|
|||||
|
|
||||||
© 2001 American Society for Clinical Oncology Body-Surface AreaBased Dosing Does Not Increase Accuracy of Predicting Cisplatin ExposureByFrom the Department of Medical Oncology, Rotterdam Cancer Institute (Daniel den Hoed Kliniek) and University Hospital Rotterdam, Rotterdam, the Netherlands. Address reprint requests to Alex Sparreboom, PhD, Department of Medical Oncology, Rotterdam Cancer Institute (Daniel den Hoed Kliniek) and University Hospital Rotterdam, Groene Hilledijk 301, 3075 EA Rotterdam, the Netherlands; email: sparreboom{at}onch.azr.nl
PURPOSE: Most anticancer drugs are dosed based on body-surface area (BSA) to reduce interindividual variability of drug effects. We evaluated the relevance of this concept for cisplatin by analyzing cisplatin pharmacokinetics obtained in prospective studies in a large patient population. PATIENTS AND METHODS: Data were obtained from 268 adult patients (163 males/105 females; median age, 54 years [range, 21 to 74 years]) with advanced solid tumors treated in phase I/II trials with cisplatin monotherapy or combination chemotherapy with etoposide, irinotecan, topotecan, or docetaxel. Cisplatin was administered either weekly (n = 93) or once every 3 weeks (n = 175) at dose levels of 50 to 100 mg/m2 (3-hour infusion). Analysis of 485 complete courses was based on measurement of total and nonprotein-bound cisplatin in plasma by atomic absorption spectrometry. RESULTS: No pharmacokinetic interaction was found between cisplatin and the anticancer drugs used in combination therapies. A linear correlation was observed between area under the curves of unbound and total cisplatin (r = 0.63). The mean plasma clearance of unbound cisplatin (CLfree) was 57.1 ± 14.7 L/h (range, 31.0 to 116 L/h), with an interpatient variability of 25.6%. BSA varied between 1.43 and 2.40 m2 (mean, 1.86 ± 0.19 m2), with an interpatient variability of 10.4%. When CLfree was corrected for BSA, interindividual variability remained in the same order (23.6 v 25.6%). Only a weak correlation was found between CLfree and BSA (r = 0.42). Intrapatient variability in CLfree, calculated from 90 patients was 12.1% ± 7.8% (range, 0.30% to 32.7%). CONCLUSION: In view of the high interpatient variability in CLfree relative to variation in observed BSA, no rationale for continuing BSA-based dosing was found. We recommend fixed-dosing regimens for cisplatin.
IN CLINICAL ONCOLOGY, dose calculations for cytotoxic drugs based on body-surface area (BSA) have become standard practice. The use of BSA has largely resulted from its use in the extrapolation of drug doses used in experimental animals to those considered safe as starting doses for phase I clinical trials in human cancer patients. However, a proper scientific rationale for BSA-based dosing of anticancer drugs in adults is lacking.1,2 One exception is docetaxel, for which BSA was reported to be a good predictor of drug clearance (CL) in a population pharmacokinetic model.3 For many other drugs, including epirubicin4 and oral topotecan,5 the pharmacokinetic and pharmacodynamic behavior cannot be predicted reliably by BSA, because several other factors, like intestinal absorption, kidney function, and activity of key enzymes in metabolic pathways, are more important denominators. For example, the finding that carboplatin CL is closely related to glomerular filtration rate has resulted in dosing of this agent according to creatinine CL by the use of an alternative dosing algorithm to achieve a target measure of systemic exposure.6 However, for most agents, including cisplatin, there still is no better method of dose individualization than the nonevidence-based (though accepted) use of BSA-adjusted dosing. The aim of the present study was to investigate the utility of BSA in dosing of the anticancer drug cisplatin. Cisplatin (cis-diaminedichloroplatinum) is a frequently applied agent with a broad spectrum of activity against solid tumors, including testicular, ovarian, bladder, lung, and head and neck cancers. With the use of optimal hydration measures, including supplementation of potassium and magnesium, the dose-limiting toxicity of cisplatin has changed from nephrotoxicity toward neurotoxicity and ototoxicity. Myelosuppression is generally moderate. Because of this toxicity profile, there is no useful short-term clinical marker for dose-limiting toxicity and dose adjustment. Earlier studies have shown that the area under the plasma concentrationtime curve (AUC) of unbound cisplatin was significantly related to important pharmacodynamic end points.7-11 For example, the AUC of free cisplatin and the platinum-DNA adduct formation measured in leukocytes were strongly correlated. The same applies to the grade of thrombocytopenia (which proved to be a dose-limiting side effect of weekly administration of cisplatin at a dose of 80 mg/m2), as well as to the tumor response.12 Here, we have analyzed pharmacokinetic data of 268 adult cancer patients who were treated with either cisplatin monotherapy or cisplatin-based combination therapy in several prospective studies. Intra- and interpatient variability in CL of total and free cisplatin was determined. In addition, the interpatient variability in BSA was calculated and compared with the interindividual variability in pharmacokinetic behavior of cisplatin.
