Knowing whether or not a fluid infusion can improve cardiac output (fluid responsiveness) is crucial when treating hemodynamically unstable patients. Generally, cardiac filling pressures (central venous pressure, pulmonary artery occlusion [''wedge''] pressure) and volumes (end-diastolic left and right ventricular volume) are used, although they are not reliable predictors of fluid responsiveness. For this reason, new indices, the so-called dynamic indices of fluid responsiveness, have been recently introduced in clinical use. If stroke volume, or stroke volume-derived parameters (pulse pressure and aortic flow) show wide variation during mechanical ventilation, a good response to fluid therapy can be predicted. As these indices are based upon the effects of controlled mechanical ventilation on stroke volume, they can be used in deeply sedated or apneic patients whose cardiac rhythm is regular. To overcome these limitations, new dynamic indices have been introduced. Among them, variation of cardiac output induced by passive leg raising (PLR) has raised particular interest since it can identify fluid responders even among spontaneously breathing and non-sinus rhythm patients. Although promising, the dynamic indices of fluid responsiveness have been studied only retrospectively in a relatively small number of patients and evidence that clinical use of these indices can improve outcome is still limited. Further investigations are needed to confirm their clinical validity.