SEARCH

SEARCH BY CITATION

Keywords:

  • medicinal herbs;
  • dose translation;
  • allometric scaling;
  • conversion;
  • equivalent;
  • correlation

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. POTENTIAL VALUE
  5. DOSE TRANSLATION OF ANIMAL STUDIES
  6. PREDICTABILITY
  7. PRACTICAL APPLICATIONS OF CONVERSION PRINCIPLES
  8. CONCLUSIONS
  9. Conflict of Interest
  10. REFERENCES

Animal studies testing medicinal herbs are often misinterpreted by both translational researchers and clinicians due to a lack of information regarding their predictability, human dose equivalent and potential value. The most common mistake is to design or translate an animal study on a milligram per kilogram basis. This can lead to underestimation of the toxicity and/or overestimation of the amount needed for human therapy. Instead, allometric scaling, which involves body surface area, should be used. While the differences in the pharmacokinetic and pharmacodynamic phases between species will inevitably lead to some degree of error in extrapolation of results regardless of the conversion method used, correct design and interpretation of animal studies can provide information that is not able to be provided by in vitro studies, computer modeling or even traditional use. Copyright © 2013 John Wiley & Sons, Ltd.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. POTENTIAL VALUE
  5. DOSE TRANSLATION OF ANIMAL STUDIES
  6. PREDICTABILITY
  7. PRACTICAL APPLICATIONS OF CONVERSION PRINCIPLES
  8. CONCLUSIONS
  9. Conflict of Interest
  10. REFERENCES

Animal studies on medicinal herbs are easily misinterpreted due to the lack of information regarding their predictability, potential value and human dose equivalent. One of the most common misinterpretations is to assume that isometric scaling, which involves a simple conversion by body weight, is appropriate for conversion of animal dose to humans. This type of scaling rarely correlates to the correct dose, and it is not correlated to oxygen utilization, blood volume, renal function, capacity to retain heat or most biological factors (Liu and Chen, 2001). Isometric scaling of animal studies may lead practitioners or scientists to underestimate the toxicity and/or overestimate the amount needed for therapy (Greaves et al., 2004). Instead, allometric scaling, which involves body surface area (BSA) should be used. Allometric scaling attempts to compensate for the fact that larger animals normally have a slower metabolic rate and therefore require a smaller drug dose on a milligram per kilogram basis (Dubois and Dubois, 1915; Valentin et al., 2009; Sharma and Mcneill, 2009).

The goal of this article is to review the most current information regarding potential value, dose extrapolation and predictability of animal studies on medicinal herbs. In cases where data on herbal extracts were not available, we applied relevant data from pharmaceutical studies.

POTENTIAL VALUE

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. POTENTIAL VALUE
  5. DOSE TRANSLATION OF ANIMAL STUDIES
  6. PREDICTABILITY
  7. PRACTICAL APPLICATIONS OF CONVERSION PRINCIPLES
  8. CONCLUSIONS
  9. Conflict of Interest
  10. REFERENCES

Animal studies on medicinal herbs can provide valuable information, even if the herb has been traditionally used over hundreds of years. The example of analgesic nephropathy emphasizes the importance of animal studies on herbs and or drugs that have been combined. Aspirin and extracts of Salix daphnoides et spp. (Willow bark) containing salicin have some inherent nephrotoxicity (Schwarz, 1993; Molland, 1978). Likewise, phenacetin, which was introduced in 1887, is associated with slight amounts of nephrotoxicity (Whelton, 1999). However, when the two are combined and used over a number of years, as was common in the first half of the 20th century in Europe, Australia and the United States, they cause analgesic nephropathy involving renal papillary necrosis and chronic interstitial nephritis. Renal papillary necrosis is readily induced in experimental animals with combinations of aspirin and phenacetin (Molland, 1978; Kincaid-Smith, 1968). This lesson from the past has taught us that combinations of herbs or drugs may result in levels of toxicity that are not seen with one agent individually and that common combinations should have some preclinical testing performed. Other pharmacological safety studies using animals can be performed to investigate the potential undesirable pharmacodynamic effects of a single substance on physiological functions (Valentin et al., 2009). There is often very little known about the pharmacodynamics of a medicinal herb until such studies are performed. Animal studies lasting 4 weeks with subsequent histologic evaluation of organs may yield important information about insidious organ damage or inflammation that the herb or drug may cause (Greaves et al., 2004).

