500P - The Leeds formula: a new, more accurate and widely applicable formula for estimating renal function, accounting for variability in creatinine assay...

Date 01 October 2012
Event ESMO Congress 2012
Session Poster presentation III
Topics Anti-Cancer Agents & Biologic Therapy
Presenter Fiona Collinson
Authors F. Collinson1, W. Gregory1, C. Twelves2, C. Handforth3, M. Bosomworth4, G. Hall3
  • 1Clinical Trials Research Unit, University of Leeds, LS2 9JT - Leeds/UK
  • 2Clinical Cancer Research Groups, Leeds Institute of Molecular Medicine & St James's Institute of Oncology, Leeds/UK
  • 3St James's Institute Of Oncology, St James's University Hospital, LS9 7TF - Leeds/UK
  • 4Blood Sciences, St James's University Hospital, LS9 7TF - Leeds/UK



Many formulae are used to estimate renal function, but none account for the significant variation in creatinine (Cr) measurement between the 26 assays currently used in the UK. The UK National External Quality Assessment Service has published assay-specific Cr adjustors, but the calculated 'standardised Cr' (SCr) dangerously overestimates GFR when used in formulae such as Cockroft and Gault (C&G) and Wright (W). We aimed to develop a simple formula to accurately estimate GFR in oncology patients, using easily available patient characteristics and the SCr, hence accounting for inter-assay variation.


Isotopic GFR (iGFR) was measured using Tc99mDTPA clearance. Serum Cr was measured using the O'Leary Jaffe assay and from this SCr derived. Clinical parameters (age, sex, height, weight, SCr, urea and albumin) were used in regression modelling (STATA) to investigate the relationship of individual parameters with iGFR. From this a novel formula was derived to estimate iGFR (the Leeds formula). This formula was then prospectively validated on a second cohort of patients.


In the discovery set 423 oncology patients were included with a range of malignancies, a median age of 49 (range 18-91) years and serum Cr between 50-130 µmol/l who underwent iGFR measurement between 1/4/06 and 31/3/09. A model incorporating SCr, age, sex, height, weight, urea and albumin predicted iGFR (median calculated GFR 92 ml/ min, range 25-217; r2 of 0.74) more accurately than alternative established GFR formulae e.g. C&G and W formulae (r2 0.4-0.6) . Prospective validation on a separate cohort of oncology patients (496 patients between 1/4/09 and 31/3/11) validated the Leeds formula with an r2 of 0.71 for correlation with iGFR.


The Leeds formula uses readily available clinical information to estimate GFR more precisely than existing formulae and has the advantage of being applicable to Cr results from any laboratory provided the assay type is known. Many oncologists do not appreciate the resultant effect of inter-Cr assay variability on estimated renal function (GFR) and hence chemotherapy dosing.


All authors have declared no conflicts of interest.