In patients with septic shock, the administration of fluids during initial hemodynamic resuscitation remains a major therapeutic challenge. We are faced with many open questions regarding the type, dose and timing of intravenous fluid administration. There are only four major indications for intravenous fluid administration: aside from resuscitation, intravenous fluids have many other uses including maintenance and replacement of total body water and electrolytes, as carriers for medications and for parenteral nutrition. In this paradigm-shifting review, we discuss different fluid management strategies including early adequate goal-directed fluid management, late conservative fluid management and late goal-directed fluid removal. In addition, we expand on the concept of the “four D’s” of fluid therapy, namely drug, dosing, duration and de-escalation. During the treatment of patients with septic shock, four phases of fluid therapy should be considered in order to provide answers to four basic questions. These four phases are the resuscitation phase, the optimization phase, the stabilization phase and the evacuation phase. The four questions are “When to start intravenous fluids?”, “When to stop intravenous fluids?”, “When to start de-resuscitation or active fluid removal?” and finally “When to stop de-resuscitation?” In analogy to the way we handle antibiotics in critically ill patients, it is time for fluid stewardship.

The clinical determination of the intravascular volume can be extremely difficult in critically ill and injured patients as well as those undergoing major surgery. This is problematic because fluid loading is considered the first step in the resuscitation of hemodynamically unstable patients. Yet, multiple studies have demonstrated that only approximately 50% of hemodynamically unstable patients in the intensive care unit and operating room respond to a fluid challenge. Whereas under-resuscitation results in inadequate organ perfusion, accumulating data suggest that over-resuscitation increases the morbidity and mortality of critically ill patients. Cardiac filling pressures, including the central venous pressure and pulmonary artery occlusion pressure, have been traditionally used to guide fluid management. However, studies performed during the past 30 years have demonstrated that cardiac filling pressures are unable to predict fluid responsiveness. During the past decade, a number of dynamic tests of volume responsiveness have been reported. These tests dynamically monitor the change in stroke volume after a maneuver that increases or decreases venous return (preload) and challenges the patients' Frank-Starling curve. These dynamic tests use the change in stroke volume during mechanical ventilation or after a passive leg raising maneuver to assess fluid responsiveness. The stroke volume is measured continuously and in real-time by minimally invasive or noninvasive technologies, including Doppler methods, pulse contour analysis, and bioreactance.

The effects of COVID-19 are highly variable, ranging from individuals who are asymptomatic to patients who develop severe acute respiratory distress syndrome, with potential involvement of almost all organs and systems.¹ These acute symptoms have been well described since the first cohort studies that were published at the beginning of the pandemic.²

In patients with acute circulatory failure, the decision to give fluids or not should not be taken lightly. The risk of overzealous fluid administration has been clearly established. Moreover, volume expansion does not always increase cardiac output as one expects. Thus, after the very initial phase and/or if fluid losses are not obvious, predicting fluid responsiveness should be the first step of fluid strategy. For this purpose, the central venous pressure as well as other “static” markers of preload has been used for decades, but they are not reliable. Robust evidence suggests that this traditional use should be abandoned. Over the last 15 years, a number of dynamic tests have been developed. These tests are based on the principle of inducing short-term changes in cardiac preload, using heart–lung interactions, the passive leg raise or by the infusion of small volumes of fluid, and to observe the resulting effect on cardiac output. Pulse pressure and stroke volume variations were first developed, but they are reliable only under strict conditions. The variations in vena caval diameters share many limitations of pulse pressure variations. The passive leg-raising test is now supported by solid evidence and is more frequently used. More recently, the end-expiratory occlusion test has been described, which is easily performed in ventilated patients. Unlike the traditional fluid challenge, these dynamic tests do not lead to fluid overload. The dynamic tests are complementary, and clinicians should choose between them based on the status of the patient and the cardiac output monitoring technique. Several methods and tests are currently available to identify preload responsiveness. All have some limitations, but they are frequently complementary. Along with elements indicating the risk of fluid administration, they should help clinicians to take the decision to administer fluids or not in a reasoned way.

We performed a systematic review and meta-analysis of studies investigating the passive leg raising (PLR)-induced changes in cardiac output (CO) and in arterial pulse pressure (PP) as predictors of fluid responsiveness in adults. MEDLINE, EMBASE and Cochrane Database were screened for relevant original and review articles. The meta-analysis determined the pooled area under the ROC curve, the sensitivity, specificity and threshold for the PLR test when assessed with CO and PP. Twenty-one studies (991 adult patients, 995 fluid challenges) were included. CO was measured by echocardiography in six studies, calibrated pulse contour analysis in six studies, bioreactance in four studies, oesophageal Doppler in three studies, transpulmonary thermodilution or pulmonary artery catheter in one study and suprasternal Doppler in one study. The pooled correlation between the PLR-induced and the fluid-induced changes in CO was 0.76 (0.73–0.80). For the PLR-induced changes in CO, the pooled sensitivity was 0.85 (0.81–0.88) and the pooled specificity was 0.91 (0.88–0.93). The area under the ROC curve was 0.95 ± 0.01. The best threshold was a PLR-induced increase in CO ≥10 ± 2 %. For the PLR-induced changes in PP (8 studies, 432 fluid challenges), the pooled sensitivity was 0.56 (0.49–0.53), the pooled specificity was 0.83 (0.77–0.88) and the pooled area under the ROC curve was 0.77 ± 0.05. Sensitivity and subgroup analysis were consistent with the primary analysis. PLR-induced changes in CO very reliably predict the response of CO to volume expansion in adults with acute circulatory failure. When PLR effects are assessed by changes in PP, the specificity of the PLR test remains acceptable but its sensitivity is poor.