The limitations of animal tests for assessing possible hazards in new foods are well documented

Hazard characterisation of chemicals in food and diet. dose response, mechanisms and extrapolation issues.

Dybing E, Doe J, Groten J, Kleiner J, O'Brien J, Renwick AG, Schlatter J, Steinberg P, Tritscher A, Walker R, Younes M.

National Institute of Public Health, Department of Environmental Medicine, PO Box
4404 Nydalen, N-0403, Oslo, Norway.

Food Chem Toxicol. 2002 Feb-Mar;40(2-3):237-82.

Section 9. Novel foods, macronutrients and whole foods
9.1. Novel foods
9.2. Macronutrients and whole foods
9.2.1. Current situation

9.2.2. Limitations
A fundamental difference between the toxicological evaluation of macronutrients/whole foods and food additives or other compounds occurring at low levels in the diet is the difficulty of using conventional toxicological safety factors in the design of animal feeding studies (Borzelleca, 1992). For whole foods and for ingredients intended for use at high levels of incorporation in the human diet, the feeding of 100-fold or more the concentration in the human diet to experimental animals is usually unachievable. Normally the limit of dietary admixture is approximately 5% (OECD, 1995), above which unintended effects due to nutritional imbalances/deficiencies are likely to occur. For nutrients, dietary incorporation at the expense of a corresponding nutrient is possible at levels as high as 60% (OECD, 1995). However, with increasing levels of dietary incorporation, an increasing number of observed effects are not toxicologically relevant and a consequence of high doses leading to secondary effects, for instance due to nutrient imbalances or metabolic overload.

Interpretation of such studies is often difficult and may give rise to erroneous conclusions.
Historically there are several examples that demonstrate this problem. The safety evaluation of modified starches, polyols and other poorly digestible carbohydrates 30 years ago demonstrated the difficulty of interpreting physiological changes on high-dose feeding to rodents. Several such studies reported toxicological effects that appeared to be common to several structurally diverse substances. Such effects included cecal enlargement, nephrocalcinosis and, in long-term studies, adrenal medullary hyperplasia and neoplasia (Roe, 1993). The cecal enlargement seen in the rat has long been accepted to be an adaptive phenomenon without toxicological relevance. It is currently accepted that the nephrocalcinosis seen in such studies may be attributed to disturbances of calcium metabolism (Lynch et al., 1996). The adrenal medullary hyperplasia and phaeochromocytomas seen in long-term studies of xylitol were also attributed to disturbances of calcium homoeostasis (Roe, 1984).

What constitutes a ‘high dose’ leading to secondary effects that are not due to overt/inherent toxicity of the test compound, however, is highly dependent on the substance to be tested. Therefore, standard approaches for dose selection in rodent feeding studies as developed for low molecular weight chemicals are generally not applicable for macronutrients and whole foods.

Cited and other references:
Borzelleca, J.F., 1992. The safety evaluation of macronutrient substitutes. Critical Reviews of Food Science and Nutrition 32, 127– 139.
Borzelleca, J.F., 1996. A proposed model for safety assessment of macronutrient substitutes. Regulatory Toxicology and Pharmacology 23, S15–S18.
Lynch, B.S., Tischler, A.S., Capen, C., Munro, I.C., McGirr, L.M., McClain, R.M., 1996. Low digestible carbohydrates (polyols and lactose): significance of adrenal medullary proliferative lesions in the rat. Regulatory Toxicology and Pharmacology 23, 256–297.
OECD, 1984. Toxicokinetics. OECD Guideline for Testing of Chemicals, No. 417. Organisation for Economic Co-operation and Development, Paris.
OECD, 1995. Giudelines for testing of chemicals, OCED Paris.
Roe, F.J.C., 1984. Perspectives in carbohydrate toxicology with special reference to carcinogenicity. Swedish Dentistry Journal 8, 99–111.
Roe, F.J.C., 1993. The Leon Golberg Memorial Lecture. Recent advances in toxicology relevant to carcinogenesis: seven cameos. Food and Chemical Toxicology 31, 909–925.

A fuller discussion of hazard identification:

Hazard identification by methods of animal-based toxicology.
Barlow SM, Greig JB, Bridges JW, Carere A, Carpy AJ, Galli CL, Kleiner J, Knudsen I, Koëter HB, Levy LS, Madsen C, Mayer S, Narbonne JF, Pfannkuch F, Prodanchuk MG, Smith MR, Steinberg P.
Food Chem Toxicol. 2002 Feb-Mar;40(2-3):145-91.

