Search

As most ecosystems are being challenged by multiple, co‐occurring stressors, an important challenge is to understand and predict how stressors interact to affect biological responses. A popular approach is to design factorial experiments that measure biological responses to pairs of stressors and compare the observed response to a null model expectation. Unfortunately, we believe experiment sample sizes are inadequate to detect most non‐null stressor interaction responses, greatly hindering progress. Using both real and simulated data, we show sample sizes typical of many experiments (<6) can (i) only detect very large deviations from the additive null model, implying many important non‐null stressor‐pair interactions are being missed, and (ii) potentially lead to mostly statistical outliers being reported. Computer code that simulates data under either additive or multiplicative null models is provided to estimate statistical power for user‐defined responses and sample sizes, and we recommend this is used to aid experimental design and interpretation of results. We suspect that most experiments may require 20 or more replicates per treatment to have adequate power to detect nonadditive. However, estimates of power need to be made while considering the smallest interaction of interest, i.e., the lower limit for a biologically important interaction, which is likely to be system‐specific, meaning a general guide is unavailable. We discuss ways in which the smallest interaction of interest can be chosen, and how sample sizes can be increased. Our main analyses relate to the additive null model, but we show similar problems occur for the multiplicative null model, and we encourage similar investigations into the statistical power of other null models and inference methods. Without knowledge of the detection abilities of the statistical tools at hand or the definition of the smallest meaningful interaction, we will undoubtedly continue to miss important ecosystem stressor interactions. [ABSTRACT FROM AUTHOR]

The benchmark dose (BMD) concept is increasingly utilized to analyze quantitative dose-response relationships in genetic toxicology. This methodology requires the user (i.e. the toxicologist) to a priori define a small increase over controls that is 'acceptable' to be induced by a genotoxic test substance. The increase is called benchmark response (BMR) or critical effect size (CES), depending on the software used. To render the metrics calculated from the data of animals treated with the test substance applicable for risk assessment, the BMR or CES must represent biologically relevant changes of parameters measured in in vivo genotoxicity assays such as the Micronucleus, Comet, Transgenic rodent or Pig-a assay. Current recommendations for CES in genotoxicology are arbitrary (10% increase over mean vehicle controls) or based on limited, usually 5-6, data points (i.e. the standard deviation of the concurrent vehicle control group). We have, therefore, analyzed historical vehicle control data of standard in vivo genotoxicity test systems with statistical methods. Based on this evaluation, we illustrate limitations of the currently recommended CES values and propose a pragmatic approach that may contribute to better defining endpoint-specific CES values for BMD software like PROAST. [ABSTRACT FROM AUTHOR]

A general theory on effect size for continuous data predicts a relationship between maximum response and within-group variation of biological parameters, which is empirically confirmed by results from dose–response analyses of 27 different biological parameters. The theory shows how effect sizes observed in distinct biological parameters can be compared and provides a basis for a generic definition of small, intermediate and large effects. While the theory is useful for experimental science in general, it has specific consequences for risk assessment: it solves the current debate on the appropriate metric for the Benchmark response in continuous data. The theory shows that scaling the BMR expressed as a percent change in means to the maximum response (in the way specified) automatically takes “natural variability” into account. Thus, the theory supports the underlying rationale of the BMR 1 SD. For various reasons, it is, however, recommended to use a BMR in terms of a percent change that is scaled to maximum response and/or within group variation (averaged over studies), as a single harmonized approach. [ABSTRACT FROM AUTHOR]

