| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
PHARMACOKINETICS |
From the Department of Pharmacology, Kitasato University School of Medicine, Sagamihara, Kanagawa, Japan (Dr Fujita, Dr Majima, Dr Kumagai); the Clinical Investigation Center, Kitasato University East Hospital, Sagamihara, Kanagawa, Japan (Dr Fujita, Dr Ozaki, Dr Kumagai, Dr Ohtani); the Department of Clinical Pharmacokinetics, College of Pharmacy, Nihon University, Funabashi, Chiba, Japan (Dr Matsumoto); and the Department of Health Science, Kitasato University School of Allied Health Sciences, Sagamihara, Kanagawa, Japan (Dr Ohtani).
Address for reprints: Tomoe Fujita, Clinical Investigation Center, Kitasato East Hospital, Asamizodai 2-1-1, Sagamihara, Kanagawa 228-8520, Japan; e-mail: tomoe{at}kitasato-u.ac.jp.
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: Phase I study • body surface area normalization • area under the time-concentration curve (AUC) • maximum drug concentration (Cmax) • linear regression analysis
In July 2005, the US Food and Drug Administration (FDA) published guidance for estimating the maximum safe starting dose in initial clinical trials for therapeutics in adult healthy volunteers.4 The guidance indicated that toxic effects should be avoided at the initial dose of an investigational new drug, with the exception of cytotoxic, anti-human immunodeficiency virus, and biological agents, because the drugs are administered to healthy volunteers. Therefore, NOAEL, the highest dose tested in an animal species without adverse effects being detected, is used as an indicator of estimating the starting dose in humans, and the human equivalent dose of the NOAEL is calculated by conversion of the animal dose to the human dose based on the normalization to body surface area in most systemically administered drugs. The NOAEL has been considered to be an important indicator for estimating the starting dose in humans, indeed, the most frequently used criterion was a dose less than 1/60 of the NOAEL in determining the starting dose in the first-in-human studies of Japanese-origin new drugs.5
It is known that the surface area rule, an equivalent dose in mg or mg/d is proportional to the body surface area, is best applied to toxicologic and pharmacologic dose extrapolation from animals to humans.6 For cytotoxic agents, doses lethal to 10% of the animals (LD10) in rodents and the maximum tolerated dose (MTD) in dogs both correlated well with the human MTD in analyses based on the surface area rule.7,8 Typically, the starting doses of such cytotoxic agents are considered to be one tenth of the mouse LD10 expressed on the basis of body surface area.9
Drug toxicity is determined by the drug concentration as well as sensitivity to the drug. In recent preclinical studies for investigational drugs, maximum drug concentration (Cmax) and area under the time-concentration curve (AUC), which are indicators of systemic exposure, have been extensively measured along with toxicologic findings in toxicokinetic studies. The ratio between drug exposure at the NOAEL in animals and that at the therapeutic dose level in humans is considered to be the safety margin of a drug in humans.10
The present study focuses on Cmax and AUC data for oral (PO) and intravenous (IV) investigational drugs that have been studied in phase I clinical trials and compares the animal-human ratios of Cmax or AUC between body weight normalization and body surface area normalization. The pharmacokinetic characteristics of investigational drugs that had animal-human Cmax and AUC ratios of one tenth or less are also reviewed. In addition, relationships between Cmax or AUC and body weight in human and animals were examined by linear regression analysis and nonparametric analysis. Also, interspecies relationships, particularly between humans and dog and rat, for Cmax and AUC were examined by linear regression analysis.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Seventy-five PO and 10 IV investigational drugs were included in the study. As shown in Table I, central nervous system drugs accounted for 67% of the PO drugs, followed by cardiovascular and metabolic system drugs (11%), inflammatory and immune system drugs (11%), digestive system drugs (6.7%), and others (5.3%). Similarly, among the IV drugs, central nervous system drugs accounted for 60%, followed by cardiovascular and metabolic system drugs (20%), digestive system drugs (10%), and others (10%). The IM drugs included 2 osteoporosis preparations and an analgesic.
