Anti-dense fine speckled 70 (DFS70) autoantibodies: correlates and increasing prevalence in the United States.

Objective: Recent studies report high-titer anti-dense fine speckled 70 (DFS70) autoantibodies in persons with inflammatory conditions, but the clinical significance remains unclear. Our goals were to estimate anti-DFS70 autoantibody prevalence, identify correlates, and assess time trends. Methods: Serum antinuclear antibodies (ANA) were measured by indirect immunofluorescence assay on HEp-2 cells in 13,519 participants ≥12 years old from three time periods (1988-1991, 1999-2004, 2011-2012) of the National Health and Nutrition Examination Survey. ANA-positive participants with dense fine speckled staining were evaluated for anti-DFS70 antibodies by enzyme-linked immunosorbent assay. We used logistic models adjusted for survey-design variables to estimate period-specific anti-DFS70 antibody prevalence in the US, and we further adjusted for sex, age, and race/ethnicity to identify correlates and assess time trends. Results: Women were more likely than men (odds ratio (OR)=2.97), black persons were less likely than white persons (OR=0.60), and active smokers were less likely than nonsmokers (OR=0.28) to have anti-DFS70 antibodies. The prevalence of anti-DFS70 antibodies increased from 1.6% in 1988-1991 to 2.5% in 1999-2004 to 4.0% in 2011-2012, which corresponds to 3.2 million, 5.8 million, and 10.4 million seropositive individuals, respectively. This increasing time trend in the US population (P<0.0001) was modified in some subgroups and was not explained by concurrent changes in tobacco smoke exposure. Some, but not all, anti-DFS70 antibody correlates and time trends resembled those reported for total ANA. Conclusion: More research is needed to elucidate anti-DFS70 antibody triggers, their pathologic or potentially protective influences on disease, and their possible clinical implications.

Generalized mean residual life models for survival data with missing censoring indicators.

The mean residual life (MRL) function is an important and attractive alternative to the hazard function for characterizing the distribution of a time-to-event variable. In this article, we study the modeling and inference of a family of generalized MRL models for right-censored survival data with censoring indicators missing at random. To estimate the model parameters, augmented inverse probability weighted estimating equation approaches are developed, in which the non-missingness probability and the conditional probability of an uncensored observation are estimated by parametric methods or nonparametric kernel smoothing techniques. Asymptotic properties of the proposed estimators are established and finite sample performance is evaluated by extensive simulation studies. An application to brain cancer data is presented to illustrate the proposed methods.

Increasing Prevalence of Antinuclear Antibodies in the United States.

OBJECTIVE: Growing evidence suggests increasing frequencies of autoimmunity and autoimmune diseases, but findings are limited by the lack of systematic data and evolving approaches and definitions. This study was undertaken to investigate whether the prevalence of antinuclear antibodies (ANA), the most common biomarker of autoimmunity, changed over a recent 25-year span in the US. METHODS: Serum ANA were measured by standard indirect immunofluorescence assays on HEp-2 cells in 13,519 participants age ≥12 years from the National Health and Nutrition Examination Survey, with approximately one-third from each of 3 time periods: 1988-1991, 1999-2004, and 2011-2012. We used logistic regression adjusted for sex, age, race/ethnicity, and survey design variables to estimate changes in ANA prevalence across the time periods. RESULTS: The prevalence of ANA was 11.0% (95% confidence interval [95% CI] 9.7-12.6%) in 1988-1991, 11.4% (95% CI 10.2-12.8%) in 1999-2004, and 16.1% (95% CI 14.4-18.0%) in 2011-2012 (P for trend <0.0001), corresponding to ~22.3 million, ~26.6 million, and ~41.5 million affected individuals, respectively. Among adolescents age 12-19 years, ANA prevalence increased substantially, with odds ratios of 2.07 (95% CI 1.18-3.64) and 2.77 (95% CI 1.56-4.91) in the second and third time periods relative to the first (P for trend = 0.0004). ANA prevalence increased in both sexes (especially in men), older adults (age ≥50 years), and non-Hispanic white individuals. These increases in ANA prevalence were not explained by concurrent trends in weight (obesity/overweight), smoking exposure, or alcohol consumption. CONCLUSION: The prevalence of ANA in the US has increased considerably in recent years. Additional studies to determine factors underlying these increases in ANA prevalence could elucidate causes of autoimmunity and enable the development of preventative measures.