New, or different, extraction processes can yield different individual phytochemicals, which may affect efficacy and or toxicity. For instance, when garlic cloves are crushed into vegetable oil, the main product is diallyl trisulphide, whereas when crushed into water, the sulphur containing components convert to allysulphides (Wojcikowski et al., 2007a). The difference in the final phytochemical makeup of garlic supplements is thought to be the reason for at least some of the discrepancy between results of human supplementation studies (Wojcikowski et al., 2007a). In other instances, the polarity of the extraction solvent may result in the extraction of different constituents which may offer more, or less, benefit or toxicity (Wojcikowski et al., 2007b). Well-planned combinations of in vitro and, when justified, in vivo animal testing of these extracts may help identify these differences prior to human trials and usage.

One of the most common types of herbal studies on animals is one that attempts to discover the mechanism by which the herb acts. Once discovered, this information can be valuable in determining whether the herb should be investigated for increasing the benefits of another herb or drug with known actions. For instance, in our own studies, we found that extracts of Angelica sinensis and Astragalus membranceous decreased renal fibrosis by decreasing fibroblast activation, collagen deposition and tubular apoptosis (Wojcikowski et al., 2010). Since the main effects did not involve the renin-angiotensin mechanism, we hypothesized that the herbal extract combined with angiotensin converting enzyme inhibitors would provide more protection against renal fibrosis than either drug separately. We have found that this is true in experimental animals and await further studies to determine how this correlates to humans (Wojcikowski et al., 2010).

DOSE TRANSLATION OF ANIMAL STUDIES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. POTENTIAL VALUE
  5. DOSE TRANSLATION OF ANIMAL STUDIES
  6. PREDICTABILITY
  7. PRACTICAL APPLICATIONS OF CONVERSION PRINCIPLES
  8. CONCLUSIONS
  9. Conflict of Interest
  10. REFERENCES

There have been numerous methods devised to predict drug effect based by body weight; however, the most accepted method is still based on surface area, as presented by DuBois and DuBois in 1915 (Du Bois and Du Bois, 1916; Dubois and Dubois, 1915). These authors postulated that the correct dose of a drug is proportional to the surface area in the two species rather than the body weight. In other words, larger animals generally require a smaller drug dose on a milligram per kilogram basis.

The history of this assumption began in 1883, when Rubner discovered that small animals utilized relatively more oxygen and produced more heat than larger animals (Rubner, 1883). Later, Dreyer and Ray studied blood volume of mammals, finding that while the ratio of blood volume to body weight in animals decreased with increasing animal weight, the relation of blood volume to BSA was constant (Dreyer and Ray, 1911). These early experiments led others to postulate about the evolution of various species. Since larger animals have less capacity for losing heat than smaller animals, they could have evolved new metabolic mechanisms that functioned at higher temperatures. Instead, the larger animal species retained similar anatomical features, but developed by functioning at a lower metabolic rate (Sharma and Mcneill, 2009).

While there has been no specific guidance for herbal medicines, the United States Food and Drug Administration (USFDA) has published a formula to calculate human equivalent dose (HED) of pharmaceuticals involving the species’ Km factors (body weight in kg divided by BSA in m2) (USFDA, 2005) Their formula for dose translation is HED (mg/kg) = animal dose (mg/kg) × animal Km/human Km. This formula was presented to determine minimum recommended starting dose for drugs entering into clinical trials; however, it is still the most accepted guideline for animal to human dose conversion of biologically active constituents at this time (Sharma and Mcneill, 2009). Examples of how this relates to a number of different species are presented in Table 1, while examples of the error that is made by using body weight alone are presented in Table 2. When safety is a major concern, the minimum anticipated biological effect level can be used, which is the lowest dose that produces a biological effect of any kind (Duff, 2006). For adverse effects of new medicinal herb extracts, a safety factor of at least 10 should be applied due to the possibility that humans may be more sensitive to the toxic effects of the extract, or that bioavailability of the active components of the herb may be higher in humans than the animal species (USFDA, 2005).

