FC 053RATIO OF MEASURED GFR TO ESTIMATED GFR MAY PREDICT EARLY DEATH AND REQUIREMENT FOR DIALYSIS

Abstract

Abstract Background and Aims It has been suggested that, in patients with CKD stage 5, measured GFR (mGFR), defined as the mean of urea and creatinine clearance as measured by a 24-hour urine collection, is a better measure of renal function than estimated GFR (eGFR), based on the CKD-EPI formula. This could be due to reduced muscle mass in this group. Its use is recommended in the ERBP guidelines. Unplanned dialysis initiation (UDI) is associated with increased morbidity, mortality, and reduced modality choice. It is generally considered undesirable. We hypothesized that the ratio mGFR/eGFR (R) aids prediction of death and end stage kidney disease (ESKD), as defined by permanent dialysis requirement or transplantation. Method All 24-hour measurements of urea and creatinine excretion were extracted from the clinical biochemistry databases in Zealand. Data concerning renal diagnosis, comorbidity, biochemistry, medical treatment, mortality and date of ESKD, were extracted from patient notes, the National Patient Registry and the Danish Nephrology Registry. Patients were included if their eGFR was <30 ml/min/1.73m2. The last available value for each patient was included. Results 1265 patients were included. In 519, body surface area (BSA) was available, and the corrected ratio (RBSA) could be calculated. The urea clearance was 49 ±24% of creatinine clearance. R was median 0.88 (IQR 0.63-1.15), RBSA 0.87 (0.68-1.06). R was not related to eGFR. Comorbidity was associated with lower R, e.g. atherosclerosis (0.90 ±0.41 vs. 0.97 ±0.49*), heart failure (0.80 ±0.37 vs. 0.95 ±0.44*), pulmonary disease (0.80 ±0.37 vs. 0.94 ±0.44***), hepatic disease (0.67 ±0.41 vs. 0.92 ±0.43***), but not diabetes mellitus. It was related to albumin (r=0.24***), C-reactive protein (-0.22***) and biochemical markers of uraemia, e.g. bicarbonate (-0.19***). Medical treatment data was available in 137 patients. R was higher in patients treated with ACE inhibitors (1.20 ±0.50 vs. 1.01 ±0.36*) and diuretics (1.09 ±0.40 vs. 0.94 ±0.35*), but no other treatment groups. Patients were grouped as high R (H, >1.25), medium (M, 0.75-1.25) or low (L, <0.75). R was not associated with prognosis at one year, but L patients had a significantly higher ESKD and mortality incidence at 3 months. For patients with eGFR 10-15 ml/min/1.73m2, ESKD incidence was L 22%, M 15%, H 5%, mortality 19, 5, and 2% respectively. Similar findings were seen in other groups, e.g. eGFR 15-20 ml/min/1.73m2: ESKD 11, 2, and 0%; death 11, 5, and 1%. UDI was higher for L patients. For patients with eGFR 10-15 ml/min/1.73m2, UDI occurred in L: 47%, M:27%, H:25%. For patients with eGFR 10-15 ml/min/1.73m2 the figures were 51, 38 and 12% respectively. Findings for the subgroup of patients with RBSA measurements were similar. *:p<0.05; **:p<001;***:p<0.001 Conclusion A low mGFR/eGFR ratio is associated with comorbidity, malnutrition, inflammation and biochemical uraemia. It a marker of early ESKD, death and unplanned dialysis initiation, independently of eGFR. Particular attention paid to patients with a low R may lower the incidence of unplanned dialysis requirement.