Expanded assessment of xenobiotic associations with antinuclear antibodies in the United States, 1988-2012.

BACKGROUND: The prevalence of autoimmunity in the U.S. has increased recently for undetermined reasons. Little is known about associations between autoimmunity and environmental causes. OBJECTIVES: In a large representative sample of the U.S. population, we expanded our prior exploratory study of how exposures to selected xenobiotics and dioxin-like (DL) mixtures relate to antinuclear antibodies (ANA), the most common biomarker of autoimmunity. METHODS: We analyzed cross-sectional data on 12,058 participants aged ≥ 12 years from three time periods of the National Health and Nutrition Examination Survey between 1988 and 2012, of whom 14% were ANA-positive. We used lognormal regression models and censored-data methods to estimate ANA associations with xenobiotic concentrations overall and in sex, age, and race/ethnicity subgroups. Our analyses adjusted for potential confounders and appropriately handled concentrations below detection limits. RESULTS: Observed ANA associations were positive for most DL compounds and nonDL polychlorinated biphenyls (PCBs), negative for most phthalates, and mixed for other xenobiotic classes. After correcting for multiple comparisons, some associations remained statistically significant. In subgroup analyses, the most significant finding was a positive ANA association with N-acetyl-S-(2-hydroxy-3-butenyl)-L-cysteine (MHB2) in males, followed by positive associations with 2,2',3,5'-tetrachlorobiphenyl (PCB 44), 2,2',4,5'-tetrachlorobiphenyl (PCB 49), and 2,2',3,4',5',6-hexachlorobiphenyl (PCB 149) in 12-19 year-olds, and with 3,4,4',5-tetrachlorobiphenyl (PCB 81), 2,2',3,3',4,4',5,5',6-nonachlorobiphenyl (PCB 206), and N-acetyl-S-(phenyl)-L-cysteine (PMA) in Mexican Americans. Negative associations were found with mono-benzyl phthalate (MBzP) in 20-49 year-olds and mono-n-butyl phthalate (MnBP) in 12-19 year-olds. In overall analyses, combining stratum-specific results across race/ethnicity strata revealed a positive ANA association with PCB 81 and a negative ANA association with N-acetyl-S-(2-hydroxyethyl)-L-cysteine (HEMA). DISCUSSION: This study identified potential associations between ANA and various xenobiotics. Further investigation to confirm these observations and elucidate effects of certain xenobiotics on immune regulation could have important mechanistic, preventive, and treatment implications for a variety of immune-mediated disorders.

Hygiene Hypothesis Indicators and Prevalence of Antinuclear Antibodies in US Adolescents.

Autoimmunity prevalence, as measured by antinuclear antibodies (ANA), is increasing in U.S. adolescents. Improved hygiene and cleaner environments in childhood may reduce exposure to infections and other immune challenges, resulting in improper immune responses to later-life exposures. We examined associations of hygiene hypothesis indicators, including asthma, allergies, and antibodies to infectious agents, with ANA prevalence, measured by HEp-2 immunofluorescence, in adolescents (aged 12-19 years) over a 25-year time span in the National Health and Nutrition Examination Survey (NHANES) (N=2,709), adjusting for age, sex, race/ethnicity, body mass index, education and survey cycle, overall and within individual time periods, using logistic regression. Prevalence of ANA in adolescents increased from 5.0% in 1988-1991 to 12.8% in 2011-2012. ANA were positively associated with diagnosis of asthma in early childhood (OR: 2.07, CI: 1.09-3.99) and the effect estimate for current hay fever was elevated but not statistically significant (OR: 1.55, CI: 0.85-2.84). Fewer than 2% of those with ANA in 1988-1991 had been diagnosed with asthma, compared with 18% in 1999-2000, and 27% in 2003-2004 and 2011-2012. ANA trended negatively with Helicobacter pylori antibodies (OR: 0.49, CI: 0.24-0.99). ANA may be useful as an additional indicator of inadequate immune education in adolescence, a critical period of growth and development.

