Heavy Numbers: When Obesity Tips the Balance between Creatinine- and Cystatin C-Based eGFR

Obesity may distort eGFR because creatinine and cystatin C are influenced by non-GFR determinants. We quantified discordance between creatinine- and cystatin C-based eGFR across BMI strata and assessed its impact on KDIGO G-stage classification. We hypothesized that higher BMI would be associated with greater discordance and more frequent KDIGO G-stage reclassification.

In this observational cohort of 2207 patients with paired routine-care measurements at a German university hospital, non-indexed eGFR was calculated using the 2009 CKD Epidemiology Collaboration (CKD-EPI) creatinine equation and the 2012 CKD-EPI cystatin C and combined creatinine-cystatin C equations. We compared BMI ≤25 kg/m2 (n=776) with BMI ≥35 kg/m2 (n=309). Clinically relevant discordance thresholds and KDIGO G-stage reclassification were assessed, and multivariable models examined predictors of discordance. Agreement between eGFR equations was additionally assessed using Bland-Altman analyses, overall and stratified by BMI.

Among 2207 participants, those with BMI ≥35 kg/m2 had significantly lower eGFRcys and eGFRcr-cys than those with BMI ≤25 kg/m2, whereas eGFRcr differed only borderline. The discrepancy between eGFRcys and eGFRcr was greater in participants with BMI ≥35 kg/m2, and clinically relevant discordance occurred more frequently. Bland-Altman analysis showed wide limits of agreement and a more negative bias in participants with BMI ≥35 kg/m2. Switching from eGFRcr to eGFRcys resulted in KDIGO G-stage up-classification in 43% versus 26%, despite similar albuminuria and proteinuria. Using eGFRcr-cys as reference, eGFRcr caused more frequent down-classification in participants with BMI ≥35 kg/m2, whereas eGFRcys caused more frequent up-classification. Independent predictors of discordance were higher BMI, CRP, prednisolone use, and current smoking, while high urea was inversely associated.

Participants with BMI ≥35 kg/m2 had frequent, clinically relevant divergence between creatinine- and cystatin C-based eGFR, leading to substantial KDIGO G-stage reclassification. Because measured GFR was unavailable, the study demonstrates disagreement between equations but cannot determine which estimate is more accurate. Combined-marker equations or measured GFR should be considered when discordance could change clinical decisions.