File(s) under permanent embargo

Reason: Restricted by author. A copy can be supplied under Section 51(2) of the Australian Copyright Act 1968 by submitting a document delivery request through your library or by emailing

Dosing of cytotoxic chemotherapy in adult oncology patients : evidence to inform clinical practice.

posted on 09.02.2017, 03:18 by Dooley, Michael
The administration of cytotoxic chemotherapy is one of the key strategies for the treatment of early and advanced cancer. In Australia, there are over 300,000 hospitalisations per year for the administration of cytotoxic chemotherapy, equating to approximately a third of all cancer related hospitals admissions. As a consequence of the variability in drug exposure and the narrow therapeutic window of cytotoxic chemotherapy, clinicians have adopted various approaches to individualise chemotherapy dosing in an attempt to maximise clinical outcome and minimise toxicity. In this thesis, a series of studies have been undertaken to examine and inform the dosing of cytotoxic chemotherapy in clinical practice. These studies focused on existing clinical practices where evidence to support the practice was lacking. The use of body surface area (BSA) as a means of individualising doses of cytotoxic chemotherapy is the most common dosing method in clinical practice. One of the historical justifications for the BSA approach is the postulated relationship between renal function and BSA. The first paper evaluated the relationship between BSA and glomerular filtration rate (GFR) measured by radiolabelled technetium-99m diethyl triamine penta-acetic acid (Tc99mDTPA) clearance. In 122 adult oncology patients planned for chemotherapy, the correlation between BSA and GFR was r = 0.24 (95% confidence interval: 0.06 -OAO). These results do not support BSA dosing based on a purported relationship with renal function. While BSA is the predominant means for dose individualisation for cytotoxic chemotherapy, there are other instances when BSA is not utilised. Carboplatin is one such example, which is dosed based on a target area under the plasma concentration-time curve (AUC) with two formulae (Calvert and Chatelut) used in practice. The Chatelut formula had been validated in a very small patient cohort, while the Calvert formula has been utilised for 2 decades in clinical practice. Anecdotal clinical experience observed that Significantly different doses are calculated depending on which formula is used and also the method of estimating renal function, for example, Tc99mDTPA or the Cockcroft and Gault (CG) formula. The second paper evaluated i) a comparison of the doses of carboplatin calculated using the Chatelut formula compared with the Calvert formula and ii) a comparison of doses calculated using the Calvert formula when modified with non-isotopic estimations of GFR. When the Chatelut formula was compared the Calvert formula in 122 oncology patients, it was found to statistically over-estimate doses for males (% diff = +22%, P<0.001) and under-estimate doses for females (% diff = -6%, P<0.05). When non-isotopic estimates of GFR where substituted into the Calvert formula there were statistically significant reductions in the dose calculated (CG formula: -8% mean difference, P<0.05; Jelliffe formula: -14% mean difference, P<0.05). This study demonstrated that the dose of carboplatin calculated to achieve a specific target AUC was very dependant on the formula applied and that the Chatelut formula had a statistically significant gender bias and should not be used in practice. Two oncology research groups have developed formulae to predict GFR specifically in cancer patients as a means of assisting the bedside dosing of cytotoxic chemotherapy. Despite only being validated in a small number of patients, these GFR prediction methods were adopted into practice. The third paper evaluated the predictive ability of the Wright formula and the Martin formula when compared to Tc99mDTPA clearance. This analysis was performed with the patient cohort from the two earlier papers. The mean measured GFR was 87mLlmin (range: 30-174mLlmin). The mean precision (mean absolute percentage error) was 18% and 19% for the Wright and Martin formulae respectively. The Martin formula under-estimated GFR for females (mean percentage error: -9%) and over¬estimated GFR for males (mean percentage error: +8%). The key findings were that the estimations of renal function from these two formulae were imprecise and do not provide an improvement on other existing bedside estimates of GFR. The practice of rounding up low serum creatinine (SCr) concentrations to 60 IJmollL in the estimation of renal function for carboplatin dosing is often practiced and also stipulated in clinical trials. This round up method is utilised on the basis that low SCr concentrations «60 IJmol/L) reflect reduced creatinine production rather than enhanced clearance. In the fourth paper an evaluation of this round up method incorporated in the CG formula was compared to measured T c99mDPT A clearance in 419 adult oncology patients. This research demonstrated that rounding low SCr concentrations up to 60 IJmollL improved the predictive ability of the CG formula using actual body weight (mean percentage error 12.9% compared to -7% without rounding). The fifth paper evaluated the practice of dose rounding at the time of prescribing and dose preparation. This study evaluated the feasibility of applying a whole vial size dose rounding approach. The study assessed the impact of this practice across five common cytotoxic chemotherapy drugs. Using data from 2828 oncology patients receiving chemotherapy (mean BSA: 1.80m2; range: 1.00 -2.56m2) this approach would not have resulted in a clinically significant difference between the mean calculated dose and the rounded dose for docetaxel, gemcitabine, oxaliplatin and vinorelbine. This study provided support for adoption of the practice for dose rounding for certain drugs when multiple vial size presentations are available. Clinical practice guidelines have been published to assist in the management of cancer patients with organ dysfunction, including recommendations for cytotoxic chemotherapy dosing in renal impairment. Clinicians are now routinely provided with estimated GFR (eGFR) through automated pathology reports in many countries calculated from the 4¬variable Modification of Diet in Renal Disease (4v-MDRD) formula. The sixth paper evaluated doses calculated for renally excreted cytotoxic chemotherapy drugs using available formulae in comparison to doses when using measured Tc99mDPTA clearance. This included the 4v-MDRD and the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula which has been recommended to replace the other formulae in routine practice. The concordance of the renal function estimates according to CKD classification with measured Tc99mDPTA clearance in 455 adults (median age: 66.0 years; range 17-87 years) for the 4v-MDRD, CKD-EPI, CG, Martin and Wright formula was 47.7%,56.3%,46.2%,56.5% and 60.2% respectively. Concordance for chemotherapy dose based on Cancer Institute New South Wales recommendations for these formulae was 89.0%, 89.5%, 85.1%, 89.9% and 89.9% respectively. The key findings of this study was that all the bedside formulae provided similar levels of concordance in dosage selection for the renal excreted chemotherapy drugs when compared to the use of a direct measure of GFR. Clinicians using any of these formulae for dosing selection must be cognisant of the potential inherent inaccuracies in dosing that may result. The findings from the series of papers presented as a thesis by publication provide evidence to guide dosing practices for cytotoxic chemotherapy. Further improvements in dosing approaches must be explored as alternatives to existing practices, which are primarily historically based.


Campus location


Principal supervisor

Carl Kirkpatrick

Year of Award


Department, School or Centre

Centre for Medicine Use and Safety (CMUS)


Doctor of Philosophy

Degree Type



Faculty of Pharmacy and Pharmaceutical Sciences