Early-life Farm Exposure and Ovarian Reserve in a US Cohort of Women.

BACKGROUND: In a previous exploratory study, we reported lower concentrations of the ovarian reserve biomarker anti-Müllerian hormone (AMH) in adulthood with prenatal farm exposure. We now examine this association as well as childhood farm exposure using enrollment data from the Sister Study, a large US cohort of women. METHODS: We collected prenatal and childhood farm exposure data by questionnaire and telephone interview. However, serum AMH data were available only for a nested subset: premenopausal women ages 35-54 subsequently diagnosed with breast cancer (n = 418 cases) and their matched controls (n = 866). To avoid potential bias from restricting analyses to only premenopausal controls, we leveraged the available cohort data. We used data from both premenopausal cases and controls as well as postmenopausal women ages 35-54 (n = 3,526) (all presumed to have undetectable AMH concentrations) and applied weights to produce a sample representative of the cohort ages 35-54 (n = 17,799). The high proportion of undetectable AMH concentrations (41%) was addressed using reverse-scale Cox regression. An adjusted hazard ratio (HR) <1.0 indicates that exposed individuals had lower AMH concentrations than unexposed individuals. RESULTS: Prenatal exposure to maternal residence or work on a farm was associated with lower AMH concentrations (HR 0.66; 95% confidence intervals [CI] = 0.48 to 0.90). Associations between childhood farm residence exposures and AMH were null or weak, except childhood contact with pesticide-treated livestock or buildings (HR 0.69; 95% CI = 0.40 to 1.2). CONCLUSIONS: Replication of the prenatal farm exposure and lower adult AMH association raises concern that aspects of prenatal farm exposure may result in reduced adult ovarian reserve.

Quantile regression models for survival data with missing censoring indicators.

The quantile regression model has increasingly become a useful approach for analyzing survival data due to its easy interpretation and flexibility in exploring the dynamic relationship between a time-to-event outcome and the covariates. In this paper, we consider the quantile regression model for survival data with missing censoring indicators. Based on the augmented inverse probability weighting technique, two weighted estimating equations are developed and corresponding easily implemented algorithms are suggested to solve the estimating equations. Asymptotic properties of the resultant estimators and the resampling-based inference procedures are established. Finally, the finite sample performances of the proposed approaches are investigated in simulation studies and a real data application.

Increasing Prevalence of Antinuclear Antibodies in the United States.

OBJECTIVE: Growing evidence suggests increasing frequencies of autoimmunity and certain autoimmune diseases, but findings are limited by the lack of systematic data and evolving approaches and definitions. This study was undertaken to investigate whether the prevalence of antinuclear antibodies (ANA), the most common biomarker of autoimmunity, changed over a recent 25-year span in the US. METHODS: Serum ANA were measured by standard indirect immunofluorescence assays on HEp-2 cells in 14,211 participants age ≥12 years from the National Health and Nutrition Examination Survey, with approximately one-third from each of 3 time periods: 1988-1991, 1999-2004, and 2011-2012. We used logistic regression adjusted for sex, age, race/ethnicity, and survey design variables to estimate changes in ANA prevalence across the time periods. RESULTS: The prevalence of ANA was 11.0% (95% confidence interval [95% CI] 9.7-12.6%) in 1988-1991, 11.5% (95% CI 10.3-12.8%) in 1999-2004, and 15.9% (95% CI 14.3-17.6%) in 2011-2012 (P for trend < 0.0001), which corresponds to ~22 million, ~27 million, and ~41 million affected individuals, respectively. Among adolescents age 12-19 years, ANA prevalence increased substantially, with odds ratios (ORs) of 2.02 (95% CI 1.16-3.53) and 2.88 (95% CI 1.64-5.04) in the second and third time periods relative to the first (P for trend < 0.0001). ANA prevalence increased in both sexes (especially in men), older adults (age ≥50 years), and non-Hispanic whites. These increases in ANA prevalence were not explained by concurrent trends in weight (obesity/overweight), smoking exposure, or alcohol consumption. CONCLUSION: The prevalence of ANA in the US has increased considerably in recent years. Additional studies to determine factors underlying these increases in ANA prevalence could elucidate causes of autoimmunity and enable the development of preventative measures.

