Heise et al developed a model for predicting if a patient would develop adequate renal function during an interval between continuous renal replacement therapy (CRRT). The authors are from the University Hospital Gottingen in Germany.


Patient selection: patient in the ICU with renal failure managed by CRRT


Restoration of adequate renal function was assumed if the patient was discharged from the ICU and no further renal replacement was needed.



(1) SOFA score at the end of the last CRRT session

(2) number of CRRT cycles

(3) urine output after cessation of CRRT in mL per hour


X =

= 1.695 - (0.174 * (SOFA score)) – (0.802 * (number of CRRT cycles)) + (0.026 * (urine output))


probability of restoration of adequate renal function =

= 1 / (1 + EXP((-1) * X))



• A value of X = 0.684 had a sensitivity of 74% and specificity of 74% for predicting recovery of adequate renal function.


To read more or access our algorithms and calculators, please log in or register.