Patient Selection All patients studied had a confirmed diagnosis of malignant solid tumor and were treated with cisplatin monotherapy or cisplatin-based combination therapy (with either oral etoposide,13 intravenous [IV] irinotecan,14,15 oral topotecan,16,17 or IV docetaxel18) in clinical trials that included pharmacokinetic analysis of cisplatin. Detailed clinical and pharmacologic profiles have been documented previously.13-19 According to the inclusion criteria of these trials, all patients were 18 to 75 years of age with an Eastern Cooperative Oncology Group performance status 2, had no previous anticancer therapy for at least 4 weeks, and had adequate hematopoietic (absolute neutrophil count 1.5 x109/L and platelet count 100 x109/L), hepatic (total serum bilirubin 1.25 times the upper limit of normal and AST and ALT levels 2.5 times the upper limit of normal or 5.0 times in case of liver metastases), and renal function (creatinine CL, 60 mL/min) at the time of study entry.
Drug Administration
Clinical Samples
Analytic Methods
Data Analysis
Statistical Evaluation
Patient Population A total of 268 patients (163 males and 105 females; median age, 54 years [range, 21 to 74 years]) were studied in the present analysis (Table 1). The predominant disease types were nonsmall-lung cancer (n = 80 patients), colorectal cancer (n = 55), head and neck cancer (n = 42), and carcinoma of unknown primary (n = 35). The patients received single-agent cisplatin (n = 17) or cisplatin in combination with either docetaxel (n = 61), etoposide (n = 76), irinotecan (n = 57), or topotecan (n = 57). In total, 485 cycles were available for pharmacokinetic analysis.
Cisplatin Pharmacokinetics In line with previous findings,12 concentration-time profiles of unbound and total cisplatin could be best fitted to a one-compartment model with elimination characterized by decay in an apparent mono-exponential manner (Fig 1). By plotting all 485 kinetic data sets, a strong linear relationship was observed between the AUCs of unbound and total cisplatin (AUCtot = [15.7 ± 1.48] + [7.49 ± 0.572] x AUCfree; r = 0.62; Fig 2). Because the fraction of unbound cisplatin to total cisplatin AUC values seemed to be constant amongst patients (overall mean, 0.074 ± 0.016; range, 0.035 to 0.147) and the previous observations that exposure measures based on unbound cisplatin are most closely associated with pharmacodynamic outcome,7-12 we focused on these measures in further analyses. In univariable and multivariate linear-regression analysis, it was observed that cytotoxic comedication, age, disease, and drug dose (in milligrams or milligrams per square meter; Fig 3) were all unrelated to CLfree (r < 0.05). Using one-way ANOVA, neither the coadministered drugs (etoposide, irinotecan, topotecan, or docetaxel), nor the administered cisplatin dose significantly influenced absolute (liters per hour) CLfree (P = .72 and P = .54, respectively). However, statistically significant differences in CLfree were noted between sexes (P < .0001); males had approximately 15% faster absolute CLfree than females (60.1 ± 13.9 v 52.4 ± 14.6 L/h; mean difference [± SE], 7.65 ± 1.77 L/h; 95% confidence interval, 4.17 to 11.1 L/h). In addition, a significant difference was observed in BSA between males (mean, 1.93 ± 0.178 m2; range, 1.45 to 2.40 m2) and females (mean, 1.75 ± 0.161 m2; range, 1.43 to 2.27 m2) in the present study (P < .00001). However, the apparent CL of unbound cisplatin remained significantly different even when expressed relative to BSA (33.2 ± 6.94 v 28.9 ± 7.69 L/h/m2; P < .001).
Kinetic Variability and Role of BSA In the entire patient population, CLfree ranged between 31 and 116 L/h (mean, 57.1 ± 14.7 L/h), with an interpatient variability of 25.6%. However, in these 268 patients, BSA varied between 1.43 to 2.40 m2 (mean, 1.86 ± 0.19 m2), with an interpatient variability of no more than 10.4%. When CLfree was corrected for BSA, the individual variability remained in the same order (ie, 23.6% v 25.6%). 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). Statistically significant differences in CLfree for the three studied BSA groups (ie, < 1.7 m2, between 1.7 and 2.0 m2, and > 2.0 m2) were noted, with mean values of 49.7 ± 14.4, 55.7 ± 11.3, and 65.4 ± 16.2 L/h, respectively (P < .05, ANOVA followed by a Dunns test). The interpatient variabilities for these groups were 29.0% (n = 59), 20.2% (n = 133), and 24.8% (n = 76), and these values remained unaltered after correction for individual BSA values (29.7%, 19.8%, and 24.5%, respectively).
