Background: Fluid therapy is widely used in critically ill patients to restore effective intravascular volume and improve organ perfusion. Recent studies have questioned the administration of colloid-based solutions, especially if containing hydroxyethyl starch (HES), in different ICU populations; however, there is still uncertainty on the use of colloids as initial fluid therapy for early resuscitation.
Methods: The aim of this study was to investigate the effect of two different resuscitation fluid strategies on the mortality of patients with shock. In a multicentric (57 ICUs), controlled, open-label trial (from February 2003 to August 2012), the authors randomized patients with signs of acute hypovolemia, defined by the combination of hypotension, evidence of low filling pressures or cardiac index and at least two signs of tissue hypoperfusion (such as altered consciousness, mottled skin, oliguria, lactate levels > 2 mmol/L), to received either a colloid- or crystalloid-based therapy. Both cohorts received maintenance fluids consisting in isotonic crystalloids and albumin in case of severe hypoalbuminemia (<2 g/dL). Exclusion criteria included previous fluid therapy, pregnancy, brain death, extended burns, chronic hemodialysis or liver disease, known coagulopathy, acute anaphylaxis, dehydration and hypotension due to sedative drugs.
Results: Among the 6498 eligible patients, 2857 were eventually randomized in one of the two groups. The 28-day mortality was 25.4% in the colloid and 27.0% in the crystalloid group (P=0.26). In the sub-group analysis, similar mortality rates were reported for shock due to hypovolemia, sepsis or trauma. Also, the use of continuous renal replacement therapy was similar between groups (11.0% vs. 12.5%, P=0.19). There were more days alive without mechanical ventilation or vasopressors during the first 7 and 28 days and a lower 90-day mortality in the colloid group.
Conclusion: These data suggest that mortality was not increased and probably decreased with the use of colloids in different forms of shock requiring early fluid resuscitation. We discussed herein some methodological issues that may explain the discrepancies of this trial with the other studies developed in the same field.