MRC Institute for Environment and Health, University of Leicester, 94 Regent
Road, LE1 7DD, Leicester, UK.

This paper is one of several prepared under the project "Food Safety In Europe: Risk Assessment of Chemicals in Food and Diet" (FOSIE), a European Commission Concerted Action Programme, organised by the International Life Sciences Institute, Europe (ILSI). The aim of the FOSIE project is to review the current state of the science of risk assessment of chemicals in food and diet, by consideration of the four stages of risk assessment, that is, hazard identification, hazard characterisation, exposure assessment and risk characterisation. The contribution of animal-based methods in toxicology to hazard identification of chemicals in food and diet is discussed.

The importance of first applying existing technical and chemical knowledge to the design of safety testing programs for food chemicals is emphasised.

There is consideration of the presently available and commonly used toxicity testing approaches and methodologies, including acute and repeated dose toxicity, reproductive and developmental toxicity, neurotoxicity, genotoxicity, carcinogenicity, immunotoxicity and food allergy. They are considered from the perspective of whether they are appropriate for assessing food chemicals and whether they are adequate to detect currently known or anticipated hazards from food. Gaps in knowledge and future research needs are identified; research on these could lead to improvements in the methods of hazard identification for food chemicals. The potential impact of some emerging techniques and toxicological issues on hazard identification for food chemicals, such as new measurement techniques, the use of transgenic animals, assessment of hormone balance and the possibilities for conducting studies in which common human diseases have been modelled, is also considered.


Examples of nutrient imbalances in rat feeding:

Methionine deficiency in rats fed soy protein induces hypercholesterolemia and potentiates lipoprotein susceptibility to peroxidation.
Metabolism. 1995 Sep;44(9):1146-52.
Moundras C, Rémésy C, Levrat MA, Demigné C.

A number of studies have provided evidence that plant proteins, especially soy protein, have a cholesterol-lowering effect as compared with casein. However, dietary supply of sulfur amino acids may be deficient when soy protein is present in the diet at a suboptimal level, which could affect lipid metabolism. Accordingly, in rats fed 13% protein diets, soy protein feeding resulted in a cholesterol-increasing effect (+18%), which could be counteracted by methionine supplementation (0.4%). In contrast, soy protein was effective in decreasing plasma triglyceride, as compared with levels in rats fed casein; this triglyceride-lowering effect was entirely abolished by methionine supplementation. The hypercholesterolemic effect of soy protein was characterized by a higher cholesterol content in low-density lipoprotein (LDL) and high-density lipoprotein 1 (HDL1) fractions, together with a marked induction of hepatic hydroxymethyl glutaryl coenzyme A (HMG CoA) reductase activity and to a lesser extent cholesterol 7 alpha-hydroxylase. There was practically no induction of these enzymes, as compared with levels in rats fed casein diets, when the soy protein diet was supplemented with methionine. Very-low-density lipoprotein (VLDL) plus LDL susceptibility to peroxidation was higher in rats fed soy protein than in casein-fed rats, which could reflect in part the lack of sulfur amino acid availability, since methionine supplementation led to a partial recovery of lipoprotein resistance to peroxidation. These findings suggest that amino acid imbalance could be atherogenic by increasing circulating cholesterol and leading to a higher lipoprotein susceptibility to peroxidation.


Dietary protein level and origin (casein and highly purified soybean protein) affect hepatic storage, plasma lipid transport, and antioxidative defense status in the rat.
Madani S, Prost J, Belleville J.
Nutrition. 2000 May;16(5):368-75.