The CLtot as measured during the first treatment cycle in all patients ranged from 2.30 to 7.98 L/h (mean, 4.09 ± 0.91 L/h), with an interpatient variability of 22.2%. Intrapatient variability in CLfree was calculated from 90 patients who had complete pharmacokinetic monitoring during three cycles of chemotherapy. The intrapatient variability in CLfree ranged from 0.30% to 32.7% (median, 10.7%; mean, 12.1% ± 7.8%). A significant carryover effect was observed for total cisplatin concentrations, particularly in patients on the weekly schedule, precluding accurate calculation of intrapatient variability in CL. For unbound cisplatin, no alteration in CL was observed after multiple courses (P = .31).
The current clinical practice for dosing of the anticancer agent cisplatin is based on BSA of the individual patient. This use, in turn, is explained by the assumption of a narrow relationship between the CL of the drug and BSA. However, no data in the literature exist to support this assumption for cisplatin. In the present report, we have studied the relationship between cisplatin CL and BSA in a group of 268 patients who were treated with cisplatin in phase I/II clinical studies. In line with previous findings, the concentration-time curves of unbound cisplatin and total cisplatin (ie, bound to plasma proteins [mainly albumin] plus unbound) did not run parallel, suggesting that protein-binding is essentially irreversible.20,21 It has been shown that the drug is primarily eliminated by the kidneys, although within 24 hours of treatment cisplatin exhibits relatively low recovery in urine (approximately 11% to 32%).22 This observation probably reflects the extensive binding of reactive platinum metabolites to plasma proteins and tissues and implies that renal CL is relatively unimportant, at least during the initial phases of drug CL. In view of this restrictive CL of cisplatin, it has been recommended to use unbound cisplatin-concentration data for representative calculation of pharmacokinetic parameters (AUC values and CL). Indeed, as mentioned previously, it has been shown that correlations exist between clinical effects and kinetic parameters of unbound cisplatin.7-12 Thus the interindividual variability in CL of the pharmacologically active unbound cisplatin is an important determinant of treatment outcome. Furthermore, it has been demonstrated in several studies (in which patients received a constant dose in milligrams per square meter) that inappropriate cisplatin AUC was the single most important determinant of toxicity or relapse.12 Here, we found that the coefficient of variation for cisplatin CL expressed in absolute measures or relative to BSA were both in the same order (23.6% v 25.6%) and that BSA was poorly related to unbound cisplatin CL. Thus, given the frequency with which cisplatin AUC will exceed thresholds associated with undesired outcomes even when dose is adjusted for BSA, unbound cisplatin monitoring and/or the need to adjust cisplatin dose based on AUC measured in the individual patient may be required regardless of body-size measures. On the other hand, it is noteworthy that the interindividual variability in BSA of the patients was only 10.4%, indicating that an effect of BSA on cisplatin CL was measurable but not highly contributory. This strongly suggests that other (unknown) factors than BSA could be considered more important predictors of CL and AUC. Using univariable and multivariate regression analysis, we found that various demographic variables or other covariates, including comedication, disease type, and drug dose, were not related to unbound cisplatin CL, which is in agreement with previous population models for cisplatin pharmacokinetics.23,24 Interestingly, it was observed in these models that only BSA and infusion duration impacted on CL, although interindividual variability as well as residual variability remained large (approximately 23% to 36%, respectively). The finding that age is not a covariate on CL is at odds with at least one recent observation of a significant negative correlation between age and CL for both total cisplatin and unbound drug.25 Using a general linear model, these authors found that age was independent of sex and tumor type as a predictor of cisplatin kinetics and suggested that with increasing age (a known determinant of renal function), cisplatin elimination by the kidneys might be reduced, leading to a reduction in overall CL. The basis for these discrepant findings is currently unknown, although they may relate, for example, to patient-specific differences in the studied populations. Interestingly, we observed that male patients had approximately 15% faster CL of unbound cisplatin compared with female patients, whose