Initial assessment on the impact of crystalloids versus colloids during damage control resuscitation

High ratios of fresh frozen plasma:packed red blood cells in damage control resuscitation (DCR) are associated with increased survival. The impact of volume and type of resuscitative fluid used during high ratio transfusion has not been analyzed. We hypothesize a difference in outcomes based on the type and quantity of resuscitative fluid used in patients that received high ratio DCR.

A matched case control study of patients who received transfusions of ≥ four units of PRBC during damage control surgery over 4 1/2 y, was conducted at a Level I Trauma Center. All patients received a high ratio DCR, >1:2 of fresh frozen plasma:packed red blood cells. Demographics and outcomes of the type and quantity of resuscitative fluids used in combination with high ratio DCR were compared and analyzed. A Kaplan-Meier survival analysis was computed among four groups: colloid (median quantity = 1.0 L), <3 L crystalloid, 3-6 L crystalloid, and >6 L crystalloid.

There were 56 patients included in the analysis (28 in the crystalloid group and 28 in the colloid group). Demographics were statistically similar. Intraoperative median units of PRBC: crystalloid versus colloid groups was 13 (IQR 8-21) versus 16 (IQR 12-19), P = 0.135; median units of FFP: 12 (IQR 7-18) versus 12 (IQR 10-18), P = 0.440. OR for 10-d mortality in the crystalloid group was 8.41 [95% CI 1.65-42.76 (P = 0.01)]. Kaplan-Meier survival analysis demonstrated lowest mortality in the colloid group and higher mortality with increasing amounts of crystalloid (P = 0.029).

During high ratio DCR, resuscitation with higher volumes of crystalloids was associated with an overall decreased survival, whereas low volumes of colloid use were associated with increased survival. In order to improve outcomes without diluting the survival benefit of hemostatic resuscitation, guidelines should focus on effective low volume resuscitation when high ratio DCR is used. A multi-institutional analysis is needed in order to validate these results.