The effects of different proportions (10, 20, and 30%) of dietary casein or highly purified soybean protein on lipid metabolism were studied in growing Wistar rats. Hepatic, plasma and lipoprotein lipid, and protein concentrations, plasma thiobarbituric acid-reactive substance (TBARS) levels, and resistance of red blood cells against free-radical attack were determined after a 4-wk dietary regimen. Compared with the 20% casein diet, the 20% soybean protein diet exhibited similar cholesterolemia but lower plasma triacylglycerol concentrations and very-low-density lipoprotein (VLDL) particle number, as measured by diminished contents of VLDL-triacylglycerol, VLDL-protein, and VLDL-apolipoprotein (Apo) B (B-100 and B-48). The soybean protein diet raised high-density lipoprotein (HDL)(2-3) particle number, as measured by enhanced concentrations of HDL(2-3) cholesterol, HDL-phospholipid, and HDL-ApoA-I.
Increasing casein or soybean protein level (from 10 to 30%) in the diet involved higher VLDL-ApoB (B-100 and B-48), indicating an increase in the number of VLDL particles. Feeding the 30% casein or 30% soybean protein diet enhanced LDL-HDL(1) cholesterol contents. Despite similar HDL(2-3)-ApoA-I levels, the 30% casein diet enhanced the HDL(2-3) mass and its cholesterol concentrations. In contrast, feeding either the 10 or 30% soybean protein diet significantly lowered HDL(2-3) cholesterol and ApoA-I levels. These effects on cholesterol distribution in lipoprotein fractions occurred despite unchanged total cholesterol concentrations in plasma. Feeding 20% soybean protein versus 20% casein involved lower plasma TBARS concentrations. Decreasing casein or soybean protein levels in the diet were associated with higher plasma TBARS concentrations and had a lower resistance of red blood cells against free-radical attack. The present study shows that dietary protein level and origin play an important role in lipoprotein metabolism and the antioxidative defense status but do not affect total cholesterol concentrations in plasma.

VLDL metabolism in rats is affected by the concentration and source of dietary protein.
Madani S, Prost J, Narce M, Belleville J.
J Nutr. 2003 Dec;133(12):4102-6.

The present study was designed to determine if changes in dietary protein level and source are related to changes in VLDL lipid concentrations and VLDL binding by hepatic membranes and isolated hepatocytes. Male Wistar rats were fed cholesterol-free diets containing 10, 20 or 30 g/100 g casein or highly purified soybean protein for 4 wk. Hepatic, plasma and VLDL lipids, VLDL apo B-100 and VLDL uptake by isolated hepatocytes and VLDL binding to hepatic membrane were determined. Increasing casein or soybean protein level (from 10 to 30 g/100 g) in the diet increased VLDL apo B-100, indicating an increase in the number of VLDL particles. VLDL uptake by isolated hepatocytes and VLDL binding to hepatic membrane increased when the protein level increased from 10 to 20 g/100 g in the diet and decreased with 30 g/100 g protein, regardless of protein type. The dietary protein source did not affect plasma total cholesterol concentrations at any protein level. Feeding 20 g/100 g soybean protein compared with casein lowered plasma triglyceride concentrations and VLDL number as measured by decreased VLDL-protein, -phospholipid, -triglyceride, -cholesterol and -apo B-100. VLDL uptake by isolated hepatocytes and VLDL binding to hepatic membrane were higher in rats fed soybean protein than those fed casein. The higher VLDL uptake could be responsible for the hypotriglyceridemia in rats fed soybean protein.


Effect of proteins from beef, pork, and turkey meat on plasma and liver lipids of rats compared with casein and soy protein.
Brandsch C, Shukla A, Hirche F, Stangl GI, Eder K.
Nutrition. 2006 Nov-Dec;22(11-12):1162-70. Epub 2006 Sep 15.

OBJECTIVE: We assessed the effect of dietary proteins isolated from beef, pork, and turkey meat on concentrations of cholesterol and triacylglycerols in plasma, lipoproteins, and liver and the composition of the microsomal membrane (fatty acids, phosphatidylcholine/phosphatidylethanolamine ratio) compared with that of casein and soy protein in rats.
METHODS: Five groups of 12 rats each were fed semisynthetic diets for 20 d that contained 200 g/kg of proteins isolated from beef, pork, or turkey meat or, as controls, casein or soy protein.
RESULTS: Rats fed beef, pork, or turkey proteins did not differ in cholesterol concentrations of plasma, lipoproteins, and liver and in composition of microsomal membrane from rats fed the casein diet. All groups fed a protein from an animal source had
higher very low-density lipoprotein (VLDL) and liver cholesterol concentrations than did rats fed soy protein. However, rats fed pork protein had lower concentrations of triacylglycerols in liver, plasma, and VLDL and lower mRNA concentrations of sterol regulatory element binding protein-1 and glucose-6-phosphate dehydrogenase than did rats fed casein. However, concentrations of plasma and VLDL triacylglycerols in rats fed pork protein were not as low as those observed in rats fed soy protein.
CONCLUSION: Proteins isolated from beef, pork, or turkey meat do not differ from casein in their effects on cholesterol metabolism. Pork protein decreases plasma triacylglycerol concentrations compared with casein but not compared with soy protein. The triacylglycerol-lowering effect of pork protein compared with casein is suggested to be caused by decreased hepatic fatty acid synthesis.


