A dose-response model for teratological experiments involving quantal responses

This paper introduces a dose-response model for teratological quantal response data where the probability of response for an offspring from a female at a given dose varies with the litter size. The maximum likelihood estimators for the parameters of the model are given as the solution of a nonlinear iterative algorithm. Two methods of low-dose extrapolation are presented, one based on the litter size distribution and the other a conservative method. The resulting procedures are then applied to a teratological data set from the literature.     

Design and analysis of experiments on mutagenicity. II. Assays involving microorganisms.

The design and statistical analysis of mutagenicity experiments involving microorganisms and a single dose of mutagen are discussed. Test statistics are derived for use in determining the mutagenicity of a chemical when survival data are available and also when such data are not available. One's likelihood (power) of correctly concluding a chemical is mutagenic is examined, and minimum total sample sizes required for 95% power are presented. It is found that one generally has greater power when survival data are available. Required precision is estimating survival is discussed in reference to type-1 and type-2 errors. The proper use of the formulae and figures presented is illustrated by examples.     

Analysis of dichotomous response data from certain toxicological experiments

In certain toxicological experiments with laboratory animals, littermate data are frequently encountered. It is generally recognized that one characteristic of this type of data is the "litter effect", i.e., the tendency for animals from the same litter to respond more alike than animals from different litters. In this paper attention is restricted to dichotomous response variables that frequently arise in toxicological studies, such as the occurrence of fetal death or a particular malformation. Various techniques for estimating the underlying probability of response are discussed. A number of generalized models that have recently been proposed to take the litter effect into account are breifly reviewed and compared to the simpler binomial and Poisson models. Various procedures for assessing the significance of treatment-control differences are presented and their relative merits discussed. Finally, future research needs in this area are outlined.     

The use of a correlated binomial model for the analysis of certain toxicological experiments

In certain toxicological experiments with laboratory animals, the outcome of interest is the occurrence of dead or malformed fetuses in a litter. Previous investigations have shown that the simple one-parameter binomial and Poisson models generally provide poor fits to this type of binary data. In this paper, a type of correlated binomial model is proposed for use in this situation. First, the model is described in detail and is compared to a beta-binomial model proposed by Williams (1975). These two-parameter models are then contrasted for goodness of fit to some real-life data. Finally, numerical examples are given in which likelihood ratio tests based on these models are employed to assess the significance of treatment-control differences.     

Kinetic analysis of experiments involving the single turnover of an enzyme.

The complete solution to the kinetic equation for nucleotide fluorescence quenching on addition of pyruvate to the late dehydrogenase-NADH complex modifies previous interpretations of such experiments.     

The presentation of treatment responses from block experiments after analysis of variance of transformed data

In pearl millet, severe water deficit during the period of panicle development delays flowering. The flowering response of both main shoot and tillers to water stress during panicle development was investigated using four hybrids. Panicle initiation of all tillers occurred in the three early genotypes despite water stress. In the late genotype, however, panicle initiation of tillers occurred only after the release of stress. The delay in flowering due to water stress was more pronounced in the tillers than in the main shoot. However, the proportion of tillers producing an inflorescence was increased by water stress. Grain yield losses on the main shoot by water stress were compensated by an increase in tiller grain yields. Delay in flowering and buffering by tillers provide an important adaptive mechanism to overcome a period of drought stress prior to flowering.     

Analysis of trinomial responses from reproductive and developmental toxicity experiments.

This paper presents a Dirichlet-trinomial distribution for modelling data obtained from reproductive and developmental studies. The common endpoints for the evaluation of reproductive and developmental toxic effects are the number of dead fetuses, the number of malformed fetuses, and the number of normal fetuses for each litter. With current statistical methods for the evaluation of reproductive and developmental effects, the effect on the number of deaths and the effect on the number of malformations are analyzed separately. The Dirichlet-trinomial model provides a procedure for the analysis of multiple endpoints simultaneously. This proposed Dirichlet-trinomial model is a generalization of the beta-binomial model that has been used for handling the litter effect in reproductive and developmental experiments. Likelihood ratio tests for differences in the number of deaths, the number of malformations, and the number of normals among dosed and control groups are derived. The proposed test procedure based on the Dirichlet-trinomial model is compared with that based on the beta-binomial model with an application to a real data set.     

