single substance to target refinement. Using an updated hazard characterisation for this substance reduced the mean exposure estimate from 0.43 to 0.28 µg kg-bw-1 day-1 and reduced the 99.99th percentile exposure from 24.9 to 5.1 µg kg-bw-1 day-1. Other retained substances contributed little to the risk estimates, even after accounting for uncertainty.

Assessment of the combined nitrate and nitrite exposure from food and drinking water: application of uncertainty around the nitrate to nitrite conversion factor.

Dietary exposure to nitrate and nitrite occurs via three main sources; occurrence in (vegetable) foods, food additives in certain processed foods and contaminants in drinking water. While nitrate can be converted to nitrite in the human body, their risk assessment is usually based on single substance exposure in different regulatory frameworks. Here, we assessed the long-term combined exposure to nitrate and nitrite from food and drinking water. Dutch monitoring data (2012-2018) and EFSA data from 2017 were used for concentration data. These were combined with data from the Dutch food consumption survey (2012-2016) to assess exposure. A conversion factor (median 0.023; range 0.008-0.07) was used to express the nitrate exposure in nitrite equivalents which was added to the nitrite exposure. The uncertainty around the conversion factor was taken into account by using conversion factors randomly sampled from the abovementioned range. The combined dietary exposure was calculated for the Dutch population (1-79 years) with different exposure scenarios to address regional differences in nitrate and nitrite concentrations in drinking water. All scenarios resulted in a combined exposure above the acceptable daily intake for nitrite ion (70 µg/kg bw), with the mean exposure varying between 95-114 µg nitrite/kg bw/day in the different scenarios. Of all ages, the combined exposure was highest in children aged 1 year with an average of 250 µg nitrite/kg bw/day. Vegetables contributed most to the combined exposure in food in all scenarios, varying from 34%-41%. Food additive use contributed 8%-9% to the exposure and drinking water contributed 3%-19%. Our study is the first to perform a combined dietary exposure assessment of nitrate and nitrite while accounting for the uncertain conversion factor. Such a combined exposure assessment overarching different regulatory frameworks and using different scenarios for drinking water is a better instrument for protecting human health than single substance exposure.

Statement on advancing the assessment of chemical mixtures and their risks for human health and the environment.

The number of anthropogenic chemicals, manufactured, by-products, metabolites and abiotically formed transformation products, counts to hundreds of thousands, at present. Thus, humans and wildlife are exposed to complex mixtures, never one chemical at a time and rarely with only one dominating effect. Hence there is an urgent need to develop strategies on how exposure to multiple hazardous chemicals and the combination of their effects can be assessed. A workshop, "Advancing the Assessment of Chemical Mixtures and their Risks for Human Health and the Environment" was organized in May 2018 together with Joint Research Center in Ispra, EU-funded research projects and Commission Services and relevant EU agencies. This forum for researchers and policy-makers was created to discuss and identify gaps in risk assessment and governance of chemical mixtures as well as to discuss state of the art science and future research needs. Based on the presentations and discussions at this workshop we want to bring forward the following Key Messages.

Equivalence Testing Approaches in Genetically Modified Organism Risk Assessment.

Since 2011, the European Food Safety Authority (EFSA) has implemented combined difference and equivalence testing of agronomic, phenotypic, and composition data in the risk assessment of genetically modified crops. A short perspective is provided on misunderstandings that have shown up in published criticisms of the approach to equivalence testing, different viewpoints regarding the questions to be answered, and new developments in statistical modeling.

Omics analyses of potato plant materials using an improved one-class classification tool to identify aberrant compositional profiles in risk assessment procedures.

The objective of this study was to quantitatively assess potato omics profiles of new varieties for meaningful differences from analogous profiles of commercial varieties through the SIMCA one-class classification model. Analytical profiles of nine commercial potato varieties, eleven experimental potato varieties, one GM potato variety that had acquired Phytophtora resistance based on a single insert with potato-derived DNA sequences, and its non-GM commercial counterpart were generated. The ten conventional varieties were used to construct the one-class model. Omics profiles from experimental non-GM and GM varieties were assessed using the one-class SIMCA models. No potential unintended effects were identified in the case of the GM variety. The model showed that varieties that were genetically more distant from the commercial varieties were recognized as aberrant, highlighting its potential in determining whether additional evaluation is required for the risk assessment of materials produced from any breeding technique, including genetic modification.

Use of omics analytical methods in the study of genetically modified maize varieties tested in 90 days feeding trials.

