Biomarkers of Kidney Function; Validity of Estimates in Clinical Practice

Abstract

Chronic kidney disease (CKD) and acute kidney injury (AKI) are two common conditions that are encountered by clinicians, yet unlike the cardiac troponins, there is currently no gold-standard biomarker that can accurately detect a reduction in glomerular filtration rate (GFR) without considerable delay, or the amount of damage that has occurred in the kidneys. Measurement of radioisotopes such as Tc-diethylenepentaacetic acid (Tc-DTPA) provides an accurate assessment of GFR, although, it is not without complexity, and therefore reserved for select populations who do not require routine monitoring. The drivers for this thesis were guided by two significant events which materialised in recent times, firstly, the standardisation of cystatin C measurements, and secondly, the publication of the CKD-EPI equations incorporating standardised cystatin C. The primary goal of this thesis was to investigate the utility of prediction equations based on standardised creatinine and standardised cystatin C in quantifying GFR, in routine clinical practice. The performances of contemporary estimated GFR (eGFR) equations comprising the CKD-EPI equations versus more traditional eGFR equations, such as the MDRD and Cockcroft-Gault, were compared with respect to gold-reference measured GFR. The CKD-EPI equation employing the combination cystatin C and creatinine, was found to be more accurate than the CKD-EPI creatinine-only, the CKD-EPI cystatin C-only, the MDRD and the Cockcroft-Gault equations in predicting renal function in select populations. These included; kidney donors, adult cancer patients, and in patients receiving gentamicin for sepsis. There was a slight improvement with the usage of contemporary CKD-EPI equations in assigning patients to correct CKD staging. However there was no obvious improvement in eGFR accuracy when plasma creatinine was measured by an enzymatic method over the more traditional Jáffe reaction. Novel observations in the body composition and renal function (BodyComp) study showed cystatin C to be associated with higher body fat and creatinine with higher muscularity. The strongest determinants of measured GFR in healthy men included lean mass and age, and in those with increased GFR, protein intake was not found to be a significant predictor. The CKD-EPI cystatin C equation was able to predict Tc-DTPA GFR with greater accuracy than equations incorporating creatinine in healthy men with extremes of body composition. Estimates of GFR are pertinent in drug dosing given that personalised treatment of individuals may require certain drugs to be dosed according to renal function. The removal of BSA normalisation in eGFR improved the performances of the contemporary CKD-EPI equations in predicting raw-measured GFR, especially at the extremes of size. The CKD-EPI cystatin C equation was found to obtain higher accuracy in terms of carboplatin dosing. Paediatric eGFR equations examined in children receiving chemotherapy, were found to significantly underestimate GFR. Notably, cystatin C concentrations were not shown to confer better accuracy than plasma creatinine in determining kidney function. The observation of increases in cystatin C independently from renal function suggests cystatin C maybe be affected by potential non-renal factors in this cohort. In relation to acutely unwell patients, the hypothesis of decreased creatinine production was explored. By integrating different cut-offs for changes in creatinine to include a 20% increase, an unchanging, or a 10% decrease, this thesis showed that individuals with increased and stable creatinine were highly associated with mortality rates in a cardiac arrest cohort. In those with unchanging creatinine, renal function can be compromised, and this suggests that current definitions of changes in plasma creatinine is not adequate in detecting AKI. In acutely unwell patients, the inclusion of cystatin C in new kinetic eGFR equations were shown to be more accurate in predicting GFR, but only in those with reduced renal function, and when there is unchanging levels of plasma creatinine. This thesis have thus examined creatinine, cystatin C and renal function equations as tools for assessing kidney function. The contemporary equations however still significantly underestimated GFR in those with normal kidney function and in obesity. In conclusion, the greatest accuracy in estimation of kidney function was achieved when cystatin C was used in conjunction with creatinine and its utility may benefit in situations such as (i) confirming stage 3a CKD, (ii) to guide drug dosing after accounting for individual BSA (if reference measurement of GFR is unavailable), and (iii) for monitoring progress in critically unwell patients. The cystatin C only equation was only more accurate than the combination equation in the context of extremes of body composition in healthy men.