Systematic reviews of animal experiments demonstrate poor human clinical and toxicological utility

The assumption that animal models are reasonably predictive of human outcomes provides the basis for their widespread use in toxicity testing and in biomedical research aimed at developing cures for human diseases. To investigate the validity of this assumption, the comprehensive Scopus biomedical bibliographic databases were searched for published systematic reviews of the human clinical or toxicological utility of animal experiments. In 20 reviews in which clinical utility was examined, the authors concluded that animal models were either significantly useful in contributing to the development of clinical interventions, or were substantially consistent with clinical outcomes, in only two cases, one of which was contentious. These included reviews of the clinical utility of experiments expected by ethics committees to lead to medical advances, of highly-cited experiments published in major journals, and of chimpanzee experiments--those involving the species considered most likely to be predictive of human outcomes. Seven additional reviews failed to clearly demonstrate utility in predicting human toxicological outcomes, such as carcinogenicity and teratogenicity. Consequently, animal data may not generally be assumed to be substantially useful for these purposes. Possible causes include interspecies differences, the distortion of outcomes arising from experimental environments and protocols, and the poor methodological quality of many animal experiments, which was evident in at least 11 reviews. No reviews existed in which the majority of animal experiments were of good methodological quality. Whilst the effects of some of these problems might be minimised with concerted effort (given their widespread prevalence), the limitations resulting from interspecies differences are likely to be technically and theoretically impossible to overcome. Non-animal models are generally required to pass formal scientific validation prior to their regulatory acceptance. In contrast, animal models are simply assumed to be predictive of human outcomes. These results demonstrate the invalidity of such assumptions. The consistent application of formal validation studies to all test models is clearly warranted, regardless of their animal, non-animal, historical, contemporary or possible future status. Likely benefits would include, the greater selection of models truly predictive of human outcomes, increased safety of people exposed to chemicals that have passed toxicity tests, increased efficiency during the development of human pharmaceuticals and other therapeutic interventions, and decreased wastage of animal, personnel and financial resources. The poor human clinical and toxicological utility of most animal models for which data exists, in conjunction with their generally substantial animal welfare and economic costs, justify a ban on animal models lacking scientific data clearly establishing their human predictivity or utility.     

Testing the effect of treatment in experiments with correlated binary outcomes.

This paper considers the problem of testing for treatment effect in a randomized experiment with correlated binary outcomes, representing success or failure for different "parts" of a randomized unit. Attention is restricted to tests that are based on a summary score for each individual randomized, and thus are valid regardless of the precise nature of the correlation among parts. The focus is on the efficiency of such tests under various correlation structures, with special emphasis on the case in which the correlation among parts within an individual differs across treatment groups. A class of summary score statistics is defined, and optimal testing is discussed for some simple situations. Three potential general-purpose tests also are described: (1) the ratio estimate test discussed by Henderson et al. (1988, Controlled Clinical Trials 9, 189-205); (2) a modified ratio estimate test with adjusted weighting based on the within-individual correlation between parts; (3) a test defined by applying the Mantel-Haenszel procedure to the proportion of individuals with at least one failure, stratifying by the number of parts. For these general-purpose tests, numerical calculations of asymptotic efficiency are presented under a wide range of designs and correlation structures. On the basis of these results, some practical recommendations for choosing a test are made.     

The analysis of cancer chemoprevention experiments

This paper is concerned with the analysis of certain cancer chemoprevention experiments that involve Type I censoring. In experiments of this nature, two common response variables are the number of induced cancers and the rate at which they develop. In this study we assume that the number of induced tumors and their times to detection are described by the Poisson and gamma distributions, respectively. Using the method of maximum likelihood, we discuss a procedure for estimating the parameters characterizing these two distributions. We apply standard techniques in order to construct a confidence region and conduct a hypothesis test concerning the parameters of interest. We discuss a method for comparing the effects of two different treatments using the likelihood ratio principle. A technique for isolating group differences in terms of the mean number of promoted tumors and the mean time to detection is described. Using the techniques developed in this paper, we reanalyze an existing data set in the cancer chemoprevention literature and obtain contrasting results.     

