The Use of Epidemiology in Risk Assessment: Challenges and Opportunities

The assessment of risk from environmental and occupational exposures incorporates and synthesizes data from a variety of scientific disciplines including toxicology and epidemiology. Epidemiological data have offered valuable contributions to the identification of human health hazards, estimation of human exposures, quantification of the exposure–response relation, and characterization of risks to specific target populations including sensitive populations. As with any scientific discipline, there are some uncertainties inherent in these data; however, the best human health risk assessments utilize all available information, characterizing strengths and limitations as appropriate. Human health risk assessors evaluating environmental and occupational exposures have raised concerns about the validity of using epidemiological data for risk assessment due to actual or perceived study limitations. This article highlights three concerns commonly raised during the development of human health risk assessments of environmental and occupational exposures: (a) error in the measurement of exposure, (b) potential confounding, and (c) the interpretation of non-linear or non-monotonic exposure–response data. These issues are often the content of scientific disagreement and debate among the human health risk assessment community, and we explore how these concerns may be contextualized, addressed, and often ameliorated.     

Carcinogenic risk assessment: some points of interest for a discussion

This paper is aimed at providing some elements of discussion about points that emerge about the meaning of risk assessment procedures. The considerations that will be developed mainly pertain to four topics. First, the feasibility of objectively checking the validity of risk assessment results on the basis of epidemiological data, particularly for very low exposures. Second, the influence of uncertainty factors and the resulting variations in risk estimates. Third, the influence of individual susceptibility factors, easily entailing differences in risk in the order of three levels of magnitude. These differences, as a consequence of genetic factors, are often in the range of three orders of negritude and are more evident for low classes of exposure. In the general population, DNA repair related susceptibility should also be considered. Fourth, the mathematical model of the risk assessment procedure may itself influence the results. The default approach adopted by the US EPA is commented.     

Uncertainty in cancer risk at low doses of inorganic arsenic

It is unknown whether inorganic arsenic in drinking water concentrations at the current maximum contaminant level of 50 μg/1 poses a cancer risk in the United States. Data from two large epidemiological studies of cancer and arsenic in drinking water in Taiwan indicate a dose‐response relationship, but the magnitude of risk at low concentrations is highly uncertain. Four sources of uncertainty are described: model choice, data aggregation, intra‐village variability of arsenic in well water, arsenic intake from food. New data from an appropriately designed epidemiological study are needed to improve dose‐response assessment.     

Epidemiology, cancer genetics and microarrays: making correct inferences, using appropriate designs

Different disciplines approach cancer with different study designs, techniques and established bodies of knowledge. This article identifies some established epidemiological data and methods, which are useful for cross-disciplinary molecular and genetic studies of cancer but which are ignored by some researchers. First, the international variation in cancer risk is accounted for extensively by variation in environmental exposures; it is unlikely that even minute characterization of individual genomes will provide the best assessment of cancer risk in the absence of comparably detailed information on the environment. Second, epidemiological study-design methods are sometimes the most appropriate to answer molecular questions, particularly when using techniques such as expression microarrays or proteomics to establish differences among cancer subtypes or biomarkers in the setting of a non-experimental study. In such studies, it is essential to avoid bias, control confounding and undertake accurate replication. Established epidemiological data and methods will contribute to the best use of the new molecular technology.     

Risk Assessment of Rotavirus Infection in Surface Seawater from Bohai Bay,China

Dose-response data indicate that rotavirus (RV) may be one of the more infective agents among enteric viruses. The major limitation at present in the assessment of infection from rotavirus is lack of quantitative data on viral infectivity. In this work, an integrated cell culture and real-time quantitative polymerase chain reaction (ICC-qPCR) method and a Beta-Poisson model for risk assessment were employed. A set of 28 surface seawater samples was collected from December 2010 to September 2011 in Bohai Bay, China, to enable a seasonal risk assessment of infective RV at recreational beaches. Thirty-two percent of the samples were positive for rotavirus, and the estimated concentration range of infectious human rotavirus was 1 to 279 PFU/L. We further confirmed that the contamination of seawater with rotavirus was higher in autumn and winter, which was in reasonable agreement with the trend observed in a prior epidemiological study. Our preliminary risk assessment indicated the daily risk of illness at almost all the contaminated sites exceeded an acceptable threshold of marine recreational water quality (19 illnesses per 1000 swimmers). The detection method and dose-response model in the current work appear useful for evaluating pathogenic risks of seawater to vacationers and can inform management actions.     

