Colloids in Cardiac Surgery-Friend or Foe?

OBJECTIVE: The right choice of fluid replacement still is a matter of debate. Recently, two large-scale studies on the use of hydroxyethyl starches (HES) in the intensive care setting have been published, which have caused a huge shift in the daily practice of volume therapy. These results have been applied to patients outside intensive care. The aim of this study was to evaluate the impact this change has had on the outcomes in a large population of cardiac surgery patients, with a focus on the type of colloid infusion.

DESIGN: A prospective, registered, observational study, using propensity score matching.

SETTING: Cohort study from 3 university hospitals using a common registry.

PARTICIPANTS: The study comprised 17,742 patients who were referred for cardiac surgery from 2007 to 2014.

INTERVENTIONS: Patients were divided in groups according to perioperative fluid replacement with either crystalloids or colloids. The colloid group was further divided into HES or human albumin (HA). Analyses were based on the following 3 subsections: HES versus crystalloids, HA versus crystalloids, and HES versus HA, with use of propensity score matching or direct matching of cases. Primary outcome parameters were 30-day and 6-month mortality, new postoperative renal replacement therapy, and new cardiac ischemic events.

MEASUREMENTS AND MAIN RESULTS: The groups were fully comparable in individual analyses. The use of HES had no impact on new dialysis and 30-day mortality. A Cox proportional regression analysis showed that HES had no impact on 6-month mortality and new postoperative ischemic events. When comparing HA with crystalloid use, a significantly increased risk in crude analysis was demonstrated on all outcome parameters; and when comparing HA with HES, a significantly higher risk was observed in HA patients in mortality parameters and new postoperative, but after adjustment, only the risk of new postoperative dialysis persisted.

CONCLUSION: This study underlined the difficulties in establishing hardcore outcome data, even in large cohort studies. The findings seemed to diminish the magnitude of risk when using HES in cardiac surgery patients and seriously questioned the choice of HA when a plasma expander is needed.