Abstract
BACKGROUND: Patients receiving chronic kidney replacement therapy (cKRT) are at high risk of admission to the intensive care unit (ICU), where their management remains challenging. Short- and long-term outcomes of cKRT patients admitted to ICU remain elusive precluding optimization of ICU admission policies and dedicated follow-up after discharge. METHODS: In this retrospective study of 216 cKRT patients admitted to a tertiary ICU in France (2013-2021), multivariable logistic regression and Cox's proportional hazard models were employed to identify predictive factors of death in ICU, at one year, and at long-term. RESULTS: The leading cause of admission were septic shock (36.1%). The mortality rate in ICU was 14.5% and was best predicted by cardiac arrest as the cause of admission, the SAPS2, the need of mechanical ventilation, and the use of a tunneled catheter for dialysis access, while 1-year survival was predicted by age, RBC transfusion and the SAPS-2 score. Median survival in ICU survivors was 49 months. In survivors, long-term mortality was predicted by the number of daily medications before the admission (HR 1.089 (CI(95%) 1.027; 1.155)), the use of a tunneled catheter (HR 1.67 (CI(95%) 1.06; 2.7)), and age at admission (HR 1.036 (CI(95%) 1.017; 1.057)). During the 6 months following discharge from the ICU, 125 patients (69%) were re-admitted to the hospital. Finally, twenty-one patients received a kidney transplant. CONCLUSION: The prediction of early death in cKRT patients admitted to ICU largely depends on the severity of the acute condition at admission, whereas a multimodal risk stratification including surrogate markers of frailty gives a better indication of long-term outcomes.