|
Data Extraction
Cmax or AUC data after single administration of a PO investigational drug and AUC data after single or drip infusion of an IV investigational drug at all doses tested in both animals and humans were extracted from the results of pharmacokinetic analysis reports described in the investigator's brochures and phase I study reports. All of the PO drugs and 8 of the IV drugs were administered under fasting conditions. The 2 remaining IV drugs were either administered 1 hour after breakfast or at an unknown time. When pharmacokinetic analyses of Cmax and AUC data did not exist, the actual Cmax was obtained directly from the mean plasma time-concentration curve. The AUC from zero time to infinity was approximated by summing all the trapezoids to the last time point and adding the AUC from the last measurable concentration to infinity. The latter was estimated by the following:
 |
where Clast is the last measurable concentration and t
is the elimination half-life. In animals, Cmax or AUC data obtained from toxicokinetic studies were not used because few studies conducted toxicokinetic studies.
Pharmacokinetic characteristics including protein binding rate, metabolite-parent drug ratio, bioavailability, and urinary excretion rate of the parent drugs were obtained from the investigator's brochures.
Calculation of Animal/Human Cmax and AUC Ratios
For each investigational drug, Cmax or AUC data was plotted against dose, and data exhibiting nonlinear patterns such as saturated absorption and metabolism were omitted by eye fitting. The regression line was then obtained using the least squares method. The percentages of residuals to the predicted data on the regression line were calculated at all doses, and at least three data points within ±20% were averaged. When the number of data points was less than 3, data exhibiting the least residual were chosen.
Cmax and AUC data thus obtained were normalized by dose per body weight or body surface area. Conversion of dose in mg/kg to dose in mg/m2 was conducted according to the method described in the FDA guidance.4 In this guidance, body surface area was approximated from an allometric equation:
 |
where S is body surface area in square meters, W is body mass in kg, and K is a value unique to each animal. The allometric exponent used was 0.67. The equation for converting doses is
 |
The following equation was then derived from the above 2 equations:
 |
To convert the dose in mg/kg to dose in mg/m2, the dose in mg/kg was multiplied by km: 3 for mouse, 6 for rat, 20 for dog, and 12 for monkey. In humans, when actual body weight and height data were available, body surface area was calculated using a modification of the formula of Dubois proposed by Fujimoto and Watanabe,11 which has been widely applied to Japanese people:
 |
where HT is height in meters. When actual body weight and height data were not presented in the reports, the average values from 59 and 50 investigational drugs were used (63.9 kg, 1.72 m).
Cmax and AUC data normalized by dose per body weight or body surface area in animals and humans are expressed as the animal-human ratio. The number of investigational drugs was plotted against the animal-human ratios of normalized Cmax and AUC. The pharmacokinetic characteristics of investigational drugs with an animal-human ratio for Cmax and AUC normalized by body surface area of one tenth or less were surveyed from the investigator's brochures and study reports.
Development of Relationships Between Humans' and Animals' Cmax or AUC and Body Weight and Relationships for Cmax and AUC Between Humans and Animals
Body weight (W) in kg, dose-normalized Cmax (ng/mL/dose), and dose-normalized AUC (ng·h/mL/dose) were transformed logarithmically for the PO and IV drugs, respectively, and fitted by the equation with linear-least squares analysis:
 |
where a is an intercept and b is a slope. Nonparametric analysis was also applied to the same data to examine the relationships between body weight and dose-normalized Cmax or AUC in humans and animals. The analysis was performed on the data subsets individually categorized by the body weight of humans (<63 kg; 63-67 kg; >67 kg) and dose-normalized Cmax (<1 ng/mL/mg; 1-5 ng/mL/mg; 5-10 ng/mL/mg; >10 ng/mL/mg) and AUC (<10 ng·h/mL/mg; 10-30 ng·h/mL/mg; 30-50 ng·h/mL/mg; >50 ng·h/mL/mg). In another analysis, the Cmax or AUC data from humans, dog, and rat, which had relatively large numbers of data sets, were fitted by the equation with linear least-squares analysis:
 |
where a and b are slopes for dog and rat and c is an intercept. Data from the PO and IV drugs were mixed in this analysis. Subgroup analyses were performed based on the bioavailability value or urinary excretion rate of the parent drugs in rat, which were shown to be the largest number of data sets among the animals. In addition, the predicted values for Cmax and AUC in humans, which were calculated from those in animals based on the equation obtained above, were plotted against the observed values for Cmax and AUC in the group of the entire data set and the subgroups, respectively.