CL values were 9.0% lower than the mean of all patients. The reason for a lower CL of unbound cisplatin in females is not clear, although it does not seem to be related to differences in prior chemotherapy. Previous studies have revealed similar differences for several other anticancer agents, including irinotecan and topotecan,26 and that major factors responsible for sex-dependent pharmacokinetics are related to differences in body composition, renal elimination, and hepatic function.27 Indeed, we found a significant difference in BSA between males and females, although most importantly, the apparent CL of unbound cisplatin remained significantly different even when expressed relative to BSA. As studies have identified kinetic thresholds for severe toxicity and antitumor activity, the current data suggest that an apparent decrease in cisplatin CL in females could result in enhanced toxicity after fixed-dosing regimens (in milligrams per square meter) as compared with males. However, the relatively small difference found in this analysis indicates that it might not be of clinical relevance and suggests that this issue should be investigated further. The variability in unbound cisplatin CL in the patients we studied is similar to that reported for cisplatin administered to 26 cancer patients at a dose of 80 mg/m2 by infusion over 2, 3.5, or 4 hours (23.6% v 22.9%).24 In general, variability in absolute CL of most anticancer drugs, including docetaxel, etoposide, irinotecan, paclitaxel, and topotecan, is substantially larger (ie, > 30%) than that observed here for cisplatin. One reason for this might be that these agents are primarily eliminated by hepatic P450 oxidative metabolic (phase I) enzymes, the expression of which is highly variable among individuals (up to six-fold) and which is sensitive to induction and inhibition by various other xenobiotics. It is also noteworthy that CL of several other platinum analogs (eg, carboplatin and oxaliplatin) is predominantly driven by glomerular filtration, which may explain increased kinetic variability as compared with cisplatin.28 The intrapatient variability in the AUC and CL of total and unbound cisplatin, measured in a subset of 90 patients sampled during three courses, was also exceptionally low (approximately 12%). Part of this intraindividual variability, particularly for total cisplatin, was caused by a decrease in CL after repeated administration of cisplatin, suggesting that this variation might even be overestimated. The mechanism of decreased cisplatin CL after repeated administration is not completely understood. One explanation is that part of the observed increase in AUC (and hence decrease in CL) may be artificial because drug levels have not fallen to zero at the time of next administration (ie, as a result of carryover effects). It is concluded that cisplatin CL is related to BSA. The relation, however, is weak and other presently unknown factors are probably more important. The effect of BSA on pharmacokinetics was measurable but had little, if any, predictive ability. In view of the relatively high interindividual variability in the CL of unbound drug and the small range in observed BSA in the entire patient population, cisplatin can be added to the list of anticancer agents where BSA-adjusted dosing does not seem more accurate. A careful study to identify alternative clinical or laboratory parameters with predictive value toward cisplatin CL and AUC by a population approach to pharmacokinetic modeling using NONMEM (S.L. Beal and L.B. Sheiner, San Francisco, CA) is currently in progress. Unless better predictors for unbound drug CL are identified, it is recommended (in the therapeutic dose range of 50 to 100 mg/m2) to apply fixed-dosing regimens for cisplatin in adult cancer patients. Currently, a further randomized clinical study is being conducted to fully explore the advantages of this approach, in which simultaneously the need for potential dosage adjustments at extreme BSA values will be investigated.
1. Gurney H: Dose calculation of anticancer drugs: A review of the current practice and introduction of an alternative. J Clin Oncol 14: 2590-2611, 1996[Abstract] 2. Ratain MJ: Body-surface area as a basis for dosing anticancer agents: Science, myth, or habit? J Clin Oncol 16: 2297-2298, 1998 (editorial)[Medline] 3. Bruno R, Vivier N, Vergniol JC, et al: A population pharmacokinetic model for docetaxel (Taxotere): Model building and validation. J Pharmacokinet Biopharm 24: 153-172, 1996[Medline] 4. Gurney HP, Ackland S, Gebski V, et al: Factors affecting epirubicin pharmacokinetics and toxicity: Evidence against using body-surface area for dose calculation. J Clin Oncol 16: 2299-2304, 1998[Abstract]
5.
