Oxaliplatin Pharmacokinetics during a Four-Hour Infusion

Colorectal cancer is the fourth most frequent malignancy in the United States, with an incidence of 50 new cases per 100,000 inhabitants per year (1). Approximately half of these cases are diagnosed at localized stages and undergo initial and potentially curative surgery. The remaining patients present with disseminated disease, for which systemic chemotherapy with palliative intention is available. For this situation, the combination of 5-fluorouracil and folinic acid comprises the treatment of first choice, with a response rate of ∼50% (2, 3). To improve on these results, new substances like raltitrexed, capecitabine, trimetrexate, and irinotecan have been evaluated in clinical trials and have shown a low to moderate activity (4, 5, 6). As a new platin analogue, oxaliplatin may offer some advantages over currently used drugs (7) because it lacks cross-resistance to cisplatin in vitro (8, 9, 10, 11, 12, 13, 14) and shows significant activity as single agent (15, 16, 17, 18) and in combination with other drugs (19, 20, 21, 22, 23, 24, 25) in patients with advanced colorectal cancers. The administration of oxaliplatin as a chronomodulated infusion seems, especially, to improve its efficacy (26, 27, 28, 29, 30). However, besides statistical deficiencies of the comparison between continuous and chronomodulated infusions (31), the pharmacokinetics of oxaliplatin have not been analyzed appropriately to select the most effective way of administration (32, 33).

Therefore, a comprehensive pharmacokinetic analysis of oxaliplatin during a constant-rate infusion applied to patients with advanced colorectal cancers was performed in this study.

To detect possible binding of oxaliplatin to plasma proteins during sample preparation, we incubated 10-ml samples of both plasma and full blood with oxaliplatin at concentrations of 100, 500, and 1000 ng/ml. The studies were carried out in triplicate and repeated three times. The temperature at the start of incubation was 37°C; thereafter, samples were stored at 20°C. The incubation was stopped every 15 min for each set of samples, with a maximum incubation period of 3 h. The plasma was ultrafiltrated by Amicon filters NMWL 10,000 (Amicon, Beverly, MA).

Patients aged over 18 years with histologically proven colorectal cancer that was refractory to combination chemotherapies of 5-fluorouracil and leucovorin (34, 35) were entered into the current trial. All patients had normal hepatic function and no signs of ascites or pleural effusion, and all but one had normal creatinine clearances.

All patients received at least two courses of oxaliplatin and 5-fluorouracil chemotherapy via an implantable chamber device as long-term venous access. Oxaliplatin was applied at a dose level of 130 mg/m² in 250 ml of 5% glucose solution as a 4-h infusion on day 1, and 5-fluorouracil was given at a dose level of 375 mg/m² in 50 ml of 0.9% saline as a 24-h infusion through a battery-operated pump on days 1, 8, and 15. Cycles were repeated every 3 weeks.

During the first and second cycle of chemotherapy, 5-ml blood samples were drawn into NH₄-heparinized tubes hourly during the infusion, every 20 min for 2 h thereafter, every hour until 6 h after the end of infusion, and at three time points on the following day. The plasma was separated immediately, and an aliquot was ultrafiltrated as described above. The ultrafiltrate and total plasma samples were stored immediately at −20°C for a maximum of 2 weeks. The total urine excreted during the infusion and, 24 h thereafter, was collected in fractions and stored immediately at −20°C for a maximum of 2 weeks.

The determination of the platin concentrations of the respective specimens was performed by flameless atomic absorption spectrometry using a GBC 904 AA (Maassen, Germany) as described previously (36). The schedule of furnace heating is outlined in Table 1. The method was linear between platin concentrations of 30 ng/ml (limit of detection) and 1000 ng/ml. Both within-day and between-day variations were <10% for plasma and urine samples (Table 2). Protein-bound platin was calculated as the difference between total and free platin for each plasma sample. Analysis of pharmacokinetic results was based on the TOPFIT computer program providing an optimized adaptation of coefficients of variation between the observed and calculated respective data (37).