Growth responses of age 0 white perch and yellow perch from field-enclosure experiments

In this paper, we investigate the potential for competition between age 0 white perch Morone americana and yellow perch Perca flavescens in field-enclosure experiments. In 1986 and 1987, different densities of white perch and yellow perch were stocked into cages in the western basin of Lake Erie and their growth responses determined. Fish were weighed and measured pre- and post-stocking. In 1986 when fish were food-limited, individual growth of both species were negatively affected by increased total fish density; yellow perch to a greater extent. In 1987, experiments were carried out later in the year at lower temperatures. Yellow perch were shown to feed and continue to grow at lower water temperatures than white perch. We conclude that competition in summer may be mitigated in the fall because of differences in thermal optima between these two species.     

Importance of spontaneous alpha-ketoacid decarboxylation in experiments involving peroxide

The potential role of spontaneous alpha-ketoacid decarboxylation as a source of interference in experiments involving peroxide was investigated. The assay of pyruvate dehydrogenase activity in isolated renal mitochondria was employed as an example. Spontaneous peroxide-induced pyruvate decarboxylation competed significantly with enzymatic decarboxylation at peroxide concentrations greater than 50 microM. Corrected values for enzymatic decarboxylation could be obtained by subtracting spontaneous decarboxylation rates from rates obtained in the presence of mitochondria. At higher peroxide concentrations (greater than 200 microM), reaction product accumulates (acetoacetate) to levels which may have regulatory effects on mitochondrial metabolism. The divalent cations, Ca2+ and Mg2+, both accelerate spontaneous peroxide-induced pyruvate decarboxylation while other components of the assay medium had an inhibitory effect on the reaction. The results are discussed in relation to the currently accepted reaction mechanism. Investigators who perform experiments involving reactive oxygen species should be familiar with this often overlooked reaction.     

Multiple comparisons in the randomization analysis of designed experiments with growth curve responses

A randomization approach to multiple comparisons is developed for comparing several growth curves in randomized experiments. The exact Type I probability error rate for these comparisons may be prespecified, and a Type I error probability for each component test can be evaluated. These procedures are free of many of the standard assumptions for analyzing growth curves and for making multiple comparisons. An application of the procedure gives all pairwise comparisons among the mean growth curves associated with four treatments in an animal experiment using a Youden square design, where growth curves are obtained on monitoring hormone levels over time.     

Graphic presentation and analysis of inhibition data from ligand-binding experiments

One technique for the characterization of receptor subtypes involves measuring the inhibition of the binding of a radioligand which is not subtype selective by agonists or antagonists which are subtype selective. Although such data are routinely calculated using computer programs, it is often useful to have a graphical representation of the data. Until now, a "modified Scatchard" or "Hofstee" plot of the form P = -(P/I)(IC50) + 1 (where P is percentage inhibition and I is the inhibitor concentration) has been used. We describe an alternate plot of the form B = -(B X I)(1/IC50) + B0 (where B is the concentration of bound radioligand and B0 is the concentration of radioligand bound in the absence of the inhibitor). This method has the important advantage that the data need not be calculated as percentage inhibition which eliminates the error involved in the experimentally determined B0 value being included in the other values.     

Phenytoin teratogenicity and midgestational pharmacokinetics in mice.

Mice of the A/J and C57BL/6J (C57) strains were dosed with phenytoin (PHT) every 48 hr throughout pregnancy by gastric intubation to test the hypothesis that maternal plasma PHT concentration may be the significant factor in determining PHT reproductive and developmental toxicity. Serial serum samples were obtained from each mouse from gestation day (GD) 10-GD 12 for determination of individual dam PHT pharmacokinetics. Maximum PHT concentration and PHT AUC (area under-the-time-concentration curve) were regressed to laparotomy and fetal evaluation endpoints to determine whether significant association existed. Although serum PHT concentrations exceeded levels associated with teratogenicity (greater than 10 micrograms/ml), few major malformations were induced in either strain. However, in the A/J strain, there was a significant increased incidence of hydrocephaly and open eyelid. Regression of pharmacokinetic parameters with embryo and maternal endpoints indicated significant associations between gestational weight gain and maximum concentration measured (Cmax) or AUC in both strains. This association was also found for fetal weight in the C57 strain. In the A/J strain, the induction of decreased ossification of the sternebrae was also associated with maternal PHT concentration; however, linear regression of hydrocephaly and open eyelid to PHT concentration was not statistically significant. These results suggest that maternal plasma PHT concentration may be a quantifiable determinant of certain aspects of PHT developmental toxicity in the mouse.     