Loos WJ, Gelderblom H, Sparreboom A, et al: Inter- and intrapatient variability in oral topotecan pharmacokinetics: Implications for body-surface area dosage regimens. Clin Cancer Res 6: 2685-2689, 2000 6. Calvert AH, Newell DR, Gumbrell LA, et al: Carboplatin dosage: Prospective evaluation of a simple formula based on renal function. J Clin Oncol 7: 1748-1756, 1989[Abstract] 7. Campbell AB, Kalman SM, Jacobs C: Plasma platinum levels: Relationship to cisplatin dose and nephrotoxicity. Cancer Treat Rep 67: 169-172, 1983[Medline] 8. Desoize B, Marechal F, Millart H, et al: Correlation of clinical pharmacokinetic parameters of cisplatin with efficacy and toxicity. Biomed Pharmacother 45: 203-208, 1991[Medline] 9. Johnsson A, Höglund P, Grubb A, et al: Cisplatin pharmacokinetics and pharmacodynamics in patients with squamous-cell carcinoma of the head/neck or esophagus. Cancer Chemother Pharmacol 39: 25-33, 1996[Medline] 10. Kelsen DP, Alcock N, Young CW: Cisplatin nephrotoxicity: Correlation with plasma platinum concentrations. Am J Clin Oncol 8: 77-80, 1985[Medline] 11. Reece PA, Stafford I, Russell J, et al: Creatinine clearance as a predictor of ultrafilterable platinum disposition in cancer patients treated with cisplatin: Relationship between peak ultrafilterable platinum plasma levels and nephrotoxicity. J Clin Oncol 5: 304-309, 1987[Abstract] 12. Schellens JHM, Ma J, Planting AST, et al: Relationship between the exposure to cisplatin: DNA-adduct formation in leucocytes and tumour response in patients with solid tumours. Br J Cancer 73: 1569-1575, 1996[Medline] 13. Schellens JHM, Ma J, Planting AST, et al: Individualization of cisplatin chemotherapy using AUC and DNA-adducts in WBC. Proc Am Assoc Cancer Res 36: 1388, 1995 (abstr)
14.
De Jonge MJA, Verweij J, Planting AST, et al: Drug-administration sequence does not change pharmacodynamics and kinetics of irinotecan and cisplatin. Clin Cancer Res 5: 2012-2017, 1999
15.
De Jonge MJA, Verweij J, De Bruijn P, et al: Pharmacokinetic, metabolic, and pharmacodynamic profiles in a dose-escalating study of irinotecan and cisplatin. J Clin Oncol 18: 195-203, 2000
16.
De Jonge MJA, Loos WJ, Gelderblom AJ, et al: Phase I pharmacologic study of oral topotecan and intravenous cisplatin: Sequence-dependent hematologic side effects. J Clin Oncol 18: 2104-2115, 2000
17.
Gelderblom AJ, Loos WJ, De Jonge MJA, et al: Phase I and pharmacological study of increased dose oral topotecan in combination with intravenous cisplatin. Ann Oncol 11: 1205-1207, 2000
18.
Pronk LC, Schellens JHM, Planting AST, et al: Phase I and pharmacologic study of docetaxel and cisplatin in patients with advanced solid tumors. J Clin Oncol 15: 1071-1079, 1997
19.
De Jonge MJA, Sparreboom A, Planting AST, et al: Phase I study of 3-week schedule of irinotecan combined with cisplatin in patients with advanced solid tumors. J Clin Oncol 18: 187-194, 2000
20.
Ivanov AI, Christodoulou J, Parkinson JA, et al: Cisplatin binding sites on human albumin. J Biol Chem 273: 14721-14730, 1998 21. Takada K, Kawamura T, Inai M, et al: Irreversible binding of cisplatin in rat serum. Pharm Pharmacol Commun 5: 449-453, 1999
22.
Bajorin DF, Bosl GJ, Alcock NW, et al: Pharmacokinetics of cis-diaminedichloroplatinum(II) after administration in hypertonic saline. Cancer Res 46: 5969-5972, 1986 23. Schellens JHM, Maliepaard M, Planting AST, et al: Population pharmacokinetics and limited sampling model (LSM) of cisplatin. Proc Am Soc Clin Oncol 16: 206a, 1997 (abstr 719)
24.
Nagai N, Ogata H, Wada Y, et al: Population pharmacokinetics and pharmacodynamics of cisplatin in patients with cancer: Analysis with the NONMEM program. J Clin Pharmacol 38: 1025-1034, 1998 25. Shelley MD, Fish RG, Maughan TS, et al: Patients age correlates with plasma and renal clearance of cisplatin (CDDP). Proc Am Soc Clin Oncol 18: 228a, 1999 (abstr 877) 26. Loos WJ, Gelderblom HJ, Verweij J, et al: Gender-dependent pharmacokinetics of topotecan in adult patients. Anticancer Drugs 11: 673-680, 2000[Medline] 27. Tanaka E: Gender-related differences in pharmacokinetics and their clinical significance. J Clin Pharm Ther 24: 339-346, 1999[Medline]
28.
Graham MA, Lockwood GF, Greenslade D, et al: Clinical pharmacokinetics of oxaliplatin: A critical review. Clin Cancer Res 6: 1205-1218, 2000 Submitted February 23, 2001; accepted June 5, 2001. This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|||||||||||