Prior to therapy all patients gave their informed consent for participation in the current evaluation after having been advised about the purpose and investigational nature of the study as well as of potential risks. The study design adhered to the Declaration of Helsinki and was approved by the local ethics committee prior to its initiation.

Thirteen patients were entered into the current trial between August 1997 and June 1998. Their ages ranged from 38 to 64 years. In nine of them, pharmacokinetic analyses were performed during two cycles of oxaliplatin, whereas in the other three patients, only one cycle was analyzed. One patient suffered from renal insufficiency due to urethral obstruction (maximum creatinine serum = 4.0 mg/dl).

Levels of free platin declined slowly, with a mean half-life of 2.21 ± 0.72 h.

Fitting the respective measurements to a three-compartment-model, plasma levels of platin decline rapidly during a first phase of distribution, followed by a slower decay corresponding to a second phase of distribution, as exemplified for one patient in Fig. 1. During the terminal elimination phase, plasma levels of free platin decline with a half-life of ∼1 day. Accordingly, an initial rapid renal excretion of platin is followed by a retarded elimination, cumulating to 35% of the applied dose at 24 h after the start of infusion.

The pharmacokinetic parameters obtained for all patients with normal renal function are shown in Table 3. With the exception of the initial half-lives, the interpatient variability is extremely low for all parameters. The terminal elimination half-life averages 27.3 ± 10.6 h, and the mean AUC² amounts to 20.17 ± 6.97 μg·h/ml. The renal clearance amounts to 121 ± 56 ml/min, which is ∼50% of the total clearance (222 ± 65 ml/min), resulting in a cumulative renal elimination of platin of 54 ± 20% of the applied dose. No changes in pharmacokinetics were observed between the first and second course of oxaliplatin therapy.

The patient with impaired renal function, who received one dose of oxaliplatin only, shows a more prolonged elimination of free platin. The terminal half-life is 56.3 h, and the AUC amounts to 53.20 μg·h/ml, both parameters being >100% above the mean values of patients without renal insufficiency. Accordingly, the cumulative renal elimination of platin amounts to 34%, which is about half of the values for the latter patients.

The pharmacokinetic parameters for platin bound to plasma proteins are also given in Table 3. The interpatient variability is greater than for free platin. The terminal half-life is 47 ± 64 h, and the total clearance amounts to 26 ± 18 ml/min.

On the basis of the pharmacokinetic parameters listed in Table 3, two simulations of a 12-h infusion of 20 mg/m² oxaliplatin were performed. For the first simulation, the infusion rate of the drug was sinusoidal, which resulted in a maximum plasma level of free platin of 103 ng/ml at 7 h after the start of infusion (Fig. 2). In the second case, a constant-rate infusion was simulated, which resulted in a marginally lower maximum plasma level of free platin of 87 ng/ml (Fig. 3). The general course of the plasma levels of free platin as well as of the cumulative renal elimination of platin was very similar for both simulations.

Insights into the metabolism and elimination of a drug and into interpatient differences in the corresponding parameters require extensive pharmacokinetic analyses and are essential to the selection of the most appropriate way of administration for new substances. Thus, a pharmacokinetically based administration scheme has been shown to enhance the efficacy of cytostatic agents in hematological and oncological disorders (38, 39). A relatively new platin compound, oxaliplatin, is increasingly used to treat patients with advanced colorectal cancer; however, adequate pharmacokinetic evaluations (40, 41, 42, 43, 44, 45), especially those during the chronomodulated application of the drug (33), are lacking. This study provides, for the first time, a comprehensive analysis of the pharmacokinetics of oxaliplatin during a constant-rate infusion.

