Interspecies Comparisons of Micronutrient Requirements: Metabolic vs. Absolute Body Size

In the paper by Prohaska and Brokate (1), “Timing of Perinatal Copper Deficiency in Mice Influences Offspring Survival,” a key observation is made that copper deficiency can have an important and negative effect on the survival of offspring. Further, several approaches are used to extrapolate from data derived from mice to potential application in humans. In this commentary, we wish to address the appropriateness of approaches that are often used for extrapolation. We suggest that when making interspecies comparisons from a nutrition perspective, the strongest case is made when a measure of metabolic body size or food intake, rather than body weight, is used to extrapolate the dosages required for a given response. Although it is a common practice to use body weight as a reference in comparative toxicology studies (2), this can lead to inappropriate conclusions when small animals are used to estimate the amounts of a given nutrient required to produce a deficiency or toxicity in a large animal. This point is illustrated by using values that are recommended to meet essential mineral requirements [see Table 1, Fig. 1, and Refs. (3–7)]. In homeothermic animals, mineral requirements are similar across species if expressed per unit of energy intake or as the concentration in dry food (assuming an energy density equivalent to 3.9 – 4.2 kcal/g or 16 kJ/g). Direct extrapolation to an adult human on the basis of dosages administered to a mouse or rat may be in error by a factor of 10 or more. As an example, most would agree that a daily dose of Zn equivalent to 0.25 mg/kg body weight is sufficient for most humans (i.e., a daily intake of 15 mg of Zn/d for a person weighing 70 kg). For a 30-g mouse, however, an extrapolation based on the data and weight for a human would suggest that only 6 –7 g Zn/d is required. Yet, the actual requirement for the mouse is closer to 60 g Zn/d, assuming that the mouse consumes daily 60 –75 kJ or 3– 4 g of dry food. As a “rule of thumb,” extrapolation from a large animal to a small animal often leads to an underestimate in dosage, whereas extrapolations from a small animal to a large animal lead to an overestimates for the large animal. This point is developed in Table 1, where requirements for selected minerals and a range of species are expressed as a concentration (amount per unit of diet) vs. per kilogram of body weight. Log plots (daily intake vs. body weight) are also given in Figure 1. It is noteworthy that the regression equations that describe the relationships are similar. The characteristics of such equations follow those described by Kleiber, Baldwin and others for energy relationships (8 –10). A review of the comparative toxicology literature also yields numerous examples in which relationships based directly on body weight have resulted in gross overor underestimations of dosages required for a given response (11–14). Mordenti (11) examined the literature for 14 antineoplastic agents and concluded that power equations were required to describe adequately the relationship between a toxic dose and animal weight. Lathrop et al. (12) came to a similar conclusion, using the retention time to reach a given concentration in blood as a correction. It was observed that the retention of technetium 99 (Tc99), injected as pertechnetate, reached 10% of the initial dose in mice in 1 d, whereas 7 d were required in humans. This observation is consistent with the use of power equations that describe the metabolic transfer rate. That is, if the