Addendum
One of the challenges of rat trials is having enough statistical power (test sensitivity) to determine if different foods make a difference to rat physiology. (Even if there is a diffrence, it still has to be scrutinised to it correlates over time and dose). But the first step is to find statistically meaning differences and for that you need statistical power. The theory for this is fortunately well developed for other biology situations:

Increasing scientific power with statistical power
MULLER, K. E. AND V. A. BENIGNUS. . NEUROTOXICOL. TERATOL.
14(3), 211-219, 1992.-

A survey of basic ideas in statistical power analysis demonstrates the advantages and ease of using power analysis throughout the design, analysis, and interpretation of research. The power of a statistical test is the probability of rejecting the null hypothesis of the test. The traditional approach to power involves computation of only a single power value. The more general power curve allows examining the range of power determinants, which are sample size, population difference, and error variance, in traditional ANOVA. Power analysis can be useful not only in study planning, but also in the evaluation of existing research. An important application is in concluding that no scientifically important treatment difference exists. Choosing an appropriate power depends on: a) opportunity costs, b) ethical trade-offs, c) the size of effect considered important, d) the uncertainty of parameter estimates, and e) the analyst's preferences. Although precise rules seem inappropriate, several guidelines are defensible. First, the sensitivity of the power curve to particular characteristics of the study, such as the error variance, should be examined in any power analysis. Second, just as a small type I error rate should be demonstrated in order to declare a difference nonzero, a small type II error should be demonstrated in order to declare a difference zero. Third, when ethical and opportunity costs do not preclude it, power should be at least .84, and preferably greater than .90.

Statistical power and design requirements for environmental monitoring
Australian Journal of Marine and Freshwater Research 42(5) 555 - 567
PG Fairweather

Abstract
This paper discusses, from a philosophical perspective, the reasons for considering the power of any statistical test used in environmental biomonitoring. Power is inversely related to the probability of making a Type II error (i.e. low power indicates a high probability of Type II error). In the context of environmental monitoring, a Type II error is made when it is concluded that no environmental impact has occurred even though one has. Type II errors have been ignored relative to Type I errors (the mistake of concluding that there is an impact when one has not occurred), the rates of which are stipulated by the a values of the test. In contrast, power depends on the value of α, the sample size used in the test, the effect size to be detected, and the variability inherent in the data. Although power ideas have been known for years, only recently have these issues attracted the attention of ecologists and have methods been available for calculating power easily.

Understanding statistical power gives three ways to improve environmental monitoring and to inform decisions about actions arising from monitoring. First, it allows the most sensitive tests to be chosen from among those applicable to the data. Second, preliminary power analysis can be used to indicate the sample sizes necessary to detect an environmental change. Third, power analysis should be used after any nonsignificant result is obtained in order to judge whether that result can be interpreted with confidence or the test was too weak to examine the null hypothesis properly. Power procedures are concerned with the statistical significance of tests of the null hypothesis, and they lend little insight, on their own, into the workings of nature. Power analyses are, however, essential to designing sensitive tests and correctly interpreting their results. The biological or environmental significance of any result, including whether the impact is beneficial or harmful, is a separate issue.

Low power, type II errors, and other statistical problems in recent cardiovascular research
J. L. Williams, C. A. Hathaway, K. L. Kloster and B. H. Layne
Department of Physiology and Pharmacology, School of Medicine, University of South Dakota, Vermillion 57069, USA.
Am J Physiol Heart Circ Physiol 273: H487-H493, 1997;
Heart and Circulatory Physiology

Frequently in biomedical literature, measurements are considered "not statistically different" if a statistical test fails to achieve a P value that is less than or = 0.05. This conclusion may be misleading because the size of each group is too small or the variability is large, and a type II error (false negative) is committed. In this study, we examined the probabilities of detecting a real difference (power) and type II errors in unpaired t-tests in Volumes 246 and 266 of the American Journal of Physiology: Heart and Circulatory Physiology. In addition, we examined all articles for other statistical errors. The median power of the t-tests was similar in these volumes (approximately 0.55 and approximately 0.92 to detect a 20% and a 50% change, respectively). In both volumes, approximately 80% of the studies with nonsignificant unpaired t-tests contained at least one t-test with a type II error probability greater than 0.30. Our findings suggest that low power and a high incidence of type II errors are common problems in this journal. In addition, the presentation of statistics was often vague, t-tests were misused frequently, and assumptions for inferential statistics usually were not mentioned or examined.