Prescription medication use and antinuclear antibodies in the United States, 1999-2004.

BACKGROUND: Clinical reports link specific medications with the development of antinuclear antibodies (ANA), but population-based evidence is limited. OBJECTIVE: The present study investigated associations between prescription medication use and ANA in a representative sample of the adult noninstitutionalized US population, first focusing on medications previously related to ANA and then considering all medications reported in the National Health and Nutrition Examination Survey (NHANES). METHODS: Based on NHANES data (1999-2004) for 3608 adults (ages ≥18 years), we estimated odds ratios (ORs) and 95% confidence intervals (CIs) to assess associations between recent medication use and ANA (overall and in sex and age subgroups), adjusted for potential confounders and the survey sampling design. RESULTS: We found no evidence that most medications previously associated with ANA in specific individuals were risk factors for ANA in the general population, although statistical power was limited for some medications. Overall, ANA were less prevalent in adults who recently used any prescription medications compared with those who did not (OR = 0.73; CI = 0.57,0.93), and likewise several classes of medications were inversely associated with ANA, including hormones (OR = 0.73; CI = 0.55,0.98), thiazide diuretics (OR = 0.43; CI = 0.24,0.79), sulfonylureas (OR = 0.41; CI = 0.19,0.89), and selective serotonin reuptake inhibitor antidepressants (OR = 0.65; CI = 0.42,0.98). Positive associations with ANA were seen for loop diuretics (OR = 1.72; CI = 1.03,2.88) in all adults, and for benzodiazepines (OR = 2.11; CI = 1.09,4.10) and bronchodilators (OR = 1.83; CI = 1.00,3.38) in older (ages ≥60) adults. Estrogens were positively associated with ANA in older women (OR = 1.80; CI = 1.00,3.23) but inversely associated with ANA in younger (ages 18-59) women (OR = 0.43; CI = 0.20,0.93). Regarding individual medications, ANA were positively associated with ciprofloxacin (OR = 4.23; CI = 1.21,14.8), furosemide (OR = 1.79; CI = 1.09,2.93), and omeprazole (OR = 2.05; CI = 1.03,4.10) in all adults, and with salmeterol (OR = 3.76; CI = 1.66,8.52), tolterodine (OR = 6.64; CI = 1.45,30.5), and triamterene (OR = 3.10; CI = 1.08,8.88) in older adults. Also, in younger adults, hydrochlorothiazide was inversely associated with ANA (OR = 0.44; CI = 0.20,0.98). CONCLUSIONS: Our findings in the general population do not confirm most clinically reported positive associations between specific medications and ANA in some individuals. However, novel positive ANA associations with other medications, as well as unexplained inverse associations with certain classes of medications and overall medication use, deserve further research to clarify the possible roles of medications as risk and protective factors in the development of autoantibodies and autoimmune disease.

Antinuclear antibodies and mortality in the National Health and Nutrition Examination Survey (1999-2004).

