Fig. 2.
Intraoperative variables and their assumed causal relationships with the exposure and outcome. Potential confounding variables (green) are those that might influence both hydroxyethyl starch administration and acute kidney injury (AKI): estimated blood loss, intraoperative hypotension, urine output, duration and category of surgery, and the anesthesiologist. Potential mediator variables, such as intraoperative vasopressor use and blood product transfusion (red), might be external variables that are mechanisms by which hydroxyethyl starch administration might cause increased risk of AKI but do not influence the decision to administer hydroxyethyl starch.