A scoping review of renal function calculation methods for the dosing of carboplatin

BackgroundCarboplatin is widely used to treat solid and haematological malignancies. The Calvert formula (area under the curve (AUC)×(glomerular filtration rate (GFR) + 25) is used globally to calculate the dose. Due to the large number of methods available to calculate GFR it is possible that disparity exists for patients regarding their carboplatin dosing depending on the policy of the organisation treating them. The purpose of this scoping review is to determine the breadth and accuracy of GFR calculation methods used in the literature.MethodsMedline, Web of Science, Embase and CINAHL databases were searched in November 2024. Eligibility criteria included randomised controlled trials, case studies, case series and cohort studies that reported renal function calculation methods for the dosing of carboplatin. Accepted studies were written in English and conducted in adult or paediatric humans from 2000-2024. A combination of MeSH terms and keywords were used. Duplicates were removed. One author reviewed the articles by title and abstract and then by full text. The other authors provided a validation of 20% of selections at the title and abstract review stage and then at the full text stage. Data were extracted into an Excel spreadsheet and included the main author, year, study design, sample size, patient age, cancer type, GFR method and any relevant findings such as accuracy values.Results454 articles were identified, reduced to 102 after removing duplicates and exclusions applied. There were 41 different GFR methods reported. The most common were Cockcroft and Gault (C&G) using actual body weight (ABW), Jelliffe, CKD-EPIcreatinine, 24-h urine collection, Wright, MDRD and radionuclide techniques with either 99mTcDTPA or 51Cr EDTA. Whilst all papers reported the use of GFR methods only 26 reported comparisons to determine accuracy. Most papers used a reference GFR such as a radionuclide technique for comparison. A variety of comparison methods were used to measure accuracy such as mean (MPE), median (MdPE) and percentage errors (PE), bias and mean bias, coefficients and root-mean-squared error. C&G using ABW and Jelliffe generally overestimated renal function and had a wide dispersion of PE readings (-10-30.1 MPE and -9.5-9.55 MdPE respectively) suggesting low accuracy to the reference GFR. CKD-EPIcreatinine adjusted for body surface area (BSA) reported low MPE values (-2-0) suggesting good accuracy. This article reported an improvement when CKD-EPIcreatinine was adjusted for BSA whereas another reporting bias values showed the reverse. Wright and MDRD reported wide dispersion of MPE readings showing both over and underestimation compared with the reference GFR. While 14 articles mentioned 24-h collection, only two compared the accuracy to a reference. These reported MPE values of 35.9 and -25 showing large over and underestimation.ConclusionsThis review highlights a variety of methods used to calculate GFR for carboplatin dosing. Further studies are required utilising the same reference GFR, and accuracy comparison methods to determine the most accurate GFR calculation method to determine carboplatin dosing. Standardisation of this method will ensure patients receive accurate dosing regardless of treating organisation location.