The purpose of this study is to determine if testing preload responsiveness, the normal physiologic state that means that changes in preload determine changes in cardiac output, allows an earlier and physiologically safer weaning from mechanical ventilation in critically ill patients with fluid overload, when compared to a strategy of fluid removal aimed at obtaining a predetermined negative fluid balance.

Fluid overload is a state of global body accumulation of fluids with a deleterious impact in organ function. This condition is frequently found in critically ill patients after acute resuscitation. Its adverse impact is well demonstrated on weaning-induced heart failure, pulmonary and visceral edema, intraabdominal hypertension, etc., which results in longer mechanical ventilation and length of stay, and worse clinical outcomes. Despite these well-known facts, there are no guidelines on how to implement depletive strategies on this phase. The usual approach is to set in advance a desired negative fluid balance for the upcoming days, initiating diuretics or ultrafiltration in preparation for weaning from mechanical ventilation. Unfortunately, this strategy frequently results in excessive and detrimental fluid removal.

A more physiologic approach to guide fluid removal is testing preload responsiveness, which is the normal physiologic state, and means that changes in preload determine changes in cardiac output, with mild or null increment in filling pressures. In contrast, preload unresponsiveness corresponds to a state in which preload increases do not increase stroke volume but produce large increments in filling pressures. This altered state is usually present in patients with fluid overload. Preload responsiveness can be tested routinely in the ICU by assessing the interactions between preload and cardiac output.

Now, in usual clinical practice, weaning from mechanical ventilation is accomplished through a process called the spontaneous breathing trial (SBT), which is a standardized test to mimic the real conditions of breathing without the ventilator, before extubation. One-third of patients fail the initial SBT, which determines a prolonged or difficult weaning and longer stay on mechanical ventilation. Importantly, one of the main determinants of this problem is fluid overload. The pathophysiologic explanation lays in that when switching from positive pressure ventilation to spontaneous breathing, intrathoracic pressure goes from being steadily positive across the ventilatory cycle to markedly negative, promoting increased preload and impeding left ventricular ejection, and this phenomenon is associated to preload unresponsiveness. Interestingly, in most patients with fluid overload, preload responsiveness can be restored just a few hours after starting fluid removal, while modifying fluid balance may take several days. Notably, some patients may persist with preload unresponsiveness even after achieving significant fluid removal.

The investigators hypothesized that in mechanically ventilated patients with fluid overload, a fluid removal strategy aimed at attaining a state of preload responsiveness associates with a decreased incidence of weaning failure from cardiovascular origin, shorter weaning time, and less depletion-induced hypoperfusion events, metabolic derangements and kidney stress compared to patients depleted with a fluid removal strategy aimed at obtaining a predetermined negative fluid balance.

To confirm this hypothesis, the investigators propose a prospective randomized study on 46 critically ill mechanically ventilated patients with fluid overload, comparing these two strategies of depletion and their impact on weaning development and other related systemic functions. Throughout all the protocol, patients will receive general monitoring and management according to ICU standards, plus protocol-specific monitoring that will be added since randomization and before and after SBT attempts, for up to 72 h. Patients will be followed-up for 28 days.

If the investigators' hypothesis is confirmed, it may generate a change in the paradigm of managing fluid overload in critically ill patients, since the physiologic endpoint preload responsiveness may suffice as the valid target and safety parameter to appropriately discontinue mechanical ventilation, shortening the days on mechanical ventilation, the ICU length of stay, and many other costs associated, among additional benefits.

Inclusion Criteria:

1. Mechanical ventilation for >24 hours and <7 days
2. Patients who on the daily evaluation to assess their potential for weaning present fluid overload and, in consequence, require fluid depletion before the spontaneous breathing trial
3. Clinical condition resolving or hemodynamically stable condition with acceptable ventilatory status that allows an spontaneous breathing trial according to attending's criteria.

Exclusion Criteria:

1. Pregnancy
2. Do-not-resuscitate status
3. Child B or C liver cirrhosis
4. Circulatory instability
5. Acute coronary syndrome
6. Active bleeding
7. Severe concomitant acute respiratory distress syndrome
8. Malnutrition
9. Muscle weakness severe enough to be considered by itself a risk for weaning failure
10. Patient should be excluded based on the opinion of the clinician/investigator (documented reason)