Fluid therapy is often used as first line treatment of acute circulatory failure, aiming an increase in cardiac output (by improving preload) and in tissue perfusion. Depending on left ventricular systolic function, fluid challenge could lead to an increase in cardiac ouput and tissue perfusion, or only detrimental consequences (by fluid overload and aggravation of lung and tissues oedema, increase of morbi-mortality). Patients are defined as responders to fluid therapy if one can observe an increase of cardiac output up to 15% after fluid therapy (500ml of crystalloids): gold standard test used in most of the studies on the subject. Literature reports on heterogenous populations a reproductible and constant response rate to this fluid challenge of 50%. It seems reasonable to dispose of indices allowing to predict fluid responsiveness without resulting in fluid intake. Statics markers have been abandonned for several years and dynamics methods have been developped.

In front of arrythmia, validated methods are scarce. Passive leg rising method appears to be the only one and it's validity seems to be less well documented than in sinusal patients. The purpose of this study is to determine a new method to assess fluid responsiveness in arrythmic patients.

In atrial fibrillation, RR interval varies widely between cardiac cylces. Systolic interval remain constant. Variations will occure at expense of diastolic interval, or ventricular filling interval. One can reliably assume that when RR is longer, preload is rising. If the patient is on the ascendant part of the Franck-Starling curve, a longer RR should cause au greater VTI (Vitess Time Integral, surrogate of cardiac output).

The evaluation by transthoracic echocardiography of the indice delta ITV / delta RR should determine the degree of fluid responsiveness in arrhythmic patients.

After decision of fluid expansion, patients will have haemodynamic and echocardiographic data measured, delta ITV / delta RR indice assessed, then passive leg rising and fluid expansion with 500 ml of cristalloids administered, with evaluation of VTI (as surrogate of cardiac output) at each time. Fluid responders will be compared to non-responders to evaluate the diagnostic performances of the indice delta ITV / delta RR.

Inclusion Criteria:

• Hospitalized patients in the units of Surgical ICU and Cardiothoracic ICU of the Universatory Hospital of Poitiers
• With atrial Fibrillation
• With spontaneous ventilation
• With acute circulatory failure and medical decision to fluid therapy (Systolic Arterial Blood Pressure < 90 mmHg and/or vasopressors agents)
• With no left or right ventricular dysfunction

Exclusion Criteria:

• Patients < 18 yrs
• Contraindictions to passive leg rising
• Clinical evidence of hemorrhagic shock
• Cardiogenic shock
• Necessity of fluid expansion in emergency
• No echographic windows