Risk assessment of neurotoxic pesticides

In order to establish safe exposure levels for toxic chemicals, risk assessment guidelines have been developed. A compilation is given by the author on the elements of risk assessment of hazardous neurotoxic pesticides, using data obtained from human epidemiological studies, from animal experiments, from the international literature and from the author's own experiments as well. Well-controlled laboratory studies of neurotoxicity have the potential to provide adequate exposure and effect data for accurate hazard identification. Animal models of neurotoxicity as highly sensitive behavioral and neurophysiological methods as a function of doses, provide data for human low dose extrapolation by using mathematical models. This procedure might be the basis for reducing risk ("risk management"), therefore some examples are given, how to handle properly neurotoxic pesticides with different- high or low-risk.     

Quantitative health impact of indoor radon in France

Radon is the second leading cause of lung cancer after smoking. Since the previous quantitative risk assessment of indoor radon conducted in France, input data have changed such as, estimates of indoor radon concentrations, lung cancer rates and the prevalence of tobacco consumption. The aim of this work was to update the risk assessment of lung cancer mortality attributable to indoor radon in France using recent risk models and data, improving the consideration of smoking, and providing results at a fine geographical scale. The data used were population data (2012), vital statistics on death from lung cancer (2008–2012), domestic radon exposure from a recent database that combines measurement results of indoor radon concentration and the geogenic radon potential map for France (2015), and smoking prevalence (2010). The risk model used was derived from a European epidemiological study, considering that lung cancer risk increased by 16% per 100 becquerels per cubic meter (Bq/m³) indoor radon concentration. The estimated number of lung cancer deaths attributable to indoor radon exposure is about 3000 (1000; 5000), which corresponds to about 10% of all lung cancer deaths each year in France. About 33% of lung cancer deaths attributable to radon are due to exposure levels above 100 Bq/m³. Considering the combined effect of tobacco and radon, the study shows that 75% of estimated radon-attributable lung cancer deaths occur among current smokers, 20% among ex-smokers and 5% among never-smokers. It is concluded that the results of this study, which are based on precise estimates of indoor radon concentrations at finest geographical scale, can serve as a basis for defining French policy against radon risk.     

Epidemiology and reporting characteristics of preclinical systematic reviews

In an effort to better utilize published evidence obtained from animal experiments, systematic reviews of preclinical studies are increasingly more common—along with the methods and tools to appraise them (e.g., SYstematic Review Center for Laboratory animal Experimentation [SYRCLE’s] risk of bias tool). We performed a cross-sectional study of a sample of recent preclinical systematic reviews (2015–2018) and examined a range of epidemiological characteristics and used a 46-item checklist to assess reporting details. We identified 442 reviews published across 43 countries in 23 different disease domains that used 26 animal species. Reporting of key details to ensure transparency and reproducibility was inconsistent across reviews and within article sections. Items were most completely reported in the title, introduction, and results sections of the reviews, while least reported in the methods and discussion sections. Less than half of reviews reported that a risk of bias assessment for internal and external validity was undertaken, and none reported methods for evaluating construct validity. Our results demonstrate that a considerable number of preclinical systematic reviews investigating diverse topics have been conducted; however, their quality of reporting is inconsistent. Our study provides the justification and evidence to inform the development of guidelines for conducting and reporting preclinical systematic reviews.

