|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2002 American Society for Clinical Oncology
Cautious Arguments in Favor of Body Surface Area–Based DosingMedizinische Klinik und Poliklinik I, Dresden, Germany To the Editor:In the September 1, 2001 issue of the Journal of Clinical Oncology, de Jongh et al1 put the following claim in the title of their article: "body-surface based dosing does not increase accuracy of predicting cisplatin exposure." I would like to criticize this article in two points. First, the article contains an obvious error at a sensitive point. On page 3,736 (second column) the authors state: "Furthermore, only a weak correlation was found between CLfree and BSA (CLfree = [1.54 ± 0.043] + [0.0055 ± 0.0007] x BSA; r = 0.42), with large variability in Clfree across all studied BSA values in the 90 patients with three pharmacokinetically assessable courses (Fig 4)." However, the line defined by this equation does not even touch the data points in Fig 4, this raises doubts whether the reported value for r might be also incorrect. The second criticism is of more principal nature. Obviously, the optimal studies that relate body-surface area (BSA)-based dosing to pharmacodynamic parameters are almost impossible to perform. Therefore, we have to content ourselves with studies that relate pharmacokinetic (PK) parameters with BSA. Plasma PK may not be the ideal compartment, which has been argued in a previous letter,2 but usually, it is the only one accessible. Three parameters come to mind: area under the curve (AUC), volume of distribution (VD), and clearance. If the assumption is right, that BSA-based dosing should reduce variability, one should expect that AUC is less variable with BSA-based dosing and more variable with flat doses. Therefore, a study using only BSA-based dosing can hardly make direct assumptions about relations between BSA and AUC. In an indirect argument, one would expect that if BSA-based dosing makes sense, either VD or clearance (or both) should increase with BSA. A drug that shows virtually no positive correlation between VD and BSA is difficult to imagine. This alone would result in some sense in adjusting doses somehow to body size if one excludes for a moment the unlikely theoretical possibility that by increasing size a clearance could actually decrease. VD is sometimes difficult to determine reliably (situations of oral absorption, active metabolites, active free fractions, and so on), which does not invalidate the basic principle: "put a drug into a bigger volume and the concentration will decrease." Figure 4, in the study by de Jongh et al,1 implies that at least some correlation between clearance and BSA exists. In terms of reducing the variability of AUC by BSA-based dosing, this adds to the meaningful adjustment to a larger VD. Therefore, the squared correlation coefficient (if r = 0.42, r2 = 0.176) could be regarded as a minimum estimate of the proportion of variance in AUC that can be accounted for by knowing BSA. But how do we interpret this figure? My guess would be, that with flat dosing, the variance of the distribution of AUCs would be expected to increase by 17%, which is probably neither negligible nor of utmost importance. We have as yet no agreement how big an r2 should be to warrant BSA-based dosing. The authors fail to provide arguments why r = 0.42 is not enough. Because BSA calculation is neither expensive nor dangerous for the patient, one could probably agree on a fairly small value. Flat dosing is a good means within a PK study to assess more objectively the relationship between BSA and AUC, but outside these studies, I would recommend using BSA-based dosing if "it is concluded that cisplatin CL is related to BSA"1 as de Jongh et al do. However, this should not reduce our efforts to explain the other admittedly more important factors contributing to the variance. Once we have them, we will think again. REFERENCES
1. de Jongh FE, Verweij J, Loos WJ, et al: Body-surface area–based dosing does not increase accuracy of predicting cisplatin exposure. J Clin Oncol 19: 3733-3739, 2001
2. Order SE, Court W: When the mirror doesn’t reflect it at all. J Clin Oncol 20: 1144, 2002 (letter)
ResponseErasmus MC-Daniel den Hoed Cancer Center, Rotterdam, the Netherlands In Reply:We thank Dr Schuler for his interest and comments on our article that was published in the September 1, 2001 issue of the Journal of Clinical Oncology.1 As pointed out, this article indeed contains an error in the equation that relates body-surface area (BSA) to the clearance (CL) of the unbound (free) fraction of cisplatin. Erroneously, the reported values for slope (a) and intercept (b) were derived from the same data with BSA as a function of CL [ie, BSA = (a x CL) + b]. The correct formula with the coefficients (± SE) from the linear-regression analysis is as follows (Fig 1): CL free (in L/h) = (31.80 ± 4.225) x BSA (in m2) + (-1.985 ± 7.897).
