Bulgarian Journal of Veterinary Medicine (2009), 12, No 1, 3−24 COMPARISON OF THE PHARMACOKINETICS OF SEVEN FLUOROQUINOLONES IN MAMMALIAN AND BIRD SPECIES USING ALLOMETRIC ANALYSIS A. M. HARITOVA & L. D. LASHEV Department of Pharmacology, Physiology of Animals and Physiological Chemistry, Faculty of Veterinary Medicine, Trakia University, 6000 Stara Zagora, Bulgaria Summary Haritova, A. M. & L. D. Lashev, 2009. Comparison of the pharmacokinetics of seven fluoroquinolones in mammalian and bird species using allometric analysis. Bulg. J. Vet. Med., 12, No 1, 3−24. Allometric analysis is used to predict the pharmacokinetic behaviour of drugs in animal species where it has not been studied yet. This method was applied to calculate total body clearance, volume of distribution and elimination half-life of seven fluoroquinolone drugs. The results showed that pro- vided information for quinolones’ pharmacokinetics was very close to real data, with the highest accuracy for marbofloxacin. On the contrast, the prediction of pharmacokinetics of enrofloxacin and its active metabolite ciprofloxacin was the most unreliable. Birds should be separately subjected to allometric scaling in order to receive more accurate results. The comparison of data among species showed that in rabbits, pigs, sheep, donkeys and wild animals as gorals, alpakas and oryxes, allomet- ric scaling of fluoroquinolones could not always provide a reasonable accuracy. Therefore, the speci- ficity of metabolism and excretion of a given drug should be taken into account. Key words: allometric scaling, birds, fluoroquinolones, mammals, pharmacokinetics INTRODUCTION Allometric analysis has been used to pre- physiological characteristics as birds and dict the pharmacokinetic behaviour of mammals. Mammalian species could be drugs and to estimate dosage regimens in further divided into carnivores, herbi- animal species that have not been studied vores, or to ruminants and others (Kirk- yet. It also has been used in drug deve- wood & Merriam, 1990; Riond & Riviere, lopment (Mahmood & Balian, 1999) and 1990; Pashov et al., 1997; Riviere et al., in comparison of pharmacokinetics of 1997; Dinev, 2008). This division in different substances between species (Di- groups is based on interspecies differ- nev, 2008). The main assumption of this ences in the physiology and it is aimed to approach is that many physiological pro- predict the specific pharmacokinetic cesses and organ sizes exhibit a power properties of the drugs with a higher ac- law relationship with the body weight of curacy. the species (Mahmood & Balian, 1999). Interspecies pharmacokinetic scaling The allometric scaling is regularly con- has been performed for a large variety of ducted using data for animal species, be- antibacterial agents (Duthu, 1985; Riond longing to taxonomic groups with similar & Riviere, 1990; Pashov et al., 1997; Comparison of the pharmacokinetics of seven fluoroquinolones in mammalian and bird species ... Riviere et al., 1997; Lashev, 1998). The some differences between fluoroquinolo- most recent data concern fluoroqui- nes. nolones as a class of antimicrobial drugs The objective of this study was to as- which is rapidly developing and widely sess the relationship between elimination used in veterinary medicine (Bregante et half-life, volume of distribution at steady- al., 1999; Cox et al., 2004; Cox, 2007). state, and total body clearance to body Pharmacokinetics of fluoroquinolones weight of seven fluoroquinolone drugs in was extensively studied in a number of different species by the method of al- animal species and after different routes lometric scaling. These data could serve of administration. These drugs have simi- for a better understanding of fluoroqui- lar distribution characteristics, however, nolone pharmacokinetics and could be elimination pathways and rates differ con- further used for prediction of pharma- siderably among species and among qui- cokinetic parameters in rare wild and ex- nolones. Less variations were found out otic species or for first-in-animal dose in rates of absorption. Fluoroquinolones selection. are rapidly absorbed to a high extent and well distributed in different tissues with MATERIALS AND METHODS volume of distribution greater than 1 L/kg in all investigated species (Haritova et al., The allometric analysis of pharmacokine- 2006a). Binding to plasma proteins varies tic parameters of enrofloxacin and its me- among species and for different gyrase tabolite ciprofloxacin, danofloxacin, mar- inhibitors, but in most cases it is low bofloxacin, difloxacin, pefloxacin and its (Zlotos et al., 1998). The major differen- metabolite norfloxacin, was performed ces between animals with regard to elimi- using data from previously published nation are connected with active trans- studies (Tables 1−4). Only data for intra- port, intestinal and hepatic metabolism, venously administered drugs, quantitated and renal excretion. Fluoroquinolones are by microbiological assay or HPLC were metabolized by oxidation, demethylation used. The matrices of interest were serum and deethylation (Lefebvre et