OBJECTIVE: Recent studies suggest antinuclear antibodies (ANA) may be related to mortality risk, but evidence is sparse and inconclusive. Thus, we investigated ANA associations with all-cause and cause-specific mortality in U.S. adults. METHODS: Our sample included 3357 adults (ages ≥20 years) from the 1999-2004 National Health and Nutrition Examination Survey with ANA measurements (1:80 dilution) and mortality data through 2011 (median follow-up: 9.4 years). We estimated hazard ratios (HRs) and 95% confidence intervals (CIs) via weighted Cox regression to assess ANA associations with mortality from all causes, cardiovascular disease (CVD), and cancer. Models adjusted for age, sex, race/ethnicity, education, and obesity. Analyses examined mortality in the full sample and in subgroups based on self-reported histories of CVD and cancer, both overall and stratified by sex and age at enrollment. RESULTS: Overall, ANA were not strongly associated with death from all causes (HR: 1.13; CI: 0.79, 1.60), from CVD (HR: 1.60; CI: 0.80, 3.20), or from cancer (HR: 1.58; CI: 0.75, 3.33), though all three HR estimates exceeded 1. In the subgroup with a history of cancer, ANA were associated with elevated all-cause mortality in men (HR: 2.28; CI: 1.01, 5.14) and in participants who enrolled at age ≥75 years (HR: 1.99; CI: 1.04, 3.80). CONCLUSION: These findings suggest that ANA are not strongly associated with mortality in the general population. Longitudinal studies with repeated assessments are needed to understand the temporal relationship between ANA, aging-associated diseases, and mortality.

Associations Between Selected Xenobiotics and Antinuclear Antibodies in the National Health and Nutrition Examination Survey, 1999-2004.

BACKGROUND: Potential associations between background environmental chemical exposures and autoimmunity are understudied. OBJECTIVES: Our exploratory study investigated exposure to individual environmental chemicals and selected mixtures in relation to the presence of antinuclear antibodies (ANA), a widely used biomarker of autoimmunity, in a representative sample of the U.S. METHODS: This cross-sectional analysis used data on 4,340 participants from the National Health and Nutrition Examination Survey (1999-2004), of whom 14% were ANA positive, to explore associations between ANA and concentrations of dioxins, dibenzofurans, polychlorinated biphenyls, organochlorines, organophosphates, phenols, metals, and other environmental exposures and metabolites measured in participants' serum, whole blood, or urine. For dioxin-like compounds with toxic equivalency factors, we developed and applied a new statistical approach to study selected mixtures. Lognormal models and censored-data methods produced estimates of chemical associations with ANA in males, nulliparous females, and parous females; these estimates were adjusted for confounders and accommodated concentrations below detectable levels. RESULTS: Several associations between chemical concentration and ANA positivity were observed, but only the association in males exposed to triclosan remained statistically significant after correcting for multiple comparisons (mean concentration ratio = 2.8; 95% CI: 1.8, 4.5; p < 0.00001). CONCLUSIONS: These data suggest that background levels of most xenobiotic exposures typical in the U.S. population are not strongly associated with ANA. Future studies should ideally reduce exposure misclassification by including prospective measurement of the chemicals of concern and should track changes in ANA and other autoantibodies over time. CITATION: Dinse GE, Jusko TA, Whitt IZ, Co CA, Parks CG, Satoh M, Chan EKL, Rose KM, Walker NJ, Birnbaum LS, Zeldin DC, Weinberg CR, Miller FW. 2016. Associations between selected xenobiotics and antinuclear antibodies in the National Health and Nutrition Examination Survey, 1999-2004. Environ Health Perspect 124:426-436; http://dx.doi.org/10.1289/ehp.1409345.

Accommodating measurements below a limit of detection: a novel application of Cox regression.

In environmental epidemiology, measurements of exposure biomarkers often fall below the assay's limit of detection. Existing methods for handling this problem, including deletion, substitution, parametric regression, and multiple imputation, can perform poorly if the proportion of "nondetects" is high or parametric models are mis-specified. We propose an approach that treats the measured analyte as the modeled outcome, implying a role reversal when the analyte is a putative cause of a health outcome. Following a scale reversal as well, our approach uses Cox regression to model the analyte, with confounder adjustment. The method makes full use of quantifiable analyte measures, while appropriately treating nondetects as censored. Under the proportional hazards assumption, the hazard ratio for a binary health outcome is interpretable as an adjusted odds ratio: the odds for the outcome at any particular analyte concentration divided by the odds given a lower concentration. Our approach is broadly applicable to cohort studies, case-control studies (frequency matched or not), and cross-sectional studies conducted to identify determinants of exposure. We illustrate the method with cross-sectional survey data to assess sex as a determinant of 2,3,7,8-tetrachlorodibenzo-p-dioxin concentration and with prospective cohort data to assess the association between 2,4,4'-trichlorobiphenyl exposure and psychomotor development.

