Predicting Vinorelbine Disposition and Toxicity: Does BSA Provide More Than a “Bad Statistical Association”?

Abstract

The concept that chemotherapeutic agents are administered at a dose to the maximum that a patient can tolerate before the onset of unacceptable toxicity, is still in wide clinical use today. However, the therapeutic range for most cytotoxic anticancer agents is extremely narrow, and in most cases no information is available before therapy on the intrinsic sensitivity of a patient’s tumor to a particular agent or the patient’s tolerability of a given dose. Hence, the dosage of chemotherapeutic agents remains largely empirical and it is typically only normalized to an individual’s body-surface area (BSA). From a pharmacologic perspective, interindividual differences in body size and body composition present a challenging role for clinicians, as their effects on the time course of drug response (ie, toxicity and efficacy) are still poorly understood. BSA does seem to be a biologically plausible size descriptor to consider in drug dose calculation as it is derived from height and weight, and because of the intuitive belief that patients with a larger BSA would require more drug to induce the same drug effects. Moreover, there are some data describing general physiologic characteristics being associated with BSA, including glomerular filtration rate, blood volume, and basal metabolic rate. In contrast, many studies have now convincingly demonstrated that BSA does not account for the marked differences between patients in drug clearance for many widely used agents, which may be due, in part, to large interindividual variability in drug-metabolizing enzyme activity that is known to exist irrespective of body size and body composition. Based on a retrospective analysis involving 33 different drugs, we have previously estimated that BSA-based dosing is statistically significantly associated with a reduction in interindividual pharmacokinetic variability in only approximately 15% of anticancer drugs. This notion has led several investigators to question the routine use of BSA in dosing strategies for anticancer agents. Overall, however, there is a deficit of knowledge on the influence of body size on interindividual pharmacodynamic variability that does not originate from differences between patients in a drug’s pharmacokinetic profile. In an attempt to overcome this information deficit, Wong et al describe the feasibility and pharmacokinetic/pharmacodynamic relationships for vinorelbine in a group of 41 cancer patients treated with a flat-fixed dose of 60 mg in this issue of the Journal of Clinical Oncology. Using rank correlation analyses, Wong et al demonstrate that BSA did not correlate with vinorelbine clearance, suggesting that BSA is not significantly contributing to explaining interindividual pharmacokinetic variability. Interestingly, among the variables tested, they found that BSA and vinorelbine clearance were the only statistically significant and independent predictors of the post-treatment fractional survival of neutrophils, used as a measure of vinorelbine-induced myelosuppression. The notion that BSA may predict toxicity based on an association with interindividual pharmacodynamic—as opposed to pharmacokinetic— variability is intriguing and warrants additional exploration for vinorelbine, as well as other myelosuppressive chemotherapeutic agents. Although the underlying mechanism for the possible link between BSA and hematologic toxicity remains to be elucidated, the previous finding that BSA is also a risk factor for severe leukopenia in patients treated with a vindesine/cisplatin-containing regimen suggests that this effect may not be limited to vinorelbine. Nonetheless, the currently observed association between BSA and vinorelbine-associated neutropenia shows rather wide scatter and hence better tools to individualize chemotherapeutic treatment with this agent are still highly desired. One of such possible tools evaluated by Wong et al is the application of probe drugs to predict the disposition characteristics of vinorelbine by determining the baseline phenotypic activity of two proteins of putative importance for drug distribution and elimination. Using midazolam, a widely used phenotypic probe for the cytochrome P450 isoforms CYP3A4 and CYP3A5, and technetium labeled sestamibi (99mTc-MIBI), which was used to determine the activity of the transporter protein ABCB1 (MDR1; P-glycoprotein), Wong et al found that only the latter was a significant independent predictor of vinorelbine clearance in the 34 assessable patients. Several factors need to be considered to understand the lack of relationship between the observed phenotypic CYP3A4 activity and vinorelbine clearance. Recent studies have indicated that CYP3A4 probably plays only a modest role in the metabolic breakdown of vinorelbine in vivo, and that this pathway is even relatively insensitive to inhibition (by ritonavir) and induction (by rifampin). Furthermore the ability of a probe drug to predict the metabolism of other CYP3A4 substrates will be compromised if differences in transporter affinities are not identified and fully taken into account. It logically follows that only once all factors contributing to the clearance of a particular CYP3A4 substrate drug are known, including nonmetabolic determinants such as reliance on JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 24 NUMBER 16 JUNE 1 2006