Glomerular filtration rate measurement and estimation : improvement and validation of existing methods

Holness, Jennifer Lyn (2019-12)

Thesis (PhD)--Stellenbosch University, 2019.

Thesis

ENGLISH SUMMARY : Glomerular filtration rate (GFR) is regarded as the best measure of kidney function. It can either be measured or estimated. This dissertation aims to provide a better understanding of GFR measurement in order to improve its performance and interpretation. It also aims to validate GFR estimation in local populations and to demonstrate the utility of simple adaptations of existing equations to improve estimation. On completion of a GFR measurement, various quality control (QC) checks are performed to ensure the accuracy of the result. However, this requires comparison with clearly defined reference ranges. In a study of healthy, potential kidney donors, reference data for two QC parameters were defined. In a study analysing the effect of measurement errors on GFR, the single-sample method was found to be the most robust technique overall, although for all methods measurement error was generally insignificant compared to expected biological variation in GFR. However, at low GFR values measurement errors were shown to affect all methods significantly. Errors in measurement of the doses were found to have the greatest impact on accuracy. Using nuclear medicine techniques 51Cr-ethylenediaminetetra-acetic acid (51Cr-EDTA) is the most commonly used and widely studied exogenous filtration marker. However, 99mTc-diethylenetriaminepenta-acetic acid (99mTc-DTPA) is gaining favour because of a few technical advantages it has over 51Cr-EDTA, its lower cost, and recent issues with the availability of 51Cr-EDTA. In response to a systematic review suggesting that GFR measurement from the plasma clearance of 99mTc-DTPA was inaccurate, a mini meta-analysis was performed that demonstrated excellent agreement between 51Cr-EDTA and 99mTc-DTPA clearance, thus supporting the use of 99mTc-DTPA as a reliable alternative. Where GFR cannot be routinely measured, it is frequently estimated using a creatinine-based equation. The use of any equation first requires validation in the population in which it will be used. In a study evaluating the Modification of Diet in Renal Disease (MDRD) and Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equations in non-cancer, mixed ancestry adults, both equations were found to perform well. However, in a study that evaluated equations in adults with cancer, the GFR estimates were found to be biased and imprecise. This study highlighted the limitations of using estimated GFR for guiding management decisions in cancer patients. A further study evaluated 11 estimating equations in non-cancer and cancer populations of South African children. The accuracy of all estimates was poor, particularly in the cancer group. Given the extensive use of GFR estimates in South Africa, these findings have profound implications for their use in the management of children and adults with cancer in this country. Developing new equations for a specific population requires large datasets and incurs costs that are impractical in most middle- or low-income countries. A simpler alternative is to adapt existing equations. This work demonstrates the application of a relatively simple approach to adapt existing equations, using modest amounts of data and a readily available Microsoft® Excel-based tool. While this approach is simple and likely to require further refinement, its utility was demonstrated in paediatric and adult cancer populations.

AFRIKAANSE OPSOMMING : Glomerulêre filtrasiespoed (GFS) word as die beste maatstaf van nierfunksie beskou. Dit kan gemeet of geskat word. Hierdie proefskrif se doel is om ‘n beter begrip van GFS meting te bewerkstellig teneinde die metode se verrigting en interpretasie te verbeter. Verder word gepoog om GFS skatting in plaaslike populasies te valideer en om die nut van eenvoudige aanpassings van bestaande vergelykings om skatting te verbeter, te demonstreer. Na voltooide GFS meting word verskeie gehaltetoetse gedoen om die akkuraatheid van die resultaat te verseker. Dit verg egter vergelyking met duidelik gedefinieerde verwysingswaardes. Verwysingsdata vir twee gehalteparameters is in ’n studie van gesonde moontlike nierskenkers daargestel. In ‘n studie wat die effek van metingsfoute op GFS ontleed, is bevind dat die enkel-monster metode algeheel die mees robuste tegniek is, alhoewel metingsfout vir alle metodes oor die algemeen vergeleke met verwagte biologiese variasie in GFS nie betekenisvol was nie. By lae GFS waardes is egter aangetoon dat metingsfoute alle metodes betekenisvol beïnvloed. Daar is getoon dat foute in meting van die dosis die grootste impak op die akkuraatheid het. 51Cr-etileendiamientetra-asynsuur (51Cr-EDTA) is die mees algemeen gebruikte en bes nagevorste eksogeene filtrasiemerker in kerngeneeskundige tegnieke om GFS te bepaal. 99mTc-di-etileentriamienpenta-asynsuur (99mTc-DTPA) wen egter veld weens ‘n aantal tegniese voordele wat dit bo 51Cr-EDTA het, soos laer koste en meer onlangse probleme met die beskikbaarheid van 51Cr-EDTA. In antwoord op ‘n sistematiese oorsig wat daarop dui dat GFS meting gebaseer op plasma opruiming van 99mTc-DTPA onakkuraat sou wees, is ‘n mini meta-analise gedoen wat uitstekende ooreenkoms tussen 51Cr-EDTA en 99mTc-DTPA opruiming getoon het, en sodoende die gebruik van 99mTc-DTPA as betroubare alternatief ondersteun. Waar GFS nie roetinegewys gemeet kan word nie, word dit dikwels geskat met behulp van ‘n kreatinien-gebaseerde vergelyking. Die gebruik van enige vergelyking vereis eers validasie in die populasie waarin dit gebruik gaan word. ‘n Studie waarin die sogenaamde Modification of Diet in Renal Disease (MDRD) en Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) vergelykings in kankervrye volwssenes van gemengde afkoms evalueer is, het bevind dat beide vergelykings goed doen. ‘n Ondersoek wat die vergelykings in volwassenes met kanker evalueer het, is egter bevind date GFS skattings sydig en nie presies was nie. Hierdie studie het die beperkings van die gebruik van GFS skattings om oor kliniese hantering van kankerpasiënte te besluit, uitgelig. ‘n Verdere ondersoek het 11 skattings-vergelykings in populasies van Suid-Afrikaanse kinders met en sonder kanker evalueer. Die akkuraatheid van alle skattings was swak, veral in die groep met kanker. In die lig van die wye gebruik van GFS skattings in Suid-Afrika, het hierdie bevindings verregaande implikasies vir hul gebruik in die hantering van kinders en volwassenes met kanker in hierdie land. Die ontwikkeling van nuwe vergelykings vir spesifieke populasies vereis groot datastelle en koste wat in die meeste middle- en lae-inkomste lande onprakties is. Aanpassing van bestaande vergelykings is ‘n eenvoudiger alternatief. Hierdie werk toon die toepassing van ‘n redelik eenvoudige benadering om bestaande vergelykings aan te pas met behulp van matige hoeveelhede data en geredelik beskikbare Microsoft® Excel-gebaseerde nutsprogrammatuur. Alhoewel die benadering eenvoudig is en waarskynlik verdere verfyning gaan vereis, is die nut daarvan in volwasse kanker- en pediatriese populasies getoon.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/107242
This item appears in the following collections: