Trajectories of eGFR decline over a four year period in an Indigenous Australian population at high risk of CKD-the eGFR follow up study

Being able to estimate kidney decline accurately is particularly important in Indigenous Australians, a population at increased risk of developing chronic kidney disease and end stage kidney disease. The aim of this analysis was to explore the trend of decline in estimated glomerular filtration rate (eGFR) over a four year period using multiple local creatinine measures, compared with estimates derived using centrally-measured enzymatic creatinine and with estimates derived using only two local measures.

Method: The eGFR study comprised a cohort of over 600 Aboriginal Australian participants recruited from over twenty sites in urban, regional and remote Australia across five strata of health, diabetes and kidney function. Trajectories of eGFR were explored on 385 participants with at least three local creatinine records using graphical methods that compared the linear trends fitted using linear mixed models with non-linear trends fitted using fractional polynomial equations. Temporal changes of local creatinine were also characterized using group-based modelling. Analyses were stratified by eGFR (< 60; 60–89; 90–119 and ≥ 120 ml/min/1.73 m²) and albuminuria categories (< 3 mg/mmol; 3–30 mg/mmol; > 30 mg/mmol).

Results: Mean age of the participants was 48 years, 64% were female and the median follow-up was 3 years. Decline of eGFR was accurately estimated using simple linear regression models and locally measured creatinine was as good as centrally measured creatinine at predicting kidney decline in people with an eGFR < 60 and an eGFR 60–90 ml/min/1.73 m² with albuminuria. Analyses showed that one baseline and one follow-up locally measured creatinine may be sufficient to estimate short term (up to four years) kidney function decline. The greatest yearly decline was estimated in those with eGFR 60–90 and macro-albuminuria: −6.21 (−8.20, −4.23) ml/min/1.73 m².

Conclusion: Short term estimates of kidney function decline can be reliably derived using an easy to implement and simple to interpret linear mixed effect model. Locally measured creatinine did not differ to centrally measured creatinine, thus is an accurate cost-efficient and timely means to monitoring kidney function progression.