Statistical Analysis
Differences in the distribution of the animal-human ratios of Cmax and AUC were compared among animals, including mouse, rat, dog, and monkey, by the non-parametric Kruskal-Wallis test and then compared between each animal using the Mann-Whitney U test. The following analyses were performed using SAS (GLM procedure). Relationships between dose-normalized Cmax or AUC in humans and animals and body weight were analyzed by linear regression and the Jonckheere-Terpstra test. Relationships for dose-normalized Cmax and AUC between humans and animals were analyzed by linear regression. P < .05 was considered to be significantly different.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
Pharmacokinetic Characterization of Investigational Drugs Exhibiting Animal-Human Cmax or AUC Ratios of One Tenth or Less
Table III summarizes pharmacokinetic characteristics such as bioavailability, major metabolite-parent drug ratio, and the protein binding rate of 15 investigational drugs whose animal-human ratios for Cmax or AUC normalized by dose per body surface area were found to be one tenth or less in any of the animals. Low bioavailability or a large first-pass effect was observed with 9 (60%) of the investigational drugs (numbers 1, 2, 3, 6, 7, 12, 13, 14, and 15). Although no bioavailability data were obtained in rats for drug number 9, drug levels in the liver around tmax were found to be low, suggesting a large first-pass effect in rats. Both animal and human bioavailability data were available for drug number 2, and the animal-human AUC ratios were correlated with the bioavailability ratios between animals and humans.
|
Higher ratios of major metabolite-parent drug or intrinsic rates of metabolism were observed for 4 of the investigational drugs (numbers 3, 4, 6, and 7) in the animals showing low AUC ratios. Pharmacokinetic characteristics were not available for drug number 10. A unique metabolite was observed in the animals with low AUC ratios for drug number 8. Similar protein binding rates for the drugs were observed between animals and humans for drugs 7, 9, 12, 13, and 15.
Relationships for Cmax and AUC Among the Animals for the Oral and Intravenous Drugs
Figures 4 and 5 present the Cmax or AUC data per dose as a function of body weight for the PO and IV drugs. Body weight (W) in kg, dose-normalized Cmax (ng/mL/dose), and dose-normalized AUC (ng·h/mL/dose) were transformed logarithmically and fitted by the equation ln Cmax (or ln AUC) = ln a + b ln W with linear least-squares analysis. Equations and correlation coefficients (R2) were as follows: ln Cmax = ln 4.6503-0.8835 ln W, R2 = 0.5574; ln AUC = ln 6.0898-0.7879 ln W, R2 = 0.4331; for the IV drugs, ln AUC = ln 6.8196-0.6838 ln W, R2 = 0.6959. However, these relationships were not significant when linear regression analysis was conducted. Analysis using a non-parametric test for ordered differences among human and animals revealed significant decreases in the relationships (Figures 4A, 4B, and 5).
|
|
Relationships for Cmax and AUC Between Humans and Animals
Results of the linear regression analysis on the relationships for Cmax and AUC between humans and dog and rat are presented in Tables IV and V, respectively. In the groups consisting of the entire data set, the relationships for both Cmax and AUC were significant between humans and dog, whereas they were not between humans and rat. In the subgroup with a bioavailability of 60% or more, the significance of the regression on the relationships for both Cmax and AUC between humans and rat increased in conjunction with increases in R2 (0.8433). Similar results were observed in the subgroup with a urinary excretion rate of the parent drugs of 10% or more (R2 = 0.9551). The predicted Cmax and AUC data based on the above equations were plotted against the observed data for each compound in the groups of the entire data set and each subgroup divided by bioavailability and urinary excretion rate of the parent drugs (Figures 6 A-F and 7 A-F). Correlations increased in the following order: the subgroups of the drugs with bioavailability less than 60% or a urinary excretion rate of the parent drugs less than 10%, those of the entire data set, and those of the drugs with bioavailability of at least 60% or a urinary excretion rate of the parent drugs of at least 10% (Table IV and V).
