Intra-individual variability of eGFR trajectories in early diabetic kidney disease and lack of performance of prognostic biomarkers

Abstract

Studies reporting on biomarkers aiming to predict adverse renal outcomes in patients with type 2 diabetes and kidney disease (DKD) conventionally define a surrogate endpoint either as a percentage of decrease of eGFR (e.g. ≥ 30%) or an absolute decline (e.g. ≥ 5 ml/min/year). The application of those study results in clinical practise however relies on the assumption of a linear and intra-individually stable progression of DKD. We studied 860 patients of the PROVALID study and 178 of an independent population with a relatively preserved eGFR at baseline and at least 5 years of follow up. Individuals with a detrimental prognosis were identified using various thresholds of a percentage or absolute decline of eGFR after each year of follow up. Next, we determined how many of the patients met the same criteria at other points in time. Interindividual eGFR decline was highly variable but in addition intra-individual eGFR trajectories also were frequently non-linear. For example, of all subjects reaching an endpoint defined as a decrease of eGFR by ≥ 30% between baseline and 3 years of follow up, only 60.3 and 45.2% lost at least the same amount between baseline and year 4 or 5. The results were similar when only patients on stable medication or subpopulations based on baseline eGFR or albuminuria status were analyzed or an eGFR decline of ≥ 5 ml/min/1.73m2/year was used. Identification of reliable biomarkers predicting adverse prognosis is a strong clinical need given the large interindividual variability of DKD progression. However, it is conceptually challenging in early DKD because of non-linear intra-individual eGFR trajectories. As a result, the performance of a prognostic biomarker may be accurate after a specific time of follow-up in a single population only.

Development of a machine learning-based prediction model for extremely rapid decline in estimated glomerular filtration rate in patients with chronic kidney disease: a retrospective cohort study using a large data set from a hospital in Japan

“…However, patterns of eGFR decline, which are called trajectories of eGFR, vary between patients. 30 A report from six large-scale, randomised controlled trials revealed that more cases presented with non-linear eGFR decline among diabetes patients. 31 Many reports have demonstrated risk factors related to the decline in eGFR.…”

A prediction model for the decline in renal function in people with type 2 diabetes mellitus: study protocol

“…However, this approach suffers from several shortcomings. First, individual eGFR slopes can only be estimated very inaccurately, as only very few data points at non-equidistant time points are available for each patient [5]. Moreover, the categorization of the resulting eGFR slopes adds another layer of possible bias as it leads to subjects close on the continuous spectrum but on the opposite sides of the cutpoint being characterized as different [31].…”

“…4 Department of Nephrology, University of Regensburg, University Hospital Regensburg, Regensburg, Germany. 5 Department of Nephrology and Medical Intensive Care, Charité Universitätsmedizin Berlin, Berlin, Germany. 6 Department of Internal Medicine IV (Nephrology and Hypertension), Medical University Innsbruck, Innsbruck, Austria.…”

“…Personalized risk assessment based on large study cohorts could therefore offer several potential benefits to the preventive care for people with DM2 such as early detection followed by monitoring to guide treatment and potentially slow the progression of the decline of kidney function as expressed by the estimated glomerular filtration rate (eGFR) [4]. Despite the accumulated knowledge with regard to CKD nowadays, identifying individuals at high risk for fast disease progression has proven to be challenging because longitudinal eGFR slopes not only vary between patients (inter-patient variability), but eGFR measurements also fluctuate within each patient over time (intra-patient variability) [5].…”

Untitled

“…[1] In our study though eGFR was not significantly different in the two groups; with and without DN, serum creatinine levels were a risk in both sexes and eGFR negatively correlated with UACR in males as shown in Table-4. As eGFR is closely related to age, baseline eGFR at the time of diagnosis of T2DM and further monitoring to see the rate of decline [16] is imperative for initiating treatment with antidiabetic agents which will protect the kidneys also. [12] In our study, the serum lipids were not statistically different in the two groups, yet studies have shown that HDL and TG are independent risk factors for cardiovascular disease, systolic blood pressure (SBP) being the measure.…”