Normal ranges are some fraction of a reference distribution deemed to represent an expected condition, typically 95%. They are frequently used as the basis for generic criteria for monitoring programs designed to test whether a sample is outside of "normal," as in reference-condition approach studies. Normal ranges are also the basis for criteria for more classic environmental effects monitoring programs designed to detect differences in mean responses between reference and exposure areas. Limits on normal ranges are estimated with error that varies depending largely on sample size. Direct comparison of a sample or a mean to estimated limits of a normal range will, with some frequency, lead to incorrect conclusions about whether a sample or a mean is inside or outside the normal range when the sample or the mean is near the limit. Those errors can have significant costs and risk implications. This article describes tests based on noncentral distributions that are appropriate for quantifying the likelihood that samples or means are outside a normal range. These noncentral tests reverse the burden of evidence (assuming that the sample or mean is at or outside normal), and thereby encourage proponents to collect more robust sample sizes that will demonstrate that the sample or mean is not at the limits or beyond the normal range. These noncentral equivalence and interval tests can be applied to uni- and multivariate responses, and to simple (e.g., upstream vs downstream) or more complex (e.g., before vs after, or upstream vs downstream) study designs. Statistical procedures for the various tests are illustrated with benthic invertebrate community data collected as part of the Regional Aquatics Monitoring Program (RAMP) in the vicinity of oil sands operations in northern Alberta, Canada. An Excel workbook with functions and calculations to carry out the various tests is provided in the online Supplemental Data. [ABSTRACT FROM AUTHOR]

Potential adverse effects of chemical substances on thyroid function are usually examined by measuring serum levels of thyroid-related hormones. Instead, recent risk assessments for thyroid-active chemicals have focussed on iodine uptake inhibition, an upstream event that by itself is not necessarily adverse. Establishing the extent of uptake inhibition that can be considered de minimis, the chosen benchmark response (BMR), is therefore critical. The BMR values selected by two international advisory bodies were 5% and 50%, a difference that had correspondingly large impacts on the estimated risks and health-based guidance values that were established. Potential treatment-related inhibition of thyroidal iodine uptake is usually determined by comparing thyroidal uptake of radioactive iodine (RAIU) during treatment with a single pre-treatment RAIU value. In the present study it is demonstrated that the physiological intra-individual variation in iodine uptake is much larger than 5%. Consequently, in-treatment RAIU values, expressed as a percentage of the pre-treatment value, have an inherent variation, that needs to be considered when conducting dose-response analyses. Based on statistical and biological considerations, a BMR of 20% is proposed for benchmark dose analysis of human thyroidal iodine uptake data, to take the inherent variation in relative RAIU data into account. Implications for the tolerated daily intakes for perchlorate and chlorate, recently established by the European Food Safety Authority (EFSA), are discussed. [ABSTRACT FROM AUTHOR]

Abstract: Assuming that temporal fluctuations in physiological parameters (e.g. haematology, biochemistry) in individual healthy non-exposed animals are non-adverse, the minimal magnitude of the Critical Effect Size (CES) for a number of continuous parameters of toxicity studies was derived. A total of 36 studies (19 pharmaceutical preclinical studies in dogs and 17 chemical risk assessment studies in rats) were analysed to determine within-animal variation in their control groups. Minimal CES-values were derived for each group of studies, differentiating where necessary between strains and sexes, using the 2.5 percentile (lower limit) and/or 97.5 percentile (upper limit) of the distribution of the within-animal variation around the mean of each parameter. We concluded that minimal CES-values for continuous clinical chemistry and haematology parameters should be established separately per species, strain, sex and study duration investigated. Grouping of minimal CES-values, leading to more or less “general” values, seems possible for those parameters that are subject to tight homeostatic control and consequently show little within-animal variation. Nearly a quarter of the proposed CES-values is ⩽5%, nearly a quarter range from 6% to 10%, a quarter is 15% or 20%, and nearly 30% of the proposed values is ⩾20% of the mean of the control animals. [Copyright &y& Elsevier]