A cross sectional study of a sample of recent preclinical systematic reviews reveals deficiencies in reporting and provides the justification and evidence to inform the development of specific guidelines for conducting and reporting preclinical systematic reviews.     

A Conceptual Framework for Evaluating the Interaction of a Chemical and Nonchemical Stressor in Human Health Risk Assessments: A Case Study for Lead and Psychosocial Stress

Recent research has demonstrated that nonchemical stressors may alter the toxicity from chemical exposures. This may have public health implications for low socioeconomic status (SES) communities that may be disproportionately exposed to toxic chemicals and various types of community and personal stressors. Nonchemical stressors may introduce an important source of variability that needs to be considered by risk assessors. Herein, we propose a framework for determining if a chemical–nonchemical interaction exists and, if so, options for incorporating interaction information into risk assessments. We use the increasingly recognized interaction between lead and psychosocial stress to illustrate the framework. We found that lead exposure occurs disproportionately in low SES groups that also tend to face high levels of psychosocial stress; that stress and lead both affect neurodevelopment and that this occurs via similar pathways involving the hypothalamic-pituitary axis. Further, several epidemiological and experimental studies have provided evidence for an interaction between lead and psychosocial stress. The implications of this interaction for risk assessment are also discussed.     

Appropriate use of animal models in the assessment of risk during prenatal development: an illustration using inorganic arsenic

BACKGROUND: Assessing risks to human development from chemical exposure typically requires integrating findings from laboratory animal and human studies. METHODS: Using a case study approach, we present a program designed to assess the risk of the occurrence of malformations from inorganic arsenic exposure. We discuss how epidemiological data should be evaluated for quality and criteria for determining whether an association is causal. In this case study, adequate epidemiological data were not available for evaluating the potential effect of arsenic on development. Consequently, results from appropriately designed, conducted, and interpreted developmental toxicity studies, which have been shown to be predictive of human risk under numerous scenarios, were used. In our case study, the existing animal data were not designed appropriately to assess risk from environmental exposures, although such studies may be useful for hazard identification. Because the human and animal databases were deficient, a research program comprising modern guideline toxicological studies was designed and conducted. RESULTS: The results of those studies in rats, mice, and rabbits indicate that oral and inhalational exposures to inorganic arsenic do not cause structural malformations, and inhalational exposures produced no developmental effects at all. The new study results are discussed in conjunction with considerations of metabolism, toxicokinetics, and maternal toxicity. CONCLUSIONS: Based on the available experimental data, and absent contrary findings from adequately conducted epidemiological studies, we conclude that exposure to inorganic arsenic by environmentally relevant routes poses no risk of the occurrence of malformations and little risk of other prenatal developmental toxicity in developing humans without concomitant and near-lethal toxicological effects in mothers.     

International Commission for Protection against Environmental Mutagens and Carcinogens. ICPEMC publication No. 13. The need for biological risk assessment in reaching decisions about carcinogens

The prudent assumption that carcinogen bioassays in rodents predict for human carcinogenicity is examined. It is suggested that in certain cases, as for example the induction of tumors against a high incidence in controls, or in situations in which high dose toxicity may be a critical factor in the induction of cancer, the probability that animal bioassays predict for humans may be low. The term 'biological risk assessment' is introduced to describe that part of risk assessment concerned with the relevance of specific animal results to the induction of human cancer. Biological risk assessment, which is almost entirely dependent on an understanding of carcinogenesis mechanisms, is an important addition to present mathematical modeling used to predict the effects of animal carcinogens that have been demonstrated after high dose exposure, to the effects of the much smaller doses to which humans are perceived to be exposed. Evidence for the conclusions reached by biological risk assessment may sometimes be supported by a careful review of human epidemiological data.     