Genetically modified (GM) maize and their non-modified counterparts were compared using MON810 varieties, the only GMO event cultivated in Europe. The differences in grain samples were analysed by omics profiles, including transcriptomics, proteomics and metabolomics. Other cultivated maize varieties were analysed as a reference for the variability that will exist between cultivated varieties. The observed differences between modified and non-modified maize varieties do not exceed typical differences between non-modified varieties. The use of these advanced analytical approaches to analyse novel plant materials as compared to the results from animal feeding trials with whole foods is assessed. No indications were observed for changes in the GM varieties that warrant further investigations. Furthermore, it was shown that such indications will be obtained if maize samples of inferior quality are analysed similarly. Omics data provide detailed analytical information of the plant material, which facilitates a risk assessment procedure of new (GM) plant varieties.

Equivalence analysis to support environmental safety assessment: Using nontarget organism count data from field trials with cisgenically modified potato.

This paper considers the statistical analysis of entomological count data from field experiments with genetically modified (GM) plants. Such trials are carried out to assess environmental safety. Potential effects on nontarget organisms (NTOs), as indicators of biodiversity, are investigated. The European Food Safety Authority (EFSA) gives broad guidance on the environmental risk assessment (ERA) of GM plants. Field experiments must contain suitable comparator crops as a benchmark for the assessment of designated endpoints. In this paper, a detailed protocol is proposed to perform data analysis for the purpose of assessing environmental safety. The protocol includes the specification of a list of endpoints and their hierarchical relations, the specification of intended levels of data analysis, and the specification of provisional limits of concern to decide on the need for further investigation. The protocol emphasizes a graphical representation of estimates and confidence intervals for the ratio of mean abundances for the GM plant and its comparator crop. Interpretation relies mainly on equivalence testing in which confidence intervals are compared with the limits of concern. The proposed methodology is illustrated with entomological count data resulting from multiyear, multilocation field trials. A cisgenically modified potato line (with enhanced resistance to late blight disease) was compared to the original conventional potato variety in the Netherlands and Ireland in two successive years (2013, 2014). It is shown that the protocol encompasses alternative schemes for safety assessment resulting from different research questions and/or expert choices. Graphical displays of equivalence testing at several hierarchical levels and their interpretation are presented for one of these schemes. The proposed approaches should be of help in the ERA of GM or other novel plants.

A probabilistic approach for risk-benefit assessment of food substitutions: A case study on substituting meat by fish.

Accounting for substitution of foods is inevitable when evaluating health impact of dietary changes. But substitution behavior and the associated health impact may vary between individuals. We therefore propose the use of probabilistic methods to model substitution and assess health impact distributions in risk-benefit assessment (RBA) of foods. We investigated the health impact of substituting red and processed meat with fish in the Danish adult population and the variability in health impact. We applied probabilistic approaches in modeling the substitution to reflect variability between individual substitution behaviors. Furthermore, when multiple intake scenarios are compared, we propose a method for adjusting intake differences for individual day-to-day variability. We estimated that 134 (95% UI: 102; 169) Disability-Adjusted Life Years/100,000 were averted per year by the substitution. The health impact varied considerably by age and sex, with the largest health benefit of the substitution observed for young women in the child-bearing age and for the older generation, mainly men. This study provides further insight in how the health impact of substituting meat by fish varies between individuals and suggests a framework to be applied in RBAs of other food substitutions. Our results are relevant for policy makers in defining targeted public health strategies.

Equivalence limit scaled differences for untargeted safety assessments: Comparative analyses to guard against unintended effects on the environment or human health of genetically modified maize.

Safety assessments guard against unintended effects for human health and the environment. When new products are compared with accepted reference products by broad arrays of measurements, statistical analyses are usually summarised by significance tests or confidence intervals per endpoint. The traditional approach is to test for statistical significance of differences. However, absence or presence of significant differences is not a statement about safety. Equivalence limits are essential for safety assessment. We propose graphs to present the results of equivalence tests over the array of endpoints. It is argued that plots of the equivalence limit scaled difference (ELSD) are preferable over plots of the standardised effect size (SES) used previously for similar assessments. The ELSD method can be used either with externally specified equivalence limits or with equivalence limits estimated from (historical) data. The method is illustrated with two examples: first, environmental safety of MON810 Bt maize was assessed using field trial count data of arthropods; second, human safety of herbicide tolerant NK603 maize was assessed using haematological, biochemical and organ weight data from a 90-day rat feeding study. All assessed endpoints were classified in EFSA equivalence categories I or II, implying full equivalence or equivalence more likely than not.