QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

Electronic Letters to:

ARTICLES: Keith W. Ward and Brian R. Smith A COMPREHENSIVE QUANTITATIVE AND QUALITATIVE EVALUATION OF EXTRAPOLATION OF INTRAVENOUS PHARMACOKINETIC PARAMETERS FROM RAT, DOG, AND MONKEY TO HUMANS. I. CLEARANCE Drug Metab Dispos 2004; 32: 603-611 [Abstract] [Full text] [PDF]

eLetters: Submit a response to this article

Electronic letters published:

Prediction of human clearance by monkey liver blood flow

Iftekhar Mahmood   (17 August 2004)

Response from KW Ward

Keith W. Ward   (17 August 2004)

Prediction of human clearance by monkey liver blood flow 17 August 2004
Iftekhar Mahmood FDA Send letter to journal: Re: Prediction of human clearance by monkey liver blood flow Mahmoodi{at}cder.fda.gov Iftekhar Mahmood	Ward and Smith, using 103 xenobiotics have proposed a method for the prediction of human clearance from animal data following intravenous administration. Their method utilizes the product of ratio of human and animal liver blood flow and the animal clearance. Based on their data analysis, the authors have concluded that monkey is the most suitable species to predict human clearance using the aforementioned approach. Their work also confirms the previous findings of some other investigators such as: 1. Human clearance cannot be predicted using two animal species (Mahmood, 1996). 2. A good correlation between body weight and a parameter of pharmacokinetic interest does not necessarily guarantee a good prediction of that PK parameter (Mahmood 1998 & 2002, Lave 1999). 3. An exponent of 0.75 or close to it does not guarantee a prediction success (Mahmood 1999). The authors approach is reasonably good and may predict clearance with fair degree of accuracy under certain conditions but it may not have universal application. For example, this approach may predict clearance poorly for drugs that are given orally or for drugs that are renally secreted or biliary excreted. Furthermore, this approach may not be suitable for highly extracted drugs or drugs with high clearance. Looking at the names of drugs and the references provided by the authors it appears that drugs with different physicochemical properties (drugs that are metabolized, renally excreted, renally secreted and biliary excreted) have been lumped together in the analysis. Although number of times most quantitatively predictive drugs were 15 by three-species allometry and were 33 by monkey liver blood flow (LBF), basically there is no difference between three-species allometry and monkey LBF when percentage of correct qualitative prediction (66 vs 71.8) and percentage of incorrect qualitative predictions that were overestimated (57.1 vs 58.6) were calculated. From the fold variance (Table 1 in the manuscript) it also appears that three-species allometry and monkey LBF predict human clearance almost with same degree of accuracy. The Table suggests that there were several drugs that were predicted with more accuracy with three -species allometry than monkey LBF or vice versa. One of the major drawbacks of this study is the application of the simple allometry across all drugs. It is now well-established fact that the simple allometry is not appropriate for all drugs. Rather, based on the exponents of simple allometry, correction factors such as maximum life span potential or brain weight are needed to improve the predictive performance of allometric scaling for the clearance. The authors incorrectly state that it is difficult to establish a priori when maximum life span potential or brain weight can be used. One of the papers cited by the authors (Mahmood and Balian, Xenobiotica, 1996) clearly outlines the conditions under which correction factors can be applied. Had the authors applied the ‘rule of exponents’ they would have found that the exponents indeed help to select a correction factor, hence a substantial improvement in the predicted human clearance. The improvement in prediction in allometric scaling must be clearly understood and defined. For example, if the error in the prediction of clearance is reduced from 50% to 30% by using a correction factor, this should be considered an improvement. From the authors’ description, it is unclear how many drugs had exponents of the simple allometry that were <_0.55. mahmood="mahmood" has="has" suggested="suggested" that="that" if="if" the="the" exponents="exponents" of="of" simple="simple" allometry="allometry" are="are" _0.55="_0.55" then="then" predicted="predicted" clearance="clearance" a="a" drug="drug" provided="provided" is="is" not="not" />liver blood flow in the animals) will be lower than the observed clearance. It does not mean that the predicted clearance will be inaccurate. In fact, the predicted human clearance may be marginally lower (e.g. by 10%) or may be substantially lower (by 50%). Overall, Ward and Smith’s work provides an approach for the prediction of human clearance that may be useful under certain conditions. However, the application of this approach to multiple classes of drugs must be evaluated. This approach should also be compared with the rule of exponents. A comparative study is ongoing in this direction. Iftekhar Mahmood, Ph.D. Food & Drug Administration Rockville, MD 20852 Disclaimer: The views expressed in this commentary are those of the author and do not reflect the official policy of the FDA. No official support or endorsement by the FDA is intended or should be inferred.
Response from KW Ward 17 August 2004
Keith W. Ward GlaxoSmithKline Send letter to journal: Re: Response from KW Ward Keith_W_Ward{at}gsk.com Keith W. Ward	We appreciate Dr. Mahmood's interest in our work, and agree that the liver blood flow method outlined in our paper (Ward and Smith, 2004) may or may not enjoy universal applicability. Indeed, as detailed in our Discussion, we are conducting additional investigations of this method to explore its applicability for oral data and for molecules with varying pharmacokinetic properties or routes of elimination. Additionally, we have been particularly interested in investigating the very allometry- related questions raised here, including the role of the 'rule of exponents' in improving predictivity. In fact, we have recently described the application of the 'rule of exponents' to the interspecies pharmacokinetic scaling of clearance of these 103 molecules (Nagilla and Ward, 2004). We explored 776 different allometric combinations encompassing 27,313 individual outcomes, including correction for brain weight, maximum lifespan potential, or the 'rule of exponents', and discrimination of molecules based on rate of elimination, route of elimination, or physicochemical property. The outcomes of these comparisons were analyzed both qualitatively (number of predictions within 2-fold of the actual value) and quantitatively (mean absolute error, MAE). Regardless of physicochemical property, route of elimination, or clearance rate, allometric techniques (including those based on the 'rule of exponents') provide inferior predictivity compared to the use of the liver blood flow method based on data from the nonhuman primate alone. The percent of compounds with predicted human clearance within 2-fold of the measured value was 68% for monkey liver blood flow compared with 49% for the ‘rule of exponents’. Furthermore, removing the 25 compounds with an allometric exponent <0.55 resulted in no improvement, with only 47% of the remaining compounds having <2-fold error in prediction. The same was observed when considering absolute error; monkey liver blood flow prediction resulted in an MAE of 5.12 mL/min/kg compared to 11.0 for the ‘rule of exponents’. Consequently, we respectfully disagree that application of the ‘rule of exponents’ to allometric scaling will help select a correction factor that would generally result in improved prediction of clearance compared to the liver blood flow method. As noted in our original publication, certainly there are some molecules for which clearance is better predicted by allometry than by the liver blood flow technique. We are very interested in further exploring the molecular features of both the “inliers” and the “outliers” based on the liver blood flow scaling approach, with a goal of ultimately being able to define a priori the chemical space that may be well- or poorly- predicted by a given extrapolation technique. These studies are underway in our laboratory, and will be reported in due course. Nagilla R and Ward KW (2004). A comprehensive analysis of the role of correction factors in the allometric predictivity of clearance from rat, dog, and monkey to human. J Pharm Sci 93:2522-2534. Ward KW and Smith BR (2004). A comprehensive quantitative and qualitative evaluation of extrapolation of intravenous pharmacokinetic parameters from rat, dog, and monkey to humans. I. Clearance. Drug Metab Dispos 32:603-611. Keith W. Ward, Ph.D. GlaxoSmithKline King of Prussia, PA 19406