The curvefit coefficients show 95% confidence limits of -17.53 to 13.56 (P = .802) for the intercept and 23.48 to 40.11 (P < .0001) for the slope, with an r value of 0.42 (R2 = 0.176). In line with our original conclusions, this suggests that despite the significant slope, BSA accounts for only a small percentage of the observed kinetic variability. We agree that there is no general consensus on optimal study designs to relate BSA-based dosing to pharmacodynamic outcome and that cutoff levels for significance and r still remain arbitrarily defined. Arguably, the best setting to test this relationship is a randomized comparative analysis with BSA-based versus flat-fixed dosing with appropriate attention to pharmacologic end points, as performed recently for paclitaxel.2 Unfortunately, for the vast majority of agents used in today’s clinical practice no data are available on the use of non-BSA–based dosing, with the exception of agents that do not induce any relevant degree of acute side effects (eg, bleomycin and novel target-based cytostatic drugs).3 Therefore, the issue of whether non-BSA dosing might offer any improvement cannot be evaluated. However, it can be anticipated that implementation of such concepts as flat-fixed dosing would have significant economic implications. The ability to manufacture a unit dose has obvious benefits for the pharmaceutical company involved. Similarly, reconstituting a fixed dose without subsequent individualization for different patients is more efficient and cost-effective than preparing individualized doses, and would eliminate a significant source of error in attempting to obtain precise dosing.4 In addition, and in contrast to the comment that BSA-based dosing is not dangerous, drug administration errors are very common in oncology and are usually the result of systematic error (eg, inaccuracy of the calculation algorithms) and inevitable convergence error (eg, use of inaccurate height and weight for BSA calculation).5 Although a large number of studies have appeared in the literature that clearly question the routine use of BSA in drug dose calculation, the message seems not to have been heard or understood, and many clinicians are still concerned. This concern is based on the intuitive belief that patients with a larger BSA necessarily require more drug to induce the same drug effects. In this context, the statement "put a drug into a bigger volume and the concentration will decrease" is a remarkable and unnecessary oversimplification. Furthermore, there is ample evidence that for some water-soluble agents (eg, doxorubicin and irinotecan) CL actually decreases with an increase in BSA, presumably as a result of poor distribution to adipose tissue.6,7 A recent retrospective investigation involving 33 different drugs and 1,650 adult cancer patients shows that dosing strategies based on BSA may be relevant clinically only in very exceptional cases (Baher et al, manuscript submitted for publication). This is when drug disposition characteristics are confined to the central blood compartment (eg, temozolomide), and to a lesser extent, when renal function plays a principal role in drug elimination (eg, troxacitabine). This can be explained, in part, by known relationships between body size and blood volume and BSA and glomerular filtration rate, respectively.8,9 In any event, we hope that performing exploratory analyses of BSA-CL relationships, such as that performed previously for cisplatin, will encourage investigators to adequately define and optimize dosing strategies for other agents. REFERENCES
1. De Jongh FE, Verweij J, Loos WJ, et al: Body-surface area-based dosing does not increase accuracy of predicting cisplatin exposure. J Clin Oncol 19: 3733-3739, 2001 2. Smorenburg SM, Sparreboom A, Bontenbal M, et al: Randomized-crossover evaluation of body-surface area-based dosing versus flat-fixed dosing of paclitaxel. J Clin Oncol (in press) 3. Felici A, Verweij J, Sparreboom A: Dosing strategies for anticancer agents: The good, the bad and body-surface area. Eur J Cancer 38: 1674-1681, 2002 4. Egorin MJ: Overview of recent topics in clinical pharmacology of anticancer agents. Cancer Chemother Pharmacol 42: S22-S30, 1998 5. Jeraj R, Keall PJ, Siebers JV: The effect of dose calculation accuracy on inverse treatment planning. Phys Med Biol 47: 391-407, 2002[CrossRef][Medline]
6. Rodvold KA, Rushing DA, Tewksbury DA: Doxorubicin clearance in the obese. J Clin Oncol 6: 1321-1327, 1988 7. Miya T, Goya T, Fujii H, et al: Factors affecting the pharmacokinetics of CPT-11: The body mass index, age and sex are independent predictors of pharmacokinetic parameters of CPT-11. Invest New Drugs 19: 61-67, 2001[CrossRef][Medline] 8. Baker RJ, Kozoli DD, Meyer KA: The use of surface area for establishing normal blood volume. Surg Gynecol Obstet 104: 183-189, 1957[Medline] 9. Dooley MJ, Poole SG: Poor correlation between body surface area and glomerular filtration rate. Cancer Chemother Pharmacol 46: 523-526, 2000[CrossRef][Medline]
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2002 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