The effective design of field studies requires that sample size requirements be estimated for important endpoints before conducting assessments. This a priori calculation of sample size requires initial estimates for the variability of the endpoints of interest, decisions regarding significance levels and the power desired, and identification of an effect size to be detected. Although many programs have called for use of critical effect sizes (CES) in the design of monitoring programs, few attempts have been made to define them. This paper reviews approaches that have been or could be used to set specific CES. The ideal method for setting CES would be to define the level of protection that prevents ecologically relevant impacts and to set a warning level of change that would be more sensitive than that CES level to provide a margin of safety; however, few examples of this approach being applied exist. Program-specific CES could be developed through the use of numbers based on regulatory or detection limits, a number defined through stakeholder negotiation, estimates of the ranges of reference data, or calculation from the distribution of data using frequency plots or multivariate techniques. The CES that have been defined often are consistent with a CES of approximately 25%, or two standard deviations, for many biological or ecological monitoring endpoints, and this value appears to be reasonable for use in a wide variety of monitoring programs and with a wide variety of endpoints. [ABSTRACT FROM AUTHOR]

In this study, the within-animal variation in routinely studied continuous toxicological parameters was estimated from temporal fluctuations in individual healthy nonexposed animals. Assuming that these fluctuations are nonadverse, this within-animal variation may be indicative of the minimal magnitude of the critical effect size (CES). The CES is defined as the breaking point between adverse and nonadverse changes in a continuous toxicological parameter, at the level of the individual organism. The total variation in the data from individual nonexposed animals was divided in variation parts due to known factors (differences in sex, animal, and day) and a residual variation, by means of analysis of variance. Using the residual variation and the estimated analytical measurement error of a toxicological parameter, the within-animal variation can be estimated. The data showed within-animal variations ranging between 0.6% and 34% for different clinical chemistry and hematological parameters in 90-day rat studies. This indicates that different (minimal) CES values may be applicable for different parameters. [ABSTRACT FROM AUTHOR]

A family of (nested) dose-response models is introduced herein that can be used for describing the change in any continuous endpoint as a function of dose. A member from this family of models may be selected using the likelihood ratio test as a criterion, to prevent overparameterization. The proposed methodology provides for a formal approach of model selection, and a transparent way of assessing the benchmark dose. Apart from a number of natural constraints, the model expressions follow from an obvious way of quantifying differences in sensitivity between populations. As a consequence, dose-response data that relate to both sexes can be efficiently analyzed by incorporating the data from both sexes in the same analysis, even if the sexes are not equally sensitive to the compound studied. The idea of differences in sensitivity is closely related to the assessment factors used in risk assessment. Thus, the models are directly applicable to estimating such factors, if data concerning populations to be compared are available. Such information is valuable for further validation or adjustment of default assessment factors, as well as for informing distributional assessment factors in a probabilistic risk assessment. The various applications of the proposed methodology are illustrated by real data sets. [ABSTRACT FROM PUBLISHER]

It is argued that dose-thresholds in a strict quantitative sense cannot exist in dose-response relationships that are studied by toxicologists. The biological, more qualitative, notion of thresholds can only be translated into quantitative terms by defining thresholds in the effect-size of continuous end-points: What size of an effect can the organism cope with and when is the effect adverse? These are key questions that toxicologists should address, for the various (continuous) end-points that may be observed in toxicological studies. Avoiding this question by resorting to a “formal” statistical test is deceptive and makes the outcome of a risk assessment fortuitous. There is no defendable rationale for different procedures of data analysis based upon assumed underlying mechanisms. The approach of data analysis should be determined by the nature and quality of the data. If the data allow for dose-response modelling, this should be done whatever the underlying biological mechanism. Quantal dose-response relationships (relating to noncancer end-points) are problematic, because their slopes are largely determined by the experimental noise. Therefore, the concept of “extra risk” is not well defined in this case. In the case of tumor incidence data, on the other hand, the definition and estimation of an extra risk level are defendable, if it can be assumed that the stochastic component of the carcinogenicity process itself is large compared to the interindividual variation and experimental noise. [ABSTRACT FROM PUBLISHER]

