Abstract Background and Aims Evidence on the risks of adverse outcomes associated with glomerular filtration rate (GFR) has been limited to estimated GFR (eGFR), due to limited availability of large-scale data with measured GFR (mGFR). These studies either used creatinine (eGFRcr), cystatin C (eGFRcys), or both (eGFRcr-cys), which can be influenced by the non-GFR determinants of creatinine and cystatin C. Here, we aimed to (1) quantify associations between mGFR and the risk of all-cause mortality and kidney failure with replacement therapy (KFRT) and (2) investigate which filtration marker most closely resembled these associations. Method We included 6059 adults referred for single-point plasma iohexol clearance between 2011-2021 in Stockholm, Sweden (SCREAM project), and who had simultaneous measures of creatinine and cystatin C. eGFR was calculated with the CKD-EPI 2012 (eGFRcys) and 2021 (eGFRcr and eGFRcr-cys) equations. Hazard ratios were estimated using Cox regression, adjusting for age, sex, medical history, medications and urinary albumin-to-creatinine ratio (UACR). Results Mean age was 59 years, 40% were women, median mGFR was 66 ml/min/1.73 m2 and median UACR was 1.1 mg/mmol. During median 3.8 years of follow-up, 1961 individuals died and 425 progressed to KFRT. Associations between mGFR and mortality increased linearly for values <90 ml/min/1.73 m2 and were not statistically significant between 90-120 ml/min/1.73 m2 (Figure). Compared with mGFR of 90 ml/min/1.73 m2, hazard ratios for mortality were 1.5 (95% CI 1.3-1.6), 1.7 (1.5-1.8), 2.0 (1.8-2.1) and 2.4 (2.0-2.7) for 60, 45, 30 and 15 ml/min/1.732, respectively (Table). eGFRcr-cys most closely resembled the mortality risk relationships of mGFR. Conversely, eGFRcr had a strong U-shaped pattern with underestimation of risks <90 ml/min/1.73 m2 and overestimation of risks between 90-120 ml/min/1.73 m2; for eGFRcys this pattern was reversed. Associations between mGFR and KFRT were also linear (Figure). Compared with mGFR of 60 ml/min/1.73 m2, hazard ratios for KFRT were 2.1 (1.4-3.0), 6.2 (4.5-8.5), and 32.1 (23.4-44.0) for 45, 30 and 15 ml/min/1.732, respectively (Table). Both mGFR and estimating equations displayed similar monotonic risk associations. Conclusion In the largest study on mGFR and outcomes to date, lower mGFR was linearly associated with an increase in the risks of mortality and KFRT. eGFRcr-cys closely resembled the relationship between mGFR and mortality, whereas eGFRcr under-estimated mortality risk and had a U-shaped association. All estimating equations had similar associations with KFRT risk.
Read full abstract