Association between Initial Fluid Choice and Subsequent In-hospital Mortality during the Resuscitation of Adults with Septic Shock

Abstract

BACKGROUND: Currently, guidelines recommend initial resuscitation with intravenous (IV) crystalloids during severe sepsis/septic shock. Albumin is suggested as an alternative. However, fluid mixtures are often used in practice, and it is unclear whether the specific mixture of IV fluids used impacts outcomes. The objective of this study is to test the hypothesis that the specific mixture of IV fluids used during initial resuscitation, in severe sepsis, is associated with important in-hospital outcomes. METHODS: Retrospective cohort study includes patients with severe sepsis who were resuscitated with at least 2 l of crystalloids and vasopressors by hospital day 2, patients who had not undergone any major surgical procedures, and patients who had a hospital length of stay (LOS) of at least 2 days. Inverse probability weighting, propensity score matching, and hierarchical regression methods were used for risk adjustment. Patients were grouped into four exposure categories: recipients of isotonic saline alone ('Sal' exclusively), saline in combination with balanced crystalloids ('Sal + Bal'), saline in combination with colloids ('Sal + Col'), or saline in combination with balanced crystalloids and colloids ('Sal + Bal + Col'). In-hospital mortality was the primary outcome, and hospital LOS and costs per day (among survivors) were secondary outcomes. RESULTS: In risk-adjusted Inverse Probability Weighting analyses including 60,734 adults admitted to 360 intensive care units across the United States between January 2006 and December 2010, in-hospital mortality was intermediate in the Sal group (20.2%), lower in the Sal + Bal group (17.7%, P < 0.001), higher in the Sal + Col group (24.2%, P < 0.001), and similar in the Sal + Bal + Col group (19.2%, P = 0.401). In pairwise propensity score-matched comparisons, the administration of balanced crystalloids by hospital day 2 was consistently associated with lower mortality, whether colloids were used (relative risk, 0.84; 95% CI, 0.76 to 0.92) or not (relative risk, 0.79; 95% CI, 0.70 to 0.89). The association between colloid use and in-hospital mortality was inconsistent, and survival was not uniformly affected, whereas LOS and costs per day were uniformly increased. Results were robust in sensitivity analyses. CONCLUSIONS: During the initial resuscitation of adults with severe sepsis/septic shock, the types of IV fluids used may impact in-hospital mortality. When compared with the administration of isotonic saline exclusively during resuscitation, the coadministration of balanced crystalloids is associated with lower in-hospital mortality and no difference in LOS or costs per day. When colloids are coadministered, LOS and costs per day are increased without improved survival. A large randomized controlled trial evaluating crystalloid choice is warranted. Meanwhile, the use of balanced crystalloids seems reasonable