A new statistical test of fitness set data from reciprocal transplant experiments involving intermediate phenotypes

Experimental biologists use reciprocal transplant experiments (RTEs) involving divergent forms to test hypotheses about fitness trade-offs across, and local adaptation to, native environments. Additional evolutionary hypotheses about diversifying selection, the evolution of specialization, and the coexistence of specialists and generalists are only testable when the RTE also includes intermediate (or alternatively generalist) forms. Environmental variation makes such RTEs challenging, and so strategies that increase their effectiveness are useful. Here, we focus on improvements to the efficiency of RTEs involving intermediate forms with respect to the experimental design and the analysis of the resulting data. We provide a likelihood ratio-based test that offers increased statistical power and robustness relative to another test involving nonlinear regression, when used both for simulated data sets and for data from a study of two divergent fish species and their hybrids transplanted between two lake habitats. The test can be used with unequal numbers of observations, unequal variances, and binomial-type survival data and other nonnormal data. Simulations suggest that having equal numbers of experimental units in each phenotype-environment combination is reasonable. The intentional pairing of observations between environmental conditions (by using clones, full sibs, or half-sibs) is beneficial when paired observations have fitnesses that are negatively related between conditions but is detrimental with positive relatedness. Our methods can be extended to study more than two divergent forms.     

Mutagenicity, carcinogenicity, and teratogenicity of acrylonitrile.

Acrylonitrile (AN) is an important intermediary for the synthesis of a variety of organic products, such as artificial fibres, household articles and resins. Although acute effects are the primary concern for an exposure to AN, potential genotoxic, carcinogenic and teratogenic risks of AN have to be taken seriously in view of the large number of workers employed in such industries and the world-wide population using products containing and possibly liberating AN. An understanding of the effect of acrylonitrile must be based on a characterization of its metabolism as well as of the resulting products and their genotoxic properties. Tests for mutagenicity in bacteria have in general been positive, those in plants and on unscheduled DNA synthesis doubtful, and those on chromosome aberrations in vivo negative. Wherever positive results had been obtained, metabolic activation of AN appeared to be a prerequisite. The extent to which such mutagenic effects are significant in man depends, however, also on the conditions of exposure. It appears from the limited data that the ultimate mutagenic factor(s), such as 2-cyanoethylene oxide, may have little opportunity to act under conditions where people are exposed because it is formed only in small amounts and is rapidly degraded. The carcinogenic action of AN has been evaluated by various agencies and ranged from 'reasonably be anticipated to be a human carcinogen' to 'cannot be excluded', the most recent evaluation being 'possibly carcinogenic to humans'. Animal data that confirm the carcinogenic potential of AN have certain limitations with respect to the choice of species, type of tumors and length of follow up. Epidemiological studies which sometimes, but not always, yielded positive results, encounter the usual difficulties of confounding factors in chemical industries. Exposure of workers to AN should continue to be carefully monitored, but AN would not have to be considered a cancer risk to the population provided limitations on releases from consumer products and guidelines on AN in water and air are enforced. AN is teratogenic in laboratory animals (rat, hamster) at high doses when foetal/embryonic (and maternal) toxicity already is manifest. Pregnant workers should not be exposed to AN. In view of the small concentrations generally encountered outside plants, women not professionally exposed would appear not to be at risk of teratogenic effects due to AN. Future research should concentrate on the elucidation of the different degradation pathways in man and on epidemiological studies in workers including pregnant women, assessing also, if possible, individual exposure by bio-monitoring.