Table 1. US Food and Drug Administration guidelines for calculation of human equivalent dose
SpeciesBody weight (kg)Body surface area (m2)Km factorRapid calculationa
  1. Examples of body weight, body surface area (BSA) and Km factors for use in calculation of human equivalent dose (HED). Km factors are derived by dividing body weight in kg by BSA in m2. The formula for dose translation recommended by the US Food and Drug Administration is

  2. HED (mg/kg) = animal dose (mg/kg) × (animal Km/human Km).(USFDA, 2005) Toxicity studies should include a safety factor of 10. This table is not to be used for calculating HED for children (refer to text).

  3. aFor rapid calculation of HED if the animal dose is known, divide animal dose in mg/kg by the number in this column. The result is in mg/kg.

  4. aFor rapid calculation of the animal dose if the human dose is known, multiply the mg/kg/day human dose by the number in the rapid calculation column and then multiply by the weight of the animal. A veterinarian should be consulted prior to administration of biologically active substances to pets.

Human adult601.6237 
Mouse0.0200.007312.3
Hamster0.0800.01657.4
Rat0.1500.02566.2
Ferret0.3000.04375.3
Guinea pig0.4000.0584.6
Domestic cat2.50.25103.7
Rabbit1.80.15123.1
Monkey30.25123.1
Dog100.50201.8
Pig200.74271.4
Gelding5406.4840.45
Table 2. Examples of the differences made in dose conversion with isometric versus allometric scaling
SpeciesAnimal dose (mg/kg)Isometric scaling of dose to 60 kg human (mg)Allometric scaling of same dose to humans (mg/kg)Allometric scaling for a 60 kg human (mg)
  1. Example of the error made with isometric scaling (direct extrapolation on a mg/kg basis) of a 100 mg/kg dose from animals to humans. A 100 mg/kg dose in any species is extrapolated to 6000 mg in a 60 kg human if isometric scaling is used. Allometric scaling is preferred.

Mouse10060008.130081487.8049
Hamster100600013.51351810.8108
Rat100600016.12903967.7419
Ferret100600018.867921132.075
Guinea pig100600021.739131304.348
Cat100600027.027031621.622
Rabbit100600032.258061935.484
Monkey100600032.258061935.484
Dog100600055.555563333.333
Pig100600071.428574285.714
Gelding1006000222.222213333.33

Sharma and McNeil have described drugs which may not be amenable to allometric scaling (Sharma and Mcneill, 2009). These include drugs that are highly protein bound, drugs that undergo extensive metabolism and active transport, drugs that undergo significant biliary excretion or renal secretion, drugs whose targets are subject to significant inter species differences and biological drugs that exhibit significant target-binding effects (Sharma and Mcneill, 2009).

Children

Isometric scaling should not be applied to children, as a child's dose of any substance is generally not a mg/kg fraction of the adult dose. In fact, adjusting dose from adults to children is often as challenging as scaling among species (Sharma and Mcneill, 2009). The processes of absorption, distribution, metabolism and excretion are all immature during infancy, making it potentially dangerous to administer unproven doses of medicinal herbs to those less than 18 months of age. For children older than 18 months, Johnson found that a specific rule involving BSA most reliable: child’ s dose = adult dose x (BSA of the child/BSA of the adult) (Johnson, 2008). Another formula involving a variation of Clark's body weight was also found to be valuable: child's dose = [adult dose × (child's weight/adult's weight]0.75 (Johnson, 2008). A number of authors state that scaling for children must only be used as a last resort, that extrapolation of dosage from animals to children should not be attempted and that doses should be carefully titrated according to response (Johnson, 2008; Lack and Stuart-Taylor, 1997; Sharma and Mcneill, 2009).

PREDICTABILITY

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. POTENTIAL VALUE
  5. DOSE TRANSLATION OF ANIMAL STUDIES
  6. PREDICTABILITY
  7. PRACTICAL APPLICATIONS OF CONVERSION PRINCIPLES
  8. CONCLUSIONS
  9. Conflict of Interest
  10. REFERENCES

There are numerous arguments against the use of animal studies on medicinal herbs or drugs, many of which involve the fact that animal studies do not necessarily predict what will happen in humans (Fisher and Tatlisumak, 2005; Pound et al., 2004). It is true that a significant interspecies variation in any of the components of pharmacokinetics or pharmacodynamics may result in inappropriate extrapolation of animal dose to humans. Another reason for concern is the lack of variability between subjects in animal studies, as the animals normally are young healthy adults that are of homogenous genetic background and are housed and fed uniformly (Olson et al., 2000). Arguments against animal studies have also included the fact that only 37%–44% of animal studies published, that had more than 500 citations, resulted in human randomized trials (Hackam and Redelmeier, 2006; Ioannidis, 2006).