Taking into account possible changes in drug concentrations of plasma samples during their preparation (46), extensive analyses have been performed to guarantee the validity of the measurements. In agreement with published data (47), with levels of free platin decreasing relatively slowly under conditions simulating the work-up of patient samples (half-lives between 1.35 and 3.11 h), the determination of false low values can be excluded. Therefore, the current pharmacokinetic analyses are valid and are not influenced by sample preparation.

With a relatively low interpatient variability, the results outlined above indicate that oxaliplatin is eliminated triphasically in humans, as has also been suggested for animals (32). This has not been demonstrated yet, as in only six incompletely analyzed patients, three elimination half-lives of the drug were reported (45). However, the latter study lacks sufficient plasma sample numbers during the terminal elimination phase. This resulted in an underestimation of the AUC, which was 11.90 μg·h/ml as compared to 20.18 μg·h/ml in this analysis. In no other study, values for the AUC were reported (33, 40, 41, 42, 43, 44), thus leaving undetermined the availability of the drug following its systemic administration. Having obtained reproducible values for the AUC, this study indicates that the basis for the dosing of oxaliplatin according to the body surface area is appropriate, at least in cases with normal renal function. The values obtained for the total clearance of platin from the plasma (mean, 222 ml/min) are in the range of published data (134–170 ml/min; Ref. 40), as are the data for the amount of renal elimination of platin [mean, 54%; 33% (45) and 53.8% (44)]. In contrast, no data have been reported yet on the renal clearance and the volume of distribution. The renal clearance of 121 ml/min, which is 50% of the total clearance, indicates that about half of the dose applied is removed from the plasma by ultrafiltration and, possibly, by tubular secretion. Accordingly, a relatively high volume of distribution of 349 liters suggests an intensive uptake of the drug into the body tissue, possibly, to a great extent, represented by plasma proteins, as has been described previously (47, 48, 49). This contrasts data on the pharmacokinetics of cisplatin and carboplatin for which volumes of distribution of 19 ± 2 and 17 ± 2 liters have been reported, respectively (50). However, for both latter compounds, values for renal clearances fit within the range of creatinine clearance and are similar to those for oxaliplatin (74 ± 29 and 81 ± 17 ml/min, respectively, versus 121 ± 56 ml/min). Interestingly, although the molecular structure of oxaliplatin differs more from that of cisplatin than does carboplatin, the values for terminal half-life and total clearance of oxaliplatin are in the range of cisplatin parameters (27.3 ± 10.6 versus 36 ± 1 h and 222 ± 65 versus 354 ± 33 ml/min, respectively) and differ substantially from carboplatin data (120 ± 11 h and 107 ± 19 ml/min, respectively).

The data obtained for the pharmacokinetics of protein-bound platin in this analysis correspond to those reported in the literature (33, 40, 41, 42, 43). However, in the latter studies, only values of total platin were given, obscuring the analyses of solely the protein-bound fraction of the drug. Furthermore, once bound to plasma proteins, the bond is nearly as stable as a covalent bond thereby questioning the clinical significance of determinations of this fraction (47).

Finally, from the pharmacological point of view, the chronomodulated application of the drug needs to be evaluated adequately, which has not yet been accomplished (33). Assuming linear pharmacokinetics for oxaliplatin during this mode of administration, the simulation of a chronomodulated application of oxaliplatin clearly results in a similar course of plasma levels of platin and renal elimination, compared to a constant-rate infusion, questioning the value of applying the drug with a sinusoidal infusion rate (Figs. 2 and 3). Thus, the possible improvement of the results of chemotherapy for colorectal cancers achievable by chronomodulated applications (26, 28) may not be based on specific pharmacokinetic behavior of the drug during a sinusoidal infusion rate. Instead, the bases for these observations may be phenomena analyzed in the setting of studies on a dependency of the growth rate of benign cells and even malignant tissue on a diurnal rhythm (30, 51, 52).

To confirm the preliminary results of the current analyses and to evaluate their validity for other application modalities, further studies should focus on the pharmacokinetics of free platin in humans following the application of oxaliplatin as long-term constant-rate and chronomodulated infusions.