Dietary cholesterol and atherosclerosis

The perceived relationship between dietary cholesterol, plasma cholesterol and atherosclerosis is based on three lines of evidence: animal feeding studies, epidemiological surveys, and clinical trials. Over the past quarter century studies investigating the relationship between dietary cholesterol and atherosclerosis have raised questions regarding the contribution of dietary cholesterol to heart disease risk and the validity of dietary cholesterol restrictions based on these lines of evidence. Animal feeding studies have shown that for most species large doses of cholesterol are necessary to induce hypercholesterolemia and atherosclerosis, while for other species even small cholesterol intakes induce hypercholesterolemia. The species-to-species variability in the plasma cholesterol response to dietary cholesterol, and the distinctly different plasma lipoprotein profiles of most animal models make extrapolation of the data from animal feeding studies to human health extremely complicated and difficult to interpret. Epidemiological surveys often report positive relationships between cholesterol intakes and cardiovascular disease based on simple regression analyses; however, when multiple regression analyses account for the colinearity of dietary cholesterol and saturated fat calories, there is a null relationship between dietary cholesterol and coronary heart disease morbidity and mortality. An additional complication of epidemiological survey data is that dietary patterns high in animal products are often low in grains, fruits and vegetables which can contribute to increased risk of atherosclerosis. Clinical feeding studies show that a 100 mg/day change in dietary cholesterol will on average change the plasma total cholesterol level by 2.2-2.5 mg/dl, with a 1.9 mg/dl change in low density lipoprotein (LDL) cholesterol and a 0.4 mg/dl change in high density lipoprotein (HDL) cholesterol. Data indicate that dietary cholesterol has little effect on the plasma LDL:HDL ratio. Analysis of the available epidemiological and clinical data indicates that for the general population, dietary cholesterol makes no significant contribution to atherosclerosis and risk of cardiovascular disease.     

Data warehouse for assessing animal health,welfare, risk management and –communication

The objective of this paper is to give an overview of existing databases in Denmark and describe some of the most important of these in relation to establishment of the Danish Veterinary and Food Administrations’ veterinary data warehouse. The purpose of the data warehouse and possible use of the data are described. Finally, sharing of data and validity of data is discussed. There are databases in other countries describing animal husbandry and veterinary antimicrobial consumption, but Denmark will be the first country relating all data concerning animal husbandry, -health and -welfare in Danish production animals to each other in a data warehouse. Moreover, creating access to these data for researchers and authorities will hopefully result in easier and more substantial risk based control, risk management and risk communication by the authorities and access to data for researchers for epidemiological studies in animal health and welfare.     

Validity of a self-administered food frequency questionnaire in the assessment of heterocyclic amine intake using 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) levels in hair

Several case-control studies have reported possible associations between heterocyclic amine (HCA) intake and the risk of cancer. The validity of questionnaires used to assess HCA intake has hardly been examined, however; in particular, no biomarker able to serve as an independent measure of habitual HCA intake has been established. In this study, we examined the validity of HCA intake estimated from a food frequency questionnaire (FFQ) using 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) level in hair as a reference method. Study subjects were 20 volunteers (7 men and 13 women) aged 25-57 years residing in Tokyo or neighboring cities in Japan. The subjects completed the FFQ, and gave 3-5g of hair twice at an interval of 1-3 months for use in establishing validity. Results showed that intakes of PhIP, MeIQ, Trp-P-1, and total HCA by the FFQ were significantly correlated with PhIP levels in hair when adjustment was made for melanin content (r=0.47, r=0.50, r=0.55, and r=0.51, respectively). The present study indicates that HCA intake estimated from this FFQ provides a reasonable ranking of individuals to allow the analysis of associations between HCA intake and risk of cancer in large-scale epidemiological studies.     