There has been no work drawing correlations from studies testing the benefits or toxicities of medicinal herbs from animals to humans. In fact, there have been few attempts to assess the correlation between the effects of drugs in animals and humans. From these studies, it is clear that there are numerous adverse drug reactions in humans that cannot be detected in animals (Greaves et al., 2004). Although the preliminary analysis of these cases might suggest a poor correlation between non-clinical and clinical data (yielding poor predictive value of animal studies), Valentin and colleagues argue that in most cases, this can be associated with inappropriate, limited non-clinical testing and other factors not associated with the true concordance rate (Valentin et al., 2009). For instance, hepatobiliary toxicity in humans was poorly predicted (50%) from animal studies (Olson et al., 2000), but Greaves explained that this is, at least partly, because aminotransferase enzyme levels were often used, which are relatively insensitive markers of liver toxicity (Greaves et al., 2004). Correlation with human toxicity was much better if the study incorporated histopathology data from animals (Greaves et al., 2004). Greaves further argued that the true concordance level of hepatotoxicity would be higher simply because it is probable that new drugs that produce severe hepatotoxicity in animals are not tested in humans. Finally, another reason for decreased predictability in some animal experiments was that they lacked ideal duration, as studies performed for less than two weeks have a poor predictive value for detection of hepatotoxicity. One month of animal studies followed by histopathological examination of the tissues to determine the expression of pathological change in tissues will detect 99% of the hepato-toxicities that can be detected in animal models (Greaves et al., 2004).

Since animals have no ability to communicate symptoms, animal studies are often criticized for their inability to detect the most common adverse drug reactions including headache, anorexia, dizziness, sleepiness, oedema and flushes (Olson et al., 2000). However, studies aimed at analysing correlations between symptoms in humans and signs in animals have found that dizziness correlates to spontaneous motor activity, oedema correlates with urinary sodium excretion, anorexia correlates to gastric emptying time and headache and malaise correlate to decreased blood pressure in experimental animals (Valentin et al., 2009). Animal studies including the results of the signs just mentioned may have increased concordance rates to outcomes in humans (Valentin et al., 2009).

While the overall concordance rate for animal to human studies was estimated to be about 70% (Olson et al., 2000), avoiding previous mistakes should help increase the concordance (Greaves et al., 2004). Concordance for the cardiovascular and hematopoietic systems appears to be the strongest (80% and 91%, respectively) (Macdonald and Robertson, 2009). The skin and hypersensitivity reactions show the least concordance between effects in animal studies and human patients (Greaves et al., 2004). Unfortunately, at this time, it is unlikely that any combination of solely in vitro studies would reach the level of predictive accuracy seen with whole animal studies; however, there is continuing work in this area, and it is hoped that this work will result in the development of new and improved predictive technologies (Macdonald and Robertson, 2009).

PRACTICAL APPLICATIONS OF CONVERSION PRINCIPLES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. POTENTIAL VALUE
  5. DOSE TRANSLATION OF ANIMAL STUDIES
  6. PREDICTABILITY
  7. PRACTICAL APPLICATIONS OF CONVERSION PRINCIPLES
  8. CONCLUSIONS
  9. Conflict of Interest
  10. REFERENCES

A brief discussion of some of the recent animal studies on medicinal herbs will serve to highlight the practical applications of the conversion principles just discussed. In 2012, Khalili et al. found that 300 and 500 mg/kg dried extract of Hypericum perforatum (hypericum; St. John's wort) once every two days reduced the number of renal calculi in rats by 40% (Khalili et al., 2012). The dried extract was originally obtained using an 80% ethanol: water solvent on the leaves of the plant. To determine the human dose equivalent, 150 mg/kg (the small average daily dose) is divided by 6.2 (as per Table 1), and the result (24.19) is multiplied by a 60 (for a 60 kg human), making the human dose equivalent for a 60 kg person 1451.61 mg/day. Clinical trials testing the extract on humans with depression have used doses of 500 to 1800 mg/day dose, and those trials were well tolerated with side effects (gastrointestinal disturbances, fatigue, dizziness, confusion, dry mouth) near placebo levels (Kasper et al., 2010). Therefore, the study cannot be criticized for using the ‘wrong’ dose; however, it may have been of greater benefit to practitioners if the low dose in the animal experiment had been 50–75 mg/day (100–150 mg/every second day). It is often difficult for scientists to choose the optimal dose in animal experiments testing the benefits of herbs, because if too small a dose is used, there may be no observed effect, decreasing the likelihood of approval for further studies.

