Fig. 2

Cut-off values of study outcomes at admission based on the kidney function and change compared to the previous year. In both receiver operating characteristic (ROC) curves X-axis shows 1-specificity and Y-axis sensitivity, while cuf-off and area under curve (AUC) values are also noted. In every figure (a–d), which represents different study outcomes, the first plot shows the eGFR cuf-off value (ml/min/1.73 m²) at admission. The second plot represents the change in kidney function (expressed in percentage), when the laboratory parameter of kidney function at admission was compared to the same individual’s eGFR from the last year (365 + 1 days). Negative sign means decrease. (a) In-hospital mortality. If the patient’s eGFR at admission was less than 25.96 was more likely to die. When the same person’s actual kidney function decreased with 18.15% compared to its’ healthy status from the last year, the risk of lethal outcome was significantly higher (Fig. 3). (b) Discharge home. Above 36.64 eGFR value, the outcome was discharge home, rather than in-hospital mortality, transfer to other ward or leave against medical advice. The actual kidney function compared to a previous laboratory result from the previous year showed that 23.65% loss is associated with discharge home (Fig. 3). (c) Need for RBC transfusion. Patients with worse eGFR than 49.61 was more likely to receive RBC transfusion. 4.3% deterioration in kidney function at admission compared to the results from the previous increased the need for RBC transfusion (Fig. 3).