al., 1998; or plasma. For analysis of each drug, data Anadón et al., 2002). They are excreted for elimination half-life (t1/2β), volume of with urine by glomerular filtration and distribution at steady-state (V ) and tubular secretion, with the exception of d(ss) total body clearance (ClB) were used. difloxacin which is found mainly in the Data for body weights were collected faeces (Fernandez-Varon et al., 2006b). from the same studies and they referred to These data were used in the allometric healthy adult animals. All values were scaling of pharmacokinetic parameters of calculated on the basis of any single pub- enrofloxacin and its major metabolite lished value of pharmacokinetic parame- ciprofloxacin, marbofloxacin, danofloxa- ters versus body weights of the included cin and difloxacin (Lashev, 1998; Bre- animal species from each study. The gante et al., 1999; Cox et al., 2004; Cox, analysis of data for enrofloxacin was per- 2007). Data about allometric analysis in formed for mammals and birds separately birds were not included in the published and together. Because of the lack of investigations. Fluoroquinolones such as enough pharmacokinetic data for avian pefloxacin and norfloxacin were not sub- species, a separate analysis for other jected to analysis. The findings indicated fluoroquinolones was not performed. 4 BJVM, 12, No 1 A. M. Haritova & L. D. Lashev BJVM, 12, No 1 5 Comparison of the pharmacokinetics of seven fluoroquinolones in mammalian and bird species ... 6 BJVM, 12, No 1 A. M. Haritova & L. D. Lashev BJVM, 12, No 1 7 Comparison of the pharmacokinetics of seven fluoroquinolones in mammalian and bird species ... 8 BJVM, 12, No 1 A. M. Haritova & L. D. Lashev BJVM, 12, No 1 9 Comparison of the pharmacokinetics of seven fluoroquinolones in mammalian and bird species ... 10 BJVM, 12, No 1 A. M. Haritova & L. D. Lashev BJVM, 12, No 1 11 Comparison of the pharmacokinetics of seven fluoroquinolones in mammalian and bird species ... The simple allometric approach has been A statistically significant relationship based on the following power function: was found between body weight and Vd(ss) as well as between body weight and Cl Y = a.Wb (1) B when all species were analyzed (Table 5). where Y is the value of the respective The highest intercept for Vd(ss) was found pharmacokinetic parameter (t1/2β; Vd(ss) or for danofloxacin and the lowest − for ClB), а is the coefficient equal to antilog marbofloxacin. The values of y-intercept of c in equation 2, W is the body weight were similar for enrofloxacin and cipro- and b is the exponent of allometric equa- floxacin when data about mammals and tion. birds were analyzed together. The same The log transformation of (1) gives: was valid for pefloxacin and norfloxacin. The lowest value of Cl was calculated log Y = logc + b.logW (2) B for marbofloxacin. where Y is t1/2β, Vd(ss) or ClB, logc is Predicted values of t1/2β, Vd(ss) and ClB the y-intercept and b is the slope. were compared to literature values (Tab- The least squares linear regression les 1−4). The allometric approach had the method was used for estimation of corre- highest predictive power with the lowest lation between pharmacokinetic parame- error with regards to the pharmacokinetic ters of interest and body weight. Statisti- parameters of marbofloxacin, danoflo- cal analysis was done by Statistica 6.1 xacin and norfloxacin. Pigs, rabbits, software (Statistica for Windows, sheep, chickens and turkeys are the ani- StatSoft. Inc., Tulsa, OK, USA). mal species with higher deviation of the predicted vs reported values. RESULTS DISCUSSION Results of the regression analysis con- ducted are listed in Table 5. The values of Simple allometric scaling is an attractive the exponent b for t½β were very low for low-cost and time-efficient alternative to all fluoroquinolones. Its value for Vd(ss) provide reliable predictions of t1/2β, Vd(ss) and ClB was between 0.74 and 1.29 for all and ClB. Despite the risk for deviation of studied drugs, the lowest (0.67) being that the estimated values from the observed of Vd(ss) of danofloxacin. pharmacokinetic parameters in some ca- There was no association between t1/2β ses, interspecies scaling in veterinary me- and body weight in all species and for all dicine could be used to analyze the phar- quinolone drugs of interest. Therefore, macokinetic behaviour of the drugs and to animals were divided into mammals and focus the efforts on providing good ex- birds for allometric scaling of enrofloxa- planations for the observed differences cin, for which enough pharmacokinetic between animal species (Mahmood, data are available. Although the correla- 2007). The experience with allometric tion was improved, a statistically signifi- scaling shows that with higher number of cant relationship between t1/2β and body the analyzed data, including number of weight was not observed. The highest animal species and number of individual value of y-intercept for t1/2β was calcu- studies for each species, the method lated for marbofloxacin. would have the best predictive value (Mahmood & Balian, 1999; Mahmood, 12 BJVM, 12, No 1 A.