Impact clinique de la standardisation des doses

Introduction
Dose – banding is a system where chemotherapy doses, calculated on BSA or by other means, are fitted to pre-defined dose ranges or “bands”. For each band a standard dose (mid-point of each band) is provided with standard, pre-prepared syringes or infusion bags used singly or in combination [1]. Standard syringes or infusions may be batch-prepared by the hospital pharmacy or could be purchased from a commercial source.
Dose-banding is different from dose-rounding.

Key advantages of dose-banding include :

• Ready to use, standard doses, virtually eliminates out-patient waiting time between assessment and commencement of treatment.
• Batch preparation of standard doses reduces or eliminates drug wastage, enables prospective quality control testing, reduces risk of compounding errors and enables workload planning in pharmacy and clinic areas. Introduction of Dose Banding in a major UK cancer centre reduced patient waiting times by >60%, reduced drug wastage to zero, and reduced specialist nurse overtime by 80% [2].

There are, however, some clinical concerns about dose-banding. An inter-disciplinary focus group1 expressed concern about the impact of a systematic error arising from dose- banding adding to pre-existing random errors in the prescribing, preparation and administration of chemotherapy. The process of dose-banding would itself influence the magnitude of these errors/deviations from calculated dose. This may be important since there is evidence that reducing the dose intensity of cancer chemotherapy by as little as 15% can significantly affect treatment outcomes [3].

For this reason, and to gain acceptance of dose-banding, the focus group decided that maximum deviation within a band from the calculated dose should (initially) be 5% [1].

However, there is still some resistance to dose-banding among oncologists and oncology pharmacists and this is reflected in a recent UK Cancer Network survey which suggests dose-banding is used for <38% of cancer treatment centres [4] despite clear benefits to patients and healthcare system.

Dose-banding has been incorporated into some clinical trial protocols, and concern was expressed by the NCRN OPCS Committee [5] in the UK that although the use of dose- banding in trials was in some ways desirable, there was a compelling need for evidence to confirm that dose banding would not influence trial outcomes, particularly those involving PK studies.

In view of these concerns, we have initiated 2 clinical studies using pharmacokinetic (pk) measures as surrogates of clinical outcome and toxicity.

Clinical Studies
Our initial studies focused on 5-FU because it was the first drug to be dose-banded, widely used, and data would be applicable to oral 5-FU pro-drugs, such as capecitabine, where tablets create fixed dose intervals. A cross-over study compared pk measures for conventional individualised doses and dose banded 5-FU in FEC regimens. Variations in observed AUC values did not directly correspond to variation from the calculated dose introduced by dose-banding, and although there was no significant difference between AUC values for conventional and dose-banded arms (1530±506 and 1470±428min.ug/ml, respectively), the large CV in each arm (29% and 33%, respectively) indicates intra-and inter- patient variability in FU handling (well documented) and the inherent limitations in BSA-based dosing. Although the study was controlled for circadian variation in FU clearance, we questioned whether the systematic error introduced by dose-banding was simply masked by the random errors and variability inherent with 5-FU.

To further examine the clinical impact of dose-banding on chemotherapy a more rigorous study was designed. With a more robust dose-strategy based on renal function, strong relationship between AUC and dose-limiting toxicity and a simple elimination mechanism, carboplatin was selected. The main outcome measure is AUC, which can be measured accurately and is an acceptable surrogate for both toxicity and effectiveness. This drug allows not only comparisons between AUC_DB and AUC_C, but also variance from the predicted (target) AUC according to the Calvert equation. Low intra-individual variability in carboplatin clearance enabled a cross-over design with individualised and banded doses being administered on successive cycles. A restricted blood-sampling protocol was adopted and medical ethics approval has been obtained.

An update on this in-progress study will be presented, together with additional pharmacoeconomic studies and our experience of more than 8 years of the clinical use of dose-banding.