The use of uncertainty factors in the standard method for deriving acceptable intake or exposure limits for humans, such as the ReferenceDose (RfD), may be viewed as a conservative method of taking variousuncertainties into account. As an obvious alternative, the use of uncertainty distributions instead of uncertainty factors is gaining attention. This paper presents a comprehensive discussion of a general framework that quantifies both the uncertainties in the no-adverse-effect level in the animal (using a benchmark-like approach) and the uncertainties in the various extrapolation steps involved (using uncertainty distributions). This approach results in an uncertainty distribution for the no-adverse-effect level in the sensitive human subpopulation, reflecting the overall scientific uncertainty associated with that level. A lower percentile of this distribution may be regarded asan acceptable exposure limit (e.g., RfD) that takes account of the various uncertainties in a nonconservative fashion. The same methodology may also be used as a tool to derive a distribution for possible human health effects at a given exposure level. We argue that in a probabilistic approach the uncertainty in the estimated no-adverse-effect-level in the animal should be explicitly taken into account. Not only is this source of uncertainty too large to be ignored, it also hasrepercussions for the quantification of the other uncertainty distributions. [ABSTRACT FROM AUTHOR]

The article offers the author's insights on the significance of employing good statistical approaches in making decisions on policies in plant science. The author mentions the drawbacks of using null hypothesis significance testing (NHST), which are based on using arbitrary threshold for evaluating statistical significance. He argues that NHST's significance threshold must be established to attain the objectives of the statistical test.

Community-scale impacts of glyphosate-based herbicides on wetland plant communities and the magnitude of those impacts that should be considered biologically relevant are poorly understood. We contrast three different thresholds for setting biologically meaningful critical effect sizes for complex ANOVA study designs. We use each of the of the critical effect sizes to determine optimal α levels for assessment of how different concentrations of glyphosate-based herbicides affect wetland plant communities over two years of herbicide application (alone and in combination with agricultural fertilizers) and two subsequent years without herbicide (or fertilizer) application. The application of glyphosate-based herbicides was found to result in a decrease in macrophyte species richness, an increase in macrophyte species evenness, a decrease in macrophyte cover and a reduction in community similarity. There was little evidence that nutrient additions directly or indirectly affected plant community endpoints. The glyphosate effects were evident in the first year of herbicide application in 2009, and became more pronounced in the second year of herbicide application in 2010. However, when herbicides were not applied in 2011, recovery was observed in most endpoints, with the exception being species evenness, for which partial recovery was not observed until 2012. Optimal α levels differed among the three critical effect sizes for each ANOVA term and endpoint combination, however regardless of differences in α levels, conclusions were generally consistent across all critical effect sizes. Image 1 • Glyphosate-based herbicides reduced macrophyte cover and species richness. • Herbicide application increased species evenness and reduced community similarity. • Cover and diversity recovered in 2011, the year after herbicide applications. • Our conclusions were consistent across different target effect sizes. [ABSTRACT FROM AUTHOR]

Endocrine-disrupting chemicals currently represent one of the major concerns and this study was aimed to investigate the effects of different doses of cadmium, widespread toxic metal, on the levels of thyroid hormones and to calculate Benchmark doses for these effects. Furthermore, the effects of co-exposure to cadmium and polychlorinated biphenyls on thyroid function were investigated. Six orally-treated groups of rats were receiving 0.3, 0.6, 1.25, 2.5, 5 and 10mgCd/kgb.w./day, five groups were orally treated with 0.5, 1, 2, 4 and 8mgPCBs/kgb.w./day, while nine groups of rats were orally-treated with different dose combinations of Cd and PCBs (0.6, 1.25 and 2.5mgCd/kgb.w. and 2, 4 and 8mgPCBs/kgb.w./day), during 28 days. Thyroid hormones were adversely affected by cadmium, with most prominent effect observed on triiodothyroxine levels indicating Cd interference with thyroid function at extrathyroidal level. Calculated Benchmark doses for Cd effects on thyroid hormones indicate triiodothyroxine as the most sensitive one that can be used as a basis for risk assessment. This study also implicates possible synergistic effects of Cd and PCBs on thyroid function as a consequence of their interference at different levels of thyroid homeostasis., (Copyright © 2013 Elsevier Ireland Ltd. All rights reserved.)