SPARSE SAMPLING APPROACH IN RODENT TOXICOKINETIC STUDIES TO REDUCE THE USE OF ANIMALS

Toxicokinetics (TK) studies is generation of kinetic data for systemic exposure and toxicity assessment of the drug. These studies help us to estimate the observed toxicity to that dose. TK evaluation is very important in drug development phase in both regulatory and scientific perspective. There are several guidelines to conduct TK study in animals recommended by regulatory bodies (OECD). TK evaluation is useful in selection of dose, dosing form, alternative dosing route, evaluation of toxicological mechanism, and also used for the setting safe dose level in clinical phases. This TK studies also used to reduces the animal number (replacement, reduction and refinement). On the other hand, TK data are practically used for the purpose of drug discovery such as lead-optimization and candidate-selection. Email:marketing@medicilon.com.cn web:www.medicilon.com

INTRODUCTION Preclinical pharmacokinetics play an important role in drug development. In the case of new human/animal drugs, safety must be established for a particular therapeutic application as regulatory agencies throughout the world review the results of preclinical pharmacokinetic investigation. Recent developments and advances in pharmacokinetics have led to the insight that drug effects can often be better correlated with concentrations of drug in blood or plasma as a measure of systemic drug exposure rather than directly with dose1 . It has been well accepted now that toxicological experiments should include the assessment of the concentration-time course of the investigated agent in order to help in interpreting the toxicological data obtained. This field is commonly referred to as toxicokinetics. For purpose of establishing a safety margin, the extrapolation of data between animals and humans, based on mg/kg doses can give very misleading results. This can be eluded by appropriate toxicokinetic studies. Invariably, there appears to be much greater safety margin when plasma concentration data, which represent actual drug exposure, are used2 . It is generally considered appropriate to determine systemic exposure from area under the plasma/serum concentration-time curve3,4 (AUC) and Cmax which are typically determined using an intensive sampling scheme (10 to 20 time-points)5 . That means multiple blood samples are taken from each individual experimental animal. This is referred to as serial sampling in each individual unit. While the withdrawal of a sufficient number of blood samples from individual animals for AUC determination is usually not difficult in non-rodent toxicity studies, blood sampling from individual animals in rodent toxicity experiments is restricted because of low blood volume6 . The trauma associated with frequent venipuncture and blood loss could therefore be expected to cause adverse changes in physiology, which could interfere with the findings of toxicity studies. Therefore, separate toxicokinetic (with 4 to 5 animals/dose) studies are often designed and conducted simultaneously with toxicity experiments to obtain multiple blood samples from individual animals to fully characterize the AUC and Cmax 7 . These animals in the toxicokinetic study form a satellite group which undergoes the same treatment as the animals in the main toxicity study group. This approach inevitably requires a greater number of animals in a toxicity study, and hence more drug substance and increased labour for dosing and veterinary care. Secondly, it does not help in correlating the plasma drug concentration and toxicity findings within the same animal group. To overcome these problems, several investigators6,8,9 suggested the use of the sparse sampling approach (SSA) within the toxicity study itself. The basic characteristic of SSA is that a complete profile is not sampled in every animal, i.e. serial sampling is not done. Rather each animal contributes to the profile with just one or two samples. Hence blood samples can be generated directly from animals in toxicity studies as one or two samples will not effect the toxicity findings6,10, thereby precluding the need for separate toxicokinetic studies. This also enables direct correlation of plasma drug concentrations and toxicity findings within the same study and hence leads to a more efficient preclinical drug development program. Moreover, it is desirable to use replicates at time points to obtain a better estimate of exposure as inter individual variability is very high in animals11. Secondly, to obtain estimates of standard error and confidence interval, replicates at time points are essential12. This, however, again increases the requirement of more animals in the main toxicity study to characterize AUC. To reduce the use of animals, the use of sparse sampling approach with fewer time points for collecting blood samples have been proposed. Tse et al. 10 have used sparse sampling approach with 5 time points (0.5, 1, 4, 8, 24). Pai et al. 13 have obtained the optimal time points on the basis of concentration-time profile. The present contribution expands on the previous reports of using sparse sampling approach with few time points. A five time point design (including 0 hr data point) was used and time points were chosen based on the inflection points in the plasma concentration-time curve. The present report examines whether one can use sparse sampling approach with few time points in animal toxicokinetics studies and avoid the use of satellite animals and yet achieve the desired pharmacokinetic objectives. MATERIALS AND METHODS Plasma concentration-time data for serial sampling Plasma concentration-time (Cp-t) data in rats using serial sampling were from pharmacokinetic studies with flurbiprofen conducted to assess interstrain and sex differences14. Flurbiprofen was administered as a solution at a dose of 2 mg.kg-1 in a solution of poly (ethylene glycol) 400, to Sprague-Dawley and Wistar rats, in two studies, one study being conducted in each sex. Blood samples were obtained at 9 timepoints at 0 hr (predose) and then at 0.25, 0.5, 1, 2, 4, 6, 8 and 12 hrs postdose from each animal. In both the studies, plasma samples were analysed for flurbiprofen by a HPLC method14. The numbers of animals in each group are summarized in Table 1. Pharmacokinetic analysis for serial sampling Individual peak concentrations (Cmax) were directly determined from the Cp-t data. AUC values were calculated by the linear trapezoidal rule. The mean AUC of all animals in the group over all time points using serial sampling is referred to as AUCtrue and similarly the mean Cmax of all animals is referred to as Cmax,true since they represent estimates of the true mean AUC and true mean Cmax, respectively. Although these values are only estimates, they are useful as points of reference for assessing the performance of SSA procedures using lesser data points. Selection of critical time points and sampling design for sparse sampling approach The full profile (with 9 time points) of one animal from each group was inspected and five time points were chosen. These points included 0 hr and the time points where inflections occurred in the profiles. The point of a curve where it changes from concave to convex or vice versa is called a point of inflection.