Scientific retractions and corrections related to misconduct findings.

We examined all 208 closed cases involving official findings of research misconduct published by the US Office of Research Integrity from 1992 to 2011 to determine how often scientists mention in a retraction or correction notice that there was an ethical problem with an associated article. 75 of these cases cited at least one published article affected by misconduct for a total of 174 articles. For 127 of these 174, we found both the article and a retraction or correction statement. Since eight of the 127 published statements consisted of simply the word 'retracted,' our analysis focused on the remaining 119 for which a more detailed retraction or correction was published. Of these 119 statements, only 41.2% mentioned ethics at all (and only 32.8% named a specific ethical problem such as fabrication, falsification or plagiarism), whereas the other 58.8% described the reason for retraction or correction as error, loss of data or replication failure when misconduct was actually at issue. Among the published statements in response to an official finding of misconduct (within the time frame studied), the proportion that mentioned ethics was significantly higher in recent years than in earlier years, as was the proportion that named a specific problem. To promote research integrity, scientific journals should consider adopting policies concerning retractions and corrections similar to the guidelines developed by the Committee on Publication Ethics. Funding agencies and institutions should take steps to ensure that articles affected by misconduct are retracted or corrected.

Do U.S. research institutions meet or exceed federal mandates for instruction in responsible conduct of research? A national survey.

PURPOSE: To explore the extent to which U.S. research institutions are meeting or exceeding National Institutes of Health and National Science Foundation mandates to provide instruction in responsible conduct of research (RCR). METHOD: In summer 2011, the authors sent an e-mail survey to officials responsible for overseeing RCR instructional programs at the 200 top-funded research institutions in the United States and Puerto Rico. They cross-classified the proportions exceeding federal mandates by the types of additional individuals required to receive training and by medical school presence/absence. RESULTS: Responses were received from 144 institutions (72%); all had an RCR program. Of these 144 institutions, 69 (47.9%) required only federally mandated individuals to take RCR training, whereas 75 (52.1%) required additional individuals to be trained as well. A greater proportion of institutions with medical schools (62.3%; 53/85) went beyond the federal mandates than did those without (37.3%; 22/59). Types of additional individuals required to receive training included all students in selected programs (23.6%; 34/144), all students participating in externally funded research (12.5%; 18/144), all graduate students (11.1%; 16/144), all faculty/staff participating in externally funded research (9.7%; 14/144), all postdoctoral students or fellows (8.3%; 12/144), all doctoral-level students (4.9%; 7/144), all faculty/staff involved in human subjects research (4.9%; 7/144), and all faculty/staff involved in animal research (2.1%; 3/144). CONCLUSIONS: More institutions with medical schools exceeded federal RCR training mandates than did those without. The authors encourage other institutions to expand their RCR requirements to promote research integrity.

Parameterizing dose-response models to estimate relative potency functions directly.

Many comparative analyses of toxicity assume that the potency of a test chemical relative to a reference chemical is constant, but employing such a restrictive assumption uncritically may generate misleading conclusions. Recent efforts to characterize non-constant relative potency rely on relative potency functions and estimate them secondarily after fitting dose-response models for the test and reference chemicals. We study an alternative approach of specifying a relative potency model a priori and estimating it directly using the dose-response data from both chemicals. We consider a power function in dose as a relative potency model and find that it keeps the two chemicals' dose-response functions within the same family of models for families typically used in toxicology. When differences in the response limits for the test and reference chemicals are attributable to the chemicals themselves, the older two-stage approach is the more convenient. When differences in response limits are attributable to other features of the experimental protocol or when response limits do not differ, the direct approach is straightforward to apply with nonlinear regression methods and simplifies calculation of simultaneous confidence bands. We illustrate the proposed approach using Hill models with dose-response data from U.S. National Toxicology Program bioassays. Though not universally applicable, this method of estimating relative potency functions directly can be profitably applied to a broad family of dose-response models commonly used in toxicology.

