<p>Fluids are given with the purpose of increasing cardiac output (CO), but approximately only 50% of critically ill patients are fluid responders. Since the effect of a fluid bolus is time-sensitive, it diminuish within few hours, following the initial fluid resuscitation. Several functional hemodynamic tests (FHTs), consisting of maneuvers affecting heart-lung interactions, have been conceived to discriminate fluid responders from non-responders. Three main variables affect the reliability of FHTs in predicting fluid responsiveness: (1) tidal volume; (2) spontaneous breathing activity; (3) cardiac arrythmias. Most FTHs have been validated in sedated or even paralyzed ICU patients, since, historically, controlled mechanical ventilation with high tidal volumes was the preferred mode of ventilatory support. The transition to contemporary methods of invasive mechanical ventilation with spontaneous breathing activity impacts heart-lung interactions by modifying intrathoracic pressure, tidal volumes and transvascular pressure in lung capillaries. These alterations and the heterogeneity in respiratory mechanics (that is present both in healthy and injured lungs) subsequently influence venous return and cardiac output. Cardiac arrythmias are frequently present in critically ill patients, especially atrial fibrillation, and intuitively impact on FHTs. This is due to the random CO fluctuations. Finally, the presence of continuous CO monitoring in ICU patients is not standard and the assessment of fluid responsiveness with surrogate methods is clinically useful, but also challenging. In this review we provide an algorithm for the use of FHTs in different subgroups of ICU patients, according to ventilatory setting, cardiac rhythm and the availability of continuous hemodynamic monitoring.</p>