Fluid Challenge During Anesthesia: A Systematic Review and Meta-analysis

Abstract

Background: Assessing the volemic status of patients undergoing surgery is part of the routine management for the anesthesiologist. This assessment is commonly performed by means of dynamic indexes based on the cardiopulmonary interaction during mechanical ventilation (if available) or by administering a fluid challenge (FC). The FC is used during surgery to optimize predefined hemodynamic targets, the so-called Goal-Directed Therapy (GDT), or to correct hemodynamic instability (non-GDT).

Methods: In this systematic review, we considered the FC components in studies adopting either GDT or non-GDT, to assess whether differences exist between the 2 approaches. In addition, we performed a meta-analysis to ascertain the effectiveness of dynamic indexes pulse pressure variation (PPV) and stroke volume (SV) variation (SVV), in predicting fluid responsiveness.

Results: Thirty-five non-GDT and 33 GDT studies met inclusion criteria, including 5017 patients. In the vast majority of non-GDT and GDT studies, the FC consisted in the administration of colloids (85.7% and 90.9%, respectively). In 29 non-GDT studies, the colloid infused was the 6% hydroxyethyl starch (6% HES; 96.6% of this subgroup). In 20 GDT studies, the colloid infused was the 6% HES (66.7% of this subgroup), while in 5 studies was a gelatin (16.7% of this subgroup), in 3 studies an unspecified colloid (10.0% of this subgroup), and in 1 study albumin (3.3%) or, in another study, both HES 6% and gelatin (3.3%). In non-GDT studies, the median volume infused was 500 mL; the time of infusion and hemodynamic target to assess fluid responsiveness lacked standardization. In GDT studies, FC usually consisted in the administration of 250 mL of colloids (48.8%) in 10 minutes (45.4%) targeting an SV increase >10% (57.5%). Only in 60.6% of GDT studies, a safety limit was adopted. PPV pooled area under the curve (95% confidence interval [CI]) was 0.86 (0.80-0.92). The mean (standard deviation) PPV threshold predicting fluid responsiveness was 10.5% (3.2) (range, 8%-15%), while the pooled (95% CI) sensitivity and specificity were 0.80 (0.74-0.85) and 0.83 (0.73-0.91), respectively. SVV pooled area under the curve (95% CI) was 0.87 (0.81-0.93). The mean (standard deviation) SVV threshold predicting fluid responsiveness was 11.3% (3.1) (range, 7.5%-15.5%), while the pooled (95% CI) sensitivity and specificity were 0.82 (0.75-0.89) and 0.77 (0.71-0.82), respectively.

Conclusions: The key components of FC including type of fluid (colloids, often 6% HES), volume (500 and 250 mL in non-GDT studies and GDT studies, respectively), and time of infusion (10 minutes) are quite standardized in operating room. However, pooled sensitivity and specificity of both PPV and SVV are limited.