Hazard Function Estimation with Cause-of-Death Data Missing at Random.

Hazard function estimation is an important part of survival analysis. Interest often centers on estimating the hazard function associated with a particular cause of death. We propose three nonparametric kernel estimators for the hazard function, all of which are appropriate when death times are subject to random censorship and censoring indicators can be missing at random. Specifically, we present a regression surrogate estimator, an imputation estimator, and an inverse probability weighted estimator. All three estimators are uniformly strongly consistent and asymptotically normal. We derive asymptotic representations of the mean squared error and the mean integrated squared error for these estimators and we discuss a data-driven bandwidth selection method. A simulation study, conducted to assess finite sample behavior, demonstrates that the proposed hazard estimators perform relatively well. We illustrate our methods with an analysis of some vascular disease data.

An EM Algorithm for Fitting a 4-Parameter Logistic Model to Binary Dose-Response Data.

This article is motivated by the need of biological and environmental scientists to fit a popular nonlinear model to binary dose-response data. The 4-parameter logistic model, also known as the Hill model, generalizes the usual logistic regression model to allow the lower and upper response asymptotes to be greater than zero and less than one, respectively. This article develops an EM algorithm, which is naturally suited for maximum likelihood estimation under the Hill model after conceptualizing the problem as a mixture of subpopulations in which some subjects respond regardless of dose, some fail to respond regardless of dose, and some respond with a probability that depends on dose. The EM algorithm leads to a pair of functionally independent 2-parameter optimizations and is easy to program. Not only can this approach be computationally appealing compared to simultaneous optimization with respect to all four parameters, but it also facilitates estimating covariances, incorporating predictors, and imposing constraints. This article is motivated by, and the EM algorithm is illustrated with, data from a toxicology study of the dose effects of selenium on the death rates of flies. Other biological and environmental applications, as well as medical and agricultural applications, are also described briefly. Computer code for implementing the EM algorithm is available as supplemental material online.

Comparing tumor rates in current and historical control groups in rodent cancer bioassays.

When evaluating carcinogenicity, tumor rates from the current study are informally assessed within the context of relevant historical control tumor rates. Current rates outside the range of historical rates raise concerns. We propose a statistical procedure that formally compares tumor rates in current and historical control groups. We use a normal approximation for the null distribution of the proposed test when there are at least 5 historical control groups and the average tumor rate is above 0.5%; otherwise, we apply standard bootstrap techniques. For comparison purposes, we show that formally basing decisions on the range of historical control rates would yield unusually high false positive rates. That is, a range-based decision rule would not maintain the nominal 5% significance level and could produce Type I error rates as high as 67%. In other cases, the power could go to zero. The proposed test, however, controls Type I errors while adjusting for survival and extra variability among the historical studies. We illustrate the methods with data from a study of benzophenone. Compared to a range-based decision rule, the proposed test has several important advantages, including operating at the specified level and being applicable with as few as one historical study.

Characterizing non-constant relative potency.

Relative potency plays an important role in toxicology. Estimates of relative potency are used to rank chemicals by their effects, to calculate equivalent doses of test chemicals compared to a standard, and to weight contributions of constituent chemicals when evaluating mixtures. Typically relative potency is characterized by a constant dilution factor, even when non-similar dose-response curves indicate that constancy is inappropriate. Improperly regarding relative potency as constant may distort conclusions and potentially mislead investigators or policymakers. We consider a more general approach that allows relative potency to vary as a function of dose, response, or response quantile. Distinct functions can be defined, each generalizing different but equivalent descriptions of constant relative potency. When two chemicals have identical response limits, these functions all carry fundamentally equivalent information; otherwise, relative potency as a function of response quantile is distinct and embodies a modified definition of relative potency. Which definition is preferable depends on whether one views any differences in response limits as intrinsic to the chemicals or as extrinsic, arising from idiosyncrasies of data sources. We illustrate these ideas with constructed examples and real data. Relative potency functions offer a unified and principled description of relative potency for non-similar dose-response curves.