Perhaps, the compromise that future studies should consider is to test one dose that is equivalent to the lowest recommended dose for humans and one significantly higher to observe adverse effects and for the scientific community interested in drug discovery. A good example of the use of normal and high doses for determining the benefits of medicinal herbs is the study by Dost et al. (2009), who investigated the effect of a hypericum extract similar to that used by Kasper et al. (2010) on the inflammatory and immune response in rats with induced inflammatory bowel disease (Dost et al., 2009). Rats received doses of 50, 150 and 300 mg/kg/day, which corresponds to HEDs of 483.871, 1451.613 and 2903.226 mg/day in a 60 kg human. The HED of the low dose in that study was near the 500 mg/kg low dose used in clinical trials for depression, while the largest dose (300 mg/kg in rats) was equivalent to over twice the high dose for humans, meaning that the beneficial effects observed in the animals may indeed be relevant to humans at relatively low doses. The safety profile of hypericum has been elucidated (Kasper et al., 2010), so an even higher dose was not necessary.

There are several worthwhile points of discussion in the recent toxicity study on Actaea racemosa (black cohosh) (Mercado-Feliciano et al., 2012). The recommended human dose for black cohosh dried extract is 40 mg per day, which is 0.57 mg/kg/day for a 70kg human or 0.66 mg/kg for a 60kg human. The researchers investigated the effects of a dried extract (obtained using a 50% aqueous ethanol extraction solvent) in rats and mice at doses of 0, 15 (rats only), 62.5 (mice only), 125, 250, 500 or 1000 mg/kg for 90 days. Table 3 extrapolates the HEDs of these animal doses, which are between 3.6 and 241 times (362 and 24193%) the recommended daily dose for a 60 kg human. The results indicated that liver weights were increased in the top two doses in mice and the top dose in rats. Two of the ten rats had mild liver necrosis with no changes in the liver function tests. This supports the recommendations by Greaves et al. (2004) who explain that liver enzyme biochemistry is not a reliable tool for evaluation of liver toxicity in animals and that histological samples should be always be evaluated (Greaves et al., 2004). The study also found a dose-dependent non-regenerative normochromic macrocytic anaemia. The lowest dose that had any effect was 62.5 mg/kg in mice, which extrapolates to an HED of 7.6 or 8.9 times the recommended human dose in a 60 or 70 kg person, respectively. Since the dose at which changes began to occur is near the FDA recommendation of 10 times or higher for toxicity studies, the authors correctly concluded that pharmacokinetic studies of several marker components should be done to obtain a better animal to human comparison. However, given that relatively serious pathology did not occur until 121 and 242 times the equivalent of the recommended human dose, it is unlikely that the extract will be found to be dangerous. Finally, regarding the conversion made in that study, the authors state that ‘in the current study the lowest dose that had an effect (62.5 mg/kg/day) was 125 times the currently recommended amount for daily consumption (≈ 0.5 mg/kg/day) for a 70 kg human’. This is an example of the mistake that is made when direct extrapolation on a mg/kg basis (isometric scaling) is used rather than the preferred method of scaling by allometry. Allometric scaling of the 62.5 mg/kg daily dose in mice yields 5.08 mg/kg in humans or 355.69 mg/day in a 70 kg human, which is 8.9 times the current recommended daily dose.

Table 3. Extrapolation (allometric scaling) of the human equivalent doses (HED)
Dose in mg/kgSpeciesHED in mg/kgHED (60 kg human)% of 60 kg human doseIsometric conversion of same dose (60 kg human)
  1. Extrapolation (allometric scaling) of the HED in the study by Mercado-Feliciano et al. (2012) (29). Isometric conversion (direct extrapolation on a mg/kg basis; final column) is provided for comparison to the HED (60kg human). Isometric scaling is not recommended (refer to text).

