Resuscitation Fluids in Septic Shock: A Network Meta-Analysis of Randomized Controlled Trials

The aim of this study was to assess the efficacy and safety of various resuscitation fluids in septic shock by adopting a network meta-analysis (NMA). Randomized controlled trials (RCTs) comparing resuscitation fluids in septic shock were carried out by retrieving electronic databases. NMAs of 28-day mortality, 90-day mortality, incidence of acute kidney injury (AKI), and the need for renal replacement therapy (RRT) were conducted using the STATA 15.0 software. Probability-based ranking and surface under cumulative ranking (SUCRA) were performed to identify the optimal resuscitation fluid. Inconsistencies were evaluated by node-splitting analysis and a loop-specific approach. Furthermore, publication bias was analyzed by funnel plots. A total of 13 RCTs were enrolled in the analysis. The NMA results revealed that no significant differences were detected in the outcomes of 28-day mortality and 90-day mortality among various resuscitation fluids. The SUCRAs (the first indicates the best) of 28-day mortality showed that the hypertonic sodium chloride/hydroxyethyl starch 40 solution ranked the highest (93.8%), followed by the balanced solution (BS) (69.6%), and albumin (61.9%). On the other hand, the SUCRAs of 90-day mortality revealed that gelatin (GEL) ranked the highest (75.1%), followed by BS (55.1%), and NS (52.4%). The NMA results of AKI demonstrated that high-molecular-weight hydroxyethyl starch (H-HES) was associated with increased risk of AKI in comparison with GEL, BS, and L-HES. The SUCRAs of AKI showed that GEL ranked the highest (74.4%), followed by NS (64.9%), and BS (58.3%). In addition, the NMA results of RRT revealed that H-HES was associated with an increased need for RRT in comparison with BS and NS, and L-HES was associated with increased need of RRT in comparison with BS. The SUCRAs of RRT revealed that NS ranked the highest (91.6%), followed by BS (74.4%) and L-HES (36.1%). No significant inconsistencies were shown by the node-splitting analysis and no publication bias was demonstrated in the funnel plots. In conclusion, BS was determined as the preferred resuscitation fluid for septic shock. Moreover, the use of GEL requires further evaluation. H-HES was associated with a significant risk of AKI and RRT, whereas L-HES with an increased need for RRT compared with BS. Thus, both resuscitation fluids should be avoided for septic shock.