Biomarkers in nutritional epidemiology: applications, needs and new horizons

Modern epidemiology suggests a potential interactive association between diet, lifestyle, genetics and the risk of many chronic diseases. As such, many epidemiologic studies attempt to consider assessment of dietary intake alongside genetic measures and other variables of interest. However, given the multi-factorial complexities of dietary exposures, all dietary intake assessment methods are associated with measurement errors which affect dietary estimates and may obscure disease risk associations. For this reason, dietary biomarkers measured in biological specimens are being increasingly used as additional or substitute estimates of dietary intake and nutrient status. Genetic variation may influence dietary intake and nutrient metabolism and may affect the utility of a dietary biomarker to properly reflect dietary exposures. Although there are many functional dietary biomarkers that, if utilized appropriately, can be very informative, a better understanding of the interactions between diet and genes as potentially determining factors in the validity, application and interpretation of dietary biomarkers is necessary. It is the aim of this review to highlight how some important biomarkers are being applied in nutrition epidemiology and to address some associated questions and limitations. This review also emphasizes the need to identify new dietary biomarkers and highlights the emerging field of nutritional metabonomics as an analytical method to assess metabolic profiles as measures of dietary exposures and indicators of dietary patterns, dietary changes or effectiveness of dietary interventions. The review will also touch upon new statistical methodologies for the combination of dietary questionnaire and biomarker data for disease risk assessment. It is clear that dietary biomarkers require much further research in order to be better applied and interpreted. Future priorities should be to integrate high quality dietary intake information, measurements of dietary biomarkers, metabolic profiles of specific dietary patterns, genetics and novel statistical methodology in order to provide important new insights into gene-diet-lifestyle-disease risk associations.     

Probabilistic Analysis of Human Health Risks Associated with Background Concentrations of Inorganic Arsenic: Use of a Margin of Exposure Approach

Substantial evidence exists from epidemiological and mechanistic studies supporting a sublinear or threshold dose–response relationship for the carcinogenicity of ingested arsenic; nonetheless, current regulatory agency evaluations have quantified arsenic risks using default, generic risk assessment procedures that assume a linear, no-threshold dose–response relationship. The resulting slope factors predict risks from U.S. background arsenic exposures that exceed certain regulatory levels of concern, an outcome that presents challenges for risk communication and risk management decisions. To better reflect the available scientific evidence, this article presents the results of a Margin of Exposure (MOE) analysis to characterize risks associated with typical and high-end background exposures of the U.S. population to arsenic from food, water, and soil. MOE values were calculated by comparing a no-observable-adverse-effect-level (NOAEL) derived from the epidemiological literature with exposure estimates generated using a probabilistic (Monte Carlo) model. The plausibility and conservative nature of the exposure and risk estimates evaluated in this analysis are supported by sensitivity and uncertainty analyses and by comparing predicted urinary arsenic concentrations with empirical data. Using the more scientifically supported MOE approach, the analysis presented in this article indicates that typical and high-end background exposures to inorganic arsenic in U.S. populations do not present elevated risks of carcinogenicity.     

Attributable risk function in the proportional hazards model for censored time-to-event

Time-to-event endpoints are often used in clinical and epidemiological studies to evaluate disease association with hazardous exposures. In the statistical literature of time-to-event analysis, such association is usually measured by the hazard ratio in the proportional hazards model. In public health, it is also of important interest to assess the excess risk attributable to an exposure in a given population. In this article, we extend the notion of 'population attributable fraction' for the binary outcomes to the attributable risk function for the event times in prospective studies. A simple estimator of the time-varying attributable risk function is proposed under the proportional hazards model. Its inference procedures are established. Monte-Carlo simulation studies are conducted to evaluate its validity and performance. The proposed methodology is motivated and demonstrated by the data collected in a multicenter acquired immunodeficiency syndrome (AIDS) cohort study to estimate the attributable risk of human immunodeficiency virus type 1 (HIV-1) infections due to several potential risk factors.     

