Continuous Central Venous Oxygen Saturation Measurement as a Tool to Predict Hemodynamic Instability Related to Renal Replacement Therapy in Critically Ill Patients

Overview

Acute kidney injury (AKI) is common in intensive care unit (ICU) and inducing a high morbidity and mortality. In severe forms of AKI (more than 25% of AKI patients admitted to the ICU), renal replacement therapy (RRT) is often necessary. Although RRT is a cornerstone of therapy, it can lead to serious adverse effects, such as intradialytic arterial hypotension. Indeed, arterial hypotension during the session - intra-dialytic arterial hypotension (IDH) - occurs frequently complication and so regardless of the RRT modality used. Its occurrence may worsen significantly the outcome as previously reported. It is therefore of parmount importance to prevent such an adverse effect. The investigators hypothesize that a decrease in the central venous oxygen saturation (SvcO2) measured related to a decreased cardiac output could precede the onset of IDH. The aim of this study is collect IDH in AKI patients and to measure continuously SvcO2 during RRT session in order to investigate its role in predicting IDH.

In ICU settings, about 25% of AKI patients underwent RRT. RRT leads to certain complications, the most frequent and severe of which is IDH. The frequency of IDH varies from 10 to 70% and from 19 to 43% in intermittent and continuous RRT respectively. The occurrence of IDH worsens the prognosis of critically ill patients. It may impair the recovery of renal function and is associated with increased mortality. Thus, early detection of IDH requires the development of preventive strategies and the implementation of adapted curative therapies. IDH results from several mechanisms. The main mechanisms involved is hypovolemia leading to cardiac dysfunction, but also cardiac dysfunction not related to hypovolemia. The identification of tools correlated with cardiac output could allow early prediction of IDH. Yet, no tool has been evaluated to predict the occurrence of IDH. Central venous oxygen saturation (SvcO2) is a reflection of the balance between systemic O2 transport and tissue O2 consumption and thus is the indicator of cardiac output. A decrease in cardiac output leads to a decrease in SvcO2 to enhance O2 transport and to maintain arterial pressure. Assuming that O2 extraction is constant, SvcO2 would then be a surrogate for cardiac output. A decrease in SvcO2 may reflect a decrease in cardiac output before IDH occurs. There are very few data in the literature evaluating changes in cardiac output during an RRT session. Continuous measurement of SvcO2, a minimally invasive and indirect marker of cardiac output, during an RRT session has never been reported to predict IDH. The investigators hypothesize that a decrease in SvcO2, indicative of decreased cardiac output, should precede the onset of IDH. In order to evaluate this hypothesis, the investigators intend to measure continuous SvcO2 and continuous cardiac output during the first three RRT sessions, regardless of the RRT modality. The investigators will assess the performance of SvcO2 values and other invasive hemodynamic parameters in predicting the occurrence of IDH using ROC curves. Once the composite score has been constructed, an optimal threshold will be determined using Youden's J statistic. Determination of the 95% confidence interval (95% CI), sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). Comparison of areas under the curve using Delong's non-parametric approach. Patients will be monitored until they are discharged from intensive care, intensive care, death or D28 max.