15Rat2.419355145.1613362.9032900
62.5Mouse5.081301304.878762.19513750
125Rat20.161291209.6773024.1947500
125Mouse10.1626609.75611524.397500
250Rat40.322582419.3556048.38715000
250Mouse20.32521219.5123048.7815000
500Rat80.645164838.7112096.7730000
500Mouse40.650412439.0246097.56130000
1000Rat161.29039677.41924193.5560000
100012.381.300814878.04912195.1260000

In another study, a dried extract of Centella asiatica (centella; gotu kola) was found to lack evidence of toxicity in mice at very high doses (Chauhan and Singh, 2012). The authors stated that extracts of centella possess antioxidant, cognitive-enhancing and antiepileptic properties. However, the studies cited were performed on extracts that used water as the extraction solvent, while the study by Chauhan and Singh (2012) used acetone as the solvent (Chauhan and Singh, 2012). There would be few (if any) overlapping constituents extracted by acetone when compared to those extracted by water (Wojcikowski et al., 2009), and therefore the results of the study do not relate to aqueous extracts of centella. Most studies use an aqueous-alcoholic extraction process. Since the polarity of an aqueous-alcoholic fluid is between the polarity of water and acetone, there is often some overlap in constituents (Wojcikowski et al., 2009). The yield in that study is also noteworthy, as the amount of material that was extracted with acetone was only 5.7% of the weight of the herb, meaning that the dried extract would be considered a 17.5:1 extract. This is can be important when considering dried solvent extracts relative to the dried plant equivalent. A dried whole plant product is considered 1:1 dry plant equivalent.

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. POTENTIAL VALUE
  5. DOSE TRANSLATION OF ANIMAL STUDIES
  6. PREDICTABILITY
  7. PRACTICAL APPLICATIONS OF CONVERSION PRINCIPLES
  8. CONCLUSIONS
  9. Conflict of Interest
  10. REFERENCES