The potential role of life cycle assessment in regulation of chemicals in the European union

Scope and Background  This paper presents the preliminary results from an ongoing feasibility study, investigating potential application of elements from the life cycle assessment (LCA) framework in European chemicals’ policy. Many policy areas affect manufacturing, marketing and use of chemicals. This article focuses on the general chemical legislation, especially issues related to regulatory risk assessment and subsequent decisions on risk reduction measures. Method  Current and upcoming chemical regulation has been reviewed and empirical knowledge has been gained from an ongoing case study and from dialogues with various stakeholders. Results and Discussion  LCAs are comparative and more holistic in view as compared to chemical risk assessments for regulatory purposes¹. LCAs may therefore potentially improve the basis for decisions between alternatives in cases where a risk assessment calls for risk reduction. In this process, LCA results might feed into a socio-economic analysis having similar objectives, but some methodological aspects related to system boundaries need to be sorted out. Life cycle impact assessment (LCIA) of toxic effects has traditionally been inspired by the more regulatory-orientated risk assessment approaches. However, the increasing need for regulatory priority setting and comparative/ cumulative assessments might in the future convey LCIA principles into the regulatory framework. The same underlying databases on inherent properties of chemicals are already applied in both types of assessment. Similarly, data on the use and exposure of chemicals are needed within both risk assessments and LCA, and the methodologies might therefore benefit from a joint ‘inventory’ database. Outlook  The final outcome of the feasibility study will be an implementation plan suggesting incorporation of core findings in future chemical regulation and related policy areas.     

Out-of-field organ doses and associated risk of cancer development following radiation therapy with photons

Innovations in cancer treatment have contributed to the improved survival rate of these patients. Radiotherapy is one of the main options for cancer management nowadays. High doses of ionizing radiation are usually delivered to the tumor site with high energy photon beams. However, the therapeutic radiation exposure may lead to second cancer induction. Moreover, the introduction of intensity-modulated radiation therapy over the last decades has increased the radiation dose to out-of-field organs compared to that from conventional irradiation. The increased organ doses might result in elevated probabilities for developing secondary malignancies to critical organs outside the treatment volume. The organ-specific dosimetry is considered necessary for the theoretical second cancer risk assessment and the proper analysis of data derived from epidemiological reports. This study reviews the methods employed for the measurement and calculation of out-of-field organ doses from exposure to photons and/or neutrons. The strengths and weaknesses of these dosimetric approaches are described in detail. This is followed by a review of the epidemiological data associated with out-of-field cancer risks. Previously published theoretical cancer risk estimates for adult and pediatric patients undergoing radiotherapy with conventional and advanced techniques are presented. The methodology for the theoretical prediction of the probability of carcinogenesis to out-of-field sites and the limitations of this approach are discussed. The article also focuses on the factors affecting the magnitude of the probability for developing radiotherapy-induced malignancies. The restriction of out-of-field doses and risks through the use of different types of shielding equipment is presented.     

Trait‐based risk assessment for invasive species: high performance across diverse taxonomic groups,geographic ranges and machine learning/statistical tools

Aim Trait‐based risk assessment for invasive species is becoming an important tool for identifying non‐indigenous species that are likely to cause harm. Despite this, concerns remain that the invasion process is too complex for accurate predictions to be made. Our goal was to test risk assessment performance across a range of taxonomic and geographical scales, at different points in the invasion process, with a range of statistical and machine learning algorithms. Location Regional to global data sets. Methods We selected six data sets differing in size, geography and taxonomic scope. For each data set, we created seven risk assessment tools using a range of statistical and machine learning algorithms. Performance of tools was compared to determine the effects of data set size and scale, the algorithm used, and to determine overall performance of the trait‐based risk assessment approach. Results Risk assessment tools with good performance were generated for all data sets. Random forests (RF) and logistic regression (LR) consistently produced tools with high performance. Other algorithms had varied performance. Despite their greater power and flexibility, machine learning algorithms did not systematically outperform statistical algorithms. Geographic scope of the data set, and size of the data set, did not systematically affect risk assessment performance. Main conclusions Across six representative data sets, we were able to create risk assessment tools with high performance. Additional data sets could be generated for other taxonomic groups and regions, and these could support efforts to prevent the arrival of new invaders. Random forests and LR approaches performed well for all data sets and could be used as a standard approach to risk assessment development.