The current study was carried out to investigate the variation in carboplatin pharmacokinetics and exposure and to determine the potential importance of therapeutic monitoring following high-dose carboplatin treatment in children. In addition, platinum–DNA adduct levels were measured in peripheral blood leucocytes, to allow a comparison of carboplatin pharmacodynamics and pharmacokinetics alongside clinical toxicity data. Clinical response data were not evaluated due to the potential role played by concomitant chemotherapy, administered either before or following high-dose carboplatin treatment, on the tumour responses observed.
A total of 28 patients were included in the study with the dose of carboplatin on day 1 based on renal function to achieve a target AUC. No significant interoccasion variability in carboplatin clearance was observed in those patients where clearance values were determined on the first and last days of treatment, although mean clearance values were lower on days 3 and 5 vs
comparable day 1 data. Interestingly, these data are in agreement with those published from a phase I study, which reported a statistically significant, but limited, decrease in clearance between days 1 and 5 of high-dose carboplatin treatment (Rubie et al, 2003
On the basis of the observed AUC values in individual patients, carboplatin dose adjustment was carried out in 75% of patients, with a range of dose adjustments up to a maximum of 67% from the initial estimated dose. These dose changes resulted in drug exposures of 84–126% of the target AUC values as defined in the two study protocols. In comparison, estimated carboplatin exposures of 65–213% of the target AUC values would have been attained without dose adjustment, with 32% of patients (9/28) receiving carboplatin AUC values above 25
.min. The likelihood of dose adjustment was not influenced by the method of GFR determination or the dosing equation used to determine initial renal function-based dosing. Based on previous studies showing correlations between carboplatin AUC and drug toxicity following high-dose chemotherapy, it is likely that patients with an AUC greater than 25
.min would have experienced serious side effects if therapeutic monitoring had not been implemented (Huitema et al, 2002
; Kloft et al, 2002
). Indeed, significantly increased frequencies of nephrotoxicity, ototoxicity and peripheral nervous system toxicities were observed in patients with AUC above 24.2
.min in a recently published study in patients with germ-cell cancer (Kloft et al, 2003
In the current study, a positive correlation was observed between renal function (GFR) and the estimated AUC that would have been observed if pharmacologically guided dosing had not been performed. Pretreatment GFR values were significantly higher in those patients whose AUC values would have exceeded the target by 25% without dose modification (n=8), than in those whose AUC values would have been more than 25% below the target AUC (n=5). There was a greater than twofold difference in mean GFR between these two groups (P=0.0139). These data suggest that the current use of dosing equations is less effective when dealing with patients receiving high-dose carboplatin chemotherapy, with children with higher GFR values being more likely to attain AUC values greater than intended, and those with lower GFR values being at a greater risk of under-exposure.
Platinum–DNA adducts were measured at 24
h after carboplatin administration, as this has previously been shown to be the time when peak adduct levels are observed following platinum drug treatment (Peng et al, 1997
). In addition, this allowed the blood sample to be drawn immediately before the next dose of carboplatin being administered in those patients receiving the drug over several days. A strong correlation was observed between the formation of platinum–DNA adducts and carboplatin AUC on day 1 of treatment, with higher adduct levels being observed in peripheral blood leucocytes obtained from patients with higher plasma carboplatin concentrations. Pharmacokinetic–pharmacodynamic comparisons with the platinum drugs have not always shown a consistent positive relationship (Schellens et al, 1996
; Peng et al, 1997
; Veal et al, 2001
). However, carboplatin seems to show a stronger trend towards such a relationship than does cisplatin (Ghazal-Aswad et al, 1999
). The correlation between this pharmacodynamic measure and drug exposure supports the use of dosing based on renal function and adaptive dosing to achieve target carboplatin AUC values. Similarly, the lack of a pharmacokinetic–pharmacodynamic relationship for cisplatin may explain why such an approach to dosing has not been established for this drug.
The current study also suggests that when carboplatin is administered over several days, there is either an increased formation of platinum–DNA adducts or an accumulation of adducts over the treatment period. Despite the complication of carboplatin dose variations between days 1 and 5 of treatment, owing to pharmacokinetic dose adjustment, adduct levels following carboplatin administration on day 5 were between 3- and 10-fold higher than the corresponding levels following day 1 of the treatment. This phenomenon may explain some of the apparent inconsistencies in published pharmacokinetic–pharmacodynamic relationships for this drug. For example, although a strong correlation was observed between platinum–DNA adduct levels and carboplatin systemic exposure following administration on day 1 in the current study, the accumulation of adduct levels observed following several days of treatment negated this relationship when data from all days of treatment were included in a similar analysis. Although we are unable to determine cumulative platinum–DNA adduct levels over the 5-day treatment period, with no information on adduct levels on days 2–4, it is possible that differences in total adduct levels may reflect differences in observed toxicity among some patients, despite comparable cumulative pharmacokinetic exposures in terms of AUC. Indeed, correlations have previously been shown between platinum–DNA adduct levels and leukocytopenia following cisplatin treatment (Veal et al, 2001
Haematological toxicity was observed in the vast majority of patients studied (90%), as anticipated with the use of high-dose carboplatin chemotherapy. The high percentage of patients experiencing CTC grade 3 and 4 haematological toxicity is likely to be associated with the high carboplatin exposures achieved in all patients through adaptive dosing. Estimated exposures as low as 13.0
.min, which would have been achieved in some of the patients studied if dose adjustment had not been carried out, may not have led to the same grade toxicity but may also have been less likely to have resulted in clinical responses. Correlations between decreased carboplatin AUC values and increased rates of relapse have previously been reported in patients with testicular germ-cell tumours (Horwich et al, 1991
). Although an evaluation of clinical response data was not included in this study, owing to the potential influence of concomitant chemotherapy, it is encouraging that no treatment-related deaths were observed in any of the patients on the current study. This is particularly the case in light of the high incidence of treatment-related deaths previously reported with the use of high-dose carboplatin in paediatric patients (Santana et al, 1992
; Jakacki et al, 1997
; Dunkel et al, 1998
Data from this multi-centre study show the feasibility of real-time monitoring of carboplatin pharmacokinetics with adaptive dosing and indicate that this approach is necessary for the attainment of consistent AUC values in individual patients receiving high-dose carboplatin treatment. This approach is now being used in clinical studies in the UK, with the aim of improving efficacy and minimising toxicity of carboplatin in similar high-dose protocols, and is likely to be relevant to the treatment of both childhood and adult cancer patients. The pharmacodynamic data presented here suggest that a strong correlation exists between the pharmacokinetics of carboplatin and the drug–target interaction.