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
According to the FDA guidance for estimating the maximum safe starting dose, a safety factor should be applied to the human equivalent dose, and the default safety factor used is one tenth. Thus, the pharmacokinetic characteristics of investigational drugs exhibiting a low animal-human Cmax or AUC ratio of one tenth or less were surveyed, and in such cases, drug exposure levels in humans may exceed the nontoxic levels in animals even after applying the safety factor of one tenth to the human equivalent dose. Fifteen investigational drugs were listed, of which 12 (80%) were categorized as central nervous system drugs. This finding may reflect the proportions of the drug population surveyed in the present study. Approximately 70% of these 15 investigational drugs had poor bioavailability and/or large first-pass effects and/or high clearance of the parent drug in the animals with a low Cmax or AUC ratio. Protein binding rates did not seem to influence the low Cmax or AUC ratio. The FDA guidance stated that variable bioavailability in several animals is one of the safety concerns that might increase the safety factor, and poor bioavailability in the test animals used to derive the human equivalent dose suggests a greater possibility for underestimating the toxicity in humans. This finding was confirmed in the present study in terms of increasing drug exposure in humans. It was also demonstrated in the present study that higher metabolic rates in animals compared to humans, as revealed in the in vitro study, should be considered as another safety concern that might increase the safety factor. When there is widely divergent bioavailability in animals, the data from such in vitro intrinsic rates of drug metabolism in both animals and humans might be useful for estimating the bioavailability in humans.
In our retrospective study on 50 first-in-human studies in Japan, we found the NOAEL was adopted in 32 studies. Ratios between starting doses as dose per kg and NOAEL as dose per kg less than 1/100 accounted for about 60% of the studies.5 Accordingly, the starting doses for most of the drugs in first-in-human studies have been determined prudently in practice by applying a large safety factor. When applying a safety factor of more than 100 to the human equivalent dose converted from the NOAEL in rat, dog, and monkey based on body weight, starting doses estimated from all 3 species result in lower levels than those obtained by the body surface area approach applying the default safety factor of 10. Thus, taking into consideration that dose scaling from animals to humans based on body surface area rather than body weight always results in more conservative doses in humans, the approach based on body surface area could fulfill the rationality requirement in determining the starting dose and appears to be superior to that based on body weight.
To find another approach with which to predict human Cmax and AUC from animals, regression analyses taking into account the bioavailability and urinary excretion rate of the parent drugs were performed using logarithmically transformed dose-normalized Cmax and AUC between humans and rat and dog. As shown in Table IV and V, in the analysis on the entire data set, the relationships for Cmax and AUC fitted by the indicated equations were not significant. On the other hand, in the subgroup of drugs with bioavailability of 60% or more and those with a urinary excretion rate of the parent drug of 10% or more that were observed in rat, the significance of the regression as well as R2 increased in the relationships for both Cmax and AUC. Therefore, the equations obtained by these subgroup analyses were used for predicting Cmax and AUC in humans from those in rat and dog. As illustrated in Figures 6 and 7, the goodness of fit in the predictions versus observed values was found in the subgroup of drugs with a bioavailability of 60% or more and those with a urinary excretion rate of the parent drugs of 10% or more compared with the group of the entire data set.
In this study, the animal-human ratios of the drug exposure levels per dose were compared between body weight and body surface area normalization for investigational drugs using pharmacokinetic data from both animal and human studies. The differences in AUC among animals, particularly between rodents and nonrodents, decreased in body surface area normalization. Poor bioavailability and high clearance of the parent drug were shown to contribute to the lower Cmax and AUC ratios, which in turn could lead to underestimating the drug concentrations in the blood in humans by extrapolating the animal data for such drugs.
Although the relationship for body weight and dose-normalized Cmax or AUC in humans and animals did not exhibit a significant regression, the study found that the relationships for the dose-normalized Cmax and AUC in humans and animals including rat and dog were fitted by the equation ln Cmax or AUC human) = a.ln Cmax or AUC (rat) + b.ln Cmax or AUC (dog) + c in the drugs with high bioavailability and a high rate of urinary excretion of parent drugs.
It should be noted that in cases of drugs that acquire toxicity after being metabolized to an unidentified active drug in humans or that have interspecies differences in drug sensitivity,14-17 human toxicity could not be predicted by extrapolating parent drug levels in the animals to humans.