Animal studies on medicinal herbal extracts are not perfect. The unique evolution of each species is reflected by interspecies differences in protein binding, transport and metabolism of the active constituents in the pharmacokinetic phase and changes in receptor expression, affinity and distribution in the pharmacodynamic phase (Sharma and Mcneill, 2009). This will inevitably lead to some degree of error in extrapolation of results to humans, regardless of the conversion method used. However in many cases, animal studies can provide information that is not able to be provided by in vitro studies, computer modeling or even traditional use or human trials. Properly designed animal studies should carefully consider the appropriate dose using allometric scaling, the best animal model for the goals of the study and adequate duration of the study. Correlative data for potential symptoms and histological evaluation of tissues will provide the most information with the lowest number of animals. While there remains a poor correlation between cutaneous and hypersensitivity reactions, the effect on most organs is usually quite similar, and animal data can be used as additional scientific information that if cautiously interpreted, can increase the safety of medicinal herbs by guiding practitioners to make the most appropriate prescriptions.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. POTENTIAL VALUE
  5. DOSE TRANSLATION OF ANIMAL STUDIES
  6. PREDICTABILITY
  7. PRACTICAL APPLICATIONS OF CONVERSION PRINCIPLES
  8. CONCLUSIONS
  9. Conflict of Interest
  10. REFERENCES
  • Chauhan P, Singh V. 2012. Acute and Subacute Toxicity study of the Acetone Leaf extract of Centella asiatica in Experimental Animal Models. Asian Pac J Trop Biomed 2: S511S513.
  • Dost T, Ozkayran H, Gokalp F, Yenisey C, Birincioglu M. 2009. The effect of Hypericum perforatum (St. John's Wort) on experimental colitis in rat. Dig Dis Sci 54: 12141221.
  • Dreyer G, Ray W. 1911. The blood volume of mammals as determined by experiments upon rabbits, guinea-pigs, and mice; and its relationship to the body weight and to the surface area expressed in a formula. Philos Trans R Soc London Series B: 133160.
  • Du Bois D, Du Bois EF. 1916. Clinical calorimetry: tenth paper a formula to estimate the approximate surface area if height and weight be known. Arch Intern Med 17: 863.
  • Dubois D, Dubois EF. 1915. Fifth paper the measurement of the surface area of man. Arch Intern Med 15: 868.
  • Duff G. 2006. Expert Scientific Group on Phase One Clinical Trials Final Report. Department of Health: London, UK.
  • Fisher M, Tatlisumak T. 2005. Use of Animal Models Has Not Contributed to Development of Acute Stroke Therapies Con. Stroke 36: 23242325.
  • Greaves P, Williams A, Eve M. 2004. First dose of potential new medicines to humans: how animals help. Nat Rev Drug Discov 3: 226236.
  • Hackam DG, Redelmeier DA. 2006. Translation of research evidence from animals to humans. JAMA 296: 17311732.
  • Ioannidis JPA. 2006. Evolution and translation of research findings: from bench to where. PLoS 1: e36. doi:10.1371/journal.pctr.0010036
  • Johnson TN. 2008. The problems in scaling adult drug doses to children. Arch Dis Child 93: 207211.
  • Kasper S, Caraci F, Forti B, Drago F, Aguglia E. 2010. Efficacy and tolerability of Hypericum extract for the treatment of mild to moderate depression. Eur Neuropsychopharmacol 20: 747765.
  • Khalili M, Jalali MR, Mirzaei-Azandaryani M. 2012. Effect of Hydroalcoholic Extract of Hypericum Perforatum L. Leaves on Ethylene Glycol-Induced Kidney Calculi in Rats. Urol J 9: 472479.
  • Kincaid-Smith P. 1968. Analgesic nephropathy and papillary necrosis. Postgrad Med J 44: 807810.
  • Lack J, Stuart-Taylor M. 1997. Calculation of drug dosage and body surface area of children. Br J Anaesth 78: 601605.
  • Liu XD, Chen J. 2001. Prediction of drug clearance in humans from laboratory animals based on body surface area. Eur J Drug Metab Pharmacokinet 26: 249255.
  • Macdonald JS, Robertson RT. 2009. Toxicity testing in the 21st century: a view from the pharmaceutical industry. Toxicol Sci 110: 4046.
  • Mercado-Feliciano M, Cora MC, Witt KL, et al. 2012. An ethanolic extract of black cohosh causes hematological changes but not estrogenic effects in female rodents. Toxicol Appl Pharmacol 263: 138147.
  • Molland EA. 1978. Experimental renal papillary necrosis. Kidney Int 13: 514.
  • Olson H, Betton G, Robinson D, et al. 2000. Concordance of the toxicity of pharmaceuticals in humans and in animals. Regul Toxicol Pharm 32: 5667.
  • Pound P, Ebrahim S, Sandercock P, Bracken MB, Roberts I. 2004. Where is the evidence that animal research benefits humans? BMJ 328: 514517.
  • Rubner M. 1883. Über den einfluss der körpergrösse auf stoff-und kraftwechsel. Z Biol 19: 536562.
  • Schwarz A. 1993. Beethoven's renal disease based on his autopsy: a case of papillary necrosis. Am J Kidney Dis 21: 643652.
  • Sharma V, Mcneill JH. 2009. To scale or not to scale: the principles of dose extrapolation. Br J Pharmacol 157: 907921.
  • USFDA. 2005. Guidance for Industry, Estimating the maximum safe starting dose in initial clinical trials for therapeutics in adult healthy volunteers. Center for Drug Evaluation and Research, USFDA, US Department of Health and Human Services, http://www.fda.gov/cder/guidance/index.htm.
  • Valentin J-P, Bialecki R, Ewart L, et al. 2009. A framework to assess the translation of safety pharmacology data to humans. J Pharmacol Toxicol Methods 60: 152158.
  • Whelton A. 1999. Nephrotoxicity of nonsteroidal anti-inflammatory drugs: physiologic foundations and clinical implications. Am J Med 106: 13S24S.
  • Wojcikowski K, Myers S, Brooks L. 2007a. Effects of garlic oil on platelet aggregation: a double blind placebo controlled crossover study. Platelets 18: 2934.
  • Wojcikowski K, Stevenson L, Leach D, Wohlmuth H, Gobe G. 2007b. Antioxidant capacity of 55 medicinal herbs traditionally used to treat the urinary system: a comparison using a sequential three-solvent extraction process. J Altern Complement Med 13: 103110.
  • Wojcikowski K, Wohlmuth H, Johnson DW, Rolfe M, Gobe G. 2009. An in vitro investigation of herbs traditionally used for kidney and urinary system disorders: potential therapeutic and toxic effects. Nephrology 14: 7079.
  • Wojcikowski K, Wohlmuth H, Johnson DW, Gobe G. 2010. Effect of Astragalus membranaceus and Angelica sinensis combined with Enalapril in rats with obstructive uropathy. Phytother Res 24: 875884.