It is concluded that the normalization of animal doses to body surface area may estimate the human equivalent dose safely and reasonably in terms of correcting the differences in AUC between animals and humans. Bioavailability and intrinsic metabolic rate data in a preclinical study could provide useful information with which to consider increasing the safety factor. With respect to the drugs with relatively high bioavailability or those with a urinary excretion rate of the parent drug of 10% or more examined in animals, the equations obtained in the present study might be useful for predicting drug exposure levels in humans from animal data.
The authors thank the pharmaceutical companies that provided the preclinical and human phase I data for the investigational drugs and Mr Stephen McKay for his assistance with editing the manuscript.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
1. Boxenbaum H, DiLea C. First-time-in-human dose selection: allometric thoughts and perspectives. J Clin Pharmacol. 1995;35: 957-966.[Abstract]
2. Thomas M. Study design and assessment of wanted and unwanted drug effects in phase I/II trials. In: O'Grady J, Joubert PH, eds. Handbook of Phase I/II Clinical Drug Trials. Boca Raton, Fla: CRC Press; 1997.
3. Nakashima M. Iyakuhin kaihatsu to rinshoshiken. In: Nakano S, Yasuhara H, Nakano M, eds. Rinshou Yakuri Gaku.Tokyo, Japan: Igaku-Shoind; 2003.
4. Food and Drug Administration. Guidance for industry: estimating the maximum safe starting dose in initial clinical trials for therapeutics in adult healthy volunteers. Available at: http://www.fda.gov/cder/guidance/5541fnl.pdf.
5. Ozaki M, Fujita T, Yokota S, et al. A retrospective study on determining starting doses [in Japanese]. Jpn J Clin Pharmacol Ther. 2006;37, in press.
6. Chappell WR, Mordenti J. Extrapolation of toxicological and pharmacological data from animals to humans. Adv Drug Res. 1991;20: 1-116.
7. Freireich EJ, Gehan EA, Rall DP, Schmidt LH, Skipper HE. Quantitative comparison of toxicity of anticancer agents in mouse, rat, hamster, dog, monkey, and man. Cancer Chemother Rep. 1966;50: 219-244.[Medline] [Order article via Infotrieve]
8. Mordenti J. Man versus beast: pharmacokinetic scaling in mammals. J Pharm Sci. 1986;75: 1028-1040.[Web of Science][Medline] [Order article via Infotrieve]
9. Collins JM, Zaharko DS, Dedrick RL, Chabner BA. Potential roles for preclinical pharmacology in phase I clinical trials. Cancer Treat Rep. 1986;70: 73-80.[Web of Science][Medline] [Order article via Infotrieve]
10. Schwartz S. Toxicokinetics. In: Evans G, ed. A Handbook of Bioanalysis and Drug Metabolism. Boca Raton, Fla: CRC Press; 2004.
11. Fujimoto S, Watanabe T. Studies on the body surface area of Japanese. Acta Med Nagasaki. 1969;14: 1-13.[Medline] [Order article via Infotrieve]
12. Mahmood I, Balian JD. Interspecies scaling: predicting clearance of drugs in humans: three different approaches. Xenobiotica. 1996;26: 887-895.[Web of Science][Medline] [Order article via Infotrieve]
13. Travis CC, White RK. Interspecific scaling of toxicity data. Risk Anal. 1988;8: 119-125.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
14. Tingle MD, Mahmud R, Maggs JL, Pirmohamed M, Park BK. Comparison of the metabolism and toxicity of dapsone in rat, mouse and man. J Pharmacol Exp Ther. 1997;283: 817-823.
15. Ammerschlaeger M, Beigel J, Klein KU, Mueller SO. Characterization of the species-specificity of peroxisome proliferators in rat and human hepatocytes. Toxicol Sci. 2004;78: 229-240.
16. Khan KN, Venturini CM, Bunch RT, et al. Interspecies differences in renal localization of cyclooxygenase isoforms: implications in nonsteroidal anti-inflammatory drug-related nephrotoxicity. Toxicol Pathol. 1998;26: 612-620.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
17. Garattini S. Toxic effects of chemicals: difficulties in extrapolating data from animals to man. Crit Rev Toxicol. 1985;16: 1-29.[CrossRef][Medline]
[Order article via Infotrieve]
CiteULike 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
), rat (
), dog (
), and monkey (
) versus humans. The distribution of A was skewed to the left compared with that of B.
